> RATAN TATA LIBRARY DELHI UNIVERSITY LIBRARY SYSTEM eae BDO Oe DEH Oe On ea > ae ee o> | : D.U.P. No. 1337—1-81—-20,000 RATAN TATA LIBRARY (Delhi University Library System) ClLNo os. ff. oY i Ac. No. . Date of release for loan This book should be returned on or before the date last stamped below. An overdue charge of Ten Paise will be charged for each day the book is kept overtime. yy . S W. Lutosfawski - Plato’s Logic Wincenty Lutostawski The Origin and Growth of Plato’s Logic With an Account of Plato’s Style and of the Chronology of his Writings 1983 Georg Olms Verlag Hildesheim : Ziirich - New York O Dem Nachdruck liegt das Exemplar der Universitatsbibliothek Erlangen zugrunde. Signatur: Phi. VII, 3700! Nachdruck der Ausgabe London, New York und Bombay 1897 Mit freundlicher Genehmigung der Longmans Green & Co. Ltd. Printed in Germany Herstellung: Strauss & Cramer GmbH, 6945 Hirschberg 2 ISBN 3 487 07336 6 PLATO'S LOGIC THE ORIGIN AND GROWTH OF PLATO'S LOGIC WITH AN ACCOUNT OF PLATO’S STYLE AND OF THE CHRONOLOGY OF HIS WRITINGS BY WINCENTY LUTOSEAWSKI LONGMANS, GREEN, AND CO. 89 PATERNOSTER ROW, LONDON NEW YORK AND BOMBAY 1897 Adi rights reserved TO LEWIS CAMPBELL ON THE THIRTIETH ANNIVERSARY OF HIS ‘INTRODUCTION TO THE SOPHISTES AND POLITICUS’ THIS WORK IS DEDICATED IN GRATEFUL REMEMBRANCE OF TRUE FRIENDSHIP AND KINDNESS PREFACE ——1O4———— In undertaking the investigations summarised in this volume, the author's chief aim was to explain the origin of Logic by a psychological study of the first logician. This required a knowledge of the chronology of Plato’s writings, not supplied by our historical tradition nor by the extant Platonic investigations. English and French scholars mostly believed this problem to be insoluble ; the prevalent opinion in Germany, represented by the suc- cessive editions of Zeller’s and Ueberweg’s handbooks on Greek philosophy, was plainly wrong. Under these cir- cumstances there was need of a new method in order to attain a greater certainty as to the order in which Plato wrote his dialogues. The method here proposed improves the stylistic tests used heretofore by formulating the theoretical principles on which a new science of Stylo- metry should be based (pp. 145-161) and by applying these principles (pp. 162-193) to five hundred peculiari- ties of Plato’s style (observed in fifty-eight thousand cases) collected in the course of fifty years by some twenty authors working independently (pp. 74-139). ‘This stylo- metric method, supplemented by many comparisons of the contents of Plato’s works (for instance, pp. 329, 333, 366, 368, 372, 396, 430, 452, &c.), and by such observa- tions and suggestions as were found available in the Vlll ORIGIN AND GROWTH OF PLATO’S LOGIC Platonic literature of all countries, led the author to determine the chronological order of about twenty among the most important of the Platonic dialogues. On this basis an account of Plato’s logical theories and of their development is given here for the first time. It is ascertained that the theory of ideas, generally believed to be the unique form of Plato’s logic, was only a first attempt of the philosopher to settle the difficulties of the relation between Knowledge and Being ; and that, when past fifty, he produced a new logical system, in which he anticipated some conceptions of modern philosophy, arriving at the recognition of the substantial existence of the individual soul and substituting a classification of human notions for the intuition of divine ideas. This being a work of research, not a general hand- book, the reader need not expect a digest of literature. The authors chiefly quoted are those who were the first to make an important observation, or who have expressed more amply the author’s own views on some subject briefly treated here, or whose remarkable want of judgment makes them instructive as examples to avoid. A full indication of the bibliography on any special question has nowhere been attempted except in Chapter III on Plato’s style. However, it has been sought to demonstrate the merits of some writers as yet insufficiently appreciated (for instance, pp. 83, 112, 352). As a Pole, the author may possibly be more impartial than the representatives of other nations more active in Platonic research. The works of British scholars are little known in Germany, and, on the other hand, many special German investigations are overlooked in France and Great Britain. Here the results obtained through unconscious international collaboration have been summed up and presented in a general outline, PREFACE 1x though without bibliographical completeness. The absence of alphabetical indices in the majority of works on Plato makes 1t hard to remember by whom a given observation was first made. These historical debts have been acknowledged in many instances, and wherever such an acknowledgment is missing, this should be attributed to defective memory. The peculiar method of research used in the present work is a result of the author’s previous study of natural sciences and mathematics (1881-1885), and he feels much indebted to his teachers at the late German University of Dorpat'!: Carl Schmidt, Arthur von Oettingen, Johannes Lemberg, Gustav Bunge, Wilhelm Ostwald, Andreas Lindstedt, and Jan Baudouin de Courtenay, all of whom in their lectures and also in private intercourse with their pupils insisted on exactness of method in scientific in- vestigation. His interest in Plato the author owes to Gustav Teichmiiller,? from whom however he now differs somewhat in his views on the method of Platonic research and on Plato’s philosophy (pp. 57-59, 102-103). 1 To acknowledge this debt of gratitude is all the more a duty, as since the change of this German seat of learning into the Russian University of Jurjew all its most eminent professors have been obliged to resign, and Dorpat University is now but a historical reminiscence, dear to all its ancient pupils. * Under Teichmiiller’s influence the author wrote ten years ago his first work on Plato: Hrhaliung und Untergang der Staatsverfassungen, nach Plato, Aristoteles und Machiavelli, Dorpat 1887 (Breslau 1888), wherein Plato’s views on political revolutions are shown to be the source of later theories on that subject. The chief contents of Chapter I of the present work have been more amply treated in the author’s Polish publica- tions: O Logice Platona, Part I, Krakéw 1891 and Part II, Warszawa 1892, condensed in the French Bulletin de l’ Académie des sciences de Cracovie, April 1890 and November 1891. Also Chapters V, VII and VIII rest chiefly on a Polish work of the author: O pierwssych trsech tetralogiach dziet Platona, published by the Académie des sciences de Cracovie, Cracow 1896; condensed in the same Bulletin for October, November 1895, and in the Archiv fiir Geschichte der Philosophie, vol. ix. pp. 67-114, October 1895. x ORIGIN AND GROWTH OF PLATO’S LOGIC The author feels deeply obliged to all who have helped him, and in the conviction that the collaboration of many is needed to bring full light to bear upon the difficult problems dealt with in this volume, he ventures to invite his readers also to assist him in his further studies on Plato by pointing out such errors or even formal deficiencies, however minute, as may be observed (address, care of Longmans, Green, and Co., 39 Paternoster Row, London). La CoruNa, SPAIN: October 1897. CONTENTS CHAPTER I PLATO AS A LOGICIAN (pp. 1-84) Progress of logic questioned, p. 1—Mill against Plato and Kant, 2—Univer- sality and permanence of knowledge to be tested by history of logic, 2— Plato the first logician, 3—Exceptional preservation of his works, 3—Its reasons, 4— Permanence of the Academy, 5—Protection by the Christian clergy, 6—Plato’s logic unknown, 7—Opinion of Plethon, 8—of Gen- nadios, 8—Champier, 9—Patrizi and other historians of the XVIth century, 10—Morainvillier, 11—Stanley and Gassendi, 12—Reaction against Plato in the XVIIth century, 13—Tennemann, 13—He did not attempt to represent the evolution of Plato’s logic, 14—Various opinions on Platonic ideas, 15—Van Heusde and other writers in the XIXth century, 16—Recent logical writers, 17—They were ignorant of Platonic chronology, 18—Susemihl first combined both problems, 19—Ueberweg first recognised the difference between Plato’s earlier and later logic, 20—Misunderstood by Oldenberg, 21—Confirmed by Peipers, 22—Jack- son, 23—Benn, 23—Aristotle still held by some historians to be the first logician, 23—-Many useless dissertations on Plato’s dialectic, 24—and theory of ideas, 25—Conditions of a better study of Plato’s logic, 27— Zeller objects to the representation of Plato’s logic, 28—Our aim is to learn about Plato’s logic more than he expressed himself in his works, 29—To explain his psychological evolution, 30—To know him better than he could know himself, 31—To find out how he progressed in his views, 33—and what was the last stage of his thoughts, 34 CHAPTER II AUTHENTICITY AND CHRONOLOGY OF PLATO’S WRITINGS (pp. 85-68) Order of dialogues proposed by Patrizi, like that of Serranus, of no import- ance, 85—First inquiry by Tennemann, 35—Schleiermacher agrees with Tennemann on important points, 36—He left uncertain the order of small dialogues, 87—He supposed that Plato had planned from the Xli ORIGIN AND GROWTH OF PLATO’S LOGIC beginning the whole of his literary activity, 37—Difference between early Socratic criticism and the later Platonic criticism, 37— Progress from ethics to metaphysics, from polemical to didactic tone, 38—Ast denies the authenticity of all Socratic small dialogues, 38—Socher recognises a gradual evolution of Plato, 39—but proclaims the dialectical dialogues as spurious, 39—Stallbaum in favour of a late date of the Phaedrus, 89—H. Ritter, 40-—-Hermann establishes a Socratic period from which such important works as Parmenides must be excluded, 40 —In many particulars Hermann agrees with Stallbaum and Schleier- macher, 41—All these authors are wrong as to the supposed early date of the dialectical dialogues, 42—First origin of the myth of a Megaric period, 42—Erroneous identification of the presumed date of a conversa- tion and the date of the composition of a dialogue, 43—-Based on equally wrong identification of the Platonic and the historic Socrates, 43— Residence of Plato in Megara based on no valid testimony, 43—but on an isolated opinion of an unknown and evidently ignorant witness, 43— There was no danger for Plato to remain in Athens, 44—The author of the Crito was not a coward, 44—-Cicero trustworthy as to Plato’s life, quotes Egypt as the first place whereto Plato travelled after Socrates’ death, 45—Schleiermacher speaks of Plato’s ‘ flight’ without quoting authori- ties, 45—Ast increases the duration of the supposed sojourn at Megara, 46—Influence of Euclides on Plato taken for granted by Stallbaum, 46 — This myth repeated by Ritter and Hermann, 47—Its acceptance a con- sequence of the same esthetical prejudice which reigned in the method of editing Plato’s text before the Zurich edition, 47—Ingenious hypo- thesis preferred to careful weighing of the evidence, 48—A change in the beautiful theory of ideas esthetically objectionable, 48— Every historian built on some wrong leading hypothesis, 49—-We must get rid of such prejudice and learn to measure probabilities, 49—Plato’s philosophical consistency more probable than his cowardice, 49—Hermann recognised that Hermodorus’ testimony deserved no confidence, 49—and dis- trusted it as to the date of Plato’s first journey, 50—Followers of Her- mann and Schleiermacher, 50—Suckow, in a work full of errors, first recognised an important truth : the late date of the dialectical dialogues, 51—He was followed by Munk, 52—True genetic method first applied by Susemihl, 52—who recognised the near relation between Phaedrus and Theaetetus, 53—Ueberweg the first logician who investigated the problem of Platonic chronology, 54—and gave strong reasons for the late date of the dialectical dialogues, 55—but he came to doubt the authenticity of the Parmenides, 55—In this scepticism he was followed by Schaarschmidt, 56—while Grote and Chaignet defended the authen- ticity of all the dialogues, 56—Jowett, 56—Philosophers begin after Ueberweg to investigate his problem, 57—Tocco defended the authen- ticity and late date of the Parmenides and other dialectical dialogues, 57—Teichmiiller exaggerated the polemical aspect of Plato’s works, 57— but he supported Ueberweg’s conclusions as to the late date of the dialectical dialogues, 58—-This confirmed by Peipers, who convinced Susemihl, 59—but Zeller and the editor of Ueberweg’s ‘History of Philosophy ’ maintain the old mythus of the Megaric period, 59—and CONTENTS xn are therein followed by other popular writers, 60—-New arguments in favour of the late date of the dialectical dialogues, collected by Bergk, Rohde, Christ, Siebeck, remain little known, 60—Diimmler confirms Ueberweg’s finding by new applications of Teichmiiller’s method, 61— Anarchy in Platonic literature, 61—-Not removed by the efforts of the French Académie des sciences morales, 61—which crowned a work in which the chronological problem is regarded as insoluble, 62—This is contradicted by the whole progress of these studies, 62—to which the comparison of the logical contents will add new conclusions, 63 CHAPTER III THE STYLE OF PLATO (pp. 64-193) Style as a mark of identity of a writer, 64—What Plato thought of it, 65— Modern science deals with problems beyond the reach of Plato, 65— Identification of handwriting, 66, not easier than that of style, 66— Peculiarities of vocabulary, 67—Kinds of words, 68—Their frequency, 69—Arrangement of words, 70—Other stylistic marks, 71— Stylistic investigations easy and useful, 72—A new Lexicon Platonicum and a full bibliography of Platonic literature needed, 73 REVIEW OF FORTY-FIVE PUBLICATIONS ON THE STYLE OF PLATO AND LIST OF 500 PECULIARITIES OF PLATO’S STYLE (pp. 74-189) Engelhardt, 74—Peculiarities 1-5 (anacoluthiae), 76— who publicly acknowledged that he abandoned his former opinions, expressed thirty years earlier, as to the date. of the dialectical dialogues. Susemuhl’s impar- tiality, which allowed him to make this confession, was compensated by the obstinacy of Zeller, who, in his authoritative work on Plato, in each successive edition maintained the old assumption of a Megaric period to which he referred the Sophist, Politicus, and Philebus, alleging them to have been written before the Republic. Also the editor of the later editions of Ueberweg’s History of Greek Philosophy, M. Heinze, adhered to the old error of Hermann and Schleiermacher, and continued to spread the conviction that the Sophist and Politicus were written before the Banquet. If we take into account that Ueberweg’s and Zeller’s works on Greek philosophy enjoy up to the present time the greatest * Peipers, Ontologia Platonica, Lipsiae 1888. mation of Ueber- weg’s con- clusions as to the late date of the dialectical dialogues. Also Peipers came to the same results by his study of Platonic ontology, which convinced Susemihl, but Zeller and Heinze in the last editions of their histories of philo- sophy adhere to the Megaric mythus, as well as Weygoldt and Pfleiderer. While the general reader is thus kept in ignor- ance of the latest investiga. tions, new detailed inquiries of Bergk, Rohde, Christ, Siebeck, Diimmler 60 ORIGIN AND GROWTH OF PLATO'S LOGIC popularity, there will be no exaggeration in saying that Ueberweg’s earlier conclusions, which he afterwards abandoned, although confirmed with new arguments by Tocco, Teichmiiller, and Peipers, remain almost unknown to general readers of Plato. Ina very popular work on Plato, written by Weygoldt, we still find the dialectical dialogues placed before the Republic, and the same order occurs® in the most recent work of E. Pfleiderer on Socrates and Plato. Since Susemihl’s conversion, however, many special investigations have fortified Ueberweg’s conclusion in favour of a late date for the Sophist and Polvticus. Besides such philological investigations as those of Bergk,®* Rohde,” and Christ,** who declared in favour of a very late date for the Theaetetws and consequently also for the Sophist and Politicus, we have in the last ten years a new confirmation, through an investigation by H. Siebeck,” author of a history of psychology. Siebeck started from the question whether Plato did not quote his own works, as is frequently done by Aristotle. He observed certain allusions which led him to affirm that Plato not only quotes the Republic and the % Weygoldt, Die platonische Philosophie, Leipzig 1885; E. Pfleiderer, Socrates und Plato, Tiibingen 1896. The views of this author have to be specially dealt with in connection with the date of the Republic, as he sub- ordinates the whole order of Plato’s dialogues to a distinction of some succes- Sive stages in the Republic, wherein he follows Krohn (see note 90). Pfleiderer’s conclusions as to the order of other dialogues are not very distant from Hermann’s views, with the difference that Pfleiderer against every pro- bability places the Euthydemus after the Sophist, and the Phaedo before the Symposium. % T. Bergk, Funf Abhandlungen sur Geschichte der griechischen Philo- sophie und Astronomie, Leipzig 1888; Griechische Literaturgeschichte, 4° Bd. Berlin 1887. *” Rohde, ‘ Die Abfassungszeit der platonischen Thediet’ in Jahrbticher fir Philologie und Pddagogik, vol. cxxiii. p. 821, vol. exxv. p. 80; also in Philologus, vol. xlix. p. 2, vol. 1. p. 1, vol. li. p. 474 (1890-1892). %° W. Christ, ‘ Platonische Studien,’ pp. 4538-512 in vol. xvii. of Abh. der philos. philol. Classe der Kinigl. bayer. Akad. Miinchen, 1886. * H. Siebeck, Untersuchungen sur Philosophie der Griechen, Freiburg i. B. 1888. AUTHENTICITY AND CHRONOLOGY OF HIS WORKS 61 Politicus in the Laws, but that he also in the Republic announces a later settling of matters dealt with in the Sophist and Philebus. Also Dummler,'” who continued Teichmiiller’s studies on supposed feuds between Plato and his contemporaries, added to the considerable stock of arguments in favour of a late date of the dialectical dialogues, by a special inquiry into the relations of Plato to Antisthenes, Antiphon, Aristippus, Aristophanes, Aristotle, Empedocles, Euri- pides, Gorgias, Heraclitus, Hippias, Isocrates, Polycrates, Prodicus, Protagoras, Xenophon, and others. Besides these works, which deal with a great number of writings, there are many special dissertations on each dialogue, which constitute, taken together, ample evi- dence for a definitive solution of the problem of their date. But this literature has grown so much that nobody has attempted to collect all such detailed observations and to give a clear picture of all arguments urged in favour of each hypothesis. We have here specially insisted on the date of the dialectical dialogues because of their exceptional importance for Plato’s logic, but on each other work, as the Republic, Gorgias, Phaedo, Phaedrus, hundreds of authors have expressed various opinions, generally based only on a very limited knowledge of other investigators. So long as all these separate observations are not summed up, every new writer on this subject runs the risk of repeating discoveries already made, or falling into errors easy to avoid. In these cir- cumstances a new general work on Plato’s dialogues, summarising all the separate observations made in this century, becomes indispensable to the progress of further investigations concerning Plato’s philosophy. This need has been felt by the French Académie des Sciences morales et politiques, a learned society which has awarded many considerable prizes for works on Plato, 100 Diimmler, Akademika, Giessen 1889; Chronologische Beitrage su einigen platonischen Dialogen, Basel 1890. increase the amount of evidence in favour of Ueber- weg’s opinion. The neces- sity of an impartial co-ordina- tion of all detailed investiga- tions has been ap- preciated by the French Académie des sciences morales, but the work of C. Huit awarded a prize does not corre- spond to the pur- pose, the author neglecting the pro- blem of chrono- logy. Chrono- logy of Plato’s works not an insoluble problem, as has been generally thought in France. It is at least easy to prove the late date of the dialectical dialogues. 62 ORIGIN AND GROWTH OF PLATO'S LOGIC among which those of Chaignet and Fouillée were not without value. But the last answer to the summons of this Academy, a work in two volumes written by C. Huit under the title La vie et lauvre de Platon, published in Paris 18993, falls short of the most modest critical requirements, and by no means satisfies its purpose. The author knows so little of the special literature of his subject that he repeats Schaarschmidt’s arguments against the authenticity of the most important works of Plato without being aware that these arguments have been often refuted during the last thirty years. He also ignores the steady progress in chronological investiga- tions since Ueberweg, and regards the problem of Platonic chronology as almost insoluble. Such is not the conviction which results from an impartial sutvey of what has been already done for our problem. There is a progress in the validity of conclu- sions, as well as in the method employed from Tennemann to Schleiermacher, from Schleiermacher to Hermann, from Hermann to Susemihl, and from Susemihl to Ueberweg. Besides these inquiries referring to the majority of the works of Plato, there has been real progress also in the special investigations referring to each single dialogue. All these results should be co-ordi- nated in a general subject index showing all arguments in favour of and against every hypothesis as to the date of each several dialogue. Then only it would be inevitably seen that there is overwhelming evidence in favour of some conclusions and against others. It is not the purpose of the present work to furnish the reader of Plato with such an extensive index, but chiefly to indicate the agreement of the chief arguments advanced in favour of a late date of the dialectical dia- logues, in order to show that the logical science founded y Plato was advanced during his own lifetime by his renewed efforts. Before we enter upon the task of tracing this logical AUTHENTICITY AND CHRONOLOGY OF HIS WORKS 63 development through Plato’s works, it is a duty to give the reader some information about a special kind of in- vestigation, subsidiary to the general study of Platonic chronology. We have limited the above review to those authors who sought to establish the order of Plato’s dia- logues on arguments taken from their contents ; because it is our own purpose likewise to compare the contents of a series of Platonic dialogues as to their logical theories. But, admitting all the importance of the contents, we must still contend that the form and style of Plato’s writings also give some indications as to their chronolo- gical order, and it is useful to compare the conclusions arrived at by both methods. The study of the style of Plato is much more recent than that of Plato’s philosophy, but it has led to very important conclusions as to the order of his writings, and it is our duty to consider these conclusions before we venture to represent the origin and growth of Plato’s logic. For this, not only the contents should be compared, but also the style, which has been in- vestigated only in recent times. 64 ORIGIN AND GROWTH OF PLATO’S LOGIC CHAPTER III THE STYLE OF PLATO Style If we wish to assure ourselves of the identity of a friend, amark 4 Whose thoughts and actions are familiar to us, the simplest of the plan is to appreciate his appearance and to verify our im- identity of pression by the tone of his voice. Could one of our best an author friends perfectly disguise his voice and his features, it indepen- would be difficult to recognise him by the manifestations dently of his thoughts or by the moral character of his actions. of the : : : ; contents NOW the external form of a writer is his style, and it be- of his trays him even when he for some reason may be profes- writings. sing thoughts very different from those which we usually associate with his name. Great A thought can be expressed in various ways in the differences Same language ; it might even be said that the notion of of style any one language includes as many languages as there between have been original writers in it. This is truer of Greek bie than of any modern language, and is especially true of ad Greek prose writing in the fourth century B.c. A student another 2&Ving read and understood all the works of Xenophon andin the Might be unable to understand many passages in Plato. worksof Plato’s language differs from Xenophon’s, though both thesame wrote Attic prose. author. That there are peculiarities of style which distinguish @ writer among many others is almost self-evident; that the style of some writers has changed in the course of years is a patent fact; yet many objections have been made to stylistic study as a means of settling problems of ascription and chronology. Everybody knows the dis- cussions which this method provoked when applied to PLATO'S STYLE: GENERAL CONSIDERATIONS 69 Shakespeare, though, as regards Shakespeare, the difficulty is diminished by the fact that metrical intricacies and the poet's resources are more varied than is the case with prose, even the prose of such a writer as Plato. But it is to be noted on the other hand that Plato’s literary activity ‘was continued through a period twice as long as Shake- speare’s. Since most readers think that style is indefinable, they infer that it must afford an insecure basis for scientific reasoning. So Plato thought concerning all physical movements in the universe. According to him, their infinite variety hindered genuine scientific investigation (Phileb. 59 ac); they could only be guessed at with ‘some degree of probability (Tum. 29 c, 48 D); and such guesses constituted ‘a pleasure not to be repented of, and & wise and moderate pastime’ (Tim. 59 D: dyerapérntov ndovnv .. pétpiov travdcay Kat ppovipov), but they did not admit of accurate determination (Tim. 68 c D). This Platonic view of natural science extended also to linguistics (Crat. 421 D), and the Master would have smiled at those who count words in his writings. But if the science of modern mechanics, by application of new infinitesimal methods, unknown to Plato, has reached a degree of certainty by which it claims rank as a more exact science than any investigation of the human soul, then we need not allow Plato’s linguistic scepticism to keep us from the ‘moderate pastime’ of investigating his style. If an exact definition be possible of the notes which distinguish Plato’s style from the style of other writers, or by which a work written contemporaneously with the Laws differs from a work written at the time when Plato founded the Academy, then we may hope to ascertain the true order of Platonic dialogues according to the stylistic variations observed in them. There is no exaggeration in this pretension, since questions of identification are generally settled by purely F Definition of style difticult, and Plato would have held it to be im- possible. But modern methods enable us to deal with many problems beyond the reach of Plato. Identity of hand- writing no more definite than identity of style. It con- sists in a number of pecu- liarities, among which only those essential need con- sideration. Essential marks of style may be found first by investiga- tion of the vocabu- lary of an author. This in- cludes his 66 ORIGIN AND GROWTH OF PLATO’S LOGIC external tests. The identity of handwriting, consisting in many minute signs difficult of definition, is held to be so far ascertainable, that on an expert’s decision in such matters a man’s life may sometimes depend. The limited number of marks of identity contained in a signature is sufficient to decide its authenticity for all purposes. A banker requires no further security for paying out the deposits left with him under his responsibility. Docu- ments written by a prisoner, despite his denial of their authenticity, may prove his guilt in the eyes of any magistrate. If handwriting can be so exactly determined as to afford certainty as to its identity, so also with style, since style is still more personal and characteristic than handwriting. But the definition of style requires a deeper study, because style is not, like handwriting, accessible to the senses. It may be objected that, since style has an almost infinite number of characteristic notes, it cannot be reduced to one fixed formula. The answer is, that a like infinity of characteristics exists in every object of natural science, and that science is possible only through the distinction of essential marks from those which are unessential. What, then, are the essential marks of style? In- dividuality of style is developed along two different lines, each of which requires special study. An author uses words as the raw material for the expression of his thoughts, and the choice of words affords him the most obvious opportunity for displaying his individual taste. There are cases when one given word, and no other, expresses a given idea; but this 1s not the general rule. In most phrases there are words which might easily be changed for others. In every language there are many words which have never been used by some authors, and other words used only once by theirinventor. The contrivance of new compounds, and even of entirely new meanings for old and simple words, is of common occurrence in the PLATO’S STYLE: GENERAL CONSIDERATIONS 67 style of great writers. A knowledge of the words invented by an author and only once used by him is an important factor in determining questions of style and ascription. We need a full index of such words invented by all authors who lived in Plato’s time. In comparing them we should probably find that Plato proceeded in some respects differently from others in his new formations. Weshould be led to observe what methods of composition were used by him in each of his works. We should be enabled to classify the occasions when he was most inclined to have recourse to such new formations, as, for instance, in employing mathematical, physical, or dialectical terms ; and we should remark a difference between the manner of expressing these notions at various epochs of Plato’s life, taking as our starting point a few productions undoubtedly written very late, as the Laws, and comparing them with other works, as to which there is ample evidence that they date earlier: for example the Apology. No- body doubts that the tenth book of the Republic was written after the first book, and many authors agree that it belongs to a much later period. In some cases there is also a general agreement as to the relative date of two dialogues ; thus it is certain that Plato wrote the Poltticus after the Theaetetus, the Timaeus after the Republic, and it is scarcely less certain or less generally admitted that the Philebus was written after the Laches and Charmdes. A comparison between such groups would lead to definite conclusions as to the direction taken by Plato in the modifications of his style. Besides this chapter on new words, we need in Platonic lexicography another chapter on rare words borrowed from poets. It is not usual to introduce into philosophical prose words which have been heretofore used only in poetry. The language of verse always differs from prose language, and the difference is exceptionally manifest if we compare the tragedians with the Attic F 2 tendency to invent new words or com- pounds for certain classes of notions. They vary in various works ad- mitted to have been written at different epochs. Poetical words used by Plato more than by other prose writers. Use of foreign words. Rare and common words used dif- ferently. Frequency of each word in Plato not yet inves- tigated. Oppor- tunities for the use of each kind of 638 ORIGIN AND GROWTH OF PLATO'S LOGIC orators. Plato is known to have used liberally words which before him were peculiar to dramatic poetry, and it is an interesting question to answer, whether this taste be equally prominent in all his works, or be chiefly apparent in some of them. Words borrowed from a foreign dialect would form a third class of rare words to be classified and enumerated. This classification could be definitely settled only after collecting all the lexicographical evidence, because it would serve no purpose to form classes out of a few chosen examples. In the above three classes we should include first of all such rare words as are used for the expression of some peculiar idea. Their use depends mainly on the thoughts they convey, and is essentially different from that of common words occurring frequently and not generally indispensable in cases where they occur. Among these common words the particles are conspicuous. The new compounds, poetical and foreign words were closely related to the contents of the text; it is not so with particles. We are still far from possessing a complete index of the Platonic vocabulary, informing us precisely how often a characteristic word occurs in each dialogue. Assuming that no word used by Plato is missing from Ast’s Lexicon'®! and Mitchell’s Index,'!” it might be easily ascertained how many different words, and espe- cially how many substantives, verbs, adjectives, etc., he used. But a separate effort would be required to calculate the frequency of each word in each work. Even if we knew the exact number of times each word occurred, there would still remain the special task of calculating the oppor- tunities for its occurrence. Such calculations are needed for but a small part of the vocabulary, because words of rare occurrence in all works form the majority. Ast’s 101 F, Ast, Lexicon Platonicum, vols. i.-iii. Lipsiae 1885-1886-1888, 102 'T., Mitchell, Index Graecitatis Platonicae, 2 vols. Oxonii 1882. PLATO'S STYLE: VOCABULARY 69 Lexicon contains on 1,975 pages approximatively 10,000 different words used by Plato, while the whole number of words in the text of all the works of Plato amounts roughly to 600,000.'™ If each word in Plato’s text be used, on an average, sixty times, we might be justified in defining as rare words, words which in all the writings of Plato occur less than sixty times, or on average less than once in twenty pages (ed. Didot). These would form the majority, and a certain natural limit of scarcity would soon be detected, by the absence of certain degrees of recurrence. Suppose for instance that, as appears from some inedited calculations by Tadeusz Micitiski, the number of words occurring less than ten times is above 7000, and that x, is the number of words occurring between ten and twenty times, gene- rally z, the number of words occurring between 10n and 10(n+1) times, then the limit of rare words will be reached when z,=0=z2,,,=2Z,,, .-.. We should at once observe that there are no words occurring more than m and less than m+y times, and with those occurring m-+y times would begin the series of common words up to such words as occur a maximum of times, possibly thousands. Such statistics of Plato’s vocabulary would require immense labour. A new Lexicon Platonicum with all the above indicated details, in spite of the utmost economy of space, could not occupy less than several volumes like Bonitz’s Index Aristotelicus. Even this would give us knowledge only of one aspect of Plato’s style: its vocabulary. But, as Plato himself observed, we should examine in a speech not only the 103 This number of different words used by Plato has been calculated by Tadeusz Micifiski upon the assumption that each 100 entries fills 20 pages of Ast’s Lexicon, as has been found by counting the entries on 20 pages in twenty-five different parts of the lexicon. The total number of words used by Plato results approximately from the consideration that the text of all the 86 works bearing Plato’s name, including the small spurious dialogues and some of doubtful authenticity, fills in Didot’s edition only 1245 pages of 54 lines, with 8-11 words in each line. words are not the same. A limit between rare and common words is given by the average frequency. Limit of scarcity reached when certain degrees of frequency are missed. Arrange- ment of words distin- guished already by Plato from their selection. Numerical ratio of the parts of speech inter- mediate between statistics of fre- quency and the proper charac- teristics of arrange- ment. Inversion very char- acteristic in Plato’s later style, as may be seen from two samples of 500 words in Prota- goras and Laws. 70 ORIGIN AND GROWTH OF PLATO’S LOGIC choice of words, but also their arrangement (Phaedr. 236a). The arrangement of words is more difficult to define than their number. The same thought may be rendered not only by different words but also by a different arrange- ment of the same words. One of the characteristics of arrangement is the numerical proportion between verbs, adjectives, substan- tives, and other kinds of words, because in many cases the same word appears as adjective or verb or substantive ; the repetition of a noun can be avoided by a pronoun, and this allows many possible variations. For instance, ‘a wise man is unable to become unjust’ and ‘ wisdom forbids injustice’ express substantially the same thought, while in the first we have thrice as many adjectives as substantives, and in the second no adjective at all. It is highly probable that Plato did not always preserve the same proportion in the use of various parts of speech. More especially the numerical relations between adjec- tives and substantives, between substantives and verbs, between these and adverbs, afford very characteristic properties of style, which might enable us to notice similarities or differences between one composition and another. The knowledge of these quantitative relations of every kind of word is intermediate between the lexicographical statistics of the scarcity or frequency of each term and the study of the construction of phrases. Here the immediate object of study would be the relative position of subject and predicate, of nouns and determinatives, adverbs and verbs, which may all occupy the first or the second place. No author follows a uniform practice in this respect, and variation is the rule; but at each period of life an author may show a certain predilection for one or another order in the phrase. Taking only the first five hundred words in the Laws and comparing them with the first five hundred words in the Protagoras, we may readily see how great are the differences between the two PLATO'S STYLE: INVERSION 71 dialogues as to the use and order of the substantives and the adjectives : In Protagoras, words 1-500 In Lates, Number of words 1-500 i Substantives . Adjectives . Verbs (including participles) . Adjectives preceding the correlated substantive . : Adjectives following the correlated substantive . . : ‘ ———: at ee If further calculations confirmed these, then it would appear that in his later style Plato used many more sub- stantives and adjectives than in his earlier writings, and that he acquired in old age a predilection for putting the noun before its qualifying words. But in order to draw such conclusions the examination should be extended to all the works of Plato, and should include the position of adverbs before or after the verb, of genitives before or after the noun on which they depend, and of all kinds of words in their mutual interdependence. If we observe that the Philebus has some hundred peculiarities In common with the Laws, and has very few constant characters in common with other dialogues, then we may be justified in ascribing the Philebus and the Laws to the same epoch of Plato’s life, with a certainty scarcely less than that which enables us to recognise that Plato and Demosthenes both wrote Attic prose. But, besides these, there remain some other classes of stylistic peculiarities: the length, construction, and interdependence of phrases; the rhythm produced in- tentionally or resulting naturally from the order of words selected; the recurrence or exclusion of certain phonetic effects, as, for instance, avoidance of the hiatus or the repetition of syllables with the same vowels or consonants; a preference for certain sounds; the use made of quotations and proverbs; the frequency of Further calcula- tions required before drawing general conclu- sions. Stylistic tests are conclusive if their number be sufficient. Variety of such tests will be found increasing with the progress of similar inquiries. Useful- ness of detailed investiga- tions depends upon the import- ance of the chrono- logy of Plato’s writings, and is far greater than that of idle discussion on Plato’s philo- sophy. Zeller’s objections based on insuf- ficient knowledge of the existing stylistic investiga- tions, which are little 72 ORIGIN AND GROWTH OF PLATO'S LOGIC rhetorical figures and tropes; and many other points which would be suggested in the course of such inquiries. Such investigations are useful, inasmuch as they lead us to a better knowledge of the mental development of one of the greatest of all thinkers. Hundreds of German dissertations on Plato contain mere repetitions and vague generalities, of no importance for our knowledge of this philosopher. Had their authors spent the same time in studying some special property of Plato’s style, they would have made valuable additions to the positive knowledge of his development. The task of investigating every detail of style seems immense, but the number of persons fit for such work is much greater than the number of those capable of passing judgment on Plato’s philosophical doctrine. Any student, with a moderate knowledge of Greek, is made richer for life by a single reading of all Plato’s works, and this requires but an hour’s study a day during a year. And if in such a reading attention be directed mainly to some special peculiarity of Plato’s style, the impression produced by the contents need not be weakened. Hach year in all countries hundreds of students dedicate their time to classical philology. If but one in a dozen undertook a study of Plato’s style, within ten years our knowledge of Platonic chrono- logy would have progressed more than in these twenty centuries. Of the foregoing programme of investigation but a very small part has been executed, and this without any systematic common aim. Zeller, criticising chronological conclusions based on stylistic investigations (Philosophie der Griechen, II. 1. p. 512), objects that the number of characteristics investigated is too small, and that only if it amounted to hundreds could we thence draw inferences as to the chronological order of Plato’s dialogues. Of all the investigations made, Zeller quotes only those of Dittenberger, Schanz, Frederking, Gomperz, and Hoefer. He is apparently unaware that besides these authors there PLATO'S STYLE: BIBLIOGRAPHICAL DIFFICULTIES 173 are many others whose study of Plato’s style does extend over hundreds of stylistic peculiarities. It is unfortunate that these studies are little known, being chiefly pub- lished in school programmes or as university disserta- tions. The authors, generally unaware of the work of their predecessors, were therefore unable to appreciate the cumulative evidence afforded by the coincidence of results obtained through different methods. A full biblio- graphy of Plato’ is as necessary and desirable as a complete Lexicon Platonicum,’™ and neither is likely to appear very soon, for such works require an amount of material resources which is rarely at the command of Platonic scholars. Important contributions to the knowledge of Plato have been buried in introductions to the text of a single dialogue, or in dissertations privately printed for the pur- 104 The bibliography of Plato is, up to the present time, very incomplete. Besides such general works as those of Ueberweg and Zeller, many indica- tions of older literature are found in: W. S. Teuffel, Uebersicht der Platonischen Literatur, Tiibingen 1874; J.Vahlen, ‘ Zur Litteratur des Plato’ (Zeitschrift fur Oesterreichische Gymnasien, 23° Jahrgang, 1872, p. 518) ; W. Engelmann, Bibliotheca scriptorum classicorum, 8th ed. Lipsiae 1880. The current literature is indicated almost exhaustively in the quarterly Bibliotheca philologica classica, published since 1873 by 8. Calvary in Berlin. For a full Platonic bibliography it would be indispensable to sup- plement the information contained in these publications by a careful comparison of the catalogues of larger public libraries, and even of smaller university libraries in Germany, France, Great Britain, and Italy. Also the numerous antiquarian catalogues issued yearly by many German second- hand booksellers contain titles of some smaller publications not easily found elsewhere. A Platonic bibliography based on all these sources would very much facilitate special investigations, if it contained not only the titles but also a short account of the contents of rare publications. But such a work implies much travelling, because all the materials could nowhere be found together. 105 The mere cost of reprinting Ast’s Lexicon, which is now very rare, has been estimated at 700/., and as the work is incomplete, a revision and thorough comparison with the text of the best edition of Plato would be indispensable. The cheapest cost of such a labour has been estimated by Dr. C. Ritter (cf. note 184) at 750/., which raises the expense of a new edition of Ast’s work to 1,450/., while the number of buyers for such a work could scarcely exceed a few hundreds. This removes the probability of such a pub- lication being undertaken in the ordinary way. known, being usually published in small tracts or in peri- odicals. No biblio- graphy of Plato exists. A survey and com- parison of detailed investiga- tions indispen- sable as illustra- tion of the above as- sertions, though it is difficult to make it exhaus- tive. First investiga- tions on Plato’s style made by Engel- hardt of Gdarisk. From his work some 74 ORIGIN AND GROWTH OF PLATO'S LOGIC pose of obtaining degrees. Many are rarely to be found in circulation or in public libraries, and for this reason writers on Plato often neglect their predecessors. In these circumstances it may be useful to give here a short review of over forty publications referring to Plato’s style, and to insist upon the lesson they teach when their con- clusions are compared. It is probable that besides these authors others have written on this subject, without being aware of the importance of their investigations. It is common to all these detailed inquiries that, con- sidered separately, they seem inconclusive, while taken together they prepare the way for a complete change of the prevailing views on the matter to which they refer. I. ENGELHARDT. The merit of priority in considering the question of Plato’s style (but without chronological applications) belongs to Friedrich Wilhelm Engelhardt, late director of the gymnasium in Gdarisk (Danzig). He published in the course of thirty years (1834— 1864) five dissertations on Plato’s style '® in five school programmes never mentioned in any later work on that subject. His aim was not chronology but grammar, and he undertook in the first three dissertations a very careful study of the examples of anomalous construction in Platonic phraseology. After a long enumeration of all ‘anacolutha’ found in the works of Plato, he classified these stylistic phenomena, and repeated very carefully for each class the indication of all passages containing an example of that particular construction. From these very interesting tables we can easily gather some indications bearing on the Platonic chrono- 106 F. G. Engelhardt, Anacoluthorum Platonicorum specimina, i. ii. iii. program. Gymnasii Gedanensis 1884, 1888, 1845. The third dissertation contains on pp. 87-46 and 47-48 two indices of the passages enumerated also in the first two. By the same author, also as programme of the same gym- nasium in Gdayisk: De periodorum Platonicarum structura, dissertatio prima (pp. 1-86), Gedani 1858, dissertatio altera (pp. 1-27), Gedani 1864 (iv-v). PLATO’S STYLE: ENGELHARDT 75 logy. In order not to increase the bulk of our refer- ences, we must limit our quotations to those stylistic marks which may be regarded as characteristic of later style, being either limited in their occurrence to the latest dialogues, or at least increasing in their frequency. To exclude characteristics occurring occasionally in earlier dialogues would deprive us of a useful measure of affinity between each of them and the latest group. With a view to clearness of exposition and arrangement we take for granted what will only appear as the ultimate result of our inquiry, namely, that the Sophist, Politicus, Philebus, Timaeus, Critias, and Laws form the latest group of Plato’s works. This, as will be seen in the course of this exposition, becomes probable beyond reasonable doubt by the totality of stylistic observations, because these six dialogues have hundreds of stylistic peculiarities which occur nowhere else in Plato, and likewise show an increas- ing frequency of peculiarities which in other dialogues are exceptional. For easy reference the stylistic peculiarities of Plato’s later style follow here in the chronologic order of their observation, and are numbered consecutively.* Among the twenty classes of altered construction * In the following enumerations the dialogues are quoted in their probable chronological order; the numbers placed after the name of each dialogue indicate the number of occurrences; where no number is given, the occurrences have not been counted. The numbers are printed in different type to show their relative importance. 2, 3, &c. mean that a peculiarity is repeated 2 or 8 times in the dialogue named, but is not frequent. 3, 4, &c. mean that the same peculiarity, occurring 3 or 4 times, must be looked upon as frequent, in view of the size of the dialogue, if each occurrence is found on average more than once in 12 pages (ed. Didot). Numbers printed thus: $34, mean that a peculiarity is very frequent, occurring once or more in every two pages. + means a word not used before Plato; (A), a word used by Aristotle; *an dwat elpnudvoy according to the author from whom the observation is taken. Dialogues of dubious authenticity (Clitopho, Minos, Hipparchus, Epinomis, Theages, Hippias Major, Alcibiades I. and II., Amatores) or of no logical import- ance (Hippias Minor, Io, Menexenus, Lysis) are omitted in this list. The writings on the style of Plato are numbered consecutively in the notes by smal] Roman numbers placed after each title: i-xlv. pecu- liarities of later style can be gathered and in- cluded in the follow- ing list of five hundred pecu- liarities of Plato's style. Changes of con- struction observed hy Engel- hardt are specially frequent in the Laws and other late dialogues. Other observa- tions of Engel- hardt are not pecu- liar to later style or do not refer to all the works of Plato. 76 ORIGIN AND GROWTH OF PLATO'S LOGIC enumerated by Engelhardt the following characterise the later style : 1. ‘Anacoluthiae genus quod ex symmetriae studio oritur’ (Anacol. Platon. spec. III. p. 89) is a change of construction produced by Plato’s increasing taste for symmetry, and consists in beginning the second part of the phrase in the same manner as the first, as for instance in Phaedr. 288 B: romaira yap 6 épas émdeixvurar: Svoruxovvras pev, A pn AUmnY Trois aAAots mapeExel, dvapa moet vouifev’ evruyovvras Oe xai ra py ndovns déa map’ éxeivwv énraivov avayxdfe: rvyxdvery. Such changes of construction were observed by Engelhardt in: Gorg. 1 Crat. 2 Phaedo 1; Rep. 1 Phaedr. 1; Soph. 1 Phil. 1 Tim. 2 Legg 8. 2. Change of construction in consequence of the more con- venient form of the continuation (ex commodiore sequentis structurae forma, p. 89), as, for instance, Euthyd. 281 D: xevduvever cUpmavra, d rd mparov ehaper ayada elva, od wept rovTov 6 Adyos atvrois elvat, Orws aira ye xab’ abra mépuxev ayaa. . . . Such anacoluthie are found: Meno 1 Euthyd. 1. Symp. 1; Rep. 5; Polit. 1 Phil. 4 Tim. 4 Legg. 4. 8. Two different constructions co-ordinated and dependent on the same enunciation (III. p. 41: anacoluthia fit duabus structuris conjunctis), as for instance, optat. with dv and infinitive both dependent on Soxet in Lach. 184 8, or os with genit. partic. and infinitive in Charm. 164 x. Such cases were found: Apol. 1 Charm. 1 Lach. 1 Gorg. 1; Rep. 4; Legg. 9. 4. Anacoluthia ex transitu orationis suspensae in directam vel contra (III. p. 41): Gorg. 1 Symp. 1 Phaedo 4; Rep. 8 Phaedr. 1 Theaet. 2; Soph. 1 Phil. 1 Legg. 2. 5. Cases of omitted apodosis are quoted (p. 44) by Engelhardt : Gorg. 2 Symp. 2 Phaedo 2; Rep. 1; Phil. 1 Legg. 8. The other kinds of change of construction enumerated by Engelhardt do not appear to be specially frequent in the latest dialogues. His collections extend over all the works of Plato, and include more than four hundred cases. Being unaware of the chronological application of his work, he perhaps did not attempt a painful completeness of quotations. But even if he collected only those changes of construction which struck his attention in a first reading, we may assume that he had no special reason to notice the actual occurrence of such cases in one work more than in another. His observations are therefore valuable, and they may be accepted as at least approximate. PLATO’S STYLE: ENGELHARDT, KAYSSLER 77 This author’s later investigations on the construc- tion of phrases are limited to the Phaedo and Republic, s0 that they afford no matter for comparison. It is interesting, however, to notice that according to Engel- hardt co-ordination of phrases prevails in Plato over subordination, and that the principal sentence generally precedes all subordinate clauses. Herein he sees a radical difference between Plato and Demosthenes, who inverted more frequently the natural order. Engel- hardt thinks (Period. Plat. I. p. 26) that this difference in the order of co-ordinate and subordinate sentences is due to the dialogical character of Plato, as opposed to the rhetorical character of Demosthenes. He would perhaps have been less confident as to the essential difference between the style of Plato and Demos- thenes, had he given as much attention to the Laws as to the Republic and Phaedo. It remains an inter- esting problem to compare the Laws and other dialogues as to the construction of phrases, and Engelhardt’s classification would be most useful for this purpose. II. KayssuErR. Of less importance is a small disser- tation by Kayssler!*’ (1847) on Platonic terminology. The author accuses Plato of inconsistency in the use of terms, even as defined by himself, and enumerates the terms which he held to be the most important, without any attempt at comparing earlier with later dialogues, or at using the difference in terminology as an instrument of chronological determination. Ill-V. J. Braun!® (1847, 1852) and A. Lanar'” (1849), quoted by Engelhardt, seem also to have left 107 Kayssler, Ueber Plato’s philosophische Kunstsprache, Oppeln (Polish Opole) 1847 (vi). The inexactitude of quotations is seen from the fact that Kayssler affirms p. 18 to have found cvvaywyf and d:alpeors only in Phaedr. Soph., while they occur also in Theaet. Rep. Phil. ea J. Braun, De hyperbato Platonico i. ii. progr. gymnas. Culmensis {Cheltmno), 1847, 1852 (vii—viii). 109 A, Lange, De Constructione periodorum, imprimis Platonis, Vratis- laviae 1849 (ix). Other authors of the same epoch are less im- portant. Kopetsch of Lyk published an in- teresting disserta- tion on a class of adjectives, among which many have been invented by Plato, but few can be included in this list because Ko- petsch’s enumersa- tion of passages is incom- plete. 78 ORIGIN AND GROWTH OF PLATO’S LOGIC chronology out of the question in their investigations on Plato’s phraseology. To the same time belongs the dissertation of F. MicHEtis'!° (1849), which deals more with Plato’s views on style and grammar than with any specialities of Plato’s own style. VI. KorrtscH. Some interesting observations are contained in the dissertation of Gustav Kopetsch (1860), teacher in the gymnasium of Lyk." He also had no chronological purpose, but his grammatical aim to collect from Plato’s writings every kind of information about the use of adjectives in ros and réos gives us an oppor- tunity to select from his enumerations such uses of this class of words as appear to be peculiar to Plato’s later style : 6. Adjectives in ros composed from a substantive and a verb are very rare. Kopetsch enumerates only (pp. 4 and 19) : Phaedr. 2 (opupnAraros 286 B, vupddAnrros 288 D) Tim. 1 (mupixavros 865 c) Critias 1 (yetporoinros 118 c), Legg. 1 (alypddwros 919 A). 1. Adjectives in rds, oxytona, formed from compound verbs (p. 6): Prot. 2 Meno 8 Phaedo 1; Rep. 1 Phaedr. 1; Polit.1 Tim. 4 Legg. 8 (mapairnrés, diaBards, éxdexrds). 8. Superlatives in roraros, beginning with $8vs or eb (p. 7): Phaedo 2 (SvaeAeyxrorarov, evappoordraroyv) Tim. 3 (S8ucadwrdraror, dvoxiwnrdérarov, evxiyyrorarov) Legg. 1 (8ucperayeipiordrarov). Super- latives in roraros of other adjectives occur besides: Apol. 1 Prot. 1 Symp. 3 Rep. 8 Soph. 1 Phil. $ Tim. 2 (with the preceding Tim. 5). 9. Adjectives in ros composed of an adjective and verb: Phaedo 1 (moAvOpvAnros); Rep. 1 (soAvOpuAnros) Phaedr. 1 (lcopérpyrov) ; Polit. 2 (6Acoxeoros) Tim. 1 (vedrynros) (p. 19). Kopetsch quotes many other uses of adjectives in ros, but without attempting completeness of quotation except in the above cases of very rare occurrence. Of some hundred adjectives quoted and classified by this author, many might be included in our list, had their ne F, Michelis, De enuntiationis natura, sive devi quam in grammatica habuit Plato (pp. 1-68), Doctor's dissertation, Bonn 1849 (x). 1 @, Kopetsch, De verbalibus in ros et réos Platonicis dissertatio, cut intextae sunt breves de Homericis adnotattones, Lyck 1860, programme of the German Gymnasium in Lyk (xi). PLATO'S STYLE: KOPETSCH, SCHONE 79 occurrence been completely investigated. This was not the aim of Kopetsch, since he was not aware of any application of his work to Platonic chronology. For his purpose it was sufficient to quote a few characteristic occurrences of each word. A full investigation of the use of adjectives in ros and réos in the works of Plato remains a very interesting problem for future special inquiry. Here we quote only two more single words which, according to Kopetsch, as well as Ast, occur but seldom in Plato: 10. dyémros (p. 27): Prot. 1; Phaedr.1; Legg. 1. 11. peurros (p. 21): Theaet. 1; Legg. 1. VII. R.ScH6nz. The first author who insisted energetically on the importance of stylistic observations as leading to chronological conclusions seems to have been R. Schéne!!? (1862) in his dissertation on Plato’s Protagoras. But he had a very superficial knowledge of Plato and of the means of defining literary style. Schone, despising enumeration of stylistic characteristics, quotes the authority of art critics, who judge whether a picture has been painted by Raphael or Murillo, without condescending to give special reasons for it, and he wishes to introduce into Platonic chronology such artistic intuitive judgment without the help of reasoned evidence. Still, Schéne is right in his fundamental argument as to the comparative value of style and contents for chronological conclusions. He declares that an author can put in each work such contents as he chooses, while his style will simply be the result of his effort to write as well as he can, if he is so careful about the form of his writings as Plato was. Hence style is the surest measure of the stage of a great writer’s evolution. Schéne quotes Lessing and Goethe as competent authorities for such a view on 112 Richard Schéne, Veber Platons Protagoras, Leipzig 1862 (xii). The author confesses his indebtedness for a great part of his theories to Prof. Weisse’s lectures on Plato delivered in 1860-1861 at the University of Leipzig. Schéne recognised the supe- riority of stylistic tests as means of chrono- logical conclu- sions, but failed to find the right method of measuring differ- ences of style. 80 ORIGIN AND GROWTH OF PLATO'S LOGIC the stylistic progress of great writers, and he concludes : ‘wir diirfen den Stil als ein schlechthin allgemeines und sicheres Kriterium betrachten, wo es sich um Echtheit und Zeitfolge der platonischen Schriften handelt’ (p. 21). But after having thus clearly set forth the importance of stylistic study in determining Platonic chronology, Schéne fails to find a right method for such investiga- tions. He believes an exact analysis of style impossible, ignoring the labours of Engelhardt, Braun, Lange, and Kopetsch ; and invokes a mysterious power, the ‘ feeling of style.’ This ‘feeling’ led Schéne to see a higher degree of stylistic perfection in narrated conversation than in dramatic dialogue. He inferred that all narrated dia- logues—the Charmides, Protagoras, Banquet, Phaedo, Republic, and Parmentdes—are later than all the works whose form is dramatic. Sch6éne did not perceive that Plato, after having used the form of narrated dialogues, grew tired of the repetitions which it involves, and returned to the primitive dramatic mode. Had Schéne limited his judgment to the relation between Protagoras and the small dramatic works, such as the Laches, Crito, Euthyphro, his observation of the stylistic perfection of a narrated dialogue could not have led him to the absurdity of placing the Laws and even Timaeus earlier than the Republic. Thus he discredited the method which he was the first to propose. He did not under- stand that for a philosopher contents are more important than form, and that the artistic skill which Plato exercised on his narrated dialogues was peculiar to a time when the deepest problems of thought had not yet absorbed the writer’s whole attention and endeavour. Schone represents Plato as struggling during his maturity for perfection in the form of the philosophical dialogue, after spending earlier years in elaborating philosophical convictions. Thus the Sophist and Philebus appear to Schone earlier than the Protagoras. He had the merit PLATO'S STYLE: SCHONE, MARTINIUS 8] and boldness of drawing extreme consequences from his theory, arriving at the untenable conclusion that Plato renounced dialectical aims for the sake of artistic perfec- tion (p. 82). VIII. C. Martinius. What Schéne attempted by a mistaken route has been more successfully carried out as regards a special characteristic of Plato’s style by C. Martinius'* (1866, 1871), who, himself a teacher, began with the conviction that Plato as a teacher must have progressed in the art of interrogating, and that therefore differences in the form of questions might lead to chronological conclusions as to the order of the dialogues. Martinius first collected what Plato himself had said upon the art of asking questions, and then pro- ceeded to classify the interrogations found in Plato’s dialogues. Enumerating not less than eighteen differ- ent kinds of questions, he invites the reader to continue the inquiry as to the occurrence of each of these in the entire works of Plato, in order to establish the pro- gress made by the philosopher in his practice as a teacher. Martinius himself published, five years after his first effort, a very short summary and continuation,'* in which he insists on the importance of ‘Suggestivfragen,’ that is, questions which take for granted something not yet accepted or discussed. Such questions seek to determine something sup- posed to be known, while it is really not known, as if a prisoner were asked the time when he committed a crime which he has not admitted. In Plato’s dia- lopues the imputed object is not an action but a know- ledge, as, for instance, when (Phaedr. 276 a) Socrates asks whether another kind of teaching is not much ns ©, Martinius, ‘Ueber die Fragestellung in den Dialogen Platos,’ in the Zeitschrift fur das Gymnasialwesen, xx** Jahrgang, Berlin 1866, pp. 97-119, and 497-516 (xiii). 4 ©, Martinius, ‘ Ueber die Fragestellung in den Dialogen Platos und iiber eine besondere Eigenthiimlichkeit derselben,’ Jahresbericht tiber das Progymnasium 2u Norden, 1871, 4to., pp. 1-18 (xiv). G Martinius knew the right method, but his work remained incom- plete, being only a pro- gramme deserving the atten- tion of future investi- gators of Plato’s style. The pro- blem of a classifica - tion of questions in Plato’s dialogues already proposed by Ueber- weg. First attempt of a methodic solution of the problem of Platonic chrono- logy by 82 ORIGIN AND GROWTH OF PLATO’S LOGIC better and more powerful, while he had not yet named that other kind and obtained assent as to its existence. Such questions were seen by Martinius (i. pp. 9-13) in Gorg. 486 D, Rep. 414 B, 421 c D, Phaedr. 276 a, Theaet. 158 B, 187 cp, 190 #, Parm. 156 D, Polit. 278 a, 290 a, 302 B, Phil. 38 pz, Legg. 646 £, 691 B. We cannot include these quotations in our list of characteristics of later style, because Martinius did not profess to give a complete enumeration but only examples of each kind of questions. He seems not to have continued and completed these investigations, which are remarkable for their method and originality, and might serve as a starting point for anybody who undertook to realise the programme proposed by the ingenious Hanoverian teacher. The problem of defining the differences between various modes of putting a question in Plato’s dialogues had been also slightly broached by Ueberweg (Unter- suchungen, p. 207), who observed that in the Sophist, the Politicus, and the Philebus, as also in the Timaeus, Critias, Laws, the play of question and answer becomes more and more conventional and more remote from the tone of natural conversation, approaching to the form of an uninterrupted lecture. The observation of such a peculiarity limited to only six dialogues (Soph. Polit. Phil. Tum. Critias Legg.) was in so far a very important first step in conscious determination of Plato’s later style, since 1t could not well be attributed to chance. IX. Lewis CAMPBELL. No single characteristic of style, however important, suffices for general conclusions, as the case of Schéne shows. It is edifying to see the great contrast between Schéne’s confidence and the modest caution with which stylistic inferences were justified by an author who alone enumerated and com- pared more characteristics of the style of Plato than all other investigators put together. This contribution to the study of Plato’s style, still after thirty years the PLATO'S STYLE: CAMPBELL 83 most important of all, is contained in the introduction to an edition of the Sophist and Politicus of Plato by Lewis Campbell, then Professor of Greek in the Uni- versity of St. Andrews (1867). Campbell''® knew none of the authors enumerated above, and he approached the study of Plato’s style quite independently, with the special purpose of determining the date of the dialogues which he edited while main- taining their genuineness. He had the original idea of going through Ast’s Lexicon Platonicum and of finding out what words are peculiar to each dialogue in common with the group of Timaeus, Critias, Laws, which are recognised to be the latest works of Plato. He assumed that a word, for which Ast quotes references only from a few dialogues, does not occur else- where. This assumption is probably correct in the great majority of cases, and is quite justifiable in a first general inquiry, though it would be desirable, after collecting such words as Ast quotes only from a few dialogues, to examine the bulk of Plato’s text in order to be certain that they occur nowhere else. When Ast prepared his Lexicon Platonicum, more than seventy years ago, he could not foresee the importance now attached to precise reference; and for some particles, which have been specially investigated afterwards, and are peculiarly characteristic of Plato’s later style (as, for instance, pnv), Ast quotes only a small number of the instances remarked by later writers. In the introduction to an edition of two dialogues, Campbell could not go into so many details as later investigators of Plato’s style; he does not quote the single passages in which each word occurs, nor even all the words observed, and he condenses the results of a long and tedious labour into a few pages of dry "5 The Sophistes and Politicus of Plato, with a revised text and English notes, by the Rev. Lewis Campbell, M.A., Professor of Greek in the University of St. Andrews : Oxford, at the Clarendon Press, 1867 (xv). GZ means of the study of Plato’s style. Campbell based his investiga- tion on Ast’s Lexicon, and had the special purpose of determin- ing the position of the dialogues he edited. His work remained entirely unknown to all later investi- gators of Plato’s style, and he did not insist on the im- portance of his dis- coveries. Thus it is necessary to explain his obser- 84 ORIGIN AND GROWTH OF PLATO'S LOGIC enumeration (Introduction, pp. xxv—-xxx), which, to be fully appreciated, needs more comments than the author cared to give. His observations are of such novelty, that, giving so many new facts, he left the reader to weigh them and to judge the correctness of the conclusions drawn with admirable sagacity by the author. Such readers as he had did not notice the importance of the evidence collected. Having brought together ma- terials sufficient to prove that the Sophist and Polwticus must have been written in Plato’s old age, Campbell con- cludes with the modest phrase: ‘If our hypothesis of the comparatively late origin of these dialogues is correct, the non-appearance of the Philosopher coincides with and renders more significant the abandonment of meta- physical inquiry in the Laws.’ He had laid the first foundations of a new solution of the problem of Platonic chronology. Twenty-two years later, reviewing a Ger- man book, which on a much smaller basis proclaimed like results with much greater confidence, Campbell said 6 with equal candour: ‘Now, if not before, it is clearly proved that the Sophistes, Politicus, Philebus, Timaeus, Critias, and Leges, in this order, or nearly so, form a separate group, and are the latest written... inquiries wholly independent of each other have led to this coincidence of result.’ An author capable of such self-effacement could not impress upon the reader his convictions as definitive truths, and, accordingly, Campbell’s investigations re- mained entirely unnoticed for nigh thirty years.” It 6 The Classical Review, February 1889, pp. 28-29, review of C. Ritter, Untersuchungen tuber Plato, by Lewis Campbell. 17 The first public recognition of the exceptional importance of Campbell’s investigations on the style of Plato is contained in the vol. ix. pp. 67-114 of the Archiv fiir Geschichte der Philosophie (October 1895) in an article ‘ Ueber Echtheit Reihenfolge und logische Theorien von Platos drei ersten Tetralogien’ and in the Bulletin del’ Académie des sciences de Cracovie, October 1895 pp. 268-277, where the Polish work O pierwsszych trsech PLATO'S STYLE: CAMPBELL 85 was also not suspected that the introduction to an edition of the text of two isolated dialogues could contain a capital inquiry into the vocabulary of all the works of Plato. Under these circumstances it may be well to recall Campbell’s chief observations, the more so as these should be repeated, in order to give them greater exactness than can be afforded by our confidence in the relative completeness of Ast’s lexicon. Assuming, with all competent writers, that the Laws, as well as Timaeus and Critias, belong to Plato’s latest period, Campbell sought for peculiarities of style which, being common to these works, are also observed in others. He found the following points in which the Sophist and Politicus, partly also the Philebus, are similar to the Timaeus, Critias, and Laws: 12. The Sophist and Politicus are both the middle pair of an un- finished tetralogy, sketched out in the second dialogue of the series; so are the Timaeus and Critias (Introduction, p. xix). In both tetralogies the plan of the four consecutive dialogues was not indicated in the first of the series. Neither in the Republic is there any hint as to the author’s intention of writing the Timaeus, Critias, and Hermocrates; nor is there in the Theaetetus any clear indication con- cerning the Sophist, the Politicus, and the Philosopher as an intended continuation. In both tetralogies the fourth dialogue remained un- written. There is no evidence that Plato ever wrote the Hermocrates announced in the Timaeus, or the Philosopher announced in the Sophist. The first dialogue of both tetralogies is conducted by Socrates, while in the second and third Socrates remains a listener, who merely proposes the subject of conversation at the outset. The idea of planning out four consecutive dialogues as one larger tetralogjach dztet Platona, by W. Lutostawski, is announced. In France Campbell’s discoveries became known only after a lecture delivered on May 16, 1896, in the Institut de France, in Paris, before the Académie des sciences morales et politiques, and published in vol. cxlvi. of the Compte rendu des séances et travaux del’ Académie des sciences morales et politiques, also apart with an additional preface : W. Lutostawski, Sur une nowvelle méthode pour déterminer la chronologw des dialogues de Platon, Paris, H. Welter, 1896. More detailed is the account of Campbell’s investi- gations in the Polish work of the same author, ‘O pierwszych trzech tetralogjach dziet Platona,’ published by the Cracow Académie des Sciences in vol. xxvi. pp. 81-195 of the philological memoirs of that society, and also in a separate volume, Cracow 1896. vations in order to enable others to repeat them. He chiefly sought for pecu- liarities which denote the similarity of Sophist, Politicus, and Phile- bus to Timaeus, Critias, and Law: . Analogy between the tetralogy planned in the Sophist and that which is sketched in the Tiunaeus. Socrates is no longer the chief teacher in Timaeus, Critias, Laws, Nor in the Sophist and Politicus, and he appears as a pupil of Diotima in the Syinpo- sium, of Parme- nides in the Par- menides. 86 ORIGIN AND GROWTH OF PLATO'S LOGIC whole corresponds to the great length of the last work of Plato, the Laws. It is also psychologically plausible that Plato, grown old, had more to say, and said it in an ampler manner. His recognised earliest productions, the so-called Socratic dialogues, are much shorter than the works of his mature age. The most obvious reason which prevented him from finishing the two intended tetralogies is the shortness of life, and this alone would lead us to ascribe the second and third dialogues of these unfinished tetralogies to a later time than both first parts: that is later than the Republic, and later than the Theaetetus. On the other side the Republic and Theaetetus being singled out among all the other works by the circumstance that a continuation to them has been given, it seems probable that this relation of both to later dialogues is due to their relatively late date, because Plato is more likely to have connected his latest works with those preceding them, than with works written very much earlier. If we take into account also that the Laws differ from all earlier dialogues by their volume, and that they may be considered as consisting of at least four parts, we may observe that the late peculiarity of uniting several] dialogues into a larger whole extends to Soph. Polit. Tim. Critias Legg. (and toa certain degree also to Rep. and Theaet.) 13. The Sophist and Politicus, as well as Timaeus, Critias, Laws, also in some degree the Parmenides and Symposium, are the only works of Plato in which Socrates is not the principal figure in the conversation, and in which other teachers take his place (Introduc- tion, p. xix). While these are named in the Symposium, Parmenides, Timaeus, and Critias, they are but unnamed abstract personalities in Sophist, Politicus, and Laws. The stranger from Elea, the Athenian stranger, are representatives of pure reason and experience, while the Platonic Socrates of other dialogues is generally a concrete personage, with a certain historic idiosyncrasy, although freely adapted to the expression of Plato’s theories. The predominance of other teachers over Socrates characterises only seven dialogues: Soph. Polit. Tim. Oritias Legg. and to a certain degree Symp. and Parm. 14. The exposition in the latest works is chiefly didactic (Intro- duction, p. xx), and the Socratic dissimulation of knowledge, still appearing in the Theaetetus, is definitively forgotten. ‘The Philosopher guides his pupil by a path familiar to himself to conclusions which he foreknows’ (p. xx). ‘The speakers are playing at a laborious game (Parm. 18738) to which they are evidently not unaccustomed, and which proceeds according to certain rules’ (p. xxi). With no sudden gust of eloquence as in the Republic or Theaetetus, but with a gravity akin to solemnity, Plato discusses in these works subjects loftier than those proposed at the outset, and displays a fixed conviction of human nothingness. PLATO'S STYLE: CAMPBELL 87 This refers to: Parm. Soph. Polit. Phil. Tim. Critias Begg. 15. From the conversational freedom of the Republic we are led to scientific exactness and compression (Introduction, p. xv); there is an air of self-imposed restraint; an appearance of studied order and arrangement becomes manifest also in the occasional reference to earlier dialogues, as in the Soph. 217 c the Parmenides is quoted, in the Soph. 2164 the Theaetetus, in Polit. 2848 the Sophist, in Tim. 17c the Republic, in the Critias 1068 the Timaeus, and less clearly in the Laws 71la, 7124, 73839BcpD, the Republic. Also the ‘preludes’ and ‘ recapitulations,’ disdained in the Phaedrus, are quite as common in the Sophist and Politicus as in the Laws, the Timaeus, and Critias (p. xxiii). This care for form, while the perfection of form wanes, may be best explained by the increasing preoccupation with the philosophical contents, peculiar to the writer’s old age. The dry light of reason accompanied the decline of poetical grace and power. A vein of refined and caustic satire succeeds to the simple and playful humour of earlier times (p. xix). This special and evident care for exactness of expression, leading to a fixed terminology, belongs to: Parm. Soph. Polit. Phil. Tim. Critias Legg. 16. The periods are more elaborate and less regular than in the Republic: (Introduction, p. xxxviii) Soph. Polit. Phil. Tim. Critias Legg. 17. The natural order of words is more often inverted, and the hyperbaton in the use of particles is specially frequent (p. xxxvli) : Soph. Poht. Phil. Tum. Critias Legg. 18. The monotonous recurrence of a certain rhythmical cadence (Introduction, pp. xx and xl) under the increasing fascination of rhythmical linguistic music : Soph. Polit. Phil. Tum. Critias Legg. 19. Careful balancing of words so as to relieve the tediousness of a prolonged phrase by the counterposition of noun and epithet, verb and participle, subject and object, and by the alternation of emphatic and unemphatic words (Introduction, p. xli): Soph. Polit. Phil. Tim. Critias Legg. 20. The adjustment of long and short syllables so as to quicken or retard the movement of the sentence. Sometimes short syllables are accumulated as in choric metres; more often a sentence is concluded with an iambic hemistich, or with a dochmiac, each generally terminating with a dissyllable, which is often divorced from the immediate context (Introduction, p. xlii): Soph. Polit. Phil. Tim. Critias Legg. For all these peculiarities Campbell quotes examples which need not be repeated here, because points 16-20 The latest seven dialogues have a more pro- nounced didactic character than all earlier works. We notice in them a methodic proceed- ing and quotations of earlier works ; a special care for form and termino- logy. Phraseo- logy more elaborate. Inversion frequent. Phonetic effects sought for. Symmetry in the order of words and evenin the order of syllables. These points should be in- vestigated again. Avoiding of the hiatus later observed by Blass. Of many grammati- cal pecu- liarities observed and quoted by Campbell only one can be in- cluded in our list because he did not attempt complete- ness of enumersa- tion. Lexico- graphy. The vo- cabulary of Plato’s later works is very original, contain- ing many words 88 ORIGIN AND GROWTH OF PLATO'S LOGIC deserve renewed inquiry, as they have not been treated exhaustively. 21. The avoiding of the hiatus, a peculiarity of the same order, though not expressly noticed by Campbell in 1867, is implied in the influence of rhetorical artifice on Plato, to which Campbell directs our attention (p. xl). According to later investigations of F. Blass !** (1874) the avoidance of hiatus is limited to the following dialogues: Soph. Polit. Phil. Tim. Oritias Legg. 22. The use of the Ionic dative plural in oc was indicated by Campbell (p. xxiv) as a characteristic of later style. Its occurrence has been later exactly determined by C. Ritter, and found only in: Rep. 6 Phaedr. 8 Polit. 4 Tim. 2 Legg. 85 (C. Ritter, Unter- suchungen, p. 9; also Jowett and Campbell, Republic, vol. ii. p. 52). Some other grammatical peculiarities of later style, observed by Campbell, as: perfects with present meaning, participles with auxiliary verb, neuter article with the geni- tive to express the abstract notion of a thing, ellipse of ro pév etc. with ro 54 etc. following, redundant or explicit use of the participle, repetition of a verbal notion which has been already expressed or implied (Introduction, pp. xxiv— XXvli) cannot be included in our list, because they are in- dicated without a complete quotation of their occurrences in all the writings of Plato. These points ought to be investigated anew by some philologer acquainted with Campbell’s work, and they would yield very interesting results. The most important peculiarity of Plato’s vocabulary in his later works is its originality, leading the author to invent many new words, or to mould old words to new ideas with an affectation of variety and minuteness of distinc- tion (Introduction, p. xxx). In the Laws Campbell found 1,065 words occurring nowhere else, on 317 pages of text (ed. Stephani; Campbell quotes 345 pages because he did not take into account the space without text at the end of each book). This yields a proportion of 336 original words to each 100 pages, an originality of vocabulary absent from earlier works of Plato. The Tumnaeus and Critias show ORIGINALITY OF PLATO'S VOCABULARY 89 the same tendency to the use of rare words, as they have on 90 pages 427 words unused elsewhere by Plato. This raises the proportion to 474 original words in 100 pages. It does not imply that Plato in writing the Timaeus and Critias tends to a greater use of new and rare words than in writing the Laws, for physics exceed politics in the opportunities for such usage. In such a political treatise as the Laws, 336 new words to 100 pages show as great a leaning to an original vocabulary as 427 new words to 100 pages in a physical treatise. Turning to the Sophist and Politicus taken together as one whole, in 107 pages there are 255 new rare words not found else- where in Plato, a proportion which corresponds to that of 239 in 100 pages. That this bent towards the use of rare words was increasing we can easily see by comparing the three dialogues which were avowedly written by Plato in succession. In the Theaetetus he employs 93 new words unused elsewhere, that is 133 to 100 pages (ed. Steph.), in the Sophist 187 to 100 pages, in the Politicus 295 to 100 pages ; but in the Philebus only 100 to 100 pages, and in the Phaedrus 326 to 100 pages. (These last numbers are given in vol. 11. of the edition of the Republic by Jowett and Campbell, pp. 53-55.) It is to be regretted that nobody has as yet calculated these proportions for the Parmenides, Republic, and for earlier dialogues. The numbers given by Campbell refer only to: Phaedr. Theaetet. Soph. Polit. Phil. Tim. Critias Legg. This originality of vocabulary is a very powerful argument in favour of the late date of the Sophist and Politicus, but cannot be included in our list of marks of later style, so long as comparative statistics about all the works of Plato in this respect are not estab- lished. The absence of a fixed terminology, which is observed by readers of the earlier dialogues, is less noticeable in the Sophist and later works. In all these dialogues a great number of rare words recur, besides those used used only once. But this peculiarity has not yet been investi- gated through- out all the works of Plato like the recur- rence of fixed terms. Some words used in Timaeus, Critias, Laws, occur besides in only one of the earlier dialogues. This allows a measure of affinity between each dia- logue and the latest group. 90 ORIGIN AND GROWTH OF PLATO'S LOGIC only once, and this repetition of new and rare words shows an inclination to ‘fix in language some of the leading generalisations of philosophy’ (Introduction, p. XXx). Taking the Timaeus, Critias, Laws, as containing Plato’s latest terminology, Campbell counted the words which each dialogue shared with this latest group, and which occur nowhere else in Plato. If we reduce the numbers given by Campbell to the proportion of 100 pages, and if we allow a correction consisting in count- ing a8 common and peculiar to Sophist and the group of the Laws also those words which, besides these four dialogues, have been used only in Politicus—then we have in the Sophist to 100 pages 108 new words common and peculiar to the Sophist and to the group of the Laws. In the Politicus the number of such words rises to 136 in 100 pages, counting also those which besides occur only in the Sophist. Of the other Platonic dialogues, the Phaedrus alone shows a vocabulary which in almost equal measure approaches that of Plato’s recognised latest writings, containing a proportion of 117 rare words to 100 pages (ed. Steph.), which apart from this dialogue are used only in the group of the Laws. This does not necessarily prove that the Phaedrus belongs to the same epoch, since, the Phaedrus being in more senses than one a pro- gramme, and a work of rare poetic richness and artistic excellence, it 1s natural that Plato should have retained in use many words there first employed. Among the other writings, the Philebus affords a remarkably low proportion of such words. They are only sixty-two to 100 pages, though in many other respects the Philebus is more nearly related to the Sophist and Politicus, and also to the group of the Laws, than the Phaedrus. This low figure is explained by the circumstance that no account was taken of such words as occur, besides in Philebus and the latest three dialogues, also in Sophist and Politicus. Assuming that Plato wrote the Philebus at about the PLATO’S LATER VOCABULARY 9] same time as the Politicus, it would be natural that he should use in both some rare words peculiar to the group of the Laws. Allowing for such words, the figure rises to ninety-two rare words in 100 pages (ed. Steph.) com- mon and peculiar to the Philebus with the latest three dialogues. The importance of these figures is apparent on comparison with those of other works in which Campbell counted the words peculiar to the group of the Laws. These are seen from the table on the next page (calculated on Campbell’s table, p. xxxii). In this table some anomalies require explanation. The Protagoras, being an early dialogue, has more words peculiar to the latest group than could have been ex- pected. To explain this we should require to know what words these are, because if they refer to some special subject treated in the Laws as well as in the Protagoras, the coincidence would be natural. Later inquiries have not confirmed such an affinity between the Protagoras and the latest dialogues. On the other side the numbers for the Theaetetus and Parmenides are remarkably low. This might be explained by the circumstance that Camp- bell according to his method did not include in these numbers those words which, besides occurring in each of these dialogues, are found in some other dialogue belonging to the same epoch. The correction of the error resulting from this omission can be made here only for the Sophist, Politicus, and Philebus, and has altered very much the proportions given by Campbell. Really, if a word is peculiar to the latest dialogues and is found besides in two other works, the occurrence of this word in these two works is as much a sign of affinity between them and the latest group as (and is perhaps more significant than) if the occurrence were limited to one dialogue besides the three latest works. The Theaetetus has many words in common with the Republic, the Parmenides many words 1n common with the Theaetetus and Sophist, as later investi- gations have sufficiently shown. All these words were This first table of affinity requires correc- tions as to the Par- menides, Theaete- tus, and Philebus, which have been shown by later investiga- tions to be nearer to the Laws. 92) ORIGIN AND GROWTH OF PLATO'S LOGIC Statistics of rare words in Plato according to Lewis Campbell. Number Number Number of | Of rare* | Proportion to | Of rare} Proportion words oc- 100 pages. worse to 100 pages. Abbre- Name of viation Dialogue.’ Euthyphro . Apology . . Crito. .. Charmides . Laches Lysis . Protagoras . Meno... Euthydemus Gorgias . Cratylus . Symposium . Phaedo .. Platonicum.’ WIV WWI Wee | * Unknown, but ascertain- able from Ast’s ‘ Lexicon PAD OD ORD 90d OD Med ORD Fad Od OW} On nd Od Republic, . Phaedrus . Theaetetus . | Parmenides . Sophistes . Politicus. . Philebus. . 43 52° Timaeus. . it jf over Critias 5 11} | ( 508° Laws .. 236% | over 1146° Menexenus . ||Menex. . Ion. . . © {{Ion Hipp. Minor ||Hipp. [. Alcibiades I. |} Ale I. . oF WI OW OW Ore vv WO SV Observations. * The dialogues are in their presumed chronological order, as resulting from the sum of stylistic observations, 1834-1896 ; in some doubtful cases, as for the first six small dialogues, the traditional order preserved in Manuscripts (tetralogies) has been main- tained. The Repudlic is placed between Phaedo and Phaedrus, with reference to the greater part of it, though it is supposed that the phir orgs Sead the Rep. was written before the Phaedo, and some other parts after the Phaedrus. Those which have no logical import- ance and will not be dealt with in the present work (Menex., Ion, Hipp. I., Alc.) are omitted and follow only in this table after the Laws. * These numbers are not given by Campbell, but are calculated on his ‘numerical ratios.’ * The pages ed. Didot are more equally printed than in any other edition ; and they form the best measure of the amount of text. * Corrected after elimination of an error resulting from the circumstance that Campbell ponte in Rep. and Legg. also some pages without text, between every book and the Ollowing. * This number contains the words common to Tim. Oritias with Legg., and those occur- ring in Tim. Critias, and nowhere else, according to J. and C., . Vol. IL. p. 57, a * This number contains the words common to Tim. Critias with Legg., and those of Legg. one. 7 Including five such words which also occur in Polit. * Including five words which are also found in Soph. * Including eight such words, which are also found in Soph. Polit. *° This number results from the ratio § given by cempbell counting 295 pp. as he counted. The proportion is increayed through the omission of pages without text. 6 - piers the ratio given by Campbell the result would be 34 words; he may have found ree or four. PLATO'S STYLE: CAMPBELL 93 excluded by Campbell from the number of words ‘common and peculiar’ to each dialogue with the group of the Laws. Thence, partly, the low figures for Theaetetus and Parmemdes. As to the Parmenides, the very peculiar and exceptionally abstract contents of this dialogue also make it impossible to find many rare words in it, because the greatest number of rare words refer to concrete objects. Apart from these easily explained exceptions, Campbell’s observations, as represented in the above table, show clearly that the Sophist, Politicus, Philebus, also the Republic and Phaedrus, have the greatest affinity in vocabulary to the latest dialogues. There results the following important addition to our list: 23. Occurrence of rare words common and peculiar to each dialogue with the latest group once or more in each page (ed. Didot) is confined to: Rep. 246 Phaedr. 61 Soph. 87 Polit. 78 Phil. 52 Tim. and Critias 508 Legg. 1146, while such words are scarcer, but still occur more than once in two pages in: Symp. 33 Phaedo 42 Theaet. 27. Campbell found*by this method over seven hundred characteristics of the later style of Plato, each word recurring in certain dialogues being as much a pecu- liarity of the style of these dialogues as any of the more general stylistic properties. He inferred that the Theaetetus and Phaedrus form with the Republic an earlier group (p. xxxix) than Sophist, Politicus, and Philebus, and that these more nearly approach Timaeus, Critias, Laws in their style than any other works of Plato. He could not have so correctly recognised the middle group of Republic, Phaedrus, and Theaetetus, had he simply considered the number of characteristic peculiarities, without taking into account also their nature. The weighing of evidence in every kind of statistics is the indispensable condition of correct con- clusions, and Campbell has shown a surprising power of divination in connecting the Theaetetus and Phaedrus with the Republic in face of the purely numerical data he The affinity of Sophist, Politicus, Philebus, with Timaeus, Critias, Laws, is evident. The group of Republic, Phaedrus, T heaete- tus also recog- nised by Campbell, though his evidence at first sight placed the Phaedrus much later and the Theae- tetus much earlier. But he avoided errors, carefully weighing his evidence, wherein his method surpassed the method of all later investi- gators. The Par- menides contains some sig- nificant terms re- curring in later works. The vo- cabulary of the Philebus is poor but very charac- 94 ORIGIN AND GROWTH OF PLATO'S LOGIC had collected. All later inquiries have confirmed this connection and removed the anomalies which Camp- bell’s statistical table still offered. Had Campbell relied blindly on numbers alone, he would have concluded according to the evidence afforded by his observations that the Parmenides is one of the earliest works of Plato, as Schleiermacher imagined ; that the Theaetetws belongs, as Zeller thinks, to about the same period as the Protagoras ; and that finally the Phaedrus is later than the Philebus. These natural errors he happily avoided and this gives to his work a methodic value far above everything done after him in the study of Plato’s style, since later in- quirers frequently discredited their method by unjusti- fiable generalisations from a single occurrence of a single expression in a small dialogue, as for instance of r/ yyy in the Lysis. The Parmenides has a poor vocabulary, but it contains, as Campbell has shown, some highly characteristic’ words (Introduction, pp. xxv-xxx compared with Ast’s Lexicon as to the number of occurrences). 24. yevos as @ logical term: Phaedr. 1 Parm. 8 Soph. 4 Polit. 1 Phil. 2 Tim. 7 Legg. 1. 25. decuds, as @ bond uniting ideas: Parm. Soph. Polit. Phil. Tim. Legg. (This special meaning has not been distinguished by Ast, and Campbell does not give the number of occurrences.) 26. péOeés: Parm. 8 Soph. 2 (A). 27. pepi(w: Parm. 4 Soph. 1 Polit. 2 Tim. 3. 28. woAdcds: Parm. 1 Polit. 1 Tim. 1. Also the vocabulary of the Philebus, though less rich than that of the Politicus, is quite sufficient to indicate the place of this dialogue. Of words used in the Philebus, Campbell enumerates the following as very characteristic terms common to later dialogues (Intro. pp. xxv-xxx): 29. yéveors, in the sense of production in general: Soph. Polit. Phil. Tim. Legg. (A). 80. cupucéis: Soph. Polit. Phil. Tim. Legg. $1. vAn, in the general sense of matter or in a sense approach- ing this: Polit. Phil. Tim. Critias Legg. $2. oxi{w: Soph. Polit. Phil. Tim. PLATO'S STYLE: CAMPBELL 95 33. Stapepifw: Polit. Phil. Legg. 34-36. duerpos, Suixpiois, odua ( = body in general): Soph. Phil. Tim. Legg. 37-38. cvyxpeors, didbeots : Soph. Phil. Tim. Legg. (A). 89. diaywpifw: Polit. Phil. Tim. 40-41. diadvyi{opuat, érixAny : Soph. Phil. Tim. 42-48. oréANopat, amideiv: Soph. Phil. Legg. 44. dynpews: Polit. Phil. Tim. Legg. 45. mAaros: Polit. Phil. Critias Legg. 46-48. ovyxeadaotpat, evaptOpos, So€oaodia: Soph. Phil. 49-51. mnkis, cvyxpact, xataravw: Polit. Phil. 52. avetXirrw, in Phil. corresponds to aveiArfis in Polit. 53. péxpurep: Soph. 1 Polit. 3 Phil. 1 Tim. 4 Critias 1 Legg. 16. (The number of occurrences for this word was found later by C. Ritter, see p. 59 of his Untersuchungen iiber Plato. In all other dialogues €woa7mep is used, which occurs also concurrently with pexpirep and oftener than this, except Tim. Critias Legg. in which both words occur an equal number of times, according to Ditten- berger.) One glance at these words shows for what kind of notions Plato sought new terms in his later writings. Hight words refer to division and reconstitution of unities (30, 32, 33, 35, 37, 39, 49, 50) which Plato had proclaimed in the Phaedrus (266 B) as a divine art, worthy of the greatest admiration. Four words indicate logical opera- tions (40, 42, 48, 46), six physical and mathematical notions (29, 31, 36, 38, 45, 52). This agrees perfectly with what we know of Plato’s latest investigations. His dream was a general theory of science and classification of human knowledge. Campbell’s study of the vocabulary of the Sophist and Politicus confirms the above enumerated general analogies between these dialogues and the group of the Laws. Striking, indeed, is the number of words used by Plato only in the Laws and in one of these dialogues. The following twenty-six words, first used in the Sophist, recur in the Laws (Intro. pp. XXv-xXxx) : 54. *dyxirpevrds in Soph. corresponds to “dyxiorpeia in Legg. 55. *évvypobnpixds in Soph. corresponds to *éyvypoOnpeurns in Legg. teristic, as it contains a great number of logical terms re- curring in Timaeus or Laws. This shows the influence of logical studies. The Sophist and Politicus are still richer in terms proper to Plato's later style, denoting 96 ORIGIN AND GROWTH OF PLATO'S LOGIC dialecti- 56-58. * voudernrixés, * cvvoporoyia, * cvvd:arova: Soph. Legg. eal, physi- 59-68. dywuoriKy, Braorixds, eixaotiKn, eipwvixds, Pappaxorogia. cal, and 64. dpwbeurixcs in Soph. corresponds to épyevrys in Legg. 65-68. Terms expressing logical operations: didyvwcis, pepis, Tapwovuwov, mporxowova: Soph. Legg. 69-78. Poetical words: ad@qos, drderos, Eévios, mapappoovry (in Soph. corresponds to mapadpov in Legg.), rAaoras: Soph. Legg. 74-79. Compounds and derivatives: axpar/s, dvdoraros, adep- pnvevo, oxorodivia, roAunpds, picOwors (A): Soph. Legg. mathe- matical notions. The following forty-three words occur in the Politicus and in the Laws (Intro. pp. xxv-xxix) : 80-84. duurrnpwos, maiyvov, mrexrinds, cxerracpa (A), craciactiKis (in Polit. corresponds to oraciwreia in Legg.) : Polit. Legg. 85-89. Dialectical terms: dropepi{w, droayxifw, éxxpivw (Exkptros Legg.), érivéuw, + yrmpiors: Polit. Legg. 90-100. Physical and mathematical: dvarodyn, ddects (A), yupvacrns, * Spvoropixn (in Polit. corresponds to * 8pvorouia in Legg.), émuorxevafopat, éemomevdo, pérpnots, petpnrds, *auurodnyodpa (in Polit. corresponds to mrodnyetv in Legg.), trepoxn, tpn: Polit. Legg. 101. aéedrns: Polit. Legg. 102-108. Poetical: dvrdgwos, yetrrovd, edovupos, jovyaios, xpyris, avvdpopuos (A), cuvrpodos (A): Polit. Legg. 109-122. Compounds and derivatives: d@uvAaxros, €yxatpos, exdoots, éuropevrixds (in Polit. corresponds to éumropevona in Legg.), evdaBns, iraudrns (in Polit. corresponds to irayds in Legg.), pnvurns, povapxia, mpoopiyyupt, mpoorvyns, ovyxatackevdtw, npeuaios (A), Gupavrctv (A), vonobernua (A): Polit. Legg. Abund- The following are found only in the Sophist or anceof Politicus, and in the Timaeus or Critias : words 128-127. Dialectical: xaBaprads (A), doxtoros (A), StaxpiBoro- borrowed youpas, mpoopodcyovpat, dtapavw: Soph. Tim. from the 128-180. Physical: dinOeiv (A), etxvedos (used first in a quota- poets and tion from the philosopher Parmenides), icoradés (also from Par- unusual roenides) : Soph. Tim. com- 181-182. Poetical: d:amepdw, xpudaios: Soph. Tim. pounds. 188-134. Compounds or derivatives: peOnuepivds, ra havnberra : Soph. Tim. 186-141. diadvrucds, caraxdopnots, carabpava, rapdAXakis, cupmira, ouvudaiva, cvAAayxavw: Polit. Tim. 142. dvaxvxAnors (in Polit. corresponds to dvaxvxAotpa in Tim.). 1438-144. diopiopuds (A), ovvarepyalerOa (A): Polit. Tim. 145-146. * xuxAnors, * pagpy: Polit. Tim. 147-148. Bpdxos, rnxrds: Soph. Critias. 149. dcadayxdvo: Polit. Critias. PLATO'S STYLE: CAMPBELL 97 Many words enumerated by Campbell are not limited to two dialogues, beiny in different ways characteristic of later style: 150-155. * dcaropa, * Srnmep, xupros, mapadopos (in Soph. Legg. corresponds t0 mapagopérns in Tim.), cupduns, xepoaios: Soph. Tim. Legg. 156-158. ¢revyopuat, dytos, yaderorns: Soph. Critias Legg. 159-165. * d:avdnots, drAavys, mpoBodn, Tporn (as an astronomical term), deomdris, tpurdovs, mapmav (A): Polit. Tim. Legg. 166-167. aypduparos, oréyacua: Polit. Tim. Critias. 168-170. ovvodos (A), gaivw, edrperns: Soph. Polit. Legg. 171. cuveéropar: Soph. Tim. Critias Legg. 172-178. cvvvopos (A), meptreirw: Polit. Tim. Critias Legg. 174-176. évvdpos (A), rou (metaphorical), rAéyya: Soph. Polit. Tim. Legg. Some words are limited to Sophist and Politicus only (Intro. pp. xxvi-xxix): 177-181. *dudioByrnrixds, * yvapeurixes, avrom@Ans, amepnuow, ouvrepve : Soph. Polit. Here we have a list of 158 characteristic words observed by Campbell in more than one of the six latest dialogues of Plato, and showing clearly the direction of Plato’s tendency to use rare and new words in his old age. Besides these Campbell enumerates 93 words used by Plato in the Sophist and nowhere else, and 157 words used only in the Politecus. Among these 250 words whose use is limited to a single dialogue, 60 have not been used by any other Greek author (14 in Sophist and 46 in Politicus), and 39 have passed into the language of Aristotle (14 from Sophist and 25 from Politicus). The numerical proportions of all these peculiarities of vocabu- lary may readily be seen from the table based on Camp- bell’s enumerations, given on p. 98. In addition to these, Campbell gives also a list of fourteen words which, without being peculiar to the later dialogues, occur with greater frequency in them than in Plato’s other writings. Among these ¢pafw, arepydlopar, mpoarpsiaOar, hirov, sudavitw, pavralecGar, adrrodacis, H Some of these words are found in more than two dia- logues. Words used only in one dialogue have no chrono- logical im- portance Campbell has thus sufti- ciently proved the ORIGIN AND GROWTH OF PLATO’S LOGIC 98 |} ‘Somskyarwe ‘somsaymyvig ‘sorkdsp ‘sonim0s: 321XdD ‘sedan ‘sosvdaxp ‘sospvdayy -A “mILmgoa ‘Solds ‘somsanys ‘Somhyrs ee ‘sgansidx ‘Sogdogo.sss ‘s9x11.22dg ‘SomLovAds ‘sowrsx0114) cagnoyoxodnanan stinahirtpspyan ‘nadoskd ‘otkddoxs “AI ‘°mRayypsanl “Ty “maropouvvods ‘mrloz0X19 ‘m prios019 ‘I :a0jstay Aq pasn BUIAOT[OJ dy} av snIinjog 344 Wolg (¥ mumyoo) 092Xpano ‘sommewkys ‘spartoaad ‘Somsky goLen ‘sow iktn ‘Somsovyow ‘Sondo1o1 ‘SomLN1291I9 ‘Sor kzvuD ‘SomuMdgaD A “SISDyYDut? “AT "madgigoy TT] “S9LORYH “TT ‘S1OYH 7: OTB | eMowstry fq psn 1978] sssydog at) TOIT SPAOM aT, “SMOTA §,]TaqdmM~ED Jo yNODOe 4901100 B aALB 04 Japi0 Ul paploas Useq SBY STqI ING “JIT “II “I zapun peqnqy]zs1p } 90 WQS" A PUB “AT Japan sploMm a4} JO aMOg ‘NOTWNOUTsSBIO sTAqduB) 07 SuIpi0csoe ‘SUL “Ev ‘Th “AI Jopun {$F ‘ge TIT JopuD ! op Be-Fe “TE ‘NE ‘eg "II Japan : OF ‘ee ‘es ‘ZE “I Jopun T WUNNJOO UT “Tg ‘ZF “AT Javan ‘gg “T]] tepun ! ge ‘Og ‘ee ‘I JOPUN {gp ‘gF SPAOM 944°] JopuN are H UGINPOO UT “9/T-¥ZT ‘OLT-B9T SpIOM 94} ‘A PUB O suTN]OU JO 98Ol[} SePlseq *peqyunoa viv f UTIN[OD UL = “[g1-2Z]1 SplOM pas ‘a ‘¥ sUMINTOO Jo UMS 944 SUTeAUOO D WuINlOO OU, “SZI-SLZ1 ‘Z91-6ST “evi “ObI-Ge1 BpXOM ary A MUMjOo Jo as04} 07 poppe ale A UUINJOo NY (141) 2807 Seno mrz, ‘ydog ut (gg1-991) ‘B8eT SByTID "Gdog Uy “(S91-091) 330] “UNL “qdog UF ‘(SF I-Z4T) seNtID "qdog UF (+ET-EZT) “WEL “qdog Ur ATWO BupLM000 SpIOM 044 A NUINTOO JO 9504} 07 Pappe ere 9 TUMNIOD Ul—suojwunidry TN " * * ) apoastry Aq pesn, * pasnohoy Foxy Suypnouy . “ : * SPIOAM 9181 MOT ITV e e sO UsLIy £ * vaariokay youn baerons : 7 SpiOA MOT IVTIO “A " * egoqstry Aq pesn . *paavighay Fox” Zurpnpouy * — * spumodazoo reunsnuyn “AT * aoystry {q posn spiom Suypnpouy ‘5 5 * * splom TBORVC ‘ITI * —* afoystry Aq pesn spo M ‘ ¢ + waarioday Josep Zuypnpour °° JeoReaaqyear paw peopshqg ‘TT ° apogstry 4q posn spi0M vasrighay youn Suypojouy ! " + Spout TwOTOOTWIG “7 ||| —— | | ——_ | ° a bate pote fe ee | 2] 2| E pl cei eee | wet were =| Pc ei Ae 2 B eo | ee | os 9 ° 5 pie ae | Sr ES a> | gF 8 || RF | RF a RF Re | ES | ono ee sor pus Woy = odog “vou Z| 22 || 23 ie s | 5 2 B/ 2 jj 8 B ono re “SOT ew “UL = HT ce I = ont | oad = ~ | "SttOlgMIAdL aa | | t : | 1 Mou of Ho} oD | « a qj,» | @ Y N n || NI ATIAO OLYId AG GH80 SCHOM NHUVA ANN JO WaGKON PLATO'S STYLE: CAMPBELL, RIDDELL 99 pnbév, mpoapnyua, the indefinite zrorepos, are characteristic of the increasing logical interest, while vepséyw, aepe- AapBave, wetpytixos, pétoyxos illustrate the fondness for compounds and derivatives. The number of stylistic cha- racteristics observed by Campbell in the latest group thus reaches 434, of which twelve are of a genera) character, 255 refer only to Sophist or Politicus, 153 are common to these two with the latest three dialogues (twenty-five to the Philebus with the preceding two groups), and fourteen refer to the increased frequency of words also used in earlier dialogues. Till it be shown that as many peculiarities unite the Sophist, Politicus, Philebus with some other dialogue, we have good reason to follow Campbell in joining them with the group of Timaeus, Critras, and Laws. X. RippELL. At the same time, another editor of another dialogue of Plato undertook an almost equally laborious investigation on the style of Plato, with this difference, that the friend who published it took the pre- caution of mentioning it in the title of the edition. James Riddell,''® late fellow and tutor of Balliol College, Oxford, buried in his edition of the Apology of Plato an appendix of 135 pages under the title Digest of Platonic Idioms. He classifies the idioms used by Plato and quotes examples of all dialogues, but without aiming at complete enumera- tion, and without being aware of the bearing of such stylistic researches on Platonic chronology. ‘Though the Apology has more readers than the Sophist, Riddell’s Digest of Idioms remained almost as unnoticed, at least out of England, as Campbell’s Introduction to the Sophist. As Riddell does not compare the relative fre- quency of each idiom in each dialogue, little can be gained from his enumerations for the chronology, because idioms are less often limited in their occurrence to a few 1s The Apology of Plato, with a revised text and English Notes, and a digest of Platonic idioms, by the Rev. James Riddell, M.A., fellow and tutor of Balliol College, Oxford, 1877 (misprinted for 1867) (xvi). H 2 late date of the Sophist, Politicus, Philebus. Riddell’s Digest of Platonic idioms, however valuable, affords no chrono- logical conclu- sions. but con- firms the authen- ticity of the Sophist and Politicus. Other authors on Plato’s style neglected chrono- logy. 100 ORIGIN AND GROWTH OF PLATO'S LOGIC dialogues than peculiar words. Still at least one very characteristic idiom observed by Riddell is peculiar to the latest dialogues alone : 182. The periphrastic use of the participle, with auxiliary verb substantive (p. 167): Soph. 1 Polit. 4 Tim. 3 Legg. 1. For those who assert with Schaarschmidt that the style of the Sophist, the Politicus, and Philebus is un- Platonic, it may be interesting to learn that Riddell found in the Sophist forty Platonic idioms belonging also to other dialogues whose authenticity is beyond even Schaarschmidt’s suspicions. In the Poltticus he found thirty-six such idioms and in the Philebus forty-five. Few dialogues are as much quoted in the 325 paragraphs of this interesting monograph as the Sophist, Poltticus, Philebus, Timaeus, and Laws. XI.-XII. ScHanz, Lincenserc. Shortly after the labours of Campbell and Riddell, Schanz!!® (1870) wrote on the hypothetical period in Plato, but at that time he, like Lingenberg !* in his dissertation (1874) on metaphors and proverbs in Plato, left the question of chronology out of sight. XIII. Imme. The same indifference to chronological arrangement appears in a dissertation of T. Imme on the forms of interrogation '! (1873) in Plato. This author limited his work to an attempt at classifying interroga- tions psychologically, and quoted for each kind only a few examples, insufficient for chronological inferences. In this case the author’s ignorance of the work of others on the same subject has done him much wrong. Had us M. Schanz, Bifurcation der hypothetischen Periode nach Platon, 1870 (xvii). 120 W. Lingenberg, Platonische Bilder und Sprichworter, Kéln, without date, but published 1874 (xviii). The author enumerates proverbs on God, men, products of human activity, proper names, uses and customs, and literary proverbs. #21 Th. Imme Culmensis (of Chetmno), De enuntiationum interroga- tivarum natura generibusque psychologorum rationibus aique usu maxime platonico illustratis, doctor. dissert. Lipsiae 1878 (xix). PLATO'S STYLE: BLASS, ROEPER 101 Imme known the dissertations of Martinius, he might have made an instructive and interesting addition to our knowledge of Plato’s style. But he quotes only examples of each kind of interrogation without aiming at an exhaus- tive enumeration. XIV. Brass. Another scholar, F. Blass,'!” the author of the History of Greek Eloquence (1874), made a very curious observation, thereby unexpectedly confirming Campbell’s conclusions, though unaware of Campbell’s work. He remarked that the hiatus is less frequent in the Phaedrus than, for instance, in the Symposiwm, and that it is still more rare in Sophist, Politicus, Philebus, Timaeus, Critias, and Laws, where the hiatus is chiefly limited to very frequent words as xai, «i, 7, 7 or the article, while all kinds of hiatus are frequent in the Republic and earlier works. Blass inferred from this single observation that Soph. Polit. Phil. Tim. Critias Legg. were the latest writings of Plato. XV. RoEPER. When four years later (1878) Roeper!” published his investigation on the dual number in Plato, he knew none of the twenty contributions to the know- ledge of Plato’s style which have been mentioned above. He distinguishes two different uses of the dual in Plato: in earlier writings, the common use as in the current language of the fifth century B.c., and in later writings, at the time when the dual fell into disuse, Plato em- ployed it intentionally to lend a phrase an air of solemnity. This usage is shown by Roeper to be frequent in Soph. Polit. Phil., though not limited to these dialogues. Very characteristic of a time when the use of the dual began to be abandoned is : 188. dvoiy with the plural of a substantive (p. 26): Prot. 1; Rep. 1; Soph. 1 Polit. 1 Phil. 1 Legg. 2 (Prot. 855 B and Rep. 546c — Se ee ee 12 F, Blass, Attische Beredsamkeit, vol. ii. p. 426, Leipzig 1874 (xx) ; also on Dittenberger in Bursians Jahresbericht, vol. xxxiii. p. 234, for 1883. 123 Augustus Roeper, Gedanensis, De dualis usu Platonico (doctor’s dis- sertation univers. Bonn), Gedani 1878 (xxi). Roeper distin- guished @ solemn and inten- tional use of the dual from the primi- tive com- mon use. Muny uses of dual forms are either limited to the latest proup or increasing in fre- quency. Teich- miiller believed, like 102 ORIGIN AND GROWTH OF PLATO'S LOGIC are held doubtful by Roeper, but these passages must be counted on the authority of the best MS.). Other peculiarities of later style observed by Roeper, but not singled out as such by him, are: 184. Article raiy (p. 17): Polit. 1 Tim. 1 Legg. 8. 185. ra dvo without substantive (p. 25): Soph. 2 Polit. 1 Legg. 2 (generally in other passages ra 8vo). 186. vov (p. 16): Symp. 1; Phaedr. 1 Theaet. 1; Soph. 1 Polit. 3 Phil. 4 Legg. 2. 187. Adjectives and participles in -av (p. 5): Rep. 1; Soph. 1 Polit. 1 Tim. 1 Legg. 8. 188. réyva as dual of réyyy (p. 5): Rep. 2; Soph. 1 Polit. 1 Legg. 1. 189. Subst. in -av (p. 6): Rep. 2; Parm. 1 Polit. 2 Legg. 6. 190. Dual of substantives neutr. in -n (p. 12): Rep. 2 Phaedr. 1 Theaet. 1; Soph. 4 Polit. 1 Phil. 1 Tim. 2 Legg. 2 (counting only indubitable dual forins; besides these Roeper quotes many passages in which such forms may be either plural or dual, occurring chiefly in Soph. Polit. Legg.). 191. Dual in @ (p. 3) ‘nominum, quorum etiam in o formas licebat praeferre’: Symp. 1; Rep. 2 Theaet. 1; Polit. 2 Legg. 1. 192. Dual rovrw gener. communis (p. 4): Rep. 2 Phaedr. 1 Theaet. 3; Phil. 1 Legg. 2. Similar to this are also zoiw Theaet. 175 c and povw Legg. 777 c. 198. Dual of nouns in -ou II decl. with augoty (p. 11): Prot. 1; Rep. 2; Parm. 1 Polit. 1 Phil. 1 Critias 1. 194. odor (p. 16): Kuthyd. 4; Theaet. 1; Legg. 12. This coincidence between Euthyd. and Legg. Roeper explains by the circumstance that in both dialogues one person is speaking to two others, intimately associated. 195. dvw instead of dvo, according to the best codices, Clarkianus or Parisinus A (p. 20): Rep. 2; Soph. 1 Phil. 1. 196. roty dvoiv (p. 25): Crat. 1; Soph. 2 Polit. 1 Phil. 1 Legg. 2. 197. Dual of verb following plural of subject (p. 30): Euthyd. 2; Rep. 1; Polit. 1 Tim. 2 Legg. 2. 198. Dual of nouns in -uy with dvoty (p. 10): Prot. 1 Meno 2 Euthyd. 2 Gorg. 3; Rep. 1; Parm. 3; Tim. 5 Critias 1 Legg. 2. XVI. TEICHMULLER. A counterpart of Schéne’s theory of perfection in style was Teichmiiller’s '* (1879) stylistic test, according to which the dramatic dialogues are written later than the narrated dialogues, because '4 Gustav Teichmiiller, Die Reihenfolge der Platonischen Dialoge, Leipzig 1879 (xxii). PLATO'S STYLE: ROEPER, DITTENBERGER 103 Plato in the Theaetetus (143 c) criticises the form of a narrated dialogue and introduces the dramatic form as more convenient. This easy way of classifying the dialogues according to a single peculiarity of style led Teichmuller to some conclusions as strange as those of Schéne, though less extravagant, because all the later dialogues are dramatic in form, and Plato seems actually to have given up the form of a narrated dialogue in his old age. But the dramatic form cannot be treated as a special invention, and to place with Teichmiiller the Meno, Gorgias, and Cratylus after the Theaetetus is almost as rash as to recognise with Schéne the Timaeus as an earlier work than the Republic. Still Teichmiiller was led by his argument to the correct conclusion that the Sophist, Politecus, Philebus are later than the Republic. XVII. DirTENBERGER. A new method of stylistic research was proposed by Dittenberger '* (1881), who, though knowing none of his predecessors, happily avoided the repetition of work already done, and directed his attention to a subject not yet investigated, namely the relative frequency of synonyms preferred or rejected in Plato’s different works. This effort brought into pro- minence some fresh peculiarities of later style : 199. xa@drep occurs (according to Dittenberger, and for some dialogues according to later corrections of C. Ritter, p. 58): Lach. 1 Meno 1 Euthyd.1 Gorg. 1 Crat. 2 Symp. 2; Rep. 6 Phaedr. 4 Theaet. 2; Soph. 14 Polit. 3@ Phil. 27 Tim. 18 Critias 5 Legg. 148. In all other dialogues Sonep is used instead, and prevails very much over xadrep even in the Republic (212 times against 6 xaddrep), in the Phaedrus (27 against 4 xa@drev), and in the Theaetetus (47 times against 2 xaOdrep). The prevalence of one synonym over another is a peculiarity of style not less remarkable than the total absence or the appearance of some rare word, and Ditten- 12 Dittenberger, ‘Sprachliche Kriterien fiir die Chronologie der pla- tonischen Dialoge’ in Hermes, vol. xvi. p. 821, Berlin 1881 (xxiii). The numbers quoted by Dittenberger have, in some cases, been corrected by C. Ritter, and are given here according to these corrections. Schone, one stylis- tic pecu- liarity to be decisive. Ditten- berger in- troduced the study of pre- valence of one synonym over another, and found that cer- tain words are pecu- liar to one group of dialogues, while other words of the same meaning are used in other works. He com- pared the changes in the style of Plato with those occurring in the 104 ORIGIN AND GROWTH OF PLATO'S LOGIC berger had the great merit of extending the stylistic study to the relative frequency of synonyms; herein he developed independently an idea to which Campbell had alluded in a footnote (p. xxxii) when he quoted fourteen words of increased frequency in the later dialogues. 200. domep is scarcer than xaédrep only in: Soph. 9/14, Polit. 16/84 Phil. 9/27 Tim. 10/18 Critias 2/5 Legg. 24/148. This scarcity of Sorep, a word which is very frequent in all other dialogues of Plato, is certainly one of the most characteristic peculiarities of Plato’s later style, and coincides with the use of péxperep for €worep noticed above (Nr. 58). 201. raya iows: Soph. 2 Polit. 8 Phil. 3 Tim. 1 Legg. 11. 202. ri unv; Rep. 35 Phaedr. 12 Theaet. 13; Parm. 6 Soph. 12 Polit. 20 Phil. 26 Legg. 48. 2038. ye unv: Euthyd. 1 Symp. 1; Rep. 2 Phaedr. 1 Theaet. 1; Parm. 5 Soph. 6 Polit. 8 Phil. 7 Tim. 7 Critias 1 Legg. 265. 204. dAAd... uyv: Symp. 2; Rep. 11 Phaedr. 1 Theaet. 1; Parm. 2 Soph. 2 Polit. 8 Phil. 2 Legg. 2. 205. cai nv: Euthyph. 1 Charm. 2 Lach. $ Prot. 2 Meno 5 Kuthyd. 4 Gorg. 9 Crat. 9 Symp. 9 Phaedo 7 Rep. 44 Phaedr. 3 Theaet. 11 Parm. 28 Soph. 2& Polit. 2 Phil. 20 Tim. 1 Legg. 36. This expression, though occurring in so many earlier dialogues, may nevertheless be counted among the pecu- harities of later style, because it becomes very frequent only in the later dialogues, in which it supplants add pny, preferred to xal nv in earlier writings of Plato. 206. akAd pny is scarcer than xal pv only in: Lach. 2/3 Symp. 2/9; Theaet. 6/11; Soph. 10/24 Polit. 7/24 Phil. 7/20 Tim. 0/1 Legg. 8/86, while in all other dialogues dAda phy prevails over kai pny (except Charm. 2 Meno 5 Crat. 9 Rep. 44 Parm. 25 Critias 0, in which both occur an equal number of times). This relative scarcity of d\\a pny is the more striking inasmuch as the strong prevalence of the shorter xai uyv cannot be accidental. Besides these Dittenberger counted 0052 yy which cannot be looked upon as peculiar to later style. He added to the strength of his conclusions by the observation that uv occurs with increased frequency also in the works of other authors who wrote about the time when Plato was over sixty. As r/ yy in the meaning of an affirmative answer was not used in the Attic dialect, Dittenberger PLATO'S STYLE: DITTENBERGER, JECHT 105 inferred that Plato brought it from Sicily. But the occurrence of ri unv in a work like the Lysis, which in all other respects has the style of earlier dialogues, tells against Dittenberger’s inference. Even granting the Sicilian origin of the expression, there had been, for some years before the death of Socrates, sufficient intercourse between Sicily and Athens to familiarise Plato with ri pny before he visited Sicily himself. His predilection for this formula, apparent in all later works, is a result of his increasing tendency to strong affirmation, because vi pnv ; has the character of a great logical certainty, excluding every doubt: ‘ What else?’ 7.e. ‘ How could it be otherwise ?’ Dittenberger’s article was the first investigation of Plato’s style which attracted the general attention of German philologers, so much so that, of late, the merit of introducing statistics of style as a method for determining the chronology of Plato’s dialogues has been frequently attributed to him. It was a happy circumstance that Dittenberger, in his conclusions from a very small number of observations, committed no greater error than the uncertain assumption that the Lysts came among the dialogues of the second group, between the Symposium and Phaedrus. But he correctly recognised the group of the latest six dialogues, and admitted that the Re- public, Phaedrus, Theaetetus preceding these are later than the Symposium, Phaedo, Cratylus, and all Socratic dialogues. XVIII. Jecut. Since Dittenberger’s publication the subject of the statistics of Plato’s vocabulary has been widely discussed by writers on the chronology of Plato. Blass '” recognised the new method as leading to the surest results, while Zeller opposed it as too superficial. Dittenberger’s pupil Jecht '* (1881) chose as the subject for his doctor’s dissertation the use of 767 in Plato’s 1%6 Ricardus Jecht, De usu particulae 43n im Platonis dialogis qui feruntur (Doctor’s diss. Univ. Halle a. S.), Halis Saxonum 1881 (xxiv). style of other authors. Ditten- berger’s inferences were correct though based on quite in- sufficient evidence. Jecht in- vestigated the use of %8n in Plato’s works, and found some dif- ferences between the various dialogues. But he did not draw the inferences resulting from his observa- tions. dialogues. following uses prevail in later dialogues : 106 ORIGIN AND GROWTI OF PLATO'S LOGIC From his observations it results that the 207. ovx fin; dn... ov“ or ovx...76n; (p 12): Lach. 1 Menol Gorg. 1; Rep. 3 Parm. 4; Soph. 2 Phil. 1 Legg. 1. 208. évrevdev #dn (p. 50) : Theaet. 1 Polit. 1 Tim. 1 Legg. 1. 209. #3n rd (or ra) pera rovro (or ravra) to effect a transition (p. 50): Soph. 1 Phil. 1 Tim. 1. 210. 78n was (p. 8): Euthyd. 1; Rep. 4 Phaedr. 2; Soph. 1 Polit. 6 Phil. 2 Tim. 8 Legg. 6, including also passages, where 48n is separated by other words from mas, gvpras, Evvaras, mamas in their various cases, with or without preposition. 211. mas 7dn (p. 8): Euthyd. 1 Crat.1; Rep. 1; Parm.1 Polit. 1 Phil. 1 Tim. 1 Legg. 2 212. avris dn or an atros (p. 9): Crat. 1 Rep. 8 Theaet. 1 Parm. 3 Phil. 1 Critias 1 Legg. 1, including also such passages where a d¢ or ye separates 757 from avrds. 213. 5n with perfect designing an action terminated only in the present time (p. 21), with the meaning of ‘by this time’ (nunmehr): Rep. 2 Soph. 1 Phil. 1 Tim. 1 Legg. 1. 214. viv fdn (p. 44): Phaedo 1 Rep. 2 Phaedr.1; Soph. 1 Phil. 2 Tim. 1 Legg. 2 (f3n viv does not occur). 215. vir...43n separated by one or more words (p. 45): Charm. 1 Prot. 1 Meno 1; Rep. 1 Phaedr. 1 Theaet. 1; Soph. 2 Polit. 2 Phil. 1 Legg. 4. 216. rér’ 73n meaning ‘then already’ (damals bereits, p. 46): Rep. 2 Phaedr. 1; Parm. 1 Soph. 1 Polit. 1 Critias 1 Legg. 3, while in some earlier passages as Lach. 181 p, Gorg. 527 p, Phaedo 87 E the meaning is ‘not until then’ (‘dann erst’=tum demum), which meaning occurs also in Theaet. and Legg. This difference of meaning, similar to the difference appearing in the use of ovrws 43n (see below Nr. 220), is very characteristic. Impatient youth complains that things were ‘not done until then’ (Fr. enjin); resigned old age is fain to be content that they are ‘done so soon,’ or ‘already’ (Fr. dé7a). 217. ror’ 8m in apodosi (p. 46): Lach. 2 Prot. 1; Rep. 2 Phaedr. 1 Theaet. 1; Tim. 1 Critias 1 Legg. 1. 218. 73n between a participle and an adjective belonging to it (p. 4): Rep. 1 Theaet. 1 Legg. 2. 219. pera ravro 78 (p. 9): Rep. 2 Tim. 1 Legg. 8. 220. ov'rws 7dn (p. 9): Crat. 1 Symp. 3 Phaedo 2; Rep. 1; Parm. 1 Polit. 1 Tim. 1, including one instance of odrws dv #3n in Parm. 145 c. It is important to notice that in the passages of Crat. Symp. Phaedo the meaning is ‘then’ or ‘not until then’ (‘dann erst’), while beginning with the Republic the four later passages are best translated by ‘thus already ’ (so bereits), which is parallel to the use of rér’ 7én. 107 221. 78n xat (p. 18): Charm. 2 Prot. 2 Meno 1 Crat. 1 Symp. 1 Phaedo 1; Rep. 8; Parm.1 Soph. 2 Polit. 1 Phil. 1 Critias 1 Legg. 8. 222. #5 with plusquam-perfectum (p. 21): Euthyph. 1 Prot. 1 Crat. 1; Rep. 1; Polit. 2 Tim. 2 Legg. 1. PLATO'S STYLE: JECHT, HOEFER XIX.-XX. FREDERKING and Hoxrrer. Dittenberger’s article aroused opposition. In order to show that statistics of particles are at times inconsistent, Frederking of Dorpat undertook (1882) to count how many times Plato used te and some other words.'*” He counted roughly and failed to distinguish the various uses of te. Hence his investiga- tion loses all importance, all the more that the counting has been better done by Hoefer! (1882), who also studied the use of te and some other particles, adding to the stock of peculiarities distinctive of Plato’s later style. Hoefer, as his dissertation shows, knew none of his predecessors save Dittenberger, though he occasionally quotes Camp- bell’s emendations of the Sophist and Poltticus, probably from the original edition. Obviously he had not read Campbell’s Introduction, yet he perceived the importance of stylistic studies for Platonic chronology. Moreover, he recognised that his observations were too few to allow of definite conclusions as to the order of the dialogues, wherein he has shown greater caution than some other authors. His careful and complete enumerations yield the following data : 223. rovyapov (p. 40): Soph. 8 Legg. 2, while in some earlier dialogues rocydproe is used instead, occurring Lach. 1 Euthyd. 1 Gorg. 2 Symp. 1 Phaedo 1; Kep. 8 Theaet. 1, and never later. Hoefer points out that Thucydides always used rocydpro: and never rotyapovr, while in Aristotle only the second form is used. Isocrates, Xenophon, and Demosthenes use both. 224. xai pny ovde (p. 40): Rep. 2; Parm. 1 Soph. 1 Polit. 1 Legg. 1. 225. ydp...dn separated by a verb (p. 25): Parm. 1 Legg. 2. 226. udv...re (p. 17): Rep. 1 Phaedr. 1; Tim. 1 Legg. 2. 227. re used after a single word (not a sentence), adding a third object after two enumerated (p. 9): Rep. 3 Theaet. 3; Polit. 1 Tim. 9. 127 A. Frederking, ‘Sprachliche Kriterien fiir die Chronologie der pla- tonischen Dialoge,’ in Jahrbucher fir classische Philologie, 28** Jahrgang, p. 584, 1882 (xxv). 198 Hermann Hoefer, De particulis platonicis capita selecta (Doctor's diss. Univers. Bonn), Bonn 1882 (xxvi). Freder- king’s objections over- thrown by a more exact inquiry of Hoefer, who inde- pendently, and with- out know- ing Fre- derking, counted the same particles, and found some uses of re and vo limited to the same dialogues in which Kabdwep prevails over Sowep. Even the simple re, which at first sight occurs in- differently in early and late dialogues, may be used for chrono- logical conclu- sions if some dis- tinctions are made between the dif- ferent uses. 108 ORIGIN AND GROWTH OF PLATO'S LOGIC 228. re...re(p. 11): Charm. 1 Gorg. 1 Crat. 1 Symp. 2 Phaedo 2; Rep. 35 Phaedr. 12 Theaet. 5; Parm. 1 Soph. 3 Polit. 3 Phil. 2 Tim. 11 Critias 1 Legg. 50. 229. re...re connecting single words, not phrases (p. 11): Rep. 5 Phaedr. 5; Polit. 1 Tim. 3 Critias 1 Legg.16. Here we see how by distinguishing the various uses of a word the affinity of dialogues belonging to the later time is made evident, even if at first sight a word’s use is not limited to them. This becomes still more instructive by the following distinction : 280. re... re connecting two words not separated by any other part of the phrase, as in Tim. 87 E: ro r’ Wv Td 1’ éora or Critias 121 B: mayxadoi re paxapsoi re (p. 12): Tim. 1 Critias 1 Legg. 1. In this way sometimes an expression which at first sight appears not to be peculiar to a group of dialogues, may by subsequent distinc- tions be used to characterise several groups. According to Frederking re...re was used indistinguishably in early and late dialogues, while according to the above distinctions established by Hoefer one particular use is limited to the Republic and dialogues later than the Republic, while another particular use exists only in the three latest dialogues, Timaeus, Critias, Laws. 281. The simple re, whose frequent occurrence according to Frederking gave no chronological indications, is also shown by Hoefer to furnish some chronological distinctions. It occurs (pp. 5-6) : Apol. 1 Crito 1 Charm. 2 Prot. 1 Gorg. 1 Symp. 1 Phaedo 1 Rep. 25 Phaedr. 23 Theaet.6; Parm. 2 Soph. 3 Polit. 6 Phil. 1 Tim. 198 Critias 27 Legg. 158. It results that it is used more than twice only in Rep. Phaedr. Theaet. Soph. Polit. Tim. Critias Legg., and more than twice in every five pages only in Tim. Critias Legg. This word appears to have two epochs of greatest frequency, the proportion being in Rep. 13 times to 100 pp. (ed. Didot), rising in Phaedr. to 54 times in 100 pp., rapidly declining in the later dialogues until in Phil. it occurred only once (corresponding to a proportion of 2 in 100 pp.), to rise again to a maximum of 878 times in 100 pp. in Tim., 245 times in 100 pp. in Critias, and to decline once more in the Laws to 65 times in 100 pp. There is no reason whatever to doubt that Plato might have twice increased and then diminished the use of a word. re being frequent in all books of the Laws, it tells against C. Ritter’s opinion that the Philebus was written at the same time as the earlier books of the Laws. Although no positive chronological inferences can be drawn from a single stylistic peculiarity, we may doubt whether Plato avoided almost completely in one work the use of a word frequently used by him at the same time in another work, especially as the use of this word is entirely independent of the contents. But such observations are never decisive so long as they remain isolated. If some other equally important stylistic differences between Phil. and Legg. are found, PLATO'S STYLE: HOEFER, PEIPERS 109 then only the presently observed difference will acquire its full value. 282. re connecting phrases, placed immediately after the verb (p. 7): Crito 1 Rep. 3 Phaedr. 1; Parm. 1 Phil. 1 Tim. 8 Critias 1 Legg. 5. 233. re adding a third phrase to two preceding phrases, which are united by kai, re, re... kui, pev...7e, or wey... 8 (p. 7): Phaedr.3 Tim. 5 Critias 1 Legg. 9. Hoefer (p. 7) quotes also two other cases of re peculiar to Timaeus and Laws only, too special for inclusion in our list, but very instructive as samples of acute distinction in stylistic statistics, showing the close relation between these two dialogues. 234. re used avaxodovbws (p. 138): Gorg. 1 Phaedo 2 Rep. 4 Phaedr. 2 Theaet. 1 Tim. 1. 235. re... xai...d€ (p. 15): Critias 1 (118 p) Legg. 1 (708 a). Other particles investigated by Hoefer, as ydp, rot, trou, 5, péevrot, and their various combinations are more characteristic of the earlier than of the later style. XXI. PreIPeRS. Following closely upon these statistics of the use of particles appeared the first special work con- cerning an important part of Plato’s terminology, the use of the words év and ovata. This philological inquiry is contained in Peipers’ * Platonic ontology (1883), and exceeds in volume all preceding treatises on Plato’s style. Of his predecessors, Peipers only knew Dittenberger, though he quotes Campbell’s commentary to the Sophist, which he used without examining the Introduction. He observed some differences in the use of the terms investi- gated, but did not build on such stylistic tests any chronological conclusions, while he correctly inferred the very late date of Parm. Soph. Polit. Phil. from their philosophical contents. From his exhaustive enumera- tions it results that many expressions may be looked upon as peculiarities of later style. 236. dvrws dy, in the meaning of metaphysical being, or ovcia dvrws, in the same meaning, generally d»rws as a metaphysical term, are found by Peipers (pp. 30-81, 514, 540) in: Rep. 8 Phaedr.3; Soph. 8 Polit. 7 Phil. 3 Tim. 8 Legg. 8. 237. o’cia meaning ‘aliquid totum et absolutum, rebus nas- centibus et incrementa capientibus oppositum ’ (pp. 88-108, 515), which is a mixed substance between ideal and material being (of mépas and drepov, au¢piorov and peptordy, ravrdéy and Odrepov). This Peipers found some pe- culiarities of Plato’s later style, though style was not the object of his study. He classi- fied the various meanings of the words dy and otcla, and found certain meanings of these terms very frequent in the latest works. Peipers’ distinc- tions obscure ; his work should be repeated from the stand- 110 ORIGIN AND GROWTH OF PLATO'S LOGIC notion is, according to Peipers, very near to the Aristotelian con- ception of substance, and is found only in Phil. 8 Tim. 2 Legg. 2. 238. ovcia=complexus omnium rerum, quas entium nomine appellare homines solent (pp. 28-29 and 512): Rep. 1 (486 a) Soph. 1 (261 =) Tim. 2 (85 a, 87). 289. 6 gore ‘pp. 88-41 and 541): Crat. 2 Symp. 1 Phaedo 7; Rep. 8 Phaedr. 1 (247 ©); Parm. 9 Tim. 1 (89 Ff). 240. dvrws nai ddAnOas (p. 124): Rep. 1 Soph. 1 Phil. 1. 241. dvrws meaning adnOas (pp. 125 and 513): Crat. 1 Rep. 3 Phaedr. 2 Theaet. 1; Soph. 6 Polit. 4 Phil. 11 Tim. 6 Legg. 49. 242. dv or ovcta=res vera, opposita fictitie (pp. 182-152 and 518): Euthyd. 1 (290 c) Gorg. 2 (472 B, 495 a) Symp. 1 (202 a) Phaedo 7; Rep. 9 Theaet. 7; Soph. 13 Polit. 1 Phil. 8 Legg. 25. 243. rd dvy=id quod tam a loci quam a temporis conditionibus liberum, neque nagcitur, neque interit, sed immutabile et constans eodem modo semper se habet, objectum philosophandi (pp. 50 and 514): Crat. 1 (424 a) Rep. 22 Phaedr. 4 Theaet. 1 Soph. 36 Phil. 2 Tim. 2 Legg. 2. 244. ra dvra in the same meaning as above (pp. 63-66) : Crat. 2 Phaedo 2; Rep.5 Phaedr.3 Theaet. 1; Parm. 2 Soph. 5 Phil. 2 Tim. 4. 245. ot’cia = substance as object of knowledge (pp. 67 and 515) : Crat. 9; Rep. 11 Phaedr. 4 Theaet. 8; Parm. 3 Soph. 6 Polit. 3 Tim. 1 Legg. 5. Some isolated passages quoted by Peipers from other dialogues, as Euthyph. 11 a Charm. 168 v p Prot. 849 B Meno 72 B, seem not to belong here, as they offer a different meaning of ovgia, as ‘nature,’ ‘object,’ ‘ property,’ ‘definition.’ 246. rd dv = what exists, opposed to pndév (pp. 11-16 and 512): Euthyd. 8 Crat. 2 Symp. 1 (205 8); Rep. 11 Theaet. 11; Soph. 33 Polit. 1 Phil. 1 Tim. 1 Legg. 5. 247. ovcia = what exists (pp. 17 and 589): Rep. 2 Theaet. 5; Soph. 7 Polit. 1 Tim. 1. 248. ra dvra=ra mpaypara (pp. 19-28, 512, 540): Charm. 3 Meno 8 Euthyd. 6 Gorg. 5 Crat. 28 Symp. 2 Phaedo 9; Rep. 4 Phaedr. 6 Theaet. 8; Parm.5 Soph. 4 Polit.3 Phil.6 Tim. 8 Legg. 6. 249. ro dv= veritas cognitione aut oratione expressa (pp. 222- 280): Euthyd. 4 Gorg. 1 Crat. 2; Rep. 4 Phaedr. 8 Theaet. 5; Parm.1 Soph. 4. Peipers’ distinctions are sometimes obscure, and the numerous quotations collected in his work are not con- veniently arranged. The Laws are treated apart in a few pages towards the end of the work (pp. 512-516). Peipers did not count the passages quoted, nor did he distinguish the number of occurrences in a single passage. His work remains a valuable collection of texts, which calls for a complete digest by some clearer expositor. PLATO’S STYLE: PEIPERS, WEBER 111 His conclusions on the order of dialogues do not precisely correspond to considerations of style. Against the purely statistical evidence, Peipers separates the Phaedrus from the Republic by the Banquet, and puts the Theaetetus later than the Timaeus, following alleged differences of ontological doctrines not easily definable. But he had the great merit of recognising the very late date of Soph. Polit. Phil., as written after the Republic. XXII. P. WEBER. After so many investigations on Plato’s vocabulary, P. Weber '” (1884) returned to the old problem of the construction of phrases in Plato. But he seems to have wholly ignored the relation between the style and the chronology of Plato’s writings, and he neither distinguishes the single dialogues nor enumerates the passages, except when dealing with some very rare stylistic peculiarity. Under these circumstances Weber’s disser- tation is chiefly of interest as contributing to the stylistic definition of Plato’s works as a whole, for comparison with other authors, but containing very few hints for distinc- tions between early and later style: 250. iva with conjunct. ‘nach Nebenzeiten, and referring to a design lasting up to the present time (p. 11): Crito 1 Prot. 2 Meno 1 Crat. 1 Symp. 2; Rep. 1 Theaet. 3 Parm. 1; Tim. 8 Legg. 8. 251. drws with conjunct. ‘nach Hauptzetten, in vollstandigen Finalsatzen’ (p. 18): Symp. 1 Legg. 9. 252. dws with optativ. praes. ‘nach Nebenzetien, in vollstand- agen Finalsatzen’ (p. 14): Prot. 1 Phaedr. 1 Tim. 5. 258. omws dv with conjunct. ‘in vollstandigen und unvoll- standigen Finaledtzen’ (pp.14, 21) : Lach. 1 Prot. 1 Gorg. 6 Symp. 1 Phaedo 1; Rep. 9 Phaedr. 1; Tim. 1 Legg. 22. Weber also gives the number of all occurrences of final sentences with 7, iva, d7ws, ws, with various tenses and moods, but without distinctiqn of single dialogues, so that his work must be repeated if it is to afford chronological distinctions. XXIII. Droste. A marked contrast to both the 128 Dr. Philipp Weber, Der Absichissate bet Plato, Wiirzburg 1884. A Doctor’s dissertation of the university of Wiirzburg. This is the xxviiith publication on this subject, Peipers’ being the xxviith. point of stylistic investiga- tion. P. Weber also ignored the rela- tion be- tween the style and the chro- nology, and gave only ina few in- stances complete enumera- tions of passages, by which omission he de- prived us of many useful in- dications. Droste first dis- tinguished classes of rare words according to the mode of their for- mation, so effecting & progress in the method of stylistic investiga- tions. 112 ORIGIN AND GROWTH OF PLATO'S LOGIC preceding writers as to the clearness of exposition and excellent method of investigation is presented in the dissertation of P. Droste ' of Diisseldorf (1886), who undertook to represent Plato’s use of adjectives terminating in ed7js and #dns. Since Campbell nobody had examined the formation of new rare words by Plato, and Droste knew none of his predecessors ex- cept Dittenberger, yet he unconsciously perfected the Scotch investigator’s method, distinguishing classes of new rare words according to the mode of their formation, and not only according to their meaning or origin. This endows Droste with a merit scarcely dreamed of by him, and manifests at the same time how progress in scientific method may be realised apart from wide knowledge. Droste dissects Plato’s art of word-building under one of its aspects, dealing with words mostly very rare and invented by Plato for the expression of his thoughts against the general usage of his times: of seventy given adjectives, forty-six are never used before (13 in evdys and 33 in ons), and thirty-seven are later accepted by Aristotle (7 in ecdys and 30 in wédns). Droste minutely compared Plato’s use of such adjectives with their employment by earlier and later authors. Before Plato these words were rare, and since Plato they became very common, as is easily seen from the following table : historians : Shae | Number of different ba 7 Thucydid, adjj. in ecéys “etn from Sée 4} prota is re-arranged accordi T together had used a he use of both kinds of adjj. rapidly in 1% P. Droste, De adjectivorum im eaidhs et in b8ns desinentium apud Platonem usu (Doctor's diss. Univ. Marburg), Marpurgi, without date, pub- lished 1886 according to Hinrich’s Catalogue (xxix). PLATO’S STYLE: DROSTE 113 This interesting comparison proves how well chosen was the use of such adjectives, as constituting an important pecu- harity of Plato’s style. The relative occurrence in various dialogues is seen from the following table, constructed from the materials given by Droste, pp. 18-19, 37-41, re- arranged in a more systematic manner than in his tables: \ i| OBbERVATIONS.—All quoted adjectives are used only once b eS Peceassties - Plato, unless the number of occurrences in each dialogue ff adjectives |jof all adjec- shown. Adjectives invented by Plato and used for the first time te cic inat- |itives termi-,| 27° printed in heavy type += not used before Plato; * = not ine in nating in used betore nor after Plato ; A = accepted by Aristotle ; Aesch, = 8 & used by Aeschylus; Eur. = used by Euripides , Her. = used by Herodotus; Xeu. = used by Xenophon; Hom. = used by Humer ; Hes. = used by Hesiod ; Iso. = used by Isocrates -w8ns||-ecd%s|-w8xs eee eo th -e.dys = (1) eveid4s (Aesch. Eur. Her. Xenoph.) occurs Crito 1 Rep. 2, A.—[1] voowdns (Iso.) is found in Plato : more often than any other adjective in dys, | occurring 24 times: Crito 1 Charm. 3 Lach. 1 Symp. 2 Rep. 9 Theaet. 1 Polit. 3 Legg. 3 Alc. I. 1) A. 1} f3) + alvtywartrddns, A, seems to be the first | adjective in #3ns invented by Plato, occurs bo | > Charm. 1 Theaet. 1 and Alc. II. 1. 1] 195 c. ' [3] dyxwdns (Xen.) A. 4| teparddns (Aristophanes) A. i (2) + &ev8%g occurs Gorg. 1 Crat. 1 Phaedo 12, A. —_—— -_ — = SE —e | | 8 2/12). + (2).—(3) * rpayoesS4s—[5] + yAouSns, A | ra] 2 +KoAAd&ns, A,—[7] + pvod&ys, A.— 8] + coxord&ns, A, occurs Crat. 1 Phaedo 1 Rep. 2 egg. 1—[9] * tnpid&ns, A, occurs Crat. 5 Legg. 2--[10] + 8&:OvpapBdsys—[11] Onpicdns (Eur. Xen.), A, occurs Crat.1 Rep. 1 Polit. 1 Tim. 1 Legg. a tia] rvevnarwdns, A. Only these 8 adjec- tives in édys are enumerated, occurring 12 times, while according to Droste’s table 9 different adjec- tives are used in the Cratylus 13 times. (4) + povoec8%s, A, used Symp. 2 Phaedo 3 Rep. 1 Theaet. 1 Tim. 1—[13] eté5ns (Hom.), used Symp. 1 Phaedr. 1 Tim. 1 Critias 1, A.—[14] dySpawodébns (Xen.) used Symp. 1 Phaedo 1 Rep. 1 Phaedr. 1 Legg. 1, A. + (2)—+ (4)—(5) wodveidjs (Thucydides) occurs Phaedo 1 Rep. 8 Phaedr. 3 Soph. 1, A.- (6) Ge0e3hs (Hom. Hes.) occurs Phaedo 1 Rep. 1 Phaedr. 1, Epinomis.—(7) xpvooedis (Xen.) A— (8) croedhs (Aristoph.) A-—(9) + @vnroer8%5— (10) + coparoer84s, used Phaedo 5 Rep. 1 Polit. 1 Tim. 2, A—(14]—[15] wnaddns (Thucyd.) A— [16] + 8nyd8ns: Phaedo 1 Legg. 1—[17] + Bop- | Bopd8ns, A.— [18] + yed8ng: Phaedo 2 Tim. 3 | | Critiag 1, A. 114 ORIGIN AND GROWTH OF PLATO'S LOGIC OBSERVATIONS.—AIl quoted adjectives are used only once by Plato, unless the number of occurrences 1n each dialogue 1s shown. Adjectives invented by Plato and used for the first time are printed in heavy type. + = not used before Plato; * = not used before nor after Plato ; A = accepted by Aristotle; Aesch.= used by Aeschylus; Eur = used by Euripides; Her. = used by | Herodotus ; Xen.= used by Xenophon ; Hom. = used by Homer ; -erdys|-8ns||-erdys!-o8ys;| Hes. = used by Hesiod ; Iso. = used by Isocrates. Number of || Total of different |/occurrences adjectives |iof all adjec- terminat- ||tives termi- ing in nating in In a eaaaamnemmned ae a A 16 || 39 | 28 || (1)—+ (4)—(5)—(6)—+ (10)—(11): @upoesdhs (Xen.) used in the meaning ‘hot-tempered,’ chiefly of restive horses: Rep. 8 Legg. 2, distinguished from the philosophical term + 6vpoer8rg : Rep. 19 Tim.1, A—(12) * dya8oer84g—(13) * hAvoErsHs: Rep. 2, A—[1]—* [8]—[11]—[14]—[19] + @pnva- 8ns Rep. 3 Legg. 1—[20] brvedns (Eur.) A—[21] + dreypaTrdSns, A—[22] + petpaxid8Sns, Rep. 2, A—[23] + &Aurnpidhns, Rep. 1 Legg. 2-—- bea + omnArAard8ns—[25] mvdedns (Isocr.) A -- 26] + endnvdSns—[27] * Acovrd8ys, A—[28] + &hed&s, A—[29] + dxAG8ns—[30] werpatns (Sophocl.) A, in the order of occurrences ; Droste counted 26 instead of 28. Ce net pel as in Phaedo and Symposium. —[3 1 i a y paae, 4 || +(4)—*[2 1] + compdé8ns, A—[32] * AnpdSns, A—[33] + Avéd8ns, A. 1 | [84] + wpayporerd8ns. — | (5)—(14) 8vee:d4s (Sophocl. Her.). ‘| + (10)—[11]—[1]—[85] + xpond8ns. | (15) * weparoer8rys—[36] + wardapiddns, A. | + (4) —* (10) — + (11) — (16) 9 aeaipoesdhs (Xen.) ' Tim, 4, A—(17) + ArOoer8%¢—(18) + depoerdijs, A—(19) + cnpoer8is-- (20) + capKoerdris, A— (21) * orepeoesd4s—[11]—[13]—* [18]—[37] dans —[38] capxétns (Her. Xen.) A—[39] xoAddns, Tim. 3, A—[40] + tpavrd8ns—[41] + olorpd8ns, Tim. 1 Legg. 1—[42] + Aurpd8s —[43] * puddns, A—[44] + QopuBd8ns, Tim. 1 Legg. 1, A—[48] +vevpd8ns has not been counted by Droste, though it is quoted p. 34; this increases the number of adjectives to 12, of occurrences to 16. 8 || [13]—+* [18]—[45] xupsdns (Aristoph.) A. 18 || (11)-—(22) * wvpoe84s—[1]—+* [8]—* [9]J—[11 , —[14]—+ [16]—+ [19]—+ [23]—+ rays 44 | —[46] + yod8ns5, A—[47] tatpaord8ns. Critias . Laws . In no other dialogue adjectives in ecd:j¢ or &8ys are found, except Epinomis (6) Alo. I. [1] [14] and Alc. II. [2], in each of which occur only 1-2 saleotives used by Plato in aut Date alalooced a4 mn Alo. I, mpewes8ns, taken from Aristophanes, New- The most interesting general result of Droste’s in- invented vestigation is that not one of the spurious dialogues adjectives contains new-invented adjectives in edie or wéns, and PLATO’S STYLE: DROSTE 115 that even those introduced by Plato are used only in four isolated instances in probably spurious dialogues, as Alcibiades I. and II. and Epinomis. This shows the originality of vocabulary to be an inimitable pecuharity of Plato’s style, and further increases the improbability of anybody but Plato having written such original works as the Parmenides, Sophist, and Politicus. In these dialectical dialogues adjj. in eds and wédns are scarce, while many new-formed adjectives in sxéds abound; Droste counted 224 such adjectives in the Sophist, and 320 in the Politicus, while only 12 occur in the Phaedo. Droste’s dissertation offers important additions to our list of peculiarities of later style : 254. New-invented adjj. in esdns occur (p. 18): Gorg. 1 Crat. 2 Symp. 2 Phaedo 21; Rep. 24 Theaet. 1; Polit. 1 Phil. 1 Tim. 8 Legg. 1. (These numbers are not given by Droste; they result from the above table.) 255. New-invented adjj. in wdns (pp. 88 and 81-35): Charm. 1 Crat. 10 Phaedo 4; Rep. 14 Theaet.4; Parm. 1 Polit. 1 Phil. 1 Tim. 9 Critias 1 Legg. 11. 256. rroAveidns : Phaedo 1 Rep. $ Phaedr. 3; Soph. 1, A (Table (5), Droste, p. 11). 257. *povoedis: Symp. 2 Phaedo 3; Rep. 1 Theaet.1; Tim. 1, A (Table (4), Droste, p. 11). In these adjectives the primitive meaning of the termination is preserved, though here, too, efdos often means species and not form. This use of adjj. in edys to designate a species corresponds to a logical tendency, as Droste well observed, and was never attempted before Plato. Plato introduced it into the Greek language ‘ex necessitate quadam et ex philosophandi angustiis’ (p. 19). 258. Adjj. in ens designating a species (p. 14) : Phaedo ( (6) (8) (9) (10) ) & Rep. ( (10) (11) (12) (18) ) 23 Polit. ( (10) ) 1 Phil. ( (15) ) 1 Tim. ( (10) (11) (18) (19) (20) (21) ) 7 Legg. ( (22) )1. Among these adjectives some are specially character- istic : 259. +cwparoedns (p. 15): Phaedo 5 Rep. 1 Polit. 1 Tim. 2, A, 12 in e:dhs and dns do not occur in spurious dialogues, while they are fre- quent in authentic works. Plato first intro- duced the use of such ad- jectives to designate & species, and he used them frequently also in other meanings well defined by Droste. 116 ORIGIN AND GROWTH OF PLATO’S LOGIC 260. dvyoedns (p. 16), in the same meaning as in Xenophon: Rep. 8 Legg. 2 (see table (11) ). 261. tOvpoedns (p. 15), as philosophical term, used also later by A.: Rep. 19 Tim. 1. 262. Adjectives in «dys or dns designating form or colour (including dedns) (pp. 10, 18-14, 81): Crito ( (1) ) 1 Gorg. ( (2) ) 1 Crat. ( (2) (3) ) 2 Phaedo ( (2) (7) ) 13 Rep. ¢ (1) (6) )8 Phaedr. ( (6) ) 1 Soph. ( (14) ) 1 Tim. ( (16) (17) [89] ) 7. (Geoedns is used in this meaning only Rep. 501 8 Phaedr. 251 a, while in Phaedo 95 c and Epinomis it designates a species.) More frequent are the adjectives in dns, which are classified by Droste according to their meaning. Those derived. from éfw form one class, containing only evwdns and awdns, of which the second is used only once (Tim. 50 £). 268. evadns : Symp. 1 Phaedr. 1 Tim. I Critias 1, A (Droste, p. 81, table [13] ). 264. Adjectives in éns designating similarity (pp. 81-32): Crat. ((10] [11]) 2 Phaedo ( [16] ) 1 Rep. ( [11] [14] [22] [26-30] ) 9 Phaedr. ([{14]) 1 Theaet. ({81] [83]) 2 Polit. ({11]) 1 Phil. ({86]) 1 Tim. ( [40] [41]) 2 Critias ([45}) 1 Legg. ([11] [16] [41]) 5. Among these the following are characteristic : 265. 6npiwdns: Crat. 1 Rep. 1 Polit. 1 Tim. 1 Legg. 3.—A. (in Tim. 91 £ it designates a species, while in Legg. 909 a it means ‘like brutes,’ and in other passages, as Rep. 571 ©, Legg. 906 B, it has a similar meaning). 266. *dnueddns: Phaedo 1 Legg. 1. 267. avdparodadns (p. 32): Symp. 1 Phaedo 1 Rep. 1 Phaedr. 1 Legg. 1 (in Symp. 215 = and Legg. 880 a it designates a species. Droste omitted Phaedo 69 B, where it means similarity). 268. *oiarpwédns: Tim. 1 Legg. 1. 269. Adjectives in odns designating a species (pp. 32-33) : Crito ({1]) 1 Charm. ([1]) [2}]) 4 Lach. ([{1]) 1 Crat. ([5-8] [12]) 5 Symp. ([1] [14]) 2 Phaedo ([{17] [18]) 3 Rep. ( [1] [24] 6 Theaet. ({1] [2]) 2 Polit. ([1} [85]) 4 Tim. ([{11]} [18] [42] [48}) 5 Critias ({18}) 1 Legg. ([14] [47]) 2. 270. tvorwédns, designating a species: Crito 1 Charm 3 Lach. 1 Symp. 1 Rep. 5 Theaet. 1 Polit. 8, A. This meaning, as for instance Rep. 488 &, is different from the following : 271. vorwdns: meaning sickly, diseased, opposed to vyrecvds : Symp. 1 Rep. 4 Legg. 8, A. (Droste omitted Rep. 556 &, and quotes therefore only three passages in Rep.) 272. yeodns: Phaedo 2 Tim. 8 Critias 1, A (in Tim. 66 B it does not designate a species, but local connection). PLATO'S STYLE: DROSTE, KUGLER 117 278. oxorwdns: Crat. 1 Phaedo 1 Rep. 2 Legg. 1, A (of these only in Crat. 412 B is a species designated, while the other passages use that word in the meaning called by Droste ‘of local connection,’ as ‘ full of darkness’). 274. Adjectives in dns indicating local connection (p. 84) meaning ‘full of . . .’: Meno ({3]) 1 Euthyd. ([4)) 1 Crat. ((8]) 1 Symp. ([1]) 1 Phaedo ([8] [15]) 2 Rep. ([1] [8] (25]) 7 Theaet. ({82]) 1 Parm. ([84]) 1 Tim. ( [18] [38] [89] [48]) 5 Legg. ([1] [8] [19] [46}) 6. This use is distinguished by Droste from the pre- ceding, and also from the following, as may be seen by comparing the meaning of O6pyvwdns in Legg. 792 B (274) and Rep. 398 £ (275), of yoAwdns in Tim. 86 (274) and Tim. 71 8, 88 B (262). 275. Adjectives in wédns denoting causal relations (p. 84): Crat. [9] 5 Rep. [19, 20, 21, 23] 6 Tim. [44] (42p) 1 Legg. [9] (650 a, 690 £) [23] (854 B, 881 £) [44] (671 4) 5. Among these the following occur in more than one dialogue : 276. ddurnpiwdns (p. 34): Rep. 1 Legg. 2. 277. Cyurmdns (p. 84): Crat. 5 Legg. 2 (Droste omitted Crat. 418 a, B). 278. OopuBodns (p. 85): Tim. 1 Legg. 1. Droste’s dissertation 1s a model of stylistic investiga- tion made for the purposes of Platonic chronology. We see that in the above enumeration the Phaedo very fre- quently occurs together with later works, and Droste in- ferred that the Phaedo was written after the Phaedrus. But this cannot be decided without considering many other peculiarities of vocabulary and style, besides the adjectives investigated by Droste ; it will then appear that the Phaedrus is much nearer to the Republic as well as to the latest six dialogues than the Phaedo, though in some respects the Phaedo may approach the style of the Republic more nearly than does the Phaedrus. The natural explanation is that the Phaedo immediately preceded the Republic, while the Phaedrus followed it. XXIV. F. Kuauer. A dissertation published at the same time as Droste’s, by F. Kugler,'*! of Basel, on roi and its compounds, shows also certain analogies between 131 F, Kugler, Dissertatio inauguralhs de particulae rol ejusque com- positorum apud Platonem usu (Doct. diss. Univ. Basel), Trogen 1886 (xxx). Droste’s disserta- tion a model of method, though one of his con- clusions is erroneous. The Phaedo not later than the Phaedrus, as Droste believed. Kugler found many uses 118 ORIGIN AND GROWTH OF PLATO'S LOGIC of row the Phaedrus and the latest group which are lacking in prevailing the Phaedo, and many others between the Phaedo and the he Republic. sisiade 279. pévroe used to oppose to each other two parts of the same Broup 5 phrase (p. 26) :jgProt. 4 Meno 1 Euthyd. 1 Gorg. 2 Symp. 1 especially © Phaedo 2; Rep. 4)Phaedr. 2 Theaet. 1; Parm. 1 Polit.1 Phil. 1, in- the syllo- cluding also some cases of opposition by means of ov pévro, and gistic use rs) peévrot. in conclu- 280. ye . . . wévrot (p. 27): Crito 1 Euthyd. 1 Gorg. 2; Rep. 8 sions, Phaedr. 1 Theaet. 4; Soph. 1 Phil. 1 Legg. 1. while 281. ro. between article and substantive (p. 7): Symp. 1 wdvrot Theaet. 1 Soph. 1 Phil. 1. hacamie ; ee rot after the verb (p. 7): Gorg. 1 Phaedo 1; Theaet. 1 oph. 1. pe 288. xairo. = et vero (pp. 17-18): Gorg. 2; Rep. 1 Theaet. 1; Phil. 1 Legg. 3. 284. roivvy in the conclusion of a syllogism or of a similar argument (p 82): Crito 1 (444) Charm. 2 (1628 syll.) Meno 2 Gorg. 4 Crat. 1 (482p syll.) Phaedo 3 (62cayll.) Rep. 18 (868 £, 6084 syll.) Phaedr. 4 Theaet. 1 (192 syll.) Soph. 8 Polit. 4 Phil. 10;(including three syll. 88 , 41 p, 56. c) Legg. 14. This increasing use of a word which was afterwards so much used by Aristotle in logical conclusions is very characteristic of the progress made by Plato in his logical terminology. 285. roivuy érs in transitions (p. 84) : Charm. 1 Phaedo 1; Soph. 3 Polit. 2 Phil. 2 (the form ér: roivvy is much more often used). 286. ére 37 roivuy: Phil. 1 (52.4) Legg. 1 (817 £). 287. xai roivuy (p. 84): Soph. 2 Polit. 1 Legg. 3 (while cai... roivuy was used earlier, in Charm. 1 Gorg. 1 Rep. 4 Theaet. 1 and © also in Phil. 1). 288. mpa@rov pev roivuy (p. 85): Crat. 1 (4260) Phaedo 1 (90D) Rep. 1 Polit. 1 Phil. 2 Legg. 8. 289. roivuy begins a new argument (p. 85): Apol. 1 Euthyph. 1 Crito 1 Charm. 3 Gorg. 1 Crat. 9 Symp. 1 Phaedo 6; Rep. 18 Phaedr. 6 Theaet. 6; Parm. 1 Soph. 10 Polit. 13 Phil. 9 Legg. 21. 290. roivvy in transitions (p. 85): Crito 1 Crat. 9 Symp. 1 Phaedo 1 ;3Rep. 14 Theaet. 4; Soph. 4 Polit. 1 Phil. 4 Legg. 9. 391. 3) roivvy (p. 86): Rep. 1 Phil. 1 Legg. 5. 292. roivuy 87: Gorg. 1 Legg. 1. 293. 73n roivvy (p. 86): Meno 1 Crat. 1; Polit. 1 Phil. 1 Legg. 1. 294. wy rolvywy (p. 86): Crito 1 Charm. 1 Lach. 1 Meno 1 Symp. 1 Phaedo 1; Rep. 4 Theaet. 8; Soph. 7 Polit. 1 Phil. 2 Legg. 6. PLATO'S STYLE: KUGLER 119 295. ov—roivur (p. 36): Soph. 1 Legg. 1. 296. roivuy, instead of being the second word of the phrase as usual, is placed in the third place or further (p. 36): Apol. 1 Kuthyph. 1 Charm. 1 Meno 1 Euthyd. 1 Crat. 1 Phaedo 1; Rep. 10 Phaedr. 1 Theaet. 1; Soph. 5 Polit. 7 Phil. 8 Legg. 8. 297. ws 3n rou (p. 12), beginning an evident conclusion: Rep. 1 Phaedr. 1 Tim. 1. 298. cairo... . dé or dpws S€ (p. 19): Apol. 1 Lach. 1 Meno 1 Kuthyd. 1 Gorg. 1; Rep. 3 Phaedr. 2; Parm. 1 Phil. 1 Critias 1 Legg. 3. 299. ddnO7y perros (in affirmative answers, p. 23): Lach. 1 Rep. 1 Soph. 1 Legg. 5. 300. Froo . . . F (p. 14): Prot. 2 Meno 2 Gorg. 2 Crat. 5 Phaedo 2; Rep. 2 Phaedr. 1 Theaet. 3; Parm. 3 Phil. 2 Legg. 2. 301. xairoc... adAd (p. 19): Lach. 3 Gorg. 1 Crat. 1 Phaedo 1; Parm. 1; Polit. 1 Legg. 2. 302. Simple pay (p. 40): Prot. 1 Meno 2 Kuthyd. 2; Rep. 1 Theaet. 1; Soph. 2 Polit. 2 Phil. 1 Legg. 5. 803. yay ody (p. 40): Soph. 2 Polit. 2 Phil. 4 Legg. 10 (includ- ing one ovy par). 304. par od (p. 40): Soph. 8 Polit. 2 Phil. 4 Legg. 10. 305. pay pn (p. 40): Phaedo 1 Rep. 2 Soph. 1 Phil. 1. 306. rocvyy more than four times oftener than pévro: (p. 45) : Soph. 55/13 Polit. 46/7 Phil. 52/8 Legg. 120/17 while in all other works roivyy is much scarcer, occurring in no other dialogue twice as often as pevrot, the proportion to pévroe being in Rep. Phaedr. Theaet. 488, in Euthyphr. Apol. Crito Charm. Lach. Prot. §#, in Meno Euthyd. Gorg. Crat. Symp. Phaedo }$#, in Parm. ¥,, in Tim. Critias 3. It would be unjustifiable to draw any inference from the absence of both particles in Tim. Critias, or from the scarcity of roévvy in Parm. The only conclusion allowed is, that Soph. Polit. Phil. Legg. have the peculiarity in common of an exceptional predominance of roévuy over pevrot. From asingle peculiarity no chronological conclu- sions can be drawn, but this peculiarity, jomned to many others, offers a measure of affinity between the dia- logues in question. 807. yévrot occurs less than once in two pages only in (p. 45) : Crito 2 Prot. 19 Meno 6 Gorg. 23 Symp. 18; Phaedr. 16 Parm. 13; Soph. 13 Polit. 7 Phil. 8 Tim. o Critias o Legg. 17 being less than once in five pages only in Polit. Phil. Tim. Critias Legg. This acquires a special importance if we consider that pévro: went These ob- servations are valu- able, but Kugler attributed too great import- anceto the scarcity of rolyuy in the Par- menides. Schanz indepen- dently confirmed Camp- bell’s con- clusions. Gomperz recognised the con- clusions reached by both Dittenber- ger and Schanz, 120 ORIGIN AND GROWTH OF PLATO'S LOGIC out of frequent use in Plato’s time as Kugler has shown by com- paring other authors, from Xenophon, in whose writings pevros greatly prevails over roivuy, down to Demosthenes, who uses pevroe very rarely. 808. roivwy is very frequent, occurring once in two pages or oftener in : Crito 5 Charm. 20 Lach. 10 Meno 13 Crat. 32; Rep. 133 Theaet. 39; Soph. 55 Polit. 46 Phil. 52 Legg. 120. From these and many other uses of ros Kugler inferred quite correctly that the Sophist, Politicus, Philebus belong to the same period as Timaeus, Critias, Laws. XXV. M. ScHanz. The same conclusion is also reached by Martin Schanz,'!” the editor of Plato, who simultaneously with the dissertations of Kugler and Droste published his article on the development of Plato’s style. Though he quotes Campbell’s emendations to the Sophist in his critical edition of the same dialogue, Schanz seems not to have read Campbell’s Introduction. Directing his attention to expressions designating truth and being, he found : 809. dvrws: Euthyd. 1 Crat. 1; Rep. 9 Phaedr. 6 Theaet. 1; Soph. 22 Polit. 11 Phil. 15 Tim. 8 Legg. 50, while in earlier works r@ dvre is used instead, which is entirely absent from Polit. Phil. Tim. Critias Legg., and occurs but once in Soph. 810. aAnOeia (used instead of r7 aAnOeia) only in Prot. 8 Phil. 1 Tim. 1 Legg. 8. 311. aAnOas (instead of as aAnOes): Apol. 1 Euthyph. 1 Prot. 1 Meno 2 Euthyd. 1 Phaedo 2; Rep. 8 Phaedr. 1 Theaet. 1; Soph. 6 Polit. 4 Phil. 7 Tim. 8 Legg. 6. XXVI. GompERZ. Only these few observations of Schanz, with those of Dittenberger, became generally known to German philologers. They did not convince Zeller, but they were held sufficient for the stylistic definition of the latest group of Plato’s works by another most competent historian of Greek philosophy, Theodor Gomperz '* (1887),of the University of Vienna. He 2 Martin Schanz, ‘Zur Entwickelung des platonischen Stils’ in Hermes, vol. xxvi. pp. 487-459, for 1886 (xxxi). 133 Th. Gomperz, ‘ Platonische Aufsitze,’ in Sitzwngsberichte der Kaiser- lichen Akademie der Wissenschaften su Wien, vol. cxiv. pp. 741-766, Vienna 1887 (xxxii). PLATO'S STYLE: SCHANZ, GOMPERZ 121 repeated Dittenberger’s observations, and insisted on their decisive importance as to the order of the Platonic dia- logues. Gomperz argued that the more or less frequent recurrence of words does not lead to such certain con- clusions as does the complete absence of certain words in certain dialogues; and in this he unconsciously agreed with Campbell, who also had chiefly directed his attention to the presence or absence of certain words in some dialogues. Yet it cannot be denied that observations on the comparative frequency or rarity of words give valu- able confirmation of conclusions obtained from complete changes of vocabulary, and also that the number of words increasing in frequency 1s vastly greater than the number of expressions replaced by synonyms. We have no reason to disdain supplementary evidence on a matter in which, as in other historical problems, even the greatest amount of testimony leads only to pro- gressive probability. XXVII. C. Rirrer. The question of comparative recurrence was the object of the first book on Plato’s style, a monument of patient labour, by Constantin Ritter '™ (1888), now teacher at the gymnasium of Ellwangen in Wirtemberg. Until the publication of this book the investigations on the style of Plato were published as academic dissertations, articles in reviews, or as with Campbell, Riddell, Blass, and Peipers, in volumes on a different subject. Ritter was the first to write a special work on the matter, but he likewise knew only a few among his predecessors. He quotes Blass, Dittenberger, Frederking, Schanz, and Roeper, out of all the authors wko had preceded him in studying Plato’s style. But, again, as with Droste, this incomplete bibliographical equipment did not prevent Ritter from achieving a great progress towards the full solution of our pro- blem, and even perfecting earlier methods. He not only corrected numerical errors committed by Ditten- 1% ©, Ritter, Untersuchungen tlber Plato, Stuttgart 1888 (xxxiii). and he insisted upon the im port- ance of negative evidence. C. Ritter knew only five out of his pre- decessors, but he achieved a great pro- gress in the study of Plato’s style by measuring the oppor- tunities for the occur- rence of different assertions and nega- tions. C. Ritter investi- gated a greater number of stylistic peculiari- ties than any of his German predeces- sors, and, though he did not 122 ORIGIN AND GROWTH OF PLATO’S LOGIC berger, Frederking, and Schanz: he introduced a new method of estimating the recurrence of words, undertaking to calculate the number of opportunities for the intro- duction of at least one important class of words used by Plato. Previous writers had only reckoned the words occurring—or the number of times each word recurred in each dialogue—or the proportion of occurrences to a page of text. Nobody had counted the number of opportunities for using a given word. This Ritter did, and found for various kinds of affirmative and negative answers a better basis of comparison than that of the proportion to a page of text. He accepted the sum of all such forms of answer as the number of opportunities for the occurrence of each special form of answer, and referred to this number the particular observations of each form. This was an important step in advance as regards method, to which corresponded also a remarkable progress in the knowledge of Platonic chronology. Before Ritter only the order of the last six dialogues was well as- certained. His merit lies in giving a detailed justification of Campbell’s earlier supposition that the group preceding the Sophist consisted of the Republic, Phaedrus, and Theaetetus. From the numerous observations of Ritter the following more especially characterise the latest group of six dialogues : $12. mperov Av ely (p. 6): Tim. 2 Legg. 16. 813. mas Kai rq (p. 67): Phil. 1 Tim. 1 Legg. 2. $14. as duvardy (p. 6): Phil. 1 Legg. 4. 815. xaOarepei (p. 58): Polit. 1 Phil. 8 Tim. 1 Legg. 1. $16. ype (p. 6): Soph. 1 Polit. 1 Tim. 8 Critias 2 Legg. 57. 817. elov predominates over éAeyov (p. 10): Symp. 3/2 Parm. 6/8 Soph. 4/1 Polit. 5/4 Phil. 5/4 Tim. 3/0 Critias 1/0 Legg. 24/6. 818. Answers such as éywye, éuovye, and the like (Soxet prot, €pot your Soxet) which denote a subjective assent, are very rare, occurring less than once in sixty answers (p. 17) : Phaedr. 1/69 Parm. 7/486 Soph. 1/215 Polit. 3/251 Phil. 3/314 Tim. 0/43 Critias o/o Legg. 0/868 (in earlier dialogues they occur very often, namely, once in five answers in Euthyph. Meno, once in six answers in Lach. Euthyd. Gorg., once in seven to ten answers in Apol. Crito Charm. Crat. PLATO’S STYLE: RITTER 123 Theaset., once in sixteen to eighteen answers in Prot. Phaedo know Rep.). Campbell, 319. xara ye THY épny (p. 68) : Polit. 2 Phil. 1 Tim. 1 Legg. 1. reached 820. Inversion of the ordinary position of Aéyers, as for instance similar A€yes dAnOéotrata instead of dAnOeatara reyes (p. 56): Soph. 4 nie Polit. 2 Legg. 8. aint: 321. ro mayrav (p. 72): Polit. 2 Tim. 2 Legg. 8. 322. eixds yodv (p. 57): Parm. 1 Soph. 4 Polit. 7 Phil. 5 Legg. 16. Other peculiarities of later style extend also over the group of Rep. Phaedr. Theaet. : 328. mavrn mdvros (pp. 67, 101): Phaedr. 1 Parm. 1 Tim. 1 Legg. 2. 324. eipnrat (p. 10): Phaedr. 1 Theaet. 1 Soph. 1 Polit. 2 Tim. 3 Legg. 11. 325. Superlatives dAnOécrara, dpOdrara déyets prevail over corresponding positives in aftirmative answers (Ritter, p. 19, corrected by Tiemann, |” p. 586) only in: Phil. 22/6 Legg 36/22 and are half as frequent or oftener in Phaedo 4/8 Rep. 29/48 Phaedr. 2/2 Theaet. 8/14 Soph. 6/10 Polit. 7/8. 326. yap odv in short answers (pp. 57, 100): Rep. 4 Theaet. 1 Parm. 22 Soph. 6 Polit. 5 Phil. 1 Legg. 1. 827. mavres kal rdvry (p. 67): Rep. 1 Phil. 1. 328. ij) was... i} was (p. 57): Rep. 1 Phil. 5 Legg. 6. 329. pupio (p. 5): Rep. 1 Polit. 1 Phil. 1 Tim. 1 Legg. 2. 330. dvayxaiov, dvayxaidrara (p. 20): Rep. 8 Soph. 1 Phil. 7 Legg. 4. 331. i mas; (p. 24, in questions exacting affirmative answers) : Rep. 1 Parm. 1 Soph. 4 Polit. 3 Phil. 5 Legg. 8. 332. m7; (p. 25): Rep. 4 Parm. 8 Soph. 7 Polit. 6 Phil. 3 Legg. 8. 883. mwas eles; (p. 25): Rep. 1 Polit. 8 Phil. 1 Legg. 1. 334. djdov ws (pp. 2-3): Rep. 2 Phaedr. 8 Soph. 8 Polit. 2 Phil. 5 Tim. 4 Critias 1 Legg. 14. 335. paxp@ (p. 5): Rep. 2 Theaet. 1 Phil. 2 Tim. 1 Legg. 4. 336. €pp76n (p. 10): Rep. 1 Theaet. 1 Soph. 1 Polit. 6 Phil. 1 Tim. 1 Critias 2 Legg. 8. 337. ¢uoi your doxei (p. 17): Rep. 4 Theaet. 1 Phil. 2 Legg. 1. 338. ovxodv yp) or adda yxp7 (p. 22): Rep. 4 Theaet. 1 Parm. 1 Soph. 2 Polit. 4 Phil. 3 Legg. 1. 389. xai ras; (p. 28): Rep. 6 Theaet. 2 Parm. 1 Soph. 6 Polit. 1 Phil. 6 Legg. 11. 840. al més dv; (p. 24): Rep. 2 Theaet. 2 Parm. 1 Soph. 1 Phil. 1. 341. €£ dvdykns (p. 67): Rep. 6 Phaedr. 2 Theaet. 1 Soph. 5 Polit. 4 Phil. 2 Tim. 18 Legg. 22. This co- incidence shows the superi- ority of stylistic determi- nation of chrono- logy over other methods which constantly contradict each other. 124 ORIGIN AND GROWTH OF PLATO'S LOGIC 342. ddnOécrara, dp6as, dpOdrara without A¢yes and épOdrara Aéyes in affirmative answers (pp. 17, 56): Rep. 57 Phaedr. 1 Theaet. 6; Parm. 22 Soph. 16 Polit. 26 Phil. 32 Legg. 38. (Arnim: Soph. 18 Polit. 29 Legg. 40; Tiemann: Rep. 55 Polit. 28 Phil, 81 Legg. 85.) 348. dor (pp. 20, and 86, 100): Rep. 24 Phaedr. 1 Theaet. 2 Parm. 2 Polit. 4 Phil. 1 Legg. 4. There remain some peculiarities, which, though more frequent in the later dialogues, occur also exceptionally in one or other of the earlier works : $44. ovdauy ov8ands or pndauy pundayds (p. 66): Phaedo 1 Theaet. 1 Parm. $ Phil. 2 Tim. 2 Legg. 8. 845. xa\Xtoros Kut dptoros (p. 7): Symp. 1 Phaedr. 1 Tim. 4 Legg. 4. 846. eis or xara Suva (p. 6): Crat. 1 Rep. 6 Phaedr. 1 Soph. 8 Polit. 11 Phil. 4 Tim. 10 Critias 1 Legg. 63. $47. elres or elpnxas in answers (p. 19): Gorg. 1 Rep. 1 Theaet. 1 Soph. 2 Polit. 7 Phil. 8 Legg. 11. 848. iméAaBes (p. 20): Rep. 2 Theaet. 1 Legg. 5. $49. ravranact pev ody (pp. 23, 86): Lach. 1 Rep. 38 Phaedr. 8 Theaet. 9 Parm. 7 Soph. 10 Polit. 4 Phil. 4 Tim. 1 Legg. 18. 850. cyedov without re (p. 8): Apol. 2 Crito 1 Charm. 1 Gorg. 8 Phaedo 2 Rep. 7 Phaedr. 4 Soph. 26 Polit. 13 Phil. 14 Tim. 9 Critias 4 Legg. 122. 351. ra viv as adverb (p. 7): Charm. 1 Prot. 1 Phaedo 1 Rep. 1 Soph. 5 Polit. 5 Phil. 9 Tim. 5 Critias 3 Legg. 79. 852. xai pada (p. 28): Euthyph. 1 Euthyd. 1 Phaedo 1 Rep. 47 Phaedr. 3 Theaet. 4; Parm. 2 Soph. 4 Polit. 2 Phil. 7 Legg. 6. 353. Questions by means of soins (p. 25): Lach. 1 Crat. 2 Phaedo 1 Rep. 48 Phaedr. 4 Theaet. 13 Parm. 3 Soph. 32 Polit.s6 Phil. 33 Legg. 47. 854. mavu péev ody prevails over rdvv ye in (Ritter, pp. 22-28, corrected by Arnim,'* p. 6): Crito 1/0 Rep. 64/40 Phaedr. 3/1 Theaet. 16/5 Soph. 14/10 Polit. 18/7 Phil. 23/9 Legg. 49/4, and is over half as frequent in Lach. 6/10 Prot. 3/8 Phaedo 21/23 Parm. 15/28. 855. ydpev (p. 59): Prot. 1 Gorg. 8 Symp. 1 Rep. 12 Phaedr. 8 Theaet. 4 Soph. 1 Polit. 8 Phil. 8 Tim. 7 Critias 2 Legg. 33. These considerable additions to the number of pecu- liarities of Plato’s later style led C. Ritter to the same general conclusions as those arrived at by Campbell twenty years earlier, namely that Soph. Polit. Phil. Tim. Cnitias Legg. are the last works of Plato, and that PLATO'S STYLE: RITTER, WALBE 125 Rep. Phaedr. Theaet. form a group preceding them. At the same time, other inquirers added new observations, all confirming this distinction of the above two groups of Plato’s works, and happily avoiding repetition of work already done. XXVIII. WauBe. The philological seminary of Bonn University, where the dissertations of Roeper and Hoefer were written, produced in 1888 a third doctoral disserta- tion on the style of Plato, by E. Walbe' (1888) who counted the occurrences of mds, 1ts compounds and the expressions containing it. Of his predecessors he only knew Roeper, Dittenberger, Hoefer, and Schanz. Among over a hundred uses of was enumerated by Walbe, the following deserve our special attention : 356. £uvaras (p. 3): Soph. 8 Polit. 1 Phil. 2 Tim. 3 Legg. 1. 857. of Evuravres or ra Evpravra (p. 11): Soph. 1 Polit. 1 Phil. 1 Legg. 8. 358. mas otros or ovros mas (p. 36): Crat. 1; Soph. 2 Parm. 2 Phil. 1 Tim. 2 Critias 1 Legg. 5. 359. was doticvdy (p. 87): Polit. 1 Phil. 1 Legg. 1. 860. ra mdvra yévn (p. 85): Soph. 1 Tim. 2. $61. ra mdvra eidn or pépn (p. 85): Rep. 1 Theaet.6 Parm. 4 Legg. 1. 362. rovvavriov dravy or Grav rovvartiov (p. 16): Polit. 1 Phil. 1 Legg. 3. 363. ro fvprayv (p. 9): Phaedr. 1 Soph. 1 Polit. 1. 364. wav (dor, meaning ‘ every animal’ (p. 20): Rep. 1 Polit. 1 Phil. 2 Tim. 3 Legg. 5. 365. fupmras prevails over das only (p. 4): Soph. 20/8 Polit. 45/18 Phil. 21/19, while in all other dialogues dias is more frequent, being in Tim. Legg. over twice as frequent as évpras. $66. was and its compounds occur over four times in a page ed. Didot (p. 4) in: Soph. 181 Polit. 239 Phil. 209 Tim. 378 Critias 67 Legg. 1290, rising in Polit. Tim. Critias Legg. to more than five and even up to seven times in a page, while in all other dialogues they are much scarcer (Euthyd. 102 Crat. 187 Symp. 142 Rep. 601 Theaet. 188 Parm. 91, elsewhere less). 867. das, évumas, Evvdras occur over once in two pages in (p. 4): Apol. 12 Crito 7 Lach. 10 Prot. 22 Euthyd. 17 Parm. 17 Soph. $1 Phil. 42, and over once ina page in: Polit. 6@ Tim. 62 iss HK. Walbe, Silesius, Syntaxis Platonicae Specimen (Doctor’s diss.), Bonn 1888 (xxxiv). Walbe’s observa- tions on was and com- pounds lead to the same results, though he made no methodic use of them, attaching chronolo- gical im- portance only to the frequency of a few words. Siebeck’s observa- tions on peculiari- ties of style explained by Plato’s psycho- logical evolution. This is a progress in the 126 ORIGIN AND GROWTH OF PLATO'S LOGIC Critias 12 Legg. 288, in all other dialogues less. being over once in three pages only in: Meno 8 Gorg. 28 Phaedo 17 Rep. 73 Phaedr. 17 Theaet. 20. 368. sav dcov (p.7): Symp. 1 Soph. 1 Polit. 1 Tim. 4 Legg. 8. 869. mavra (oa or (Ga rdvra (p. 31): Phaedo 2 Rep. 2 Soph. 1 Phil. 3 Tim. 2 Legg. 8 (including two occurrences of (6a gvpruvra in Legg.). 870. das or drav without article or substantive (pp. 5, 7): Symp. 1 Phaedo 1 Phaedr. 1 Parm. 8 Tim. 2 Legg. 4. 371. rd nav, meaning the universe (omniwm rerum universitas, p. 10), is limited to: Crat.3 Symp. 1 Rep. 1 Theaet. 3 Parm. 1 Soph. 8 Polit. 7 Phil. 10 Tim. 38 Legg. 11. 872. rd wav Stadépew (pp. 10-11): Polit. 1 Legg. 2. 878. maca or Graoa dvdyxn (p. 28): Phaedo 2 Rep. 5 Phaedr. 2 Theaet. 2 Soph. 2 Phil. 1 Tim. 4 Legg. 2. 874. mas or compounds used together with éxaoros (p. 37): Euthyd. 1 Rep. 2 Theaet. 1 Parm. 1 Soph. 1 Tim. 6 Legg. 1. 875. was used with odos (p. 88): Rep. 2 Soph. 1 Legg. 3. XXIX. SrEBECK. In the same year as Walbe’s dis- sertation and Ritter’s work was published an’ original investigation on Plato’s style by H. Siebeck,'* author of the History of Psychology. Siebeck, as a psychologist, sought for characteristics of Plato’s style revealing changes in the author’s state of mind which are capable of psychological explanation. He chose for his purpose the different classes of affirmative answers, and made a step further in the right method of calculating opportu- nities for the occurrence of each particular answer, not taking, as Ritter did, the sum of all answers as a com- parative measure, but the sum of all affirmative answers only. Siebeck, moreover, classified all these answers and distinguished problematic, assertive, and apodictic afirma- tions. The apodictic affirmations, as for instance a\76éo- tata, opOcrara, mavrdract, &c., are, as Siebeck shows, in all cases when the chronological order of two dialogues is known from other certain sources, more numerous in the later works. They form in the Republic 18% H. Siebeck, Untersuchungen zur Philosophie der Griechen, 2° A., Freiburg in B., 1888, pp. 253-266: ‘Nachtrige die platonische Frage betreffend, I. Sprachstatistisches ’ (xxxv). Siebeck knew among his pre- decessors Dittenberger, Frederking, Hoefer, Schanz, and Gomperz. PLATO'S STYLE: WALBE, SIEBECK 127 fifty per cent. of all affirmative answers, and in the Laws fifty-four per cent. A similar relation is observed in the dialectic trilogy. In the Theaetetus Siebeck found thirty- eight per cent. apodictic answers, in the Sophist forty-two per cent., in the Politicus forty-nine per cent.—while in the Protagoras, generally recognised as an early dialogue, such answers form only fifteen per cent. of all. More- over, in the separate books of the Republic we notice the like progress from a more problematic to an apodictic certainty. In Book I we find thirty-eight per cent. apodictic affirmations, as in the Theaetetus ; in Books II- IV they mse to forty-six per cent.; in Books V-IX to fifty-four per cent.; in Book X thcy decline a little, being fifty-three per cent. of all affirmative answers. It would be an exaggeration to affirm that these numbers correspond precisely to the chronological order, because the special subject of each work gives greater or fewer opportunities for apodictic certainty, and if the Phaedo contains forty-nine per cent. apodictic replies, this is no sufficient reason for inferring that this dialogue was written after the Sophist ; still, Siebeck’s method of cal- culating the opportunities for different kinds of answers marks a progress over Ritter’s first attempt. Siebeck also counted the number of simple direct questions, with- out any interrogative particle, or with 7 or dpa or pov only, in order to find the relative recurrence of these par- ticles ; and he found the percentage of questions with dpa or pay to be very high in the dialogues of the latest group. These investigations increase our list by some charac- teristics whose importance outweighs their number : 876. Over forty in each hundred affirmative answers are apodictic (p. 260) only in: Phaedo 83/168 Rep. 669/1342 Phaedr. 42/76 Parm. 189/394 Soph. 140/329 Polit. 130/268 Phil. 298/323 Legg. 312/878. In other dialogues the proportion is much smaller, coming nearest to the later style in Euthyd. 45/180 Gorg. 105/821 Crat. 77/288 Theaet. 101/263 (in these dialogues over 80 %). $77. To each problematic answer correspond at least four method of stylistic study for chrono- logical purposes, as can be tested on those works whose chrono- logical order is otherwise known. Siebeck’s calcula- tions add very im- portant informa- tion to our knowledge of Plato’s style. They show that apodictic affirma- tions and certain kinds of interroga- tions in- crease in frequency in the latest group. Tiemann supple- mented Ritter’s observa- tions and corrected them on some points, giving more detailed informa- tion on the use of participles and of some kinds of answers peculiar to later style. 128 ORIGIN AND GROWTH OF PLATO'S LOGIC apodictic answers or more: Phaedo 20/83 Rep. 141/669 Phaedr. 10/42 Soph. 31/140 Phil. 32/198 Legg. 69/312. In other dialogues the problematic answers occur much oftener, being less than one to three apodictic answers only in Euthyd. 12/45 Gorg. 82/105 Parm. 52/159 Polit. 35/130. 878. Interrogations by means of dpa form 24 % or more of all simple interrogations: Parm. 80/207 Soph. 46/271 Polit. 31/106 Phil. 56/186 Legg. 98/329, while in all other dialogues dpa is much scarcer, the proportion being above 15 % only in: Prot. 27/140 Crat. 34/172 Phaedo 31/161 Rep. 183/931 Phaedr. 11/72 Theaet. 39/229, and in other dialogues less. XXX. TIEMANN. Stylistic investigations on Plato became better known after 1888; those of Dittenberger, Schanz, Ritter, and Siebeck receiving most attention, but still they met with obstinate opposition, and Zeller con- tinued to disdain them. J. Tiemann,'*’ under the influence of Ritter’s work, investigated the use of some participles with evar, and noticed among others the following peculiarities : 379. Particip. aorist. with eivac (p. 559) : Polit. 2 Tim. 1 Legg. 1. 880. mpemov with eiva: Lach. 1 Gorg. 1 Symp. 1 Tim. 2 Critias 2 Legg. 7. 881. mpoonxwy with evac: Rep. 8 Phaedr. 1 Tim. 1 Legg. 2. 382. Part. praes. with efvac: Euthyph.1 Prot.1 Meno 2 Gorg. 2 Crat. 2 Symp. 1 Phaedo 1; Rep. 8 Phaedr. 3 Theaet. 3; Soph. 6 Polit. 8 Phil. 8 Tim. 4 Critias 1 Legg. 11. 383. Pleonastic use of participles (p. 556) : Lach. 1 Prot. 1 Meno 8 EKuthyd. 1 Gorg. 8 Crat. 1 Symp. 2 Phaedo 2; Rep. 14 Phaedr. 4 Theaet. 3; Soph. 7 Polit. 12 Phil. 7 Tim. 12 Critias 4 Legg. 24. $84. Periphrastic impersonal expressions (p. 556): Symp. 1 Rep. 1 Soph. 1 Polit. 2 Tim. 7 Critias 2 Legg. 10. $85. aAn6n without Aéyes in affirmative answers (p. 586) : Charm. 3 Lach. 1 Prot. 1 Gorg. 1 Phaedo 1 Rep. 29 Theaet. 9 Parm. 28 Soph. 7 Polit. 5 Phil. 2 Legg. 4. (The occurrence of adn 67 in Prot. and Gorg. has not been noticed by Tiemann, nor by C. Ritter, but is mentioned by von Arnim'‘ p. 9, and has been admitted here on his testimony, because an involuntary omission 137 J, Tiemann, ‘Zum Sprachgebrauch Platos’ in Wochenschrift fir klassische Philosophie, 1889, columns 248-253, 862-366; also in his exten- sive review of C. Ritter’s work in the same journal, columns 791-797, 839-842, Berlin 1889 (xxxvi). The numbers for Parmenides omitted by Tiemann have been in some cases added from Arnim’s (see note 144) publication. PLATO'S STYLE: TIEMANN, LINA 129 appears more probable than a wrong observation, unless Arnim counted as simple dAn67 some adnOq r€yets.) dAnOq AeEyes, Very common in earlier dialogues, is scarcer afterwards. Already C. Ritter had noticed that the abridged forms op0ads, adnbéorata, opOorara without Aédyers, as well as opOorara even with Ayes, were limited to Rep. Phaedr. Theaet. Parm., and to the six latest dialogues, occurring nowhere earlier (342). Tiemann counted the occurrences of each of these forms of affirmative answers, and found that op0ds, dpOorara, and dAnOécrara, with or without Aéyevs, though not limited to the latest works, occur in them with increased frequency, and may therefore be looked upon as peculiarities of later style: 386. dpdas with or without Aé¢yes in affirmative answers (p. 586): Euthyph. 1 Charm. 1 Meno 1 Crat. 1 Phaedo 1 Rep. 35 Phaedr. 1 Theaet. 5 Parm. 18 Soph. 10 Polit. 17 Phil. 13 Legg. 24. (Arnim agrees generally with these numbers, but he found no 6p6és in Meno and Crat., and only two in Phil., 25-26 in Legg.) 387. adnOeorara with or without Adyes in affirmative answers (p. 586): Lach. 1 Crat. 1 Symp. 1 Phaedo 4 Rep. 28 Phaedr. 2 Theaet. 7 Parm. 6 Soph. 5 Polit. 7 Phil. 16 Legg. 23 (Arnim Legg. 24). 388. dp@drara with or without A¢yes in affirmative answers (p. 586): Rep. 10 Phaedr. 1 Theaet. 1 Parm. 1 Soph. 4 Polit. 8 Phil. 10 Legg. 12. Arnim: Rep. 11 Phaedr. 2 Soph. 5 Polit. 12 Phil. 12 Legg. 15 or 16. (In this and the preceding Nos. 385-387 the numbers for Parmenides, omitted by Tiemann, are quoted from Arnim, who slightly differs from Tiemann and Ritter in other numbers.) XXXII. Lina. Simultaneously with Tiemann, Lina published at Marburg a dissertation wherein he classi- fies no fewer than twenty-one thousand eight hundred and eighty-one occurrences of prepositions in Plato’s works. From his lists the following confirmation of earlier results is gathered : us T. Lina, De praepositionum usu platonico; dissertatio inauguralis Marpurgi 1889 (xxxvii). Of his predecessors Lina knew Dittenberger, Schanz, and Ritter. K Lina counted all uses of preposi- tions but failed to find a difference of fre- quency in their use between early and late dialogues, because he used &® wrong measure of text. His obser- vations teach us, however, & great number of pecu- liarities of later style, among which various kinds of inter- positions are pro- minent. 130 ORIGIN AND GROWTH OF PLATO'S LOGIC 889. xara with the accusative prevails over all other prepositions except év (p. 9): Crat. 75 Polit. 1380 Critias 50 Legg. 697, and over ev in Soph. 128 Tim. 283. In these dialogues xara cum acc. forms 12-15 % of the whole number of prepositions, while in other works it is much scarcer, reaching 9 % only in the Theaet. and falling to the fourth rank in Parm. (after é¢v, mpds, éx), Phil. (after év, eis, mepi), Legg. B. vi. x. xii. (after é:, eis and ex or mepi). The prevalence of xard in some dialogues is so much the more characteristic, as in the whole of Plato’s text ev (4148), mepi (8267), mpos (2292), prevail much over card (2065). 890. Twenty-one or more prepositions on each page (ed. Didot) occur only in: Phaedr. 819 Polit. 916 Tim. 1733 (82 in one page) Critias 363 (83 in one page) Legg. 5249 (22 in one page), over 19 in: Lach. 352 Phaedo 945 Rep. 3865 Soph. 757; over 88 in two pages in: Prot. 678 Symp. 737 Theaet. 885 Parm. 512 Phil. 778, elsewhere less. (In this case the superiority of Didot’s edition over Teubner’s, as a measure of text, is manifest. Lina gives for Polit. the proportion of 11 prepositions to one page, the same as for Prot., while from the numbers he quotes it results that one page ed. Didot contains in Prot. 17°4 prepositions, in Polit. 21°38. This should be carefully borne in mind by all future inquirers, who wish to determine how often per page a word occurs. The proportion of 11 prepositions to one page ed. Teubner is given by Lina also for Lach., with 19°5 preposi- tions on one page ed. Didot; according to his calculations Symp. [18°9 prepositions on one page ed. Didot] and Phaedo [19-3 prepositions on one page ed. Didot] would contain more preposi- tions [12 on each page ed. Teubner] than the Politicus [11 pre- positions on one page ed. Teubner, and 21°3 prepositions on one page ed. Didot], while they really contain two prepositions less on each page ed. Didot. It follows that the standard of a page varies, and that we must be cautious in selecting a measure of text. So long as the ideal measure, the number of words of each dialogue, remains unknown, there is no safer standard than the pages of Didot’s edition for measuring Plato’s text.) $91. epi c. accus. prevails over rept c. gen. (p. 12): Symp. 40/39 Soph. 76/71 Polit. 92/53 Tim. 116/88 Critias 29/21 Legg. iii. v. vi. vii. 182/147. This is a very characteristic peculiarity, because in all other dialogues the predominance of epi c. gen. over mepi c. ace. is so great that in the dialogues not specified by Lina 1552 mepi c. gen. correspond to 804 epi c. acc. $92. xard sundered from the corresponding accus. by 8n (p. 14): Meno 1 Rep. 1 Parm. 2 Soph. 2 Polit. 1 Phil. 1 Tim. 1 Legg. 1. 393. idem, by pev (p. 14): Gorg. 1 Rep. 2 Theaet. 2 Polit. 1 Phil. 1 Tim. 1 Legg. 2. $94. idem, by 3¢ (p. 14): Gorg. 1 Crat. 2 Rep. 4 Theaet. 4 Parm. 8 Soph. 1 Phil. 1 Tim. 5 Critias 3 Legg. 2. PLATO'S STYLE: LINA 131 395. idem, by ré (p. 14): Rep. 2 Phaedr. 1 Theaet. 1 Polit. 1 Phil. 1 Tim. 4 Legg. 8. $96. idem, by yeé (pp. 14, 75): Charm. 2 Gorg.1Crat. 1 Phaedo1 Rep. 1 Theaet. 1 Soph. 1 Polit. 4 Phil. 5 Tim. 1 Legg. 4. 397. idem, by a genitive (p. 14): Crat. 1 Rep. 1 Phaedr. 1 Parm. 1 Polit. 1 Legg. 8. 398. idem, by more than one word (p. 15): Gorg. 1 Crat. 1 Symp. 1 Parm. 1 Polit. 1 Critias 1 Legg. 2. 899. mepi sundered from the corresponding genitive by 47 (p. 16) : Theaet. 1 Soph. 1 Polit. 1 Tim. 2 Legg. 2. Ss 400. idem, by dé (p. 16): Lach. 2 Prot. 3 Crat. 2 Symp. 2 Rep. 1 Phaedr. 2 Theaet. 2 Soph. 2 Polit. 2 Phil. 2 Tim. 2 Legg. 12. 401. idem, by yé (p. 16): Euthyph. 2 Phaedo 1 Rep. 1 Theaet. 1 Phil. 1 Legg. 4. 402. idem, by ré (p. 16): Euthyph. 1 Crito 1 Charm. 1 Symp. 1 Phaedo 1 Rep. 7 Theaet. 1 Soph. 1 Polit. 3 Phil. 2 Tim. 2 Critias 1 Legg. 4. 408. idem, by a genitive (p. 16): Euthyd. 3 Gorg. 8 Phaedo 1 Phaedr. 2 Theaet. 2 Polit. 2 Critias 1 Legg. 4. 404. idem, by pev ydp (p. 17): Rep. 1 Theaet. 1 Polit. 1. 405. idem, by three to five words (p. 17): Crat. 2 Rep. 1 Phil. 1 Tim. 2 Legg. 3. 406. mepi, sundered from the corresponding accus. by ye 18): Crat. 1 Legg. 2. 407. idem, by 8¢ (p. 18): Gorg. 3 Rep. 1 Phaedr. 1 Theaet. 2 Soph. 8 Polit. 3 Phil. 2 Critias 2 Legg. 2. 408. idem, by pev (p. 19): Gorg. 2 Rep. 1 Phaedr. 1 Theaet. 2 Soph. 3 Polit. 1 Phil. 1 Tim. 1 Legg. 8. 409. idem, by ré (p. 19): Phaedo 1 Rep. 2 Phaedr. 1 Polit. 4 Phil. 1 Tim. 4 Legg. 9. 410. idem, by a genitive (p. 19): Euthyph. 1 Lach. 1 Euthyd. 1 Crat. 1 Symp. 1 Rep. 3 Phaedr. 2 Polit. 1 Phil. 1 Legg. 8. 411. idem, by two or three words (p. 19): Symp. 1 Rep. 2 Phaedr. 1 Tim. 1. 412. wepi placed after the substantive which depends on it (anastrophe) was not very much used by writers earlier than Plato (as for instance Thucydides), while in Plato it forms over 17 % of all occurrences of this preposition, and after Plato it became still more common. But this use is not equally frequent in all dialogues; it does not occur in Crito Charm., forms under 5 % of all occurrences of mep{ in Prot. Euthyd. Crat. Phaedo, rises above 6 % in Apol. 2/24 Euthyph. 3/37 Meno 5/50 Gorg. 9/92 Symp. 8/89 Parm. 2/80 Critias 2/21, above 10 % in Lach. 10/78 Theaet. 14/123 Tim. 13/88; and above 20 % only in: Rep. 6o K 2 The very frequent use of wépr in ana- strophe begins with the Republic, and some 132 ORIGIN AND GROWTH OF PLATO'S LOGIC special (22 %) Phaedr. 18 (21 %) Soph. 16 (22 %) Polit. au (21 %) interposi- Phil. 21 (82 %) Legg. 139 (29 %) (calculated from the table tions are given by Lina on p. 29). inter core 418, Between a genitive and a following wép: belonging to it, is placed a y¢ (p. 26): Gorg. 1 Phil. 1. 414. idem, d¢ (p. 27): Gorg. 1 Rep. 1 Legg. 9. 415. idem, dy (p. 27) : Prot. 1 Phaedr. 2 Soph. 1 Polit. 1 Phil. 1 Legg. 8. 416. idem, ré (p. 27): Euthyph. 1 Gorg. 1 Symp. 1 Phaedo 1 Rep. 17 Theaet. 1 Soph. 1 Polit. 2 Phil. 8 Tim. 1 Critias 1 Legg. 12. 417. idem, a genitive (p. 27): Phil. 1 Legg. 2. 418. Between a genitive depending on wép: and the following méptis placed another word (not one of the above particles (413— 416), but including the genitives counted in 417) or more words (p. 27): Apol. 1 Lach. 1 Rep. 1 Phaedr. 8 Theaet. 1 Soph. 8 Polit. 1 Phil. 3 Legg. 17. 419. ava Adyov (in the same meaning as xara Adyoy = in pro- portion) or ava rév adrdoy Adyov (p. 85): Phaedo 2 Rep. 2 Tim. 6 Legg. 1. 420. xara c. genit. after a verbum dicendi in the same mean- ing as epi (p. 87): Charm. 1 Meno 2 Euthyd. 1 Crat. 1 Symp. 1 Phaedo 1 Rep. 1 Phaedr. 1 Soph. 1 Legg. 2. 421. idem, after a verbwm agendi (p. 87): Meno 2 Phaedo 1; Soph. 1 Phil. 1 Legg. 1. 422. xara c. accus. to designate the direction of a movement meanings (for which generally the genitive is used) meaning towards or to or are pre- in (pp. 89, 40): Symp. 1 (190 £: xara rnv yaorépa) Phaedo 1 (114 a: valent in xara Thy Aimy) Rep. 1 (614 pv) Tim. 8 Critias 4 Legg. 1 (905 a). the latest 423. idem, metaphorically (pp. 89-41) : Crat. 1 Symp. 4 (205 p: works, a8 cara xpynpariopoy etc.) Rep. 1 (896 p) Phaedr. 2 Theaet. 2 Parm. 2 Soph. 4 Polit. 1 Phil. 1 Tim. 2 Legg. 4. frequent. Also some i lt ag : 424. xara c. accus. to designate the diffusion of something over Lina’ or through some space or place (p. 41): Prot. 1 (818 D: xara ras eral wédes) Crat. 1 Symp. 1 Phaedo 2 Rep. 1 Phaedr. 2 Theaet. 1 enumera- Tim. 18 Critias 1 Legg. 2 (undicatur aliquid per aliquem locwm tions. diffundi). 425. idem, metaphorically: Rep. 1 Theaet. 1 Parm.1 Soph. 2 Polit. 1 Phil. 2 Legg. 6. 426. xara c. acc. to designate a place (= in) in such phrases as xara rdrov, OF xara xopay, or car’ doru (card mow is not counted, because Lina does not quote all the numerous occurrences of this phrase) (p. 48): Gorg. 1 Rep. 1 Tim. 4 Critias 4 Legg. 6. 427. xard pecov (p. 48): Phaedo 1 (118 a) Rep. 1 Soph. 1 Tim. 1 Critias 4 Legg. 2. 428. xara Oadarray (p. 44): Rep. 1 Polit. 1 Legg. 9. PLATO'S STYLE: LINA 133 429. kat’ ayopdy or xar’ ayopas (p. 44): Rep. 2 Theaet. 1 Parm. 1 Polit. 1 Legg. 7. 430. xara xatpov (p. 47): Polit. 1 Legg. 2. 481. xar’ éxeivoy rov xpovov (p. 47): Polit. 2 Tim. 8 Legg. 5. 482. xa@ Urvoy (p. 47): Tim. 8 Legg. 1. 433. xara Bpayv = paulum, non multum (p. 57): Soph. 2 Tim. 1 Legg. 2 (In Prot. and Gorg. the same words mean according to Lina breviter). 484. xara (rd) ép6cv (p. 57): Soph. 1 Tim. 1 Legg. 1. 435. xara pépos (p. 59): Soph. 1 Legg. 2. 436. card pépyn (p. 59): Rep. 1 Theaet. 1 Parm. 4 Tim. 8 Legg. 2. 437. 438. xara piva (p. 60): Rep. 1 Legg. 8. kara ruxny (p. 63): Rep. 1 Phaedr. 1 Soph. 1 Legg. 2. 439. dpoiov card tiva (p. 67): Phaedo 1 Polit. 1 Tim. 1 Legg. 1. 440. ro (or ra) xara te (rd apa, ras éemiornpas, &c.), meaning ‘ampliorem quam simplex substantivum notionem’ (p. 71): Euthyd. 1 Gorg. 1 Phaedo 1 Rep. 1 Phaedr. 1 Soph. 1 Critias 1 Legg. 8. 441. xara c. accus. without any grammatical relation to any part of the phrase, and meaning ‘ quod attinet ad,’ is found only (p. 72): Meno 1 (72 A: xara rv eixova) Rep. 1 (614 D: xara ra érépw) Theaet. 1 (158 D: xara ra dppara) Phil. 1 (17 C: xara réxvnv) Critias 1 (109 Cc: xar’ dAXous rérovs) Legg. 1 (812 A: card Thy vrdbeow). 442. xara c. acc. meaning ‘quantum attvnet ad’ (p. 72): Symp. 1 (185 B: xa’ airdv) Legg. 2 (715 b, 928 B). 443. xara with the accus. meaning ‘according to somebody,’ or after somebody’s fashion (p. 56): Apol. 1 Meno 1 Euthyd. 1 Gorg. 2 Symp. 2 Phaedr. 8 Theaet. 1 Parm. 1 Legg. 2. 444. xara mapddevyya or xara ovvnOeav after a verbum dicend or agendt (p. 52): Meno 2 Soph. 1 Polit. 2 Tim. 2 Legg. 1. 445. xara forming a hiatus with a following a, e, 7 or o (pp. 22- 28): Meno 2 Gorg. 1 Symp. 1 Rep. 5 Phaedr. 2 Parm. 8 Polit. 1 Critias 1 Legg. 5. 446. card Gedy (p. 63 divina quadam sorte): Euthyd. 1 Rep. 1 Soph. 1 Legg. 3. 447. xara c. acc. in the distributive meaning after a verbum dividendi (except nar’ eidn dcatpeioGac which is too frequent for enumeration, p. 58): Meno 1 Rep. 1 Phaedr. 1 Soph. 8 Polit. 3 Tim. 8 Legg. 3. XAXXII.-XXXITI. Baron. Van Creer. After so many investigations on the Platonic vocabulary in three years (1886-1889), the subject remained untouched during the following seven years, though some authors wrote on But he has made no chrono- logical use of these observa- tions, as generally all his pre- decessors except Campbell ignored the me- thodical value of ap- parently accidental pecu- liarities. Van Cleef’s investiga- tions on the use of attraction in Plato are very valuable, but as he did not distin- guish the single dia- logues he prevented our apply- ing his work to chro- nology. The same applies 134 ORIGIN AND GROWTH OF PLATO’S LOGIC other aspects of Plato’s style, ignoring the relation between style and chronology. Compared with the laborious German dissertations, the French thése of C. Baron'® on the form of Plato’s writings appears almost a rhetorical exercise. A student of the Univer- sity of Bonn, Van Cleef,'? of Ohio, spent much time in minute research on the use of attraction in Plato, but he deprived us of some additional characteristics of Plato’s later style by mixing in his statistical tables dialogues of different dates without any distinction of single works. He followed Christ in uniting Rep. Parm. Soph. Polit. Phil. Tim. Critias Legg. into one class of so-called constructive dialogues; and he observed that the use of attraction, while occurring in the sum of Plato’s works about thirty-eight times in every one hundred pages, is reduced in this group to only fourteen cases in one hundred pages of text. This result tends to show that attraction generally was not a peculiarity of later style, but we are left uncertain whether this refers equally to all the eight dialogues of the group, or only to some of them. The group which Van Cleef calls con- structive dialogues contains, besides the recognised 81x latest dialogues, only Republic and Parmenides, so that we may admit as probable that the use of attraction decreased in Plato’s later style; and as all the passages are enumerated by Van Cleef, whoever cared to under- take the task of a new classification and methodic dis- position of the materials collected by him might draw very interesting chronological conclusions, or at least afford fresh confirmation to the chronological conclusions arrived at otherwise. . XXXIV.-XXXV. GRUNWALD. BERTRAM. From the instructive collection of proverbs found in Plato by ‘% ©, Baron, De Platonis dicendi genere, Paris 1891 (xxxvilii). 1 F. T. van Cleef, Ohianus, De attractionis in enuntiationibus relatiwis usu platonico (Doctor’s diss. Bonn University), Bonn 1890 (xxxix). PLATO'S STYLE : CAMPBELL 135 E. Griinwald ’ it is again impossible to draw any chrono logical inferences, because proverbs are seldom repeated, and cannot be regarded as peculiar to any given period of Plato’s style. Also Bertram’s interesting contribution on the use of metaphor in Plato ' contains nothing that could be included in our list. XXXVI. CAMPBELL. All the foregoing writers on Plato’s language, from Roeper to Van Cleef, ignored Campbell’s Introduction to the Sophist and Poltticus, though after the publication of Ritter’s book Campbell again on several occasions recalled his first investigations. But he published these later articles in journals of limited circulation on the Continent, as the Transactions of the Oxford Philological Society, or the Bzrbliotheca Platonica.'** Consequently the coincidence of results between Campbell and the German style statisticians was known to none but the Scotch philologer himself, while the few generally known German dissertations naturally failed to secure a general recognition of the results obtained by them alone. There is reason to think that Campbell’s more recent investigations on Plato’s use of language, filling 175 pages in the second volume of the monumental edition of the Republic by Jowett and Campbell (3 vols., Oxford 1894), will likewise escape the attention of German and French students of Platonic style, unaccustomed to look for such original 41 Dr. Eugen Griinwald, Sprichwirter und sprichwoiritliche Redensarten bei Plato, Berlin 1893. (Programme des Cours du Collége Royal Francais de Berlin) (x1). 1 Heinrich Bertram, ‘ Die Bildersprache Platons,’ Betlage zum Jahres- bericht der kiniglichen Landesschule Pforta, Naumburg a. 8. 1896 (xli). 43 Transactions of the Oxford Philological Society, 1888-1889, pp. 25- 42, June 14, ‘On the position of the Sophistes, Politicus, and Philebus in the order of the Platonic Dialogues, and on some characteristics of Plato’s latest writings,’ by Professor Lewis Campbell of St. Andrews (xlii); and on the same subject in Bibliotheca Platonica, an exposition of the Platonic Philosophy edited by Thos. M. Johnson, Osceola, Mo. U.S8.A. vol. i. July, August 1889,'N. 1, pp. 1-28: Prof. L. Campbell : ‘On some recent attempt towards ascertaining the chronological order of the composition of Plato’s dialogues ’ (xliii). also to publica- tions by Grunwald and Bertram. Camp- bell's recent publica- tions de- serve the attention of Platonic scholars not less than his first con- tributions to Platome literature thirty-five years ago. Von Arnim, without knowing Campbell or even Ritter, came to the same results, though his method of joining 136 ORIGIN AND GROWTH OF PLATO'S LOGIC labours in the Appendices to an edition of a single dialogue. It would, however, exceed the limits of the present survey to epitomise this last work of Campbell, which should stand on the shelves of every philological library. Enough to state that this new publication of Campbell is of no less importance for our knowledge of Plato’s style than his Introduction to the Sophist and Politicus written thirty years ago, and forms a splendid continuation of the work he began in 1861 by his edition of the Theaetetus. A full syntax of Plato’s language, illustrated by quotations not only from the Republic but from other dialogues, it confirms in many details the close relation of the Phaedrus and Theaetetus to the Republic on one side, and of the Sophist, Politicus, Philebus to Timaeus, Critias, Laws on the other side. XXXVII. Von Arnnim. The want of centralisation in Platonic studies is illustrated by the curious fact that quite recently an author who undertook researches on one aspect of Plato’s vocabulary, J. von Arnim '* (1896), Professor at the University of Rostock, not only knew nothing of Campbell’s publications, but even ignored Ritter’s book, having read nothing on the style of Plato but the articles of Dittenberger and Schanz. On the other hand, it is very instructive to note that von Arnim, after careful comparison of twenty-six cha- 4 Joannis ab Arnim, De Platonis dialogis, Quaestiones chronologicae, ad scholas quae in hac uniwersitate Rostochiensi per semestre hibernum inde a d. XVI M. Octobris A. MDCCCXOVI habebuntur invitant Reetor et concilium. Rostock 1896 (xliv). The numbers given by Arnim are in some cases different from the numbers given by C. Ritter. In such cases the larger number has been included in our list, because an omission is more likely to happen than that one passage should be counted as two, if the work is done carefully. But von Arnim sometimes changes his classification, so that he quotes different numbers for the same dialogue, as, for instance, twelve épOérara Aéyes in the Laws in § 18, and thirteen in § 14; two dAnOéorara Aéyeis in the Politicus in § 10, and five in § 14; one dp0ds Adyes in the Politicus in § 14, and none in § 11, &c. Also his numbers for the pecu- liarities which have been collected by OC. Ritter and Tiemann show some considerable differences, as, for instance, he did not find dp0és in the Philebus, while C. Ritter and Tiemann found it eleven times. PLATO'S STYLE: VON ARNIM 137 racteristic marks of Plato’s style, came independently to the same conclusions as Campbell in 1867, and as Ritter in 1888. He recognised that Soph. Polit. Phil. Tim. Critias Legg. are the latest of Plato’s works, and that the group preceding them contains the Republic, Phae- drus, Theaetetus, and Parmenides. Many of Arnim’s observations are new, and furnish us with several additional peculiarities of Plato’s later style : 448. vai, mdvv ye, mavy perv ovv form less than one-third of all affirmative answers (p. 6): Rep. 195 Phaedr. 11 Theaet. 58 Parm. 97 Soph. 71 Polit. 54 Phil. 52 Legg. 76, being in Rep. Phaedr. Phil. Legg. even less than one-fourth of all affirmative answers, while they form in all earlier dialogues over one-third, and in Meno Euthyd. Gorg. Crat. even over one-half of all answers. 449. xados and xadas ravra ye, as affirmative answers (p. 9): Rep. 1 Soph. 2 Polit. 6 Legg. 6. 450. cdAAcora and xdA\tord ye as affirmative answers (p. 9): Phil. 1 Legg. 1. 451. Rhetorical interrogations meaning affirmative answers (as : ri pny ; dAda ri pny; Ti yap KwAvEr; GAA Ti pedANe ; Ti yap ov pedAeL ; ri dn yap ov; ri yap ov ; ri 8’ ob pédda; ri 8’ od; mas yap dv Aas ; mwas yap ov péAAe; mas yap ov; mas 8 ov peAAe; Kai Tas Ov; TaS é’of ;) were increasing in Plato’s later works. They form over 20 % of all interrogations in (p. 14): Phaedr. 24/62 Soph. 49/240 Polit. 46/2210 Phil. 89/287 Legg. 105/409, over one- tenth in Euthyph. 6/44 Crito 2/14 Rep. 125/925 Theaet. 23/198 Parm. 38/298, over 5 % in Lach. 4/49 Gorg. 16/239 Phaedo 12/181 and less in Charm. 3/67 Meno 3/180 Euthyd. 1/68 Crat. 6/176. 452. Interrogations by ri prevail over those by més only in (p. 15): Phaedr. 12/2 Theaet. 15/8 Phil. 34/25 Legg. 58/55, while they are in all other dialogues much scarcer (being in Rep. 49/71 Parm. 9/29 Soph. 15/384 Polit. 22/24). 453. Interrogations asking for a better explanation of something said before (p. 16) are missed in many dialogues. They are found in: Lach. 4 Gorg. 1 Crat. 7 Rep. 62 Phaedr. 6 Theaet. 15 Parm. 8 Soph. 37 Polit. 42 Phil. 43 Legg. 63. 454. xad@s, xadN\toTa, dptota, 6pOas, épOcrara, Sixatérara, cai pad’ eixdros in affirmative answers with other verbs than Aéyets, namely, with eles, elpnxas, dv deyors, elmdy, eipyrat, form a class missed in earlier dialogues, but found in (p. 11): Rep. 8 Phaedr. 2 Soph. 8 Polit. 7 Phil. 8 Legg. 17. 455. xadas, xaddora, dpiocra, spOds, - dpOdrara, cadéorara, dAnOeorara, dvayxaérara, used as affirmative answers without verb, many ex- pressions in one class and counting them to- gether is somewhat arbitrary. He found that rheto- rical inter- rogations and also interro- gations asking for a better explana- tion of something said before are pecu- liar to later style. Camp- bell’s recent paper on the Par- menides shows that this dialogue has many words recurring only in the latest group and charac- teristic of Plato’s later studies or of his in- clination to intro- duce poetical 138 ORIGIN AND GROWTH OF PLATO'S LOGIC are limited to (p. 11): Rep. 59 Phaedr. 2 Thenet. 7 Parm. 18 Soph. 23 Polit. 38 Phil. 19 Tim. 1 Legg. 36. 456. eixds used in affirmative answers (p. 12): Lach. 1 Prot. 1 Meno 1 Gorg. 1 Crat. 3 Phaedo 5 Rep. 20 Theaet. 8 Parm. 2 Polit. 5 Phil. 8 Legg. 12 (in earlier dialogues go:xey prevails). 457. Instead of the ordinary formula gyovye Soxet appear later a class of other similar expressions (8oxet pot, Soxet ydp pot, pos Boxel, kat ewot Soxet, épot pev Soxei, kai €pot ovr@ Soxei, ovd eyol ciAAws Soxet, €uot your doxet), which are found in (p. 12): Lach. 1 Meno 8 Crat. 7 Symp. 1 Phaedo 1 Rep. 18 Theaet. 2 Phil. 2 Legg. 1. (See above No. 887.) XXXVIIT. CaMPBELL’S LAST OBSERVATIONS. AS Campbell was the first to apply the study of Plato’s vocabulary to Platonic chronology, so it happens that he also added thirty years later the final supplement to these investigations.’ The position of the Parmenides had been one of the most difficult problems, and had been recognised as such by C. Ritter, who was even led to doubt the authenticity of this dialogue. Campbell recently undertook to prove that, however exceptional the stylistic character of this dialogue may be, it contains a consider- able number of words peculiar to the latest group, or at least not used before the Republic, namely : 458. dreipia meaning injinitas: Parm. 1 Phil. 1 Legg. 1 (numbers according to Ast). 459. duapedera: Parm. 1 Critias 1 Legg. 1. 460. icov as adverb: Parm. 2 Tim. 2 Critias 1. 461. ioriov: Parm. 1 Legg. 1. 462. civdvo: Parm. 1 Tim. 1 Legg. 1 (in Symp. cv re dv’ quoted from Homer). 468. peptoréds: Parm. 2 Tim. 1. 464. pdvws: Parm. 1 Tim. 1. 465. rappeyeOns: Parm. 2 Legg. 1. 466. ravrovaras: Parm. 1 Legg. 1. 467. yupvacia: Theaet. 1 Parm. 1 Legg. 2. 146 J, Campbell, ‘On the place of the Parmenides in the order of the Platonic Dialogues,’ in the Classical Review for April 1896, vol. x. pp. 129- 186. This closes the list of forty-five publications on the style of Plato here reviewed, out of which only twenty contained materials suitable for our chronological purposes, and included in our list of peouliarities. PLATO'S STYLE: CAMPBELL 139 468. éuoiwpa: Phaedr. 2 Parm. 2 Soph. 1 Legg. 1. 469. axivnros: Rep. 2 Theaet. 2 Parm. 2 Soph. 4 Tim. 6 Legg. 8. 470. avdravAa: Rep. 1 Parm. 1 Phil. 1 Legg. 3. 471. dvopowrns: Rep. 2 Phaedr. 2 Theaet. 1 Parm. 8 Polit. 8 Tim. 2 Legg. 1. 472. avopom: Rep. 1 Theaet. $8 Parm. 8 Tim. 1. 473. dmewpos = infinitus: Rep. 5 Phaedr. 1 Theaet. 3 Parm. 9 Soph. 2 Polit. 2 Phil. 13 Legg. 8. 474. amépavros: Rep. 1 Theaet. 2 Parm. 1 Soph. 3 Polit. 1 Phil. 1 Tim. 1 Critias 1 Legg. 1. 475. dwéyw = disto: Rep. 1 Parm. 2 Tim. 1 Critias 2 Legg. 2. 476. damidavos: Phaedr. 1 Parm. 1 Legg. 1. 477. ampenns: Rep. 1 Parm. 2 Legg. 1. 478. BéBnxa=insisto: Rep. 1 Parm. 1 Tim. 2 Critias 1. 479. yodupa=liber: Rep. 1 Phaedr. 8 Parm. 7 Polit. 2 Phil. 1 Tim. 5 Critias 4 Legg. 10. 480. decroreia: Rep. 1 Parm. 8 Legg. 1. 481. diaxovw: Rep. 1 Parm. 2 Soph. 1 Polit. 1 Tim. 1. 482. dvahopdrns: Rep. 1 Theaet. 4 Parm. 1 Phil. 2. 483. e£icuvpar: Rep. 1 Parm. 1 Legg. 1. 484. eraveip = revertor, repeto: Rep. 3 Theaet. 2 Parm. 1 Polit. 8 Tim. 1 Legg. 4. 485. ov« edxodos = difficult: Rep. 1 Parm. 1 Legg. 2 (while in Rep. I 329 p, 330 a, edxoAos is used in another meaning). 486. evmerns: Kep. 3 Phaedr. 1 Parm. 1 Soph. 2 Legg. 2. 487. iyveuw: Rep. 1 Phaedr. 1 Parm. 1 Polit. 1 Legg. 2. 488. peOiorapa: Rep. 4 Parm. 1 Legg. 1. 489. pndapov: Rep. 1 Parm. 2 Soph. 1 Phil. 1 Legg. 8. 490. puxrds: Rep. 2 Parm. 1 Phil. 5 Tim. 1 Legg. 1. 491. mdmos: Rep. 3 Theaet. 2 Parm. 1 Legg. 3. 492. eoxcaypapnpevos: Rep. 3 Parm. 1 Legg. 1. 493. orépoyat, Med. : Rep. 4 Phaedr. 2 Theaet. 1 Parm. 2 Soph. 1 Phil. 1 Legg. 4. The following words occur also exceptionally in some earlier dialogue : 494. dvucos: Phaedo 1 Rep. 2 Parm. 5 Phil. 1 Tim. 5 Legg. 5. 495. auaorns: Phaedo 1 Parm. 8 Tim. 2. 496. deomo{w: Phaedo 8 Rep. 2 Parm. 1 Polit. 1 Legg. 5. 497. mavrehas: Phaedo 2 Rep. 9 Parm. 2 Soph. 1 Polit. 1 Phil. 1 Tim. 1 Legg. 2. 498. ovyxpiveoOar: Phaedo 2 Parm. 2 Tim. 4 Legg. 2. 499. coosrep: Gorg. 2 Rep. 2 Parm. 8 Soph. 1 Tim. 1 Legg. 1. 500. cipperpos: Meno 1 Theaet. 8 Parm. 2 Phil. 4 Tim. 5 Critias 1 Legg. 7. words into the language of prose. This valu- able addi- tion to our list makes the num- ber suf- ficient for a& more methodic interpre- tation of stylistic observa- tions than has been attempted heretofore. Limita- tion of our choice of peculiari- ties of later style to those for which complete enumera- tion of occur- rences might be found in the authors to whom we owe our facts. 140 ORIGIN AND GROWTH OF PLATO'S LOGIC Some other words quoted by Campbell, as pérpor, opnold, opyn, mwépas, wepiéyo, might be included in our list, as they occur besides the Parmenides only in later dialogues and occasionally in Meno and Cratylus. But for the purpose of drawing our conclusions from these long enumerations, a round number of five hundred stylistic peculiarities (including more than fifty-eight thousand observations) is more convenient, and suffices to show by what method correct chronological conclu- sions can be obtained from such observations. On the interpretation of stylistic observations. In selecting the above five hundred peculiarities of Plato’s style from the much greater number found in the writings of so many authors, the choice has been limited to characteristics occurring in one or more of the six dialogues held independently by Campbell, Dittenberger, Schanz, C. Ritter, and von Arnim to be the latest. Another limitation was imposed by the circumstance that the great majority of authors, ignoring the chronological bearing of their researches, often failed to state ex- pressly whether a collection of passages containing a certain word or expression was intended to be exhaustive, and such enumerations could not be included in our list, though they might have been very suitable for our pur- pose, and were perhaps looked upon as complete by the investigators. A further deficiency of our list results from the circumstance that nobody has made such a special study of the vocabulary of other dialogues as Campbell has of the Sophist and Politicus. This gives in the above enumeration a prominence to these two dialogues slightly exceeding the real proportional measure. Otherwise we may offer the above list as a fair and impartial sample of observations made on Plato’s style, prepared without any preconceived aim other than the knowledge of facts necessary for a methodical inquiry INTERPRETATION OF STYLISTIC OBSERVATIONS 14] into the question as to how far stylistic observations afford means of settling chronological difficulties. The method of interpreting stylistic observations has been heretofore very defective in almost all the authors reviewed. Generally little care or thought has been given to the logical co-ordination of results obtained through tiresome philological labour. It seems that the elementary conditions of a calculation of probabilities by their numerical evaluation were utterly ignored by all except Lewis Campbell. tic method in the eyes of impartial thinkers like Zeller. In order to obtain correct conclusions, future inquirers should avoid the following errors common to the majority of the authors above mentioned : 1. While a general notion of the necessity of mea- suring the length of each dialogue before comparing stylistic peculiarities was universally accepted, nobody tried to compare methodically the different possible measures ; and the pages of Stephanus or of Teubner were considered nearly uniform, while they differ widely, according to the number of notes in Stephanus and the more or less dramatic character of the text in Teubner’s edition: so much so that in the latter one page may contain twice as many words as another (see, for in- stance, p. 7 or 48, ed. Teubner, in Parmenides correspond- ing to thirty-four lines in Didot, and p. 425 in Politicus, occupying only twenty-one lines in Didot’s edition). That this may greatly influence our conclusions, we have seen specially in the case of Lina’s statistics of prepositions. Here for the first time a more precise measure has been found by comparing all the editions of Plato from Stephanus up to the present time. The pages of the editio princeps (Aldina 1503), though uniformly printed, are too large for a measure. Among modern editions the most equal pages convenient for comparison are those of the edition of Didot. These are used in the following calculations ; though the best measure would We need a better method for the interpre- tation of facts than our pre- decessors, who did This discredited the stylis- much logical labour. No uniform measure of text has been used, the pages of the editions generally quoted being unequal. Of all existing editions that of Didot has been found to afford the most uniform measure of text. The num- ber of peculiari- ties com- pared by most authors was insuffi- cient for valid con- clusions. Isolated observa- tions were given an exagger- ated im- portance regardless of the nature of statistical evidence which always requires great numbers. Even the greatest number of observa- tions used heretofore by Camp- bell would have been insuffi- cient if 142 ORIGIN AND GROWTH OF PLATO'S LOGIC be a hundred or a thousand words. This has not yet been applied to the text of Plato. 2. Nobody except Campbell had a correct idea as to the number of peculiarities required for correct conclu- sions. Campbell had compared hundreds of peculiarities and he was cautious enough to look upon his conclusions as only probable, not certain. Dittenberger and Schanz believed that a few important observations were sufficient for a stylistic classification of dialogues, wherein they came near to Teichmiiller and Schoene, who decided the question of style on a single stylistic peculiarity. C. Ritter was so confident after an observation of forty pecu- liarities of later style that he declined further discussion with those who did not recognise the correctness of his view. Even such a methodical author as Droste was led to a wrong conclusion about the Phaedo by a very small number of observations. Kugler doubted the authenticity of the Parmenides because he found a dozen more occurrences of pévro. than he expected in this dialogue. Von Arnim placed the Lysts after Sym- posium and Phaedo because he found ré/ pv once used in this small dialogue. All such conclusions are based on an erroneous conception of the use of statistics. Style statistics, like all statistics, require great numbers. Even nearly seven hundred peculiarities observed by Campbell were insufficient to determine the place of Theaetetus, Phaedrus, and Philebus. If Campbell avoided in an admirable way the smallest error in his conclusions, he owes it not alone to the number of his observations, but to his intuitive estimate of their importance. He dealt chiefly with very accidental peculiarities, words occurring only in two or three dialogues; and this explains why his great numbers were only sufficient for a determination of the latest group. In our own list we have many peculiarities of great importance, and thus, though the total number of peculiarities is smaller than in Campbell’s calculations, our conclusions INTERPRETATION OF STYLISTIC OBSERVATIONS 143 not only confirm his results, but extend over some earlier dialogues, as to the order of which nothing could be inferred by previous authors from stylistic observations. We must lay it down as a rule for future inquirers that no inferences from less than some hundred pecu- liarities are valid, and that the correctness of the inferences from smaller numbers of observations made by Ditten- berger, Schanz, C. Ritter, von Arnim, is due to the cir- cumstance that they selected exceptionally important peculiarities. 3. Nobody has hitherto observed that only exactly equal amounts of text should be compared in order to give precise conclusions. Dialogues of different size were compared, instead of taking as a standard measure a certain amount of text of each dialogue. For this purpose it is necessary to quote the passages in which every observed peculiarity occurs. As this has been done neither by Campbell, nor by Dittenberger, nor Schanz, nor C. Ritter, nor Tiemann, nor Siebeck, on whose observations a great part of our list is based, we are unable to introduce the required completeness into our calculations, but we shall make due allowance for the size of the compared dialogues, admitting as a rule that the stylistic comparisons are inconclusive unless the pre- sumed later work is equal or smaller in size. A greater number of later peculiarities in a longer work can lead to valid conclusions only under exceptional circumstances. 4, The different importance of stylistic peculiarities has not been accounted for, except by Campbell in one way, and by C. Ritter to a certain extent, when he dis- tinguished the repeated peculiarities contained in Republic, Phaedrus, and Theaetetus, as well as in the latest group. This distinction is quite insufficient; and at least four degrees of importance must be accepted in order to give us the full advantage of the existing observations. 5. Nobody except Campbell made a proper use of accidental peculiarities, which are far the most numerous not sup- plemented by a keen apprecia- tion of their 1ela- tive im- portance. Samples of text differing in extent were com- pared while only equal po- tions of text are com par- able. A greater number of peculiari- ties of style may be ex- pected in a larger work. A classifi- cation of peculiari- ties ac- cording to their degree of Import- ance is necessary. Accidental peculiari- ties were generally disre- garded, though they afford very valu- able mate- rial for statistical calcula- tions. Each pe- culiarity should be observed apart, not united with many others into one artificial class, ex- cept when the class as such is charac- teristic. This has been found only in a few cases, while some pe- culiarities counted together differ widely 144 ORIGIN AND GROWTH OF PLATO'S LOGIC class of observations. Very important peculiarities are very few, while accidental coincidences may be found by the thousand. And their accidental character, even if fully recognised as accidental, does not deprive them of chronological importance, if sufficient numbers of such accidental coincidences are taken into consideration. The single occurrence 1s accidental, though it may be ex- ceedingly significant, as, for instance, the occurrence of péOcEs in Parmenides and Sophist. But if one dialogue has twice as many accidental coinciden¢es with the Laws as another, this result is no more accidental than the difference of mortality between England and Spain. 6. The tendency to limit observation to peculiarities appearing to be important had the result that artificial classes of similar peculiarities were counted together. Sometimes such divisions are justified, as, for instance, Sie- beck’s classification of answers into apodictic, problematic, and assertive, or von Arnim’s rhetorical interrogations and interrogations asking for a better explanation. Also the classes of newly invented adjectives, or of adjectives desig- nating a species, are perfectly natural and characteristic. But in all such cases the single peculiarities forming a class should also be counted apart, whereby a much more exact numerical evaluation of affinity between different works might be secured. This has not been done by C. Ritter, nor by von Arnim, or at least they only give the total number of occurrences of different expressions not forming a natural class, as, for instance, vai, ravu ye, qavu pev ovv, which have nothing else in common than that they are the most frequent answers. This should be avoided in future investigations. Many very valuable observations were cast away as useless, because they did not show at once an evident difference between one group of dialogues and another. C. Ritter confesses to having traced through all the works of Plato many expres- sions, which he did not include in his tables, merely because they appeared not to be peculiar to well-marked METHOD OF STYLOMETRY 145 groups. All these observations have their value if they are treated by the right method. Method of measuring stylistic affinities. The above critical observations on the work of our predecessors are made in the hope that future inquirers will turn them to account. Our aim is not to add new facts, nor even to give an exhaustive survey of facts found by others. From Riddell’s digest of idioms, from van Cleef’s long enumerations, specially from Ast’s Lexicon, and from nearly all the publications above quoted, it would be easy to collect some thousands of style-charac- teristics, instead of the half thousand included in our list. But the mere enumeration leads to no valid con- clusions, unless we attempt an exact numerical defini- tion of the affinities existing between several dialogues. For a first attempt to find a numerical equivalent of stylistic affinity between various works not by mere counting but also by weighing of the evidence, we ‘needed a greater number of facts than has been known heretofore to any single author; but we found that five hundred peculiarities, selected at random from the special investigations, were sufficient for our purpose. We feel also justified in limiting the comparison to twenty-two dialogues of unquestionable authenticity, which at the same time happen to be the only works containing some hints as to the logical theories of Plato, while the remain- ing spurious or doubtful dialogues are of no logical importance. Still, so far as these other dialogues have been taken into account by some of the authors to whom we owe our facts, it appears that they contain a surpris- ingly small number of Platonic idioms. It is extremely exceptional to find a rare use of language illustrated by examples from other dialogues than those of admitted authenticity, even on the part of inquirers who had L from each other. A much greater number of pecu- liarities could easily be gathered from the authors reviewed, and a perusal of Ast’s Lexicon would raise the number to some thou- sands. But the aim of the present investiga- tion is only to improve the method of interpre- tation of facts, not to give an exhaus- tive survey of the facts observed. The most numerous class is formed by accidental pecu- liarities, occurring only once in a dia- logue. Such pecu- liarities acquire a chronolo- gical im- portance only when found in great numbers, though even & single oc- currence is some- times more or less signi- ficant ac- cording to the mean- ing of the word, and to the as- 146 ORIGIN AND GROWTH OF PLATO'S LOGIC searched all the texts bearing Plato’s name, including those which are generally recognised to be spurious. In order to draw our conclusions, we begin by recog- nising four degrees of importance, distinguishing stylistic peculiarities : I. The most numerous class are accidental peculiarities, such as words or idioms occurring only once in a dia- logue. Asa word cannot occur less than once, it is not less rare or less accidental when occurring once in a small dialogue than in a large one. In all such cases the ob- served coincidence is liable to be removed by some emenda- tion, or might be due to an alteration of text, this being less improbable with small words than with longer ones. Therefore great numbers of such accidental peculiarities are needed to afford a measure of comparison. Within this class it would be easy to distinguish several degrees of importance. Really accidental is the recurrence of a word which was generally used by other authors, but which denotes some object about which Plato had no opportunity of writing except in two or three of his works. If, for instance, Plato uses guddn only in Symposium, Critias, and Laws, this has no deeper reason than the accidental opportunity for the use of a word denoting a thing not usually spoken of by Plato. Such words have been generally excluded from our list, though they are not quite without value if they occur in very great numbers, as in every epoch the familiar circle of objects selected for examples is characteristic of the author’s turn of thought. It is, for instance, not quite accidental that xadxos is used six times in works later than the Republic, and only once in a work earlier than the Republic. Some- times a word used only once in a dialogue may be very significant, as, for instance, perdoyeois in the Phaedo (101 c). Thisis highly characteristic of a time when Plato was fond of inventing new logical terms, many of which were soon abandoned, like einacia, Stavora, mristis in the special logical meaning which was given to these terms METHOD OF STYLOMETRY 147 in the Republic. This period could not be that imme- diately following the death of Socrates, and it would be impossible to find a similar accidental occurrence in the Apology, while such new-formed words abound in the Phaedrus much more than inthe Phaedo. A word occur- ring only once in a dialogue is still more characteristic if it is of constant use in some other work recognised as late. But in order to avoid complicating our evaluations, and to eliminate from them as much as possible every subjective element, we count as accidental all peculiarities occurring only once in one dialogue, including in this class also those peculiarities whose number of occurrences 1s unknown, as for instance all rare words observed by Campbell in the Sophist and Poltticus. II. The next degree of importance belongs to pecu- liarities repeated, or occurring twice in a small dialogue (Euthyph. Apol. Crito Charm. Lach. Critias), twice or thrice in an ordinary dialogue (Prot. Meno Kuthyd. Crat. Symp. Phaedo Phaedr. Parm. Soph. Polit, Phil.), and two to four times in a large dialogue, such as the Gorgias, the Theaetetus, and the Timaeus. As to the Republic and the Laws, in dealing with these exceptionally large works we include in the class of repeated peculiarities every word or idiom which occurs twice or more, but less than once in twelve pages, as then it will be termed frequent. Thus the difference of extent is taken into account, although imperfectly, because the best method would be to take as a sample of style exactly the same amount of text from each dialogue. So long as we deal with each dialogue as a whole—and we are obliged to do so in consequence of the absence of detailed indications of passages in most of our sources—we are bound to the inconsequence of including in one class peculiarities of widely different degrees of frequency. A peculiarity occurring twice in the Huthyphro is found on average once in five pages, while one occurring twice in the Phaedo is found once in twenty-five pages. But all these repeated L 2 sociations it evokes. But these distine- tions must be left for more special investiga- tions, as they would introduce a sub- jective element. Another class is formed by repeated occur- rences. This class includes different degrees of repetition according to the length of each dialogue. It will in- volve no exaggera- tion to count each repeated pecu- liarity as equivalent to two accidental peculiari- ties, There is also in- cluded a greater frequency of some very common expres- sions. Important peculiari- ties are words occurring frequently and a number of special observa- tions on the pre- valence of one 148 ORIGIN AND GROWTH OF PLATO’S LOGIC peculiarities may be assumed to be more important than the accidental peculiarities, and for the sake of simplicity we count each as equivalent to two accidental peculiarities. If two hundred peculiarities of the first class were admitted as denoting a certain degree of affinity between two dialogues in which they are found, then we shall estimate @ common occurrence of a hundred peculiarities of the second class as equivalent evidence for an equal affinity. Here we include also the following special peculiarities : 354. mavv pev ody more than half as frequent as wavv ye, but not prevailing over it. 367. dras, cvpras, ovvaras more than once in three pages, and less than once in two pages. 390. Between 33 and 38 prepositions in every two pages. 412. mrép: after the substantive, forming between 6 and 10 % of all occurrences of zepi. 448. vai, wavy ye, wavu pev ovv being less than one-third but more than one-quarter of all affirmative answers. 451. Rhetorical interrogations between 5 and 10° % of all interrogations. These peculiarities might easily be thought more important than other repeated peculiarities, so that we do not incur the danger of exaggerating observed affinities if we count each of them as equivalent to two accidental coincidences between an earlier dialogue and the latest group, III. There must be recognised a difference between a peculiarity occurring repeatedly and one that occurs much oftener. Peculiarities occurring more than twice in a small dialogue (Apol. Huthyph. Crito Charm. Lach. Critias), more than thrice in an ordinary dialogue (Prot. Meno Euthyd. Crat. Symp. Phaedo Phaedr. Parm. Soph. Polit. Phil.), more than four times in a large dialogue (Gorgias, Theaetetus, Tvmaeus), once in twelve pages or more in Republic or Laws, form a class of important peculiarities. This class will include a word occurring 20-117 times in the Laws, 5-26 times in Theaetetus or Timaeus, and generally any frequent repetition up to METHOD OF STYLOMETRY 149 once in two pages (ed. Didot), when we shall gall it very frequent. Besides such peculiarities we include here the following special observations whenever they refer to a dialogue : 12. Being the first member of a tetralogy projected later—this refers only to Republic and Theaetetus. 18. Partial prevalence of other teachers over Socrates. This refers only to Symposium and Parmenides. For in Sophist Politicus Timaeus Critias Laws Socrates is already completely supplanted by other teachers, and this constitutes a more important characteristic. 16. Periods less regular. 17. Natural order of words inverted, as generally observed by Campbell. 18. Recurrence of rhythmical cadence, as generally observed by Campbell. 19. Balancing of words to achieve harmony and symmetry. 20. Adjustment of longer and shorter syllables, idem. 28. Words common and peculiar to Timaeus, Critias, Laws more than once in two pages, but less than once in a page. 200. Somep less frequent than xaOarep. 206. GAdad pny less frequent than xai pnp. 306. roivvy more than four times oftener than pévror. 307. yévro: less than once in two pages, but over once in five pages. 308. roivvy more than once in two pages. 317. eiov prevailing over éXeyov. 318. Answers denoting subjective assent less than once in sixty answers. 325. Superlatives in affirmative answers more than half as frequent as positives, but not prevailing over positives. 354. mavu pév ovy prevailing over mavu ye. 365. vymras prevailing over dmas. 366. ras and compounds between four and five times in one page. 367. das, EUpras, vvdras more than once in two pages, but less than once in a page. 376. Apodictic answers between 30 and 40 % of all answers. 377. To each problematic answer between three and four apodictic answers. 378. Interrogations by means of dpa between 15 and 24 % of all interrogations. $89. xara c. accus. prevailing over all other prepositions except év. 390. Between 19 and 21 prepositions in one page (ed. Didot). synonym over another or on some ger ral properties of style or literary composi- tion. This class includes also higher degrees of frequency of very common words, and other peculiari- ties enu- merated, observed by various authors. Each of such im- portant peculiari- ties will be counted as equivalent to three accidental or to one repeated and one accidental pecu- liarity. 150 ORIGIN AND GROWTH OF PLATO'S LOGIC 391. epi c. accus. prevailing over zepi c. genitive. 412. rép: placed after the substantive between 10 and 20 % of all occurrences of repi. 448. vui, ravu ye, wavu peév ovv less than one-quarter of all affirma- tive answers. 451. Rhetorical interrogations between 10 and 20 % of all interrogations. 452. Interrogations by ri prevailing over those by mas. All these peculiarities are much more important than those of class II, and each of them will be estimated as equivalent to three peculiarities of class I, or to one of class II and one of class I. A fourth IV. There remains a class of peculiarities still more class is significant, of which a small number is equivalent to formed by more than thrice that number of peculiarities of class I. piss To this belongs first a very frequent occurrence of any currence WOrd or idiom, as for instance, 118 times or more in the of any Laws, 97 times or more in the Republic, generally more word. than once in every two pages (ed. Didot). Besides, we Very fre- include here a small number of the most characteristic anent we peculiarities of style, namely : term the occur- 12. Belonging to a tetralogy as second or third member. rence of 13. Complete substitution of other teachers for Socrates. any word 14. Didactic and authoritative character. snesin 15. Quotations of earlier dialogues, preludes and recapitula- two pages. mone. Biue é To this 21. Avoiding of hiatus. 23. Occurrence more than once in a page (ed. Didot) of rare class be- words common and peculiar to a dialogue with Timaeus, Critias, long also Laws. some 307. yevro: less than once in five pages. special ob- 818. Answers of subjective assent entirely absent. servations 325. Superlatives in affirmative answers prevailing over corre- enu- sponding positives. merated. 366. was and its compounds over five times in a page. Each very 367. das, Evyras, fvvdras more than once in a page. important 876. Apodictic answers more than 40 % of all answers. pecu- 377. Problematic answers fewer than one to four apodictic Nirity wil S78. Tn tions b than 24 % of all interrogati be enents . Interrogations by dpa more than 24 % of all interrogations. 889. xard with the accusative prevailing over ¢v. METHOD OF STYLOMETRY 151 390. More than 21 prepositions in a page. 412. mépc placed after the word to which it belongs forming more than 20 % of all occurrences of epi. 451. Rhetorical interrogations forming more than 20 % of all interrogations. All these peculiarities being very important, it will be fair to count each as equivalent to two repeated, or to three accidental, or to one accidental and one important peculiarity. In the above classification of peculiarities we have endeavoured to reduce to a minimum the relative im- portance of each peculiarity, in order to avoid every exaggeration of the measure of affinity uniting two dia- logues. Any error committéd will thus rather diminish the apparent affinities than increase them. If a word occurs once in each page, or more than two hundred times in the Laws, this will be counted as only four times more important than a single occurrence. Later inquiries may prove that this is a very low estimate of the importance of frequency. But any classification of stylistic pecu- liarities according to their importance must take into account that importance is very far from being propor- tional to frequency. If one word occurs ten times in one dialogue and ten times in another, this is very far from being a link equivalent to ten single occurrences of ten different words in both dialogues. Our classification is here proposed not as definitive, but only as a first attempt at a numerical evaluation of stylistic affinities. Future inquirers dealing with many thousands of compared pecu- liarities may find reasons for a different classification. As our purpose is only to find the lowest figures, which may be increased later, but can never be diminished, the above distinction of four degrees of frequency and importance is sufficient. Now, in order to apply our method, we must state clearly the highest hypothesis on which it is founded and define its terms. This highest hypothesis has been here- as equiva- lent to four accidental pecu- liarities. The above standards of equiva- lence are minimal, in order to avoid ex- aggeration of affini- ties. Im- portance is not pro- portional to fre- quency, and in- creases at @ much smaller rate. The above classifica- tion is not definitive, and aims at deter- mining the minimal value of affinities. Law of stylistic affinity. The num. bers of observed peculiari- ties in two works must differ at least by one tenth for valid chronolo- gical infer- ences. The total number of peculiari- ties dealt with should ex- ceed 150 in a dia- logue of ordinary length. The Laws standard of com- parison. 152 ORIGIN AND GROWTH OF PLATO’S LOGIC tofore tacitly admitted, but has not been methodically dis- cussed. It is the following LAW OF STYLISTIC AFFINITY. Of two works of the same author and of the same size, that is nearer in time to a third, which shares with wrt the greater number of stylistic peculiarities, provided that their different importance is taken into account, and that the number of observed peculiarities ts sufficient to deter- mine the stylistic character of all the three works. As to the meaning of terms in this psychological law the following may be observed : 1. Nearer in time implies nothing as to priority, unless independent evidence is forthcoming that some one work of the author is the latest. In Plato’s case the Laws are generally admitted to be such a work. But even were this doubted, a very great number of peculiarities observed would finally lead also to the de- termination of an order of priority, because the more varied style of an author has every chance of belonging to a later time. 2. A greater number of peculiarities does not mean any greater number, because if the difference is insig- nificant, no valid inference is allowed. We accept pro- visionally, as a minimum of difference between two works justifying chronological inferences, a difference of one- tenth of the observed peculiarities, and in some special cases we shall even require a greater difference. 3. A sufficient number to determine the stylistic character must be a greater number than has been used generally heretofore, except by Campbell. But this de- pends upon the importance of each peculiarity. In the present case we shall assume that the occurrence of fifty out of five hundred peculiarities allows a probable infer- ence, but that this probability approaches certainty only when a hundred and fifty peculiarities of later style are found in an ordinary dialogue. 4. The Laws are our standard of comparison for the next latest five dialogues, and for earlier works the group LAW OF STYLISTIC AFFINITY 153 of the six latest dialogues, Sophist, Politicus, Philebus, Timaeus, Critias, Laws. If now we ask how the law of stylistic affinity can be verified, the first and nearest answer lies in the psycho- logical property of style as a mark of identity, entirely depending on the totality of familiar expressions at any time in the writer’s consciousness. Every writer could find easily in his own experience sufficient evidence in favour of this psychological law. It has been suggested that it ought to be tested on the writings of a great modern writer like Goethe, as we know exactly when he wrote each of his works. But this way of testing it would cost an immense additional labour, and would still remain inconclusive, since an obstinate sceptic might object that the psychological development of Goethe differed from that of Plato—that the German language has peculiarities distinct from those of the Greek language, &c. We propose, therefore, another and better way of test- ing, with special reference to Plato, the law of stylistic affinity, and at the same time also our own classification of stylistic peculiarities, which is subsidiary to our chronological conclusions, and requires even more strict verification than the psychological law, which will appear obvious to many readers. We have sufficient means of testing our method, if we take into account that, however little is positively known in Platonic chronology, there are some works connected by Plato himself into tetralogies, and there- fore necessarily following each other, though perhaps at intervals. Further, there can be no doubt that the successive parts of a larger work, as a rule, must have followed each other, at least if the later part contains clear allusions to the preceding text. If, then, our method yields conclusions in agreement with these evident facts, we may confidently apply it to the solution of more difficult problems in Platonic chronology. We submit, The law of stylistic affinities must be tested. A test on another author than the author to whom we apply it would be incon- clusive. We have means of testing our prin- ciples and their con- sistency on the works of Plato, among which some are positively known to be later than others, for instance, the con- tinuation of a larger work is later than its be- ginning. Such tests could be collected in great numbers if all the authors had quoted the passages counted. A certain number of tests is taken from the com- parison between earlier and later books of the Republic except the sixth and seventh books, with the following eighth and ninth, as their order is not quite certain. Any book is later than the first, the 154 ORIGIN AND GROWTH OF PLATO'S LOGIC therefore, to the impartial judgment of our readers the following tests : 1. The first tetralogy sketched out by Plato consists of Republic, Timaeus, Crittas (unfinished), with the Her- mocrates, which was projected but never written. We begin by comparing the first with the last book of the Republic, because some intermediate books have been con- sidered by certain critics as later additions, while nobody doubts that the tenth book must be somewhat later than the first. We find in the first book 28 accidental, 6 repeated, and 3 important peculiarities of later style, amounting together to 49 units of affinity. In the tenth book, which is a little smaller and offers therefore even fewer opportunities for the occurrence of each peculiarity, we find 35 accidental, 14 repeated, 15 important, and 6 very important peculiarities of later style, equivalent to 132 units. For the sake of conciseness and easy com- parison we express this stylistic relation in the follow- ing formula : 1. Rep. I 827-854 (20} pp. Did.) : 28 (I) 6 (II) 8 (III) = 49 (1) ~-»Rep. X 595-621 (19} pp. Did.): 85 (1) 14 (II) 15 (III) 6 (IV) = 182 (1). 2. It is equally certain that the fourth book of the Republic must be written later than the second. If we take for comparison two samples of text of a size nearly equal to the Symposium, we find the following stylistic relation : [Symposium 172 4-228 p (89 pp. Did.): 42 (I) 16 (II) 8 (III) = 98 (1).] 2. Rep. II 857 a—III 412 a (874 pp. Did.): 47 (1) 20 (II) 22 (III) 2 (IV) =161 (1). —>Rep. III 412 s—V 471 o (89 pp. Did.): 45 (1) 28 (II) 81 (III) 2 (IV) = 192 (1). (Cf. Phaedrus (89 pp. Did.): 54 (I) 86 (II) 22 (IIT) 7 (IV) = 220 (I).] 3. The above two tests can be confirmed also by com- parison of larger samples. If we compare the last three THE STYLOMETRIC LAW TESTED 155 books of the Repwhblic, equal in size to the Theaetetus, with an exactly equal amount of B. II-IV, we find the following stylistic relation (the indications about the style of other dialogues are of course quoted here not as tests, but only for comparison) : 8. Rep. II 868 a—IV 445 & (58 pp. Did.) : 47 (I) 30 (II) 32 (III) 2 (IV) =211 (1). —>Rep. VITI-X (534 pp. Did.) : 54 (I) 36 (II) 29 (ITT) 5 (IV) = 238 (I). [Theaetetus (53 pp. Did.) : 58 (I) 41 (II) 81 (III) = 288 (1).] 4—7, As there is no doubt that the single books of the Republic were written in their present order (except B. V-VII, which are supposed to have been completed last of all), we may compare different parts of almost equal length, in order to see whether the later text always offers more peculiarities of later style. Such comparison will be easily appreciated in the following short enumera- tion : 4. Rep. I 827-II 367 = (28 pp. Did.): 86 (I) 10 (II) 8 (III) = 65 (1). — Rep. II 868 a-412 a (80 pp. Did.) : 42 (I) 17 (II) 22 (III) 2. (IV) = 150 (1). Cf. Euthydemus (28 pp. Did.): 22 (I) 5 (II) 7 (III) = 53 (I). 5. Rep. II-IV (60 pp. Did.): 47 (I) 87 (II) 82 (III) 2 (IV) = 225 (1). —Rep. V-VII (60 pp. Did.): 56 (I) 29 (II) 40 (III) 7 (IV) = 262 (1). Cf. Gorgias (60 pp. Did.) : 31 (I) 20 (IT) 6 (ITT) = 89 (I). Cf. Phaedo (49 pp. Did.): 48 (I) 26 (II) 17 (III) 2 (IV) = 154 (1). 6. Rep. II 857 a-III 412 a (874 pp. Did.) : 47 (I) 20 (II) 22 (IIT) 2, (IV) =161 (1). —Rep. VIII-IX (84 pp. Did.): 47 (I) 22 (II) 27 (IIT) 3 (IV) = 184 (I). 7. Rep. II 868-IV 445 = (58 pp. Did.): 47 (I) 80 (II) 82 (III) 2 (IV) = 211 (1). Rep. V 471 p-VII 541 (44 pp. Did.) : 50 (I) 21 (II) 88 (III) 7 (LV) = 234 (1). In the above seven test cases the earlier part has always fewer peculiarities of later style, and in every case fourth is later than the second, etc. In each case the earlier text has fewer peculiari- ties of later style, the evi- dence as to priority being given by Plato himself. The same results from a com- parison between parts of the Republic and the dialogues which are later, for instance, the Timaeus and Critias. The Critias being much smaller than any book of the 156 ORIGIN AND GROWTH OF PLATO’S LOGIC the evidence of priority is given by Plato himself, as we compared the acknowledged continuation with the pre- ceding text. We excluded from our comparisons the relation of B. V-VII to the following books, because this part of the Republi in its present form has been supposed to be later, and cannot therefore be used as a test case. Many other parts of the Republic could be compared with equal lengths of text undoubtedly later, but the above seven samples give a sufficient notion of the text of the Republic, and we may now proceed to compare the Republic with Timaeus and Critias. A direct comparison between Timaeus and Critias is im- possible, because the size of the two dialogues differs too much. 8. In order to compare the Republic with the Ti- maeus, a good test is afforded by the last three books, which are equal in size to the Timaeus: 8. Rep. VIII-X (58} pp. Did.): 54 (I) 86 (II) 29 (III) 5 (IV) = 288 (I). a (58 pp. Did.): 128 (I) 58 (ITI) 44 (IIT) 14 (IV) = 4217 (I). 9. The Critias is almost too small for any comparison, being scarcely longer than half a book of the Republic. It is certain that the Critvas is later than the last book of the Republic, and if notwithstanding its small size the Critias has more peculiarities of later style, this gives an evident confirmation to the law of stylistic affinity, and to the rules above admitted. We find: 9. Rep. X (194 pp. Did.): 86 (1) 14 (II) 15 (III) 6 (IV)= 182 (I). ->Critias (11 pp. Did.): 61 (1) 8 (II) 18 (III) 12 (Iv)= 169 (1). This test is specially important, because we have taken the last book of the Republic, apparently separated from the Critias only by the length of the Tumaeus, and we have found that to the chronologic distance there corresponds a considerable stylistic distance between the THE STYLOMETRIC LAW TESTED 157 two works. We might add as test comparisons each of the other books of the Republic, and we should find that the Critvas exceeds them all in number and importance of peculiarities of later style. But this being evident after our preceding comparisons, we need not insist upon it. 10. In order to compare the Laws with the Republic, we must allow for the difference of size, the Laws being 43 pp. (Did.) longer. If we add the Gorgas to the Republic, we obtain a whole slightly exceeding the Laws in size and affording a convenient comparison, because nobody doubts that the Gorgias and Republic are both earlier than the Laws. It results: 10. Gorg.+ Rep. as one whole (256 pp. Did.): 76 (I) 124 (IT) 80 (III) 4 (IV) = 430 (1). ~»Laws (288 pp. Did.) : 175 (1) 176 (II) 87 (III) 20 (IV) = 718 (I). The Laws being acknowledged as the latest work of Plato, many new tests would result from a comparison of the Laws with different combinations of other dia- logues equal together in size to the Laws. But as our list has been compiled on the principle of a selection of peculiarities of later style, and the standard of later style has been taken from the Laws and those other works which in style come nearest to the Laws, it might be denied that such tests confirm the law of stylistic affinity. 11. We turn to the other tetralogy indicated by Plato himself, and beginning with the Theaetetus. We com- pare first the Theaetetus with the Sophist, which is its recognised continuation according to Plato’s own indis- putable testimony : 11. Theaet. (58 pp. Did.) : 58 (I) 41 (II) 81 (III) = 288 (1). — Soph. (40 pp. Did.) : 189 (I) 86 (II) 59 (III) 20 (IV) = 468 (I) 12. The Sophist and Poltticus are as closely connected as if they were one dialogue, and still there is a difference Republic shows & style later than even the last book of the larger but earlier work. To com- pare the Republic with the Laws, we must add some text to the smaller dialogue. Then we find that the style of the Republic is much earlier than the style of the Laws. The two dialogues which were written by Plato as continua- tion of the Theaete- tus also show a much later style. Further tests are given by compar- ing those dialogues about the relative date of which there is a general agree- ment, for instance Phaedo with the preceding Meno, or Philebus with the preceding Par- menides. Also in this case our method confirms the best informa- tion obtained otherwise. 158 ORIGIN AND GROWTH OF PLATO'S LOGIC of style between them, the latter having more peculiarities of later style : 12. Soph. (40 pp. Did.): 189 (I) 86 (II) 59 (III) 20 (IV)= 468 (I). -»Polit. (43 pp. Did.): 168 (I) 48 (II) 56 (III) 19 (IV)= 498 (I). 13-14. The above twelve test comparisons refer to samples of text, for whose chronological order Plato himself has given clear indications. They confirm the law of stylistic affinity as well as the rules laid down for the application of this psychological law, including our classification of stylistic peculiarities according to the degree of their importance. We need not pause here to test our fundamental principles. There are some pairs of dialogues, which, though not forming one whole or not continuing each other as the above, are recognised as standing in a certain chronological relation because one of them contains allusions to an exposition which appears in the other. Many of such allusions are disputable, but there are at least two which are sufficiently recognised by all competent authors, includ- ing Zeller, to justify their use as tests. These are the allusion found in the Phaedo (72 E) to the theory of reminiscence first set forth in the Meno (82 B-86 a), and the allusion of the Philebus (14 c) to the difficulties of defining the relation between the One and the Many which are nowhere treated with such consciousness of the complexity of the problem as in the Parmenides (129 B-E and the whole dialogue). If now we compare the style of these four dialogues we find again a complete agreement between our own method of settling chrono- logical difficulties and the most certain hints about the order of some dialogues obtained otherwise : 18. Meno (28 pp. Did.) : 20 (I) 16 (II) 8 (IIT) = 61 (I). —»Phaedo (49 pp. Did.): 48 (I) 26 (IT) 17 (ID 2 (IV) = 154 (I). THE STYLOMETRIC LAW TESTED 159 Here the difference of size could not be accounted for, but 1s compensated by the very great difference of style. 14. Parmenides (831 pp. Did.): 56 (I) 42 (II) 21 (III) 10 (1V) = 248 (I). —>Philebus (48 pp. Did.) : 100 (I) 38 (II) 55 (ITI) 16 (IV) = 405 (I). Here also the difference of size 18 more than com- pensated by the great difference of style. 15. Other similar allusions are too uncertain, and sometimes evidently mistaken, so that we cannot use them as tests. But to remain within the limits of the greatest probability, we may take for granted that the three small dialogues referring to the death of Socrates— Apology, Euthyphro, Crito—are earlier than the Sym- postwm which nearly equals them in size. We find: 15. Apology Euthyphro Crito as one whole (41 pp. Did.) : 21 (1) 7 (II) 6 (IIT) = 58. —Symposium (89 pp. Did.) : 42 (I) 16 (IT) 8 (IIT) = 98. 16. It were easy to increase the number of similar tests by many others, taking the whole of Socratic dialogues as certainly earlier than Philebus, Timaeus and Critias, and our list offers sufficient material for comparisons which can be readily made by those of our readers who think that the above fifteen trustworthy tests are insufficient. We add only one test of a different character, in order to show how the coincidence of accidental characters operates on greater agglomerations of texts. Were our method wrong, it might happen that a certain number of single dialogues, each of which has been found earlier than one of the dialogues of another group, taken together as one whole and treated as to the distinction of degrees of importance in the same way as the Republic, would appear later than the group consisting of dialogues which taken individually are later. Now, a good test of the consistency of our method A similar test is offered by the three short dialogues referring to the death of Socrates, which are earlier than the Sym- postum. An im- portant test of con- sistency is found by comparing groups of dialogues A group of dialogues which in- dividually contain more peculiari- ties of style need not neces- sarily con- tain also & greater number of peculiari- ties if the greater number were not caused by the later date. Our method thorough- ly tested, as no stylistic 160 ORIGIN AND GROWTH OF PLATO'S LOGIC is to form two groups of dialogues, one consisting of dialogues which by individual comparison have been found to be earlier than the Republic, and the other of those which have been found to be later. Then, if our method and our rules are correct, the later group must show a greater number of peculiarities of later style than the Republic, while the earlier group must also have a smaller equivalent of affinity with the later style. Those dialogues which according to individual stylistic tests precede the Republic are the Phaedo, Symposium, Cratylus, Gorgias, amounting together very nearly to the size of the Republic. On the other side, the Theaetetus, Parmenides, Philebus, Timaeus and Critias form a group also equal in size to the Republic, and consisting of dia- logues of which each has been found later than some part of the Republic. If we compare both groups with the Republic, counting as important only the peculiari- ties which occur in each group, at least so many times (17) as is needed to call them important, if they occurred in the Republic, then we find the following results : 16. Gorg. Crat. Symp. and Phaedo as one whole (191 pp. Did.): 60 (I) 84 (II) 8 (III) = 242 (1). -»Republic I-X (195 pp. Did.): 81 (I) 110 (II) 80 (III) 4 (IV) = 407 (I). —Theaet. Parm. Phil. Tim. and Critias as one whole (191 pp, Did.) : 107 (I) 210 (II) 40 (III) 9 (TV) = 688. This test of consistency has also an independent value for many competent Platonists who recognise that the Republic is later than Gorgias, Cratylus, Symposium and Phaedo, but earlier than Theaetetus, Parmenides, Philebus, Tumaeus and Cretras. Now, having thoroughly tested our instrument of in- quiry, it is fair to apply it to those more difficult pro- blems of Platonic chronology, on which other investi- gators have heretofore failed to agree. First as to the date of the Theaetetus tetralogy, it results from the above, THE STYLUMETRIC LAW TESTED 161 that the Theaetetus must at least be later than the first books of the Republic (see test comparison No. 3). The difference of style between the Theaetetus and the last books of the Republic is too insignificant to allow direct chronological conclusions, though it shows also that the Theaetetus has a greater number of peculiarities than B. VIII-X. In order to decide whether the Theaetetus is later than the whole of the Republic, we shall be obliged to have recourse to a ‘longer way’ than our present method. For the present we must be content to say that the Theaetetus is evidently later than the Symposium and Phaedo, as can be seen from the above tests 3 and 5. A further important result from the validity of our method is that the Phaedrus is undoubtedly later than the Phaedo, and the Phaedo later than the Symposium (see above tests 2 and 5). For the relation between the Phaedrus and Theaetetus the above obser- vations afford no sufficient basis. Many new investigations are needed to settle all details with the complete certainty which the above reasoning shows to be possible in chronological inferences from stylistic observations. The present calculations, based on the work of others, are by no means sufficient to determine the order of all the works of Plato. For this it would be necessary to have a list of stylistic peculiarities ten times longer than our list of 500 stylistic characters, among which only very few are important, the majority being accidental. In order to enable the reader to extend comparisons similar to the above to other dialogues and groups of dialogues, the measure of relative stylistic affinities 1s given in the following table, which supple- ments Campbell’s and C. Ritter’s similar tables by a methodic co-ordination of over fifty-eight thousand facts, collected by twenty authors, of whom none knew more than a few of his fellows : method before, has been found con- sistent and trust- worthy. It holds good in doubtful cases. But cer- tain diffi- culties remain unsettled. What our calcula- tions prove may be seen from the following table of affinities, @ con- densed ex- pression of over fifty- eight thousand facts hitherto little known. ORIGIN AND GROWTH OF PLATO’S LOGIC 162 “(m9 oP *n'? °°) mr GPT mI : L96—TII | ‘day ut pejonb 40u | sesesseg *(,7 doy) Gory] »qdéqing : SIF I (@ np * * "q) “MIBIO OD :OgS—TT | ‘pATynG OUST : S¥P | + a "(*q) p'tpeeyd “Yow : Sip ‘(up $0 8) yorg ‘qddyqing : ITE ‘(89 ,,8) OUST “GOVT : 96S | one. (0 oP or? 1-10) OFFD soul} 6-G Suto “(® mP wr? orn) wMAIVYD OFID : 1ES | 90 serqLTeTNO4 (pte *q 8) ‘qoey ‘mueqg :¢ —T | goo | ot | —|z 18 | 6 |tedy=4boj0dp 1 | dOO0dr) OMLVHOOS ‘I | | ‘queqz0d ult ' ‘91q8} OU} SULMOT[OF SUOTyeUYTdxe 098 S[THIEp | ‘ 4 TGQ)“ °q) MO “3080N], : I6T ‘yovoyL, “Ipesyg : 98I—'I ee eressettomnaseteamentions ‘yuejzoduar 7 ‘poyeed | seul} 61-7 Supumm9 |} | -00 sepLIEINoeg Pid ag 8-68 —|8 | 9L| @& wnisoduhg “TT | { PI-0 86 -O1 G ‘TeymepIoe ET : SojleNoed Mou JO [BIO], | nr 310d ar Gdog : 68g wS90T : LLB (0 'D 8q) ‘B80r] WHY, : $28 “(PhO "q) pr" }OPOML mIPVY : S7E—TII ‘(p) ULL ld : SOP “() ,govoqy, ‘dardg : Tg ‘(9 mp ° ' *q) ‘rpeeyg opesyg : $9Z “(m9) IPs ,OPevyd : HH ‘(m® mr?) mopesyd “duxdg : ggg—TI *(7q) yy AP2VUd mn duke - SOP 1239] : 90h V(eq) ured “apeedd +268 wd 1 "qdog : ggg | . (pf 2 49) p'ydog -rpeeqg : OPE "| | | | (0) ‘IITog opesyg : ggZ | } “yuBz10d ult ‘guejzodum 7 ‘peyved ! 170 ORIGIN AND GROWTH OF PLATO’S LOGIC ‘(0 4-14) p'}ORONL py Apeeyd Eze | | (n°) wd “qdog : 69g | | ‘SBD 1 WEL ? ZL3—I1 n Oly, p UIvd : o6p | ‘(p °0) "TY m'WIBd : HEF "WL “Id * BSP "(9) "wry, “‘qdog : Lap ——— _ yp UIs “Jovouy : (oe) “Tad “qdog : "Ba0'T : -‘qureq10dur (0 p) ‘qdog ,spoeyg : 99g | 8001} +h Sapo ‘(2 p) ‘ydog “zpoeqd : F1Z Baie oh ate) ‘PIC “pe “dd ¢.67 (Pp) 310d “apeeyd :¢ —'T 10 | ET | S | 22 | 98 | SF ‘Opmvud “SI SS i § | See ‘81439 94} SutAoyoy suorsusidxe oes syreqop || . = S FA i | nS XT TITA Sra =P ‘TIA-A 8400g = 0 eT} WO fo Hy < Ee : 2 AT-TT 400g = q ‘J 0G “Coy = 8 “seqoe1q Tt POPPE £1049°[ || Home |! egswn Jo oe! 5 3 -1q¥s Srq3 Ur e+.| 3S = UWIOY} 10} posn SsuoTFVIAGIG B 7| & -q8 pus ‘Jepio peoLsofou enh peumserd 2116q3 UT oppep 943 fo suUDN Trews Aq poxreur st oqndey oy} jo yred yoo ut Azizeroed yoo “cour |! zequrou JO 90UeLM900 oT} OSTY ‘“sIMoe1 Aqrret~oed cures By} Goma Ut dnox8 | Sutmo sonsO[VIp 4xou O44 OY} JO SoUIvE poyBIAeIqg’s ey} rvedd¥ 1oqunU 98048] -{0} os ST} Purge” §‘seryirerfnoed onsyAys porpuny eay JO 48T] eA0gs ou} 04 | 04 quo] om} Ut toquinu yuexmMo s3t Aq poysusisop st AqreITNOed TVA || Keay || arabe ‘enZorerp yous ut "aug pif oy, Lof aauynposy,| payoy, || PUROF e[A9s 1048T Jo sasz enboynyg yove ur Gursveddy sairsoynoad fo uovsoMmnug ° | ~Loynsad fo LIQuuenn "Teyaeprooy ‘T i H “penuiyjuoo—OLVIg 4O SHODN0TVIG OML-ALNGMT, HOd ALINIGGY OMLSTIALG JO WIV, 171 THE STYLE OF PLATO | ‘SOT]LIUITNOed MOU [VI Ueploo” 7 SopIseq UTe}W09 pNoA | | | qI WON, ‘SBISION 944 S1OJoq U9z}IIM JT -o1 g ‘[eyUepioo’ Z : SojteMoed Mou Jo [e,O7, *(o) ‘uareg 1 90R0UY, : ‘(x9 mP oy? kG) mJ VOTL 1 IPOVT : ‘(§ $9) ‘govoqy, ,Apeeyd ‘ydog ,,UuIeg mn qdog ,,"Ipeeyd m 0d 1’ qdog mr WIS *JoV0T" TT, "(q) ‘wre ,J0v0T], ‘GaP £9) md n'Ipevqd : | ‘quejiodu ¢ ‘poyved -O1 { ‘fayuepro0"e JJ : Seytrernoed Mou Jo [e{O7, * 968 | | “(4x9 moP Sar? SEQ) py" QOVOTT, py TPOVTT "(o) WEY, “0d ‘(axP mh? tq) “Td “}10d nn UL, py UAB (P f9 {-n9) ‘Ydog ,,"maeg ‘(oP) “qOog ‘weg "(2) *S8er] py WALT, ‘8Sé—TII : 867 ‘ LOF : 96% ‘6IP ‘696 *‘poyeod : | "gueq}20d um 16% | | och—II ! : $65 ; : 18h : $9F : OFF ay || | Seton : OFF | ! ) Mes 2 —I | 400 | 6 |—|F& 19 | 8 ———S ee "PEC “Pe ‘dd 9.0% ‘s=1‘g onqndary “St 172 ORIGIN AND GROWTH OF PLATO’S LOGIC ‘() Td ‘yevoq], : Lee ‘(p) pS8or] ‘UY, : 61S 7 380T “JOR0UL, STZ ‘(*q) ‘yd ‘Ydog : gTZ (0) ,'QoBeUL, “APOVY : SET *(') ‘arog ‘ydog : ggI—T : (‘dd 46%) V2Iy—ve98 ="q || 12-0 cote gray C2) UTR “THO | 987 | (x9 mP ‘nr? Sq) 2'9°POUL mod * SHH — II S80] 28h © ‘(°4) m'Adog n'tpeedd : HEE “UL, “Ipeeyd = L63—T | : (‘dd $2) 329g—V1L¢¢ = '4 | £0-0 | I; T&-0 "e1qB} 04} Butmoyjoy suorysusdxe ses s[reyop i SMB] Tey} 19g = “X FOO = 9 ‘XI-TITA *HO0g = P “TLA~A 8¥00 = 2 | ey} U0 ‘AI-II syoog = q ‘J qoog ‘doy =e ‘s7e_81q Ul pepp¥ 819349] porns Trews Xq peyszeur et orqudey oq} jo 41ed qove uy Aqurermoad yowe || you, jO eoueLMI00 ey} OsTy ‘simoer AyBITMOed oUIUS O43 YOM UI |} GA oig SONFO[VIP {xXou OA} OY} JO SolUBU pPoyetAoIqqs oq} avedds 1equnu 80} Sig} purged ‘saryiiermoed onsriyAys porpuny ea jo 4Si[ eA0gs 2 a 04} Ul Joquinu yuermo sy Aq poysusisep st Aqirermoed youg figungfo amr yest ayz sof enborwrg yove ur Bur.svaddy sarizeninosd fo uoivsounugy OST 1 ‘fijzruyfo fo 872Un JO sequinu [| SULMOT -[O} oq 0} 7UO] -sainbo NR ee NN EN ee et ee ey ae : — mH : ° 6 | 62 | LI | O ¢ {Ir “FUBIOC UIT | SOUT} GZ—G SurM9 “90 SOFFUTNIe | GG || 3 | GF | LE | LP ‘TeyAepIooy ‘T 24vryndad fO AOQUINNT . i | Punoy o[44s 1048] JO £993 ‘pig ‘pe ‘dd $6z = VoLt—Veos ="q "pig ‘pe ‘dd $4= 1/98—VL98 = "q NOU OINOL -Vig Wiaaly ‘III "O[Q¥} STqy UT UOT} IO} POSN SUOTZBIAGIG qe pue ‘xepzo peorsojou -orqo peunseid r10q4 UI sonBorosp 94} fo sousoN ‘penurjuod- OLVIg 40 SHNDOIVIG OMIL-ALNAMY, AOI ALINIAY OMSITALG 40 ATV], 173 THE STYLE OF PLATO SCO. ‘soljlzerjnoed Mou jue}IOduIt ¢ ‘pozeed e]OyM eu s8q Uy ‘soul [[-7 *q ul ‘som F[-f Sut1N900 gaT}IeIpnoed ®q ul ‘yuBjOduI O18 SOUT, G SuUTIINIIO : -o1 G ‘TeyUeproo" ge Imo00 (°q + °q + "q) | sonjternoed ‘q uy “*q ut payeodel 7 ‘[eyUepIoOe 2 -*q ur guejzodun 7 ‘poyveder 7 ‘Teyuaptooe 77 ‘ 'q | ut poyeoder g ‘yequeprooe ¢ :soryternoed Mon "MIs “Josey sy, "("") ‘ULB 1,°4889T (at ae Devore “(ry yo) “ULIB 7 fOvouy, Td nd ‘eee "ydog ‘wmreg -('9) "qdog ‘MIS g "UIBg “Ipesyd (70) yWog “weg ‘Cn? mP on 4 m4) nr 4O8OULL eagle oe a Mais : ; - see ni a oT mP mn? m4 mm 49 mt 7P Ud (P &- rn? at ea ipesyd ° ‘WULY, “Ipesyd oe 980] “ORT, "(2 mP *n'? £9) mr UOTL ‘Ipesyd “YaOS 7 Ue *(p £9) ‘S8ery ‘wae "MIB ‘“Ipesyg ‘S80T , WIeVg "(P 79 °q) ,yov0qT, “rposyg | r350T Od Ssh “Ue “JOBE, | ‘(9 p 0) ‘maisg “jovouy, 66h —‘Tl 8h 6LP - 686 - TS§ YEG - 913 -68I—T : (dd ¢z) ac¢¢h4—aZ Ip = *q : OTS 096 - 606—TII 888 - 188 SE $vé— II ‘687 ‘ 887 : L8P “LLY SLY - GOP 885 = C—O oe a rn ee ren er ee ee ae _ 02-0 PFI 7 | OT | @ "piq ‘pe ‘dd ¢g = AGh}—AS LF = “q ORIGIN AND GROWTH OF PLATO’S LOGIC 174 "(© P) “FTO 17}OBVONT, * LBS "UY, “Ipesyd : 9ZZ—I : (pig ‘pe ‘dd 444) atpe—orzp="9 || ge-0 | Pes || LZ | se | te “l ‘(P £9) ‘10d mIPeeTd * 68S—TI nr S8or] “Ue : CSF ‘WYOq “JOVOU,L, : HOF (0) youu], “Ipeeqd : G6g (0) Bder] “qdog : g2g nd ‘1d : $98 (0 £9) mrqdog "Weg : ZEg (0 *) , WIV “YouONT, :9BE 1'580T : 912 (0 0) py qdog ;,Ipeeyg : 983—I : (‘dd 6.¢1) S1L-—V6err7 ="0 ——e z9y}0 Jog “KX FOo_ = © ‘XIJ-TITA 800d = P ‘TLA-A ®400g = 0 ‘AI-II 8q00gq = q ‘J qo0g ‘dey = 8 ‘syoHOVIG UI pepp¥ £10940] Trews Aq poyreur 81 orfqudey oy yo yzed yowo ur Aquerfnoed Gove yO eoueLIMI00 oY} OSsTy ‘sImoer AjLrer~noed ours oY} Yor UI SONZOTSIp JXOU OMY OY} JO SOTABU PozBIAGIggs O43 vedds 1equInU sTy} puryeg = “serpirerfnoed onsAys perpuny oag JO 4st] eAoge ey} UI Jeqummu yuo syt fq poysusisop st Aqureroed qougq ouisy qsaef oy} 40f enboyrrg yove ur bursveddn saizrsvynoad fo uo1viownuy | ! { f | pipe ‘dd 9.47 = ELF9—OTL = "0 Dususiaibehdavusteumennen adiaisatedt ‘pig ‘pe ‘dd ¢.q, = 11-0 | Bt | S | of | IT | se) sTL¥—VErh="0 ‘queqzodun ||. nn. seUll} 6Z—9 Suto ns ba A “00 SO LIST MIE ‘TIA-A ‘Gf 98:0 | 292 || LZ | oF | 6% | 99 ouqndesy “g] "oTqe} OY} SurMOTIOF SuHONeTVTdxO S98 STIEIOP | gmary |l-Areuasn 4). Be "3 > ey wo || fo [OF ‘B30r] SVD : EGF "839'T “TIUd : §Gh nrI9d mGdog : eze 1 S80T : OBZ "ULL “IMOd :9Z "qmog ‘ydog :¢g —T "soTqIBTNOed MOT quejzodumt 7 ‘payeeder ¢ ‘[eyueplooe 77 suTeyuOD qt meq} ‘9 e10Joq WoPIIM SI SNyoJOVONT, FI yng ‘MOU SB qUNOO soryirerfnoed yeyueprooe F ATU soot ‘WIV : "Mod “qdog : “OTD, “MTOd : "Sd0rT : : | SOUIT} G[—f SULLIMNO -00 SoT}LIBINIeg "4084100 Ur SOUT} 9Z—G SutLINd -00 SOTPIIVINIOG le | Tr | 8¢ ‘piq ‘pe ‘dd %.18 ‘ue = SOPUIUIDT “OZ "piq ‘pe ‘dd 0.g9 ‘youoq,], = $n10,9004,L ‘BI ORIGIN AND GROWTH OF PLATO'S LOGIC 178 "BBer] BVIILID : SOT-OST | "aBarT “ULL, : GGT-OST ‘SBHUD = ShT-LPI “WILL, - PET-SSVT 880T : L-HG | THId w'tTd *§G 89:0 | SIP Td -8-OF S50. TMd + 8-OP | "OL Td : Th ‘Op ‘es-¥8 Td “Hd * 38 O&-66 —I | 99-0 | 89F | $60 | §89 "81q8} O49 ZULMOTIO; SuOTZBUBIdxo COs BTBIOp || . . . 10490 JOG = “X FOR = 9 ‘XT-TITA 840g = p ‘TTA-A 8H00g = 9 |: acts ae ‘AI-IT S008 = 4 ‘T Joog ‘doy =~ -syoqowsq ut poppe £10932] || Doing || ssgtun zo [reurs Aq pexrvur st oqndoy oy} Jo 428d yoxe ut Ayirerjmoed qove sour || zoqumu JO 90ueLINI00 C4} OsTyY sIMdeI AqLTerfnoed cures oy} OTM UI dnoa3 || 3u1oy SenZo[sip 4XCU OA} OY} JO SoUTeU PozELAIqqe ony de zequmu || qseqe{ || -[0; om} sry? purged ‘seniuvifnoed onsyAys pexpuny oAy JO 9sT] GA0ge || oy} 04 a qu9| ou} Uy Jequina yuexino sy Aq peywasrsop sy AqueyMoed qoug lt gy | -aambo "our ys ou7 sof eee , anboyouy yove ut bursveddy sarqesvenoed fo uorzosownusy | | | "gureqz10d ult soul} G-p FuliaM9 | -00 SOTITVIMIe a f 02 | a¢ | vé | oor > BAUTT OTF OE F Ul ander ae jo PId set dd,9-68 06 | 62 | 96 | 6&I s1Y OS’ “1S dNOUDH LSALVT “AT | | “434200 unt | sou} 6-11 Suuzmo | 44 TET ‘ejoy oo | ; | 6 | OF | O18) LOT | 40d ‘ensoqovoyT, Pg=| Fla] e a Ey e e. PA) PE | E |B | sou sopporattonenon BL P| EB | E |\-ae paw ‘sopso peoutoton -o1yo peurmseid meq} UI enSoperp yoso wr || 970018? 943 fo someon “tipepnced fo soquenar TPO], | punoy o[44s 104%] Jo 8927 *PONUYU0I—OLVTId JO SHODOTIVI(] OML-AINEM], FOUL ALINIGAY OILSTIALG dO MIAVy, 179 THE STYLE OF PLATO ‘gusjiodun Area 7 ‘Queqiodun gz ‘peeved -O1 9 ‘Teyueplooe oY : sorplIBIMoed Mou Jo [BIOT, ar Ud artTd - 416 —— Ad m Ud arthd ‘99 m 4d m 31d + G98 1999] 1 IM0d 3 O68 mld on }19d : 908 mE4d md * 006 mld md -0¢-9. —'III n290'] “WT, : SSP nl4d ‘319d -9g¢g md w31°d : $08 md nd *§08 m' S99] “40g : 188 wS50'T : EZ nlYd nid : 10S mn 930] “40d : GST =I n'20'T : SSF n WEY 09S "WHEL ‘Id : 602 rOLL mom Md : Sgt ‘MOd : 18I-LLT "IT, “MO : 9LI-PAT “BBIYIID “WIT, : TLT “BST “30d : OLI-89T wn 2 —— oe ST TR > = NN en a ee Se _———————oeeeaeasen®gns=s=$«=0909DD0u0:.0.0 0 eee eee 180 ORIGIN AND GROWTH OF PLATO'S LOGIC — a ee a ‘S3ery ‘Gd : 6c "LL n' Ud :* STS 1 330'T “UY, * FBT "SBWIO “UIT, : SLT 1'239T “IG : 29g | ‘SUID “ULE, : SLT | ‘SUID “OL, : LOT | BUILD “OIL, : SOT | ‘BBor] “WL, S9T-6ST SBD : BPT queyzodunt | Td : 89-65 | WL :ObT-G8T seul [Z-F SuLLIN ‘SBD “Md a OFF | 6T | cg | 68 | 281 OnE, 4d | ey “OIL “Td: 68 : ay moor vorea yo | “PIC Ped 8.65 "‘B30r] “IM d : eS ; : "a110q = ‘mn Wd 118 —T sep | 6t | 99 | Sf sor MUN “BS , Sarao]jo; suoysuydxe oes streyop < Hh) # a Set ETE Gehl tet t Set | Re | le o| bE Ee] | | ‘AI-II_ s400q = 4 ‘doy = : szeun yo|| && | 3 Ss | & “O}qQ’9 SIYy WI fk ced qi : LSgeotied A ea hain bat usd Geeno aoa a Paap P 5 3 5 atk 10} pa rigtiir om SonSOTCIP FXOU OM4 OY} JO SOUTTU PoqsIAerqq’ eq} rvedds requnu srg} purged ‘serytrerqmoed onsi{Aqs any Cay JO $s] eA0gs Bay peumnserd z16q3 ut orpep 243 fo sowmmyy ey} Ur Joqumu quote sy Aq pozeuswop 81 Aqrrermnoed qougq rego nbs enZOTVIp YOCS UI "oury yeu 942 40f gasynjogy|. pejoy, || Punos ef4gs 104%] JO #027 onborwig youve us Bussveddy sowrspynoed fo worynioumnusy -unynoed fo 4aQuUunpT “penusjzuoo—OLVIg 40 SHADOTVIG OML-ALNAMY], BOX ALINIAZY OLLSTIALG JO FIavVy, 181 THE STYLE OF PLATO te a egy rc PA SAG TCP OREN S re | ‘guj 10durt seUIT} 9Z-9 SuLLINO -00 seen beg : 8 | 6-0 | $96 |! ST | Ih | 6F | Ll 7 599T -3IS— II | | : SMB] OT} "B30TT : SLZ | ‘B80rq 89 | UI INd81 FOIA JO ‘BSaTT SBD :OS3—I |: 09-0 || Lem | HT | 77 | 8S | SUT *po}e1eumnus | jou ezom Aoyy osneoog 4st] INO UI pepnyour you ‘Jeyey ZULLINIGI PUB SNQeTIYg oY} Ul sig SuLivedde SPpIOM CIB GE PUNOJ svy [[eqduIED esoq} Seprsog ‘qure}.10d wat ‘quvjzodut 7 ‘[eyuepiooe ¢ :serytrernoed MeN B30] UL, ng ery saery : - 1 380r] "BOOT UL . "9907 “poyveder F ‘feyueptooe ZE : sorprTerfnoed MoNy n'S80T WHEY, : “B3ery “WILY, : 1 890r] UY, : “wn “Td: 1oaeT : wsae7T] : | :988—I || 96-0 | gop || 9F | 29 | se | Oot | BiG ad | seul} 1Z-F SurtmM9 -00 SOT}IIVITNIGT - LE@— TI 09 200 | sxe || ot | 90 | 98 | 62 ‘ATF > BAB] OT} ae ! UI IMVG1 YOIYA Jo —_——————. fo >) re o | 4 ‘piq ‘pe ‘dd ¢.g¢ ‘OE = SNIDUWAT, “VG ‘pid ‘po ‘dd ZF Td = S2Q7U%d °SE Se EEanEIEEREEEEEEneeierememeneett 182 ORIGIN AND GROWTH OF PLATO’S LOGIC "BG-0 SBIFLID 10} “G9.0 SnOBmaLY, 10¥ “$9.0 SNqeTTYT Io oq [TA Ayruyye saryeper oy} |; ‘A[SNOOWS][NUIIS OPVUL EY SUOTIPPS O0IY} OY} [Te JT |! Wg “GS-0 OF SVIFLID OY} 1OF “G9.0 0} SROVUIT], OF} 10} ‘9¢.9 0} SnqeTTyg oy} roy ‘AjTUN SB SMB] OT} uo pommsvem ‘dnois ysoze] of} 09 AyIUyTe OATYTOI OY} OSV SUOT{IPpe esoyy, ‘Tjoqdueyp Aq poysr0Ur -nue you e10M Aoq} osnvo0qg 4ST] INO Ut pepnyour woeoq +4OU CABY YoIyM ‘ATUO SMB pus SBD | 0} WOMUIOD spIoM g7 pus ‘(gg ‘d ‘m ‘Joa *da7) Teqdureg 07 Surpioooe ‘Ayuo sMery pues snevUTy, 0} UWOUIUIOD SPI0OM 99 poppe eq FYUSIUL S10, "B30°T : 9g3—'I "8[qQ8} OY} BULMOTIO] SUOCIyVUB[dxe O88 STIBZap H ; royjoIOT "KX YOO = © ‘XT-TITA SHO = P ‘TIA-A 8HOOG = 0 | ‘AI-IT 8y00g = q ‘J yoog ‘dey = 8 ‘s7eX0vIq Ul poppe 810999] | Trews Aq poxreur st o1qnday oy} jo red yove ul Aq1e1~noed yoo ies Jo eoweLM900 ey} OsTY “smoer AqirerfMoed ours OY OIA UT , dnoza senZO[VIp }xXoU OJ OY} JO SoUIvU po}eIAeIqqs ou} Ivedds Jequmu | sIq} pulyog ‘seryirwrmoed onsryAqs perpuny eay Jo 4xSIT eAogs pa ey} Ul JequmMuU yuemMo 841 Aq pozeusisep st Aqurermoed Wow hgnungfo aung pe4f 943 tof anboyma yova ur Gursvaddn sarzruvajnoad fo worpsaUnug lees, ae tec so me $92UN JO zequmu SULAO| “TOF OU 07 7a9| -sambo T390.L — | 06 "4UBIOC UIT SOUT LIT-0% Sutzmo “00 SOMMBITNO I &§ | 9LT | SLT | seul G—g SuiLINd -00 SOMLIVINOAT or TY ats : BAMBI] OT} UI NdeI YOIUA JO at | 8 em eS ce | Cee ‘TeIUOPIOOY ‘T ‘ondorerp youo ur puno of 498 1038] JO 8997 ~sryndad [0 LOQuUin|y "pig “pe ‘dd 7.98% “BZer] = smvryT °9% ‘pig “pe ‘dd ZIT "SDUUD “SS “OT Ge4 Stq} Ur WSY3 10] pesn 8U014STAGAq -q8 pus ‘1epx0 [eorZoyou peunseid req ut sonboporp oy7 fo sun . UWIU0O—OLVIGT dO SHADOTVIGG OMI- a0d ALINIGIY OLLSIIALS fO FIV é d 183 THE STYLE OF PLATO Oe rn a er Rr eh eR YG ‘sorjirermnoed yureqzodurt AT0A = AT ‘gueqzodunt = JT] ‘peyeedor = J] ‘Teyueprooe = J "snyojovoyy, = ‘jovoyy, ‘Snevuiy, = ‘mI, ‘unisodusg = ‘duikg ‘“stqdog = ‘ydog ‘onqndey = ‘dey ‘snapesyg = ‘zpeeyg ‘snomyog = “4yy0g ‘snqeTtag = ‘Tyg ‘sBi0sej01g = 01g ‘sopluemivg = ‘wIeg ‘sMery = ‘38ar] ‘soqoery = ‘yoery ‘s¥1di0H = ‘3104 ‘snmephyyng = ‘pAqyng ‘orqdAqyny = ‘qdéyyng ‘seprmieyqy = ‘weg ‘sn[fyeig = ye1Q ‘ASojody = ‘jody—sNOILVIAgaaay "Sq UI [eJUeproos “q Ul peyveder suvew F:Fq : UGATZ O18 SOXOPUI ow} ‘orfqndey oy} jo 41ed eUO JO SUOISIAIpqns O44 Ul yUOAeyTp si AoUoNboy yo eosZep 04} JT = *o Ul JueAOdUII ‘p ur poyeeder “o yaed o1jqndey ul yuejiodu ‘snyojeveyy, Ul [eUepioos sv ‘snipeeyg oy} ul yueyz0duUlt 8B smMode1 (sn[44BI ey} UI AT[e}Uepioow’ ZuLtmo900 ysI1y) gpz AgIerNoed yey} suveut (,,0 ,p ,f9) “yoReuy, mr Ipeeyd ‘EPs ‘oouvjsur JOq ‘n000 Aoy, TOI ut oiqndey oy} jo yred Yoo Jo SMOTyeIAGIQge OY} TILA poyreM 018 oTqndoey oy} Ul FULIIN00 seryIeTTNIeg = ‘sMBTT ONY UI ATpeyeedor pus snqe[Iqg ey} Ul eoulO BMOeI (p yaed ‘orjqndey on} ut Aypeyuepiooe Surzin900 ys1y) 16g Ayiaeroed 4eq4 suBeUT 350] "TY : [Eg ‘eoueysur JOq ‘[eyueprooe st Aqrvtjnoed oY} OAL SI XOpUI OU oIOyAA ‘“pouTeU enZorerp oy} ur Aqtrernoed sITqy yo Aouanbery jo oo1sop oy} suvom popuedde xeput uy ‘eouvreedde 481g 891 Joye CUNT} Yxou oT} UI permMbow Ayirernoed Mou ey} eoueyIoduaT jo esldep ey} MOUS 0} Jepso Ul ‘JequINU YORE pUTyeEq UOATs ore sIndeI AQLTeT] -nood youo YOIYA UI sendoyVIp OM} 4xeU EN Y, “JepIo [eoGojouor1yo peumsead s10q} UL MOTIOJ SeIyLTeTTNOed Ppearesqo OOF ey} ‘endo[eIp yove Ul out} 4sIy 04} I0y Sutreadds somirerMoed jo UOYeIOUNUS Of} UT “ONSO[SIp YOR OJ OPVUA ST UONB[NOTo ous OUT, ‘ZO-0 = SIL/YT st Asofody OY} JO ApUye OArefor oT} FBY} SMOTIOF 41 ‘STL, Sa1eq SMury oY} Jo yuaTeAINbe oy], ‘s#TUN OT 04 yUeTeAINbe ore [44s 1048] JO sorqlavipnoad yueyiodut { ‘peysedea g ‘teyueprooe g AZorody eq} Ul ‘oouRIsUI 10,7 “poATEsqo sorjlierjnoed Jo Jaquinu eues ey} JO} SMBTT OY} JO yUoTeAMMbe oy} 0} ongorerp B JO yUOTeAMMbe oy} JO NOH -rodoad ot} st Ajrugge oatgeper oy, ‘AUgye Jo syIUN INO; JO; 4UBIIOduAT AI9A YOO ‘ooIq} Joy yUBJIOdUIT YOR ‘omg 103 Aqirernoed poyeedoez yove sutyun0s Aq poyslnoreo st Ay1uyye Jo yueTeAINbe oy, “lequIMuU sTq} Surpecoxe eousrine00 Aue si yunjiodwm fizaq ‘onSopetp Powe 1Oy poyeorpul SI pus ‘onZopsip ey} JO 9ZIs ey} 0} SUIpI0008 SOLIBA 7UDILOdwW? potAIE} SedUEIINN00 Jo IOqUINU ey, ‘yUL{IAOdUAT UIE} [TBO 0} ‘ENSOTVIp ey JO OZIS OT} OF Surpioooe ‘Aresseov 41 SOHVU UBT} SSO] PUB GOTO UB} e1OUL SULLINII0 80} ‘pazvede4 : OnZOTSIP 8 UI GOTO SULIINIDO SULOIPI IO SpIOA O1V SOTyLIBITMOed yoPWepLIOOp—'ALINIZAY JO AIAV], HHL JO NOMYNVIdXY ee The length has & great influence on the equivalent of affinity, but the number of peculiari- ties found in each sample of text is not pro- portional to the size. Single peculiari- ties are insignifi- cant, and the order of small dialogues remains uncertain. The increase of the equivalent of affinity is not pro- portional to the size of the sample of text inves- tigated. Only equal amounts 184 ORIGIN AND GROWTH OF PLATO’S LOGIC Among the inferences which can be drawn from the above table, nothing is of greater importance than the great influence of the size of a dialogue on the number of stylistic peculiarities found in it. We see that the Critias on its eleven pages contains less than half the number of peculiarities found in the Timaeus, which, being nearly five times larger, was written immediately before the Critias. Hence it results that eleven pages, being more than the size of the Crito and some other small dialogues, are insufficient for a stylistic determination, so long as we deal only with a few hundred stylistic tests. The difficulty might be removed by extending stylistic obser- vation over a far greater number of particulars, a task which requires only additional research. But we under- stand at once that our equivalents of affinity for such small dialogues as the Euthyphro or Crito are very far from the truth, and that for instance no valid inference can be drawn from the apparently greater affinity of the Crito with the later style. This shows also the insignifi- cance of a single test applied to such a complicated problem. One rf pv, or one xaOarep occurring any- where proves nothing, if even seventeen peculiarities of later style found in the Laches and missing in Charmides are according to our rules no sufficient evidence for the priority of Charmides. We are warned also against the error of supposing the opportunities for the occurrence of a greater number of peculiarities to be proportional to volume. In this respect the subdivision of each part of the Republic into several samples of text is very instructive. Even those who believe the Republic to have been written during many years cannot deny that B ITII-IV are the immediate continuation of B. I, and with it form one whole. The style of equal samples of text in these books is also very uniform. But the influence of the size becomes evident if we compare a small sample with a larger one. Part b, (857 a—367 B) of 74 pp. (ed. Didot) contains only an STYLOMETRIC INFERENCES 185 equivalent of 21 units of affinity, while the following 294 pages, being four times larger, have seven times more peculiarities. In another case two succeeding samples of text differ much less, namely, c, (471 c—541 B), being nearly thrice as long as c, (449 a—471 8), has less than twice as many peculiarities of later style. The whole of the Republic, being ten times larger than the tenth book, contains only a little more than thrice as many pecu- liarities of later style. From these examples, which might be indefinitely multiplied, it becomes evident that only equal amounts of text should be compared. Future Inquirers should base their calculations on an amount of text equal for each dialogue, or divide each dialogue into such equal samples of text, for instance, of ten thousand words each. Another lesson of the highest importance is taught by the stylistic comparison of the first book of the Republic with the following books. Nobody doubts that the Republic in its present shape is one whole, and that the first book, even if mainly composed much earlier, has been revised and worked into unity with the following text. Now it has a surprisingly early style, having less than half as many peculiarities of later style as the first sixteen pages of the fifth book, even fewer than the Laches, which is inferior in size. This shows on one side the early date of the first book, and on the other side it shows that no revision can substantially alter those peculiarities of style which are the subject of our investi- gation. Therefore all explaining away of the late style of the Phaedrus and Theaetetus by the supposition that we possess these dialogues in a late and revised edition is of no value whatever for chronological purposes. If later revision could alter stylistic affinities, then the first book of the Republic, which must have been revised, emended, and corrected in order to be absorbed into the larger work, could not have remained as remote from the later style as the Laches, while already the of text are compar- able so long as we deal with a few hundred peculiari- ties observed. The author's revision does not alter the essential stylistic character of the text, and stylistic compari- son shows the rela- tive date of a work even if applied te a later edition corrected by the author. This is very important for a knowledge of the date of the Phaedrus and The- aetetus which were re- vised by Plato later. The first book of the Re- public un- doubtedly revised and cor- rected has & surpris- ingly early style. Relation between Philebus, Timaeus and dia- lectical dialogues remains also un- certain because s0 many special peculiari- ties of the Sophist and Poli- 186 ORIGIN AND GROWTH OF PLATO'S LOGIC second book shows a style later than the Phaedo and Symposium. This conclusion is quite independent of any speculation on the exact date of the Republic, or on the date of the Laches. If anybody supposes that the first book of the Republic could have been written as early as the Laches (as Siebeck does), then he is bound to account for the difference of style between the Laches and the second book of the Republic. At all events, we have here a work which has been left by Plato as one whole, and which nevertheless betrays by stylistic tests the differ- ence of the times in which it was begun and continued. According to our rules the number of peculiarities of later style found in the first book of the Republic is insufficient for an exact determination of its place among the early dialogues, and it may be even later than the Gorgzas. To settle this question it would be necessary to collect a much greater number of observations, and to compare with the first book of the Republic a part of the Gorgias exactly equal in size. This we are unable to do, as a great number of authors from whom we have taken the number of occurrences of each peculiarity did not enumerate all the passages. The relation between the Philebus and Timaeus on one side and Sophist and Poltticus on the other side cannot be decided according to our table, because we have included in our list more than one hundred words observed by Campbell in the Sophist and Poltticus, while no such special study has been made of the Philebus, Tumaeus, and Critias. These words were included in the list because the late date of the Sophist and Politicus is less generally recognised than the late date of the Timaeus, Critias, and Philebus ; it therefore appeared necessary to bring out with the greatest clearness this late character of the two dialectical dialogues, even at the risk of making them appear later than the Philebus, Timaeus, and Critias. As soon as these later dialogues shall have been investi- gated with as much care as Campbell spent on the two STYLOMETRIC INFERENCES 187 continuations of the Theaetetus, the true chronological order will not be obscured as it is now in the later part of our table. Even now it is easy to eliminate a part of the error by excluding a number of peculiarities which have been first observed in Sophist and Politicus. If we omit peculiarities 12, 13, 54-181 of our list, reducing thus the total number to 370 peculiarities under investigation, then the Philebus will not be affected by this change, while the Laws lose 102 units of affinity, the Timaeus 53, the Politicus 86, and the Sophist 69. The relative affinity calculated on these reduced numbers will be 0°65 for the Sophist, 0°66 for the Pol:- ticus, 0°66 for the Philebus, and 0°61 for the Tumaeus. This calculation shows that the most important figures of our table are those of the relative affinity, which are very constant, and change little if they are calculated on a very much reduced or very much increased number of observations, changing less with the increasing number of observations. We see that the relative affinity of the Sophist, which was found to be 0°65 for 500 peculiarities, is just the same for 370 peculiarities. It 1s probable that increasing our list to 5,000 peculiarities, this constant relation would not be altered by more than a small percentage. We have therefore in the relative affinity a powerful instrument for chronological purposes, of the same constant character as the physical constants measured in natural science. If the density of pure iron has been found by a series of experiments to be 7°8, everybody understands that further experiments of a greater exactness can only alter this constant relation very slightly, adding new decimals and showing it to be more exactly 7°84, but never 7°5 or 8:0. We claim the same permanent character for the relative affinity, calcu- lated on a sufficient number of observations. Comparing these numbers, calculated on a smaller or greater part of our materials, we have found that relative affinities under 0-1 have no value whatever, and can be changed to the ticus have been in- cluded in our list. Reducing our list by 130 pecu- liarities, the relative affinity of the So- phist will not be affected, and that of the Philebus rises above the So- phist and even the Timaeus. The relative affinity has the character ofa natural constant like the constants in physical gcience. This gives an unpre- cedented strength to our con- clusions. The latest group of Plato’s works con- sists of the Sophist, Politicus, Philebus, Timaeus, Critias, and Laws, with a relative affinity of over 0°5 in samples of text exceeding forty pages. 188 ORIGIN AND GROWTH OF PLATO'S LOGIC extent of at least half their value by calculations based on a greater number of observations. But the remaining relative affinities in our list are exact in their first decimal, and any number of observations added can increase them only in the second decimal, except in the Philebus for the reasons explained above. But even the Philebus, if we measure its relative affinity by one decimal, will maintain it, whatever number of new observations may be added. Thus we claim to have proved the following general conclusions about the order of the works of Plato: 1. The latest works of Plato are: the Sophist, Politicus, Philebus, Timaeus, Critias, Laws. This group is distinguished from all other works by a relative affinity of over 0°5 in samples of text exceeding 40 pages (ed. Didot). This means that out of any number of stylistic peculiarities investigated (provided those peculiarities are selected which are not limited to one dialogue, and provided the number of peculiarities so investigated exceeds 300) more than half the number found in the Laws will be found in any sample of text of 40 pages of these dialogues. As the Critias has only 11 pages, for the investigation of the Critvas the preceding 29 pages of the Timaeus must be added. Or, if we calculate the relative affinity of the Critias apart, it must be com- pared with an exactly equal amount of text of the Timaeus ; then it cannot be expected that the relative affinity of such a small portion of text should exceed half the value of the affinity of larger units, as the relative affinity is in close relation to the amount of text to which it is applied. ‘With an increasing number of peculiarities observed, the influence of the size of a sample of text would be less important, and the size of the Critias is insufficient to define its stylistic affinity only so long as we deal with a reduced number of observations. The number of possible peculiarities of style is practically infinite, and may easily exceed the number of words contained in a sample of text. STYLOMETRIC INFERENCES 189 2. The latest group is preceded by a middle group, consisting of Republic B. II-X, Phaedrus, Theaetetus and Parmenides. In these the relative affinity is under 0°5, and even under 0°4 for samples of text of 30-60 pages. The mean affinity of dialogues belonging to this group is only 0°3, or only half as much as the affinity of equal dialogues of the latest group. The middle group is distinguished from all earlier dialogues by a great number of important and very important peculiarities appearing here for the first time, as may be seen from the table. 3. The middle group is preceded by a first Platonic group, consisting of three dialogues, Cratylus, Symposium, and Phaedo, which are characterised by a relative affinity inferior to that of equal samples of text of the middle group, being about 0:2, and not exceeding 0°21 for samples of text of 40-50 pages. The first Platonic group is distinguished from all Socratic dialogues by many special peculiarities appearing here for the first time, and indi- cated in our table. 4, Among the Socratic dialogues, which show an apparent relative affinity of 0:1, or even less, the Gorgias appears with probability to be the latest, having 18 peculiarities in common with the first Platonic and later groups, which are missed in other Socratic dialogues. But this number, which was held to be sufficient by C. Ritter to define the middle group, is according to our improved method insufficient, and affords only a certain probability, increased by internal evidence resulting from the comparison of contents, but requiring further support by a much greater number of observations. 5. Last, not least, we repeat the important conclusion, which is perhaps the greatest gain of our investigations, viz. that stylistic tests if properly directed afford cer- tainty as to the chronological order of Plato’s dialogues ; and conclusions from stylistic comparisons cannot be in- validated by assuming fictitious later editions, corrections The middle group shows & relative affinity of about 0:3. The Cratylus, Sym- posium, Phaedo are earlier, having a relative affinity of only 0°2. That the Gorgias is the latest of all Socratic dialogues is pro- bable. Certainty of stylis- tic con- clusions indepen- dent of supposed revisions. Phaedo later than Sympo- sium and Cratylus ; Par- menides later than Theaete- tus and Phae- drus ; Philebus later than Sophist. Other minor questions will easily be decided by the same method if applied on a larger scale. Our con- clusions confirm earlier conclu- sions arrived at by stylis- tic study and com- pletely change 190 ORIGIN AND GROWTH OF PLATO'S LOGIC and revisions, a8 it has been seen on the first book of the Republic that such later changes cannot affect the essential characteristics of style as these are now known. The above five conclusions are worth the labour spent on our study of Plato’s style. We do not pretend to give for certain anything more about the order of dialogues within each group, except that the Phaedo is later than Symposium and Cratylus, the Parmenides later than Theaetetus and Phaedrus, the Philebus later than the Sophist. The relative position of Republic, Phaedrus, and Theaetetus, of Politicus, Philebus, and Tiumaeus, can- not be decided on the above observations alone. These problems are of less importance than the distinction of groups, and now that the method of stylistic calculation has been shown on a small example of five hundred pecu- liarities, it will be very easy to apply it on a much larger scale, and to settle all the minor difficulties left to future inquirers. It is to be hoped that nobody hereafter will attempt to judge about Plato’s style from small numbers of observations. Any new observations ought to be added to those existing, in order to achieve a progress of knowledge in these matters. The group of the latest six dialogues, recognised independently by Campbell, Dittenberger, C. Ritter, and von Arnim, is now still better defined and is established beyond all reasonable doubt. The anomaly observed by Campbell as to the Philebus, Parmenides, and Theaetetus, is removed, and the true place of these three dialogues found in accordance with their style. This entirely changes the current traditional conception of Platonism, as taught by Schleiermacher and Hermann, and still in our own day represented by the great name of Zeller. The differences between these authors become insignificant in view of their grave and common error in placing the dialectical dialogues before thé Symposium and Republic. This error produces a complete distortion of the true view of Plato’s philo- STYLOMETRIC INFERENCES 191 sophical career. It is as if some eminent critics pro- posed to look upon Kant’s Kritik der reinen Vernunft as a juvenile eccentricity, and to seek the chief contents of Kant’s philosophy in his Principiorum primorum cognitionis metaphysicae nova dilucidatio, published in 1755, and written under the influence of the then pre- vailing philosophy of Leibniz and Wolff. We should fall into the error of premature generalisa- tion if we pretended to go further in our conclusions and to decide anything about the order of Socratic dialogues in which the relative affinity sinks below 0°1. Our in- strument is not fine enough for these small differences between dialogues probably removed forty years from the Crittas and from the latest books of the Laws. To determine their order, another standard is required than the Laws, with which they have too little in common. The Gorgias being the latest and also the longest of the group of Socratic dialogues, the best plan would be to collect and classify peculiarities common to each of them with the Gorgias. But if five hundred peculiarities were needed to fix the order of dialogues later than the Gorgias, for those earlier a much greater number of observations 1s required, and can be reached only through well-organised labour of many scholars. A distinction of only four degrees of importance of stylistic marks might ultimately prove insufficient, but even if we classify the peculiarities observed otherwise, it will always be indispensable to make due allowance for the different importance of accidental, repeated, frequent, and very frequent peculiarities, as well as for the more or less essential character of certain observations. One of the most immediate aims for further inquiry is to investigate peculiarities in the order of words and in the construction of phrases. By means of a great number of such peculiarities it will be possible to determine the relative affinity of all dialogues among each other, and this alone will probably lead to the the current concep- tion of Platon- ism. Proposed subjects for future inquirers, They must compare each dia- logue with all others, and spe- cially all Socratic dialogues with the Gorgias. The classi- fication of peculiari- ties can be improved. The order of words and con- struction of phrases should be investi- gated. A revision of work already done is also necessary. Number of words in each work should be counted. Our method can easily be applied to other problems than Platonic chrono- logy, and leads to & new science of stylometry subsidiary to historical research, like palaeo- graphy. Greater certainty of results obtained by the investiga- tion of style than 192 ORIGIN AND GROWTH OF PLATO’S LOGIC definitive solution of all difficulties of the Platonic chronology. There is no reason to fear that the amount of time spent on such inquiries will be lost. In every science there arises at certain points a necessity for much detailed research leading to no new conclusions, and only confirming previous generalisations. The familiar example of modern organic chemistry shows that valuable investigations were made by beginners, following a method already fixed, with results foreseen by general theory. Such investigations, though they teach us few new truths, increase the certi- tude of the general theory which they illustrate. Further study of Plato’s style will probably not change our know- ledge as to the order of the three groups which are now found, but it may modify our views concerning the order of dialogues within each group, and may help to fix the order of earlier dialogues, which is at present uncertain. Besides further research on the lines here indicated a systematic co-ordination of the results already ob- tained is also necessary. There are discrepancies between the numbers given by various authors for the occurrence of the same peculiarity, and the calculation of proportions between different uses might be very much improved. The number of words contained in each dialogue should be taken as the true measure of text and of the opportunity for the occurrence of expressions for which no better calculation of opportunities can be found. When once the importance of this field of research is generally recognised, it will very soon appear that the exact determination of style 1s the safest way of settling the difficulties, not only of Platonic chronology, but also of the chronology of other authors, the date of whose writings is unknown. There will be scarcely another case in which the mere question of the chronology of some writings would be .of such unparalleled importance for the history of human thought as in the question of THE FUTURE OF STYLOMETRY 193 Platonic chronology. This exceptional importance of one particular case will have produced a new science of style, which will enable us to decide questions of authen- ticity and chronology of literary works with the same certainty as palaeographers now know the age and authenticity of manuscripts. This future science of stylometry may improve our methods beyond the limits of imagination, but our chief conclusions can only be confirmed, never contradicted by further research. That the dialectical dialogues are later than the Republic is now as clearly demonstrated as any other fact in his- tory can be. Equally certain is the conclusion that the Republic, Phaedrus, Theaetetus, and Parmenides are later than the Phaedo and Symposiwm. These facts must be accepted now as if they were supported by the clearest testi- mony. The certitude attainable by a consistent theory is even much greater than the certitude of the best evidence; every astronomer believes himself to know more of the present and past movements of the moon than an historian can know of the movements of Caesar's army. Historical testimonies have always but the value of the sensible evidence on which they are based, while our results as to the order of Plato’s works rely on the higher authority of reason, producing, according to Plato, infallible knowledge whenever a good method is followed. by any in- formation based on mere testi- monies. 194 ORIGIN AND GROWTH OF PLATO'S LOGIC CHAPTER IV SOCRATIC STAGE OF PLATO’S LOGIC Small dia. WHEN the Platonic works are compared with regard to loguesdis- their volume, we find a numerous class of dialogues tinguished which do not attain to half the size of the Protagoras, aa ae and which can be distinguished from the rest as small ie 7 dialogues. No fewer than eight among them, the Clitopho, Minos, Hipparchus, Epinomis, Theages, Ama- tores, Alcibiades II. and the Greater Hvuppias, have since Schleiermacher been generally regarded as spurious. They represent seventy-two pages of text (ed. Didot), less than one-third of the Laws, and contain nothing that could be included in Plato’s logic. Their The Io, Hippias Minor, Lysis and Menexenus, though chrono- successfully defended against doubts as to their authen- logical ~~ ticity, remain outside the pale of our inquiry. All these pian! small dialogues offer greater difficulties than larger works, determine Pecause their limited volume makes a complete appre- onaccount Clation of their style and doctrine less easy. They oftheir require a special study through which their mutual size. relations might be determined and a certain place assigned to each of them. Such an inquiry would alone fill a volume, if it were intended to lead to definitive con- clusions, based on a careful weighing of many details. So long as their chronological order has not been determined by patient and impartial stylometric inquiry, we must for our part abstain from all attempt to fix this order from the few logical hints which they contain. A Socratic The existence of a Socratic stage in Plato’s logic is stagein far more probable than the myth of a Megaric period. SOCRATIC STAGE OF PLATO’S LOGIC 195 We have the clear testimony of Aristotle (Metaph. 987 b 1) that Plato owed to Socrates the tendency to form exact definitions of ethical notions. It is precisely in the small dialogues that we see the illustration of this tendency. In another passage Aristotle teaches us that the direct philosophical merits of Socrates were inductive reasoning and definition by means of general notions (Metaph. 1078 b 27). In the small dialogues we find accordingly the constant employment of inductive reason- ing and repeated attempts to define: by means of the nearest general notion, in application chiefly to ethical purposes. Though faithful even in his later period to induction as a method of investigation, Plato gave in his dialectical works a far greater importance to deductive classification. The thoroughly inductive character of the small dialogues is more Socratic than Platonic. The influence of Socrates on Plato is not, like the alleged Megarian influence, attested only by a late and un- trustworthy witness: it 1s known from numerous passages in the writings of Aristotle, and results also from the manner in which Socrates is again and again represented by Plato as the teacher of true wisdom. Were it not for Plato’s strange desire to represent, in more than twenty literary masterpieces, his own thoughts as enunciated by Socrates, we might have given to the latter no more credit than to Anaxagoras, Heraclitus, or Parmenides, nor would his name even to-day be synonymous with Sage. Hence it is natural to suppose a Socratic stage in the development of Plato’s philosophy, and to seek for the vestiges of this period in his works. These vestiges are precisely found in the small dialogues, and in the four works in which Socrates is represented as triumphant over the sophists. These are the traditional sixth tetralogy, consisting of Prota- goras, Meno, Euthydemus and Gorgias, which form a natural group, though they have not been connected by 0 2 Plato’s develop- ment very probable, because we know from Aristotle that Plato owes much to Socrates, and it results also from the pic- ture of Plato’s teacher in his dialogues Socratic influence specially visible in the smal! dialogues and in the sixth tetralogy. All these works have chiefly moral aims. Also Socrates’ philo- sophy had & pre- dominant ethical character. Socratic dialogues are the earliest. Of all small dialogues only Euthy- phro, Apology, Crito, 196 ORIGIN AND GROWTH OF PLATO'S LOGIC Plato himself into one series. They have in common with the small dialogues the predominating ethical aim, and they deal with the definition of virtue and various parts of virtue, as well as with the question whether virtue can be taught. Such ethical questions are abandoned in later works: even in the Philebus, where the avowed aim is the solution of an ethical problem, the whole argumentation takes a metaphysical and logical turn, which is wholly absent from the small dialogues and from the four others above named. The character of Socrates’ philoscphy was also mainly ethical, and this authorises us to see the pre- dominance of Socratic influence in those dialogues which are limited to ethical inquiry. Plato’s own philosophy had another character: he was rather a politician, a metaphysician, and a logician, than a simple moralist. He set perfection above mere virtue, and even despised the traditional virtue of the common citizen, which was the starting point of Socratic ethics. We shall not be far from the truth, if we admit that the small dialogues are earlier than the logical investi- gations which commence with the Cratylus, and are continued in the Phaedo and Republic. For an exact determination of their order the data are not yet col- lected, because their style 1s very much less characteristic than the style of the latest group. We can only observe, that of all peculiarities of later style only very few and unimportant examples are 1o be found in the small dialogues. For the investigation of the development of Plato’s logic only five among them are of any importance: the trilogy about the death of Socrates, consisting of the Euthyphro, Apology, Crito, and the two companion dialogues of the Protagoras, namely the Laches and Charmides. We omit the first Alcibiades, though its authenticity has been sustained by Socher, Stallbaum, 197 Hermann, Steinhart, Andreatta,“® and Kophiniotes,!” against Schleiermacher, Ast, Susemihl,'® R. Hirzel }4° and many others. Strong suspicion is roused by the noticeable contradiction between style and contents in this dialogue. According to its style the Alcibiades would be later than the Symposium, while the general contents place it among the small dialogues, as has been recognised by all defenders of its authenticity. Quite recently Ivo Bruns,' by com- paring the characterisation of persons in Plato’s dialogues, came also to the conclusion that the first Alcibiades could not have been written by Plato. As to logical contents, the Alcibiades presents, besides some theories sufficiently known from other works of Plato, a singular identification of the soul with man (180 c: pndsv adAXO Tov avOpwrov RéEirrerar cvpBaivev 7 yuy7nv), which recalls a passage from a _ notoriously spurious dialogue (Aziochus 365 E : jets éopev uyn). In the Gorgias (464 a) every man is supposed to consist of soul and body, and at all times Plato defined man as an animal (Crat. 399 c, Polit. 2718, Legg. 7658, &c.), with a soul (uy) avOpwrov Prot. 312 B, Symp. 192 D, Phaedr, 249 £, Rep. 590 a, &c.) ; the identification of man and soul seems to belong to some later Academicians. This contradiction between the first Alcibiades and the current Platonic teaching on an essential point is not of the same kind as many quite superficial contradictions quoted by those who oppose the authenticity of some of Plato’s other works. Man as consisting of body and soul is a familiar notion to Plato’s readers, and if the author of SOCRATIC STAGE OF PLATO'S LOGIC 146 Andreatta, Sull’ autenticita dell’ Alcibiade primo, Roveredo 1876. 47 J, K. Kophiniotes in vol. iv. pp. 289-296, 310-315 of the Hphemercs, Athens 1881. 48 Platons Alkilnades I. und II. tibersetzt von F. Susemihl, Stuttgart 1864. 49 R.. Hirzel, ‘ Aristoxenos und Platons erster Alkibiades,’ in Rhein. Museum, vol. 45, pp. 419-435, Frankfurt a. M. 1890. 1% I, Bruns, Das literarische Portrdt der Griechen Berlin 1896, p. 339. Char- mides, Laches, contain logical hints. Alcibiades is pro- bably spurious. Identity of man and soul as pre- sented in the Alcibiades is unpla- tonic, and contra- dicts the per- manent teaching of Plato. Authen- ticity of Euthy- phro suc- cessfully defended against doubts. Logical contents of Huthy- phro. 198 ORIGIN AND GROWTH OF PLATO’S LOGIC the Alcibiades takes the trouble to give a demonstration of the identity between man and soul, he must have felt that this was an innovation against the general opinion. If Plato had given this demonstration himself, he could scarcely have disregarded it throughout his other works, from the Protagoras to the Laws. Therefore we are justified in excluding the first Alcibiades, as well as the second, from the list of Plato’s works. The doubts raised ugainst the authenticity of the Euthyphro, chiefly by Ast, Ueberweg, Schaarschmidt, and J. Wagner,'"! have been sufficiently refuted by Stallbaum, Hermann, Yxem,'? Wells,* Adam,'* and Jezierski,'” so that there is no need to return to this question. All arguments against the authenticity of this and many other works can be reduced to two principal heads: 1. Plato would have written otherwise ; 2. Analogies with other dialogues show an imitator’s hand. Such argu- ments are necessarily subjective, and we can only affirm with certainty that Plato would have written otherwise, if we notice, as in the Alcibiades, some essen- tial contradiction to well-known and constantly expressed Platonic teaching. Nothing of that kind can be said of the Euthyphro. The logical contents of this little dialogue'* corre- spond to what might be expected of a work written while the influence of Socrates on Plato still remained un- altered by further philosophical progress. The rule of definition of terms by general notion and specific difference is applied to a particular case: (12 D: ei pépos 51 J. Wagner, Zur Athetese des Dialogs Euthyphron, Brinn 1883. 12 Yxem, Ueber Platos Huthyphron, Berlin 1842. 8 The Euthyphro of Plato, with an introduction and notes, by George Henry Wells, London 1881. ‘4 The Euthyphro of Plato, with introduction and notes, by J. Adam, Cambridge 1890. 18 A, Jezierski, Platona Eutyfron, Tarnopol 1890. 1% On the logic of the Huthyphro, see also V. Poggi, L’ Hutifrone di Platone, Roma 1891. SOCRATIC STAGE: EUTHYPHRO 199 TO Govov Tov Scxaiov, Set . . &Eevpety To crotov pépos), but without any methodic digression on logical theory such as appears in all the dialectical dialogues. Induction and analogy are used frequently (as 13 a, 14 a, &c.) and the necessity of establishing permanent notions is insisted upon (11 D: éBovAduny adv pos Tovs Adyous pévev Kal axuvy- tws iSpdcbar padrov % mpds tH Aatdddov aodia Ta Tavrdrouv ypnpata yevéoOa: see also 5D). Enumeration of examples is shown to be insufficient to give such per- manence to a notion (6D: ovy gy ty % S00 TaY TOAN@Y ociwv, AAN éxeivo avTo TO éldos, O TavTa Ta bola GoLd éoriv) and the characteristic mark is sought for. This characteristic markis here named eléos, in the sense in which Thucydides used this word when he spoke of an eldos vocov (Thucyd. 2, 50). Some authors, as for instance M. Waddington,'” thought it possible to draw chrono- logical inferences from the absence of the words «tédos or iifa in many small dialogues. M. Waddington is evidently not aware of the fact that both words are anterior to Plato, and are used by Thucydides and other earlier writers in the same sense as by Plato in his early dialogues. In the Huthyphro as in the Charmides they both occur, ié¢a in the meaning of form, property, or characteristic mark (6 E: pid idéa Td Te avoolia avoo.a elvat), but not in the later meaning of a metaphysical entity. From the occurrence of these words, which are not yet used as logical terms, we cannot infer that the Huthyphro is later than any other small dialogue, such as the Apology or Crito, from which these words are absent. There is a greater difficulty in the circumstance that in the Euthyphro (6 E: ypwpevos airy (TH tdéa) mapadelypatt) the idea is said to be a paradeigma, as this seems at first sight to approach the later theory of eternal forms or paradeigmatic ideas. But such eternal 87 ©. Waddington, ‘ Observations sur le Mémoire de W. Lutostawski,’ Compte rendu des séances et travaux de Vacadémie des sciences morales et politiques, vol. cxlvi*. N. 7. See above, note 49. Induction and defi- nitions, frequent use of analogy. Though elSos and idéa both occur in the Euthy- phro, these words have not yet their technical meaning. The same refers to wapd- Seryua. Qualities distin- guished from causes. Date early but un- certain. 200 ORIGIN AND GROWTH OF PLATO’S LOGIC forms are ‘rapadeiypata év tH duce’ (Parm. 132 D =Rep. 597 B, cf. Theaet. 176 £), while here Plato only speaks of using the characteristic of holiness as a standard for distinguishing holy actions from sinful deeds. Such a use of the word rapdderyya does not essentially differ from that of Thucydides and the early orators; it cannot be regarded as peculiar to Plato. An important logical distinction is made in the Euthyphro between activity and quality: the quality is a result of a determinate activity, but never cause or ground of this activity (10 c: e re yirxyveras, 7 Te Tao xKEL, ovy OTL yiyvomevovy soti, yiryveTtat, GAN Ste yiyverat, yuyvouevov eotiv: ovd Ott wadoxyov éoti, maoyE, adr dtu mdoyel, tacyov éoriv). This is here explained by a number of analogies before it is expressed in a general form. These few hints of a logical character offer no means of determining the date of the Huthyphro. The scene of the dialogue proves that it could not have been written before the accusation of Socrates. With regard to the later limit of time we can infer nothing beyond that the Euthyphro precedes the Meno and Gorgias on grounds of style,’ composition, and contents.'° 188 Stylistic observations place the Huthyphro at the beginning of Plato’s literary career. It contains many peculiarities of earlier style : écmep used always instead of xaSdwep, rq dvr: instead of dyrws, wévro: prevailing over tolvuv ; Eywye, Epovye, Sones wor forming 19 per cent. of all answers, wep! with the genitive prevailing over all other prepositions, &c. (See table of affinity, p. 163.) 18° Schleiermacher, Socher, Schierenberg (Ueber die Zeit der Abfassung des platonischen EHuthyphro, Lemgau 1830), Stallbaum, Steinhart, Zeller agree in placing the Huthyphro before the death of Socrates; Susemihl, Georgii (Platos Euthyphron iibersetst von Georgii, Stuttgart 1875), Bergk, Diimmler believe the Euthyphro to be written some years later chiefly on account of holiness being here a part of justice, while in the Gorgias it isa fifth virtue besides justice. Also H. Ritter, Brandis, Michelis, Ribbing, Mistriotes, Peipers, Weygoldt, Windelband, Christ, who admit the date of the ELuthyphro as uncertain, agree, however, as to the Socratic character and early origin of this work. Only Teichmiiller (ii. 355) places the Huthyphro SOCRATIC STAGE: APOLOGY, CRITO 201 The Apology shows, like the Huthyphro, a frequent use of induction and analogy (e.g. 25 Bc), and contains several repetitions of the well-known Socratic principle, that he who knows his own ignorance is wiser than those who believe themselves to know what they do not know (21 cp, 22 c, 29 a, 33.c, 41 B). This principle is carried to the extreme consequence, that all human knowledge is of little worth and that only God is wise and infallible (21 B, 23 A). Such a scepticism, bearing even upon the future life (29 a: olde ovdeis tov @avarov....also 42 A), does not extend to ethical con- victions (80 D: to do injustice is worse than death—- 30 B: virtue imports more than all besides). The uncertainty manifested as to a future life shows that the Apology was written earlier than the Meno and Gorgias, in which asin all later dialogues Plato professes the greatest certainty on this subject. Also the style of the Apology, very similar to the style of the Huthyphro, makes it probable that both dialogues were written not later than within the first years after the death of Socrates, and though the Huthyphro represents an earlier scene, there is no decisive reason to place it before the Apology.'® The Crito forms the third act in the tragedy of which the Euthyphro and Apology represent the first scenes. We remark here a curious distinction between honest (ypnoras) and immoral opinions (474: movnpas do€as), which is parallel to the later constantly repeated contrast between mere opinion and knowledge. This way of estimating a judgment according to its moral value, with- out asking for a logical standard of truth, is peculiar to the Socratic stage of Plato’s logic, and shows us how after the Symposium and even after the Theaetetus, under the influence of his wrong theory of the stylistic criterion (see above, p. 102). 10 Zeller and Ueberweg believed the Apology to be a faithful account of what Socrates said before his judges. But Riddell (see above, p. 99) and Stock (The Apology of Plato, with introduction and notes by S. G. Stock, Oxford 1887) have sufficiently demonstrated the improbability of this supposition. In the Apology we find a frequent use of analogy. Charac teristic uncer- tainty about future life. In the Crito honest and wicked opinions distin- guished. Compe- tent au- thorities trusted. The absolute authority of reason not yet es- tablished. Crito probably later than Apology. 202 ORIGIN AND GROWTIL OF PLATO'S LOGIC Plato was led from the moral teaching of his master to his own logical investigations. When he wrote the Crito, he seems not yet to have arrived at his later ideal of objective knowledge: he is satisfied with an ‘honest’ opinion of a competent expert (éralwy 47D) whom he trusts more than the opinion of the many (d0fa ray jov\Xov 47 C). In agreement with such a practical standpoint, funda- mental differences of opinion between men are recognised as inevitable, and here, as in the Huthyphro, are admitted to produce hatred and contempt, if they touch upon ethical subjects (Crito 49c D, EHuthyphro 7vD). This view 18 very characteristic, because in the Gorgias and all later dialogues the Platonic Socrates is represented as possessing objective truth about ethical as well as about other matters, a truth which can be proved and com- municated even to such enemies of philosophy as Kallikles. Here we see only competent opinion or the authority of the ‘ best’ reason (46 8B: pndevi drAr@ reiOecOar 7} TE AOYH Os dv pot Aoyilouévm BérTicTOs gaivnra). This ‘ best ’ reason is not yet ‘the reason’ familiar to the readers of later dialogues. From these logical particulars we can only infer that the Crito,'*' forming with the preceding two dialogues a natural group, is earlier than the Meno and Gorgias. There 1s @ great probability that the Crito is later than the Apology, because in p. 458 Plato makes a clear allusion to his Apology. This allusion might also refer to a coin- cidence between the Platonic Apology and the historical defence of Socrates, but if we consider that also the style 1) The doubts as to the authenticity of the Crito expressed by Ast, and later by Schaarschmidt, have been sufficiently refuted by J. H. Bremi (Philologische Bettrdge aus der Schweiz, Ziirich 1819, vol. i. p. 181 sqq.), Georgii (Apologie und Krito tbersetst von L. Georgii, Stuttgart 1883), J. Adam (Platonis Crito, with introduction, notes, and Appendix, Cambridge 1888), and many others. The relation of the Crito to the Gorgias is dealt with also in Plato’s Apology of Socrates and Crito, on the basis of Cron’s edition, by L. Dyer, Boston 1885. SOCRATIC STAGE: CHARMIDES 2038 of the Crito shows a slight advance over the style of the two preceding dialogues (see above, p. 163), we have good reason to admit that Plato himself intended this work as the supplement of the preceding. Less evident is the chronological relation of the Charmides '® to the above three dialogues. It 1s charac- teristic of the stage of logical advance which Plato had reached when he wrote this small work, that his So- crates commits a paralogism, inferring from the beauty of both temperance and quickness that quickness 1s temperate (159 D). Such logical blunders occur also in other small dialogues, and we have no reason to suppose that Plato was conscious of them. So long as the logical interest was not awakened, even a thinker lke Plato might unconsciously commit logical errors. On the other hand, we notice a correct syllogism (1614: aidas ovK ayaboyv . . . awhpoctvn ayabov . . . odK apa cwdpo- ovvn av ein aidws) of the form Cesare, introduced by the word cudAAoyiodmevos (1605), which, however, has not yet the meaning of a logical term. The allusion made by Critias to a possible division of sciences into practical and theoretical (165 E : ris NoyloTtinns .. Ti eat ToLodToy Epyov oloy oixia olKodopwKys), carried out later in the Gorgias, is not developed here; ériotnun and téyvn are used as synonyms (165 £), but theoretical knowledge, independent of personal considera- tions, is recognised as a great advantage to mankind '62 Doubts as to the authenticity of the Charmides put forth by Ast, Socher, Suckow, Schaarschmidt, and recently by Troost (Inhalt und Echtheit der platonischen Dialoge auf Grund logischer Analyse, Berlin 1889) have been sufficiently refuted by Schleiermacher, Ochmann (Char- mides Platonis num sit genuinus quaeritur, Vratislavie 1827), Stallbaum, H. Ritter, Hermann, Steinhart, Munk, Susemih], Spielmann (Die Echtheit des platonischen Dialogs Charmides, Innsbruck 1875), Alberti (‘Gesichtspunkte fir angezweifelte Platonische Gespriiche,’ Philologus, 3¢* Suppl. Bd. p. 101, Gottingen 1878), and Georgii (Laches und Charmides, tibers. von L. Georgit, Stuttgart 1882). Also Zeller, who formerly believed the Charmides to be spurious, has since defended the authenticity against Troost (Arch. f. Gesch. d. Philos. vol. iv. p. 184). In the Char- mides the term ovAdAo- ytoduevos. Import- ance of theoreti- cal know- ledge re- cognised. But certitude of know- ledge doubted. General logical question about activities acting on them- selves ad- journed as requiring a future great thinker. Char- mides is early, though the exact date un- certain. Laches belongs to 204 ORIGIN AND GROWTH OF PLATO'S LOGIC (166D: Kxowov dyabov elvar ocyedov te Tacw avOpwrors, yiyverOar xatadhaves Exactov TaY bvTwY Ory FyEI). Again, a sceptical tone is perceptible in the doubt whether certitude as to knowledge is possible (1724: ayabov ein To eidévas & te oldév TLS Kal & pn oldgev.. - ovdapod ériotnun ovdemia TovavTn ovca mépaytat). Very characteristic of the Socratic stage of Plato’s logic is the appreciation of the knowledge of knowledge according to a standard of usefulness (169 B: ov« azrodéyopar .. . mpiv dv drickeyrwpat, eite TL av Has operol, 172D: cKxepo- peOa, si apa te Kal ads ovnoe . . TO eidéva & TE oldev cai & py? oléev). This would not occur in any dialogue after the Meno, but is very natural at the time when Plato had not entirely emancipated himself from the prevailing ethical preoccupations of his teacher. A beginning of later Platonic tendencies appears in the care with which the question, whether an activity can have itself as its object, is discussed. By many examples Plato tries to prove that most human activities have not this property, that, for instance, there is no perception of perception (167 c), no desire of a desire, no will of a willing (167 &), no love of love, no fear of fear, because each of these activities has an object different from itself, but the general question of the existence and possibility of a know- ledge of knowledge is here not settled, only adjourned as a problem requiring for its solution a great thinker (169 4). Nearly all investigators agree in placing the Charmides among Plato’s early works. Many believe that it may have been written even before the death of Socrates, to which it contains not the slightest allusion. But an exact chronological determination in this case requires further stylistic research, and the attempt of Teichmiller to dis- cover in the Charmides allusions to the Memorabilia of Xenophon has failed. At all events the Laches '}* belongs to the same period. 1 Ast, Schaarschmidt, and to a certain extent Giltbauer (Philologische Streifetige, Freiburg 1886) doubted the authenticity of the Laches, but SOCRATIC STAGE: LACHES 205 It is noteworthy that Plato mentions here as objects of knowledge truths which are conceived as independent of time (198 D: qepl dowv sory ériotiun, ovK GAA pev elvar aepl yeyovoros, eidévar barn yéyovev, AdAn 52 Tept yeyvopévev, Sin yiyvetat . .. GAN 7 avtn). Such truths are found more easily by a single competent man than by an incompetent majority (185 a: ei gorw Tis Teyvixds. . . éxeiv@ wetOec Oar évi dvi, rovs & adXovus 2av), because know- ledge is a safer criterion than great number (184 E: ériotypun Se? xplvecBar AX ov AGEL TO pwédrrov Kaos xpLOncec Oar). This short acknowledgment of knowledge as superior to opinion rises above the moral standard of honest opinions required in the Crito. But Plato does not yet pretend, as in later works, to possess such a knowledge. He advises his readers to seek the best teacher, without sparing money or anything else (201 a), but he offers no definitive solution of the proposed difficulties. In all the above small dialogues we see discussions leading to a Socratic confession of ignorance, and not to a definite doctrine. Opinions of others are criticised, but not defi- nitely corrected. The character of Socrates is similar in these works to what we know about the historical Socrates : he is repre- sented as a friend of young men, detecting their errors, not yet as the ideal master of wisdom. Of a similar critical character is the first larger work written by Plato, the Protagoras. In this dialogue also logical questions are only incidentally touched upon, and it is evident that the author cares chiefly for ethical problems. These are treated in a manner which presupposes the previous particular inquiries given in the small dialogues, and the logical power also appears increased. The inconvertibility of general affirmative judgments is insisted upon these suspicions have been refuted by Stallbaum, Georgii, Bonitz, and Tatham (The Laches of Plato, with introduction and notes, London 1888). Also Zeller abandoned his earlier doubts as to the authenticity of the Laches. the same period. Personal authority of the best teacher. In the small dia- logues no definitive doctrine. Also the Protago- ras has & polemical character. Incon- vertibility of general affirma- tions. Law of contra- diction prepared. The Pro- tagoras seems to be later than the small dialogues. Defini- tions of courage. 206 ORIGIN AND GROWTH OF PLATO'S LOGIC (350 c-351 B) by means of several analogies. If we observe that this logical lesson is put into the mouth of Protagoras, and not of Socrates, we must admit as probable, that the discovery was made outside of the Socratic society. The perfect knowledge vainly sought for in the Charmides is not yet found by Plato. He still expects progress from discussion (848 D). His certitude is increased by the acquiescence of others, and not by its own absolute infallibility, as in later times, when he condemned to death those who thought otherwise (Laws 909 a, 958 a: cf. Polit. 8308 £). Still he recognises knowledge as the chief power in man, reigning over all feelings (852 c, 357), and settling all doubts (356E : Snrtwcaca TO adnOis Hovyiav dv eroincey Eyew THY Yuyny pévovoav él t@ adnfet ...). As one of the logical means of arriving at knowledge, Plato states the principle that each notion has only one contradictory to itself (832.c: él dxdorm Tadv eévavriwy by povov éorly évavriov xal ov odd) and exemplifies this rule by many instances, but without making any distinction between contrary and contradictory terms. These observations seem to indicate a further stage of logical development than is seen in the small dialogues. In the Charmides the subject, though restricted to one form of virtue, was to a great extent the same as in the Protagoras, and it seems more plausible that the greater work should contain no allusion to the smaller than that Plato should have written the Charmides after the Prota- goras without some allusion to the more general discussion on the same problem. The special subject of the Laches also is contained in the Protagoras, and the definition of courage (Lach. 195 a: rav Sewav nal Oapparéov émirrnn), arrived at in the Laches after a long conversa- tion, and shown by Socrates to refer not only to courage but to every virtue, is repeated in the Protagoras (3860 pD: 4 codla trav Sewav kal py Sewadv dvépeta éoriv), and remains unrefuted (see also Rep. 429 c). SOCRATIC STAGE: PROTAGORAS 207 Some allusions to contemporaneous facts, contained in the Protagoras, seem to show that this dialogue was written at least seven years after the death of Socrates. Kroschel '* and after him Teichmiuller have supposed that the mention of zeATacriny (8504) as a familiar example was not probable before the introduction of this arm into the Athenian army by Iphikrates, between 393- 391 B.c. Also Teichmuller and after him Dummler see in the Protagoras (347 c-350 B) allusions to Xenophon’s Memorabilia, which appear to have been published some years after the death of Socrates. This agrees with our supposition that the Protagoras followed the above five small dialogues, and also with the observations on the style, according to which the Protagoras 1s intermediate between the small dialogues and the Gorgias (see above, p. 165). The Meno is generally held to be a continuation of the Protagoras.' Theories of the greatest importance, amounting to logical discoveries, are for the first time expressed in the Meno, which in size exceeds only by a very little the limits of a small dialogue, and amounts to less than two-thirds of the volume of the Protagoras. Logical exercise, so often recommended in the’ dialectical works, is here first introduced as a methodical way of progressing on the path of truth (754: in order to enable Meno to find the definition of virtue, Socrates proposes the definition of form: f%va kal yévntal cou werérn). The aim of logical definition is indicated as the deter- mination of the substance (72B: ovcia) of things, that 14 J, §. Kroschel, ‘Studien zu Platons Protagoras’ (Jahrbiicher fur classische Philologie, vol. 87, p. 825, 1863), also in his review of Cron’s edition of the Protagoras (Zeitschrift fur das Gymnasialwesen, vol. xx., for 1866), and in his edition of this dialogue (Gotha 1865, as 8¢ ed. of Stallbaum). 1% Nearly all investigators agree that the Meno is later than the Prota- goras:.Tennemann, Schleiermacher, Hermann, Susemihl, Ribbing, Steinhart, Zeller, Ueberweg, Pfleiderer, Natorp, Siebeck, Gomperz, Ritter, J. Bartunek (Ueber die Aufeinanderfolge der Dialoge Protagoras, Gorgias und Menon, Progr. Rzeszéw 1897) &c.; only Stallbaum, Schéne, and F. Horn ad- vocated the priority of the Meno on quite insufficient grounds; R. Hirzel (Bheinisches Musewm, vol. 42, p. 249) sees in the Meno allusions to Polykrates’ xarnyopla Zwxpdrovs. Allusions to known events as chrono- logical indica- tions confirm the later date. Meno con- tinues the question raised in the Pro- tagoras. Unity of species. Dia- lectical require- ments. Hypo- thetical method of investi- gation. Opposi- tion of general and par- ticular judg- ments. 208 ORIGIN AND GROWTH OF PLATO'S LOGIC which brings unity among the variety of external appear- ances (72C: avrd rodro @ ovdev Svadgpovat, adr} TadTov siow dracat). This unity is called eféos, not yet the later Platonic idea, but already a distinct logical term, cor- responding to species (72 Cc: & ye rt eidos ravTov aracat Eyovaw, 8.’ 8 eiciv aperatl). The unity of species is the true essence of the things which it embodies (100 B: av’rd nad’ avro ri mor’ got apern). Having thus established the aim of research, Plato proceeds to give some rules as tothe method. Here appear for the first time the ‘dialectical’ requirements. Xeno- phon had once applied (Memor. iv. 5,12: avdpas d:ane- xrucwtarovs) the word ‘dialectical’ in the sense of ‘ best able to conduct conversation,’ but Plato, converting it into a logical term, requires of all who wish to discuss dialectically that they should base their reasoning on recognised notions or premisses (75D: gore 52 tows To Siarextiunorepov py rovoy tadnOn arroxpiverOar, adra cal 80 éxeivwv dv Av mpocoponroyy eidévat 0 epopevos). As a method of verifying doubtful suppositions, Plato proposes to look for the consequences following from each hypothesis. This method he describes as hypothetical rgument (86 E: && i7robécews cKorreicbar . . . datrEp of yewpérpa.), and transfers it from geometry to philoso- phical inquiry. He applies it successfully to the problem which he could not yet resolve in the Protagoras, and finds that virtue, so long as it is not taught, but merely practised according to common traditional experience, appears not to be, as was supposed in Charmides, Laches, and Protagoras, a kind of knowledge. Another sign of the awakened logical interest is the careful distinction between particular and general affirma- tion (73 E, 89 4). Such progress in respect of formal reasoning corresponds to an equally remarkable develop- ment of some fundamental logical doctrines about which neither in the Protagoras nor in any of the small dialogues had Plato expressed any opinion. The theory of innate SOCRATIC STAGE: MENO 209 ideas is not only introduced with a striking audacity, but founded on so general a metaphysical axiom as the unity of nature (81 D: dre yap tis fucews araons cuyyevovs ovons, Kai peuabnnvias rhs wuyns atavra, ovdey Kwrvee Sy povoy dvapvncbiyta .. . TadNvNa TavTa... aveupeiy). The metaphysical certainty of a priort knowledge, proclaimed by Plato in the Meno, is a new principle in the light of which the old Socratic irony and ignorance are disappearing. Still the author condescends to give an experimental and inductive proof of his assumption, after the caution that such a proof is not easy (824). The choice of the experiment and the manner in which it is executed show an educational Mastery far greater than that visible in the small dialogues (82 B-85 c). All doubts about the possibility and reality of infallible science have been removed; the Platonic Socrates boldly asserts his absolute certainty of the existence of a science far above right opinion (98 B: Stu S¢ éoriv Te dAXNotov OpO7n Sofa xal eriornun, ov mavu poe Sonam TodTo eixalety, GXN’ elrep TL AAO hainv dv eidévas, Oriya & dv hainy, tv § ody Kal Todo éxsivwy Oeinv dv oy olda), and that this science may be awakened in everybody by means of skilful inter- rogations (86 A: ddrnOeis Sofa: epwrnce: ésareyepOeioat éricthpas yiyvovra). The difference between nght belief and scientific knowledge consists in the co-ordination and causal relation peculiar to true knowledge (98 A: ad@eis Sofas . . . ov woddod Gkvai eiow, bws dv tis adras Shon aitlas Noytop@ ... areddav 58 SeOdou, mparov pev émiothpas ylyvovrat, krera povipor cal Sia radra .. « bsadsper Seopw sriornyn opOns Sokns). Science is therefore more valuable than mere belief, even if it be right belief. Armed with his new weapon, Plato enters upon its application to the ethical field, and introduces the immortality of the soul first as a true and beautiful tale of priests and poets (81 A), which he then confirms by a reflection on the nature of human thought (86 B: ovxoby P Innate ideas. A priort knowledge made pro- bable by experi- ment. Know- ledge pro- claimed as essentially different from opinions because it is founded on grounds. Applica- tion to im- mortality. Date of the Meno after 395 B.c. The Euthy- demus directed against unknown enemies. Philo- sophy and dialectic 210 ORIGIN AND GROWTH OF PLATO'S LOGIC si det adNOeLa huiy rdav dvtav sorly ev TH ~ruyy, abavaros ay h sux ein). This far-reaching logical importance of the Meno, noticed already by Guggenheim '® and Oldenberg,'” tells against those who like Socher and Stallbaum believe that the Meno could have been written before the death of Socrates. The allusion to the bribery of Ismenias, indicated by Boeckh and Schleiermacher, shows that the Meno is later than 395 B.c. Less evident is another allusion to Polykrates, maintained by Hirzel and Diimmler, who place the Meno after the Symposiwm, an order which appears impossible, if we take into account the stylistic tests (see above, p. 166). What may be safely affirmed is that the Meno is later than the Prota- goras and all smaller dialogues. The logical interest awakened in the Meno bursts out only occasionally, but with great intensity in the Huthy- demus,'® which has all the appearance of a polemical work written for a certain practical purpose, and against enemies whom it is not quite easy for us to identify. Plato 1s so proud of his acquired certainty of knowledge that he would not give it up even for immortality, if not accompanied by knowledge how to use it (289 B). While in the Protagoras the word philosophy was still used in the meaning of love of wisdom (335 D, 342 p), here we see it defined as acquisition of knowledge (288 D: girocodla xrijois imiotnyns), and the dialectician, who had received his first rules in the Meno, becomes the highest judge of every particular knowledge (290 c). 10 M. Guggenheim, Die Lehre vom apriorischen Wissen in ihrer Bedeu- tung fur die Entwickelung der Ethik und Erkenntnisstheorie in der Sokratisch-Platonischen Philosophie, Berlin 1885. 167 H. Oldenberg, De Platonis arte dialectica, Géttingen 1878. 1 Doubts as to the authenticity of the Huthydemus, emitted by Ast and later by Schaarschmidt, have been sufficiently refuted by A. Polzer (Ueber die Echtheit des Huthydemos, Olmiitz 1874) and Bonitz (Platonische Studien, Berlin 1886). Bonitz gives also an elaborate classification of more than twenty sophisms contained in the Huthydemus. SOCRATIC STAGE: EUTHYDEMUS 211 These dialecticians, thus placed so high above the mathematicians and all other inquirers, are evidently Plato himself and his school. For the writer of the Euthydemus is clearly a teacher, though probably not yet the head of the Academy. Philosophy is the subject of his teaching, and he passionately defends his science against those who call philosophy a worthless and vain occupation (304 £). To the right belief, explained in the Meno, Plato adds in the Huthydemus his explanation of error and wrong belief, whose existence is proved against the Sophists by the hypothetical method taught in the Meno (Euthyd. 284 A, 287 ©). Plato gives an interesting collection of current sophisms resulting from the use of the same word in two different meanings, the misinterpretation of predi- cation, the omission of limiting determinations, and the double meaning of phrases according to their grammatical construction. The date of the Huthydemus can be approximately determined by its admission of the possibility of teaching virtue (as in the Republic and Laws), whence we conclude that it was written after the Protagoras and Meno, in which the same question is discussed. Those who, like Tennemann, Stallbaum, Steinhart, C. Ritter, believe the Euthydemus to have been written before the death of Socrates cannot account for the logical enthusiasm which is here manifested and is absent from all earlier dialogues. Those who, like Bergk, Siebeck, and Weygoldt, place the Euthydemus after the Symposium are not aware of the great difference in style between the Huthydemus and all dialogues later than the Cratylus and Symposium (see above, p. 166). There is no contradiction from the standpoint either of logical or of stylistic development in admitting the close relation between the Huthydemus and Isocrates’ discourse against the Sophists. This relation, first P 2. defended and placed above particular sciences. Many sophisms refuted and their origin explained. Date of the Eu- thydemus has been supposed by some writers to be very early. Allusion to Iso- crates’ discourse against the Sophists is a safe indica- tion, and this con- firms in- ferences from style. Gorgias represents the tran- sition from the Socratic 212 ORIGIN AND GROWTH OF PLATO'S LOGIC noticed by Spengel,'® and Thompson,!”? has been since investigated by Teichmiiller, Sudhaus,'”) Diimmler, and recognised by Zeller and Susemihl, without any note- worthy opposition. According to these investigations, the Euthydemus must have been published not before 390 and probably not much later. Another allusion to Lysias, although supported with great ingenuity by Teichmiiller, is not quite so evident, and also the references to Anti- sthenes, alleged by Teichmiiller, Urban,'? and Diimmler, are possible, but not certain. If we admit that Plato wrote the Euthydemus ‘? about 390 B.c., this agrees very well with the general character of the dialogue, which directs the most acute polemic against wrong education, thus seeming to indicate that the author had already acquired some educational experience, and gathered around him a number of pupils, preparing the foundation of that philosophical school which achieved such an un- paralleled importance in the history of human thought. This educational character reaches a still higher level in the Gorgias, which represents the transition from the Socratic to the peculiar Platonic philosophy. In its ethical character the Gorgias is still Socratic, but the method of argumentation and the apodictic certainty with 1 Spengel, ‘Isokrates und Plato,’ Abhandlungen der Akademie zu Munchen, vol. vii. pp. 729-769, Miinchen 1855. 170 The Phaedrus of Plato, with English notes and dissertations, by W. H. Thompson, London 1868, p. 179. 11 Sudhaus, ‘Zur Zeitbestimmung des Euthydem, des Gorgias und der Republik,’ Rheinisches Musewm, vol. xliv. p. 52, Frankfurt a. M. 1889. 172 Urban, Ueber die Erwahnungen der Philosophie des Antisthenes in den platonischen Schriften, Konigsberg 1882. 173 Some authors inferred from the use of wdpeors Euthyd. 8014 that Plato when he wrote the Euthydemus had already produced his theory of ideas. But this is by no means probable, because wapetvac is used in exactly the same manner in some of the small dialogues, as Charm. 1594 and Lys. 217p, like wapaylyvecda: in the Laches 1898. This use does not correspond to the terminology of ideas. Instead of wdpecrs ndAdos (Huthyd. 8014) Plato would have said later wdpeor: rd ndddos (adrd nab’ airé). Generally wapeiva: is very little used by Plato in connection with ideas. SOCRATIC STAGE: GORGIAS 213 which ethical principles are proclaimed (509 A: ovdeis olds stage to 7 gotiv dddNws Ayr 7) ov KatayéNaoTos elvat) belong to original Plato, are his own creation, and are manifested constantly Platonic in all his later works. The literary skill displayed in the PB Gorgias reaches a higher perfection than in the small sp: dialogues, and even than in the Protagoras, Meno, or Euthydemus. Plato has now arrived at a mastery of form, which approaches to the highest beauty attained by human language, and has been exceeded perhaps only by Plato himself in the Phaedo, the Phaedrus, and parts of the Symposium, the Republic, and the Theaetetus. The teaching of those dialecticians, who were indicated Philo- in the Huthydemus as treasurers of knowledge, is now sophy per personified and attributed to ‘Philosophy.’ This Philo- *nified. sophy is loved more than all human beings, and is * sae credited with eternal truths, which never change (482 a: ie jy ) pirocogia asl rév avtav sort Aoywv). The power of tality, these truths is based on our own consciousness, nor can any man contradict them without contradicting himself (482 B: 7 didrocodiay 2&éreyEov . . . 7} 0 cou oporoynoes Karrxrdijs, & Karrtxras, adda Scadwvyce dy arravtt To Bi). And to all faithful followers of this his Queen, Plato promises after death a happy life, apart from other human beings (526 c). In this he still betrays a juvenile egoism, which was abandoned later, when he bade the philosophers descend like gods among mortals to teach them a better life. The difference between right belief and scientific Difference knowledge, found in the Meno, is here applied to the art of between persuasion, and leads to the distinction of two kinds of >elief and rhetoric, one based on knowledge, the other on faith oe (454 &: Sv0 eldn Odpev meBods, To pav riot Tmapexopevoy . a dvev tod eidévar, 7d 8 érreoryjpnv): knowledge alone is in- pied, fallible (454 D: dmructnyn ovdapas éotiv ~pevdys), while belief may be true or false. In full accordance with this increasing separation between science and opinion, Plato distinguishes more clearly than in the Charmides between Logical terms. To do wrong is worse than to auffer wrong. Pleasure is not the aim of life. Great politicians treated with contempt. This shows independ- ence of tradition 214 ORIGIN AND GROWTH OF PLATO’S LOGIC theoretical and applied or practical sciences (450 c-451 D), and he insists on the importance of the division of con- cepts (500 D: BéAriordy éotw . . . Starpsio Oar, dvedopévous 52 xal oporoyncavtas ddAnAOs . . . oxépacOat, Ti Te dia- déperov adAndow). The reasoning proceeds on granted premisses, according to the rule given in the Meno, and the logical connection is carefully shown by means of logical terms (498 E: gcvANoytcal, TL Hutv cupRawe &k ToVY @poroynpEVvor). Inevitable repetitions are excused by the logical aim (499 a: cal dls yap rou Kal Tpis hacw Karov selva Ta Kara Neve TE Kal &ricKxoretoOat, cf. 508 D). This gives the im- pression of an author who is used to personal teaching, and has already found the truths he wishes to convey to his hearers, but professes to seek them again in company with his pupils. Whatin the Apology (80D) and Crito (49 ac) has been expressed as a personal belief, that one should by no means do wrong, is here affirmed as a well- founded scientific truth (509 A: radra ... npiv otto gavévta Kxatéyerar Kal Séerar oidnpois nal adapavtivors Noyous), and is so far extended as to imply even the necessity of punishment if one has done wrong (482 B, 527 B). The aim of human life is not, as it seemed to be in the Protagoras, pleasure but ‘the good’ (513 D: év’ Epapev elvar tas wapackevas earl To Exacrov Oeparrevew, .. . play pay rpos 4 Soviv opirely, thy érépav 58 mpos Td Béxtecrov). The politician’s duty is to make better the people whom he leads. In the Protagoras and Meno Plato still maintained the popular belief that Pericles and Themistocles were great and wise men. He only complained that they were unable to impart their greatness and wisdom to their children or others. But now, from the height of the newly founded philosophy, Plato dares to say that these idols of the Athenians were bad politicians and corrupters of the people (515 £). This bold contempt of the men who had generally been esteemed greatest among the SOCRATIC STAGE: GORGIAS 21a citizens of Athens shows how rapidly the breach is widening for Plato between vulgar common sense and the teachings of philosophy. He has risen from Socratic ignorance and irony to that full independence of tradition and public opinion which in all ages characterises a great philosopher. Another indication of the later date of the Gorgias is the hatred of tyranny (525 D) here expressed and henceforth maintained by Plato throughout his life. Stylistic inquiry places this dialogue after all the above-mentioned works, and between the Huthydemus and the Cratylus (see above, p. 167). If we admit with Teichmiiller that the Pro- tagoras and Huthydemus were written between 393- 390 B.c., we are not obliged to accept his supposition that the Gorgias is fifteen years later. Teichmiiller (i. 357) as well as Sudhaus '! place the Gorgias after Isocrates’ discourse to Nicocles, which 1s supposed to have been written 376 B.c. But the allusions to this discourse supposed to be contained in the Gorgias are not evident, while Dummler, who also specially investigated Plato’s relation to Isocrates, assigns to the Gorgias a much earlier date. The most certain conclusions as to the date of the Gorgias that can be drawn from the contents have been indicated by Natorp '“: the Gorgias is probably later than the Protagoras, Meno, and all above-mentioned small dialogues. This is also the result reached by Horn in his comparison of the ethical theories of these works. The Gorgias '® closes the Socratic stage of Plato’s 14 P, Natorp, ‘ Ueber Grundansicht und Entstehungszeit von Platos Gorgias’ (Archiv ftir Geschichte der Philosophie, vol. ii. p. 394, Berlin 1889). 178 The Gorgias is one of the few works of Plato which has escaped the searching criticism of those who have doubted the authenticity of many other dialogues. Voluminous and instructive commentaries on the Gorgias have been published by Findeisen (Platonis Gorgias, Gothae 1796, 624 pp.), D. Coray (Hevopavros ’Amopynuovedipara cal TAdrwvos Tépyias éxdlSovros Kal Biopbotvros "A8auxayriou Kopay, évy Mapiolos 1825), Ast (Annotationes in Pla- tonis Opera, tom. ii. Lipsiae 1882), Woolsey (The Gorgias of Plato, Boston 1842), Cron (Beitrdge zur Erklarung des Platonischen Gorgias, Leipzig and public opinion. Gorgias the latest of all Socratic dialogues, as results from its contents and style. This confirms the con- clusions of Natorp and Horn. Plato’s progress from moral pro- blems to logical investi- gations. His dis- covery of scientific certainty. Rules for dialectical discus- sion. Infallible knowledge attained. 216 ORIGIN AND GROWTH OF PLATO'S LOGIC philosophy, and leads from the ethical problems which occupied him in the first years after the death of his master to the logical and metaphysical inquiries which filled the greatest part of his manhood. Looking back over the above survey of Plato's first steps in logic, we see that he started from ethical problems, agitated by his teacher, and that his first attempts to find a definition of particular virtues and of virtue generally were made with moral purposes. In order to be temperate it seems to be indispensable to know what temperance is, and where is the limit separating this virtue from in- temperance. Among such inquiries on particular virtues Plato became interested in the more general problem of a definition of virtue. This he began to seek, and after some vacillation recognised the identity of virtue and know- ledge. But he was still unable to attain certainty of knowledge ; only after years of educational practice he found that such certainty is possible, and not to be sought for in the assent of any majority, nor in tradition, nor in idle discussion, but in the inward power of the soul which sees the truth with absolute certainty. To trace the origin of this power, felt by him when he imparted his moral convictions to his pupils, he recurred to the hypothesis of a previous existence of the soul, and deduced also the soul’s immortality. We see the influence of his activity as a teacher in the rules for dialectic discussion, consisting in starting from re- cognised premisses, in dividing and distinguishing notions, in following up the consequences of each hypothesis, and avoiding unjustifiable generalisation. By these means Plato reached a degree of certitude not experienced before. He created an ideal of infallible knowledge, far above traditional opinions, and he distinguished this scientific knowledge from common belief by his ability to show a reason for each assertion. The methodic connection of 1870, G. Lodge (Gorgias, edited on the basis of the Deuschle Cron’s edition, Boston 1891, 308 pp.), and many others. SOCRATIC STAGE: GORGIAS 217 thought gave to his conclusions a permanence and con- sistency which unscientific opinion never reaches. The new power of philosophy, acquired by logical exercises undertaken with ethical purposes, reacted first on the moral problems from which Plato started. He applied his logical method first to the great questions which had been unsuccessfully discussed in his earlier writings, and he produced a consistent theory of virtue and of the aims of life in the Gorgias. But the logical progress achieved will not be limited in its effect to the subject for which it has been devised. We see already in the Meno, in the Euthydemus, and in the Gorgias, that Plato begins to feel an interest in logical method in- dependently of its applications, and this logical interest, once awakened, will lead him to special logical investiga- tions, and to further development of methods in order to acquire and communicate to others an infallible know- ledge. An almost fanatical enthusiasm and love of absolute science explains certain exaggerations: the new know- ledge referred only to very few principles, but Plato is as proud of it as if he had already extended it to all depart- ments of Being. He obtained a glimpse of a world different from the world in which he lived, and he had the audacity to believe more in the reality of this new world of his thoughts than in all other authorities. Thus he progressed out of the Socratic stage to his own philosophy, and created the theory of ideas, which has been so often identified with Platonism. We cannot agree with Zeller who sees vestiges of this theory of ideas already in the Meno, Euthydemus, and Gorgias. Here we have only the germ from which the theory of ideas was afterwards developed. This germ is the consciousness of infallible knowledge arrived at when Plato wrote the Meno, becoming @ special science in the Euthydemus, and in the Gorgias entrusted with the direction of human life. This consciousness was in the New method first ap- plied to the theory of virtue led then to other subjects. Reality of the world of thought prepared in the Gorgias. Buttheory of ideas not yet expressed. Its germ is the con- sciousness of intui- tive infallible know- ledge. 218 ORIGIN AND GROWTH OF PLATO'S LOGIC beginning purely personal and based on experience in teaching. Plato enjoyed it as a new sense, a feeling of higher life, and he did not yet undertake to explain it fully. The absolute certainty was reached in his own mind, and referred really only to a few ethical truths; he had im- parted it to some of his pupils, and he generalised the faculty of absolute knowledge, postulating such knowledge for all departments of being. The complete theoretical explanation of the possibility of such knowledge was not yet given—scarcely asked for. But the consciousness of absolute knowledge, created in the soul of Plato, was transmitted from generation to generation, and since his time has never deserted European philosophy. 219 CHAPTER V ORIGIN OF THE THEORY OF IDEAS WHEN Plato had discovered in his own consciousness the existence of an infallible knowledge (a priort) and applied this knowledge to the ethical problems which were the chief subject of his teacher’s philosophy, it was natural for him to seek an explanation of the nature of knowledge itself. A priority of knowledge with its accompanying certainty appeared to him first as a psycho- logical fact, a feeling concerning certain thoughts. This feeling from a psychological point of view might still be an illusion. The logical standpoint was not yet reached, or at least 1s not known to have been reached by anybody before Plato. The fact of an @ priors know- ledge proclaimed by Plato in the Meno was for him a psychological fact, the difference between the state of mind of one who knows and knows reasons of his know- ledge, and that of one who believes, and does not care to find out why he believes. The dialectician, whom Plato had described in the EHuthydemus as the master of every knowledge, distinguished his knowledge from other people’s opinions by the circumstance, that he had reasons to quote for his judgments. The doctrine of an absolute morality was presented in the Gorgias as a knowledge above and beyond all changes of opinion ; but Plato had not yet inquired into the ultimate founda- tions of the certainty which he experienced and imparted to his pupils. The antenatal existence mentioned in the Meno was rather an inference from the fact of a priori knowledge than the explanation of it. Certainty of know- ledge first accepted as & psy- chological fact, then investi- gated as & logical problem. Not all the steps of the inquiry recorded. Three points of view ap- pearing in the Cratylus, Sym- posium, Phaedo. Cratylus presents 220 ORIGIN AND GROWTH OF PLATO'S LOGIC This explanation was the next task undertaken by Plato after giving his definitive solution of the moral problem in the Gorgias. We cannot expect Plato to record for us every step of his new investigations. We must ourselves supply the connection between one work and another, because the works themselves do not exhibit a continuity of evolution. The dialogues were not intended as a diary of investigations, but as an artistic embodiment of certain conclusions with an ideal indica- tion of a method by which they might have been reached, not necessarily coinciding with the actual steps through which the author had arrived at them. Such artistic reminiscences of a long inquiry were the Protagoras, Meno, Euthydemus, and Gorgias; they were never connected by Plato into one whole, nor are they a progressive account of the development of the author’s theories, but represent only occasional mani- festations of his original thoughts. The next movement in advance of these ethical dialogues is visible in the Cratylus and Symposium, which approach the solution of the logical problem of a priori knowledge from two different sides, which may be described as the linguistic and the esthetical. A third note is struck in the Phaedo, and it is really only in the Phaedo that the theory of ideas takes a definitive shape, and is based on meta- physical considerations. All these three dialogues are undoubtedly later than the ethical series, because their style has many more characteristics peculiar to the latest group (see above, pp. 168-169). I. The Cratylus. (Relative affinity to the latest group, measured on the Laws as unity, =0°16; see above, p. 168.) The Cratylus, which recalls the Euthydemus by the humour displayed in it, offers many difficulties to the difficulties interpreter, because it is not quite easy to distinguish ORIGIN OF THE THEORY OF IDEAS: CRATYLUS 221 what is meant seriously from what is a parody of con- temporary linguistics. Cratylus, who is here represented as debating with Socrates, might be the same about whom Aristotle! says that he was a follower of Hera- clitus and a teacher of Plato. But while Aristotle repre- sents Plato as faithful in an essential point to the doctrine of this his first teacher, we see in the present dialogue how he frees himself from a prejudice main- tained by Cratylus, according to which philology took the place of philosophy, and the truth about being was to be sought in etymology. It is very characteristic of the dialogue which makes the starting point of Plato’s logic, that in order to prove that things are not necessarily as they appear, that there is an existence independent of appearance, and a certainty not liable to doubt, Plato uses an ethical example, and quotes as one of such certainties the existence of bad and good men (386 8B). Thus the existence of things is treated as independent of the words we use to define them, and they are viewed as having their own permanence of substance (8864: &yew atta avTav twa BeBaidrnta Tijs ovcias—423 D: ovcia Soxet elvas éxdotw, Womep Kal NPO"A.. + TPATOV aVT@ TH ypwpate Kai TH povyn ori ovaia Tis éxaTépw avTar, Kal Tois ddXOLs TacLW boa HEiwTat TavTns Ths mporpnoews Tod elva). Neither is Protagoras right in affirming that everything 1s as it appears to everybody (386 c), nor Euthydemus in believing that everything is for everybody the same always (386 D), for in either case no room would be left for the distinction be- tween good and bad, and this distinction Plato since writing the Gorgias looked upon as incontestable. The opinion here ascribed to Euthydemus is found in the dialogue of 6 Aristotle in the Metaphysics (987a 32) quotes Cratylus as Plato’s teacher, and says that he was a follower of Heraclitus. Proclus in his commentary on the Cratylus of Plato (ed. J. F. Boissonade, p. 4) identifies with this Heraclitean Cratylus the Cratylus of Plato’s dialogue. of inter- pretation. Moral judgments taken as standard of certainty. Protago- ras and Euthyde- mus con- demned ; with a reference to the dialogue Euthy- demus. Substance perma- nent, while ap- pearances are changing. Perma- nence of notions a condition of know ledge. 922 ORIGIN AND GROWTH OF PLATO'S LOGIC this name, and if we compare the passages, the Cratylus seems to refer to the Huthydemus : Cratyl. 886 D: ovdé xa’ EVAu- ? “ a Snpov ye otpar wot Soxet mace mavra ¢ , ? . 2 8 HA ‘ dpotws etvat dua kai dei ovde yap ? dv ovrws elev of pév xpnoroi, of be novnoot, et dpoiws dao. xal det dpetn Te Kai Kakia e€in. Euthyd. 2948: morepov mavra viv pdvov ericracbov f Kal dei ;— kai aei—answers Euthyd. and he says: 2054: émbdetEw xai o€ ravra Ta O@avpuord éxovra. After a sophistical argument he concludes with saying to Socrates: 296pD: det yap @podrdynxas éxictar Bat Kai duatavra, This is then proved by Socrates to be wrong 297 a by the example of the evident falsehood of a judgment such as ‘good men are unjust.’ What this substance or nature of things and even of actions (387 D) is, Plato does not yet fully explain. His first step is only to ascertain that it must be per- manent, while appearance is changing. The permanence of the substance of things results from the possibility of knowledge, which, since it has been established in the Meno, is no more liable to doubt, and is here accepted as a basis of reasoning. If things never remained the same, there would be nothing in them whereof Being might be predicated (489E: mas otv av ein ti éexéivo, 3 pnddtrore DoauTws tye; ... eb b8 del dcavTws eye Kal TO auto got, Tas Ay TOvTO ye peTaBaAXKOL 7 KLvOiTO pNdeV éEtordpevoy Ths avrov ideas ;). When a thing changes it becomes another, and no longer corresponds to the idea we first conceived of it. In such continuous changes know- ledge becomes impossible, because knowledge refers to a determinate being, and if that being becomes another, then our knowledge can no more refer to it, since know- ledge cannot know an indeterminate object (4404: yvaou 8 tov ovdepla yuyvacKke 8 yiyvacKke pndapas Fyov). Knowledge itself, if it be knowledge, must remain unaltered and without change, because if it changes and no longer corresponds to the notion of knowledge, then it ceases to be knowledge at all (4404 B: GAN odds yvaow élvat davat eixds, ci petatinre mavTa ypnpata ORIGIN OF THE THEORY OF IDEAS: CRATYLUS 223 Kal pndav péver . . . 8 TovTOV TOU AOYoU OUTE TO YvwaO- pevov ovTe TO yvoacOncomevov dv cin). This reasoning is of fundamental importance for Plato’s logic, and for the origin of logic generally. It returns many times in later writings ; the existence of a knowledge that is different from mere opinion is an axiom and the foundation of science. But knowledge cannot deal with ever-changing matter. The aim is to discover fixity in its objects, and these, the notions of our mind, if grasped by real know- ledge, cannot undergo change. If they change, then they were not at first obtained by knowledge but by a wrong opinion. It is inconceivable how Schaarschmidt (pp. 262-263) could believe that the objects of knowledge referred to so frequently (as ra évra) in the Cratylus were material things. Plato says clearly that the substance of things, as being invariable, is different from material appearances, and he quotes as illustrations of such substances the knowing subject, the known object, the beautiful, the good (440 B: & d& gore péev aei TO yiyvaoxKon, zatu dé To ylyvwoKopevoy, tote 5& TO KaXOV, ott SF TO ayabov, gore b& bv ExactTov THY dYTwWY, ov pot daiverar Ttadta Gpota dvta, & viv jpeis Néyopuer, pon ovdev ovde dopa). He expressly warns his disciples that the beautiful is not the same as a beautiful face, since the beautiful face can change, while the beautiful remains always the same (439 D: avdro To Kadov TovovToy ae éoTw olov éorwv). If it did not remain the same, we could not even name it or think of it. The negative determination of the substance as different from particular things leaves open the inquiry whether this substance has an ideal or a real existence. The beautiful might be independent of our own individual reason, and might still exist only in some personal reason, being a necessary form of thought, as has been admitted by Kant. Or the beautiful might have a separate existence as a power independent of any personal Know- ledge can- not deal with ever- changing matter. Material things are not true Being. Only general notions or the knowing subject exist really. Substance deter- mined ne- gatively. No trace of sub- stantial ideas in the Cratylus. Further investi- gation invited. Allusions to earlier exposition uncertain, and could notrefer to Phaedrus or Theae- tetus. 224 ORIGIN AND GROWTH OF PLATO’S LOGIC being, the origin and cause not only of all beautiful particular things, but also of our personal notion of the beautiful. If we look at all the places in the Cratylus where the existence of an idea is postulated, we find in none of them any hint as to whether Plato in wniting this dialogue was aware of the above alternative and whether he had already made a choice between the two possible answers to the question in which manner the substance of things exists. In every passage where he uses the words eldos, iS¢a or similar expressions (as 389 D: avro éxeivo 5 Forw, 389 A: rocodrov ti 6 wépuxe) We can render them by ‘notion,’ ‘form,’ ‘idea,’ and we need not have recourse to the supposition that Plato had already imagined a world of self-existing ideas, as in his later teaching. He is very cautious in taking his first steps in logic, and he confesses that the definitive solution of these problems is very difficult (440 c), but he exhorts his readers to investigate courageously and well, and not to desist from that investigation (440 p). He seems to promise further exposition, because Socrates and Cratylus at the end of the dialogue mutually advise each other to consider the matter. This is in perfect accordance with the position of this dialogue as introductory to Plato’s special logical studies. The necessity of a substance of things, as the true object of knowledge, is ‘here alluded to as dreamt of many times (439 ©: wodAdas dvetpwrrw). Some inter- preters have inferred that this implies earlier exposi- tions of the same problem, and have accordingly placed the Cratylus after other dialogues, as for instance Pfleiderer '" held it to be ‘indubitable’ that the Phaedrus and even Theaetetus preceded the Cratylus. Use of the But we must be cautious in such inferences, because words Plato did not look upon his works as a continuous series 17 E. Pfleiderer, Socrates und Plato, Tibingen 1896, p. 818 sqq. ORIGIN OF THE THEORY OF IDEAS: CRATYLUS 225 of handbooks, in which each presupposes all that precede. An allusion to frequent discussions on a particular subject may refer much more probably to Plato’s oral teaching than to his previous works. The use of ovoia in the meaning of the true substance of a thing as opposed to its appearance is not found in the ethical dialogues preceding the Cratylus, and appears here for the first time.’ It cannot easily be taken in the later meaning of a transcendental idea, because the only marks of substance here insisted upon are its permanence, and its difference from appearance and opinion. Both can be predicated of concepts of our mind, and when Plato began to understand something else by an idea, he said so expressly in quite different terms. If anybody from the mention of the form of a shuttle (889 B: ei8os xepxidos) infers that Plato in the Cratylus admitted ideas of manufactured articles, then of course he would find the Platonic theory of ideas already in Thucydides. But in the light of an impartial interpretation, the theory of ideas is only prepared in the Cratylus, not yet formulated.'!” The power of the dialectician, assumed in the Euthydemus, 1s again asserted in the Cratylus. The dialectician, however, is here defined as ‘he who knows how to ask and to answer questions’ (890 C: 0 épwrap érictdpevos Kal amoxpivecOa); this definition is not given here as something new, but as well known and 8 Peipers (Ontologia Platonica, p. 67) quotes some passages from earlier dialogues, where according to him ovcla refers to ideas, but on con- sideration, in all these passages another meaning is obvious. Huthyph. 114 ovcla écfov=definition of holiness (Jowett: essence); Charm. 168 p obcla = nature (Jowett) or quality; Protag. 3498 ovcia (dv‘uaros) kal wpayna= object and thing (Jowett: ‘essence and thing’); Meno 728 ovala weAlrrns = definition of a bee (Jowett: nature of a bee); Gorg. 472 8 exBddrAAew ex ris ovolas (Jowett : inheritance). In none of these passages is otcla opposed to appearance, asin the Cratylus and in many later works. i” That the Cratylus is introductory to the theory of ideas has been also recognised by Susemihl (see note 54), who observed that the words eldos and i3éa bave in all passages of the Cratylus whenever they occur the meaning of ‘ species,’ ‘kind,’ ‘form,’ but not the later meaning of Platonic ideas (Genetische Entwickelung, vol. i. p. 161). Q el3os and is8éa not technical. The dia- lectician directs the creation of new words. First ele- ments of every- thing must be first ex- plained. Natural divisions of things. Origin of error in the wrong use of language: it is the privilege of the dialec- tician to use words 226 ORIGIN AND GROWTH OF PLATO'S LOGIC recognised, though it had not been given in any earlier work of Plato. In the Huthydemus, the only earlier dialogue where the dialectician is mentioned, the term was also assumed as known, and it may have been used by Socrates, as it occurs in Xenophon’s Memorahtla. But here the privilege of the dialectician to judge every kind of knowledge is extended also to the art of creating words. The maker of words has to recognise as his master the dialectician (390 D: vopodérou Epyov dvoma, értatd ny ExovTos SuaXextixov avdpa), and here Plato is clearly conscious of his dialectical superiority over contemporary philologers, and, as he expressly states, over the sophists (891 c) and poets (391 D-393 B). Related to this is the demand that the first elements of everything must be explained unless the whole is to remain unexplained; which is here applied to the origin of language (426 A: qep. Tov mpwTwv dvoudTov .. paduord te Kal KaBapwrara Sei yew arrodetEa, 4 ed eidévar, Ott ta ye VoTtepa Hon Prvapyjoe). Things have their natural divisions, according to which we must divide them if we do not wish toerr (387 A: xara riv dvaow tod réuvew te Kat TEeuvecOar xal @ wépuxe). Things are as they are, according to their own nature (886 E: xa& avira mpos THY avTav ovciay EyovTa hrep wépvxev) and not according to our imaginations (886 EB: ov mpos nas ovde ud’ jyav, EAXxopEva avw Kal KadTw TO TpEeTép@ gavtdopari), which produce error and wrong belief as opposed to truth (385 8B). Against those who pretended that error is impossible (429 p) Plato shows the origin of error in the incompetent use of language. Words are instruments (388 A: dpyavov) of thought, for educational purposes and for logical distinctions (888 c: SédacKadixdv cal Svaxpirixoy Tihs ovcias); they imitate things (430 B: Svopa piunua Tov wpdyparos) as their symbols (433 B: dyrwpa ocvrAd\aBais nal ypdpyact mpdyparos, also 435 8B), and yet are not always similar to them (432 bp), because a good word-maker is the rarest of all artisans (389 a: ORIGIN OF THE THEORY OF IDEAS: CRATYLUS 227 Snulovpyav omavwwrartos), and if he does not work after the dialectician’s directions, he may have named things not according to their nature (4325). The competent use of right words is the dialectician’s privilege (390 c) and those who do not possess the dialectical power are liable to employ words in a manner contrary to their intention, whence mistakes arise (431 8B). Thus truth differs from falsehood (885 B). The worst source of error is self-deception, because the deceiver never abandons the deceived (428 D) and makes him disagree with himself (433 B: cf. Gorg. 482 B). Here Plato confirms what he said in.the Gorgias about contradiction as the mark of error, and consistency as the condition of truth. Truth is found in the unity and similarity of things (488E: padeiy (Ta gvra)... 8: ddANAD, Et In Evyyevn soriv, Kal avta 80 ab’tav). What method should be used for ascertaining truth Plato de- clines to explain (439B: petfov tows éoriv eéyvwxévar 7 car’ éue xal oé), but he insists that knowledge is not to be gathered from words (489 B: dyamnrov 82 nal TovTo oporoynaacbar Sti ovK &E dvopatwrv, GAAA TOAD paddov avta && avutav cai paOnréov Kai (nrnréov), for the first word-maker, if he named things according to their nature, must have had a knowledge of them not gained through words (488 B). Plato thus claims for his philosophical pursuit the authority to judge about the propriety of words (425 4), to change their meaning and to make new words accord- ing to the requirements of his dialectic. He has largely used that liberty in his later works, whereas but few new words occur in the Socratic dialogues. The Cratylus pro- claims the philosopher’s independence of and power over language. Faithful to the a priort character of his knowledge, Plato despises statistics (437 D) and inferences from a majority of cases. He wants a sound basis and beginning for each theory (486D: de rept ris apyis mavTos wpdyparos trav) dvdpt Tov moAvv doyov slvat Kai Q2 in the proper way. Consist- ency & condition of truth. Philo sopher’s independ- ence of language, and power over language. Divine origin of language not ac- cepted as sufficient explana- tion. The special problem here brought forward is not decided. Only ex- tremes re- pudiated. 228 ORIGIN AND GROWTH OF PLATO'S LOGIC THY TOAAHY oKxérv) and betrays his geometrical predilec- tions by adducing the familiar analogy of a small error unnoticed in the commencement of a geometrical con- struction (436D: Siaypaypdtwv éviote Tod mpwrou cpixpod kal adnrou yevdous yevouévou, Ta AovTa TdauTrOAAa HON dvTa éTropeva Opmoroyety GAANHAOLs). He does not recognise a reference to divine origin as an explanation of anything, comparing it with the introduction of gods on the dramatic stage, when no better solution is forthcoming (425 pD), and calling it a clever evasion of the duty of giving reasons and proofs (426 4). Still, the religious spirit of the Gorgias is not extinct, and God remains free from human con- tradictions (438c), while the future life is assumed as a matter of course (408D), with the addition, that it is dominated by philosophy (404 a). It is curious, however, to see that this increasing con- fidence in the power of dialectic and philosophy seems to fail him in the concrete problems with which he is chiefly concerned in the Cratylus. The avowed purpose of the Inquiry is to ascertain the origin of language, and the dis- cussion, not invariably quite serious, of many etymologies ends in a compromise between two conflicting theories. As a result of the Cratylus we must recognise the view that there is a certain natural phonetic expression of thoughts, but that this is adulterated through the word- maker’s errors, which remain in the language by tacit consent of the people speaking any dialect. Both extreme theories of language, as the result of an agreement, or as a product of divine inspiration, are here repudiated. Plato in this dialogue employs a method very familiar to the readers of his later writings, consisting in beginning a dis- cussion with some secondary topic, and passing from this to a deeper consideration of some problem not thought of at the outset. Here the question of the origin of language is a pretext leading to the metaphysical distinction between substance and appearance, and identifying the substance of a thing with the object of true knowledge. ORIGIN OF THE THEORY OF IDEAS: CRATYLUS 229 This is a logical investigation, widely different from the simpler ethical inquiries which pervade the Socratic dialogues. The importance of the Cratylus as a first chapter in Platonic logic has not been always recognised. Plato has even been supposed to imply that consistency is no test of truth (Jowett, 1. 263). This inference is based on the passage in which Plato explains by a geometrical analogy the possibility of concealing an initial error of reasoning beneath a subsequent ‘enforced’ consistency (486D: to TpaTov chanreis o TUOéuEvos TANNA HON Trpos TOT’ EBidleTo kal Evpdwveiv nvaycavev). Such an artificial and only apparent consistency was clearly distinguished by Plato from true self-consistency, which had been proclaimed already in the Gorgias (482 B: ot cot oporoynoes Kadrxrjs, @ KadvXixrets) as a test of truth, and is again used as such a test in the Cratylus (48383: e raira audotepa épeis, ovyx olos + gaer cuppwvety cavt@). The familiar example of a wrong consistency was adduced only in order to show the decisive importance of the first principles in every science (436 D). The ideal consistency required by philosophy is not expected by Plato to be found in a language (435 c), though he affirmed that language to be the most beautiful in which the greatest consistency reigned (485D). To build such an ideal language by creating a philosophical terminology was a task which Plato subsequently undertook in part, but which he almost ridiculed when he wrote the Cratylus (433 E; cf. Polit. 2618: pH crovdaley dri rots ovopact). The Cratylus, a literary masterpiece comparable in its originality to the Parmenides, was held by the successors of Plato in an esteem attested by the commentary of Proklos,'®° and has up to the present day exercised the perspicacity of numerous commentators, as can be seen 1” Ke Procli scholiis in Cratylum Platonis excerpta e. codd. edtt. J. F. Boissonade, Lipsiae 1820. Logical consist- ency ex- plained by a geo- metrical analogy. Only arti- ficial and wrong consist- ency is not a test of truth. Ideal con- sistency not found in lan- guage. The Cratylus has been esteemed by many commen- tators, and its authen- ticity certain. Etymo- logies quoted to a great extent justified by the state of linguistic knowledge 230 ORIGIN AND GROWTH OF PLATO'S LOGIC from the writings of Dittrich,'*' Benfey,'*? Hayduck,'* Rosenstock,'* Heath,'!® P. Meyer,!*® and Bonitz'* on this dialogue. What Schaarschmidt (p. 245 sqq.) said against the authenticity of the Cratylus has been suff- ciently refuted by Alberti,'** Lehrs,'® Luckow,'® Drey- korn,'®' and H. Schmidt,!*? so that even Huit (ii. p. 187), who popularised in France Schaarschmidt’s doubts as to many other dialogues, thought it advisable to dissent in this respect from his master, and to defend the authenticity of the Cratylus. One of the grounds alleged by Schaarschmidt, the apparent absurdity of the etymologies proposed, has been explained by Schaublin,! who compared these etymo- logies with other evidence about the knowledge of etymology accessible to Plato, and found that among 120 etymologies attempted by Plato over sixty were perfectly justified according to the knowledge of his times, and twenty stand even the test of our present knowledge of Greek. Schiaublin has also carefully compared the 181 Ki. M. Dittrich, De Cratylo Platonis, Berolini 1841. 192 T. Benfey, Ueber die Aufgabe des platonischen Dialogs Cratylus, Gittingen 1866. 188 W. Hayduck, De Cratyli Platonici fine et consilio, Vratislaviae 1868. 18 P, E. Rosenstock, Platos Cratylus und die Sprachphilosophie der Neuzett, Strassburg 1893. 188 D. Heath, ‘On Plato’s Cratylus,’ in the Journ. of Philol. for 1888, vol. xvii. p. 192. 186 P, Meyer, Quaestiones Platonicae, Leipzig 1889, pp. 12-25. 187 Bonitz, ‘ Ueber Platos Cratylus,’ Monatsber. Berliner Akadem. 1869, p- 703. 188 Alberti, ‘Ist der dem Plato zugeschriebene Dialog Cratylus acht?’ in Rheiwnisches Museum, vol. xxi. p. 180 sgg., and vol. xxii. p. 477 sqq. 1866-67. 16 Lehra in Rheinisches Museum, vol. xxii. p. 436, 1867. 10 R. Luckow, De Platonis Cratylo, Treptow 1868. 1 Dreykorn, Der Kratylus ein Dialog Platos, Zweibriicken 1869. 18 H. Schmidt, Platos Kratylus im Zusammenhange dargestellt, Halle 1869, an excellent commentary. 198 F, Schaublin, Ueber den platonischen Dialog Kratylos, Basel 1891. The same subject had been treated very differently by C. Lenormant (Commentaire sur le Cratyle ‘de Platon, 316 pp., Athénes 1861), in his voluminous edition and commentary. ORIGIN OF THE THEORY OF IDEAS: CRATYLUS 231 etymologies given in the Cratylus with other etymologies occasionally indicated by Plato, and he demonstrates their similarity and good faith against Steinthal!®* who be- lieved all the etymologies given in the Cratylus to be arbitrary. Competent writers disagree widely as to the date of the Cratylus. Even C. Ritter, notwithstanding his stylistic observations, believed it possible for the Cratylus to have been written before the death of Socrates, as has been thought also by the poet Gray,'® by Socher, Stallbaum, and others. This opinion is opposed by those who believe the Cratylus to be later than the Phaedo and Phaedrus, as for example by Ast, and in recent times by Peipers and Bergk. The style would (see above, p. 168) place this dialogue clearly between the Gorgias and Phaedo, and the logical contents also point to the same result, the Cratylus being intro- ductory to the logical theories of the Phaedo, while pre- supposing the conclusion of ethical inquiries summed up in the Gorgias. Some other hints confirm the position here given to the Cratylus as the first logical work sub- sequent to the complete series of ethical dialogues. Diimmler '* observes that the allusion (433 a) to the early closing of the gates in Aegina presupposes a time of peace in which Athenians and more especially Plato's students could make excursions to the neighbouring town. But such a time of peaceable intercourse between Athens and Aegina was not possible, so Dimmler thinks, before the peace of Antalcidas, or 387 B.c. The Cratylus must then have been written later, after Plato’s return from his 1% Steinthal, Geschichte der Sprachwissenschaft bei den Griechen and Rémern, Berlin 1862. 1% Thomas Gray, Notes on Plato, in vol. iv. pp. 67-3888 of the Works, edited by E. Gosse, London 1884 (first published 1814), p. 164, calls the Cratylus ‘ the least considerable’ of the works of Plato. 16 Diimmler, Chronologische Beitrdge su einigen platonischen Dialogen aus den Reden des Isokrates, Basel 1890, p. 48 ; Christ, Platonische Studten, p. 8, made it probable that Plato had money transactions in Aegina. in Plato’s times The date of the Cratylus has been differently deter- mined. Mention of Aegina gives & useful in- dication. Hellenes and foreign nations spoken of as equal, as in later dialogues. This seems to imply that the Cratylus was written after Plato’s voyages, perhaps at the begin- ning of his teaching activity. 232 ORIGIN AND GROWTH OF PLATO'S LOGIC first voyage to Sicily, and also after his captivity in Aegina,—if the story of this captivity 1s true. Another confirmation of this view is given by the 1m- partiality with which Plato treats foreign nations in the Cratylus as equal to the Greeks (383 A: dpGornra ovo- patroy kal “Endnote xai BapBdpots thy avrny dracw.... 90 A: rov vopobérny tov tre évOdde Kai Tov av ois BapBapos .. .. 425 EB: sick 88 av dpyaworepor BdpBapo., see also 385 E, 390 c, 409 Z). This concep- tion remains unchanged in many later works, as the Symposium, Phaedo, Republic, Theaetetus, Politicus, Timaeus, while it is opposed to the narrow Greek and even Athenian patriotism, shown in the Protagoras, in which Athens is called the seat of wisdom (Prot. 337 D: mpv- tavetoy THs codias) by the non-Athenian Hippias. In the Gorgias Athens is praised as the place in Hellas where the greatest freedom of speech is to be found (461 E : ’A@nvate adixomevos, ov THs ‘EAXdOos tAsictyn gotiv s€ovcia Tov Afyevv), without any mention of foreign countries, such as occurs repeatedly in the Cratylus, whenever the whole of Greece or the Greeks are named. This frequent mention of foreign nations in the Cratylus seems to belong to a time when the horizon of Plato’s experience had been considerably enlarged by his travels abroad, while the subject of the origin of Greek language, generalised into the inquiry about the origin of human speech and the relation of thought to it, would seem to have been specially debated in Plato’s school. The moral problems discussed in the preceding dialogues were inherited from Socrates, though their solution in the Gorgias is already Platonic: the problem of language as a source of know- ledge has been attributed to Antisthenes,’” and the peculiar proof that philosophic truth is independent of 17 The very uncertain allusions of the Cratylus to this philosopher are treated by Diimmler, Akademika, pp. 148-161; K. Barlen, Antisthenes und Plato, Progr. Neuwied 1881; K. Urban, Ueber die Erwadhnungen der Philosophie des Antisthenes in den platonischen Schriften, Kinigsberg 1882 ORIGIN OF THE THEORY OF IDEAS: SYMPOSIUM 233 language, contained in the Cratylus, 1s a worthy inaugu- ration of Plato’s own philosophical career, in which he was distinguished from all predecessors by his power over language as an external instrument for conveying thought. Plato, the great word-maker, could not better begin his new philosophy than by this inquiry into the relation between thought and speech. The counterpart of this, the inauguration of Plato’s logic, is to be found in the Symposium, where the philosopher was led to a new vision of truth as consisting in eternal and self-existent, independent ideas, those Platonic ‘ideas’ which have been accepted by so many readers as the quintessence of Platonism. II. The Symposium. (Relative affinity to the latest group, measured on the Laws as unity, = 0°14; see above, p. 169.) Nearly every other work of Plato admitted of discus- sion as to the author’s purpose and the chief contents. The Symposium, however, is distinguished by a clear announcement of its aim, and deals apparently only with one subject, love, teaching the first lesson of that new feeling discovered by Plato and in its first stage known even to-day as Platonic by some people who know nothing else of Plato. It would appear that in this lesson of love no room could be left for logic. But Plato, who is at once a great poet and a great logician, initiates us into the mystery of his first logical discovery through this triumphant poem of victorious love. It is love, he says, that leads to the highest knowledge of truth. But not the love of a single person, however pure, nor the love of a single city, be it the greatest on earth, nor the love of a single science. There is far above all these feelings a new and powerful love, difficult to under- stand even for Socrates, who has heretofore been repre- sented as the wisest of men. The explanation of this feeling, expressed by nobody before Plato, he puts Chief sub- ject of the Sympo- sium love, but a new kind of love, leading to know- ledge. For the first time Socrates is sup- planted by another teacher, but not a historical person. Diotima unknown to Thucy- dides : probably invented by Plato in order to give apparent historical authority to his own teaching. theory 234 ORIGIN AND GROWTH OF PLATO'S LOGIC poetically in the mouth of a woman. This woman, Diotima of Mantinea, is invented by Plato, though he gives her an historic appearance by the assertion that through her prayers she preserved the Athenians from the plague. If she had been, as Plato makes his readers believe, a well-known and inspired priestess, Thucydides could not have failed to mention her. But no Greek writer!®® before Plato knows anything about a Diotima of Mantinea, and all later mentions of her are based on the Symposium. We may therefore assume that the new theory, here ascribed to Diotima, is Plato’s own invention. For the first time in all Plato’s dialogues, Socrates ceases to be the sole teacher of wisdom, and Plato unmistakably implies that his new wisdom may be above the under- standing even of his teacher (210 A: radra pév ody ra épwrixa lows, & Lwxpates, Kav ov punOeins ra dz TérEa Kal drromruxd, dv &vexa Kal Tadta eoriv, dy Tis opOds pein, ovx old 2 olos r av eins... . wee érecOas, adv olos re 7s). He clearly hesitates to expose the treasure found in solitary meditation to the unprepared adherents of vulgar love. He apologises repeatedly for the admitted obscurity of his teaching (201 D: wetpdoopac SuedOciv ... , Srrws dv Svvapar. 204 D: meipdoopar Siddka . . . cadtotepov épw (also 206 c) . . . 206 B: pavreias Seirar &6 ri more réyes .. . 207 C: ph Oavpate (also 208 B) 210 A: ép@ pav ody Kai mpoOumias ovdév arroreinro . . . 210 BE: wea 54 pou Tov vodv mpocdyew ws olov Te padtora). It is evident that the new-found knowledge is looked upon as far more important than anything which has been said in earlier dialogues. Like a precious gem, it is set in the poetical gold of the Sympostum—the most consummate work of art which even Plato’s genius has produced. If we ask wherein consists the new logical knowledge #8 The unhistorical character of Diotima was made evident by Her- mann, De Socratis magistris et disciplina juvenili, Marburg 1837, p. 12 sqq. ORIGIN OF THE THEORY OF IDEAS: SYMPOSIUM 2356 immortalised by the Symposium, we see it condensed in a few pages of the highest eloquence, which may be read as a record of personal experience, and as the result of the long previous development of Greek art. Lévéque'™ in France and Cohen” in Germany have noticed the near relation between the origin of Plato’s theory of ideas and this preceding growth of Greek art. What Plato says about his discovery amounts to this: if somebody grows accustomed to generalisations and to the progress from particulars to general notions, then, at a certain moment of his life, he will become suddenly (210 HE: 2&aidvns) aware of the existence of the general idea as something which does not depend upon particulars, but is the true origin of all particular qualities. This sudden vision, here pictured with the natural delight of a first discovery, is the aim of all intellectual development (211 4a: todro éxeivo ov 87 Evexa nal of EumpocOev mravtTes movot joay), a marvellous beauty (210 BE: Qavyacroy tiv dvow xandov) leading to every kind of virtue and to the immortality of man (212 A: rexovre dpernv adyOH Kal Opevrapev@ trdpye Ozopirget yevéoOar, cal eirep TH AAW avOpwrrav, dbavary Kai éxeive). What kind of existence the idea of the beautiful possesses is difficult to express in human language, according to Plato’s own confession. But this existence was clearly meant by Plato, when he wrote the Sym- postum, to be a solution of the problem of substance proposed in the Cratylus. In that dialogue he limited his indications as to the substance of things to a few '*® Carolus Lévéque, Quid Phidiae Plato debuerit, Parisiis 1852, p. 60 : ‘Quaecumque Plato de pulchritudine scripsit . . . haec in Phidiae deorum vultu expressa et ut ita dicam sculpta invenerit, ita tamen ut ad intelli- gendum penitus Phidiae ingenium ingenio Platonis opus fuerit. Ab illo qua via ad summae pulchritudinis ideam perveniatur didicit.’ 20 Hermann Cohen, ‘ Die platonische Ideenlehre, psychologisch ent- wickelt,’ in vol. iv. pp. 403-464 of Zeitschrift fur Voilkerpsychologie und Sprachwissenschaft, Berlin 1866, p. 418: ‘Platos That wie sein Geist wiachst hervor aus dem gemeinsamen Samen der hellenischen Weltarbeit.’ of ideas based on the pre- ceding growth of Greek art. Theidea as the cause of parti- culars. Existence of ideas difficult to explain. It is inde- pendent of opinions and ap- pearances. The idea is more perfect than a work of art. Intuition of ideas is acquired by exer- cise in generali- sations. 236 ORIGIN AND GROWTH OF PLATO'S LOGIC determinations, such as permanence and objectivity. Now he has ‘suddenly’ perceived a beauty not only eternal (211 A: del dv xal obre yuyvomevov ove dtrodAv- pevoyv, ovTe avEavopuevov ovtTe POivov) and objective, but also absolute, that is, independent of time and space, and of concrete appearances as well as individual opinions (ov TH pav Kadov, TH 8’ aicypov, ovde ToT? pév, ToTe dé od, ovdé Tpos pev TO Kadov, pos b2 Td aicypor, OVS’ EvOa péEv xanrov, vOa 82 aicypov). It is natural that Plato, being himself an artist and living in an age when art had reached an ideal perfection, should formulate this first assertion of a self-existent absolute idea with reference to the idea of beauty. He saw the distance between all human models and such a creation of art as the Olympian Zeus of Phidias. He imagined that even the most perfect work of art is only a particular instance of the ideal beauty, which he did not claim to perceive with the mortal eye, but with the divine insight of an en- thusiastic soul. He recommends his readers to acquire this superior faculty of intellectual intuition by exercise in generalisation. He says clearly that the idea is not only immaterial (211 4: odd’ ad davtacOyjcetar avt@ 16 Kanrov olov mpoowmov te ovde yetpes ovb2 aAXO OvdsY aw chpa peréyet) but not even intellectual (2114: ovdé tis Aovyos, ovdé Tis eriaotHun, ovdé Trov by ev érépw Tit, olov év Com } dv yh 7 &v ovpave 4} tv Tm add), Nor inherent in the soul as a notion, nor in anything else. Here we have an evident indication that Cohen’s™! doubts as to the separate existence of Platonic ideas, however justified with reference to other works, are inadmissible so far as concerns the Symposium, and the idea of Beauty, the first discovered by Plato and the only idea spoken of in 2! H. Cohen, Platons Ideenlehre und die Mathematik, Marburg 1879, p. 9: ‘Diese Auffassung des xwpiocuds ist einmal des Aristoteles eigenste verantwortliche That. Ob wir sie hatten, ob Jemand aus den Platonischen Dialogen sie herausgelesen haben wiirde, wenn Aristoteles sie nicht als die legitime gelehrt und—unerschrocken verhdhnt hatte, das darf wenigstens bezweifelt werden.’ ORIGIN OF THE THEORY OF IDEAS: SYMPOSIUM 237 the Symposium. This idea is certainly not immanent, but separated from concrete things. The relation of single beautiful things to the idea of beauty is expressed here by the word peréyew, not used in any earlier dia- logue to express the relation of a particular thing”? to a general notion. It means that all beautiful things owe their beauty to the idea of Beauty. This idea is not here called eldos or (d¢a, but 1s named ‘the beautiful’ (211 8° ro xadov). It1s self-existent, needs nothing else to enable it to exist eternally (211B: avro xa6’ avto pel’ avrod povoetoes aet dv), and Plato has invented the term povo- evdés, first used in the Symposium, to mark its simplicity. According to modern terminology Platonic Beauty is then a simple substance, the original cause of all in- dividual beauty, suffering no alteration through its action on the particular things, to which it imparts its own quality, though in a lesser degree. Plato admits this ideal Beauty to be an object of sclence and knowledge (211cC: pd@npa), but he leaves it uncertain whether he pretends to know it as it really is, or only as it appears to him. When, however, he 22 This term in Symp. 2118 ie paraphrased rather than translated by Jowett in the words: ‘Beauty absolute ... which ... is imparted to the ever growing and perishing beauties of all other things.’ seréyew meaning the relation of things to ideas is used for the first time in the Symposium, while in earlier works it had the current meaning with which it is found in other authors, translated (Jowett) by ‘share in’ (Prot. 322 p, 323 a, virtue and other arts), ‘take part in’ (conversation, Huthyd. 271 3, danger, 279 &), ‘are intermediate between’ (philosophy and politics, Huthyd. 306 a 8), ‘is proficient’ (in an art, Gorg. 448 c), ‘ partake’ (of good and evil, Gorg. 467 2). The technical meaning of neréxew as designating the relation of things to ideas is limited almost entirely to the Symposiwm and Phaedo (100 c, 101 c, cf. Rep. v. 476 p), while in the Parmenides (where the abstract noun “éGef:s also occurs) it is mentioned and criticised. In other dialogues peréxew is used in the ordinary meaning (as for instance Rep. 482 8, 455 p, 465 5, 520 s, &c.; Phaedr. 247 3, 2498, 2720; Phil. llc, 548, 560; Tim. 27 c,58c, 582; Legg. 721 p, 755 a, 963 8, &c.). The peculiar use of perdxew in the Sophist (as 251 ©) to mark the relation between two general notions is quite different from the meaning of a participation of things in ideas. An alternative term for pueréxew is peradauBdverw. Cf, Jowett and Campbell, Rep. vol. ii. p. 809. The idea of Beauty is the source of all beautiful things, exists eternally, being sub- ject to no changes in its sim- plicity. Ideas as objects of knowledge appa- rently identical with our subjective notions. Know- ledge remains right opinion based on grounds or sufficient reason, in the Sym- posium like in the Meno. Distine- tion between Wisdom and Philo- sophy common to 238 ORIGIN AND GROWTH OF PLATO’S LOGIC speaks of exercise, as enabling us to improve our sight of absolute Beauty (211B: drav ... eave éxeivo To Karov dpyntrar xabopav, oyeddv dv Tt Grotto Tov TéXovs), we must infer that he allowed the possibility of an immediate intuition of absolute Beauty as it is, without subjective error. He did not yet see the peculiar diffi- culties of such a position. Though Plato in the Symposium thus presents a new object of knowledge, he seems not to have progressed as to the definition of knowledge itself beyond the dis- tinction given in the Meno, according to which knowledge differs from right opinion by the reasons which we are bound to give when we know something. Here he recalls this distinction : Meno 98a: ScfacarnOeis ... od woAdod aftai elow, ews dy ris avras dnon airias Noyope. éredav 8¢ debaoww, mparov peyv érurrnpat ylyvorrat, éreira povipor*. . . . Kal &a ratra 31 riyptorepoy éemornun épOns ddéns oriv, cal diapdpe Symposwwm 2024: 1rd dp6a do€dlew dvev rov Exe Adyor Sovvat, otre énicracbai éorw: ddoyoy yap mpaypa was dy ein émornun; ovre duadia: ro yap tov Gyros ruyxa~ voy mas dy etn dyabia; forw 8 87 mov rowovroy 7 opby Sofa, perate dpomcews Kat dpaias. If right opinion without reasons is not knowledge, yet knowledge might still be for a modern logician something else than right opinion with reasons for it, but if Plato had changed his view of the nature of know- ledge expressed in the Meno, he could not conceal it here, because every unprejudiced reader infers that knowledge, not being right opinion without reasons, is right opinion based on reasons, a8 had been stated expressly in the Meno, and denied only much later in the Theaetetus. A fresh point is gained in the distinction between wisdom and philosophy, which is repeated later in the Phaedrus, and here founded on the etymology of the name ‘philosopher,’ as one who desires wisdom and therefore does not yet possess it. It is noteworthy that even in the etymologies of the Cratylus Plato did not allude to this new meaning of ‘philosophy,’ which is ORIGIN OF THE THEORY OF IDEAS: SYMPOSIUM 239 first explained in the Symposium (203E: Oedv ovédeis dirocogel odd’ errvOupst coos yevéoBar* sore yap’ ovd’ e! Tis GAXos codos, ov girocode?). This exaltation of a wisdom above philosophy, which in the Euthydemus and Gorgias was still the highest science, corresponds to the new power of intuition of Beauty, which is placed above all other knowledge. Plato became conscious of the limitations of that purely ethical knowledge of which he was so proud in the Gorgias. He felt an artistic long- ing for a perfection beyond pure logical investigation and reasoned knowledge, even beyond knowledge based on full consciousness of all reasons. He was thus led to this almost unthinkable conception of absolute Beauty. Another consequence of the new idealism is the change of position as to personal immortality. It is not clearly denied, at least for the philosopher (2124), but the religious faith as laid down in the Gorgias has been converted into a pantheistic view according to which immortality consists in the eternal reproduction of the same ideal form (208A: rovtTw T@® Tpom@m way To Ovnrov opterat, ov TH TavTdtTacw TO avToy aél elvat WoTEp TO Osiov, GANA TH TO amiov Kal TadatovpEvoy ETEpov véov éyxatanéitrew olov avro hv, cf. Legg. 721 c). This renovation of particulars is applied even to knowledge (208 A: zrodv 68 atomwrepoy ert, OTL Kal at érre- oThpat py OTe al pay yiyvovrat, at b8 admoAAvyTaL Hyiy, Kal ovdérote of avTol éopev ovd8 KaTa Tas émioTiuas, GANG Kat pia éxdotn Tay émictTnpay TavTov tacxel). Exercise keeps knowledge apparently the same, yet constantly renewed, and creates new knowledge which seems to be the same as that which we had before (2084: perérn wdadw Kawnhy éumoovca avti THs atiovons cater THv briotnunv, Bote THY auTHnv Soxety elvac). This surprises Plato himself more than the exchange of elements in the body, and it seems to contradict the identity of knowledge admitted in the Gorgias. But the contradiction is only apparent, as the identity referred to the objective know- Sympo- sium and Phaedrus. Different view of immor- tality in Gorgias and Sym- posium. Peculiar view of in- tellectual exercise compared with reno- vation of matter. Literary merits of the Sym- posium. Date of the Sym- posium 885 3.c., almost 240 ORIGIN AND GROWTH OF PLATO’S LOGIC ledge, and the successive substitutions are attributed to the individual. It was a consequence of the growing admiration of Plato for knowledge, that at this stage the subject disappeared as compared with the object, which became the only true reality. Thus was founded the system of idealism, known as the Platonic theory of ideas. In the Symposium it appears as a first attempt and is limited to the idea of Beauty. This logical importance of the Symposium has been little noticed up to the present time, being overshadowed by its literary perfection. Such poets as Racine ™ and Shelley ** have attempted to render it in modern lan- guage, and many editors and commentators have spent their leisure on the text.” There is an almost general agreement as to the date of the Symposium, the mention of the recent partition of Mantinea, which occurred 385 B.c., being admitted as a sufficient indication that the dialogue cannot have been written much later. This conclusion was suffi- 2» Te Banquet de Platon, trad. par J. Racine, M™¢ de Rochechouart et Victor Cousin, Paris 1868; also in Guvres de J. Racine, ed. L. Aimé Martin, Paris 1844, vol. v. pp. 95-186. Racine’s translation extends only up to the speech of Eryximachus. #4 Percy Bysshe Shelley, The Banquet of Plato, London 1887 (first ed. 1840). Shelley held the Symposium to be ‘ the most beautiful and perfect’ among the works of Plato. 28 Besides modern editions of F. A. Wolf (Lipsiae 1782, also 1828), Ast (Landshut 1809), P. A. Reynders (Groningae 1825), L. J. Riickert (Lipsiae 1829), A. Hommel (Lipsiae 1834), de Sinner (Paris 1834), Jahn (Bonn 1864, re-edited by Usener, Bonn 1875), C. Badham (London 1866), G. F. Rettig (Halle 1875-76), it is worth noticing that the Symposium (ed. Salamanca 1558) was the first Greek publication of the famous Salamanca University Press. An extensive commentary on the Symposium was written already by the second French translator Loys Le Roy (Le Sympose de Platon, Paris 1559), who omitted the discourse of Alcibiades as too indecent for his French readers of the sixteenth century! The first trans- lation was Le Banquet de Platon, trad. par M. Heret, Paris 1556, a beauti- ful specimen of typography. *¢ However, Plato sometimes refers with a veworl to events over twenty years old, as for instance in the Gorg.508 o the death of Pericles is called recent (vewor!), while from Gorg. 4788 it results, that the conversation between Gorgias and Socrates is assumed to have taken place 405 5.c. or twenty-four ORIGIN OF THE THEORY OF IDEAS: SYMPOSIUM 241 ciently established in the last century by F. A. Wolf and has been successfully defended *” against some attempts at another interpretation. The mention of this event comes out so naturally that 1t cannot be regarded as a later interpolation added by the author or by his copyists. But it would still leave it open whether the Symposium was written in the same year or some years later, because for any contemporary reader an historical fact which occurred four or five years ago is still quite recent. Other considerations, however, make even the year 385 B.c. seem a late date for the Symposiwm, so that there is no probability in favour of a later time. The chief reason which makes it improbable that Plato could have written the Symposiwm much after 385 B.c. is the great number of works which, as our further inquiry will show, are later than the Symposium, and which also must be earlier than the change characterising the latest stage of Plato’s authorship. On the other hand, the number of works which precede the Symposium is very small for the space of fifteen years since the death of Socrates. Admitting the Euthydemus to have been written about 390 B.c., as has been made very probable by Spengel, Teichmuller, Sudhaus, and Dummler, we have for the five following years only the Gorgias and the Cratylus, which is not much for a gifted author about the age of forty and at the height of his literary power. This years after the death of Pericles. But in referring to a time so far back Plato is careless of the exact dates. 27 Besides Wolf in his edition of the Symposium (1782), also J. Spiller (De temporibus Convivii Platonici, Glivitti 1841), Ueberweg (Untersuch. p. 219), Teichmiiller (ii. p. 262), L. v. Sybel (Platons Symposion, Marburg 1888), Kassai (Meletemata Platonica, p. 859, Budapest 1886), have shown that the Symposium must have been written about 385 s.c. 28 A. Hommel, in his edition of this dialogue, tried to get rid of the anachronism by an emendation of the text. Diimmler believes that the reference to the partition of Mantinea might have been made also about 371, when the reunion of the separated parts of Mantinea was intended. Re- cently U. v. Wilamowitz-Moellendorff (Hermes, vol. xxxi. p. 102) suggested the dissolution of the Arcadic Union in 418 3.c. as the event alluded to by Aristophanes. R unani- mously accepted on account of a well- known anachron- ism. Great number of works later than the Sym- posiun. Small number of larger works pre- ceding it. The Sym- posium as an acade- mical pro- gramme. Position inter- mediate between Cratylus and Phaedo. First introduc- tion of absolute Beauty or the idea of Beauty. 242 ORIGIN AND GROWTH OF PLATO'S LOGIC difficulty is avoided by those who place before the Sym- postum such dialogues as the Phaedo and Phaedrus,™ not to speak of the dialectical works. But our subsequent exposition will prove beyond all doubt that these works must have been written after the Symposium. Besides, the Symposiwm, according to the very plausible reason- ing of Sybel and also of Teichmiiller, bears the character of having been written under the fresh impression of the successful beginning of Plato’s Academy, which was probably founded in 387 B.c. Various other indications confirm the intermediate position of the Symposium between the Cratylus and Phaedo, after the Gorgias and the other Socratic dialogues. In the Cratylus, Plato did not advance beyond a general distinction between substance and appearance, without any close determination of substance. In the Symposium this determination is given in regard to the substance of Beauty in a manner which makes it very probable that Plato is for the first time announcing his discovery of absolute being. In all earlier dialogues Socratic notions were ‘ present’ in the things, or immanent (Charm. 159 4 : wdpeors cwhpoovvn, Lys. 217D : NevKorns, 217: wapovoia ayabod, Euthyd. 801 4: «ddXos, Gorg. 497 B, 498 D: ayaba@v mapovcia) ; in the Symposium the higher doctrine of a participation of particulars in the idea is taught. The doctrine of pre- existence, which had been formulated in the Meno, is If C. Huit (Hiudes sur le Banquet de Platon, Paris 1889) believes that all competent writers agree in placing the Phaedrus before the Symposium, except Ritter and Teichmiiller, he betrays his ignorance of many authors, as Suckow, Munk, Thompson, Campbell, Blass, Dittenberger, Schanz, Droste, Kugler, Gomperz, Lina, Tiemann, who all agree in placing the Phaedrus after the Symposium. I¢ is true that against these fourteen authors, who up to the time of Huit’s strange assertion held the Symposium to be earlier than the Phaedrus, many others, as for instance, Schleiermacher, Stallbaum, Steinhart, Susemihl, Ueberweg, Liebhold, Teuffel, Peipers, Windelband, Christ, Zeller, were of the contrary opinion. But majorities cannot decide such questions, and since 1889 the proportion is reversed, so that the majority of new investigators take the later date of the Phaedrus for granted. ORIGIN OF THE THEORY OF IDEAS: SYMPOSIUM 243 here only alluded to casually in the discourse of Aristo- phanes.?!° The rule laid down in the Protagoras (847) to exclude flute girls and similar artists from philosophical banquets is repeated in the Symposium (1768), with the recommendation to find the best entertainment in con- versation (Prot. 347c: 8a To py SdvacGas adrXndOLsS dv gavTav ovveivat... dia THs sgavTaY horas Kal Tov Noyov Tov éavT@v bo arraidevoias, tiplas Tmotovor Tas avrntpidas: cf. Symp. 176 E: elonyotpar ryv pav apte eloeAOovcav avrAntpida yaipew dav . . . Has 52 Sid royov adrAjAots ouvelvar). This appears, if we compare the passages, to be said in the Symposium as a matter of course, while it is explained at length in the Protagoras. Some other references to earlier dialogues are of the same kind : Gorg. 490£: Socrates says :— Symp. 221: ef eOédAoe tis Tov rév oxurordpov tows peyora det brodhpara cal wAciora Urodedenévoy mepirareiv, to which Kallikles answers : @Avapeis . . . and ael raira eyes,—491 A: ov oxvro- rdépous eyo... Crat. 888D : dp’ ovv mas xad- Kevs ROTH réxmy Exar, also 889 E dmas xadkevs. Huthyd. 278 v: pn pov Kxarayedare. Gorg. 478 E: Socrates complains of Polos, who is represented as dmetpos kal avdnres : @do av rovro eidos ekéyxou... xarayekav. Gorg. 512D: xara- yéAaorés co & Wdyos yiyverac refers to 484 p: (pidcogor) xara- yéAaorot yiyvorrat. S@xpdrovs dxoveryv Adyar, daveiev av yedoio: rd mpGrov . . . dvous yap ... Ayer nal yaAneas riwds kaiaxurordpovs... xaldet da TOY avtav ra avrd paiverat déyew, Sore dmetpos Kat dvdnros dvOpw- mos mwas av trav Aéyov Karaye- Adoetev. 198 ©: xat ydp pe Topyiou 6 Adyos Gvenipvnoxev ... ehoBovpny un pos teevrav 6 ‘Ayd0wv Topyiov xedha- Anv Gewod A€yeww ev TH Ady emi rov éudv Adyoy weuwas avrov pe Aidov ry abwvia romoeev’ nar €vevénoa tore Gpa xartayé\aotos oy. Gorg. 4568: éreoa, ovx DAg réxvy f tH pytopiKy. 20 This seems to have been overlooked by Grote (vol. iii. p. 17) when he says that in the Symposium no such doctrine is found. It is important to notice this, because the entire absence of the pre-existence theory in the Symposium might lead to wrong chronological conclusions, at least as to the date of the Meno. R 2 Refer- ences to some earlier dialogues. The vul- garity of examples usually quoted by Socrates defended. Hellenes and Bar- barians. Relation to Isocrates’ Busiris made pro- bable by Teich- miiller ; this con- firms our conclu- sions. Relation to earlier dialogues generally admitted. cannot be earlier than the Sym- posium. 244 ORIGIN AND GROWTH OF PLATO'S LOGIC It seems as if the examples chosen in the Gorgias and Cratylus had provoked some critics, whom Plato answers in the Symposium, though the description given by Alcibiades corresponds also to the historical Socrates as represented by Xenophon. The mention of Hellenes and Barbarians (209 £) as equal to each other also places the Symposium above the Gorgias, and on a level with the Cratylus. Teichmiiller (I. p. 120) made it very probable that the Symposium must be later than Isocrates’ Busiris, in which we read (222c) that nobody except Polycrates had ever asserted that Alcibiades had been a disciple of Socrates. This could not be said by Isocrates if he knew Plato’s Symposium, in which the near relation and friendship between Alcibiades and Socrates is clearly represented. Teichmiiller infers that Plato in introducing Alcibiades answered Isocrates’ pretension to place Alcibiades above Socrates, and at the same time defended Alcibiades against the calumnious attacks of Lysias. The Busiris was written, according to Blass, some years after 391, and this would well agree with the admitted date of the Symposium, 385 B.c. We need no further evidence as to the priority of the Cratylus, Gorgias, and all Socratic dialogues, because these have generally been admitted to be earlier than the Symposium. The proof that some other dialogues, as the Phaedo and Phaedrus, which were also held by many critics to be earlier than the Symposium, are later, will be given when we come to deal with the date of each of them. For the present we may admit as certain, that the Symposiwm was written about 385 B.c., and after the Cratylus, Gorgias, Euthydemus, Meno, Protagoras, and all smaller dialogues. This result is not new; it is one of the few points of general agreement among writers on Plato. The comparison of logical contents has confirmed it, and also the style of the Symposium (see above, p. 169) is clearly intermediate between Gorgias and Phaedo. ORIGIN OF THE THEORY OF IDEAS: PHAEDO 245 III. The Phaedo. (Relative affinity to the latest group, measured on the Laws as unity, = 0°21; see above, p. 170.) The Phaedo is less artistically simple than the Sym- postum; it contains many threads of argument united with such skill that there is room for various opinions as to the chief purpose of the author and the main subject of his work. The dialogue has been regarded as an historical account of the death of Socrates,?!! as a treatise on the immortality of the soul,” as the poetical tragedy announced at the end of the Symposiwm,?" as a general psychology,” as an ideal picture of the true philosopher,?" and even as a treatise on the underground rivers.26 There is some truth in all these assumptions if not taken absolutely ; but for our present purpose the Phaedo deserves particular attention as containing the theoretical substantiation of Plato’s first logical theory. We have seen in earlier works many allusions to logical problems discussed by Plato with his pupils. In the Cratylus the subsidiary problem of the relation between thought and language led to the hypothesis of an existent substance of things, different not only from all appear- ances, but also from all possible expression in human 211 This exceedingly improbable opinion has been sustained in recent times by T. Bergk (Griechische Literaturgeschichte, 4°" Bd. Berlin 1887). 213 This is the ordinary view, represented in our century especially by Steinhart. 213 The well-known passage, Symp. 228 p, has been interpreted as refer- ring to the Symposium as comedy, and to the Phaedo as tragedy. 4 Plutarch (Moral. 120 ) quotes the Phaedo by the title rep) yuxiis, which appears also in the manuscripts. 18 Schleiermacher believed the Phaedo and Symposium to be the con- tinuation of the Politicus, and to constitute between them the definition of the philosopher which had been promised in Sophistes 21743 and Poltit- cus 2574. This is impossible, the Politicus being much later than both Symposium and Phaedo. 16 This would result from a doubtful interpretation of Varro, de lingua latina, lib. VII. cap. iii. 88. The Phaedo has been inter- preted in different ways, but is chiefly important as the first attempt to sustain by logical argument the theory of ideas, which had been only poetically repre- sented in the Sym- postum. Beauty was the first idea, extended in the Phaedo to a system of ideas. Value of sense per- ceptions investi- gated. They are found to be mis- leading. Ideas per- ceived by the soul alone without help of the body. They are more evi- 246 ORIGIN AND GROWTH OF PLATO’S LOGIC language. In the Symposium one aspect of such a substance was displayed as an ecstatic vision insufficiently pictured by the witness who experienced it but found himself unable to give expression in words to this unique and marvellous revelation. The first substance thus discovered by Plato was Beauty, bearing some relation to the Good, or ethical Beauty (Symp. 205 E: o & éuds Aoyos oUTe Huiceos dno élvat TOV Epwra OTE GXroUu, sav py TUYXar'N yé ou, @ éraipe, dyaOov bv. 212A: opwvre @ opardv Td KaNOV, TixTay ovK Eldwra apetis, are ovK eidwAou épar- Topev@, GAN adnOH, dre TOU aNnNOods eharrropévm). This Beauty, called already in the Symposium the Good, Truth, or reality, appeared in the first moment, suddenly raised above all human standards, as the only substance of the Universe. Soon, however, growing accustomed to the ideal existence of Beauty, he generalised this experience, extending it to other notions. This he does for his readers first in the Phaedo. He builds a system of ideas and gives an account of the way leading to his idealism ; so resuming the inquiry commenced in the Cratylus. After refuting all attempts to find truth in words, he discusses the value of knowledge gained by sense percep- tion, and held by ordinary ‘common sense’ to be the most certain of all. Heat once distinguishes sight and hearing as the best of all senses (65 8B), but finds that even these give us no correct notions, as has been already recognised even by the poets (65 B: of qroural tiv det Opvrodcw), and, we might add, by such philosophical predecessors of Plato as Heraclitus and Parmenides. As in the Symposium the ecstatic vision of Beauty was independent of the senses and different from any material representation, so now in the Phaedo appear many other ideal substances, perceived by the soul alone, without help of the body (65 Bc: 7 wuyy . . . Stay perd Tov cm@patos srivyeiph Te oxoTreiv, Sprov Stu rote &atratarat tn’ adrod). This is done through reasoning (65 0: év Ta AoylSeoPar) in moments when neither sight, nor hearing, ORIGIN OF THE THEORY OF IDEAS: PHAEDO 247 nor bodily pain or pleasure affect us, and when we feel as if we had left the body in order to approach true being (65 c: Noyiferac . . . KdANOTA, GTay... édoa yalps Td Tama, kal kal’ dcoy Svvatat py Kotvwvoica a’T@ pnd’ dmroudvn opéyntat Tov dvros). Such substances as ideal Justice, or Beauty, Health, or Power, have an existence more evident to our reason than is the existence of particular things to our senses (65 D), though we can neither grasp them with our sight nor any other sense. We know them best by pure thinking (65 E: 6s dv wddwota nud Kal axpiBéotata Tapackevacntar avtTo Exacrov SvavonOjvar rept ob oxKorei, ovtTos dy eyyurata tot Tod yva@vat ExacTov), emancipated from the influence of sense perception (65 E: Ssavoia .. . MeTa TOU AoYiop“OD . . . NTE THY Oy TrapaTLOépEvos bv TO Stavosta Oat pyre tiva ANAnv alaOnowr). Human passions, illness, and physical necessities or desires put obstacles in our way to knowledge (66 B: pupias . . . aoxoNias Trapéye TO o@pa), and lead to wars or other conflicts (65 c). Thence Plato infers that ideal knowledge will be attainable for us chiefly after death (66 £), and that in earthly life our only way to approach truth is to limit the activity of senses to what is indis- pensable (67 a: év & adv Capev .. . ayyutatw écopeba Tov sldévat, av 6 Te pddAtoTa pnday OmA@pEev TO C@patt). Only the pure soul can reach pure truth (67 B: uy Kabap@ xabapod ébdrrecBar pn ov Oepirov 7). We can learn nothing from our senses, because our soul possesses eternal innate wisdom, and all our learning consists in remembering what we knew before this life (72 B: paOnous ouK GAO TL H avdapvnots TUyYavE ovoa . . jas ev amporépp tii ypovm peyabnnévar & viv avayipry- oxdus0a). The reminiscence depends upon similarity or dissimilarity of absolute ideas with the concrete objects of earthly experience (74 A: cupBaive tiv avapvnow elvar pav ad’ Gpoiwy, elva. $8 nai am avoyolwv). Still we notice in every case the difference between a perfect idea and the sensible experience which reminds us of this idea (74 4: . GvayKn Tov dent to reason than material things to the senses. Body puts obstacles in our way to know- ledge. Ideal knowledge expected after death. The pure soul reaches pure truth, and possessed it before entering the body. Difference between idea and particu- lars illus- trated by a mathe- matical example. Thereis no material equality in the world, though it is easy to under- stand ideal equality. This truth was not gained by induction or experi- ment, but appears to be the result of pure thought. 248 ORIGIN AND GROWTH OF PLATO'S LOGIC dvayKatoy rode mpoomdoyew, évvoeiv eire te EAXdElTrEr TobTO Kata Ti opooTnta Eite yn) éxalvou ob avepvycOn). Here Plato introduces a classical example of this radical difference between an idea and sensible particulars : an example which has lost nothing of its logical import- ance up to the present time, and which also shows a far-reaching apprehension of the sensible world. This example he finds not in the distance between a concrete work of art and the artist’s ideal, but in the perfection of a mathematical notion. He knows equality as the basis of all mathematical reasoning, and dares to assert that there is no such equality in the physical world. We know in our times, after many difficult measurements, that no one grain of sand is equal to another, nor a drop of water to another drop. But Plato had no microscopes and micrometers at his disposal, and it was a deep in- sight into the nature of physical phenomena that allowed him such an audacious generalisation against the evidence of his senses. He quotes as examples stones and pieces of wood, which only appear to be equal (74 8B), but are not. He certainly knew physical objects which, accord- ing to all his means of observation, were really equal to each other, as for instance two stars of the same size and brilhancy, two wings of a small insect, or even two coins of the same mint. He could not ascertain the small existing differences between such objects by exact measure- ments and observations as we are enabled to do now; he had not arrived at his conviction of the impossibility of physical equality by Socratic induction. It was for him a knowledge a priori, quite as much as the knowledge of moral ideas. Huis reasoning was not built upon attempts to establish differences between apparently equal objects. He knew beforehand that the idea of equality was too perfect to be realised in the physical world. And this a priort knowledge of Plato has been confirmed by the experience of all the generations which have come after him. ORIGIN OF THE THEORY OF IDEAS: PHAEDO 249 Plato had never alluded in his earlier writings to that difference between idea and appearance. In the Cratylus he mentioned things corresponding to the notion formed of them, and even derived general notions from particular experiences. In the Symposium he reached the sight of absolute Beauty by progressive generalisations which might be described as a continuation of Socratic induction. It is only in the Phaedo that he undertakes to construct a knowledge entirely independent of concrete particulars, and shows us the first model of such absolute ideas in the mathematical notion of equality, not derived from ex- perience. We have already seen in the Meno the theory of tran- scendental knowledge exemplified through a psychological experiment. But in the Meno there is no mention of a difference between ideal and physical equality. The figures were assumed to be equal, and their equality known. Here in the Phaedo we meet the assertion that there are no two equal objects in this life’s experience, and that therefore all notion of equality is older than the present life. The apparent equality of two material objects approaches indefinitely the limit of absolute equality (75.4: dpéyerat mavra taira eivat olov to ioov, eye 42 éydceotépws), and offers us the only opportunity of recalling the notion of absolute equality (754: dporoyodpev, p7 GNoGev abro évvevonxévar ... GAN’ 7) dx TOD ietv 7 ArxpacBa } a Tivos Gd\Ans TOV alaOncewv). This reluctant concession leaves a certain importance to the activity of the despised senses. Without their perceptions we could not find an opportunity of remembering general ideas as the object of our transcendental knowledge. But once remembered, absolute equality is known to be radically different from any equality observed, and cannot therefore proceed from particular instances of approximative equality. This principle is extended to other ideas, not only of mathema- tical relation but also of justice, holiness,and everything that is predicated of particulars (75D: qept amdvtwv ols A process not ob- servable in earlier writings than the Phaedo. Even in the Meno ideal equality is not thus distin- guished. Still, sense per- ception remains & necessary condition for our training in the intui- tion of ideas. We are re- minded by con- crete ap- pearances of eternal ideas, which were known be- fore our birth. Ideas more real than material objects, and they explain them. Perma- nence of ideas a condition of un- change- able know- ledge. Privileges of philo- sophers who become equal to gods. 250 ORIGIN AND GROWTH OF PLATO’S LOGIC ériappayitoue0a 76 5 #ort). All these ideas must have been known before we began to see, to hear, and to receive other impressions of our senses (75 B: mpd Tod dpa dp£a- o6at Huds opav Kal dxovew Kai Tara aicOavecOas Tuyely Ber mov eiAnportas ériotnunu avrou Tov toou 6 Ti Zo TL) In order to enable us to refer every sense perception to such eternal ideas (75B: ta é&& Trav aicOncewy loa éxeioe avolcew, Ott mpoOupstrar pay wavra totadta elvat oloy éxeivo, oti 8é avtov pavAoTepa). These ideas have an eternal existence, independent of the changes of sensible things. Beauty and similar ideas have the most real kind of existence, much more than any material appearances (77 A: ovx dyw Bywrye ovdev otTw ot avapryés by ws TOTO, TO TaVTa Ta tovavr’ elva. ws olov Te paddiora, Kadov Te Kab ayabov Kai Ta\XNa rayra & ov viv Sn Eeyes). Only through these ideas do we begin to understand the outward world (76D: ote... Tacay ToLavTn OvGia, Kal art TavTnY Ta éK THY alicOncewy rdavra avadépoper). Kverything that exists belongs to one of these two kinds (79 A: Oa@mev Svo eldn tav SvTwv): the visible material world, continuously changing, and the invisible ideal world, eternally the same, consisting of ideas and souls. No permanent and durable knowledge can refer to any but eternal objects, ideas without change. When the soul investigates ideas, certainty and knowledge are attained, and this we call activity of reason (79 D: qrepi axsiva dal Kata TavTa WoavTws eye, dre TovovTwy éparropevn: kal todro avtis TO mdOnua dpovnots xéednrar). Such an activity implies happiness, and frees us from error and all human sufferings (81 4). And far more than even this, the victory over illusions of the senses leads a philosopher to become after death equal to the gods (82C: eis Osdv yévos un Pidrocopynoavre Kai Travrénas xabap@ arriovte ov Oéuss adixvetoOar ddr’ 4 TO gtdouade?). A philosopher holds to be true only what he knows independently of the senses, through the pure activity of his soul, which gives an immediate, intuitive ORIGIN OF THE THEORY OF IDEAS: PHAEDO 251 knowledge of ideas (834: 7 g¢tAocodla . . . qwapapvbeirar .. . Thy spuyny Tapakenevonévn Tiotevew pnodsvi GW GXN h abrny avrTn, 6 Tt dv vonon aitn Kal’ avrny avto Kad QUTO TAY OVYTMV). Here we read for the first time about a science of thought or logic (90 B: 4 aept rovs Adyous téyvn), which 1s Indispensable in order to preserve us from utter scep- ticism. He who trusts his own thoughts without an objective logic will often change his opinion, and this will lead him to a general distrust of human thought (90c: TéNEUT@VTES OlOVTAaL . . . KATAVEVONKEvaL povoL OTL OUTE TAY mpayparwy ovdevos ovday inytds ovde BéBatov oTE TOV AOYwv). Such men believe themselves to have discovered that there is no truth, and that any and every opinion may be suc- cessfully defended by arguments among which none is decisive. This is an abnormal state of mind resulting from over-confidence, and similar to the misanthropy born of trusting men without knowing how to distinguish the good from the bad (89p). If, with this unlimited confidence, a man should be deceived by those on whom he relied, he will fall straight into the contrary extreme, and cease to put any trust in his fellows. To this wrong conclusion he is brought by his ignorance of psychology (89H: dvev réyvns ths wept tavOpwmTea), and in like manner ignorance of logic may lead to a general distrust of human reason (90 CD: oietpov dv etn 70 1d Oos, & SvTos 5 twos arbors Kai BeBaiov Aoyou xal Suvatod Karavojaat, Hrevta . . . pot) Eavrov tis aiti@ro unde Thy Eavrod arexviay, GNA... el TOUS NOyous ad’ éavTOD THY aiTiay amwcaLTe . « » ToY 58 dvTwY THs adnOEias TE Kai ErioTHENS OTEpNOEin). In this case they lose, by their own fault, the opportunity of knowing Truth and Being, and have no right to accuse human reason generally of imperfection. Plato himself is certain that human reason possesses the power of an infallible knowledge, and that we owe our errors, not to the weakness of our reason, but to the influence of the senses. Necessity of logic insisted upon. Origin of scepticism in the want of logic. Scepti- cism com- pared with misan- thropy. Ignorance of logic similar to ignorance of psycho- logy. Logic unjustly accused. Power of human reason to attain truth by means of logic. Final causes esteemed above me- chanical causation. Ultimate aim of Being produces unity of knowledge and ex- istence. As thought reflects reality, we can in- vestigate reality in our thoughts. 252 ORIGIN AND GROWTH OF PLATO’S LOGIC To find absolute Truth our thought must be pure thought, and we must take care not to trust other expla- nations of reality than those based upon an understanding of the ideal aim of everything (97.c: é ovv tis BovAotTO Thy aitiav evpeiy mept éxdorov . . . TovTo Sev mepl avrod evpeiv, 6rrn BéXtLaTOv avT@ sori 7H elvas 7 AAO OTWdY mdoyew i troy). This ideal cause is esteemed by Plato very much above all mechanical causation, which for him 18 nO causation at all, but mere succession of events, or at most a necessary condition of real causation (99B: a@AAo pév tie gore To altuov T@ dvtt, GAro S2 éxeivo avev ov TO aitiov ovx av trot’ ein airov). Plato rises here to the summit of his new idealistic metaphysics, despising all mechanical explanation of Being as quite unsatisfactory and criticising his great predecessor Anaxagoras (98 B—E) for not having understood the importance of final causes. The only true cause appears to be that divine power which leads everything to the best, and according to the aims of the whole as well as of all parts (990: Thy Tov ws olov te BedXticta avTa TeOnvar Svvamw . . . Satpoviav taxyvv). But the immediate knowledge of this ideal cause is beyond the scope of mankind, and Plato seeks an indirect way in order to find out the causes of things (99C: aya pav ovv Ths TovavTns aitlas, bry Trot? #yet, uaOnrns OToVOdY jovor’ av yevoluny® éredyn 88 tavrns éorepnOnv Kai ovr’ autos eipsiv ore Tap GAdov pabsiy olos te dyevoynv, Tov Sevrepoy mAobv érl thy rhs aitias Sntnow mTempayyatev- pat). This second-best choice is based on the reflection that human thought is, as it were, an image of reality, and that exact knowledge of thought leads to a know- ledge of truth (99E: woke 8) wos yprvas sis Tous Aoyous xataguyorta év éxsivors cxoreiy tov SvTwv Thy adjOaay), just as we can observe the image of the sun reflected in a well, thus avoiding the injury to our eyes attendant upon looking at the sun itself. Once on this path Plato soon recognised that thought ORIGIN OF THE THEORY OF IDEAS: PHAEDO 253 is more than a mere image of Being (1004: od avy ouyxape Tov év Tols NOyoLS GKOTrOUmEVOY TA GrTa &v #iKOCt paANov oKoTrely 7) TOV av Tots Epyors), a8 he had explained it in the Symposium. He now applied the hypothetical method proposed in the Meno, seeking for the safest hypothesis on which he could rely, admitting as true everything in agreement with, and rejecting as false anything contradictory to this fundamental proposition (100 A: drroépevos éxdorore NOyor by Ay xplvw Eppwyevéora- Tov eat, & pev ay pot Sonn TovT@ ouvpdovelv, TiOnw ws ann67 dvra, & & dv un, ws ov« adnOy). As such a funda- mental hypothesis he proposes to accept the independent existence of Beauty as set forth in the Symposium, and also of other ideas (100 B: wtaroOéwevos slvat tt Kadov avTo al’ avo Kal ayaOov Kal péya Kal Tada rravta). This he calls here, nothing new (1008: ovdév xarvov), but already frequently spoken of. It would be, however, an error to infer that another written exposition of the theory of ideas preceded, besides the first initiation in the Symposium. If the Platonic Socrates asserts that he constantly repeats the same truth in other as well asin the present conversa- tion (100 B: del cai GddotE Kal év T@ TapedyrAvOoTe Aoye), this is a rhetorical artifice by which, on the one hand, Plato brings his new ideas into close relation with the old Socratic notions as subsisting still in the Huthydemus (301 a) and Cratylus (489 D), while on the other hand he refers to conversations with his pupils which may have been suggested by the argument of the Symposium. Among the literary works of Plato none can be thought of as referred to in this passage of the Phaedo, because none contains a more elementary and fundamental ex- planation of the theory of ideas, the Phaedrus and Republic being undoubtedly later, as will be seen from their psycho- logy, and as has been already made evident by their style. An earlier written exposition of this theory would have rendered superfluous the painstaking didactic tone of the Phaedo, and the difficulty of understanding expected by Thought is even more than an image of Being. Concep- tion of a highest principle or hypo- thesis, which is here the existence of ideas. Apparent allusion to earlier exposi- tions ex- plained. The only earlier ex- position in the Sym- postum. Relation between Phaedo and Sym- postum in the theory of ideas. Relation between idea and particu- lars simi- lar to that between a general notion and particu- lars. 254 ORIGIN AND GROWTH OF PLATO'S LOGIO the Platonic Socrates, and admitted by his hearers (1004: BovrAopar S4 cot cadtorepoy eimeiy & Adyw° olpar yap ce viv ov pavOdvev—ov pa tov Ala, tpn o KéeBns, ov opodpa). This reminds us of the admitted obscurity of the speech of Diotima in the Symposium, and gives the impression of a first attempt at a written account of the new theory. The theory as it stands in the Phaedo is a generalisa- tion of the esthetic experience related in the Symposiwm. Particulars are what they seem to us to be, through their participation in the idea, and not only in the idea of Beauty but also in the ideas of all other general notions. The term peréyev used here (100 c: dalverar ydp pot, & ti gorw ddAXo Kadov, ovde de’ By GAXO Kadov elvar A Sidte peréyes éxeivov Tod Kadod * Kal wavta $1 obTrws Asyw), a8 1D the Symposium, is already felt to be not quite sufficient, and is supplemented by other terms, zrapovola and Kxowwvia (100 D: od« GAXo Te Tote? avTO KadOV fH 4) éxeivou TOD KaNod elte Trapovota sire xowwvia . . . ov yap Ett TobTO ducxupl- Coat, GAN’ Ste TH KAA@ Ta Kaa yiyverat Kada). The idea is present in the particulars, or is shared by them, this makes no difference for Plato : the only expression of his hypothesis which he believes to be perfectly certain is that beautiful things become beautiful through Beauty, or owe their particular beauty to the general idea. This relation between idea and particulars is formally similar to the relation between a Socratic notion and the particulars ; as expressed already in the Huthyphro (6E: eldos, 6 ravra Ta ova Said iorw . . . pid idéa Td Te dvoota avoows Kal ra dca dora). But the Socratic notion was immanent (Huthyph. 5D: radvroy éorw dv waon wpake 7d Savoy avTd atte, Kal Td dyootoy av Tod péy oolov Tavros dvavrioy, avTd 58 atTr@ Suowv Kai Byov play twa Stay ray & ri wep dv weddAy avdovov slvas), found in the concrete things as their point of similarity, while the Platonic idea is self-existent, independent of particulars, perceived by pure reason against all illusions of the senses. More- ORIGIN OF THE THEORY OF IDEAS: PHAEDO 259 over, the terms efSos and idéa, which were freely used to designate general notions in earlier dialogues, up to the Gorgias and Cratylus, preserve generally the same mean- ing in the Sympostwm and Phaedo,”" while the transcen- dental ideas are chiefly designated by the neuter of the adjective, sometimes with such determinations as éxefvo (Symp. 210 E, Phaedo 108 c) or adro xa’ aito (Symp. 21138, Phaedo 100 8B) and by the verb elva: and its deriva- tives. The direct and constant use of éeldos or idéa to designate a transcendental idea belongs to a somewhat later stage of Plato’s logic. In the Sympostwm and Phaedo he still hesitates, and this hesitation produces great variety of terms for the peculiar relation between idea and particulars.2® He says expressly that he does not insist upon any of these terms,?!° and that the only thing he is sure of is the priority of the idea, or that the given idea 217 el8os as well as i3¢€a means shape, form, or appearance in such passages as Symp. 1898, 196, 204c, 2158; Phaedo 73a, 104d, 108p, 1098. The meaning of a Socratic species or notion appears in Symp. 2055p; Phaedo 91p, 100 B, &c. Campbell has shown in §§ 24-32 of his essay on Plato’s use of language (Plato’s Republic, vol. ii pp. 294-305) that both words have been used frequently by Plato in the same meaning as by earlier writers besides the new applications, chiefly illustrated from later dialogues. In the formula rd én’ efSe: naddy (Symp. 2108) we also miss the specific Platonic use of efdo0s. The possible identity of efdos and the absolute idea seems to be admitted in the formula: elval ri exacrov tay eid3éy (1028). But here also the ef5y mean ethical notions of which substantial existence is predicated. Only Phaedo 1048 iSéa and 104 c ef5y might be equivalent to the Platonic ‘ ideas.’ 218 Besides peréxew, wapovola, xowwvla we read: peradauBdvew 102 n, npordéxerba: 102D, mpoordva: 102, 103d, SéxerOar 1025, 103 D, eveivas 108 B, perdoxeors 101 c. 219 Phaedo 100 p: ob yap Eri rovro BuexvplCoua: has been interpreted as a, reference to an earlier different opinion by Diimmler (Akademika, p. 204), P. Natorp (Philosophische Monatshefte, vol. xxvi. p. 467), and Pfleiderer (p. 395). But this interpretation is based upon the assumption that Plato wrote about the theory of ideas before the Phaedo. If the Phaedo, as results from the present inquiry, is the first methodic exposition of Plato’s theory of ideas, then ‘od yap rx’ does not signify ‘no longer,’ but ‘not further,’ ‘not moreover.’ The whole phrase would then mean: I am only sure that beautiful things are beautiful through Beauty, but I do not go so far as to affirm anything definitively about the exact manner in which this occurs. Use of terms. Variety of terms for the ideas. Priority of the idea. Logical rule as to the judg- ment on an hypo- thesis, and its conse- quences. Progres- sive gene- Talisation up toa highest principle. Supposed polemical reference. Hypo- thetical method extended, and de- fended against unknown 256 ORIGIN AND GROWTH OF PLATO'S LOGIC is the cause of the corresponding quality in each particular thing in which it is recognised. On this fundamental hypothesis, according to Plato, a consistent system of science can be built up (101 pb: dyopuevos exelvov tod aodadois tis vrobécews). He develops the hypothetical method given in the Meno, and recommends his disciples always to distinguish between an hypothesis and the consequences drawn from it. Ina skilful discussion, the agreement of all consequences with each other must precede any inquiry as to the truth of the hypothesis on which the consequences depend (101 D: «e 6 tis avris tis trobécews ehoito, xaipew sons Av Kat ovK diroxplvato, gws dv ta am’ éxelvns opunbévra axéyraio, ef cou addAndOLS cUUHwvel H) Stadwvei). He advises rising from one hypothesis to another until irrefragable transcendental axioms are reached, which have no further need of demonstration (101 D: éredy 82 éxeivns abtis Séoe oe SuSovar AOYor, dcavtws Av diSolns, GNAnv av brrdbecw vrobguevos, Aris TaV dvwbev Bertlaty dalvouro, tws eri tt ixavov AOos). He warns us against coupling illogically (101 E: @o7ep of avrTiAoytKoL), IN One and the same discussion, arguments for or against the hypothesis itself with arguments for or against the derived consequences (101 C: dua ovw dy gupots repli te Tis apyns Siadeyousvos cal rev 8 sxetvns opunudvov, elrep Bovr0od rt rev Sytwv evpsiv). This exhortation to a methodic investigation is aimed, as Diimmler thinks, against Antisthenes, and is emphati- cally assented to by Cebes and Simmias simultaneously (102 a), and by Echecrates who hears Phaedo’s report of the conversation. Phaedo adds that to all present, even to those who had the least understanding of philosophy, it seemed to be wonderfully clearly expressed (102 a: elrrep ei tav Girocddav . . . Oavpacrids Soxsi ds svapyds 7®@ kal opixpov vodv Byovrt . . . maou rots tapoicw Woke . Kad yap huiy rots airotct, viv 8&8 axovovew). This insistent asseveration of the importance of the logical ORIGIN OF THE THEORY OF IDEAS: PHAEDO 257 rule—to distinguish the consecutive steps of each argu- ment, and to require internal consistency before criti- cising the foundations of a course of reasoning, shows that Plato is introducing a new method (uéod0s, 79 E, 97 B), with full consciousness of its bearings. This new method is generalised from the inductive process by which, in the Symposium, he reached his vision of absolute Beauty. As he then proceeded from particulars to the idea, he now wishes through hypothetical argumentation to reach absolute certainty. Every successive hypothesis must be ‘better’ or logically more evident than the preceding, until by such approximations the goal is attained—namely, certainty. Even then he will not indulge in the self-conceit of those who are delighted with their own circular reasonings (101 E: of avriNoyixot . . . ixavol bro aodias ouovd mdvra KuK@vres Gpws avtol avtots apéoxav). The true philosopher is obliged to examine again and again even the highest generalisations or first principles (107 B: «ai ras brroBécets Tas mpwtas, nal ei mictal tpiv siow, Guws emioKeTrtéa aadeotepov: . . . Kad” Soov duvatov pdadior’ avOpwre) in order to advance as far as human reason may. Plato acknowledges that his own highest hypothesis, when he wrote the Phaedo, was the independent existence of ideas as true substances, always the same, eternal, divine, simple, and representing the highest reality of Being. Were it not for the repeated assertion of the independence of the ideas, we might identify them with general notions. We have no clear indication either in the Phaedo or in the Symposium of any distinction between our subjective notions and the corresponding transcendental ideas. Everything confirms our supposi- tion that Plato, at the time of writing the Phaedo, as well as when he wrote the Symposium, believed it to be possible for the human soul to know ideas as they are, and in such absolute intuition the general notion would be identical with the idea, while the idea remains equally the same g adver- saries. Progress from hy- pothesis to know- ledge, avoiding circular reasoning. First prin ciples re- examined. Probable identity of ideas and their subjective represen- tation. Intuition of ideas dispenses with ex- perience. Law of contra- diction. Solution of the problem proposed in the Cratylus. Stability and inde- pendence of ideas. Relation between ideas. 258 ORIGIN AND GROWTH OF PLA'IO’'S LOGIC both when manifest in us and outside of us (103 B: avro 76 évavriov éauT@ évavtiov ov av wore yévouTo, ovTE TO av Huty oure TO SY TH HUGE). The logical consequence of this doctrine was the power of reason to acquire all truth accessible to mankind by pure intuition, by contemplative meditation without or almost without external experience. In other words, our reason is able to discover the nature of things by intense reflection on the nature of her own ideas, which ideas are common both to human reason and to every other possible reason of any superior being here called God. The logical side of this doctrine culminates in the law of contradic- tion, expressed here as one of the chief arguments demonstrating the existence of ideas (102 E: 7d opixpov ovx éérev Trot? péya yiryverOat ovde sivat, ovd’ adAXo ovdéy Tay évavriov ert dv Sep hy da tovvavtiov yiyverOac re Kal slvat). Kach idea is only what it is, and, therefore, per- fectly simple (wovozdés, 83 E). We see that Plato in the Phaedo gave his solution of the problem proposed in the Cratylus, and definitively decided against Heracliteanism. In the Cratylus he recognised the extreme difficulty of the problem and announced a further inquiry; in the Phaedo he communi- cates the results of this inquiry, postulating not only the stability of notions, already acknowledged in the Cratylus, but their independence of human intelligence. He goes so far now as to deny every process of becoming in the world, or at least to decline any explanation of changes (97 B: ovdé ge Sv’ & te bv yiyverat ws adrriotapas Ste melOw Enavrov, ovd’ ddNo ovdev Evi Ndr Ou’ 6 Te ybyverar 7) drroANUTAL f FoTL, KATA TovTOY Tov TpoTroy THs weOddov). ‘Two unities added to each other cannot become two; it is not the addition which could produce a new idea. Addition is only the subjective side of the eternal relation subsisting independently of our reason between unity and the idea of two. Thesame explanation of all apparent changes through eternal rela- tions between immutable ideas is the result of the absolute ORIGIN OF THE THEORY OF IDEAS: PHAEDO 259 reality attributed to ideas and opposed to the phenomenal appearance of all material things. The Phaedo brings Plato’s Idealism to its highest point and contains a conscious representation of all con- sequences deriving from the fundamental hypothesis sought for in the Cratylus, perceived in the Symposium, and demonstrated here, so far as it could be, for Plato’s followers. We shall meet this theory in later works, while there is no clear trace of it in works that were certainly written before the Phaedo. The importance of the Phaedo for the development of Plato’s logic is increased by the circumstance that the authenticity of this dialogue has passed unquestioned, even by such sceptical critics as Ast and Schaarschmidt. It has been advanced ™ that the Stoic Panaetius in the second century B.c. doubted the authenticity of the Phaedo, but Zeller has clearly shown the untrustworthi- ness and even the contradictoriness of the testimonies adduced in favour of that assumption—the first mention of these pretended doubts occurring some centuries after the death of Panaetius and betraying a complete igno- rance of Panaetius as well as of the reason of his imputed scepticism. The Phaedo has been so frequently quoted by Greek and Latin writers that we must admit that this work was generally regarded as undoubtedly authentic. The extreme idealism here professed has provoked severe criticisms, as for instance those of Crawford **! in the eighteenth and of Prantl * in the present century. But even these criticisms show that, if Plato’s idealism was mistaken, such mistakes can be made only by a 220 R. Hirzel, Untersuchungen zu Crceros philosophischen Schriften, 1877, vol. i. p. 232; Chiappelli, ‘ Panezio’ in Filosofia delle scuole italiane for 1882; also Teichmiller, vol. i. p. 126. 221 ©, Crawford, A Dissertation on the Phaedo of Plato, London 1773. The author evidently had a very superficial knowledge of Plato and pro- fessed a shallow materialism. 222 Prantl, Geschichte der Logik im Abendlande, p. 78, Leipzig 1855; also in his translation of the Phaedo, Berlin 1884. s 2 The Phaedo contains the first repre- sentation of Ideal- ism. Unques- tionable authen- ticity increases its impor- tance. Plato’s idealism provoked strong op- position, but was of lasting import- ancein the history of human thought. New demon- stration of immor- tality. Proofs not entirely sufficient, but meant seriously, though consistent idealism would abolish individual immor- tality. 260 ORIGIN AND GROWTH OF PLATO'S LOGIC philosopher of genius, and they are indispensable for the progress of philosophy, just as many failures of expe- ditions undertaken with the aim of discovering the sources of the Nile were indispensable for the progress of geography. Idealism is one obvious solution of the metaphysical problem, and it was necessary to follow out all the consequences of this solution in order to decide upon its value. In the Phaedo Plato is still struggling against some consequences of his idealism. His love of the religious traditions about the immortality of the soul, as set forth in the Meno and Gorgias, and indirectly confirmed in the Huthydemus and Cratylus, is really not quite consistent with the doctrine of idealism, and though we have no direct evidence whether he was aware of this inconsistency, we see that in the Symposium, to- gether with the first glimpse of eternal ideas, there appears almost a substitution of immortal influence for the immortality of the person taught in the Gorgias. Now in the Phaedo the avowed purpose of the Platonic Socrates is a demonstration of immortality, and he connects this demonstration with the exposition of the theory of ideas, which really might have impaired the religious belief in immortality. But if we examine the arguments in the Phaedo, we see that those from the beginning up to the objection of Cebes (87 4) prove only the persistence of the individual soul for some time after death, not for all time. The remaining arguments refer more to the idea of soul than to the individual soul, though they seem intended as a defence of personal immortality. Archer Hind * argued this question against Hegel and Teichmuller, and made it very probable that Plato in writing the Phaedo still really believed in a prolongation of individual existence after death, without any suspicion of inconsistency. If we look at the final conversation of Socrates with Crito (115 D: éredav iw 7d 2% The Phaedo of Plato, edited by R. D. Archer Hind, London 1883, pp. 18-26. ORIGIN OF THE THEORY OF IDEAS: PHAEDO 261 pappaxor, ovKére Vuiv qrapapeva@, ddr’ oiyncopas aTriwv.. .) we must admit that Plato perfectly understood the con- sequences of personal immortality and believed them. The inconsistency between immortality and idealism arises only if by immortality is meant, according to our modern notions, absolute eternity of the soul, while an indeterminate continuation of the soul’s existence after death is not inconsistent with idealism. It is difficult to doubt that Plato meant his arguments as sufficient to establish individual immortality, because his conclusion does not admit of another interpretation (107 A: mavros MaANov dpa Wuyn aObdvarov Kai avddeOpoy Kat Te dvTt EgovTat Nuav ai yuyai av "Acéov). It is also a natural psychological consequence of the profound dissatisfaction with the present conditions of life, manifested by Plato in the Phaedo, that he could not easily throw off his hopes of a better state, and of a deliverance from physical limitations., The physical theory of the Phaedo, representing the insignificance of the world accessible to our knowledge as compared with a wider world even physically more perfect, confirms the place assigned to the Phaedo in the de- velopment of Plato’s thought. Here he appears no longer as an Athenian, nor as a Greek, but rises even above the standpoint of international equality between Hellene and Barbarian attained in the Cratylus. In the Phaedo there speaks a philosopher whose interests are not limited to the earth, but extend over the universe, though maintaining still the position of the earth at the centre, in conformity with the traditional religious beliefs which Plato afterwards discarded. The position of the Phaedo after the Symposium is evident from all the above comparisons, but additional evidence is not wanting as to the relation between these two dialogues, a relation generally admitted by writers on Plato since Schleiermacher. This relation allowed by Stallbaum, Hermann, Steinhart, Susemihl, and many Physical theories of the Phaedo confirm its place after the Sym- posium. Some direct con- firmations of this view are found by comparing both dia- logues, which are closely connected. The order of writing might in this case agree with order of events re- presented. 262 ORIGIN AND GROWTH OF PLATO’S LOGIC others,?* has been in recent times very successfully defended by Teichmuller against the older view of Tennemann, Ast, and Socher, who thought that the Phaedo must have been written soon after the death of Socrates. Besides the logical theories in the Phaedo, which are found to be a continuation of those in the Symposium, there are some other indications of the priority of the Symposium. The mention at the end of the Symposium of a discussion about the identity of the tragic and comic poet has no visible aim at that place, but is very well explained if we take it as an apology for the prevalent comic character of the Symposium, and an announcement of a More serious encomium on Socrates to be delivered in the Phaedo. Also the words of Alcibiades, that nobody has yet praised Socrates as he deserves, if referred to Plato’s own time, are better justified if the Phaedo had not then been written. There is a further proba- bility that the picture of Socrates in the Symposium, if planned about the same time as that of the dying Socrates, should have been executed first, however 1m- probable may be the generalisation of similar reasonings as carried out by Munk. Also the view on immortality implied in the Symposium presents an earlier stage than in the Phaedo. In the first moment of the con- templation of absolute Beauty, Plato could look upon immortal fame as an equivalent of immortal life. But so dear had been the belief in immortality to the author of the Gorgias that it became a natural task to base this personal immortality on the new logical theory emancipated from traditional authority. Philosophic reasoning in favour of immortality 1s a new departure, compared with the earlier representations of immortality as a traditional belief, a beautiful tale, true and worthy 44 A, Bischoff, Platons Phaedo, Erlangen 1866, pp. 282-306; L. Noack, Philosophisch-geschichtliches Lexicon, Leipzig 1879; also Michelis, Ribbing, and others, while Peipers, Diimmler, Christ, and Pfleiderer still believe in the priority of the Phaedo. ORIGIN OF THE THEORY OF IDEAS: PHAEDO 263 to be believed in, but not within the scope of positive knowledge. The recognition in the Symposium of immortal fame as desirable shows a regard for human opinion far greater than that professed in the Phaedo, which in this respect approaches nearer to the disposition of mind shown in the Phaedrus, Theaetetus, and Parmenides. On the other hand, while his esteem for public opinion was decreasing, Plato’s consciousness of his own power was undoubtedly growing, and here again we have an argument in favour of the later date of the Phaedo. In the Phaedo the philoso- pher is equal to the gods (82 Bc), while in the Symposium the gods are very much above the philosopher. The proportion of apodictic affirmations, such as aAnOéorata, in the Phaedo is an additional proof of its later date. These form here 49 per cent. of all affirmations, while in no earlier dialogue do they exceed 35 percent., which ratio they surpass in all later dialogues, rising above 50 per cent. in the Laws. This places the Phaedo in a line with the later works, and 1s a very characteristic sign of the increasing certainty which Plato professed to have attained—a certainty which remained with him through life, together with his conscious- ness of the high and divine vocation of the philosopher. This growing confidence is specially evident if we compare two predictions about his own philosophical career putin the mouth of Socrates once in the Apology,” and again much later in the Phaedo™ : Phaedo 78: rodXh pév 4 EAXGs, . moda 8€ kal ra ray BapBapwr Apology 81 A: rotovros oty ddXas ov padlas tpiv yeryoerar . . . 89 CD: mAelous Evovrat Upas ol edéyxovres, obs viv yw xaretxorv, € “A A > 9 4 A vpets S€ ovx poOdveore: wai xade- mweTEepot €govrat Gow vewrepoi elow, kat dpeis paGdAov ayavaxrncere, yen, obs mavras xpn Stepevvacba (nrovvras rovovroy er@dor, .. - Cyretv 8€ xpi kal avrovs per’ dAANAWY tows yap dv ode padios evpotre pad- Aov Upa@y duvapevous Tovro sroteiy. aaa 228 This passage has been already understood as a prediction about Plato by Natorp (Philosophische Monatshefte, vol. xxvi. p. 453); Sybel (De Platonis prooemiis Academicis, Marburg 1889) and others. 226 On Phaedo, 78 a, see Teichmiiller, i. 123. Disregard of human opinion. Increase of Plato’s certainty visible in the pro- portion of apodictic affirma- tions. Plato’s growing confidence illustrated by the allusions to his own activity contained in Apology and Phaedo. Allusions to Plato’s travels and teaching. Difference between Phaedo and Sym- posium in the treat- ment of poets. Also in style the 264 ORIGIN AND GROWTH OF PLATO’S LOGIC According to the Apology there was no hope of finding a worthy successor to Socrates; in the Phaedo it is admitted as probable that such a successor, even if sought for all over the world, could not be easily found outside the circle of Socrates’ disciples, and this is said with a clear reference to Plato’s travels in search of truth. In the Apology Plato speaks of the indignation which will be produced by his writings; in the Phaedo he is already conscious of the charm exercised by his philosophy, and he calls himself a charmer. We shall see how Plato progressed even to a further point in the consciousness of his own power. Another indication of the priority of the Symposium is the different treatment of Beauty. While in the Symposium Beauty is the highest ideal, it is in the Phaedo only one among many ideas, as in the Phaedrus. In the Symposium Plato quotes poets and lawgivers as truly eminent men, deserving immortality of fame ; in the Phaedo (65 8B) the poets are quoted with a certain irony, as if Plato meant that any truth observed by them must be clear even to a child. While in the Symposium Aristophanes is represented as a friend of Socrates, and Plato thus forgives the gibes of the great comic poet against his master, he refers in the Phaedo (70C: ovxouy y’ dv oluat, ciety twa viv dxovcarta, ovd’ et KwpwdioTroLos sin, @s adorscy®@ Kal od rept mpoonkovTwy Tos Adyous movovuat) to comic poets with a certain air of superiority and contempt; this reminds us of the Republic, and seems to be directed against comic poets of Plato’s own time who criticised, perhaps, the wild and playful tone of the Symposium. The position of the Phaedo after the Cratylus and Symposium is fully confirmed by the considerable number of peculiarities of later style, which bring the Phaedo nearer to the Republic and to the latest group than any of the preceding dialogues (see above, p. 170). If we take into consideration that no other work of Plato is likely to ORIGIN OF THE TIIEORY OF IDEAS: PHAEDO 265 have been composed between the Symposium and Phaedo, we must infer that the two dialogues were not separated by a great interval, since it is unlikely that Plato would remain long unproductive as an author at the period of his life in which his chief works betray such incomparable ease and mastery of form. The stylistically well-defined group consisting of the Cratylus, Sympostum, and Phaedo, contains the first exposition of the theory of ideas, and shows us how Plato was led to this theory from different starting-points. In these three dialogues the ethical questions so much dis- cussed before become secondary, and the logical problem of knowledge, blended with the metaphysical inquiry about Being, begins to occupy the philosopher’s attention. He reaches a degree of certainty and a consciousness of his power forming a remarkable contrast with the incon- clusiveness and modesty of the Socratic dialogues up to the Meno. Also his literary skill, admirable already in the Huthydemus and Gorgias, arrives in the Symposium and Phaedo at a perfection not exceeded by himself in later writings, and equalled only in the Republic and Phaedrus. The polemical tone of the Euthydemus and Gorgias is disappearing, and the didactic character begins to prevail. The aim of life, which in the Gorgias was defined as justice founded on knowledge, becomes chiefly knowledge, with virtue as one of its consequences. The stage reached by Plato in the Cratylus, Symposium, and Phaedo is introductory to that of the Republic and Phaedrus, which represent the doctrine taught by Plato during the mature years of his life. Stylistic and logical comparison agree in connecting the Cratylus, Sym- posium, and Phaedo into one group of works succeeding each other in the first years of Plato’s activity in his Academy. The great number of works later than these reduces the limits of time for their composition to a few years. If the Symposium was written about 385 B.c., we Logical character of the Cratylus, Sympo- sium, and Phaedo contrasted with inconclu- siveness of Socratic dialogues. Preva- lence of didactic over polemical aims. Connec- tion of these three dialogues confirmed by their style. They were written in the first years of the existence of the Academy. 266 ORIGIN AND GROWTH OF PLATO'S LOGIC have no reason to put the Phaedo later than about 384, or between 384 and 383 B.c., as will be seen from a com- parison between the Phaedo and later works, proving that after the Phaedo Plato must have written more than twice as much as he had written before. CHAPTER VI MIDDLE PLATONISM WHEN Plato reached the development of his logical theories as these are known to us from the Phaedv, he was anxious to apply them to practical aims with the purpose of promoting the moral progress of his contemporaries. He was not satisfied with knowing the truth for himself, and he wanted to impart it to others. Two practical appli- cations of philosophy occupied his attention: politics and education. We have the results of his meditation on these subjects in two works, the Republic and the Phaedrus. The Republic no longer deals with the moral pro- blem in the fashion of the Meno or Gorgias. In these Socratic dialogues Plato asked and tried to answer moral questions referring to the conduct of the individual, in any given state, without expressly contemplating an altered condition of the state. He still professed Socratic ignorance as to politics, while he already had resolved the problem of individual conduct and indivi- dual relations between citizens, seeing therein the true politics (Gorg. 521 D). But having gained a higher meta- physical knowledge, Plato no longer dared to decline the responsibilities it implied. He was deeply interested in the reasons of the general decay of Greek states in his time, and he understood that the Socratic precept to ‘mind one’s own business’ (Ta éavtotd mpdrtrey) would not work, if the political conditions of the state offered constant opportunities for the perversion of the individual. If the state was acknowledged to be a necessity, the citizen and especially the philosopher could Philo- sophical theories applied to practical aimsin the Republic and Phaedrus. Politics succeed to indi- vidual ethics. The state has an influence on the in- dividual’s conduct, and thus moral re- form must begin by the reform of the state, and by the reform of education. The excep- tional size of the Republic must be taken into account if we wish to reckon the time spent in writing this dialogue. According to some authors this was very con- siderable. 268 ORIGIN AND GROWTH OF PLATO’S LOGIC not remain indifferent to the mode in which the state was to be ruled. Plato’s interest in this problem led him to write one of his greatest works, the Republic, in which educational and political topics are skilfully blended. Having recognised education as one of the chief instru- ments of political reform, he dedicated another dialogue, the Phaedrus, chiefly to educational questions. I. The Republic. Every reader of Plato is familiar with the fact that the Republic is very much larger than any other work of Plato except the Laws. This impression led even Grote to a curious exaggeration, when he said (vol. iv. p. 1) that each book of the Republic is as long as any one of the preceding dialogues. He was thinking chiefly of the small spurious dialogues held by him tobe authentic. In reality four of the preceding dialogues, Gorgias, Cratylus, Symposium, and Phaedo, contain in all about the same amount of text as the Republic, and it is important to bear in mind this relation if we wish to arrive at correct conclusions on the much-debated question of the unity of the Republic. An incidental observation of Hermann (p. 539), that B. V-VII appear to be written later than B. VIII-IX, and that B. X must have been added later still, has been more recently developed by Krohn, and after him by KH. Pfleiderer, into a theory which breaks the continuity of the Republic, by supposing different parts of it to have been pro- duced at intervals during the greater part of Plato’s life. For anybody who wishes to understand the growth of Plato’s philosophy it becomes a very important pre- liminary question whether Krohn was right in _be- lieving that Plato wrote much of the Republic before he had written any other dialogue. This view has been recently carried by Pfleiderer to the extreme of placing the first five books of the Republic even before the Apology, MIDDLE PLATONISM : REPUBLIC 269 which heretofore had been almost unanimously held to be one of the earliest writings of Plato. If we consider that the Republic contains one-sixth of the texts bearing Plato’s name, and that it is generally admitted that he was occupied with literary labours for at least fifty years, it becomes evident that even the continuous production of the Republic could not have been the work of a short time. In our own century a volume of this size and on such an all-important subject is rarely written in less than several years, and there are immense differences between our methods of writing and the mode of literary composition which probably prevailed in Plato’s time. Without referring to fountain pens and typewriting machines, the superiority of our ordinary writing materials over those that were available two thousand years ago has diminished many times the mechanical labour involved. The invention of printing and the custom of revising proofs affords an infinitely easier and quicker way of correcting and maturing our works than was practicable on old papyrus rolls, with an all too limited space for additions. But besides all these mechanical and material improvements, there are also deep psychological differences between an ancient Greek writer and ourselves. Any ordinary student of the present day has read ten or even a hundred times as much as Plato could have done at the same age; we are also generally far more practised in writing from our earliest years: even our elementary education includes besides gymnastics and music many literary studies. Keep- ing all this well in mind, we must ask the question: how many years must the composition of the Republic have required even if it were not interrupted by other labours ? We suppose that in the first ten years after the death of Socrates Plato wrote about half-a-dozen small dialogues, and only two larger works (Protagoras and Meno), not amounting together to more than about three-quarters of the extent of the Republic. This was the beginning, The Republic being one- sixth of all the works written in fifty years, it is pro- bable that it took some years to write. Difference in the mode of literary produc- tion. We have in the present time many advan- tages en- abling us to write at & greater speed. Illustra- tion from the above survey of the works preceding the Republic. In about six years five dia- logues had been written exceeding slightly the size of the Republic. B. II-X of the Republic and the Phaedrus are equal in size to the works written 390-384 B.C. It is pro- bable that the bulk of the Republic and the Phaedrus were written in the last six 270 ORIGIN AND GROWTH OF PLATO’S LOGIC and it is reasonable to expect that the author’s speed in composition was increasing. In fact the next six years (390-384 B.c.) produced five dialogues (Huthydemus, Gorgias, Cratylus, Symposium, Phaedo), which taken together slightly exceed the size of the Republic. Be- sides, there is ample reason to suppose that some work preparatory to the Republic had been already done at the time of writing the Symposiwm, and the tenour and language of the first book have an obvious affinity to those of the Gorgias. Taking this for granted, there is on the other side the Phaedrus, which could not have been written before the Phaedo, as will be seen, and which also is probably not much later than the last books of the Republic. The Phaedrus, together with B. II-X of the Repwblic, corresponds very nearly to the total amount of the works which we place between 390 and 384 B.c. These works are so important and betray such a wonderful facility of composition, united with so complete a mastery of the language and of the subjects, that we have no reason to expect that Plato in the next period still further increased the speed of his writing, especially while his oral teaching must have occupied more and more of his time. Thus it becomes consistent with probability to suppose that the Phaedrus and Republic occupied him for another six years after 384, and this brings us to his fiftieth year, completed in 3t7 B.C. If we say that according to the above reasoning Plato worked on his Republic nearly up to the age of fifty, this remains only a probable inference. But where we have no direct evidence as to facts, we are justified in weighing probabilities and admitting provisionally the greatest probability, in order to obtain a distinct conception of im- portant events. For a knowledge of Plato’s philosophy it is sufficient to settle the consecutive order of his works, and it is not indispensable to name a date for each work or each part of a work. But dates are useful MIDDLE PLATONISM: REPUBLIC 271 as an illustration of results arrived at by the detailed com- parison of each work with all the others, and it is only in order to convey to our readers a clear representation of what results from the above inquiry that we say: if Plato wrote the Republic as one continuous work, and after the Phaedo, as we shall attempt to prove, this work probably filled his time for about six years before he reached the age of fifty. We know he was forty when he formally founded his Academy. His Huthydemus and Gorgias had prepared the way for this, and the first years of the existence of the Academy brought out the Cratylus, Symposium, and Phaedo, enouncing the new theory of ideas. The Republic and Phaedrus were then written within the first ten years of the existence of the Academy. If this be so, one important point of discussion is at once dis- missed. It is natural that an author between forty and fifty, labouring at one production during about six years, while his thoughts were still maturing, should insensibly alter something in its original plan, adding new matter and even falling into some trifling contradictions. Cor- rections were not then so easy as they are to-day, and the standard of literary consistency was, even for Plato, not so high, as we can see from nearly all his works. He was above everything an educator, and he did not feel obliged to say all things at once. He had taught in the Symposium a progressive exposition of truth, and he conformed to these precepts in preparing the Republic. In B. I-IV we see no direct allusion to the theories explained in the Phaedo, and we might receive the 1m- pression that the author did not yet know the eternal ideas. At the beginning of B. V we have a clear indication that what follows is an expansion of the original plan, and at the beginning of B. VIII the thread of B. IV is resumed. B. I is called in B. II expressly an introduction (mpooipmiov 357 A), and B. X has distinctly the form of a conclusion, somewhat years before Plato reached the age of fifty. An author between forty and fifty may change gone things in the plan of a work continued for seve- ral years. Educa- tional aim explains why no mention of the theory of ideas occurs in the first books. Natural partitions of the Republic do not prevent its unity. Different moods in Plato’s works and different aims have a limited influence on the style. First book of the Republic closely related to the Gorgias in contents and style. Probable allusion to the 272 ORIGIN AND GROWTH OF PLATO'S LOGIC loosely tacked on to what precedes. There is no possible discussion about the existence of these partitions, which are evident to every reader, and have been acknowledged generally. But on the other side frequent hints unite these parts into one whole (see Jowett and Campbell, Republic, vol. 11. pp. 11-20). For our purpose, we must consider each part separately, before drawing inferences as to the whole, and we recognise in the Republic five chief divisions: B. I, B. II-IV, B. V-VII, B. VIII-IX, and B. X. Boox I Plato’s mind during many of the best years of life seems to have alternated between a resolute withdrawal from the world, indulging contemplation with a few disciples, and the endeavour to go forth and influence the world and bring the results of contemplation to bear on the social life of humanity. It is natural that his style should alter with the alteration of aim. Yet such alteration of style has limits, and it is hardly conceivable that in a single work produced without intermission he should approach the characteristic form in part of earlier and in part of later writings. The first book of the Republic, equal in size to the Apology, presents a strikingly close affinity to the Gorgias both in matter and form. The gentle treatment of Cephalus may be compared with the ironical respect for Gorgias, the puzzling of Polemarchus with the easy refutation of Polus, the sudden onslaught of Thrasy- machus with the brusque interposition of Callicles. And the presumption raised by these comparisons is confirmed by the stylistic evidence, which yields very few i of later peculiarities. We see here Thrasymachus nising to defend a position which had to be abandoned by Polus in the Gorgias. Polus had admitted that injustice though advantageous MIDDLE PLATONISM: REPUBLIC I 273 is uglier than justice (Gorg. 475 B: to ddixeicOar Kdxcov . 70 88 dduceivy aicywov), and this led to his defeat in the discussion with Socrates. Now Thrasymachus, as if he had been present then, dares to assert that perfect Injustice is beautiful (Rep. 348 D E), whereby he places himself above traditional opinion. Socrates recognises the greater consistency of this position (848 E: rotdro 76n OTEPEWTEPOY . . . EL YAP AVOLTENEiV fev THY AdLKlay éTiBEco, xaxiay péyro. 7) aiaypov avTo wuoroyes slvar, doTep GaNNov Teves, elyowey av te Ayer Kata TA vowlopeva) ; we might take this as an allusion to the earlier work, and as a sign that, however the first book might be earlier than the other books, we need not admit it to be earlier than the Gorgias. The standpoint of the author is far more advanced, since he acknowledges that his argumentation, though sufficient to overthrow a sophist’s impudence, 1s not satisfactory to himself, so long as he has not given a definition of justice, which accordingly becomes the pro- fessed aim of the whole work. In the small dialogues no definition of any virtue is accepted as definitive, and in the larger ethical dialogues the question whether virtue is teachable overshadowed the logical inquiry as to the nature of virtue. It is only in the Republic that this problem is undertaken, and with a new purpose, to apply it to politics. There are some hints which show that the first book was not, as Hermann (p. 538) thought, originally meant as an independent whole, to which the following was added later. The mention of this life as preparing us for death (330 E: dyyutépw dv trav éxst wadXdv Te KaOopa aura . . . Sol A: deta adres asi mdpecti) shows us that Plato, even when he began to write his Republic, had passed beyond the stage of the small dialogues, and perhaps planned already in writing the first book the final myth concluding the tenth book. Also the threefold partition of the soul, which is the most important doctrine of the fourth book, is here as in T Gorgias in the first book of the Republic. More advanced stand- point. Polemic against the sophists secondary. First book not an in- dependent small dialogue. Close relation to the following books. Classifica- tion of rewards here mentioned fully ex- plained in the seventh book. Duty of philo- sophers to accept political power denied in Gorgias here accepted. Classifica- tion of men as in the Phaedo. Difference of terms places the first book nearer to the Gorgias. Relation to the Phaedo 274 ORIGIN AND GROWTH OF PLATO'S LOGIC the Phaedo already prepared, when Socrates says that the rulers of a state are paid in money, honour, or the advantage of escaping a penalty for refusing to rule (847 A: pioOoy rots pédXovow eeryoew apyew, ) apyvptov fh tony t Cnulav, dv py apyn). This is here a riddle for Glaucon, and is fully explained only in the seventh book, where the obligation of the philosopher to rule a state against his inclination is clearly expounded. This doctrine is in advance of the Gorgias, where Plato said that in order to get political influence the ruler must be like the people (Gorg. 518 B: dots oe rovToLs opotdratoy atrepyd- Cera, ovTOs oe Troince, ws eriOupets elyat, TrodLTLKOV Kal pntopuxov). He then saw true politics only in individual educational influence (521 D: olwat . . . ésreyeupsiv TH ws GdnOas troditixh Téyvn Kal mpdtTEew Ta TodTUd), and rejected Callicles’ exhortations to him to take an active part in the rule of the state. Now we see that already in the first book of the Republic Plato is conscious of the duty of obtaining political power in order to avoid the penalty of being ruled by his inferiors (847 C: ris nuias pweylorn ro bd movnpotépou apxecOa:). The three different kinds of men are also in the same passage opposed to each other (347 B: didoripov Te kal durdpyvpov—oi dyaGo/) very much as in the Phaedo (82 c: girocodoivres—giroypypatoi—qircripor). As in the Phaedo we see here the origin of the threefold partition of the soul. In the Phaedo Plato puts on one side the philosopher, and on the other side those who are not philosophers, almost identifying the ambitious and the money-lover (Phaedo 68 ©: the opposite of the philosopher is named dudocmparos and subdivided: 6 avros 84 grou ovros Tuyydver dv Kal diroxypnpatos Kal didrotipos, Hrot Ta &repa tovTwy 7 auporepa). Here, likewise, we have not a direct trichotomy but a dichotomy with a subsequent division of one of the two parts, without a definite name for the third part, for which in the Phaedo the term girocogos is used. This seems to show that B. I is MIDDLE PLATONISM: REPUBLIC I 275 earlier than the Phaedo, and we find a confirmation of it in the circumstance that for the lover of money the word giAdpyvpos 1s used, as in the Gorgias (515 E: Tlepucdda merronkévat A@nvaiovs dirapyvpovs), while in the Phaedo this word is replaced by qvAcypijyaros, which also fre- quently recurs as a constant term in the later books of the Republic. The same relation between the Phaedo and the first book results from the comparison of the following passages : Phaedo 105 c: without any pre- liminary explanation of what épyor means, or of what activities of the soul constitute life, comes the question: ’Amroxpivov .. . @ ay ri eyyévnrat wapart, (av éorat; “Qe dv Wuxn, pn. ovxotv dei rovro oUTws exec; mas yap ovxt; 7 8 Gs. Rep. I. 858 p, after a long enu- meration of épya (immov, 352 E, 6POaApay, etc.) follows: era travra rode aoxéWar: Wuxns gore re épyov, & dAA@ roy dvrwy ovd’ Av évi mpagas, ... TO emepedeicOa «al dpxyew nat BovdeverOa xai Ta rot- adta mdvra, €08’ dro GA@ fh} Wry] Sixaiws av avra dmodoiper, kai paipev ida éxeirys evar ;—ovderi dXo.— ri 8 avd rd Cyv; Wuxns pncopey épyov etvat; pdaduora. It seems improbable that Plato should have explained his thought about life as a peculiar power of soul with such a series of inductions, if the result had been earlier stated to be evident, and on the other side, the short statement of the Phaedo is best justified by the more elementary exposition preceding it. Itis not the length of an explanation which decides the question of priority, because a longer elucidation might be a supplement to a previous short statement of the question. But here we have on one side an elementary induction, and on the other side the result of this induction quoted as evident truth. Under these circumstances the longer explanation may be reasonably held to be the earlier. The position of Book I between the Gorgias and Phaedo is further confirmed by the notion of the peculiar virtue of the soul, which appears here as a development of what in the Gorgias was named the peculiar order in a soul : 4 tT 2 results also from the differ- ent ex- pression of the view that life is peculiar to soul. This, based on long in- ductions in the Republic, stated as evident in the Phaedo. Order in the soul, mentioned in the Gorgias, here de- veloped. Specific energies of the senses re- cognised, but with- out insis- tence. The second, third, and fourth books of the Republic represent the primi- tive state. 276 ORIGIN AND GROWTH OF PLATO'S LOGIC Gorg. 506 E: rdée reraypévov cat Kexoopnpevov éeotiy 4 apery éxdorov . . . Kdapos Tis dpa éyyevopevos év éxdorm 6 éxacrov olxetos ayabov mapéyes exacrov Tay Gvrov ... Kal Wuxi) xoopov €xovea rdv éaurns dpeivoy ris dxoopyrov. Rep. 858 &: Gp’ odv more Wuyxn Ta avris épya «0 dmepyacera areponévn tis olxelas dperijs, i. advvaroy ; — ’Aduvvarov— Avaykn dpa xaxn Wuyi Kxaxas dpxew kat eripedciaba, ty 8€ adyab7 mavra raira ev mparrev. Cf. 885 B: rav Kuvav apery, also rav trmav. The notion of a peculiar power of the soul is intro- duced in connection with the observation that each kind of perception also depends upon a peculiar faculty, resulting in a special activity, which cannot be fulfilled by any other instrument than the corresponding organ of sense (852 E: é8 Srp dp GAr® dots 7} opOarpols ; od Sita . axovoais GAX@ 7) aoiy ; oddapds. 353 BC: ap’ av wore éppata TO avrav Epyov Kada@s arrepydoatvtTo un EyxovTa THY avrov oixelay apernv...). This is a clear statement of the theory known in our century as the law of specific energies of the senses. But Plato did not give any special im- portance to this observation, and it served him only as an analogy tending to establish his general view of human faculties. However, a variety of psychic faculties 1s not yet discovered in the first book, and the soul as in the Phaedo is spoken of as one indivisible whole. Booxs [I-IV These three books, together equal in size to the Gorgias, form one whole, and represent the primitive state, including some considerations on poetry and primary education. The end of this part does not exactly coincide with the end of the fourth book, because p. 445 B begins a new argument, the explanation of a variety of states corresponding to the variety of souls, very soon interrupted at the beginning of the fifth book by the digression on the equality of the sexes. [If we disregard this last page of the fourth book, connecting it with B. V and preparing for B. VITI-IX, we are justi- MIDDLE PLATONISM: REPUBLIC II-[V 217 fied in treating B. II-IV as representing one important division of the Republic, independently of the question whether the following parts were added immediately afterwards or later. We see here chiefly one theory which belongs more to psychology than to logic, but which is indispensable for an adequate appreciation of Plato’s logical progress. This is the theory of the threefold partition of the soul, intro- duced here for the first time and based on the logical law of contradiction. Plato discovered a truth of which he evidently was not yet aware in writing the Phaedo, namely that the soul has multiple opposed activities unified only through constant efforts (443 BE: gva yevouevov é qwoA\@v). He acknowledges the great difficulty of deciding whether the different activities do not belong to one and the same soul (436 AB: yadsema diopicacba: akiws Aoryou ... ef GAN TH Yruyn Kal’ ExacTov avTay TpaTropey .. . 4 tpiclv ovo GddXo GAAw). But he invents a safe method for the solution of his new problem. He puts it down as an unquestionable truth, that the same thing cannot act or be acted upon simultaneously in contrary ways (436 B: tavrov Tavaytia Toeiv } Tdoyev Kata TadTOV YE Kal mpos Tavrov ox &ednoes Gua, repeated 437 A, 439 B). This sharp and general formulation of the law of contra- diction not only as a law of thought, as in the Phaedo, but for the first time as a law of being, as a metaphysical axiom, repeated several times with great insistence, is a very important step, not easily to be accounted for by those who believe the first part of the Republic to belong to about the same time as the Protagoras. Also the ter- minology used to express this truth betrays a stage much more advanced. Plato speaks here as a philosopher already accustomed to exact definition, not the youthful inquirer hesitating and declining the definitive solution of every proposed problem, as he appeared in the Protagoras and earlier dialogues. He is now familiar with the hypothetical method (437 a: vro@éuevort ws Tovrou obtws Threefold partition of the soul based on law of contra- diction, which ap- pears here not onlyas a law of thought, but as a meta- physical principle. Plato admits that each hypo- thesis is taken for granted 278 ORIGIN AND GROWTH OF PLATO'S LOGIC provision- Byovtos eis ro mpdaOev mrpolwpev, opohoynoayvTes, &dv TOTE ally,and @\Ay havi taita 4 tavtn, mdvta Hiv ta dro TovTOU maybe = Ey Baivovra Nekupéva soecGaz) and proceeds according to revoked the logical rule given in the Phaedo (100 a), arguing out later. ; the consequences of the most probable hypothesis. Three This leads him to the conclusion that as our sensual faculties desires are frequently in contradiction with our reason, of the desire and reason must be different from each other soul, (489c D). He thus establishes three powers or faculties oe of the soul for which he does not yet use the term dvvayis arta: (B. V 477 c: dycouev Suvdpes elvar yévos te tTav bvTwr), do not calling them «én (402 0, 437 D, 439 z, 440 8), yévn (443d), exactly or wépn (442.C), with some hesitation as to their relation corre- to the whole. Heseems to have looked upon the faculties spond to as organs or instruments of the soul, according to the val analogy of the senses, which are instruments of the body. nee The three Platonic faculties do not exactly correspond to will, feeling, and reason, which have been later generally used for the classification of psychical acts. Plato’s Xoyiore- xov (439 D: To @ Noylberar NoyLaTLKOY TpoaayopEvOYTES TIS uyjs), though it is apparently the organ of reasoning, includes also the will-power, because it could otherwise not command (441 E: T@ XAoyioTu@ dpyew mpoc- Know- yxav). Plato did not distinguish between pure objective ledge and thought and the decisions of will resulting from a the will to Certain intellectual knowledge. For him knowledge and act ac- . . ° the will to act according to this knowledge were one. cording to it belong 82in, he did not link into one all kinds of feelings, but toone § separated sensual feelings, under the general appellation faculty, of desire, from the moral feeling. Thus two of his faculties while (rrvOujia and Oupos) correspond to one of later psychology, sensual while he finds one faculty where later the will has been feeling 8 distinguished from the intellect. This union of will and rime intellect, as taught by Plato, is preserved in the current use f ene of the word reason, even in the philosophical theories of feeling Spinoze, and in the ‘ Prektische Vermunii’ of Kant. Plato assumes a gradation of faculties, placmg first MIDDLE PLATONISM: REPUBLIC LI—IV 279 reason, then the moral feeling (439 E: © OupovpeOa— 441 E: T@ Ovpoeidet mpoojne imrnxdp elvac xad Evpudyo tovrou (rov doyiorsxov), also 441 A), and at the lowest stage the sensuous desire (439 D: 70 & épa re nal rey Kal Sippy xai wept tas adAdas ériOupias érronras adoyioTov Te cal ériOvuntixov). He argues from the contradictions and conflicts of these three faculties to their independent existence. First, the sensuous desires are frequently opposed to reason and moral feeling, then the moral feeling itself is developed earlier than the reason (441 B: Ovuod pev evOvs yevoueva pecta sort, Aoytopot & evioe pev Euovye Soxodow ovdérote petarapBavev, of 5 Troddol od more). Here we notice that 76 @uyoedés does not entirely correspond even to the notion of moral feeling, because it could not then be attributed to animals (441 8B: éy rots Onpiots av tes ioe 6 Ayers). It is a very curious circumstance that the term Oupoadés, very frequent in this part of the Republic, and also in B. VIII and IX, is entirely absent from B. V-VII and from B. X, recurring besides these parts of the Republic only once in Plato in the Tzmaeus, in connection with a recapitulation of the contents of the Republic. It seems that Plato had a passing fancy for this term and soon recognised it as insufficient, as he clearly avows later in B. VI (5044: rpitta ein suyis diacrn- odpevor... B: éppn0n ta Tore ths pav axpiBelas, ws &uoi &paivero, BAX. . . +). Here also he already con- fesses the imperfection of the method used (435D: ed y’ loOe . . . axpiBas pav toto &k TolovTwy peOddwy, oiats vov ay tots NOyos ypwpmeOa, ob wy mote AGRwyuev) and announces a ‘longer way’ (435D: a@dAn yap paxporépa Kar rsiwy odds H él TodTO dyovea) leading with a greater certainty to truth. This longer way, however, is not fully shown in the Republic, and when later, in the sixth book, Glaucon insists on having it explained (506 D: dorep Sixasoovvns més nal owdpoovvns cal tay adror O1Oes, ofrw Kai rep) rob dyabod 5:éAOys), Socrates con- Opposi- tion of each faculty to the other. Difference in their growth. Limita- tion in the use of the term Oupoedés. Imperfec- tion of method confessed and a longer way an- nounced, but not fully shown. The idea of the good could not be taught by Socrates. The men- tion of a longer way is an allusion to the theory of ideas. In the Sym- postum and the Phaedo, also, the theory of ideas was not constantly referred to. It is an artifice peculiar to Plato to introduce succes- 280 ORIGIN AND GROWTH OF PLATO'S LOGIC fesses himself unable to do it (506 EB: avro péy rl ror’ éotl rayabov, sdowperv Td vov elvar’ mA#ov ydp pot paiverath Kata THY Tapovoay opuny epixéo Oat Tod ye Soxodvros sol ta viv), and returns to his beautiful allegories and metaphors. It was really beyond the reach not only of the historic Socrates, but even of the Platonic Socrates. When Plato set himself to expound the ‘longer way,’ he selected as his spokesmen Parmenides and the Eleatic Stranger, and made Socrates a hearer of their wisdom. The allusion to the longer way in B. IV is very valuable as a chronological sign, because it dismisses at once the supposition that this part of the Republic could have been written before the discovery of the theory of ideas. Plato looked upon his newly discovered treasure as a mysterium too deep to be constantly and familiarly referred to. In the Symposiwm the greatest part of the dialogue does not contain any allusion to the avro ro xadov, and then by a surprise the beautiful vision is pre- sented in the speech of Diotima, suddenly as it had appeared to Plato himself in his meditations. The same order and method were observed also in the Phaedo. In the beginning (up to p. 65D) there is no mention of ideas, then the ideas are mentioned as notions (Sécavov avro 65D), these notions are slowly worked out into indepen- dence of the senses (74: ov ravroy dp’ éotiv tatra Te TA toa xal avto TO ioov), and only after the final objections of Simmias and Cebes, after the criticism of Anaxagoras and other philosophers, appears the theory of ideas introduced ironically as something well known and implied in the preceding argument (1008: ovd&y xawov, Grr’ Grep aad wal GAnore Kal dy TH TwapednrAvOore oy@ ovdev méravpaL Afyov). This rhetorical artifice of Plato, which deceived some inquirers so far as to make them doubt the fact that the Phaedo is the first written exposition of the theory of ideas, is repeated on a larger scale in the Republic. Campbell (Rep. II. p. 11) compares the late revelation sively new Of the ideas in B. V with the peripeteva of a drama. MIDDLE PLATONISM: REPUBLIC II-IV 281 Sybel*” explained this way of proceeding by educational motives. It is quite natural that Plato should reserve the application of the theory of ideas for special occasions, and he found no such occasion in the first sketch of his political views. It was sufficient for him to allude to the longer way. The threefold partition of the soul is not introduced as a psychological problem, nor as subsidiary to some logical investigation, but simply in order to show the parallelism between the three classes in a state (rulers, soldiers, and middle class) and the parts of an individual soul. This analogy between the individual and the state, which can boast of such a long history after it had been invented by Plato, is not the idea of a young Socratic pupil, but of the Master of the Academy, and is a con- sequence of the theory of ideas. When he began to generalise widely and to seek in everything the ruling idea, he thought that he discovered an identity of principle between the state and the individual, and this led him from the individualistic ethics of the Gorgias to the politics of the Republic. The transition is already indicated in the Symposium (210c: tixrew Adyous ro.ov- tous nreiv, olrwes troijcovot Bert iovs Tous véous, lva avay- xacOn ad VedoaaOat 76 av Tots érreridevpact Kal Tots vdpots Kanov xal todr’ dey Sts TaY adTo a’T@ Evyyevés sori), and this indication has been taken for an allusion to the Republic by those who cling to the belief of a Republic written very early, within the first ten years after the death of Socrates. Such a belief is founded on a gross misconception of the relations between Plato and Aristophanes, and illustrates the uselessness of interpreting Plato from 27 L. von Sybel, Platons Technik an Symposion und Euthydem nachgewiesen, Marburg 1889; of the same author on the same subject: Platons Symposion, ein Programm der Akademie, Marburg 1888; on some smaller articles of the same author, see a review by Natorp in Philosophische Monatshefte, vol. xxvi. p. 449. points of view of greater impor- tance. Analogy between state and individual denotes a later stage of thought than the purely individual point of view of the Socratic dialogues. Relation between Plato and Aristo- phanes. Argumen- tation of Schultess in favour of later date of the Republic never refuted. Partition of the soul common to the Republic with the Timaeus, while absent from the Phaedo. Subtle logical distinc- tions begin with the Phaedo and Republic. State- ments in @ very condensed form requiring logical training 282 ORIGIN AND GROWTA OF PLATO'S LOGIC uncertain allusions found in the works of others, instead of explaining him from his own writings. There is much to show that, though the method in the first books of the Republic is avowedly elementary, the threefold partition of the soul represents a later stage than the Phaedo. This has been best proved by Schultess** (p. 55), whose arguments have never been refuted. The theory of three parts of the soul, maintained by Plato in the Tvmaeus, is a later theory than the simplicity of the soul affirmed in the Phaedo, and could not be left out of consideration in the Phaedo if Plato professed it at that time. We havein the tenth book of the Republic a sample of the manner in which Plato deals with this subject afterwards. Though he speaks of the immortality of the soul generally, he adds there expressly that the true nature of the soul, its multiplicity or simplicity, will best be seen in the next life (6124: ror’ dv tis idor avtns thy adnOH dvow, site modveldns elite wovostoys). A similar allusion to the parts of the soul would certainly be found in the Phaedo, if the Phaedo had been written after the first books of the Republic. The later date of this work is also seen in another peculiarity of Plato’s later writings, already visible in the Phaedo but further developed in the Republic and even later. Plato takes every possible opportunity to establish subtle logical distinctions in which we may discern the trace of his oral teaching in the Academy. He is de- lighted to bring such distinctions into a very concise form, which requires an explanation and is repeated afterwards. Any unprejudiced reader will recognise that a phrase like: ‘éca y’ gor) rovaidra ola slyai tov, Ta pay Trova atta movov Twos gory, Ta 8 avTa sxacTa avTov sxdorou povov’ (438 B, repeated 438 D) requires some logical training to be understood at first reading. Such phrases would be vainly sought for even in the Cratylus or the Symposium, and they are far above the sophisms of the Huthydemus. 28 Fritz Schultess, Platonische Forschungen, Bonn 1875. MIDDLE PLATONISM : REPUBLIC II~IV 283 The Platonic Socrates delivers this logical riddle as if it were something quite natural, but Plato’s experience as a teacher showed him that it was too difficult for the ordinary reader, and Glaucon answers at once that he does not understand, in order to get the necessary ex- planation from Socrates. Socrates explains by a number of examples that correlated terms remain correlated after the addition of a qualification to each of them. If a science is the science of a knowledge, then mathematical sclence will be the science of mathematical knowledge. Plato pushes his caution so far as to observe that the qualification of both terms need not consist in the same word, as for instance the science of health is not healthy. After this lengthy explanation he repeats his logical theorem almost in the same words, and concludes with another example, until Glaucon is satisfied and acknow- ledges himself to have understood (438 £). This digression was not indispensable to the progress of the argument, and appears to have been introduced not to meet ob- jections really made by somebody, but only as a result of Plato’s increasing fondness for logic, and his experience about wrong inferences from dictum simpliciter (amas 438 B) ad dictum secundum quid, a sophism exemplified already in the Huthydemus, but treated methodically for the first time in the Republic. A similar logical digression gives us the method of exclusion or of remainders, by which one part of a whole is investigated through elimination of the other parts (428.4: @orep GAXwv twav TeTTdpwr, ei iw te entodpev aviay ay orpoty ... i Ta tpla mporepoy syvwploaper, avT@ av TovTp syvwpioro TO fnrovpevoyv). This is here introduced as leading to the definition of justice after separating from the general notion of virtue the three other virtues which together with justice constitute, according to Plato, the whole of virtue, namely temper- ance, courage, and wisdom. But if we look at the end of the discussion we see that the method of exclusion to be un- derstood, and ex- plained by examples, then repeated. This pro- duces di- gressions not indis- pensable to the progress of argument, and shows increased interest in logic. Method of exclusion intro- duced for the pur- pose of a definition of justice, then not used, be- cause at the end justice appears to be the general source of other virtues, not co- ordinated to the three other virtues. In earlier works wis- dom had the first place, now given to justice, as also in the first book of the Laws. 284 ORIGIN AND GROWTH OF PLATO'S LOGIC has not been applied to the particular case for which it was introduced, because when justice appears at last, it 1s not discovered as the remaining part of virtue. After the elucidation of the three virtues corresponding to the three parts of the soul and to the three classes of citizens, Plato pretends to be still in the dark about justice (4820: SvcBaros ye tis 0 Toros galverar Kai émiaktos* katt yoby cxotewos kal Svadtepevvntos) and takes this opportunity to invent one word and to use another in a new meaning for describing this special darkness. It is the same laborious play as later in the Parmenides : justice is found not as a virtue co-ordinated to the three others, but as the source of them (433 B: O maou éxewors THY Svvayw Tapéoyev, Bore eyyevéoOar). Here also we find a point of view in advance of the Phaedo, in which wisdom was the chief virtue, and every other virtue to be exchanged for wisdom (Phaedo 69 B). The prevalence of wisdom is proper to the earlier thought of Plato, as we see in the Protagoras (352 D, cf. 357 c) and Euthydemus (282 4). In the Symposium likewise the first place is given to ¢pdvnows (209A: yuy7 mpoonxe texety gpovnow te kal THhv aGdXnv aperyv), and it is a new de- parture in the Republic to recognise the peculiar position of justice as a link between all other virtues. This view, maintained also in the first book of the Laws (63l1c: & (ppovicews xal awppoovvns) per’ dvdpelas xpabévtav tpitov av sin Siaavocvvn . . . Tdv Oelwy ayabdv), is the later view of Plato, while in his earlier works justice was only a part of virtue, co-ordinate with holiness or temperance (Prot. 329 c). In the Meno (79D: muy rolvuy pndé od ire Cnroupévns aperns SrAns & te sot olov dua tav Tavrns poplwv arroxpivopevos Snracew abthv oTpody, } aXXo OTLOby TOUT® TH avT@ rpoT@ Asywov) the identification of virtue with justice is even expressly denied, while already in the first book of the Republic justice appears to be the essence of virtue (853 E : dpetny yuyfs Sicacoovvnv—335 C : Sixauo- auvn avOpwireia apery), & position which seems to have been MIDDLE PLATONISM: REPUBLIC II-IV 285 again modified in favour of vods and dpovnors in the Tumaeus and the later books of the Laws. We may admit that the increasing importance of justice in the Platonic ethics is one of the practical results of the theory of ideas, which required at the summit of Being an idéa ayaGod, prepared already in the Symposiwm (212 a) and in the Phaedo (99c). Also in the second book of the Republic we meet the conception of good as a self- sufficient aim (857B: towvde re ayabov, d SeEaivel? av iyew ov THY atroRawovtwr epiémevol, GAX* avTO avTod EvEexa domratouevot), Closely related to that of Aristotle in his Ethics. For the date of this part of the Republic as coming next after the Phaedo and the preceding dialogues, we find some other hints which it will be sufficient to mention briefly : 1. Speech as an imitation of thought (382BC: to ye év rois Adyous wind Te TOU ev TH Wyn sort TaOnpaTos Kai datepov yeyovos eidwdXov) seems to refer to the Cratylus (430 B : dvowa plunua Tov mpdypyarTos). 2. émrvotnun is opposed to dofa (4444: codiav thy éristatovoay TavTn TH mpaka. eriotnpny ... dpadlay .. Sofav) as in the Meno (86 A, cf. 97.c,988B). In the Meno the distinction is introduced as new, and in the Republic it is assumed to be generally known. That Plato again in the Republic also currently uses ériornun in a primitive meaning, equivalent to réyvn, signifies nothing, because a careful fixity of terminology was not yet acquired by Plato, as we see even later in B. V—VII. 3. God is free from error and lying (882 D: romrns pay dpa vevdns dv Oe@ ove tu. . . HE: mdavry dpa ayevdes To Saydvov te Kad 7o Oeiov). This agrees with the Cratylus (488 Cc: ote évavria dy ériero abtos atte 6 Oels, dv Saluov ris } Oeds ;), only here the unity and simplicity of God is insisted upon, which marks an advance beyond the traditional polytheism of earlier dialogues, which still survives in some expressions (38810: advvarov Oe sOérevv Import- ance of justice results from the theory of ideas. Relation to the Phaedo. Speech and thought. Know- . ledge and opinion well dis- tinguished though the terms changed. Simplicity and unity of God, along with incidental mention of tra- ditional gods. Doctrine of one God peculiar to later Pla- tonism. Division of labour. Change of attitude towards the poets is defini- tive, and remains up to Plato’s latest age. No recon- ciliation possible, and thus 286 ORIGIN AND GROWTH OF PLATO'S LOGIC autov adXoLoby, GAN’, ws Fouxe, KANAOTOS Kal Apiotos wv es To duvatov &xaaTos avTay péver dei amr@s ev TH avTod popoy). But an occasional mention of more than one god, occurring in a criticism of traditional polytheism, is no evidence against Plato’s progress towards mono- theism, as we see from other passages in which 6 20s is used in a monotheistic sense (882 E: o Oeds dou Kat adnlas ty te tpyp Kal gv AGyo, Kal ovte adros peOlotatar outs dAXous éEatrara ; also 379 C : 0 Osos, arrecdy dyads, ... tav ayabav aitios, and elsewhere 379 4, 380D, etc. Cf. Phaedo 62c: @e0s). The doctrine of one God, a perfect Being, developed in the Republic, is adhered to in the Timaeus and Laws, while in earlier dialogues up to the Symposium a plurality of gods is either tacitly implied or expressly admitted. 4. A curious contradiction to a statement .of the Symposium is contained in the principle ‘one man one work’ (894: els geaoros Sv wey av erirndevya xaros étrern- Sevou, oN 8’ ov) when applied specially to the production and acting of comedy and tragedy (395 A: ovd2 ra Soxobvra dyyls GdAnAev eivar v0 pinata Svvavrat of avtol dua ew pipeicOat, olov nwppdiay Kal Tpaywdiay mo.odvres), while in the Symposium Socrates is made to prove the identity of the comic and tragic poet (223D). This discrepancy is in close relation to the change of Plato’s attitude towards the poets. While in the Symposium the tragic poet and the comic poet are represented as friends of Socrates ; and Homer and Hesiod, as deserving immortal fame, are placed in one line with Lycurgus and Solon (209p 2), Plato now despises poetry as a mere piwnovs and banishes Homer from his state. It 1s strange that some erudite critics, who readily believe in an irreconcilable enmity between Plato and Isocrates, and take such a supposition for a firm basis of Platonic chronology, at the same time admit the possibility of Plato’s reconciliation with the poets, which would have taken place if the Symposium were written after the Republic or Phaedrus. It is much MIDDLE PLATONISM: REPUBLIC II-IV 287 less probable that a philosopher like Plato should remain all his life hostile to a living man, than that he should become untrue to fundamental principles once recognised and repeatedly urged. We know from the last books of the Laws (9418, 967 c, cf. 890 4, 964c, and many other passages) that Plato up to his latest age thought poets dangerous, and we have no reason whatever to believe that he changed his opinion after he had written the Republic. Thence it results that the Republic, at least from B. II onwards, must be later than the Sym- posium. 5. While in the Symposiwm the educational influence of Beauty began with the love of beautiful bodies (210 a), in the Republic harmony and rhythm are acknowledged to be the chief factors in education (400 p-401 c) and are said to creep into the soul unobserved (401c). If we remember that the same view recurs in the Laws (665 £) and Timaeus (47D), it will be easy to recognise that also in this respect the Republic is later than the Symposium. 6. The purification of the senses (411 D: d:axadacpo- pévov Tov aicOncewr) 18 & Very concise term scarcely used before the Phaedo, where the necessity of such a purifica- tion is explained at length. 7. The love of the Beauty of the soul (402 D) is here mentioned as entirely independent of corporeal Beauty, while in the Symposium (2108) such a love is a higher degree to which the pupil is led, after beginning with the love of physical Beauty. In the Symposium sensual love as a lower degree is almost excused, and here we find it absolutely condemned (403 8B). 8. Although the method of exposition is a popular one and not based on the theory of ideas, in some passages terms first explained in the Symposium and Phaedo are employed as if they were familiar. This occurs apparently against the author’s intention, but furnishes us with a valuable evidence against Krohn’s opinion that the theory of ideas was entirely ignored by the author of the first the Sym- posium must have been earlier. Educa- tional im- portance of har- mony and rhythm. Purifica- tion of senses. Love of the Beauty of the soul. Terms taken from the theory of ideas, occurring probably against the author’s intention, betray later date of com- position. Definition of courage compared with that of the Laches. Praise of justice in the Republic. Caution necessary with rhe- torical figures. Relation of the Republic to Aris- tophanes’ Ecclesia- susae not justified. 288 ORIGIN AND GROWTH OF PLATO'S LOGIC books of the Republic. This would leave no room for a distinction between aidta Ta Tis cwppoovuns 215 (402 Cc) and eixdvas avtov, nor for card On dv TO eld5ee oporo- yoovra éxelvors cai Eunhovotvta, ToD avToU METEeY OVTA TUTOU, as axddNotoy Oéapa Ta Suvapeve@ OeaoOa (402d). This power of superhuman vision here invoked is certainly the same which we know from the Symposium and Phaedo. Nor are traces of the theory of ideas limited to these passages. We read also ri 7’ totw avTo xa abto (358 B)—avro Sixatoovvny (863 4)—adro 8 éxarepov TH avTod Suvdper av 7H Tov EyovTos wuy7 (366 E). 9. It need scarcely be added that the definition of courage (430 B: duvapus cai cwrnpla Sa travros So&ns opOijs Ts Kal vopisou Sear téps Kal un), which has been held by an eminent critic to be earlier than the Laches because of the promise to treat this subject again (430 c), shows a marked advance beyond the discussion on courage in the Laches. And the supposition that the Gorgias is later because Glaucon says that nobody has as yet praised justice as it deserves (358 D) is likewise based on a mis- conception. The Gorgias cannot be looked upon by Plato at this stage as an adequate encomium on justice, because it deals with the more special question whether to suffer wrongs is better than to do them, not to dwell on the absurdity of drawing matter-of-fact inferences from a rhetorical figure. Such assertions as that about the praise of love in the Symposium or the praise of justice in the Republic cannot be taken literally; any more than Isocrates’ saying in the Ewagoras that nobody before him has written an encomium on a living man. The above considerations fully confirm the conclusions about the date of the first part of the Republic which resulted from our study of style. There cannot be the smallest doubt that the first part of the Republic, except the first book which is probably earlier, was written after the Symposium and Phaedo, and that therefore it is impossible to admit that Aristophanes in 391, when he MIDDLE PLATONISM: REPUBLIC II-IV 289 produced his Ecclesiazusae, meant Plato’s (IV. 424 4) short allusion to the community of wives, or his later exposition in B. V. If this comedy were a parody of Plato’s Republic, then Plato would not have represented Aristophanes a few years afterwards with all the sym- pathy and friendship which are evident in the Symposium. It 1s a strange inconsequence to believe that Plato on one side would feel a lifelong resentment for the insignificant attacks of Isocrates, and then to represent him as in- different to a ribald parody of his most cherished ideals. Between equality of women and the rule of women there is a great difference. If Plato in the Timaeus (18c Dp) and Aristotle in his Politics (1266 a 34) both clearly say that Plato was the first, and according to Aristotle the only writer, who advocated community of wives, then it is evident that neither Plato nor Aristotle recognised the similarity which some modern critics have seen between the absurd caricature of mad women in the Ecclesia- zusae and the plea for equality of sexes brought forward by Plato as the result of his meditations. The chief point for Plato was the unity of the state and the equality of the sexes. He was no advocate of abnormal sexual relations. The progress of mankind has not con- firmed Plato’s view, but his opinion cannot have been such an absurdity in the eyes of Aristophanes as it appeared to some modern readers. That the conception of a community of wives, on which Plato laid no special stress, was not a wholly novel conception, we see from a fragment of Euripides (quoted by Clem. Alex. Strom. vi. p. 751). The coincidences quoted between the Ecclestazusae and the Republic refer chiefly to the fifth book, and are not very striking. The subject need not be further dis- cussed, as all consideration of it is precluded by the date of the Republic, which is placed after 385 B.c. according to our comparisons of style as well as of logical theories. So long as it is supposed that the U Otherwise we should have to change also the date of the Sym- posrvum. Similarity between Republic and Eccle- stazusae not Lecog- nised by Aristotle or by Plato. Commu- nity of wives suggested before Plato. Coinci- dences between Aristo- phanes and Plato irrelevant. Fxagge- sated im- portance has been given to them. External allusions less cer- tain than the result of a com parative study of style and contents. Interrup- tion of the classifica- tion of constitu- tions by the ques- tion about position of women. This form of intro- ducing a new sub- ject might be inten- tional, or indicate 290 ORIGIN AND GROWTH OF PLATO'S LOGIC Ecclestazusae were produced 391 B.c., there is no possi- bility whatever of admitting that they refer to Plato’s Republic. And if some eminent writers accepted this supposed relation, they acted lke Schoene and Teich- muller in the question of style: giving an exaggerated importance to a single observation of doubtful value. It is an error of method to rely upon uncertain external allusions more than on the study of contents or style. If our information seems to involve contradictions, we must carefully weigh against each other the evidence in favour of both contradictory views. We have seen above a great number of sound arguments proving that the Republic is later than the Phaedo in style and contents. This gives us a consistent view of Plato’s evolution which cannot be overthrown by the very uncertain supposition that a play in which Plato is not at all mentioned, written by one of Plato’s friends, could be intended as an attack on Plato’s greatest work. Booxs V-VII At the beginning of the fifth book Adeimantos inter- rupts Socrates’ classification of constitutions by a question about the position of women in the ideal Republic. The thread of the argument here interrupted is resumed only in B. VIII, and thus B. V-VII form a natural division of the whole and deserve to be considered apart. The view has been advanced that a more important division begins towards the end of B. V, p. 47lc, where the question of the rule of philosophers is raised, which fills the whole of B. VI-VII, offering many opportunities for logical reflections. But the transition from the particulars dealt with in the first part of B. V to problems of the highest philosophy is made quite plausible and natural, while the interruption at the beginning of B. V might be intentional and made in order to attract the reader’s special attention to the MIDDLE PLATONISM: REPUBLIC YV—VII 291 new subject, by the rhetorical artifice of an apparently un- expected difficulty. The subjects dealt with in B. V-VII belong to the plan of the whole, and are not an afterthought, though this part of the Republic, if we trust stylistic comparisons, seems to have been completed somewhat later than the following books. If it is once recognised, as it must be on the authority of the same evidence, that there could not be any considerable distance of time between this part and the preceding fourth book, it becomes almost indifferent whether B. VI-VII were completed later or earlier than B. VITI-X. Admitting that they are probably written after B. [IX and even after B. X, we do not agree for that reason with those who deny the unity of the Republic and the architectonic skill with which the parts of the whole structure are co- ordinated. The Platonic Republic would not be complete without the rule of philosophers, and it is irrelevant whether the explanation of this condition of the ideal state is better dealt with before or after the investigation of imperfect governments. As it stands, it crowns the picture of the ideal state and prepares the way for a representation of less perfect states. Even the discussion about the equality of sexes and the digression about international limitations of warfare (in B. V) are not out of place as an introduc- tion to the central part of the Republic. These essential peculiarities of the ideal state could be realised only under the rule of philosophers. Thus we are justified in leaving to this part of the Republic the place given to it by Plato, and in limiting our inquiry for the present to the relation between B. V-VII and the preceding, with reference to what has been already proved of earlier writings. The theory of ideas no longer takes the form of an hypothesis, as in the Phaedo, but appears as a well- established truth, and the terms é5os and idéa begin to be used currently to denote ideas, along with the familiar U2 that B. V-VII have been inserted later at that point, though they he- longed to the plan of the whole. The rule of philo- sophers an essential condition of the Platonic Republic. Theory of ideas appears to be familiar, and the terms el30s and idéa are freely used. Probable reference to the Phaedo and Sym- posmum compared with a similar allusion in the Phaedo to earlier ex- position. No fixity of ter- minology. Frequent use of meta- phors. 292 ORIGIN AND GROWTH OF PLATO'S LOGIC terms avo «a0’ avto, or avto, or 8 ati. We have here an idea of beauty (479 a: idday tivd avtod Kaddovs), of each Being (486 D: rot dytos idéay éxdotov), of justice (479 E: avto 76 Sixasov), of injustice (476 A), of the good (505 A, 517 B, 584 c: tod ayabod idéav), and of all other general notions. These ideas remain always the same (479 E: det Kata TavTa wvavTWs bvTa, repeated 484 B), and each of them is the unity of many particulars (507 B: avro 67 Kadov Kal avTo ayaboy Kai ottw tepi TayvtTwy, & ToTE @s TONG éTiOzmev, TAAL av Kat’ idéay play ExdoToU os pias ovons TiOévtes, 6 Eat Exactov mpocayopevopev). This is here stated to have been already frequently repeated (507 a: ra 7’ av rots gumrpocbev pnOévta Kali ardote Hon TodANdKis sipnuéva). Such a reference to the theory of ideas as familiar to Socrates can only allude to the Sym- postum and Phaedo, and is more explicit than the famous designation of the ideas in the Phaedo as ta moAvOpvAnta (100 B), which has appeared to some critics a reason for placing the Phaedo after the Phaedrus and Republic. In the Phaedo the mention ‘& @puvdAvimev ast’ (76 D) does not even necessarily refer to the theory of ideas, but only to the notions of the beautiful, the good, &c.: ‘if the beautiful, the good, and all similar attributes, about which we are always talking, have real existence,’ not: ‘if, as we are always repeating, the good, &c., have real existence.’ In the same way ‘ta moAvOpvAntra’ (100 B) may refer to moral ideas generally, and not to their transcendental existence as substances. But in Rep. V the theory of ideas is manifestly referred to. Moreover, no special stress is laid in the Republic on the separate and independent existence of ideas. The ideas are an object of thought (507 c: rds idéas vozicbai ghayev, opacOas § ov). The relation of things to ideas is still described with the same terms (476 D: avro xal ra dxeivou petéyovra) as in the Phaedo and Symposium, but how careless Plato was about the fixity of terms is evident if we consider that he speaks also of ‘seeing’ the idea of MIDDLE PLATONISM: REPUBLIC V—VII 293 the beautiful (476 B: ot ém avté 16 Kadoy Suvurol tevat Te kat opav xa@’ avro . . . cmdvot ay elev). This 1s obviously a metaphor, which had been used also in the Symposium (210 E: xatowerat re Oavpactov thy diow xadov), and means that the intellectual intuition of ideas is quite as immediate and objective as the sight of visible things. This knowledge of ideas is even much clearer than the ordinary knowledge based on perception (511 0: cadéarepov TO UTO THS TOV StarsyecOas erLatHuNS TOD SVTOS TE Kal VONTOD Gewpovuevov . . ). Plato insists that the ideas are inde- pendent of the senses (532 A: oftw drav tis Ta SiaddéyerOar emiyetpy, dvev Tacav Tev aicOncewy dia TOU Noyou ar’ avo & tatu ExacTov oOppua, Kal wn atrooth, mply dv avtd 6 zotw ayalov avti vonoe AdBy, cf. 537 D), and it seems as if the senses no longer enjoyed even the merit of remember- ing ideas through the similarity of our perception to absolute notions. This marks a development in the direction of pure idealism beyond the Phaedo. The similarity between concrete things and the ideas, how- ever, continues to be maintained (476 C: o «ada pep m@payuata vopitwy, avto de Kdddos pt) vouilwr . . Spovov TM pH Gpotoy GAN’ avTo HyHras elvac @ Zorxev), as the cause of errors, because every idea seems to be many, while it is really one (476.4: qwavrwv Tov eidey trépt o avTos Aoyos, avTO pay 8v Exacrov elvat, TH 58 Tov mpakewy Kal cwpdatov Kal GAAnwV KOWWwria TavTaxod davTatopeva TOANA haiverOat Exacrov). The power or faculty of knowing the ideas as they are is here presented under different names, as yvwun (476 D), yuoots (478 Cc, also 508 E), éamriotiun (478 A), vonows (532 B, 511 E), vois (511 D), rob StardyerOae S¥vayuis (511 B). This variety of vocabulary need not awaken suspicion as to the perfect unity of thought in the theory. It was Plato’s usual manner in that time, to use many names for his new ideas, and he blamed those who stick to names (4544: war’ avrd ro dvopa Simca tod AsexyOdvros THv évayriwowy) as eristics, unable to classify notions accord- \ « TO Know- ledge of ideas clearer than ex- perience of the senses. Similarity between things and ideas a cause of errors. Power of knowing ideas has different names. Variety of vocabu- lary a result of Plato’s position with regard to language. Sub- division of intel- lectual faculties. Intuition of ideas the highest faculty ; it depends on the idea of the Good. Idea of the Good identical with final cause of the Phaedo. Plato shows only the way lead- ing to it. In order to under- 294 ORIGIN AND GROWTH OF PLATO'S LOGIC ing to natural species (4544: 8: 70 pH SvvacOar nat’ edn Sarpovmevor TO Aeyouevoy éricxoretv) and therefore using the art of contradiction (4544: 9 Suvaus Ths dvtiAoyiKns réyvns) inferior to true logic. Apart from the diversity of names it 1s evident that Plato has progressed since his first attempt ata classifica- tion of psychical acts, and that the reason (Noyotixor) of B. IV is now subdivided into several distinct faculties (Suvapes 477 Cc, cf. 443 B, 518 c) among which the highest is the science or vision of ideas, or of true Being (70 dv qavrer@s 477 A, etdtxpivods, ibidem, ovola 525 B, 534 A, &c.). This knowledge is infallible (477 £: dvaydptytov), and is no longer as in the Phaedo based upon an ultimate hypo- thesis as the most probable truth, but upon a principle above every doubt (510 B: dpynv dvuToberop, cf. 511 B: péypr Tov avuTroOérou eri Thy tod TravTos apyny, cf. 533 Cc, 534 B). The knowledge of this principle is not an inference, but an intuition, and Plato constantly uses metaphorical expressions taken from the senses of sight and touch to denote the immediate character of his highest knowledge (ety 511 a, 5830, amrecOar 511 B, wWuyjs dupa 533 D, opay 476 B, OsaaOas 518 ©, &c.). The principle itself, being the foundation of all this highest science, is the idea of the Good (% Tod dyaOob idea péyrotoy udOnpa 505 a), identical with that Sarpovia ioyus mentioned in the Phaedo (99 c) and there held to be beyond the reach of mankind (Phaedo 99 cD: ravrns totepnOnv kal ovr’ avros evpety ovte Tap adXov pabeiv olds TE éyevounr). Now Plato has found it, but he feels unable to show it to his readers (53838 4: oveér’ olds 7° goes axodovbeiv) otherwise than by indicating the method of training, which leads to the evolution of the dialectical faculty. He says enough about his idea of Good to enable modern readers, who have gone through the prescribed training, and are familiar with abstraction, to distinguish what has been said meta- phorically from the abstract meaning of his thoughts. If we wish to understand Plato’s idea of the Good, we MIDDLE PLATONISM: REPUBLIC V-—VII 295 must bear in mind that mythical falsehoods have an edu- cational value (382), and that he was carried off by the novelty and the sublime beauty of his subject into some exaggerations, which he confesses clearly towards the end of the whole logical digression (536c: éredaBopuny ore érralCouev, Kal paddov avrewdpevos elrrov. Néyov yap aya EBrelra ampos didrocodiar, isov mpomndaxiopévny dvatiws, ayavaxrnoas pot Sox nai @oTep Ovuwlels Trois aiTiots omrovdsaoTEpov eitreiv & elrrov). In his indignation at the degraded condition of philosophy, Plato exalted her power and dignity. He does not add, in what particulars this exaggeration was contained, because the trifling correction introduced by this strange confession, nainely the question of the most convenient age for dialectical studies, would not justify his apology. One property, at least, attributed to the idea of Good cannot be taken literally.2” Plato says the idea of Good exceeds even Being itself in power and dignity (5098: ov« ovcias OvTos TOU adyaOov, AAN Ere eréxEwa THs ovoius mpecBeia xal Suvape. vrepéyovtos) and is the first cause of all Being as well as of all knowledge and truth (508E: aitiav § érirtijpns ovoay xa arnOelas os yryvwoKopmevns.. . cf. 509 8). Having thus brought the expectation of his hearers to the highest point, he not only refuses any explanation of the dialectic power which perceives the idea of Good (533 A) but declines even to insist that his view of it is correct (533 a: ovKér’ adfvov toiTo ducxv- pl{erbar, cf. Phaedo 114 pv). Here he employs much rhetorical artifice with the aim of inducing his readers to attempt the long and tedious training which according to his indications leads to this vision of overwhelming Beauty, the idea of Good. But this idea of Good in the Republic, with all its brilliancy and grandeur, cannot be anything else than the final cause depicted in more Kat 329 See Paul Shorey, ‘ The idea of Good in Plato’s Republic: a study in the Logic of Speculative Ethics,’ in vol. i. pp. 188-289 of the Studies nm Classical Philology of the University of Chicago, Chicago 1895. stand it we must distin- guish mythical repre- sentation from reasoning. Exaggera - tion in- evitable and con- fessed. The idea of the Good above Being, as the cause of Being and know- ledge. Nearer ex- planation declined. Some rhetorical artifice used with an edu- cational aim. Ideas if placed above Being could not have a separate existence. Truth sought in thought rather than in reality. Condi- tions of philo- sophical training. Love of all knowledge not of sights or sounds. 296 ORIGIN AND GROWTH OF PLATO'S LOGIC sober language in the Phaedo. That it is raised above all hypotheses as an unconditioned principle means only that since the time when he wrote the Phaedo Plato had grown so much accustomed to his highest hypothesis that it has lost for him every hypothetical character. It had also become more substantial through intimate association with the practical aspirations which now absorbed him. At the same time, if he placed the idea of Good beyond Being, he made a very decisive step towards a return from the conception of the separate and independent existence of ideas. An idea as a necessary notion of every possible conscious mind is not a substance, and yet limits and shapes the existence of substances. We have no sufficient evidence for saying that Plato when he wrote the Re- public had fully realised this truth, but if he did so, he had no need to change anything in his revelations about the idea of Good and the other ideas. His doctrine that truth is rather to be found in thought than in actual life (473 a: gvow tye wpakw rAsEews Frrov adnOeas épar- tecOat, «dv 2 pn te Soxet) is & sign that he went still farther away from his starting point referred to in the Phaedo, that thought is an image of Being. The conditions for an actual development of the faculty by which we see the idea of Good are depicted with glowing eloquence. Not everybody is able to follow the path, even if he has a leader (479 H: tovs avro To Kanrov pn Opavtas, und’ GdA@ ar’ alto ayovTs Suvapévous GrecOar...d0&dfeuv dycouev). A philosopher is born, and when born, he must also be made and have a strong will to develope his innate power (518 c). He has a golden nature (415 a), and loves wisdom and knowledge above everything (475 B: rdv giddcodoy codlas dyjcopmev driOuuntiy elvar, ov ris udv, ths § ob, adda waons—cf. 376 B: td ye Giropabis Kai dirocodpoy tavrév, also Phaedo 82 0 diAouabys is parallel to diAccopycas) ; he is insatiable of every kind of knowledge (475 c). Therein he is opposed to the sight-lover and others who care only for MIDDLE PLATONISM: REPUBLIC V-VII 297 concrete things (476 B). A philosopher betrays already in his childhood the greatest love of justice (486 B: Wuxnyv cxoTav dirocopov xal wn evOvs véov Gyros emicxéeyret, si dpa Sixaia Te Kal Fyepos 7 Svoxowwvyntos Kal aypia), an excellent memory, a great facility of learning, he is generous, kind, truthful, courageous, and temperate (487 A: vce pynpwrv, edpadns, peyarorpenns, evyapts, gdidos te xai Euyyevns adrnteias, Stxacocvvns, avdpstas, swdpoctvns). From his youth upwards he loves truth beyond everything (485 D: rov t® drte didroualh waons aGdnOelas det evOds ex véov 6 Te wddtoTa dpéyecOa). He grows accustomed to consider the whole of the universe in his meditations (486 A: yuy7 pedrAovon Tod Grov Kal mavros ael érropétecOae Yeiov xai avOpwrivov), which reach far beyond the limits of his own time and include the totality of Being (486 a: 4 tidpye Siavoia peyado- mpérea xal Oewpia tmavtTos pev yxpdvov, mdaons 8 ovcias, olov te oles Tovrm péya re Soxeivy elvac tov avOpwrivoy Biov;) whereby human life appears insignificant, and death loses all its terrors (4868). Through all ephemeral appearances he perceives a substance free from changes (485 B: éxeivns tis ovoias THs ast ovans Kal wn TWAaVwp_EVNS Uo yevécews nal gdlopas ... . Kal mdons avtns) and neglects no manifestation of eternal Being, having an open eye for the smallest detail as well as for the whole. His faculty by which he sees the ideas (479 E: aura Exacta Kal dei Kata TavTa @oavTws SvTa) does not impair in any way the exercise of all virtues and the capacity for acquiring practical experience (484 D: éureipia pndev éxsivwy sdrrelrrovras pnd’ ev GAdm pndevi pdpe aperis VorEepovyras). This image of the philosopher is made still more attractive by the contrast to the merely practical ordinary man (476 a) who esteems vulgar opinions (480 4), ignoring the certitude of science. He is dreaming, because he is unable to distinguish concrete things from the ideas, being deceived by their similarity (476 4, cf. Early de- velopment of moral qualities. Good memory, facility of learning. Philo- sophers possess all virtues. Contempt for the limita- tions of human life, which appears to be insigni- ficant, as compared with the total existence of the universe. Philo- sopher contrasted with the practical man who has only blind opinions. Idea of the Good the clear- est in all existence, compared with the sun as the brightest object of sight. Reason and truth produced by the philo- sopher. Highest level of knowledge attainable through highest training. Mathe- matical training required from the 298 ORIGIN AND GROWTH OF PLATO'S LOGIC 534 c). Plato calls such would-be practical persons blind (484 c: 4 odv Soxotai re tudrav Siadépew ot te Gyre Tov bvtos é&cdorou éorepnusvot THs yvocews ; cf. 506 C), their opinions are sophisms (496 a), and if they hit the truth by accident they do it like a blind man following the right road (506 c). The power of the philosopher (511 B: 7 tod &a- AsyecOar Svvapts) is directed towards the idea of the Good which is the clearest idea in existence (518cD: rov évros TO davoratov . . . elvai dapev tayabov). Whatever else Plato says about the idea of Good, as cause of truth, reason, and Being (517 c: dya0od idéa vobv Tapacyouévn ... 509B: Kal To elvai te Kal Thy ovciay im’ éxeivou avrois mpoceiva:), does not exclude the idea of Good from the system of ideas. Something is sacrificed to the defective comparison of the good with the sun, the light with truth (508 a-509 pb). Plato had himself admitted, in agreement with the common psychological experience, that truth and reason are a product of the philosopher’s own activity (490 B : 6ye dvrws duropabys, . . . yevvnoas voov Kat arnBerav, yvotn te Kat addndas tin), and if afterwards for the purpose of drawing a parallel between the material and intellectual world he attributes truth to a power independent of the individual mind, this must be counted among the exaggerations into which he was led by the greatness of the subject. In the whole Platonic doctrine of the ideal philosopher there is a permanent truth embodied: that the highest level of objective knowledge can be reached only by the highest subjective training of the best individuals. Looked at from this point of view, Plato’s indications as to this special training deserve the attention of the logician, and belong really to the logic of Plato. The way of initiation proceeds no longer, as in the Symposium, through esthetical contemplation, but is prepared, as in the Phaedo, by a course of mathe- matical propaedeutics. The power of mathematical ».. GdAnOevav Kai MIDDLE PLATONISM: REPUBLIC V—VII 299 studies in developing abstract thought is illustrated by two fresh examples, taken one from arithmetic and the other from geometry. The identity of units, which is fundamental in arithmetical inquiries, does not exist in our sensual experience, where each unit is different from every other. This identity can only be understood by the action of thought (526 A: dpiOuav av ols ro By too Te ExaoTov Trav Twavtt Kal ovdé cutxpov Siadépov, wopiov Te Fyov év gauT@ ovdév . . . StavonOjvact povov éyywpel, GrdrA(ws 8 ovdauas perayerpitecOar Suvarov). We owe it to the clear- ness of numbers that we distinguish things which to our senses appear confused (524 c: péya pny Kal dys Kat Oputkpov swpa . . . aouyKeyupevoy tr. dea Se THv toUvTOU cadivevay péya av xai opixpov 7 vonars nvayxdoln dein, ov TUYKEXUMEVA GAAG SiwplLapéeva, TovvavTion 1) ’KEiyN). This difference between numerical exactness and the Inexactness of sense perception is the origin of rational inquiry about the nature of quantity (524 c: évredtey molev mparov emrépyerat épecOat juiv, ti ody oT’ goTl TO péya ad xal To oputxpov). A similar difference exists be- tween the material models of geometrical figures and the ideal figures which they represent. Even Daidalos or another most skilful technical genius could never draw or form figures corresponding to our ideal notion of them (529 ©), and it would be ridiculous to make geometrical inferences or to endeavour to learn the truth about geo- metrical properties of figures from such models, and not from the models of ideal figures that exist only in our thought, surpassing in exactness everything visible to the eye. On these examples Plato shows that mathematical studies lead from ever-changing perceptions to the true substance of Being (521 D: pdOnua uyijs odxov ard Tod yvyvopévou ari 70 dv), from the twilight of vulgar experi- ence to the daylight of philosophy (521 Cc: é« vu«repivijs Tivos nptpas eis adnOivny tov dvtos ovcay srdvodoy, fv on girocopiay ddAnO% gdycomev elvar). But the philosopher will not content himself with such a knowledge of mathe- philo- sopher. Mathema- tical units differ widely from the units of sense ex- perience. Difference between ideal geo- metrical figures and their material represen- tation. Study of mathe- matics for philo- sophical training indepen- dently of practical considera- tions. Solid geometry recom- mended. Astro- nomy not limited to observa- tion of the stars. Striking antici- pations of the modern progress of astro- pomy. 300 ORIGIN AND GROWTH OF PLATO’S LOGIC matics as is useful for a practical man; his immediate alm is not any practical application, but theoretical knowledge (525 B). He will push his investigations far enough to understand the nature of quantity, without caring for practical advantages (525 c: gws dv él Oéav THs Tov aplOpay hvoews abixwvtas TH vonae avTH, ovK avs ovd’ mpdcews yap, GAN évexa adris ths yuyns pactwvns te peTaotpodys ao yevécews én’ adjleray TE Kai ovciay... ). Such theoretical studies develope an organ of the soul more valuable than a thousand eyes, because it 1s the only eye which beholds truth (527 D E: év rovros rots pabypacw éxactov dpyavoy te wWvyns éxxabaiperar.. . xpgirroy dv awOnva pupiov oupatwy' move yap avTO addnOea opata). Plato complains that solid geometry was in his times very much behind plane geometry, and believes that it 1s in the power of the state to further such inquiries by honouring them as they deserve (528 B). He recommends also astronomy to the future philosopher, but adds that a philosophical astronomer will not expect very much from mere observation of the stars. He will use the sight of the stars just as a mathematician uses roughly drawn figures with aview to the discovery of general laws. Plato shows here a deep insight into the logical nature of theoretical knowledge. His very words can be applied even to-day to investigations about the possibility of which he could not have a definite idea. When he says that through all the apparent movements the astro- nomer should reach the true velocity and the true orbits and movements of heavenly bodies, and that this can be done only by thought, not by sight (529 p), the modern reader involuntarily remembers how Adams and Leverrier discovered Neptune without the use of a telescope, by following out purely theoretical considerations. When Plato further decides a priort that the movements of the stars must undergo periodical changes and cannot remain always the same (5308: drorov jynoceras Tov vouitovta MIDDLE PLATONISM: REPUBLIC V—VII 301 yiyvecBal tz tadta ae M@oavtws Kai ovdayR ovdav Tapad- AdTTEV, CHa Te ¥yovTa Kai opwpeva), this appears a still more striking example of true physical knowledge acquired by pure thought. But our illusion is destroyed when we read that the details of the movements of the stars are not worth careful search, precisely because they undergo changes. Here the whole distance between Plato’s logic and the modern logic becomes evident. For Plato science could only refer to knowledge, while we have learnt to deal scientifically with probabilities. Plato was perfectly right in holding that absolutely exact knowledge is impossible in astronomy and every other investi- gation of nature. But he was wrong in supposing that therefore these subjects cannot be dealt with scienti- fically. The whole natural science of to-day, though few persons are always aware of it, is a science of approxi- mations and probabilities. We have learnt to estimate the possible amount of our errors, and to reduce them to units of such low degree that we can neglect them. We owe this power chiefly to the infinitesimal calculus, which marks the essential advance of science from Plato’s days to the present epoch of scientific progress. Plato had no instrument for such evaluations, and he therefore could not admit an exact knowledge of astronomy. He went so far as to say that looking up at the stars not only does not exalt the soul, but does not even teach us any- thing, because the soul rises upwards only through inquiries about invisible Being (529 B: od Svvapar dAXo tT vouicat dvw mowiv vpuxiy BAsrav pabnua th éxeivo, 8 av mepi TO Oy Te % Kal TO doparor). The eyes must in no way be esteemed above reason, nor the ears, and Plato despises equally those who believe in learning music by hearing tones and distinguishing them as sharp and flat (531 4). The true theory of music has higher problems to resolve, and studies the harmony of numbers and its reason (53810: émricxoreiy rives Evpdovor Contempt for actual observa- tion carried very far, because Plato was not aware of the pos- sibility of a scientific investiga- tion of probabili- ties. This became possible only through the infi- nitesimal calculus. Only rational inquiry belongs to science for Plato. Even music not studied on tones. Every particular science useful only as introduc- tory to dialectic. First prin- ciples must be best known, and this is the pr vi- lege of dialectic. The dia- lectician is able to give the ultimate reasons of his convic- tions, and refers all hypo- theses to their source, distrust- ing the testimony of the senses. General system of 802 ORIGIN AND GROWTH OF PLATO'S LOGIC apiOuoi nai rlves ob, kat dua ti éxarepor). Such higher music and higher astronomy, making use of the stars and of sound-harmonies only as matter for generalisations which show the unity of the whole, are recommended by Plato as useful in the preparatory training of a philo- sopher (531D). But even such studies are only intro- ductory to dialectic. Mathematicians, astronomers, musicians are only dreaming about true Being; so long as they rely on hypotheses, without being able to give reasons for them, their studies do not deserve the name of true science (5330: ovepmértovor piv rept TO dv, rap 5 advvarov avtais (deity). A true science cannot be based on unknown or un- knowable first principles (533: @ yap apyn pay 6 yr olde, tedevT S& Kal Ta pEerakd 8 ov pn) olde cuprémdextat, Tis PNXaVN THY ToLAVTHY Gporoylay Tore emioTHuny yevecOat ;). Such apparent sciences rest on mutual agreement, while only Dialectic rises above all hypothetical beginnings (5838cD: 9 SiarexTinn péGodos porn ratty topeverat, Tas imrobécas avaipotoa, er’ avriny thy apynv, iva BeBataonrar) up to the absolute principle to which it gives the highest stability. The dialectician seeks the substance of each thing (5384B: dvarextixov xadsis Tov AOoyov éxdoTOV Aap- Bavovra ths ovacias) and conceives himself to know some- thing only in so far as he is able to give reasons for it (5384B: tov uh syovta, Kal? dcop av pn Fyn Oyo avT@ Te wal GrAXrAp Siddvat, Kata TocodToy vobv rept TovTOU ov dyaes Hye). Dialectic, then, or as we should now term it, metaphysic, is at the summit of all other sciences (5384). This summit is reached through the ability of asking questions and answering them (531 £, 5384p), and through using the hypotheses with a full consciousness of their hypothetical character, until the highest principle is found, without any reliance on the testimony of the senses (511 Bc). Plato had then already conceived a general system of human knowledge, including all sciences and uniting them into one whole (537 Cc: ra te yvdnv paOnpara MIDDLE PLATONISM: REPUBLIC V—VII 303 . TUpAaKTEOV Eis TUVOYLY OiKELOTHTOS GAANAWY TeV paOn- patwv Kal ths Tov dvtos dvcews). Only those who are able to perceive the unity of things are dialecticians (537 C: o cuvortixos StadExTLKOs). This picture of the subjective training, which is in- dispensable if the highest objective knowledge is to be attained, betrays a point of view far more advanced than the Symposium, in which the subjective training was also recognised as indispensable, but started not from reason but from esthetical and ethical experience. Though in the Phaedo the importance of mathematics was already accepted, and one highest principle alluded to, we see here a greater certainty manifested as to this highest principle. We find the philosopher enraptured over his discovery ; it was Plato’s own discovery that all the details of existence can be brought into relation to one final cause of the universe. His great predecessor Par- menides had only recognised the unity of the whole, and declared the ‘many’ an illusion. Plato was the first to bridge over that abyss between the one and the many, and his metaphysical discovery is one that has never since been refuted. Plato’s conception of one final aim of the universe, of the connection between the highest idea and the most minute particulars even of sensible experience, remains unchanged after a long progress of particu- lar sciences and of philosophy. This conception he caught sight of in the Symposiwm, declared it beyond his understanding in the Phaedo, and affirmed confidently its existence in the Republic, though he still declined to explain it fully (506 D, 533 4), alleging as one reason that Socrates is unable to give that full explanation, and as another that Glaucon is not yet sufficiently prepared to understand it. But enough is said to enable the modern reader to see that Plato was in full possession of his highest principle when he wrote his Republic. He called it a model contained in the soul (484C: dvapyds dy 17 knowledge based on pure thought. Subjective training of the dialec- tician beginning with mathe- matics and lead- ing to the conception of the final cause of the universe. This con- ception is a dis- covery of Plato remain- ing in all later philo- sophy. Greatest exactness in the highest generali- sation. Compari- son of the idea of the Good with the sun, and of the earth with a cave. Explana- tion why the philo- sopher is liable to err in practical life, though he has a higher knowledge of Being. 304 ORIGIN AND GROWTH OF PLATO'S LOGIC sruy7 tapddevypa), and he required the greatest exactness in the highest generalisations of science (504D E: yeAotov . TOY payioroy wn peyiotas akvody elvat Kal Tas axpl- Betas). Two allegories used by Plato in the Republic to illustrate his thoughts are deservedly famous. The comparison between the sun and the idea of Good is deficient and contradictory, as truth, according to Plato’s own acknowledgment, comes not to us from without like the light of day. But the other allegory in which this world is represented as similar to a cave (514-518) 1s one of the most beautiful and consistent answers of a philosopher to practical people who deride philosophy as useless. Plato here explains why the philosopher, accus- tomed to the most difficult problems of Being, appears at first sight liable to error in practical life, and how he, better than the merely practical man, very soon acquires a certainty in action impossible for those who know only practical life and have never measured the depth of the world of thought. Nearly every image in the allegory of the cave has a deep meaning. We spend our life in chains, being limited in the possibility of our movements, and prevented by our situation from knowing the truth. Those who succeed in liberating themselves from the chains of earthly passion and human ignorance, and explore a world much wider than the cave in which the others are living, have laid on them, according to Plato, the duty of returning among their former companions in misfortune and of instructing them so as to set free as many as possible. They will not be believed at first, and people will laugh at their tales about the beauties of the upper world, and they will sometimes commit slight errors about objects seen in the cave, which are like shadows of the realities above. Their sight, after long dwelling in full daylight, requires some time to get accus- tomed to the darkness of the cave, in order to distinguish the shadows, which to the prisoners appear to be the MIDDLE PLATONISM: REPUBLIC V-—VII 305 highest realities. But once accustomed, the philosopher will judge more correctly than others, even about those shadows, because he knows the realities which produce them, and he has seen the sun of Truth, which does not shine in the cave. This beautiful allegory need not be repeated in all its details, as it may be assumed to be familiar to our readers. It has a very great logical importance, as it shows that for Plato at that time sensible experience was the shadow of the ideas. This is also the only hint which the Republic contains that the ideas might be independent of the human mind and indeed of any existing consciousness. In many passages, as we have seen, the ideas are spoken of as existing in the philosopher’s soul and even as a product of the activity of his thought. It seems that Plato no longer attached such importance to their separate existence, and that he had to a certain extent reconciled himself to the identity of ideas with general notions. The theory of ideas and of the dialectical faculty occupies the largest place in this part of the Republic, while the remaining intellectual faculties are briefly disposed of. The second rank is taken by the mathe- matical knowledge termed here d:dvora (534 A). The difference between this faculty and dialectical knowledge consists in the use of hypotheses (510 8B), which re- main untouched by the mathematician. As such hypo- theses Plato quotes arithmetical properties of numbers and geometrical properties of figures, which are ad- mitted to be the ultimate foundations of mathematical science (510 c). Both d&:dvoca and driotypn are called in one passage vonois and opposed to the inferior faculty of opinion (Sofa 534 A), which is again subdivided into ziotis referring to things and eixacia to images (511 £). It seems that this division, mentioned here only and never again used by Plato, had a purely occasional character and served the purpose of an elaborate parallelism x Once ac- customed to the darkness of the cave, the philo- sopher begins to distin- guish even ap- pearances better than those who never saw the light of truth. Mathe- matical knowled ge has the second rank as compared with dia- lectic. Sub- division of intel- lectual faculties irrelevant. It was in- troduced for the sake of analogy. Not main- tained consis- tently. Also sub- division of opinion into two different faculties has no import- ance. 306 ORIGIN AND GROWTH OF PLATO'S LOGIC between the sense of sight and the intuition of the soul. To correspond to the difference between things and images a division of ideas was wanted, and the mathe- matical figures best corresponded within the ideal world to the images of the physical world. So far the analogy was plausible, but the subdivision of the two chief faculties of opinion and science into four was not justified and is frequently contradicted by Plato in the same text, as he uses Sidvora, vonots, sriornun and diadrextinn Suvauts indifferently one for another. Even in the sixth and seventh books the distinction is by no means con- sistently followed, and in some passages (500 B: rod yé @s ddnOas mpos tots ovat Thy Siavoray FyovTt—511 A: Enrotvrés te avta éxsiva ideiy A ove Aap Ars ida Tis TH Stavola— 529 D: ro dv Tayos.... & 5) ACy@ pay Kal Stavoia Anwra, We 8 ov) Suivora means pure thought, and not the special faculty of mathematical knowledge which had been named d:avora (511 D: Siadvorav Karsiy por Soxets Thy TOV yewperpixnav ... FE.v). Equally irrelevant is the subdivision of opinion (d0£a) into an opinion about things (aéoris 511 E, 534 a), and an opinion about images (sixacia 511 B, 534 a). This division is of no importance and proves only Plato’s fondness for symmetrical dichotomies.“* He never again alludes to these distinctions, and the old bipartition of intellectual activity into opinion and knowledge remains here as in all other works of Plato fundamental. Opinion is intermediate between ignorance and know- ledge (477 B, cf. 478 D), and it refers to what in one respect is being and in another not-being, and appears as intermediate between substance and nothing (478 D: olov dua bv te Kad p11) dv). 2” It has been attempted to find a relation between the four intellectual faculties of the Republic and the degrees of perfection in the Symposium (Carl Boetticher, ‘Eros und Erkenntniss bei Plato in ihrer gegenseitigen Férderung und Ergénzung,’ Jahresbericht des Luisenstddtischen Gymna- siums su Berlin, Ostern 1894), but the exposition is by no means con- vincing. MIDDLE PLATONISM: REPUBLIC V—VII 307 For the first time Plato here investigates the object of opinion as differing both from the object of knowledge and from that of ignorance. While the ideas are the proper object of science, they are not accessible: to opinion, and Plato defines with great logical acuteness what is sus- ceptible of opinion. It is anything that could be other- wise than it 1s (479 a). Wesee here clearly established the difference between accident and substance, opinion and science. This very important logical theory was prepared by the law of contradiction, stated in the Phaedo, where Plato observed that apparent contradictions are found in things but notin ideas (Phaedo 103 8B). But neither in the Phaedo nor in any earlier work had the difference between the object of science and that of opinion been recognised. It is interesting to observe that Plato employs this distinction between accident and substance to justify his conviction of the mental equality between the sexes, wherein he was so much in advance of his own times, and even of the reigning prejudice of our own century. It is one of the deepest thoughts in Plato’s Republic, that the sexual difference is accidental and exterior as compared with individual intellectual differences among men as well as women (454 B-455 a). And this thought is one of the most interesting practical applications of Plato’s logic. Plato thus proclaimed the truth that thought is independent even of such fundamental bodily conditions as the difference of sex. Many times later philosophers have been drawn by the strength of appearance to credit organs of our body with pure thought, and thus to destroy the soul’s independence and permanence. Plato had within his limited experience many inducements to admit the popular belief that some part of the body is active in thought. He resisted this temptation and was the first to understand clearly and to affirm confidently that thought is an activity of the invisible, incorporeal soul, which does not need material organs for its exercise. That the body’s only aim is to supply us with sensations x2 Object of opinion : everything that could be other- wise than it is. Recogni- tion of the mental equality between the sexes. Thought indepen- dent of the body, even of the most essential bodily con- ditions. Plato the first to under- stand that thought is an activity of the soul. Not-Being object of ignorance, which is identical with wrong opinion. Relation to the Phaedo in the conception of the ultimate aim of life, above pleasure and even know- ledge. 308 ORIGIN AND GROWTH OF PLATO'S LOGIC and to act on the outward world according to our own will, is a truth which remains even to-day incredible to some physiologists unjustly called psychologists. This truth was discovered by Plato and constantly reaffirmed by him, from the Phaedo onwards to his latest works. A consequence of the doctrine that the objects of Opinion and science are not the same led Plato to his theory about Not-Being or Nothing as the object of igno- rance. Ignorance is a state of the soul, and consists in believing what is not (478B: dédvvatov nai §.Eucas 76 py Ov... C: My Svte pny ayvoray & avaynns arredoper, dvte 88 yvouw). Therein ignorance is distinguished from mere opinion and coincides with ‘wrong opinion’ (éd£a arevdys) called also duaOla by Plato (Prot. 358 c: dpaiav ro rowovde Néyere, TO pevds Exetv SdEav Kai ApeboOar repli tav Tpaypatwyv Toy moNXod afiwv, cl. Huthyd. 286 D, and also Theaet. 170 B, Polit. 309 a, Crat. 429D: robdrd:éotw TO yrevdy Adyetv, TO uw) TA Gvia Adyev). Opinion as inter- mediate between ignorance and knowledge had been already mentioned in the Meno and Symposium (2024: To opla Sokdfew . . . tori Te ustakv codias Kai apaias) but then with the predicate of ‘right’ which is dropped here, with an intention of exact terminology not after- wards maintained. In the above exposition of the logical theories con- tained in B. V-VII we had already occasion to see that Plato has here advanced beyond the stage of the Phaedo. But perhaps a more evident proof of this position is found in an ethical hint about the highest aims of life. It was a current theory of earlier dialogues that true happiness is the aim of each individual, and the tale of rewards and punishments after death was in agreement with this conception of the aims of life. Even in the first books of the Republic this was tacitly admitted, and in the ninth book Plato attempts to prove that the philo- sopher is happier than anybody else. Intellectual pleasure or knowledge (¢povnats Phaedo 76 c, 79 D) was the highest MIDDLE PLATONISM: REPUBLIC V_VII 309 ideal of Plato before the Republic. Now he declares that the aim cannot be pleasure, nor even knowledge (505 BC), because there sre bad pleasures, and because the knowledge, if defined, will turn out to be the knowledge of the good. The aim of life is higher than this, and must be clearly known by the leader of men (5058: 8 89 Seanee wev atraca Wuyn Kal Tovtov évexa tavta mpatte). The conception of an aim of life above every kind of pleasure and happiness, even above knowledge and wisdom (509 4), is new, and arises here as a consequence of the new knowledge of ideas and their hierarchy leading to the one highest principle of Being. Some hints show us Plato’s educational experience at the time when he wrote this part of the Republic. He says that the young must be taught through play (537 a: maitovras tpepe), and warns us that no teacher should treat his pupils as slaves (586 E: ovdéy padOnywa pera Sovrgias tov érevOepov xp pavOdvev ... >:uyy Biaov ovdav Eupovov pdOnua) because knowledge is never durable if imposed by violence. Plato is so confident as to the power of youth that he credits the young with the greatest labours and undertakings (536D: véwy mdvres of weyadot Kat of rroAAoi Tovor), but he has already experienced the logical abuses of youth, which he complains of later in the dialectical dialogues. Young men are not serious in reasoning, and delight in contradictions, playing with the argument like young dogs with our clothes (539 B). Here again, asin the Phuedo, Plato sees the origin of scepticism in the abuse of reasoning : Phaedo 908: emedav ris mo- revon Ady Twi dAnOei elvar . . . Kdrera GAlyov Uarepov ara ddfn Wevdns elvac . . . xai avis érepos Kat €repos . + . TreAevra@vres olovrat . ovdéy byes odd€ BESatov. Rep. 589 Bc: drav moAXovs pev atrot éAdyfwow, ind modd\av Be édeyxGaor, oddpa nat raxyv ép- nixrovow eis rd pydev nyeioda dbvrep mpdrepov. It is characteristic that this abuse was explained in a general and somewhat lengthy way in the Phaedo, while Traces of teaching activity. Judgment about youth. Logical abuse leading to scepti- cism. Plato’s view of youth. Simi- larity of the soul to the ideas. Relation to the Sym- posium illustrated by com- parisons. Progress beyond earlier dialogues. Milder view 310 ORIGIN AND GROWTH OF PLATO'S LOGIC here it is briefly mentioned as well known (5398 : olwasce ou Aeh7/Oévas) and attributed specially to the young, which confirms the impression that Plato was grown older, as in his latest works he frequently speaks of the inconsequence of youth (cf. Phil. 15 Dx). Some minor coincidences between this part of the Republic and the earlier dialogues may be briefly mentioned. The affinity of the soul to the ideas, affirmed in the Phaedo, is here shortly referred to (4908B: avtod 6 koTw sxdorov ths diaevs GryacOa. & mpoone wuyhs ébdarrtecOar tod rotovrov’ mpoonxes O# Evyyeve’), With the difference that according to the new division of faculties only a part of the soul is distinguished by this affinity. The metaphor Arye wdivos, used in this passage to describe the suffering of a soul in search of the Truth, would be scarcely natural in this abridged and familiar form if the theory of intellectual fecundity in the Symposiwm were not assumed as known (Symposium 209 a). A similar allusion to the Symposium appears in the assertion of the fewness of those who are able to seek the idea of Beauty, and to follow when they are led to it: Symp. 210 a: de rdv dpbas Rep.476c: 6 cada pev rpdypara lovra éml rovro ro mpGypa .-. vopifwv, ard 38€ kdddAos pire edv dp6as nyjras 6 nyovpevos vopifwy pyre, dv ms yyirat én - kaTavongat ort rd KdAAOosS Thy yraow abrov, Surduevos ére- 76 énl dr@oiy gopari Te éni éréep~ Bat, dvap 4} Umap Bonei wor Cv; oopare adeAddv cor... It would be useless to enumerate all such hints, which become convincing to anybody who reads the dialogues in the order now proposed. Only a boundless indiffer- ence to the philosophical contents of Plato’s works could allow the supposition that Plato wrote the Republic about the same time as the Huthydemus, while in every respect we find here a thought more mature, and a positive philosophy which was only a desideratum when he disputed with the Sophists. He now no longer appears so anxious about the bad influence of bad teachers generally, because MIDDLE PLATONISM: REPUBLIC VIII-IX S3ll he has found in the fundamental differences of human nature a deeper reason for the natural evolution of states as well as individuals. A weak mind is not capable either of great virtues or of great crimes (491 E: dodevy 58 vow peydrwy ovTe ayab@yv ovTE KAKa@V aiTiav ToT? ZoecOar). He denies that the Sophists could have the power of perverting their pupils (492 a). The eloquent picture of the influence of impersonal public opinion on a young man (492 BC) reveals an author who is himself very much above these dangers, and no longer in the first stage of his activity. All this agrees perfectly with our supposition that Plato was approaching the age of fifty when he wrote about the future reign of philosophers over the world. Booxs VIII-IX A strange contrast to the preceding digression is formed by the two next books, which resume the con- tinuation of the fourth book broken off at the beginning of B. V, and except the recapitulation at the outset contain no direct allusion to B. V-VII. The contents of B. VIII-IX are chiefly political, and give a peculiar application of the classification of human faculties to the classification of states and the demonstration of the happiness of the philosopher. The philosopher has a better experience of the pleasures of other men than they can have of the pleasures of knowledge, and he alone is competent to compare different feelings and to judge which of them gives the most satisfaction. Thence it results that he must be believed when he affirms that the pleasure of knowledge is the highest of all human plea- sures (580 D-583 a). This demonstration, repeated after- wards by Aristotle (Hthica Nic. X. vii.), 1s here stated with a certain insistence, and might appear superfluous after what has been said in the seventh book on pleasure as utterly indifferent to the true aims of life. While in the preceding books only contempt is ex- of the Sophists. Influence of public opinion on youth. Classifi- cation of states. Happi- ness of philo- sopher demon- strated by his ex- perience and his compe- tency. True opinions more appre- ciated but always opposed to know- ledge. Ideas more truly existent than bodies and can be better known. B. V1II- IX a con- tinuation of B. I-IV. In the tenth 312 ORIGIN AND GROWTH OF PLATO'S LOGIC pressed for mere opinion as opposed to science, here true opinion and science are placed together almost as if they were synonyms (585 c). This shows not a difference of views, but a difference of exposition. The opposition of opinion and science was already so faimiliar to Plato when he wrote the Republic that he did not always insist upon it in his most popular writings, and the eighth and ninth books are from the nature of the subject-matter very much more popular than the sixth and seventh. The author’s own aim was always pure and certain science which he valued above mere opinion; but he recognised the value of right opinion above ignorance, as he had done already in the Meno. In the same passage in which he puts right opinion along with science as opposed to sensuous gratification, he makes a direct and unmistakable allusion to the theory of ideas, and even to the special account of it given in the Phaedo: Phaedo 80 B: r@ beim cai dbava- tT... kat det @oauTws kara ravrd éxovre éaur@ dpotdrarov elvac ux}... 77 a: mavra ra rowir’ ? € cr , etvat @s olov re padtora... Rep. 585 c: rd rov dei opoiov ps > o . »* é . 3 ‘ €xopevoy cai aBavarov kai adneias kal avré rowodrov by Kal ev rowovTe ytyvopevov, parrAov elvai aoe Soxet ; Another allusion to earlier expositions is the assertion that what continually changes is less susceptible of knowledge and truth than the eternal (585 D: ta mepi rny rod oapatos Oeparelay yévn THY yevav ad tev Tepl THY Tis uyis Oepanelay }rrov adnOeias re nat ovolas pereyet). Generally this part of the Republic is not only formally but also in its philosophical contents a continuation of the fourth book, and seems not to refer in any way, unless perhaps at the end of B. IX, to the high metaphysical speculations of the immediately preceding sixth and seventh books. Book X This last part of the Republic is introduced at first as a supplement to the judgment on the poets proffered in MIDDLE PLATONISM: REPUBLIC X 3135 the second and third books. Plato seems to defend him- self against some polemical attacks on his severe criticism of poetry, and he gives a deeper justification of his con- tempt by a general definition of art asan imitation. This part of the tenth book has its peculiar place in the history of esthetics ; we are here concerned only with the logical theories alluded to in connection with other pursuits. We see here the theory of ideas treated as familiar to all readers (596 a: eldos qov te &y ExacTov eimOayev TiOzcOar méol Exucta Ta TOAAG Ols TavToY dvoua éridéepopev). But in the formulation of this method we perceive a stage of the theory unknown from earlier works. Heretofore, only general mathematical, esthetical, and ethical notions were ideas. There is no trace whatever in preceding parts of the Republic (except in the allegory of the cave, which may have been written later), nor in the Phaedo and Symposium, of ideas of manufactured things, or of any and every group of things bearing one name. Then the ideas were contemplated, known or found as existent. Now they are posited (ri@ec@a:); this term has been applied earlier to names (as, for instance, Crat. 384 D), but never to ideas, though a distinction of species was posited in the Phaedo (79 A: Oauev dv0 edn Tov dvTwr). Here also we might at first suppose that e/dos means only species, as in the similar passage of the Phaedo, but in what immediately follows the word iééa 1s used in its unmistakable technical meaning (596 B: uia idéa), and applied to a table or a chair. Thus it appears that ideas of manufactured articles are admitted. Also the popular objection to the unity of ideas is dealt with, namely, the supposition that the same process which leads to the positing of one idea could be repeated indefinitely, producing an infinity of ideas of the same thing. Plato says that God being the creator of ideas, either his will or some other necessity—of course a logical necessity—prevented the possibility of a plurality of identical ideas (597 c). This logical necessity is further book esthetical consider- ations based on the fami- liar use of ideas. Ideas of manu- factured things appear for the first time, thus initiating a change in the primitive form of the theory. Infinity of identi- cal ideas denied on the same grounds as Aris- totle justi- fied the simplicity of per- ception. One idea of each thing. God as maker of ideas is & metaphor, meaning the logical necessity. New proof of immor- tality, begin- ning by a general statement of the conditions of inde- structi- bility. A class of inde- structible things is shown to 314 ORIGIN AND GROWTH OF PLATO'S LOGIC explained exactly in the same way in which Aristotle afterwards justified the simplicity of perception (Aristot. De anima ii. 2, 425 b 15 sqq.). If there were two ideas of the same thing, then the true idea would be the common type of the two primitive ideas (597 c). This would impair the perfection of ideas, and to avoid it, God, who is not a chairmaker, but the maker of the idea of a chair, made one idea of the chair (597 D: o Ogos, ... piay puoe Fpucer). Now if we consider the deeper meaning of this ex- planation, we recognise a certain advance beyond the Phaedo and perhaps even B. VI-VII. The God who makes the ideas is not the same God who js mentioned in earlier dialogues. God makes the ideas—this is a metaphorical expression which translated into abstract speech means: the ideas are a product of pure thought-— not necessarily of men, but of a thinking subject. This is a consequent development of the theory about the idea of Good which was the final cause of all other ideas. Now this idea of good is supplanted by God, not by some god nor by a god, but by ‘the God’ (o @e0s). The mono- theism appears well established and a matter of course. Also the immortality of the soul is reaffirmed, and a proof added to those of the Phaedo, which could hardly have been omitted in the Phaedo if Plato had then been in possession of it. In the Phaedo the problem was represented as very difficult and further research invited. Now it is an easy thing (608 D: ody yap yaderrov) to prove that the soul is immortal. The proof is no longer based on the ideas, but on the substantiality of the soul. Kach existing thing has its own virtue and its own evil, and can be destroyed only by its own weakness and evil (609 a). If there is anything in existence which suffers from its own evil, without danger of being destroyed, as metals are by rust, then this substance, if any, is inde- structible (6098). To this description the soul is found to correspond. This kind of proof is the converse MIDDLE PLATONISM: REPUBLIC X 315 of all the proofs given in the Phaedo. There immortality was found as @ property of the soul, through a definition of the idea of the soul. Here Plato begins by consti- tuting a class of indestructible substances, and then shows that the soul belongs to it. We shall see that this new logical expedient is used by Plato also later, and it is certainly superior to the method of the Phaedo. We have here an application of the principle that truth is to be found in thought, that our speculation is always concerned with our own ideas, and not with the things outside. Still, from our ideas we draw inferences about the things, and Plato, after representing immortality as a necessity of thought, goes a step further and concludes that the number of souls in the universe remains invari- able (611 A: évvoets Ste aed dv elev ai avtai’ ovTe yap av mou é\adTTous yévowTo . . . ovTe ad wAgiovs). This simple conclusion, which we shall find again in a later writing, was missed in the Phaedo, and leads to very important consequences. In the Phaedo the unity of the soul was one of its properties because the threefold partition was not yet propused. Now, after the repeated exposition of a division of faculties, the parts of the soul can no longer be ignored (603 A), but Plato defends himself against a misinterpre- tation of his view. The soul is in its true substance not full of contradictory powers (6118). The eternal is simple in its own nature, and cannot be composed out of many elements (611 8B: ov pddioy aldcov elvar ovvOeroy Te &c mroAXov). The partition referred to the imperfect transitory earthly state, not to the soul’s eternal existence. We contemplated it under the modifications produced by union with the body, and failed to perceive its eternal nature. This 1s a manifest correction of the theory of three- fold partition as taught in B. IV and IX, and exempli- fies Plato’s manner of revising his earlier writings. He did not alter anything in what had been written, but he include the soul. Immor- tality as & neces- sity of thought. Conclu- sion about number of souls. Unity of soul de- spite its different parts. Simpli- city of eternal elements when set free from the bonds of the body. Example of revision of earlier writings without any altera- tion in the earlier text. Subject of immor- tality alterna- tively mentioned as new and as already dealt with. Clear allusion to the Phaedo in the tenth book of the Republic. Each dialogue of Plato stands apart, 316 ORIGIN AND GROWTH OF PLATO'S LOGIC adds his correction in the continuation of the same dialogue, just as he added his confession of a certain exaggeration in the picture of the philosopher at the end of the seventh book. This way of correcting and criti- cising his own views confirms our supposition as to the technical difficulties which stood in the way of many changes in the original drafts of Plato’s writings. Some other examples of such self-criticism will appear in later works, and it is exceedingly characteristic that this pro- ceeding begins already with the Republic. Plato’s habit of considering each work in turn as one independent whole is apparent from the fact that the subject of immortality is introduced in B. X as new and never heard of before (08D: ov« qoOnoar Sri dOdiaros Hav ny Wuyn cal ovdéote aroAAvTaL; Kal ds euBrepas por wat Oavyacas elre* Ma Ai’, ove kywye* av 88 Tobr’ Exes Aéyewv ;). Some readers of Plato saw in this passage a proof that the tenth book of the Republic had been written before the Phaedo, without noticing that a few pages later there occurs a perfectly clear allusion to the Phaedo, which cannot refer to any other work of Plato but the Phaedo only. He says (611 8B) : ére roivuv d@dvarov yuxn, xal o dpte Noyos kal of GANOt dvayndoeay av. This means that in an earlier writing there had been given a number of arguments (Acyor) of a logically neces- sary or apodictic character (dvayxd{orres) proving the soul’s immortality. Now a plurality of such arguments is not given in any other work of Plato besides the Phaedo. The Phaedrus, which might be thought of here, contains only one argument, and other dialogues, such as the Meno, Gorgias, &c., do not contain arguments (Adrvyos) but tales (uiOo1, cf. Phaedo 61 B: the poet invents pvOous, GAN’ ov Aoyous, cf. also Gorg. 523 4). That Adoyos is used in the tenth book in the meaning of a logical argu- ment can easily be seen from many passages (611 8B: o Oyos ovK éacet—609 D: droyov—610A: Kata Aoyov, &c.). Thus we see that Plato, even alluding in a general way to MIDDLE PLATONISM: REPUBLIC X 317 his earlier writings, sometimes ignored their particular contents in a new exposition. Each dialogue was meant to stand apart, as if it were written expressly for the new generation of students entering the Academy, or, in the case of the Republic, possibly for a wider circle. The illusory character of sense perception, as repre- sented in the Phaedo and in the earlier books of the Republic, is here maintained (602cC: tavroy wrov piv péyeBos syyibev re nai woppwOev Sia tis Syrews ove ioov daiveras), and is illustrated by a skilful enumeration of optical illusions produced by distance, colouring, and reflection of light. _But the distrust of the senses is no longer so unlimited as in the Phaedo, and is subject to a distinct modification. We have a means of correcting their illusions, says Plato, and this consists in measuring, counting, and weighing (602D: To perpeiy nal dpiOpeiv kat iotdvat BonOerar yapiéotatar . . . wore py apyew 7d dawopmevov . .. ad\dAa TO Aoytoduevov). This intuition of the mathematical power of correcting the illusions of sense seems to be a Pythagorean notion, and betrays also the fact that since the first understanding of the distance between appearance and ideas Plato had been working to bridge it: over partially by physical research. His pro- gramme is constantly realised in our own days, and we witness many subtle corrections of primitive sense illu- sions by the power of number, measure, and weighing. This power of correcting the illusions of the senses is ascribed to the cognitive faculty, which is the best part of the soul (6034: ro pérpw ye Kal Aoyiou@ mioTEiov BértoTov dv sin TH Yvyis). The opposition between opinion and knowledge thus alone remains out of the whole fabric of the four sub- divisions of the cognitive faculty in B. VI-VII. And opinion is here more sharply distinguished from know- ledge than ever; it becomes quite another part of the soul, like feeling or desire (6034: To mapa ta pétpa Sofdlov rhs yruyhs T@ Kata Ta pétpa ovK dy ein ravtor). though he sometimes refers to earlier ex- position. Dlusory character of the sense per- ception, corrected by mea- suring, counting, and weighing. Physical research bridges over the distance between ideas and appear- ances. Reason corrects errors of the senses. Four sub- divisions of cog- nitive faculty not main- tained. Instability of termin- ology. Law of contra- diction as & law of thought. The user more compe- tent than the maker, as he has the know- ledge. Poets deprived of right opinion : Homer below Pro- tagoras. 318 ORIGIN AND GROWTH OF PLATO'S LOGIC Hence opinion probably will not partake in immortality. The instability of Platonic terminology at the time when he wrote the Republic is seen from the circumstance that even here, where opinion 1s condemned so strongly, the same word, d0£a, is used for both opinion and knowledge, in the meaning of a judgment which might be wrong or right (6028, cf. Theaet. 190 4). Here for the first time occurs a formulation of the law of contradiction as a law of thought, while in the Phaedo and earlier books of the Republic it was a metaphysical law : Phaedo 102 £: Rep.4868: ravrovra-| Rep. 602 8: épaper ovdev rav évavriwy eri | vavria roveivy 7} rdoxev | ro avr@ dua wept ravra by Grep fv dua rovva- | xard rairdv ye Kai mpos | évavria dogdgew advvu- vriov ylyverOai re xai|ravrov otk éOeAnoe| rov eva. Cf. Theaet. elvat. dua. 190 a. This is also an indication of Plato’s advancing logical preoccupation. There are besides other hints of the re- lation of the tenth book of the Republic to earlier dialogues. Here, as in the Cratylus and EHuthydemus, the competent judge about anything is he who makes a proper use of it (601 c: Scmep érictata: ypjoPa) not the maker (601D: WodAn avayen tov “pwpevov éxadotm dutretporutov Te elas, Kad aryyerov yiyverOar t® trownth, ola ayaba fh KaKd trovee év ™ xpela @ ypnta). Here this principle is generalised, while in the Cratylus it was applied specifically to word- making. The opposition between user and maker is parallel to the contrast between knowledge and faith (601 B: 0 pay sid@s eayryédrec wept ypnotay nai Tovnpdav . 0 68 Tio TEVM@Y ToLncE). The poets are now shown to have neither knowledge nor even right opinion (602 A: ore dpa sicetas ove opOd Sokdces 6 puuntns wept wy Av punta: mpos KdddX\os 4 aovnpiav). Homer, who was named in the Symposium as holding the first place among those who deserved immortal fame, is now not only esteemed below Solon and Pythagoras, but even humiliated by comparison with SfYLE AND DATE OF THE REPUBLIC 319 Protagoras and Prodikos (600 c) who succeeded better in life, says Plato, because they had more knowledge than the king of poets. We see here a pitiless condemnation of what had been the chief element in Plato's own education. He knows well the temptations of the poet, and remains still poet enough to degrade poetry with poetical exaggeration. The future writer of the Laws appears here already with his boundless contempt even for the most refined pleasures, asking for deeds not words, choosing rather to deserve praise than to praise others (599 B), and proudly conscious of his own productive activity. The style and date of the Republic. We have found a natural progress of doctrine from the beginning to the end of the Republic, but no such fundamental differences between the first books and their continuation as to make it necessary to recur to such adventurous suppositions as Krohn and Pfleiderer made about the composition of this work, which is remarkable for its unity in spite of its unusual volume. A comparison of contents alone, however, 1s insufficient for a decisive solution of the question, and we must turn to our eilwOvia pé8o0d0s of stylistic differentiation in order to find a trustworthy confirmation of the view resulting from the study of theoretical development. As to the single books of the Republic the point of main significance is the very early style of the first book. This has none of the important peculiarities common to all the following books, neither the scarcity of vai, wdvu ye, wavy wav ovv which characterises B. II-X (these answers form in B. I over one-third of all answers, just as in Charm. Lach. Prot.)\—nor 7té pv; nor addnbécrara, nor op8as, nor opOorata, nor dpGorara Aéyets, nor 8/A0v—all these being important peculiarities charac- terising all the following books, and missed in the first book certainly not by chance, as all the usual opportuni- Unity of the Republic resulting from the study of its contents. Style of the first book very early. Many important peculi- arities common to all later books are missed in the first book. Peculi- arities of later style occurring in the first book are gener- ally found also in earlier dialogues. 320 ORIGIN AND GROWTH OF PLATO’S LOGIC ties for their use were given. Also some important peculiarities which were introduced into the style of Plato in earlier dialogues, and remained up to the latest works, are absent from B. I. Such is for instance the general prevalence of superlatives over positives in all affirmative. answers, common to the Phaedo with nearly all later dialogues and all books of the Republic (325), the great frequency of questions by means of dpa (378) common to the Cratylus with all later dialogues and all other books of the Republic, new-invented adjectives in edns (254), beginning with the Gorgias and frequent in all parts of the Republic except B. I, re singly (231) frequent in all other books and occurring already even in some Socratic dialogues, interrogations asking for better explanation (453), great frequency of prepositions (390, found already in the Laches and common to all other books of the Republic with the latest group), questions by means of motos (353); many other less important peculiarities are absent from the first book, being common to all other parts of the Republic with the latest group and, in the case of the last enumerated, even with some Socratic dialogues. In the above enumeration no accidental peculiarity has been included, and of these a certain number can easily be found in the table of affinity (pp. 162-171), quoted as occurring in various parts of the Republic except in the first book. If now after this long enumeration of peculiarities vainly sought for in the first book we ask what kind of peculiarities of later style are found in it, we find chiefly accidental occurrences of peculiarities known already from the earliest dialogues, and only one unique peculiarity of some importance later than the Phaedo, namely a double occurrence of xa\@s or @ similar adverb without verb in an affirmative answer. This is the only important peculiarity common to all other parts of the Republic, found in the first book and not in dialogues earlier than the Republic. Other important peculiarities of the first STYLE AND DATE OF THE REPUBLIC 321 book belong to an earlier time, as for instance rodyuy in conclusions (284) beginning with the Crito, frequency of apodictic answers equal to that in the Huthydemus (376). and diy without Aéyers as in the Charmides (385). This proves the very early date of the first book, which however may still be as late as the Gorgias. We cannot compare it with the Gorgias, because the latter dialogue is thrice as large, and we have no evidence as to the occurrence of the investigated peculiarities in a part of the Gorgias equal to the first book. That it is earlier than the Cratylus appears very probable if we consider the great difference of the equivalent of affinity, which is sufficiently considerable to include a reasonable allowance for the difference of size : Rep. a (204 pp.) : 28 (I) 6 (II) 8 (III) = 49 (1). —> Crat. (42 pp.) : 88 (I) 16 (II) 15 (III) 1 (IV) =114 (. As here the equivalent of affinity of the later work is over twice as large, and more than proportional to the size, while generally the equivalent of affinity increases less than proportionally to the size, we may fairly infer that the Cratylus is later. This inference is confirmed by the fact that certain peculiarities absent from the first book are sufficiently frequent in the Cratylus to be pre- sumed to exist in all its parts, and therefore also in any part equal in size to Rep. I. Such are ovcia in the meaning of substance (245), adjectives in wns denoting causal relation (275), card with accusative prevailing over all prepositions except év (389), interrogations asking for better explanation (453), new-invented adjectives in dns (255), and great frequency of roivuy (308). Of these peculiarities, all frequent and important in the Cratylus, none is found in the Gorgias except one question asking for better explanation, and therefore they show the later date of the Cratylus, while they cannot be used for a determination of the relation between Gorgias and Rep. I. The Gorgias has only three important peculiarities (253, Y Compa: i- son of the first book with the Craty- lus and with the Gorgias shows thie probable position of the tirst book between these two dialogues. Close relation between second, third, fourth, eighth, and ninth books. B. II-IV and B. VITI-IX differ in the com- 322 ORIGIN AND GROWTH OF PLATO’S LOGIC 307, 377) absent from the first book of the Republic, which happen to be absent also from the Cratylus, and these have generally less importance than those found in the Cratylus and absent from the Gorgias and the first book of the Republic. Thus it may be regarded as probable that the first book is earlier than the Cratylus, while nothing can be said from purely stylistic comparisons about its relation to the Gorgias, to which its contents show it to be subse- quent. As to the following books of the Republic, stylistic comparison proves that there is no such great distance between the first four books and the following as has been sometimes supposed by those who believed in a very early publication of the first four, five. or even six books.4! The style of B. II-IV is not very different from the style of B. VIII-IX, if equal samples are compared. Take for instance B. II-B. III 412 4, slightly exceeding in size the total of B. VITI-IX. Both appear evidently later than the Phaedo, to which they are inferior in size: Phaedo (49 pp.) : 48 (I) 26 (II) 17 (III) 2 (IV) = 154 (1). Rep. b 12 (874 pp.) : 47 (I) 20 (II) 22 (IIT) 2 (IV) = 161. ->Rep. d (84 pp.): 47 (1) 22 (I) 27 (III) 8 (IV) = 184. The advance beyond the Phaedois considerable if we take into account the difference of size, and also the nature of those peculiarities which are common to B. II-IX being absent from the Phaedo. These include nearly all the 731 The separate publication of the first four books has been advocated by Hermann and later by Krohn, Chiappelli (‘Sopra alouni capitoli della vita di Dione di Plutarco,’ Torino 1883, Rivista di jsilologia, anno 12), Siebeck (Jahrbicher fiir Philologie, Band 181, 1885, p. 229), and many others. Pfleiderer laid great stress on the division at 471 0. Teichmiiller, under the influence of the prejudice about the relation of B. V to Aristophanes’ Ecclesiagusae, supposed the first five books to be one whole published about 892 so. Finally Rudolf Kunert (‘Die doppelte Recension des Platonischen Staates,’ Wissenschaftliche Beilage sum Jahresbericht des Kdtniglichen Gymnasiums su Spandau, 1893) believes that B. II-VI form one indivisible whole, published before 390 3.c. STYLE AND DATE OF THE REPUBLIC 323 peculiarities enumerated above as characteristically absent from the first book. Among these the following have a special prominence: ri pyy ; (202), Quuoedyjs as a philo- sophical term (261), ovcody py (338), dAnOecrata, dpbas (342), d7%Xov (843), dpPorara with or without Ayes (342, 888). darevpos (473), peOlorapar (488), all found in both parts of the Republic, but not in the Phaedo. On the other side B. VIII-IX contain not a single new important peculiarity absent from B. II-IV. The advance in style from the earlier to the later part is only due to a greater number of accidental peculiarities, and to an increase of the frequency of all kinds of peculiarities. Thus generally speaking B. VIII-IX belong to the same time as B. II-IV, showing a later style only to such an extent as might be expected in a continuous work of these dimen- sions. We have therefore no stylistic reason whatever to admit a great distance of time between the earlier and the later part, as has been also shown by the comparison of the contents. Naturally this does not imply that both parts must have been written in the same year, or in the same couple of years. Style 1s changing slowly, and even the small advance in style observed may correspond to two or three years, if we allow for the whole of the Republic an average term corresponding to its size, anything between 5-7 years. As to B. V-—VII, there is some stylistic evidence to place it after B. TX, at least its chief part designated in the table of affinity as c, (471 c-541). We find: Rep. d (B. VIII-IX = 84 pp.) : 47 (I) 22 (II) 27 (1II)8 (IV) = 184. _,Rep.c, (471 o-541 = 44 pp.) : 50 (I) 21 (II) 88 (III) 7 (IV) =284(1). The comparison seems at first sight, in view of the differ- ence of size, to be insufficient for chronological purposes. But if we add to B. VITI-IX a part of B. V to increase its size, then we obtain : Rep. c,d (50 pp.) : 62 (I) 28 (IT) 86 (IIT) 8 = 228 (1). —>Rep. 0, (44 pp.) : 50 (I) 21 (II) 88 (III) 7 (IV) = 284(1): Y 2 parative frequency of identi- cal peculi- arities. Both parts of the Republic belong to the same time. The inter- mediate part of the Republic appears to be later. Slight advance of style of B. VI-VII, over B. VITI-IX confirms the suppo- sition of their later date. Relation hetween beginning and end of B. V uncertain. Position of the tenth book remains uncertain, though it is probably the latest. 324 ORIGIN AND GROWTH OF PLATO’S LOGIC a small difference of few units in favour of the smaller sample of text, very significant through the prevalence of important and very important peculiarities. Some pecu- liarities appear in B. V-VII, which are missed alike in B. II-IV and in B. VITI-IX as in all earlier dialogues. These include yap ovv in short answers (826), m7 (332), mas used with dros (375), cata separated from the corre- sponding accusative by re (395), dvtws dv (236), ddw (195), dxivntos (469), and other words of a more accidental character. This confirms our conclusions from the con- tents, and makes it probable that the bulk of B. V-VII has been added later, at least after B. IX. From the observations it is, however, not easy to ascertain whether the beginning of the fifth book forms one indivisible whole with the picture of the philosopher from 471c to the end of the seventh book. The first part of the fifth book, dealing with the equality of sexes, and with international relations between Hellenes and Barbarians, might still be earlier than the eighth book, while the larger portion filling the sixth.and seventh books might have been added later. This point can only be decided by a more minute comparison of a greater number of peculiarities in samples of text absolutely equal. For our purpose it has no importance whatever, as this part of the fifth book contains no contribution to the know- ledge of Plato’s logic. It is equally difficult to decide whether the tenth book is later than all parts of the Republic, or only later than B. VITI-IX. It contains a considerable number of pecu- liarities of later style for its small size, but only three accidental peculiarities are new (438, 475, 478), while B. V-IX contain a greater number of peculiarities which are absent from B. X. But a definitive solution of these difficulties can only be expected from further stylistic research. Meanwhile it remains certain that B. X is later than B. IV, probable that it is later than B. IX, and possible that it is later than all other books of the STYLE AND DATE OF THE REPUBLIC 325 Republic. This possibility, a mere possibility so far as our stylistic comparison reaches, becomes a probability when the contents are carefully considered. All the parts of the Republic, except the first book, being later than the Phaedo, and differing not very much in style among each other, we may conclude that they were written continuously in the time next following the Phaedo, and as we have no reason to suppose that in that time Plato increased the speed of his writing, or the average amount of text produced yearly, it remains probable that the Republic occupied him for about six years, up to nearly his fiftieth year, as we supposed. This refutes all the suppositions about a possibly early date 2 of the Republic, and shows that Plato wrote his great work after his return from the first Sicilian voyage, and after the foundation of the Academy. Chiappelli”! (p. 16) believes that Plato had already formed his political convictions when he vame to Syracuse. If we accept the traditional account of his adventures, the reverse is far more probable: that the personal experience and observa- tion of the consequences resulting from the abuse of tyrannical power gave an opportunity to Plato for political reflections. This may have brought him from a position of individualistic ethics to a socialistic political theory such as is set forth in the Republic. 782 Among all the artificial arguments in favour of an early date of the first books of the Republic, none has been invented with such remarkable imagination as Pfleiderer’s contention, which deserves, for the sake of curiosity, to be here quoted in his own words: ‘nach meiner Ansetzung in den neunziger Jahren des 4** Jahrhunderts feiert Rep. A (=‘5}’ accord- ing to Pfleiderer’s strange designation, or B. I-V 471 c) zugleich das zweihundertjéhrige Jubilium der Gesetzgebung des Solon von 594, welche ja als Leistung seines von ihm so hoch geehrten Verwandten dem Plato Zeitlebens als spornender Vorgang vorgeschwebt’ (Sokrates und Plato, p. 248). Equally bold is Gymnasialdirector Carl Schmelzer (see above, p. 25, note 66) who declares that Plato did not mean seriously his political theories, and that for instance xowda 74 tay piAwy means: ‘es muss jeder Mann eine jede Frau achten und schitzen als sei sie die seinige.’ The Republic written after the Phaedo in the course of about six years. Voyage to Syracuse might have led Plato to reflect on political matters. The Phaedrus shows the conditions of a philo- sophical rhetoric. Plato’s natural eloquence made rhetoric super- fluous for him. He saw the use- fulness of some 326 ORIGIN AND GROWTH OF PLATO’S LOGIC Il. The Phaedrus. (Relative affinity to the latest group, measured on the Laws as unity, =0°81; see above, p. 176.) The Phaedrus, beyond any other work of Plato, has been misunderstood by interpreters who devoted more attention to indifferent details than to the philosophical contents of the dialogue. It has been ascribed to a young man of twenty-five, while it contains notions and theories which Plato could scarcely have advanced before he was fifty. Some critics, and among them Grote, saw in the Phaedrus an erotic dialogue, either supplementary or even preparatory to the Symposium—though the evident aim of the Phaedrus is to establish the conditions of a philosophical rhetoric, chiefly applicable to educational purposes. In the preceding dialogues we have seen Plato rising to the highest principles of knowledge with- out any attempt to reason about the best way of imparting them, except the few precepts given in the Republic. He tacitly assumed that any one possessing knowledge can impart it to others, if they are able to receive it. We may suppose that Plato attracted chiefly very gifted pupils, and to begin with he had such a great power of teaching that he felt no need of rhetorical artifice. His eloquence, which we admire even in such early dialogues as the Apology, was the natural outburst of his genlus progressing spontaneously from the Apology to the Gorgias, from the Gorgias to Symposium, Phaedo, and the dialectical books of the Republic, apparently without effort or study (Phaedr. 2488). This explains why he contemptuously defined rhetoric in the Gorgias as a kind of flattery, and why he condemned tragic poetry in the Republic as an imitation. His first opportunity for noticing the usefulness of some rhetorical artifice must have arisen at a time when his pupils began to teach, and he first observed that some of them, with all the knowledge inherited from the Master, MIDDLE PLATONISM : PHAEDRUS 327 were less capable of imparting it than others. Though we admit some educational activity of Plato before the foundation of the Academy in 387 B.c., the teaching by others under his direction could scarcely have begun earlier, and even probably began later, than the first years of the existence of his own school. When the number of his pupils increased, and some of them had remained with him a longer time, it is natural that the elder pupils should begin to teach; and their deficiencies in teaching may have led Plato to some reflections on rhetoric, which he embodied in the Phaedrus. This view is here not given as a reason for the late date of the Phaedrus, but only as an explanation of the origin of this dialogue, which becomes probable when once we know its late date, as resulting from the study of its style (see above, p. 176). For the purpose of a discussion on rhetoric, Plato had to select a speech as an example to illustrate his views. His choice of a speech of Lysias ** was natural, inasmuch as Lysias was thought one of the greatest rhetors of “88 Much erudition has been spent on the question whether the speech attributed by Plato to Lysias is authentic or only invented by Plato in imitation of other writings of this orator. We have no reason to disbelieve Plato if he clearly credits Lysias with this speech. To criticise his own invention and to accuse Lysias of the greatest moral degradation on the ground of a forged document, would certainly be below Plato’s dignity. That the speech is read by Phaedrus, and not repeated from memory, adds to the probability of its authenticity, which has been maintained also by Haenisch (De oratione quae sub nomine Lysiae in Platonis Phaedro legitur, Ratibor 1825), Spengel, Franz, Westermann, Holscher (quoted by Hermann, p. 675, note 554), L. Schmidt (Verhandlungen der 18° Philologenversammlung, Wien 1858), Ueberweg (Untersuchungen, p. 262), and by many others, while it has been opposed by Hermann and Jowett. A certainty in this question can only be arrived at by very minute stylistic comparison. So long as an evident proof of the spuriousness is not forthcoming, we must admit the authenticity. Plato has never quoted by name an author attributing to him words or opinions which were invented by himself. So far as the works alluded to by Plato are preserved, all his quotations from Homer, Parmenides, Protagoras, have been confirmed, and the natural assumption is, therefore, that he included in the Phaedrus an authentic speech of Lysias. The onus probandi is entirely on the side of those who deny it. rhetorical rules when his pupils began to teach and mani- fested some de- ficiencies in their teachiny. Selection of aspeech of Lysias &8 an example of wrong rhetoric. Choice ‘limited because it had to be a speech to the young. The second speech improves the form, the third speech the contents. Subject- matter secondary, and not limited to love. Widened horizon. 328 ORIGIN AND GROWTH OF PLATO'S LOGIC those times (228 4: Avoias év T1oAA@ yYpovp Kata ayoAHV cuvéOnxev, Seworatos dv Tav viv ypadev). The subject of the speech to be selected was accidental and secondary. It could obviously not be a forensic speech, because Plato’s aim was an investigation of educational rhetoric, not of forensic oratory. He had to choose from speeches which were designed for the young, and it was not his fault that such speeches did not attain a very high moral standard. He could not select a model speech, even if one could be found outside the Socratic circle, because the artistic purpose required a sharp contrast between his rhetoric and the wrong rhetoric of contemporary orators. His choice of a discourse of Lysias, written in apology of illicit sexual relations, must be, therefore, recognised as perfectly fit and proper for the purpose. Before any theoretical discussion followed, a better example had to be opposed to the example taken from [uysias. This better example was at first to be better in the form, and then afterwards to be made better and more elevated in the contents. Plato chose to oppose to the first speech two speeches of his own: the first on the same subject, but better composed—the second directed against the contents of both the preceding speeches. Thus it resulted as a necessity of composition that the three speeches, intended to exemplify the theory, occupied a great part of the whole writing, being nearly equal in size to the remaining dialogue. The three speeches are avowedly examples of good and bad eloquence (262 bv, cf. 264 z). The subject- matter is of secondary importance, and is by no means limited to love, since the myth in the second speech of Socrates deals even more with immortality, reminiscence, and human perfectibility than with the particular subject of love which formed the accidental starting point. We see here in every respect a very much widened horizon ; in the Phaedo the scene of the mythical digression was limited to the earth’s depths and heights, and even in 329 Rep. X the Earth is still the centre of interest. Here we see Plato, in accordance with his recommendation in the Republic as to the study of astronomy, taking the universe up to the fixed stars as the scene for the periodical migrations of each soul. The allegory of the cave is re- peated on a much larger scale. The whole earth now takes the place of that subterraneous dwelling, and instead of the world outside the cave, where Truth can be seen as we see here earth and water, we have now the supramundane region beyond the most distant stars, a metaphorical expression which means beyond space and matter. Those who remember their vision of Truth, and act accordingly, are deemed to be mad (249 D: é&tarduevos Tav avOpwrivev orovdacpdtwv ... vovOeteitat WTO THY TOAA@Y WS Trapa- xwav, cl. Rep. 517 D: & dro Oeiwy Oewpi@v eri Ta avOpwrea tis EAOwy . . . paiverat odhodpa yeroivs), because vulgar people are unable to understand philosophy. While in the Phaedo even the murderer of his father could be pardoned after one year’s punishment (114 4), here, as in the Republic, the period of probation lasts a thousand years after each life on earth, and a free choice of a new fate is left to each soul: MIDDLE PLATONISM: PHAEDRUS Rep. X 615 a: Siyyeioba de GdAnAats . . . dvauipvnoKxopéevas Ooa Te Kat ola maOouy Kai idciev ev T7 urd ys ropeia—eiva d€ Hv 1o- petay yitdtérn—ras 8 av éx rod ovpuvovd evrabeias SinyeioBat nad Geas dunxavous To Kaddos. . . . 617 D: mpodnrny . . . AaBovra xAnpous te nat Biwov rapade’ypara . elmety" suxal ednuepo .. . vyeis Saipova aipnoec Oe. Phaedr. 249 aB: drav roy mpa@rov Biov redeutnoworr, kpioews Ervyoy, - ‘ € b} : LY ¢ ~ kplOeioa b€ ui pev els ra UO ys Sexacwrnpia €AOovoat Sixnv eéxrivov- ow, ai 8 eis rovpavod riva ronoy ¢ ~ ’ rd , bird ris 8ixns Kovgiobeioa didyovow Ly hd e 3 > , ay 20° agiws ov ev avOpamou «ide éBiwaay Biov. ra dé xytAtoagr@ auddrepas . 7g de x @ dupérep ? o > A LA e adexvoupevat eri xAnpwotv re Kat aiperwv rou Sevrépou Biov aipodvrat bv dv OeAn éxdorn. This denotes a deeper understanding of the responsi- bilities of life, and agrees with the doctrines of the latest works, such as the Timaeus and Laws. Even the philosopher, who, according to the Phaedo, reached the happiest state immediately after death, being Place of the mythical tale extended far beyond the limits of earth. Increase of the duration of punjsh- ments or rewards. Cycle of ten thousand years for all souls except philo- sophers. Indovl- gence towards persons treated severely before : Polos, Prota- goras, Pericles, Anaxa- goras, Isocrates. All of them being esteemed less than philo- sophers. Contempt of poets 330 ORIGIN AND GROWTH OF PLATO'S LOGIC freed from the body (Phaedo 114 c: ot dtrocodia ixavas Kalnpdpevor dvev Te copdter Coot To TWapdtray sts Tov srEvra “povoy), is now obliged to return twice to life on earth in three thousand years before he can again reach perfection (Phaedr. 249 a). For other souls a cycle of ten incarnations during ten thousand years, unknown in the Republic, is now imagined at each fall into matter (249 a). This shows that Plato progressed in emancipating his thought from the narrow limits of time as known on earth. We see also other signs of the greater height from which earthly affairs are looked upon. Those against whom Plato wrote some of his earlier dialogues are here judged with the indulgence of one who is too sure of his superi- ority to deny small merits in others (247 A: $O0vos yao ¥Ew Osiov ywpod ioratac). Thus Polos, who had been treated so severely in the Gorgias, also Protagoras, and many others (267 a 8B), are recognised here as inventors of certain rhetorical artifices, not quite as important as they pretended, but useful and even necessary to those who know how to use them (269 B: 1a mpo rhs Teyvns avayKaia padnpara). This concession, though supplemented by the announcement that this preliminary knowledge should not be taken for the true art of rbetoric, is certainly a sign that the earlier hate is now changed into indulgent compassion. Pericles, too, who was treated with such severity in the Gorgias (516 A), is now represented as a model orator. Anaxagoras, who in the Phaedo was accused of having deceived Socrates by his unfulfilled promise of explaining everything through the power of reason, is now credited with the merit of teaching true eloquence to Pericles (270 a). In the same line comes also the very moderate recognition of Isocrates, only in so far as his character is said to be superior to that of Lysias and other orators (279 a), with the addition that even the greatest merit in this direction is infinitely inferior to true philosophy. In one respect Plato’s severity remained unchanged : the poets are here placed very low in the scale of human MIDDLE PLATONISM: PHAEDRUS 331 fates, below the gymnasts, money-makers, and sooth- sayers (248 D). That poets are imitators, 1s here assumed without any further explanation (romnrixos 4) tev rept pipnoly tis ddXos), aS if the reader were supposed to be familiar with the tenth book of the Republic, no earlier general definition of poetry as imitation being known.™ Plato’s progress from admiration of poetry to contempt of it began only after the Symposiwm, and was first justi- fied in the Republic; it 1s manifest in the Phaedrus and all later works. Also the low place assigned to the tyrant in the ninth book of the Republic remains here unchanged (248 BE). In some other respects we notice a development of earlier views. Love was in the Symposium the universal creative power in nature, and is here only one of four kinds of madness; Beauty was the highest idea, and is here only one among many ideas, of which justice occu- pies the first place (247 D: év 82 77 Tepi0d@ Kabupa pev aitny Stxaocvvny, xabopa 5% owdpoovrnv, KaBopa &e émi- oTinv ...),a8is natural after the long dialogue on justice (276 EB: mayKxarny, Tadiav,... Sixavocvvys ... Tépt puOo- Aoyouvra). Some important terms used in the Republic are here applied as quite familiar : thus dvvaycs in the meaning of a faculty (246 D: wrepov Sdvamss), Svadextexn Meaning metaphysical science (never used before Plato, and by 74 In Rep. II 373 B woimral are named as co-ord‘nate to mmnral, and the latter term applies to interpreters of poetry. In the third book of the Republic only a part of poetical works is done ‘ by imitation’ (394 c: ris worhoews . .. h uty 81d nihoews An eoriv, 7 8t 8: dxayyeAlas abrot tov roinrod, H 8 ad d:’ augorépwv), the term uiunots being never used as a general class to which poetry belongs. This is for the first time explained in the tenth book of the Republic and then applied, in the same manner as in the Phaedrus, in the Laws (668 a: povoixhy ye waody papev elxacrichy re elvat kal uiunrichv). To an evidently earlier stage corresponds the definition of poetry as creation in the Symposium (205 B: 4 Tot éx rot uh Syros els 7d by idvri dtqoty airla xacd dors wolnots . . . ded 8t xdons ris roithoews by Béptov dgpopiaOey 1d wep) thy povorkhy Kal Td wérpa TE TOU BAov dvdpar: wpoca- yopevera:), while in the much later subdivision of xo:nrixh in the Soph. 265 3, the primitive meaning of the word seems to be already forgotten, and poetry is not even named as one of the subdivisions. as imita- tors pre- supposes the Republic. Love and Beauty have lost a part of their import- ance. Use of terms in troduced in the Republic. Proof of the soul’s immor- tality compared with the proofs given in the Republic and in the Phaedo. Phaedrus supple- ments the tenth book of the Republic. 332 ORIGIN AND GROWTH OF PLATO’S LOGIC Plato first in Rep. VII, cf. Phaedr. 276 E); dcarextexos meaning, not as in the Cratylus, Huthydemus, and in Xenophon, one who knows how to ask and answer ques- tions, but the philosopher able to discover unity in the variety of particulars (266 B: Suvarov es 8y wal dal Toda mepuxo0’ copay . . . mpocayopeiw . . . Siadrerrixov, cf. Rep. 537 C: 0 ovvorrixos Siadextixos, cf. Crat. 390: spwray xal amoxpivecOa eriotdpevoy .. . Svadextixor) ; apyn as first principle of Being (Phaedr. 245 p). There are two special psychological theories of the Republic which recur in the Phaedrus, and offer some opportunity for an instructive comparison. The most important is the proof given of the soul’s immortality. Formally the proof differs here as in the Republic from the arguments of the Phaedo: a substance which must be necessarily immortal is first defined, and then the soul is shown to correspond to the notion thus determined. The proof given in the Phaedrus is supplementary to that of the tenth book of the Republic : there the question was asked, what can be the cause of destruction of something existing, and it had been answered by the supposition that only a thing’s own weakness and evil can destroy it. Here the corresponding positive question is asked, what is the cause of life and its external manifestation—move- ment, and it is answered, that the true cause must be a self-moving principle, all other things moved from without having no certainty of continued movement. There the only thing which is not destroyed by its own evil was the soul; here also each self-moving principle 18 found to be a soul. That the proof of immortality given in the Phaedrus is the later of the two, is evident from the fact that it is the only proof recurring in the Laws, and that no other new proof is given in any later dialogue. For the purpose of a further discussion of this definitive Platonic theorem, the two similar demonstra- tions in Phaedrus and Laws ought to be carefully com- pared with the last proof given in the Republic : MIDDLE PLATONISM: PHAEDRUS 533 Rep. X (abbreviated). 608 pD : dOavaros nay n Wuyy xat obdérore amoAAvrat (1). 609 a B: rd Evpurov kakov ékdorou «ai momnpia éxagroy dmé)- Avow (2), 7 ef py TovTO aoe, ovx Av GAdo ye avré ért Siapbeipecer (8). ob yap Td ye dyabev pn mworé Tt admodéon, ovde kat av rd pnre Kaxdv pyre ayabor (4). 609 B: éay dpa re eipioxwpev Tav dvrwy, @ fore pev xaxdv, 8 motes avTdo poxOnpor, Touro pevrot ovxX oldy Te avré Avew adroddvor. .. 78n elodpeOa Gti Tov mebuxdros ovrws dhe- Opos oun nv.... (5) 609 D: Wuyny... évovca ev avty adtxia 1+. TH Evetvat ovdapas . . Pbeiper (6). @oyov tiv péev dddov mompiav amodA\vvat Th, rnv € avrod pn. 610 c: ovdeis more|..-- Seifec Os TaV aroOvn- oxdvray adiorepa ai uxai dia rov Odvarov ytyvovrat (7). 610 E: drdre 8) py ixavt) if ye olxeia rrovnpia kat Td olketov Kaxov dmoxretvat kat drro\éoat uxny, axXoAy TO Ye én’ Ddov 6rA€6pq reray- pevov Kaxdyv uxny if TE iANo dod... . érdre pnd’ bd’ évds amdddvrat Kauxov, pyre olketov pre dddXorpiov, SnAov Gre Phaedr.245 c—-246 a: Wux) maca déavaros (1), Td yap detkivnrov aOavatrov - to 8 Addo Um’ dAXov Kivoupevoy, mavAav éxov Kunoews mavAay €éxet Cans (2): povoy 8 rd eo 4 auto KivoUY Kat Kuvouv, dre ovK ? ~ € U 7 amroXetroyv €auTo, ov Tore Anyes Kevovpevov, aAda Kai d\Aos = Ooa Kuveirat ToUTO myn Kal ? ‘ U > A Gpx7 Kwvnoews (8). apxn 8€ dyévnrov. &€& apxijs yap avaykn mav To yty- vopevov yiyver Oa, avrny dé pnd’ €& évos .... > 4 \ > 4 , ? éretdn 8€ ayémroy eéo- Ti, cat adiabopov avrd dvaykn elvat. apyis yap 81 amodopuemns obre atirn more éx tov ovre Gro €& éxeivns yevn- cerat, etrep €€ apxns Set Tos .|ra wavra yiyverOae (4). ourw 07 Kiynoews pev Gpx7 To avrdé avré Kuvovv. Touro O€ ovr’ dmdAAvo Oat otre yiyverOar duvaror. . @avarov 8€ reda- opevov Tov vd’ éavrov xevouzevou (5), Wuxns ovotay te Kai Adyor rouroyv avuroy Tis Aéyov aioxuvetras (6). nav yap oa@pa, ® pev eLwbev uveio Oat, dpuxov, @ 8 ev8obev alr@ é€ atroi, éuypuxov, @s ravTys otons pucews Wuxns (7): ef 8? €or TovTo oUTws €xov, 7 @AdXo rt eivat rd adro daurd xivody hf ypuynr, é& dvdyxns § ayé > OUK & TO dyémrév Laws: 894 £ : dray érepov GAXo new peraBadn, kat rovro aAXo €repov > * a ? Gel, THY TOLOUTMY . . OUK . . €oTat Wore TL TPaTov peraBadXov. GAA’ Grav auro avTO. = kta €repuvy aAXNowon, Td 8’ érepov Go . . . dpxn res érrat ris Kunoews dmaons .. . THs aUTNS avTnY KiWn- odons peraBoAn. 895 B: dpxnv " dpa Kwnoewvy mwacav Kat mpatny .. . Pnoopev dvayxaiws eivat mpeaBu- TaTnv... auTav C: (nvaird mpowepau- pev, Gray avrd avro Kivi. 896 a: @ 87 Wux7 rotvopa, Tis Tourov Aoyos ; €xouev diddy mAnv tov voy 8 pnbevra, Tyy = =Suvapévny = avrny QUuTHy KivelyKimmow; ... ixavas dedeixOar yruynv ravtov by nat thy mpa@rtny yéveow Kai xivnow Tov Te Gvrwv Kal yeyovorwy kai écopévov ... gews dndons drractv. 904 Cc: peraBadrAa péev roivuy mavé’ dca péroxa dare Wuyns, év €avrois kextnéva THY 17s peraBodns airiay. 959 a: weiberda 8 cori rp vouobéry xpedy ... Adyorrt. . . . Ov air@ rp Bip rd mape- Xopevoy nudy éxacrov rovr’ elvar pndev GAd’ Thy Wuxny .. . B: Tov xuvn- >» 4 airia Parallel- ism between the proof in the Republic and the Phaedrus. Both proofs posterior to the Phaedo. They show a greater certainty, an ad- vance in the form 334 ORIGIN AND GROWTH OF PLATO’S LOGIC dvdykn aird dei Sy eivat, [re nal &Odvarov Wuyi | dvra quay exacrov dvras ei 3’ det dv,dbdvarov(8). | av ely (8). mept pév| aOdvaroyv, uyny éravo-~ 611 A: Totro peévjovy dOavacias adris|patdpevov, dmévat da- roivuy, ovtws exérw (9). | ixavas (9). govra Noyov... The most striking parallelism is evident between the two first proofs. Both begin by a short statement of the theorem which has to be proved in what follows (1). Both then name a kind of things subject to destruction (2), contrasted with another indestructible kind (3). The indestructibility of this second kind is then proved by elimination of other possible suppositions (4). The next step in both arguments is the conclusion that a thing corresponding to the above definition is inde- structible (5), and the identification of such a thing with the soul (6). This identification 1s brought about in the Republic by a longer digression on the possible analogies between soul and body (609 B-D) which. has been here omitted. In the Phaedrus the identification of the soul with the self-moving principle is briefly intro- duced as a conviction of which nobody needs to be ashamed. After this identification in both passages follows the special indication of the opposition between body and soul (7), the conclusion that the soul is immortal (8),in the Phaedrus supplemented by the additional determination that it has no beginning, and the whole argument concludes by an express statement that the proof is deemed sufficient (9). If Plato knew any one of these arguments when he wrote the Phaedo, he could not have omitted such proofs, which are far superior to anything which the dying Socrates had to offer to Simmias andCebes. That those proofs were not yet deemed sufficient by Plato himself is seen from the exhortation at the end of the Phaedo to investigate the subject further (Phaedo 107 B : dvaryxafopnar amiotiav art yew tap’ suavt@ wept Tov eipnutvwv, says Simmias, and Socrates answers: «al tds Uro8sces tas mpatas, Kal e& micral diy sioww, Suws imioxenréas capéorte- MIDDLE PLATONISM: PITAEDRUS 335 pov) which is the opposite of the confident assertion in the Republic as well as in the Phaedrus, that the above proof is sufficient (Phaedr. 2464: ixavads, cf. Phaedo 1O1L EB: &ws ésri te ixavov éOors). The logical method of beginning with the enunciation of the theorem which has to be proved, and then stating the axioms on which the proof rests, is also an advance beyond the method used in the Phaedo. Thus the arguments both of the Republic and Phaedrus are clearly later than those of the Phaedo. And almost equally probable is the priority of the Re- public as compared with the Phaedrus, the latter being distinguished by a greater conciseness, by the avoidance of induction based on analogy which is used in the Republic, by its deductive character based on necessities of thought, by the exact co-ordination of immortality or infinite future with an infinite past, and above all by its agree- ment with the only proof given in the Laws. This is a point of the greatest weight : Plato laid great stress on the immortality of the soul in the Laws, and out of all his arguments in favour of this doctrine he selected the proof given in the Phaedrus as adequate (txavor) and worthy to be repeated in his latest work. This confirms our view that the Phaedrus is nearer to the Laws than the Phaedo and Republic, which are the only other works of Plato containing logical argumentation about immor- tality. After the Phaedrus Plato thought it superfluous to look for new arguments, and whenever he spoke about immortality he took it as well established and certain, or he added only, as in the Timaeus, mythical representa- tions fit for popularising one of his favourite theories. The comparison with the Laws disposes also of every doubt about the author’s intention to apply his proof to the individual soul of every man.** Whatever Plato 288 Some ancient interpreters thought that Yux}) waca means ‘the whole soulin the universe,’ and this artificial interpretation has been accepted also by Teichmiller (I. 63), who contends that Plato did not admit individual immortality, against the evidence of the texts. But Walbe’s very special of expres- sion, carried further in the Phaedrus than in the Republic. Coinci- dence of Phaedi us with the Laws. In both cases the individual soul is meant, as results from a com- parison between Phaedo and the Laws. Doctrine of the parts of the soul in Phaedrus and Republic. 336 ORIGIN AND GROWTH OF PLATO'S LOGIC thought later about the relation of individual souls to the whole or to God, there is no possible doubt that he taught individual immortality as a rational theory from the Phaedo up to the Laws. There is no need to infer with Teichmiiller that those who read this teaching in Plato’s works make him an adherent of atomism or monado- logism. Individual souls can have a common origin and an universal direction, remaining all the same immortal, and always equal in number, as we read in the Republic as well as in the Tvmaeus. The Platonic doctrine was that the inward personality by no means needs the body for its existence : Phaedo115cD: ov reidw Kpirwova, Legg. 9590: ov8érote oixo- ws éym eius ovros 6 Zaxparns, 6 vurt Scadeyopevos . . . GAN’ olerai pe exeivoy elvar, dv dyerat ddtyov Uorepov vexpdv, kat épwra 8n, was pe Odrrn .. . émecdav wim rd adp- > o € a ~~ > b puxov, ovxére Upiv mapapera, AAA olynwopat amdy .. . Pbopeiv xp7 Siahepovras, voy Covra Tov avrod rovrov elvat rov Trav “ ? s capxav dyKov GAA > - @ , U 9 €xeivoy . . . Ovriva tis padiob nyetrat mwobav Odrrev, olyerGat mepaivovra Kal éeuminddvra ry avrov joipay. Garrdpevov, This doctrine, common to the Phaedo with the twelfth book of the Laws, unchanged in the course of thirty years and more between these writings, results with equal stringency from the Phaedrus as from the Republic, since in both the soul is opposed to the body, and immortality predicated of the pure soul. We had to dwell at some length on these comparisons, because of their importance for the order of the dialogues, and also because they illustrate a logical progress of method. Plato’s increased power of exact argumentation did not prevent him from indulging in his favourite manner of mythical allegories, as we see in the shape which he gives in the Phaedrus to the other chief psychological doctrine of the Republic, namely the three- investigation on the use of ras in Plato (see note 185) proves that here ux) waoa means ‘each soul’ or ‘all individual souls,’ and not, as Thompson translates, ‘the vital principle in general’ (Jowett: ‘the soul through all her being’). MIDDLE PLATONISM : PHAEDRUS 337 fold partition of the soul. But even in this mythical shape a certain development of doctrine is noticeable. The @uyos was defined in the Republic as 16 & Oupodrat (580 D), and we have there interpreted it as the moral feeling. This interpretation finds its confirmation in the Phaedrus. Plato must have felt the terms @uyos and Ovpoedés to be too narrow, and this explains why Oupoedés as @ faculty of the soul has never been used by Plato after the Republic except in the recapitulation of the Timaeus (18 4). In the Phaedrus the moral feeling is represented under the image of a beautiful and good horse of noble breeding (2468), full of ambition, but also the lover of temperance and honour, following right opinion and amenable to reason (253 D). This is a wider determination than that given in the Republic, and also the classification of men according to their capacities is much enlarged. There we had only three kinds of men, divided according to the prevalence of one or another faculty. Here we find twelve kinds of souls, each of which has its own different ideal (247 A) allegorically represented by one of the Olympian gods. We need not attach any special importance to the number twelve, which is here accommodated to the mythological form. But it is certainly characteristic that Plato admits a great variety of souls not only in the myth of the dialogue, but also in the following conversation (271 B: uyfs yévn), and this reveals an enlarged view of human nature. Here, as in former writings, the philosopher is placed above all other kinds of men, as following the band or chorus of Zeus (248D: ri mreiata idovaar eis yovny ... dtirooddov, cf. 252%). He is here named a leader of men by his very nature (252 E: ¢cAdcodds te xal Hrysuovixos tHv dvow) whereby the result of the long explanation of the Republic about the leadership of philosophers is briefly assumed as certain. A still stronger sign of the increasing educational influence of Plato is that he once uses ‘we’ (250 B: jets) without any nearer Z Wider determi- nation of the moral feeling. Classi- fication of men exceeds thenarrow limits drawn in the Republic. Philo- sopher assumed to be a leader of men. Authority of the philo- sopher above all other men. Ideal philo- sopher similar to God. Divine philo- sophy. Ideas con- templated by reason not the senses. 338 ORIGIN AND GROWTH OF PLATO'S LOGIC determination, in the meaning ‘the philosophers.” The writer has already a sufficient public of readers among his pupils to feel certain that he will not be misunderstood. But he insists repeatedly on the scarcity of philosophical natures (250 A: oAdyas Aghrovtat, als TO Tihs wvHpns ixavas wdpecti,. . . 250B: poyis avray Kai orLyou ari Tas eixovas iovres Oe@vtas TO TOU etxacOévTos yévos). No authority is binding for the thinker but his own reason (270C: ypy wpos To ‘Imoxpate: tov Noyov e€erd- Covra oxorety, ei cuudovel), and the philosopher proclaims his superiority not only above the poets, as in the Republic, but above the law-givers and orators; only when they are philosophers do any of these deserve our esteem (278cD). This contempt for the eminence of fame and vulgar opinion (274C: dvOpwrivev dofacudtov) shows a great distance from the Symposiwm. That the term dirccogos is here introduced in opposition to codgus (278 D) is a rhetorical artifice, like the novelty of immor- tality in the Republic, while in another passage the dialectician is compared to a god, whom even Socrates would follow with delight (2668). Moreover, the ideal of the philosopher appears here, as later in the Laws, more and more supplanted by the ideal of a God, to whom the philosopher is similar. But in so far as any comparison of a philosopher with other men is made, the superiority of the philosopher accentuates itself more and more. Philosophy is divine (2398: ela ¢iAocodia) as in the Timaeus (474: dirocodlas peitov ayabov obr’ FrOev oO? HEet rota re Oynr@ yéver SwpnOav ex Oeav, cf. Phil. 16C: Oedv eis dvOpwrrovs Sects), and leads her votaries to please gods not men (274 4). In the mythical part of the Phaedrus the ideas are still spoken of as contemplated by reason (247 0: 1) aypa- paros Te Kal aoynpariotos Kal avagns oiala évTws otca,. . . pov Ozarr v@), and appear to be objective (247 DE: xadopa . 2 « Otxaloovvny .. . driotHny, ov H yéverss TpdceoTLV ... GNA thy bv re & bor bv Bvrws ériotnuny odcay), with MIDDLE PLATONISM : PHAEDRUS 339 the express caution that they cannot be seen through our bodily eyes (250 D: dye Ppdvyois ody oparas). At all events the theory of an immanence of the ideas, as taught in the Symposium, and to a certain extent in the Phaedo, is supplanted by the view of a similarity or imitation of the ideas by the things which has been already indicated in the Phaedo and accepted in the Republic. The particular thing is an image of the idea (ouoiwya, 250 A, B) which it imitates (251 A: @eoedés mpoowmov KddXos ev pEmipn- pévov 4 Twa cwpuaTtos idgav). We must translate this metaphorical speech into abstract thought in order to learn whether the writer of the Phaedrus continued in his belief of separate ideas. And the metaphors here used might well be applied to general notions. There are some hints pointing in this direction. Amidst all the imagery of the space above heaven appears a very dry explanation of the difference between man and animals. Man must understand general notions which are the result of the union by means of reasoning into one concept of what appears to the senses as a manifold variety (249 B: dei dvOpwroy Evriévat To Kart’ eldos Asyomevov, & TONA@Y lov aicOncEwr Eis By NOYIoL@ Evvaipoupévwv). This is given as an explanation of the preceding metaphorical assertion that no soul is incarnated into the form of man without having enjoyed the super- celestial vision of true substance and science. If we follow this example set by Plato himself in the interpre- tation of his allegories, we soon get quit of the riddle of self-existing ideas. Plato does not require us to take his mythical allegories literally: he says clearly that he does not insist on everything said in the myth (265 B: icws pev aAXNNGods Twos éparrropevar, Taya 8 av nal ddNdNoGE Tapadbepopevot, kepdoavres ov Tavratacw «amlGavov dOyor, puOcxov Twa duvoy mpocenaicapev petplws te nal edpnyuos), and confesses to have mixed truth with fiction. Thus we are at liberty to interpret the allegories and to distinguish truth from fiction. That ‘beyond the limits of the stars z 2 Ideas as models of Being might well be identical with general notions. Faculty of per- ceiving unity of species in the variety of appear- ances & privilege of man. Metaphors about ideas could refer to general notions. Ideas of Plato and of Kant. Ideas formed by ' the study of par- ticulars. Subdi- vision of ideas into natural 340 ORIGIN AND GROWTH OF PLATO'S LOGIC exist pure ideas without shape or colour, intangible and invisible, not fixed in sensible particulars, but free and independent,’ means only: that pure concepts of reason are never fully realised in the things to which they apply, as for instance, absolute equality 1s never found identical with physical equality. Our interpretation is appliable even to the ideas of the Phaedo, though there we had not such an express authori- sation of free interpretation as in the Phaedrus, where the whole mythical account is called a pleasant play (265 C: daiverar ra pay adda TO SvTt Tabi TreTaicBat) in which the only serious thing is the double way from particular things to the general idea, and from the idea to all its particular kinds. Here éd¢a and élSos are used in @ meaning which is identical with the idea as conceived by Kant, a necessary concept of reason. The synthetic union of scattered particulars is clearly a condition of consistent definition for the purposes of teaching (265 p: eis play te Stay cvvopavta ayay Ta ToANAXH Stecrrappeva, tva scacrov opilouevos Sihdov troth, Tepl ov ay dei Siddonew a0#\n ...). The test of self-consistency is already stated in the first Socratic speech as the indispensable condition of knowledge (237 C: tovs qwrodAovs AsAHOev GTi OvK toact Tv ovaiay éxdotov’ ws ody siddtes ov SioporoyoovTar év apyh THs oKtrews, mpoedOovres 52. . . odte gavTois ovTE G@AMjAots Oworoyovow). Substance is even used as a synonym of definition (245 E: ovalay te Kai Aoyoy, cf. 270 E: rhv ovoiav SelEa axptSas tis pvcews TovTOU, Tpos § rods Noyous mpocoicet). The ideas appear as a result of the study of particulars, not found in the particulars, nor taken from the particulars, but discovered by reason in the act of defining each par- ticular (273 &: war’ eldn Svaipeiobar ta byta Kal wa idéa xa’ by Skacrov TepiiapBdvew). When once a general idea is formed, it becomes the dialectician’s aim to subdivide it into kinds, not artificially, but into natural kinds (265 £: ro maddy Kai Ta edn SivacOar téuvay nat’ &pOpa, 5 wépucer) MIDDLE PLATONISM : PHAEDRUS 34] which are distinguished from accidental parts. This division and classification must proceed to the point of indivisibility (277 B: av épitecOat, opicdmevos te mad kat’ edn wéypt ToD atynrou Téuvew). This method (126o0dos, 269 pD, 270 c, D) shows the relation between each particular and the whole, neither soul nor body nor anything being perfectly known if studied apart from everything else (270 c: Wuyis pvow akiws Noyou xaravonoas (advvatov) ... ovo? Gwpatos ... avev THS ToD ONov gdvaews). The first step of investigation is to ask whether a proposed object is simple or manifold, indivisible or divisible (270 D: aparov fev, AtrAovv H woAvEtdés gotiy . . . Set StavosioGar). The method of definition and division of notions differs from the divine intuition of ideas. And along with this transition from metaphysic to logic, the efficient cause, despised in the Phaedo, regains its rights. We are asked in the case of a simple element to investigate its active or passive capability in relation to other things (270 D: dy Mey aTrrODY 7H, TKOTrEy THY SUvapW avTov, Tlva Tpos Tl Tépu- xev sis to Spay yoy f) Tiva eis TO TaGeiy vrro Tod), While in dealing with a compound whole, we have to divide it into its kinds or elements, and then to look for the activity and passivity of each of them (270 D: dav dz Asi eidy éyn, Tatra aptOpnodpevov, Strep &p svos, trodr’ ideiv ap’ éXaoTOU, TH Ti TovEty AUTO TWépuKEV 7) TH Tt raGeiv Vird Tov;). The recognition of efficient causes corresponds to the higher esteem of Anaxagoras, and to the definition of the soul as a self-moving principle. This removes at once all possibility of believing the Phaedo to have been written later than the Phaedrus, as the importance of efficient causes 18 constantly recognised in all later works, for example in the Tymaeus and the Laws. The Phaedrus is a work of the greatest inspiration; 1t contains in the most natural exposition the germs of much that was later worked out by Plato, and it betrays also a greater range of study than the Phaedo. Dhalectic as a science of Being based on definition and division 1s the kinds brings them into mutual relations to each other. Transi- tion from meta- physics to logic. Division of things into their elements. Efficient causes recog- nised. Dialectic based on definition and divi- sion ap- plied to teaching, leads to the greatest human happiness. Natural conditions of eloquence contrasted with the art of rhetoric. Suitable disposi- tion of the parts of a dis- course. 342 ORIGIN AND GROWTH OF PLATO'S LOGIC fulfilment of what had been postulated in the earlier work. Its chief application is clearly shown according to the ethical rules explained in the Republic: he who knows is bound in duty to teach. The teacher writes imperishable lines in his pupil’s immortal soul, imparting a living know- ledge, together with the ability to defend it against errors (276 A: rov eiddros Novos Cov Kal Eurpuyos. . . ypdderar év 79 Tov pavOdvovtos uy, Svvatos pév dpdva: gavt@, eriotThpwy 53 Adyew Te Kal oiryady pos ods Sei). For this he requires dialectical art (276E: 7H Siarextingn téyvn ypw@pevos) and must make a proper selection of receptive souls (AaBwv puyny mpoonxovoay). Then his activity will yield eternal fruits and procure the highest happiness attainable by man (277A: Noyous, of ovyl Axaprro: GAXa exyovTES oTréppa, S0ev Gdroe sv arrose HOgoe Gvoduevor tToir’ asi AOdvatov saptyew ixavol, tov FyovTa eddatpoveiy TrotobvtTes ets Goov avOper@ Suvatov wadiora). An art of rhetoric is recognised as useful, but the essen- tial conditions of a good speaker are: innate ability, exercise, and knowledge of the subject on which he intends to speak.26 If to these conditions we wish to add the guidance of art, then we are asked to look for much more than has been offered by rhetors and grammarians, who were able only to invent such elementary rules as are preparatory to the art, much as the rules for tuning a musical instrument are preparatory to a theory of harmony (268 x). True eloquence requires, besides a perfect know- ledge of the subject dealt with (262 c), also an excellent formal arrangement of the contents (236 a). Hach speech must consist of well-proportioned parts, and have a proper beginning as well as a suitable conclusion, with such a disposition of the contents that each part shall have 6 Strangely enough this knowledge (¢m:orfun 269 D) has been mis- understood by many interpreters as if it meant knowledge of the rules of rhetoric. Even E. Holzner (‘ Platos Phaedrus und die Sophistenrede des Isokrates,’ Prager Studien, Heft IV. Prag 1894), who corrects the error of those who identified this éxiorfuy with the following réxvn, falls into an almost worse error in asserting the identity of ér:orfun in this passage with Th xpd rhs Téxvns dvaynaia uabhuara 269 B, This misconception is due to MIDDLE PLATONISM: PHAEDRUS 343 its proper place, being the continuation of what precedes and preparing what follows (264 C: péca nai dpa, mpémovrTa GAAHAOLS Kal TO Ow yeypappéva). A speech must not be like those verses which can be read in any order (264 D). There are rhetorical necessities which determine the placing of each part of a discourse (2648: 7d dedTEpov eipnuévov x Tewos avdryKns Sevtepov tEOHvat). In order to arrive at this perfection, an art is required far above anything known heretofore by the name of rhetoric (266D). This art will teach us to lead souls by means of speech (261 A : wuyayoyla 51a Aoywv) not only in tribunals and on the market place, but in every circum- stance of life, small or great (261 B: % adtTn optkpay te Kai peyddXwy trépt, Kal ovdev éytimotepov TO ye do0ov TEpl oTrovbaia wept daddra yeyvouevov). The true speaker must begin an abuse of comparison with Isocrates’ oration against the Sophists, without taking into account his later works, and earlier opiniuns of others: Isocrates, im sophist. (Or. xiii.) § 17 (294 p): Sety roy pey padnrhy wpbs tH Thy obow Exew olay xph, Td nev €13n Ta TaY Adywv wadery, wep) Se ras Plato, Phaedr. 269 p: ei wév cor bwdpxe: puces pnropixg elvat, Eve: phrwp eAAdyimos, TporrAaBwy émt- orhunv re wal mederny. Isocr. Antzdosis (Or. xv.) § 187: (Steph. p. 93). Sez Tous wéAAovTas Si0f- oew h wept trovs Adyous H wepl ras mpdies ... Hp@arov wps TovTe wWedu- Kévat KaAQ@s twreira wa:dsev07vac Kal AaBeiv Thy emorhunvy jris by i wept éxdorov, tplrov Se 0 evrpiBets = -yever Bat kal yupvacOjvat About 353 B.c. Cf. 237 c: eidévar Ser wept ob by 3 7 BovAh.. . 259 x: iwdpxew Se? Trois eU ye Kal Karas pnOnoomevois 9 thy ‘Tov A€yovros Sidvoiay eldviay TaAnOes dy dy dpety wéps BeAAD. Xphoes abtray yupvac- Ojvar. (About 390 3.c.) Cf. Plato Prot 323c: ob ptce: GAAA Bidanrdy Te Kal e& éwsmeAclas wapa- ylyver8at. Xenoph. Memor. II. vi. 389: dperas wdoas habhoe te nal peddry abfavoudvas. IV.i. 3: al &psora Soxotoa evar pvoes pddtocra xadelas Séovras. See also Alkidamas’ rep! cogiordv as quoted by Gercke (Hermes, vol. xxx1i. pp. 862-364, Berlin 1897) who is, however, inclined to invert the chrono- logical relations. Here it is by no means certain that Plato had in view the much earlier work of Isocrates, as the three conditions of success were a commonplace and needed not to be invented by Plato or by Isocrates. Now it is very important to observe that Isocrates thought, in 390, that only the know- ledge of rhetoric is required, while thirty-seven years later he agrees with Plato in asking for a knowledge of the subject. That in the Phaedrus éxiorf un means knowledge of the subject is evident from the other passages and from the opposition of this knowledge to réx»n. Bach part from the beginning to the end has its prope) place. New con- ception of rhetoric. Classifica- tion of souls and of kinds of oratory. Plato did not write a hand- book of rhetoric, being therein very diffe- rent from Aristotle. Plato’s works retain always their dramatic character. They re- cord his own oral teaching. He was able to speak better than he wrote, 344 ORIGIN AND GROWTH OF PLATO'S LOGIC by studying all kinds of souls and their classification (271D: dvdynn eidévas Yuyn doa edn Fyet). Plato does not enumerate here these kinds, but those enumerated by Aristotle are probably due to a great extent to Plato’s teach- ing. The next step is to determine what can act on a soul, and what are the limits of the soul’s action (271a: OT@ Ti TroLEety 7) Tadety Uae Tov méeducev). After a careful classification of souls and of kinds of oratory, a special inquiry is needed to show what kind of speech acts on each kind of soul, and why it has this power (27183: &a- TaEdwevos Ta Noywr Te Kat ruyis yévn kal Td ToUTw@Y TAOnpLaTa diese Tas aitias, mpocapportwy ExacToy Exact Kal bidacKwy, ola ovca ud’ olwy royov & Hv aitiay & avayens 7) bev mevOsrau, 7 5& atreGet). This clear programme of a future rhetoric has been so exactly followed up by Aristotle in his work on the same subject that probably Plato’s special teaching on that matter is preserved in his pupil’s exposition. Plato himself left no written system of rhetoric, because he did not write for the purpose of teaching, but for the artistic reminiscence of some new thoughts, or in order to refute the enemies of philosophy. All his works, even the Laws, preserve that character of art which is absent from the works of Aristotle. Systematic teaching was probably given by Plato to his pupils, and transmitted by them to the following generations in the Academy. But he appears not to have thought it a convenient subject for written exposition. There is some dramatic character in his works even when they contain such dry enumerations as we find in the Sophist and Timaeus. Also in the present day, though writing 1s so much easier, some eloquent men write little. We may well believe Plato when he says that his eloquence was still greater than his literary skill (278 o: Adsywv avros Suvaros ta yeypapptva padra drobvei~ar), as he expressly asks every great writer to be able to speak better than he wrote. This is certainly not a common faculty, 345 and many great writers would disagree with Plato. He judged evidently according to his own experience, and his power of oral eloquence has been unanimously praised by the tradition of his times. This explains why Plato in a much longer life wrote much less than Aristotle. Aristotle, owing his initiation to Plato, may have been older when he began to write than Plato was, since Plato at the age of twenty-eight was emancipated from the influence of his teacher, while Aristotle remained under some influence of Plato up to the age of thirty-seven. If we assume that both began their literary activity about the same early age of twenty-eight, then Plato wrote during fifty-two years, and Aristotle only during thirty-four. But the extant works of Aristotle are considerably more than twice as long as all the works of Plato, though many works of Aristotle are lost, while we have all the works of Plato. This leads us to the conclusion that Aristotle wrote four or five times more copiously than Plato, and this implies a great difference of views about the use of writing. It is clear that many things written out by Aristotle were not held by Plato as fit for literary representation. We must not judge about this from our present point of view, accustomed as we are to learn chiefly from books. In Plato’s times, and in his own opinion, oral teaching stood much higher than written handbooks, and this was a natural consequence of the difficulty of writing and reproducing written matter. It has been frequently argued from the celebrated passage on literary composition at the end of the Phaedrus that Plato despised writing altogether. This 1s certainly a very exaggerated inference. He calls writing a play, but at the same time insists on the superiority of this philosophical play over the vulgar diversions of other people (276 D: wayxdAnv mratdiav, Tob év Noyous Suvapevov rrailey, Sucacocuvvns Ts Kal GAAwY Ov Adyars wépt pvOorcyodvra). To fable about justice, as had been done in the Republic, is one of the most beautiful MIDDLE PLATONISM: PHAEDRUS and he wrote much less than Aristotle, though he lived longer. Difference of Plato’s and Aris- totle’s views on literary compo- sition. Plato did not despise writing. This was for him the noblest play. Reference to the Republic in a passage of the Phaedrus. The Phaedrus invites the reader to join the school. Invitation to learn philo- sophy ex- tended to Isocrates. 346 ORIGIN AND GROWTH OF PLATO'S LOGIC amusements for a divine man. There is no reason to think that Plato would not have spoken so lightly of writing after his great work. On the contrary, it is psychologically probable that he would not have spoken thus without the full consciousness of being a great writer (cf. Laws 968 £, where the same thought recurs at the end of Plato’s largest work). It would not suit his artistic intention to despise writing if he had not already proved that he is a master in it, and that his contempt is not a consequence of impotence. And he has a very definite rhetorical and artistic purpose in this passage. After an encomium on his own written myth put in the mouth of Phaedrus (257c: rov Aoyov 54 cov madat Gavypdoas %yw), adinitted even by Socrates with the poetical pretext of inspiration due to the Nymphs (263D: dc Aeyels Teyvinwrépas Noudas . . . Avalov mpos dAoyous eivat), it is his purpose to raise the reader’s expectation to the highest pitch by announcing that this beautiful sample of written eloquence is nothing as compared with his oral teaching. The Phaedrus, like the Symposium, Euthy- demus, and some other works, 1s written not only for the pupils, but also for those who followed wrong paths out- side of the Academy, inviting them to join the School. Invitations are extended even to those about whom no hope could be left. lLysias is told to learn dialectic, and what has been held for a eulogy on Isocrates is rather an ironic invitation to learn true philosophy. Anybody who reads Isocrates’ Panegyricus, written in 380 B.c., or about the same time when Plato was occupied with the Republic, will understand that Isocrates could not be flattered by such a form of recognition as that which we see in the Phaedrus. The recognition was meant sincerely, as also the merits attributed to Pericles (269 a), Prodikos, Polos, Hippias (267 3B), Protagoras (267 o: IIpwraydpaa.. . WoNAG Kal xadd), even to Sophocles and Euripides (268 c) MIDDLE PLATONISM: PHAEDRUS 347 who are named as the greatest poets, without any reference to the general low appreciation of poetry. In no other work of Plato is that same spirit of benevolence and conciliation shown, and this disposition of mind is best explained after a great success, like the production of the Republic. But certainly Isocrates pretended to more than to be preferred to Lysias (279 a), to hear that ‘some philosophy’ 1s manifest in his character (2798: éveoti Tus htAocodia 77H 700 «wvdpos Stavoig) and to be advised to take a more divine start (279 A: opur Georépa) than his present pursuits, if he cares to do better than to excel ‘in later age’ all orators (279A: ovdev adv yévorro Gavpacrov mpoiovons THs HAuKias, ef... WEeoV 1} Taldwv Suevéyxo: TOV TetoTe araysvwy AOywv). This prophecy is at once shown in its relative value, when we read in continuation that there is something far greater (u=fw) than to excel all orators, something requiring a divine power, and this is nothing else than Plato’s educational activity. Isocrates repeatedly pretends to be a representative of true philosophy (for instance Panegyric. §10) and he must have felt humiliated by Plato’s judgment of his relative merits. Thompson!” has shown at least one passage of the Phaedrus which clearly criticises a preten- sion of Isocrates as proffered in the Panegyricus : Isocr. (Or. iv. p.42.cb) Panegyr. § 8: émedy 8 of Adyot rocavtny éxovor THy hvow dof oidy 7’ eiva Te€pt TOY avTav TmoAdaxas eEnynoac- Gat, kat ra peydda rarewk roijoa Kal Tois puxpois peyebos meptOerva, kat rd re mada xavas ScedGetv nat Tepi Trav vewort yeyernpevov apyaios eireiv, odxért euxréoy ravr’ dori, mept dv érepot mpdrepuy eipnxace, GAN’ duewoyv éxeiveoy elmreiv retparéoy. Plato Phaedr. 2674: Ticiay S€ Topyiay re ddgopev evdev, ot mpd Trav dAnOav ra eixotra el8ov os Tuunréa paddov, ra Te av optxpa peyddka xat rad peyddta opixpa haiver ba rocovow 8u popny Adyou kava Te adpxaiws tad tT evavria Kaw@s, ouvropiay tre Adyov Kat dretpa pny wept mdyvrwv aipov; ravra 8€ dxovwv more pou Lpaddecos éyédacev. Here we see that an artifice which Isocrates recom- roended as useful is attributed by Plato to Isocrates’ Spirit of benevo- lence and concilia- tion. Moderat«: recogni- tion of Isocrates could not be accepted as satis- factory by him. Relation between the Phaedrus and Tsocrates’ Panegyri- cus dis- covered by Thomp- gon, though already alluded to by Cicero. Mention of Isocrates as at- tempting & new kind of speeches refers to the Pane- GYTICUS a8 Teich- miiller demon- strated. This in- ference confirmed by an ob- servation of Diimmler, though he professes another opinion as to the Phaedrus. Pretended early date of the Phaedrus. 348 ORIGIN AND GROWTH OF PLATO’S LOGIC teacher Gorgias, and condemned by the remark that he who would follow this advice would be obliged to esteem probability more highly than truth, and deserved the laughter of Prodikos. This relation of the Phaedrus to the Panegyricus, already implied by Cicero (Oraéor, xiii. 37), and again dis- covered by Thompson, gives a precious chronologic indica- tion, as the Panegyricus is known to have been published in 380 B.c. The date of the Phaedrus is thus indicated by an anachronism of Plato almost similar to that of the Symposium, because he puts in the mouth of Socrates the prediction that Isocrates would easily excel all orators if he continues to write such speeches as those on which he works now (279 A: doryous, ols viv érreyerpet). This ‘now’ cannot refer to the lifetime of Socrates, as then Isocrates wrote forensic speeches not deserving even that restricted recognition which Plato expresses in the Phaedrus. And, as Teichmiiller (ignoring Thompson) demonstrated in a most convincing way, no earlier work of Isocrates than the Panegyricus could educe from Plato any appro- bation. This is unexpectedly confirmed by Dimmler, though he continues to believe in an early date of the Phaedrus (Chronologische Bettrdge, p. 11). Dimmler sees in a later work of Isocrates (Antidosis, §62) a clear allusion to a conditional approbation of the Panegyricus, though he does not refer this mention to the Phaedrus, but to the Republic (426cDE). Itis more probable that Isocrates when he wrote the Panegyricus already knew Plato’s views on the relation between MHellenes and Barbarians (Rep. 470), though this cannot easily be made evident. The date of the Phaedrus is one of those problems in Platonic chronology on which a great wealth of ingenious supposition has been spent in vain. The strangest of all possible errors was the thought that the Phaedrus could have been written in the lifetime of Socrates. This is a result of purely philological combinations, without any MIDDLE PLATONISM: PHAEDRUS 349 consideration of the philosophical contents, which betray a date at least as late as the Republic, and undoubtedly later than the Phaedo. What reasons Diogenes Laertius had for his observation that the subject of the Phaedrus has something juvenile in it, is unknown. Probably he held with many superficial readers the subject to be illicit love, not philosophical rhetoric. In our century Schleier- macher was the first to proclaim that the Phaedrus must be one of the earliest works of Plato on the ground of its philosophical poverty. He says that the philosophical contents in the Phaedrus are not yet mature for a dialec- tical exposition (vol. 1. p. 67), for which were substituted strength of passion and questions of method. He seems to believe that investigations of method are particularly proper to the youth of a philosopher. According to such a standard Kant’s Kritvk might have been written thirty years earlier than it was. Another argument is the poetical language of the Phaedrus, which reminds Schleiermacher of the tradi- tion about Plato’s verses which he was said to have burnt when he knew Socrates. This argument is fully refuted by the great number of stylistic coincidences with the Laws which are found in the Phaedrus. Schleiermacher sees also a sign of early date in the triumphant confid- ence of the dialogue. If Plato had such confidence at the age of twenty-five, how could he have lost it in the Protagoras and Meno? This question is left unanswered by Schleiermacher. The contempt of writing, argues Schleiermacher, is unthinkable in a man who has written already very much. But Plato does not despise wniting at all, and he states it expressly (258 D: aavri dyAov Stu ovUx aiaypov aitd ye TO ypddety Noyous) —he despises only bad writing (éxetvo aicxpov 45n, TO pt) KarAdS Adyey TE Kal ypdépew) and the cult of mere literary erudition (275 D: mrsov Tt oldmevos elyat AOyous yeypappdvous Tod Toy sidora dropvioas wept dv dv 9 ra yeypaupéva) which substitutes opinions for knowledge (275 B: moAunxoo yap cot yevouevor Schleier- macher complains of the philo- sophical poverty of the Phaedrus Poetical language. Trium- phant con- fidence. Contempt of mere erudition, not of literature. Mention of Tsocrates is not a sign of un- reserved approba- tion, only @ recogni- tion of his superior- ity toother orators. Solemnity of style. Mention of Sophocles and Euripides used by Ast as chrono- logical indication. 35() ORIGIN AND GROWTH OF PLATO'S LOGIC dvev Sidayijs moAvyvmpoves sivat Sofovow . . . S0€dcodar yeyoveres avi copay), and leads men to spend all their attention on the form, making it impossible for such mechanical writers to have a clear view of general ideas (248 B: qodvy Fyovea movoy dtedeis Ths Tod Svtos Beas amrépyovTat). What Plato wanted, is that anybody who pretended, like Isocrates, to be named a philosopher, should be able to impart to his pupils something better than speeches corrected over and over during many years like the Panegyricus of which that rhetor was so proud. The Platonic Socrates recommends Phaedrus to say that to Lysias (278 D E), but Phaedrus asks whether the same does not apply to Isocrates, and the answer is not in the negative: more talent (279 A: ra TAs ddcews) and a nobler character (#Oe. yevvixwtépe) are not denied to the author of the Panegyricus, but he is left only the first place among orators, not allowed to rank among philosophers until he shall yield to a more divine inspiration. What Schleiermacher quotes besides as a sign of youthfulness, an exaggerated solemnity in some passages, has been demonstrated by Campbell to be a peculiarity of later style. The mention of Polemarchos, Lysias’ brother, as a pupil of Socrates appears to Schleiermacher most probable in the lifetime of Polemarchos, who was poisoned four years before Socrates (Lysias contra LHratosth. §§ 17,18). But Polemarchos is also introduced in a work written long after his death (Rep.). Ast saw in the Phaedrus Pythagorean influence, and a great similarity to the Timaeus (pp. 106-107), but this did not prevent him from following Schleiermacher in identifying the supposed date of the conversation with the date of the composition. He added to Schleier- macher’s arguments only one very curious reason: Sophocles and Euripides are spoken of as living, and therefore the Phaedrus must be written before 406 B.c. Ast did not notice that the same reasoning would lead MIDDLE PLATONISM: PITAEDRUS 351 him to place also the Timaeus and Critias before the death of Socrates. What has been said in favour of an early date of the Phaedrus by Krische” and Volquardsen,”* who has dedicated a whole volume to the subject, is only a para- phrase of Schleiermacher with such insignificant additions as the acute observation of Krische that the death of Socrates is not alluded to in this work (this would rather speak for a late date) or the unfounded fancy of Vol- quardsen that the philosophical contents of the dialogue are purely Socratic. These authors have not thought it of any importance to explain why Plato in the Phaedrus despises poetry or how he could so early have arrived at the conviction of a periodic migration of souls, contra- dictory to the very cautious statements on future life in the Apology, Crito, and all purely Socratic dialogues. A more recent attempt to represent the Phaedrus as written some years before the death of Socrates has been made by Usener ”° and accepted for a time by Wilamowitz- Moellendorff,“° but the latter has expressly revoked this opinion (Hermes, vol. xxx. p. 102). The case of the Phaedrus in one respect resembles that of the Sophist. As Campbell’s investigations on the Sophist have waited thirty years to be at last acknow- ledged by a competent authority as an ‘immortal feat in Platonic chronology,’*' so Thompson’s equally im- mortal investigations on the Phaedrus—published in 1868, 21 A, B. Krische, ‘ Ueber Platons Phaedrus,’ in Géttinger Studien for 1847, pp. 930-1065, Gottingen 1848. 2388 ©, R. Volquardsen, Platons Phaedrus, Erste Schrift Platons, Kiel 1862, 321 pp. 2399 H. Usener, ‘Abfassungszeit des Platonischen Phaidros,’ in Rhewz- sches Museum fiir Philologie, 35 Band, p. 131, Frankfurt a. M. 1880. 20 Wilamowitz-Moellendorff, Philologische Untersuchungen, Band i. p. 213. “1 Th. Gomperz, ‘Die Jowett-Campbellsche Ausgabe von Platos Re- public,’ in Zeitschrift fir Philosophie und philosophische Kritik, Band cix. p. 163, says: ‘Lewis Campbell’s Name wird in der Platon-Forschung unverganglich dauern.’ More recent defenders of the early date of the Phaedrus. Thomp- son’s edition of the Phaedrus still un- known in Germany. Both Thomp- son and Campbell stated their con- victions with great modesty, which produces the im- pression of incer- titude. Thomp- son confined the date of the Phaedrus within the narrow limits of 880-378. This agrees with the above considera- tions. 352 ORIGIN AND GROWTH OF PLATO'S LOGIC but written and read in the university of Cambridge as early as 1859—remain up to the present time a dead letter to continental philologers. Two reasons have acted in this case as well as in Campbell’s: first that on the Continent nobody expects important original investiga- tions to be buried in the Introduction and Appendices of the text edition of a single dialogue ; and second that Thompson, like Campbell, did not use the confident language which is necessary to make an impression on a reader accustomed to the confidence of Schleiermacher, Hermann, Zeller, Teichmiller—and maintained even by such paradoxical authors as Schaarschmidt or Pfleiderer. What Teichmiiller developed into an important chapter of his work, without knowing Thompson, was given by the Master of Trinity College in footnotes, with a modesty which even on a reader accustomed to the incomparable modesty of English scholars leaves an impression of in- certitude. Thompson has made it evident to the attentive reader of the four dissertations accompanying his edition of the Phaedrus (Introduction and three Appendices) that this dialogue must be written after the Panegyricus of Isocrates, that is after 380; and before the death of Lysias, that is before 378. This is such an exact deter- mination of date as 1s possible only for a very few Platonic dialogues. The same argument has been independently and with far greater assurance produced by Teichmiiller in 1881 (Literarische Fehden, vol. 1. pp. 57-82) and has never been refuted. This agrees perfectly with the place assigned by us to the Phaedrus in the development of Plato’s logic, and with the limits of the probable time necessary since the Sympostwm for the composition of the Phaedo, Republic, and Phaedrus. That the Phaedrus must be later than Phaedo and Symposium has been also recently recognised by Th. Gomperz and must be acknowledged by all who know the investigations on the style of Plato which have so completely confirmed Thompson’s view. MIDDLE PLATONISM: PHAEDRUS 353 Yet up to the present time, many eminent German scholars, as Zeller, Susemihl, W. Christ, P. Natorp and others, persist in the opinion that the Phaedrus is earlier than the Phaedo and Symposium, so that some supple- mentary observations on the evidence for the priority of these and other dialogues are perhaps not out of place. As to the Phaedo, the arguments of Schulthess are decisive, and Schedle,™? Liebhold,4? Kassai,244 who advocated the priority of the Phaedrus, were unable to refute them, while Bury *° supplemented them in the best manner. The comparison of the arguments for im- mortality has shown equally that the Phaedrus must have been written after the Phaedo. The priority of the soul to the body appears in the Phaedo (804) as a new thought and ‘is already familiar in the Phaedrus (246 B: Taca i Ypuyn tavros émipedeitar tod axypvyov) ; the theory of reminiscence, which is in the Phaedo mentioned with the caution ‘ei dAnOyns gorw’ (72 E), 1s n the Phaedrus assumed as certain (250 A); that ideas or notions are the substance of things is in the Phaedo a probability (76 D: e@ pev Eotev & Opvrovper dei, KaNov Te Kai ayaboy Kai waca % TocavTn ovcia ...100 B: troOéuevos elvai te Kadov avro xa’ avro . . .), in the Phaedrus the common inheritance of all philosophers (247 c : 4 .. . ovoia évrws ovoa xuBepynTh ove Oat) vp). More important points of comparison are afforded by some characteristic differences between Phaedo and Phaedrus, which show the Phaedrus in agreement with other later works. It has been ob- served by Ueberweg (Untersuchungen, p. 285) that an important doctrine is common to Phaedrus and Timaeus, “2 ¥. Schedle, Die Rethenfolge der platonischen Dialoge Phaedros, Phaedon, Staat, Timaeus, Innspruck 1876, 2 Liebhold, Ueber die Bedeutung des Dialogs Phidon fiir die Platon:- sche Erkenntnisstheorie und Ethik, Rudolfstadt 1876. 24 G, Kassai, ‘ Meletemata Platonica,’ in Egyetemes Philologias Kbsliny, pp. 857-870, Budapest 1886. “> J. B. Bury, ‘Questions connected with Plato’s Phaidros,’ in Journal of Philology, N* xxix. for 1886, AA Relation of the Phaedrus to other dialogues. Relation to the Phaedo. Ueber weg’s ob- servation which led him to suppose 1 later date of the Phaedo can be inter- preted otherwise. Coinci- dences between Phaedrus and Trmaeus as com- pared with Phaedo. Allusion in the Phaedrus to the Phaedo and Syi- posium with preference shown for the Phaedo. 354 ORIGIN AND GROWTH OF PLATO'S LOGIC while not yet recognised in the Phaedo, namely the axiom that what is unconditioned is indestructible, while everything that has a beginning must have an end. Ueberweg was led by this observation to place the Phaedo after the Phaedrus and Timaeus, wherein he departed from his ordinary sagacity and caution, as the natural inference would have been that the Phaedo is earlier, the more so as Phaedrus and Timaeus agree in this respect with the Laws (see above, p. 338), a fact which seems not to have been noticed by Ueberweg. The view of the sense perceptions offers another coin- cidence between Phaedrus and Timaeus against the Phaedo. In the Phaedo as well as in the Symposium true Beauty was inaccessible to the senses (Phaedo 65 D), while in the Phaedrus not only Beauty is accessible to the physical sight (250 D: xddXos . . . dedpo éXOovres Kateidr}- dapyev Sid THs svapyectdatns aicOncews . . . Ppdvycis ovy Opatat KaANOS povoy TavTnv Boxe polpav, war’ éxhavéotaroy slvat Kai épacuimtaroy), but the sense per- ceptions lead to the formation of general notions (249 B : TO KaT eldos AEyOuEvov, &x TTOAN@Y LOY aidOnaewn eis bp hoyioue Evvaipouvpévwv). This agrees with the view expressed in the Timaeus metaphorically (44 B: pos to Kata dvow idvtwv oxjpa skdoTwy TOV KUKNwWY ai TrEpLpopal xatevOuveuevat, TO Te OarEepov Kat Td TavTOY mMpocayopevoucat xa’ opOov, Eudpova Tov tyovta abras yuyvouevov aTroteNovatv). The relation of the Phaedrus to the Symposium can be easily shown by many comparisons, and it is now evident that the Phaedrus is later, though the majority of authors think otherwise. The mention that Phaedrus has been the cause of many speeches (242) in peculiar connection with a similar mention of Simmias (242 B) may with some probability refer to the Symposium, in which Phaedrus is represented (177 a) as the initiator of the series of speeches on love proposed by Eryximachos. This allusion is in so far probable as Simmias named in the same passage has in the Phaedo a principal share in MIDDLE PLATONISM: PHAEDRUS 355 initiating the dialogue on immortality (Phaedo 61 c). And if Plato in the Phaedrus credits Simmias with a greater merit, this means that he preferred his Phaedo to the Symposium, and that he looked on both dialogues as his masterpieces, very superior to speeches of other orators (242 AB: Oetos el mepi tous Noyous, @ Paidpe, cal ateyvas Bavpdoros’ olwar yap ym . . . pndéva trEious fH G8 TreTrolN- xévar yeyevijcOar Tow avTov Néyovta 7) adNAOUS Evi yé TH TpoT@ Mpocavayxatoyvta. Xppiav OnRaiov &Eaipw dAoyou" Trav 6& GNAwy WautroNV KpaTeis). The mention of Lysias’ brother Polemarchos as con- verted to philosophy (Phaedr. 257 8B) might be a direct allusion to the Republic, in which Polemarchos is repre- sented as convinced by Socrates that nobody ought to do wrong to his enemies (Rep. 885). This would be an allusion similar to that which is contained in the mention of Simmias and Phaedrus, and would tend to show that Plato looked upon the persons of his dialogues as more real than their living models, who were dead when he wrote. He says at least that there is more truth in thought than in action (Rep. 473), and he takes many times such a liberty with Socrates that he puts in his master’s mouth allusions to his own written dialogues, or even to his experiences, without any consideration whether such allusions were suitable to the historical Socrates. Teichmiiller sees (ii. pp. 22, 272) in the erotic speech of Lysias allusions to the speech of Pausanias in the Sym- posium, and believes the speech to have been written as a criticism of the Symposiwm by Lysias, thus provoking Plato’s pitiless criticism in the Phaedrus. This ingenious supposition, if it could be proved, would sufficiently ex- plain why Plato selected just this speech of Lysias as a sample of bad rhetoric, and why he criticised it with more than usual insistence and irony (248.c: dvasdds eipnoOov ta éy vavra.s qrov TeOpaypévwv Kat ovdéva ErevOepor Yowra twpaxdtwv). The parallel passages quoted by Teich- miller deserve our attention, but they seem not to be fully AA 2 AOyw ... Mention of Pole- marchos’ conversion might refer to the Republic Teich- miiller’s supposi- tion about the speech of Lysias uncertain unless new evi- dence is forth- coming. View on poetry in both. Com- parisons with earlier dialogues super- fluous. Thomp- son’s deter- mination 356 ORIGIN AND GROWTH OF PLATO'S LOGIC sufficient to prove his supposition (Symp. 188 © compared with Phaedr. 281 4, 184c with 233 a, 182D with 2344, 2188 with 231D). These allusions are not quite evident, but they might be confirmed if some independent testi- mony about Lysias’ Eroticos should ever be found: therefore they deserve to be remembered. The relation between Phaedrus and Symposium appears also in the mention occurring in the Phaedrus that physical beauty provokes an admiration which can become a germ of per- fection (251 a): this seems to refer to the corresponding explanation in the Symposium (210 4). But the most decisive argument for the priority of the Symposium turns on the difference of views about poetry. In the Symposium poets are still esteemed, in the Phaedrus the poet takes one of the lowest places, and Homer is parodied (252 8B) by two verses in which the inconstancy of his gods is ridiculed. It would be easy to show in the Phaedrus also many points of comparison with the Cratylus, with which it has in common a certain etymological tendency, with the Gorgias, which Thompson showed to be’earlier when the majority of German scholars were still of the contrary opinion, and with other dialogues. But the priority of the Gorgias has been lately recognised by some of its former opponents, especially by Zeller, and has been made evident also by Natorp, Siebeck, Duimmler, after Socher, Stallbaum, Hermann, Steinhart, Susemihl, and Ueberweg, so that it may be admitted as sufficiently proved.” Thus Thompson’s determination of the date of the Phaedrus as written between 380 and 378, or about 379 B.c., is confirmed in every respect, and not the least important of all these confirmations is given by the stylistic investi- * The recent attempt of Gercke (Platons Gorgias, erkldrt von Sauppe, herausgegeben von Gercke, Berlin 1897) to prove that the Phaedrus preceded the Gorgias is based on the assumption of uncertain allusions to writings of other authors, and without regard either for the philosophical contents or for the style of these two dialogues. See above, note 286. MIDDLE PLATONISM: PHAEDRUS 357 gations. Already Campbell found in the Phaedrus a sur- prisingly large number of words common to the latest three dialogues, exceeding in relation to the size not only the number of such words to be found in the Phaedo and Symposium, but even those of the Republic, Sophist, Parmenides, and Philebus. This peculiarity of the voca- bulary of the Phaedrus has been since outweighed by other peculiarities observed, so that in our list the Phaedrus exceeds in stylistic affinity with the latest group only those works which are really earlier, as the Symposium, Phaedo, and equal samples from the Republic. The only part of the Republic which has a slightly greater number of important peculiarities of later style than the Phaedrus is the picture of the philosophers in B. VI-VII. But the difference is too insignificant for chronological con- clusions (116 peculiarities equivalent to 234 units of affinity on 44 pp. in Rep. VI-VII against 118 peculiari- ties equivalent to 220 units on 39 pp. in Phaedr.). The more so since only the greater frequency of peculiarities occurring is superior, and not their number. This might be a consequence of the much more varied contents of the Phaedrus. If we compare the peculiarities of later style found in this part of the Republic only and absent from the Phaedrus and all earlier dialogues with those found in the Phaedrus and absent from the Republic, we see that the Phaedrus notwithstanding its smaller size has more exclusive affinities with the latest group than the latest part of the Republic : : Peculiarities of later style found in Rep. 471 c-541 B (443 pp. Did.), Phaedrus (89 pp. Did.), and in and in no earlier dialogue nor in no earlier dialogue nor in the any other part of the Republic. Republic. wavrws xal mavry (827) once wavry wavres (828) once épp7n6n (886) once elpnrat (824) once ra mavra eidn (861) once rd £vpray (868) once pupip (829) once yévos as a logical term (24) dvéravia (470) once driavos (476) once of the date of the stylistic evidence. Only B. VI-VII of the Republic might be later, though this remains uncertain. Date of the Phaedrus about 379 B.C. The period of Middle Hatonism produced ax much us One half of the amount of text written after- wards, 358 ORIGIN AND GROWTH OF PLATO'S LOGIC Rep. (continued)— Phaedr. (continued) — ovocia =complexus omnium rerum adjectives in ros formed of sub- (288) once stantives (6) repeated aximnros (469) repeated re, adding a third phrase (283) repeated dvw (195) repeated Gpotwpa (468) repeated great scarcity of answers denoting subjective assent (318), import- ant Interrogations by ri prevailing over those by mas (452), im- portant Thus it is probable, though not yet certain, that the Phaedrus is later than the Republic, taken as a whole, and it is quite certain that the last three books of the Republic preceded the Phaedrus. This results both from stylistic comparisons and from the comparison of con- tents. At all events the date of the Phaedrus as written about 379 B.c. (380-378) is now quite as well confirmed as the date of the Symposiwm about 385 B.c. Middle Platonism We have seen that in the time between 384 and 378 B.c. Plato dedicated his leisure only to the Republic and Phaedrus, all other works being either earlier, as has been already shown with respect to those preceding the Republic, or later, as will be seen in the continuation of our inquiry. This short epoch of middle Platonism lasting up to Plato’s fiftieth year produced, therefore, an amount of text equal to one half (233 pp. ed. Did.) of all the works written in the remaining thirty years of the philosopher’s life (476 pp. ed. Did.). Thence it results that Plato’s literary activity was on the decrease after the Phaedrus, and that he followed the maxims expressed at the end of this dialogue, according to which writing is by no means the most important of the aims of a philo- sopher, in contradiction to his rival Isocrates, to whom MIDDLE PLATONISM 359 nothing appeared more important than his written speeches, in which he pretended to teach also a philo- sophy, condemned by Plato. The doctrine of the ideas, invented in the first period after the foundation of the Academy, is maintained during the time of middle Platonism, but the same stress is no longer laid on the independent existence of the ideas, and the relation between particular things and the ideas, first designated by the term peréyewv, becomes a mere similarity (udunpa, pspetoOar, opolmpa, opocodv), which allows us also to form ideas by the observation of simi- larities in sensible objects. In several passages the ideas and knowledge appear as created by the philosopher, though the earlier conception of a vision of self-existing ideas is not yet wholly abandoned, and reappears in the myth of the Phaedrus accompanied by its logical inter- pretation, according to which the ideas become identified with general notions. It is fully in accordance with this later stage of the doctrine, that ideas are no longer limited as in the Symposium and Phaedo to ethical and mathematical objects, but are equally supposed to exist for manufac- tured things. Thus a transformation of the primitive theory of ideas is already prepared though not yet carried out. While the ideal of the first Platonic stage was a state of subjective perfection and separation from the vulgar surroundings of common life, a passive contempla- tion of ideas, we see in middle Platonism an increasing confidence in the necessity of applying philosophy to life, and also of investigating particulars. The search for definitions was a Socratic inheritance, but the fondness for classifications appears not earlier than in the Republic, though it is prepared by the Phaedo. This direction taken by Plato had a great influence on the development of his logic. So long as only definitions are sought for, the supramundane independence of ideas can easily be maintained. But once on the way of Doctrine of ideas during Middle Platonism developes into the view of a mere similarity between things and ideas, Ideas not limited to ethical or mathe- matical notions. Applica- tion of philo- sophy to life. Logical classifi- cations lead to a recogni- tion of the subjective element in ideas. Ideas in- depende nt of par- ticulars but not outside individual conscious- ness. Logical independ- ence of ideas the founda- tion of science. Plato’s objective idealism lasted a short time. 360 ORIGIN AND GROWTH OF PLATO'S LOGIC systematic classification it is impossible not to observe the subjective character of subdivisions, and this leads to the conclusion that the existence of ideas is only possible in a soul: not necessarily the soul of the thinker, but a soul of an individual being. The objectivity of ideas resulting from an agreement between souls is different from the objectivity based on the feeling produced by passive contemplation. The first impression of a philo- sopher who notices the distance between an idea and the particulars 1s to exaggerate the objectivity and independ- ence of the idea, and to assert emphatically its independ- ence and incommensurability with the particulars, which seems to imply its existence outside individual conscious- ness. The belief in its independence of particulars is lasting, because it is true, and has been proved by Plato in the Phaedo and in all following works, remaining the cardinal truth of all later philosophy, ignored only by thinkers who were not sufficiently versed in the history of logic, like Comte and Mill. But the existence of ideas otherwise than in some individual consciousness is an illusion, similar to that more familiar illusion which makes colours and sounds appear objective, though they have no existence outside of us. The illusion of objective idealism is, however, one of those illusions which are necessary steps in philo- sophical progress. It is only a metaphoric expression of the truth that ideas are logically independent of the individual, and this logical independence (dvdyxn) must be recognised as a foundation of objective knowledge and science. Thus Plato at the beginning of his logic and during the middle period of his literary activity was idealist: he believed in the objective existence of the ideas outside particulars and outside the individual soul. This belief found its clearest expression, at the beginning, in the Symposium and in the Phaedo, combined with a vague uncertainty as to the relation between things and ideas. MIDDLE PLATONISM 361 During middle Platonism, so far as we can guess from the hints given in the Republic and Phaedrus, the same conviction was maintained with a clearer definition of the relation between things and ideas as consisting in their similarity. But less emphatic stress was laid on the independence, and if due allowance is made for meta- phorical language, the whole mythical part of the Phaedrus may well agree with a conception of ideas in the meaning they had for Kant. We need not fear to deprive Kant of his originality if we come to the conclusion that Plato towards his later age understood the ideas in very much the same way as Kant. The truth is one, and once found cannot be changed. There is no impossibility or even improba- bility in supposing that a thinker like Plato, having no other aim in his life than thought, arrived at a correct notion of ideas after a long educational career. It would be astonishing to find the contrary. And Kant cannot lose any substantial merit in consequence of this discovery, as the notion of ideas forms only one of the points of Kant’s philosophy, while in many other points he progressed, as might naturally be expected, beyond Plato and other philosophers. There is one very striking analogy between Kant and Plato. Kant undertook a critical reform of his earlier convictions after having reached the age of fifty, and the same was the case with Plato. It is not surprising that philosophers arrive late at the full maturity of their thoughts. Every more perfect being requires a longer development, and men’s childhood lasts longer than the childhood of inferior animals. A philosopher in Plato’s Opinion must excel other men almost to the same extent as any man is superior to other animals. This is not an extraordinary pretension, if we bear in mind that for Plato the activity of a philosopher is by no means limited to abstract thought, but extends to all departments of human life; so that he would certainly have included in Already less certain in the Republic and Phaedrus. Coinci- dence between the later stage of Plato’s thought and Kant not acci- dental. Analogy between Kant and Plato. 862 ORIGIN AND GROWTH OF PLATO’S LOGIC this class some of our contemporaries, not asking them to write philosophical dissertations in order to legitimate their pride in belonging to the ruling class of mankind, formed of more perfect beings than the average citizens even in an ideal state. 363 CHAPTER VII REFORM OF PLATO’S LOGIC WE have seen in the above exposition of middle Platon- ism a theory of knowledge according to which the ideas were perceived by intuition, and constituted eternal models of everything in the phenomenal world. The chief point was the independence of ideas, not involving, however, their separate existence. That no phenomenal appearance can fully correspond to a pure idea is a great discovery of Plato, made by the consideration of mathe- matical as well as moral notions. Whether such ideas have any existence out of the human mind, or generally outside an individual consciousness, was a question not discussed, and perhaps not clearly formulated by Plato: when he speaks of the beauty of ideas outside the physical universe, he does it in such metaphorical language, that we cannot draw certain inferences from his images. The true meaning of all these visions is the conviction that ideas are independent of material things, and that the existence and changes of physical objects must be ruled by immaterial and invisible ideas, often spoken of as objects of thought. The relation between things and ideas—whether defined as a presence or immanence of ideas in the things, or as a similarity between things and ideas, or as an imitation of ideas by particulars— was the first question that occurred when once the existence of the ideas had been established. While a personal training was deemed necessary in order to attain the vision of ideas, their Ideas perceived by in- tuition and inde- pendent of appear- ances. Substan- tial ex- istence of ideas not certain. Existence of ideas evident for the initiated. Once their existence recognised their order and hierarchy deserve the at- tention of the philo- sopher. Classifi- catory tendency appears at a later stage. The dia- lectical dialogues carry out the pro- grantme 364 ORIGIN AND GROWTH OF PLATO'S LOGIC existence needed no other proof than the personal ex- périence of the initiated. This initiation by means of mathematical, astronomical, or musical studies, and subsequent discussion of political or educational problems, proved a sufficient aim for many years of teaching. But at last a new problem became inevitable. Suppose we have arrived at the intuitive knowledge of many ideas, and are aware of the difference between an idea and a particular object of sensible experience, the next question to ask is about the order of ideas and their mutual rela- tions. These can be well explained only through a distinction of similarities, leading to an universal classifi- cation of notions. Already in the Republic it was asked how many kinds of reasoning are possible (532D: ris o Tpotos THs Tov StadéyerOar duvapews, cal Kara rota dn Eidy Siéornxe), but the question was left unanswered. In the Phaedrus (266 8) the complete classification of ideas from the most general kinds down to the indivisible logical units was proclaimed as the chief aim of the dialectician. This classificatory tendency is absent from earlier works, where specific problems were discussed, without any allusion to a contemplation of all time and all exist- ence, which we find first in the Republic (4864). But even in the Republic the classifications and divisions are limited to a few subjects, and no attempt is made to bring all the possible objects of knowledge under a certain number of heads. Nor is this fully carried out even in the Phaedrus, where the importance of such a logical method is so warmly insisted upon, and the power of building up general notions and dividing them is pro- claimed divine. An attempt to realise this programme is made in the series of dialectical dialogues, among which the Theaetetus and Parmenides are the earliest, as we have seen from stylistic comparisons, which are confirmed by the exami- nation of their logical contents. They share with the REFORM OF PLATO'S LOGIC 365 later works of this group another important peculiarity, the historical method of comparing impartially and judging according to their merits the theories of other philosophers. The primitive theory of ideas is no longer the object of such ecstatic admiration as in the Sympo- swum and Phaedrus. It 1s subjected to a critical exami- nation in the Parmenides and almost ignored in the Theaetetus, so much so that many readers have believed this to be an early dialogue. This impression vanishes at once upon a close consideration of some philosophical terms familiarly used both in Theaetetus and Parmenides which had been elaborated during the period of middle Platonism. To these belong the notions of dialectic (Theaet. 161 £, Parm. 135 c), of substance (ovcla, Theaet. 186 p, Parm. 135 A), power or faculty (8vvauis, Theaet. 158 z, 159 a, 185 c, Parm. 133 £, 135 c), the one (Theaet. 152 p, Parm. 187 c, &c.), Not-Being (Theaet. 185 o, Parm. 142 a), and the opposition of activity and passivity (Theaet. 157 a, 174 B, Parm. 138 B). Both Theaetetus and Parmenides have further in common two important distinctions, which could not have been ignored in the Republic, nor in the Phaedrus, if the author had already become familiar with them. One of these is the well-defined notion of movement, including qualitative alteration as well as change of position in space. This meaning of «ivnots, accepted by Aristotle, and many later philosophers, is a result of the increasing importance of this notion for Plato, and would necessarily have been alluded to in the Republic and Phaedrus in those passages in which «évnois is used in its primitive signification of movement through space. It is a far-reaching generalisation to identify movement with qualitative alteration, because both are a manifestation of change. The comparison of corresponding passages shows that this unity was not yet noticed in the period of middle Platonism : of the Phaedrus. Theory of ideas ignored in Theaetetus criticised in Par- mendes, In both occur terms elaborated during the preceding period. Both Theaetetus and Par- menides contain & new notion of move- ment, including change of position as well as of quality. The dis- tinction of two kinds of move- ment, first introduced in the Theae- telus, stated to be a personal discovery of the Platonic Socrates. 866 ORIGIN AND GROWTH OF PLATO'S LOGIC Rep. 454. D: éxeivo ro ei8os Tis G\NoLooEws re Kai duowmoews povoy €pvdrarropey rd mpos avra rewov ra emirnded- para. 580 c: mreio etd mapéxera 1) opd.. . 588 E: rd ye ndv év Wun yryvopevoy xal rd Aumnpdy Kimois tis auorépw error. Phaedr. 245 oD: Kivnoews apxy . . . oor” droAAvoba otre yiy- ver Oar Suvardv, } mravra re ovpaviy nacdy re yéverwy = wUTTEWOVC'AY orivas Kai prrore avis éxyew bev = eu Oévra yevnoerat. dfwov Erepov eiSos advar xujoews ; Theaet. 156 A: xiw- cews 800 €i8n, B8vvapey 8é rd pev moteiy Exor, ro 8€ mdcxew (quoted as & view to be criti- cised). 181 GO: mérepov é€v re el80s avrns Aéyovow 7 Gowep uot aiverar 800; py peévroe povoy énoi Soxeirw, GAAa oup- perexe kai al, . . . dpa xivetoOac xadeis Gray Te Xopay ék x@pas pera- BadrAn 7 Kal ev tp a’r@ otpepyra ;— eywye — ... Grav b€ 7 pev ev TE air@, ynpdokn 8c, Fj péAay ex Nevxod } oKAn- pov éx padaxod yiyvyrat, # twa GdrAnv addoio- ov addAowrat, dpa ov« ... 600 87 heyw rovrw el3n Kunoews, adroiwor, rHv de hopav. 158 a: rd peév elvas cal rd yiyverOa ximors rapéxet, rd 8 py elvas Kal amrddAvaGat Hovxia. Parm. 188 B: xvov- pevdy ye éepoiro Fj addowoiro dy + abrat yap povat Kiwnoers—vai, 162 D: ovx dpa rd év py dv orpeper ba dv duvaro ev exeivep ev & ph €orw ... ovdé pny d\Aowovrai mov rd ey €avrov, ovre rd by odre ro pn Ov. ef 8é parr’ addootrat unre €v rabr@ oTpéperat pre pera- Baiver, dp’ dv mp ére Kwotro ; mas yap; Legg. 894 &: Gray aird avro Kwhoay €repov ad\Xotwon, 7rd 8 Erepov Go... pay apxn tis avray éorat rhs KUWT- cews dragons aAAn mAnv THs auTns avrny Ki- modons petaBorAn ; We see that in the Republic the distinctions introduced in the Theaetetus are not yet known. The use in the Republic of xivnots in its metaphorical meaning as move- ment of the soul is transitional to the later generalisation, but does not yet imply it. In the Phaedrus xivyots means movement through space, and this is very charac- teristic if we remember that in the later dialogues the distinction of two kinds of movement is represented as quite essential. This distinction is first made in the Theaetetus, and recurs as familiar in the Parmenides and Laws (where it is assumed as a matter of course that the first movement produced is a qualitative change) as well as later in the works of Aristotle. The distinction of two kinds of movement is introduced in the Theaetetus as a new theory, after another division had been incidentally REFORM OF PLATO’S LOGIC 367 referred to. It is stated expressly to be a personal dis- covery of the Platonic Socrates, which he is anxious to see accepted and to share with others the risk of an error (cuppétexe wal ov). After its acceptance, it is repeated as logically necessary (181 D: dvayxaiov). The starting point of this theory was the recognition of movement as a principle of Being, justified in the Phaedrus, mentioned as known in the Theaetetus, and finally reconciled with the stability of Being in the Sophist. This discovery is related to the increasing interest for physical science, which is manifest through Plato’s later works, while it is absent from his earlier writings. It need hardly be observed that here we have not to do with such an ephemeral distinction as between ioris and eixac/a in the Republic, but with one of the greatest generalisations of philosophy, continually discussed by later thinkers up to Trendelenburg and Lotze. It is one of Plato’s wonderful anticipations of ideas which have been better explained only in modern times. The identification of physical movement with qualitative change is a truth which could scarcely be fully realised before Kant, and yet it is taken for granted in the Theaetetus, Parmenides, and Laws of Plato. Had the Theaetetus and Parmenides, being the two most critical works of Plato, no other new theory than the inclusion of qualitative change and physical move- ment under one primary kind, with the subtle sub- division of physical movement into a movement through space, and revolution on the same spot—this would alone be a strong reason for placing them after Republic and Phaedrus. But we find in these two dialogues another theory of cardinal importance, yet introduced quite as incidentally as the theory of movement. In the time of middle Platonism the favourite examples of ideal existence were moral or mathematical notions, the former being specially fit for allegorical representation as objects of enthusiastic vision. When the first enthusiasm was It is one of the great generali- sations of philo- sophy. Implies aubjec- tivity of space. List of categories first attempted by Plato in the Theae- tetus. Enumera- tion of highest kinds indepen- dent of esthetical considera- tions. A very important step in philo- sophy. Its im- portance under- stood by Plato. 868 ORIGIN AND GROWTH OF PLATO'S LOGIC over, it became very natural to attempt a general enumeration of highest kinds, independently of the esthetical feelings of awe and admiration which first led to the perception of such ideas. This problem of categories has remained ever since a permanent depart- ment of philosophy and has been cultivated from Aristotle onwards by all logicians. But the first table of cate- gories in the history of logic is found in Plato’s Theaetetus, repeated and enlarged in his Parmenides and Sophist. It is not wrapped in such emphatic language as the sovereignty of philosophers or the precept that to suffer wrongs is better than to inflict them. It is the historian’s duty to show the incomparable importance of this first step in a new direction. There is reason to believe that Plato was conscious of this importance, though he did not insist on it, because he felt the incompleteness of his table of categories (ra xowvd). The enumeration in the Theaetetus is introduced at a culminating point of the dialogue, and followed by ‘ an unwonted outburst of admiration’ (Campbell, Theaeé. p. 160) of the pupil who discovered it; also by the significant observation that a long discussion has been avoided by this happy intuition, a result of good natural capacity (144 B) and a training in mathematics, music, and astronomy (145 a) according to the precepts laid down in the Republic. A careful comparison of similar passages in later dialogues and of Aristotle’s account of the same problem shows very clearly that the first attempt at such an enumeration is that occurring in the Theaetetus, not, as has been sometimes supposed, that in the Parmenides. The list is increased by some notions in the Parmenides and Sophist : Theat. 185c: 4 8¢| Parm. 1864: xpn{ Soph. 254p: peéy- 3) 3: rivos Sivayis té|oxomely ... el mroAAd|tora roy yevav . . . +d re by atrd nal ordots cal xivnots. E: 76 re ravrdoyv xa Odrepov. r édmi maou xocvdvidor:... xat av el py caird émt rovros dndot|¢ort modAd ... Kai go, Prd Zorcy dnovo-|. .. el dorw dpordrns pafes xaird obxtoriy|. ..xalmepl dvopolov nal & ved) qpwrapev|... xiyvnoews kal ord- REFORM OF PLATO'S LOGIC mept avray ;—ovciay Ayers kai TO py elvat, kat «dpotornra- xa avopovdrynra, Kai To ravrov re xat TO €re- pov, ére 8€ év nai rdv Gov = apiOpov ss trepi avrayv. Syrov 8é€ ore Kat Gpridy te Kat wep- (TT Ov épwras Kai Tad\Aa 60a rovrots émerat, dia rivos oTé Taev_ Tov owparos TH Yux7 aic- Gavdépueba ;—trépev axo- Aovbeis, kai é€orw a €pwT® aura ravra. wews, Kal mept yeve- a n a weas cat POopas, kat ~ ~ > Wept aUTOU TOU elvat kal . ? rou py e€tvat Kat €évi ’ A a ee Ady, mept drov dv dei e ~ Ly w ‘ e Ur00n ws Svros Kal os ovK Gyros «kal @do rddos macyxortos, Set oxomety ra ~EvpBac- vovra mpos aurdé kai mpos Ld “ a” éy exagTov Tov G\dov. v 129 E: ra ci3n, otov Opolvornrda Te kai avo- pocotnra kal wAnOos kai TO év Kal ordowy OTLOUY 6 cat Kivnaoev, 369 Aristoteles Categor. 1 b25: ovcia, rocov, TOLOY, POS Tt, TOU, TOTe, xetoOar, eye, srocety, mwdoxetv. Metaphys. 1029 b 24: wowWyv, moacyv, Tore, tov, xiynots. See also below, p. 480, on the categories in the Timaeus, produced by the movements of the soul. The first place is given in all enumerations to sub- stance and Not-Being. The same and the other, and similarity and dissimilarity, are also common to the three enumerations. One and the many form a third pair in the Theaetetus and Parmenides, but are dropped in the Sophist. A fourth pair is movement and immobility, omitted in the Theaetetus, but appearing both in Parme- nides and Sophist. The differences are not necessarily due to a change of views, but to the incompleteness of enumeration, also frequent in Aristotle, who often men- tions only six Categories even in passages where it would seem that the enumeration might be complete. These highest kinds, which denote what is common to many particulars, are different from the ideas admired in the Republic. There is no place among these common notions for Truth or Beauty, nor for the idea of Good, though these are mentioned as also perceivable by the soul alone (186.4). These are not entirely supplanted by the new ideas, but they no longer attract the philo- sopher’s chief attention. The intuitive vision of trans- cendental ideas is exchanged for a discursive investigation of a given universe. This may be explained by the natural evolution of Plato’s activity in his Academy. BB Differ- ences in the three lists. Among categories Truth, Beauty, the Good omitted. Variety of actual ex- perience had to be submitted to classi- fication. Extension of the field of thought beyond the limits of moral ideas. Reform of Plato’s logic carried out in the 370 ORIGIN AND GROWTH OF PLATO'S LOGIC The training recognised to be necessary in order to develope intuition had to be directed, and the variety of material appearances, at first despised as irrelevant, had to be considered and classified. The astronomical and mathematical studies recommended in the Republic tended to promote not only dialectical ability, but also some recognition of sensible experience, and of the reality underlying physical phenomena. If in earlier times the power of the soul over the body was chiefly seen in moral determinations, it now appeared that the body, though subordinate to the soul, is a useful instrument for the purpose of increasing even ideal knowledge by forming new ideas. The moral ideas, being few in number, afforded no sufficient scope for the dialectical tendency to distinguish and classify. The field of logical exercise was first extended to a classification of states and men; but even this did not satisfy that philosophical curiosity which is accustomed to consider all substance and all time, neglecting nothing, however small or insignificant it may appear to the vulgar mind. Among such pursuits, which seem to have occupied the greatest part of Plato’s time after the Phaedrus, the general problem of knowledge was reinvestigated, and this led to an important reform of earlier logical con- ceptions. Of this reform we have a record in two works which more than any preceding them may be termed critical, though at first sight they appear almost as in- conclusive as the Socratic dialogues. These works, the Theaetetus and Parmenides, are of decisive importance for an appreciation of Plato’s philosophy, and deserve our attention not only for their main subjects, but also for seemingly casual allusions to doctrines of the greatest gravity. REFORM OF PLATO'S LOGIC: THEAETETUS 371 I. The Theaetetus. (Relative affinity to the latest group, measured on the Laws as unity, = 0°32; see above, p. 177.) The aim of this dialogue is a definition of knowledge, which, however, 1s not given, in spite of several unsuc- cessful attempts made by Theaetetus. Among the defi- nitions which are recognised to be insufficient is one which had been provisionally received in some previous dialogues: namely, that knowledge is true opinion founded on sufficient reasons. This had been proposed in the Meno (98 a) and tacitly admitted in Symposium (see above, p. 238) and Phaedo, whereas it is refuted in the Theaetetus (210 4): Phaedo 968: moAXdxts éuavrov iiv@ xatw peréBaddov oxonay .. .« éx pynuns xal 80ns AaBovans rd npepetvy Kata rattra = yiyverOat emiornpny. Theaet. 210 a: odre dpa aic6n- ots, obre Sd£a aAnOis obre per’ aAn- Bois 8dEns Adyos mpooytyvdpevos émotnyn av ein. In the Cratylus (426 a), Symposium (202 a), and Phaedo (76 B) Aoyos had the meaning of a sufficient reason, while here it is more exactly analysed, and each of its three meanings is shown to be incapable of changing opinion into knowledge. What Plato’s real conviction about know- ledge was, is known from the Republic, and also from later works : for him the difference between opinion and know- ledge ultimately consisted in the difference of their objects. In this respect there is no change from the Phaedo to the Theaetetus: the activity of reason is an activity of the soul, not wanting the help of the senses and of the body: Phaedo 65 Bo: 4 uy ris dAnGetas dmrerat . . . év r@ doyiferOa.. . Aoyiferar 8 yé mou rére kadXora, Srav Ort padwora avry Kab’ avriy yiyynra dioa xaipew ro cdma. Theaet. 186D: év peév dpa rots mwabnpacty ovK tu éemornpn, ev dé T@ wept exetvov cvdAdoyiop@: ovolas yap xal ddnOeias évravda pév, os Zouxe, Suvvardv dao bat. BB 2 Defini- tion of knowledge sought in the Theae- tetus not given. Difference between opinion and knowledge consists in the objects to which they refer. Know- ledge is acquired by the soul’s own activity. Unity of conscious- ness indicated in the Republic is here more clearly expressed. Senses instru- ments of the soul. 872 ORIGIN AND GROWTH OF PLATO'S LOGIC The same term is repeatedly used in both dialogues (avrn Kal’ airny Phaedo 65 c, 79D, 8384, Theaet. 186 4, 187 a) to denote the soul’s independence of the body. Also the distinction between attaining knowledge and possessing it is already prepared in the Phaedo. Phaedo 75D: ro yap eidéva ‘Theaet. 197 c: Spa 87 wai éme- rovr’ é€oriv, AaSdvra rou émornunv ornunv ei Suvardv oUTw KeKTNMLEVOY Exe al py aro\wdexévat. py eye. But it is only here that the unity of consciousness is insisted upon, as resulting from the variety of perceptions. It had been already observed in the Republic that each sense is used only to convey one kind of impression. This observation is here generalised and affirmed as certain : Rep. 852 £&: @78 drm dv Ado Theaet. 185 A: & di’ érépas Bors } 6@Oadpots;—ovd dpra-—ri be; Suvdpews aicOdver, advvarov etvat dxovoas GAM 4} daiv ;—avdapas: 80 GAAns rar’ aiabérOa, oiov & di —ovxoty 8xaiws av raira rovrwy axons, 8 dWews, 7} a SU Beas, 80 daipev epya elvac; mavu ye. axons ; ATT C: A€yw Bu Kai dxonv a ’ ? >” ’ a ‘ , ‘ ? ray duvdpewy eivat, et dpa pavOaves 6 BovAoua éyew Td etdos. There is a certain progress in the formulation of this principle from the Republic to the Theaetetus. In the earlier work the term évvays as appliable to the senses was first introduced ; here 1t 1s used without hesitation, and the observation that it is possible to see only by means of the eyes is supplemented by the general rule: it is im- possible to perceive through one faculty the proper object of another sense faculty, as can be verified through the familiar example of sight and hearing. The application of this law of specific energy of the senses, given in the Theaetetus, goes far beyond what we found in the Republic and Phaedo. Already in those earlier works the senses were defined as instruments used by the soul, and this is here maintained : REFORM OF PLATO'S . co Phaedo 79 ©: n uxn, Grav T o@patt mpocxpyrat eis ro axoreiy te fh Sia rou dpay fh da rov > ’ “a 9 4 ? U dxovewv 7) Oe GdAns Tivds aicOnoews —rovto yap Sia row ca@paros, TO 80? aicOncewy oKoretv Ti—rore .. . mwAavarat. Rep. 508 B: dupa . . . mrdw- edeorarov Tay mepi tas aicOnoes Opyavarv. > \ €oTt TO opme LOGIC: THEAETETUS 373 Theaet. 184 c: oxére yap, amoxpiots motrépa dpOorépa, @ Op@pev, tovTo elvac odbadpous, jj 3: ob Spaper, kal @ adxovopev, dra, a 8 ’ f° id aa +) b cd 7 80 ov axovopey ;—de bv Exacta aig Oavoueba, Eporye Soxet, paddov f ois'—Aewov ydp mov, ef moddAai Ties év nuiv, Gomep ev Sovpetos @ 9 ? » é 3 4 imo, atoOnoets eyxaOnvrat, adda A} % 4 & 947 ww A pn eis play ria idéay, etre Wuxny ww ” ”~ a , a etre G te Set xadeiv, mavra ravta , t ? Evuvreive, 7 Sua rovrwy oloy opyavar atcbavopeba doa aigGnra. But we find here a new conclusion, not thought of before. If all senses are but instruments, they must be the in- struments used by one and the same thing, be it named soul or otherwise. In earlier works Plato used the term soul as free from every ambiguity. Here we see already a trace of doubts about the existence of the soul, against which he guards himself by the caution that it does not matter whether we call by the name soul or otherwise that substance which is the necessary recipient of all particular impressions. A further proof of the existence of this substance and its peculiar activity is given by the argument that impressions of different senses are comparable among themselves, and no single sense could bring about these comparisons. If we think about two different perceptions of two different senses, this could not be done by means of one of the senses concerned (185 A: ef re wept dudotépwy Stavosi, ovn dv dia ye tov étépou dpydvou, ov8 ad bid Tod érépou rept appotépwy ataOavoe av). Plato proceeds to give well-chosen examples of thoughts, which are possible with reference to different per- ceptions. He observes in the first place that all perceptions have in common existence (185 a), then that they differ from each other, and are identical each with itself (185 a), then that each of them is one, and both are two (185 B), and finally that there may be similarity or dissimilarity between them (186B: «ire dvopoiw eira opolw adAndowW). Concep- tion of the soul developed. Its power of com- paring impres- sions of different senses. Attributes of dif- ferent per- ceptions form the list of categories. No special faculty for per- ceiving categories. They are recognised by the soul alone; though this is a truth not easy to prove. Illusions of the senses difficult to dis- cover. Colour and heat explained as result- ing from motion. Traces of physical 374 ORIGIN AND GROWTH OF PLATO'S LOGIC This enumeration of general notions which can be applied to a variety of concrete objects is not accidental, because it is repeated by Theaetetus nearly in the same order, and forms really the most ancient table of categories. Plato asks by what faculty the soul can perceive those general notions. The answer that such general notions can be known only immediately by the soul’s own activity (185 DE: aury Sv abtijs? uy? Ta Kowd pot haiverar reptdvtwy emicKoreiv) is recelved as a truth which can be at once understood only by the better class of intellects, and would require a long proof, had not this been made superfluous by the natural capacity of Theaetetus (185 E). These general notions, here distinguished as the proper object of knowledge, are placed in close relation to the particulars observed by means of the senses, and this denotes a change in Plato’s attitude towards physical phenomena. He no longer despises them as in the Phaedo and Republic: he recognises the difficulty of discovering the illusions of the senses (179 c: qepi d# 76 tapov éxaoTe mados, && dv ai aicOjoeas cal ai Kata Tavras d0£at yiyvovrat, Naretrwrepov srsiv ws ovK adnGets). He has made a very special study of these appearances and has arrived at sur- prising intuitions of physical truth. Thus for instance he states clearly that colour does not belong to objects out- side us nor even to our eyes (153 D). That light is a result of movement and affects different persons in a different way, and that itis a pure quality out of space, appears to be a truth attainable only by the methods of modern physics, and yet any reader can find it in the Theaetetus (153 EB: pmdé tw’ adt@ yopay arrorad€ys). Another of the great discoveries of our own time is here anticipated, the explanation of heat as a mode of motion (153 a: 76 Oeppov te xal rip, 3 8) Kal Tara yevvd Kab éaretporever, avro yevvata: tx opas Kal tpiyews* rodro 58 Kivnots). This is certainly said with another meaning than it might have for the modern reader. But it betrays the fact that REFORM OF PLATO’S LOGIC: THEAETETUS 3875 Plato had already begun those physical reflections which led him later to the theories expounded in the Timaeus. It seems that a thorough-going materialism had made its appearance within the Academy or outside it and decided him to a full refutation. For the Theaetetus, no doubt, is meant above everything as a refutation of materialism and sensualism. The materialists are men- tioned as very uneducated men, not initiated into the mysteries of a refined philosophy (155 E). With these are contrasted the subtler sensualists (156 A: xopdrepor, Ov wé\Xw Gor TA pvoTHpia Aéyew) Who explain everything by movement and make everything relative, destroying thus all fixed notions, which are indispensable in laying the groundwork for a system of science. Plato seems to admit so much of their theory of the relativity of sensa- tions as agrees with his own views. He argues that the reality of dreams for the dreamer 1s equal to the reality of waking for men awake (158 cD), and he leaves the difficulty for the time unsolved. The same might be said of illness (158 p) and madness (157 §£), but only in so far as sensations are concerned, which have always a subjective character (1544: 4 ov Suioyupicaio ay ws, olov col daiverar ixaotov xpaua, TovdToy Kal Kuvi Kal Ormodv Sam—pa Ac’ ov« fywye). This proves that true knowledge cannot be sought in sensations. Though the true nature of knowledge is not stated in clear words as the result of the inquiry, we can easily gather from certain allusions that knowledge was no longer conceived to be a mere intuition of pre-existing ideas, but a product of the mind’s activity. Knowledge is to be found in that state of the soul, in which it con- siders being, or in its judgments (187.4: amruorjun . . . av dxeivp TH dvdpati, 5 Ti trot’ tyes % spuyy, Stay adth xa? auTiy mpayyarevntat wept ta Svta). Here knowledge is brought under the head of 8d£a, not in the meaning of Opinion, but of judgment (187 a: Sofalayv). ToUTO KaNéiTal .. . This position is not contradicted in the investi- gations. Refuta- tion of material- ism. Con- trasted with a subtler sensual- ism which had produced some physical theories accepted by Plato. Relativity of sensa- tions under different condi- tions. Plato’s view of knowledge as a kind of judg- ment. Thought as moving between affirma- tion and negation, according to the law of contra- diction. Opposi- tion of contra- dictory ideas. Judgment is @ new unity differing from its elements. Know- ledge of a whole 376 ORIGIN AND GROWTH OF PLATO'S LOGIC following discussion and may be accepted as Plato’s true conviction. He explains thought as a conversa- tion of the soul with itself (189 E: ro 82 StavosicOar ap’ Srrep &y@ Kadeis :—Tt Kar@v ;—doyov Sy adTy mpos avrny 7 spuy? SueEepyeras wept dv dv cxoTy .. . abty dauTnv gowrdca Kal amoxpwordvn, nal pdacKovoa Kai ov ddcKouca), lead- ing to a choice between affirmation and negation, wherein judgment consists (190 A: éray 82 dpicaca, .. . ro avTo Hoy b7 xal un Sioratn, SoEav ravtny Tibeuev adris). This duality of affirmation and negation begins to attract Plato’s atten- tion more than ever before. The beautiful and the good are not merely associated as in Republic and Phaedrus, but paired with their opposites (186 a): so also the four pairs of categories in the same passage, and other notions (186 B: oxAnporntra Kai padraxdrnta, 180 D: gordvar .. . xweicOar, &c.). Thus he quotes as one of the objects of judgment the essence of the opposition of beings among each other (186 B: rv odotay Ths évavTioTyTOS avTH } yuy7) xplvew qeparar), and he insists on the impossibility of identity between opposite notions (190 B: dvapipryoKou ei wwrot’ slires mpos ceavtoyv St. TavTos paddov ... Td &tepov Erepov dort). The nature of judgment is further analysed and found to be essentially different from the notions of which it consists. While according to the earlier theory the sight or intuition of ideas was knowledge, it appears now from the example of letters and syllables that the judgment is not the sum of its compo- nents, but a new unity (203 E: ypiv yap icws thy ovrraBnv ridecOar un Ta orovyeia, add’ bE adxelvov Fv re yeyovos eldos, Seay wiav aitd abrod tyov, Srepoy Trav oro- xstwv, cf. 204 a). This conception is repeated with insistence several times (203 z, 204 a, 205 c, 205 pb) in order to refute the supposition that the elements can be less knowable than the whole. He who pretends to know a whole without being able to account for its parts is declared not to speak seriously (206 B: ddv tis O79 avA-~ REFORM OF PLATO'S LOGIC: THEAETETUS 377 AaBny pev yrwotov, ay:warov 2 TEebveevar oToLyEiov, éovTaA ) axovta trailew Hynoouz6’ avrov, cf. Crat. 426 a). This postulate, to base the knowledge of everything upon the knowledge of its ultimate elements, agrees with what has been said in the Phaedrus on the same subject (270 D), and corresponds to a stage in which the chief interest attaches to those notions which are built upon the obser- vation of actual appearances. The question of analysing everything into its elements or kinds was superfluous in dealing with absolute ideas which were supposed to be simple in their perfection. It corresponds also to the new classificatory tendency that Adyos is distinguished into its three kinds: speech (206 D), enumeration of parts (207 a), and definition (208). The three degrees are declared insufficient to guarantee knowledge, but it may be taken for granted that each of them is held indispensable for knowledge. Nobody knows who cannot explain in words the object of his knowledge, enumerate its parts, and give a definition of each of its elements. This last point is stated here with greater fulness than anywhere before. Definition should consist in the indication of the specific difference which distin- guishes a given object from all others (208 c: ro éyew te onusioy sitrety w Tov amavrav Siadépes tro épwrndév, .. .« cf. 175 c). We are warned to avoid circular definitions, which pretend to explain a notion by its synonym (147 B, 210 a), and the enumeration of examples is also declared to be an insufficient substitute for a definition. When Theaetetus began by an enumeration of different kinds of science instead of giving a definition of science, Socrates detained him and appeared to imply at this stage of the dialogue that knowledge is based on definitions (146 E: 70 8 drepwrn Gav ov Todro Hy, Tivwy H émioTnun, ov! orrocat Tiwés * ov yap apiOpnoat avtas BovrAdpevor jpopefa, adda yua@vat driornpny avto 6 ti ror éoriv, cl. Huthyph. 5d, 6£; Meno 72 a). Some models of definitions are given, as for instance ‘ clay is moistened earth’ (147 c), or ‘the presup- poses the knowledge of its elements, while absolute ideas are simple in their perfection. Three kinds of Adyos indis- pensable for know- ledge. Defini- tion by indica- tion of a specific difference, equivalent to know- ledge at the beginning of the dialogue. Defini- tions common to know- ledge and opinion. Consis- tency condition of know- ledge. Heraclitus refuted while the investi- gation of Parme- nides is ad- journed. Dramatic opposi- tion of two yiews on Being. 378 ORIGIN AND GROWTH OF PLATO’S LOGIC sun is the brightest of the heavenly bodies which revolve about the earth’ (208 p). Though at the end of the dia- logue the definition is supposed not to be a peculiarity of knowledge alone, there is no doubt that it has been admitted as an essential condition of knowledge, common to knowledge and true opinion (209 D: sept thy Siadopornra dpa xal % op0y Soka av ein sxaorou wéu). It is very surprising that among the possible meanings of Adyos enumerated, precisely that meaning which this word appears to have in connection with knowledge for Plato (=airia) is omitted, except in one passage in the familiar phrase dvivai te nal SeEacGae Adyou (202 c) in which dAdyos is identical with sufficient reason, as in similar passages of the Cratylus (426 a), Phaedo (76 B, 95 a), and Republic (5381 £). Consistency is here, as already in earlier works, expressly stated to be a necessary condition of knowledge (154 EB: BovrAnoopsta OsdoacOat avta mpos abtd, Ti tor dorly & Stavoovpeba, worepoy nuiv adAndrots Evpdeve: 7) odd’ omrwatioov.—200 D: rf dv avto padtora elrrovres HKoT’ dv Hpiv avtois avavtimOeiuev ;), and the fixity of notions is represented as a condition of consistency (183 A) against the Heraclitean theory of eternal change of everything. This theory had been declared in the Cratylus to be too difficult for refutation, and only here it is refuted, while the criticism of the opposite view of Parmenides is left for a future occasion under a similar pretext to that which in the Cratylus accounted for the postponement of the criticism of the Heraclitean doctrine, namely that the philosophy of Parmenides is too deep for a superficial digression, while it would lead away from the chief pur- pose of the present conversation, the definition of knowledge (184 4). Wesee here the same dramatic opposition of two conflicting views as to the whole of universal existence, which was represented later with such pathetic solemnity in the Sophist. Only here the conflicting views are not materialism and idealism as in the Sophist, but Hera- cliteanism and Eleaticism (180 DB). REFORM OF PLATO'S LOGIC: THEAETETUS 379 This comprehensive survey of the great conflicts in human thought could have been reached by Plato only after a full elaboration of his own philosophy. Thus speaks the head of a school, who has pupils from all parts of the Hellenic world, and observes in them the natural tendencies towards different aims. What has been said in the Republic about the necessary training of a philosopher is here repeatedly mentioned with reference to Theaetetus, who has been prepared for the present inquiry by mathematical, musical, and astronomical studies (145 Aa, c), and also, according to the recommendation given in the Republic, by investigation into stereometry (148 B). His mind cor- responds in every point to what has been required from a philosopher in the Republic: he learns everything as easily as oil spreads silently over a smooth surface (1448), and besides this intellectual development he is courageous and gentle (144.4). This picture of the natural gifts of a future philosopher agrees perfectly with that given in the Republic, as also Plato’s confidence in youth expressed through the person of Theodorus (146 B: 7@ yap dvTe 7 veotns sis wav éridoow #yer). Thusin one important point the psychological rule of earlier logic 1s maintained: the highest level of knowledge can be reached only by excep- tional natures, which have the privilege of being born rulers and teachers of men. For the ideal of the philosopher rises above the rest of mankind, and finds its own model in the ideal of divinity, to which the philosopher approaches as near as possible (1764: 6:0 xat wetpacOar ypy evOevds éxeioe pevyew 6 TE TayLoTA. puyn 52 cpoiwors Gee Kata TO Suvutoyv’ opolwors 52 Sixavoy Kai Soroy peta povyicews yevéoOa). The philosopher is represented as indifferent to the political affairs of his country (173 D), and no stress is laid on his duty to go down into the struggles of vulgar life, and to apply his higher knowledge to the necessities of his countrymen. | The philosopher is here conceived in that stage of Historical stand- point reached. Training of the Philo- sopher illus- trated by the Theae- tetus, 80 as to con- firm the precepts laid down in the Republic. Philo- sopher near the divinity, far from the actual political life, dedi- cated to abstract specu- lation. Enlarge- ment of Plato’s mental horizon. Human measures of time and space insigni- ficant. Antici- pation of modern views. Antiquity of Man. Myriads of genera- tions meant more seriously than twenty- five ancestors. Reason slowly 380 ORIGIN AND GROWTH OF PLATO'S LOGIC abstract speculation which was limited in the Republic to a few years of his life. His mind expatiates over the whole heaven, and all manifold objects forming different wholes, without caring any longer for what is near at hand (173E: 9 dtavowa radta tayta Hynoapevn ouiKpa Kal ws ovdey atysdcaca Travtayyn dépetar Karta Ilivdapov, ta te yas urrévepOe wai ta etriveda yewpustpoica, ovpavod te Umep daoTpovoyotca, Kal wacav mdvtn vow Siepevvwpévn Trev dvtwy éxaoTov Grov, eis TOV syyvs ovdey avTny cuyxaieioca). Accustomed to look upon the whole earth, he despises the greatest landowner as insignificant (174 £), and he equally thinks little of human measures of time, because he knows that even this poor earth (1764: Tovdse tov Torov) has already a past of innumerable millions of years (175.4: wdamwv kai tpoyovwy pupiddes éxdoro yeyovaciww avapiOuntot, év als mrovavo Kal mrwyol Kat Bacwras wal S0dr0. BapBapot re nal “EXAnves moddanes pupiot yeyovaciy otmodv). We see here an horizon of thought extending beyond even that of the Phaedrus. With his wonderful intuition, Plato credits the earth with an age which modern geology for the first time made probable, and leaves far behind him those primitive chronologies which counted only thousands of years since the appearance of the first man. It is strange that acute critics, who took quite seriously the number of twenty-five ancestors quoted here as an example of ouxpordoyia, and counted with the greatest care the ancestors of various con- temporaries of Plato in order to ascertain whom he might have meant, did not perceive that ‘innumerable myriads of generations’ evidently was not a rhetorical exaggera- tion, but a quite serious view of Plato about the antiquity of mankind, in agreement with the cycle of ten thousand years alluded to in the Republic and the myth of the Phaedrus, but entirely absent from the Phaedo and all earlier dialogues. The theoretical tendency is increasing here, and the dif- ferences between men still more clearly recognised than in REFORM OF PLATO'S LOGIC: THEAETETUS 381 the Republic. Few reach a full development of reason : true knowledge can be acquired only by long endeavours under the best guidance, while man and beast alike have sense perceptions from their birth upwards (186 c). The impartial pursuit of truth is here contrasted with eristic discussion, and this exhortation is curiously enough put into the mouth of Protagoras, against whom Plato fought earlier not quite impartially in the dialogue bearing his name. Here Protagoras recommends justice in every discussion, and explains for us some of Plato’s own contradictions, avowing frankly that in polemical writings every one seeks the appearance of being right, while convicting his opponent of as many errors as possible (167 E: aducety 8 éotiv év rd ToLvovTe@, Stay Tis pH ywpis pev @: aywvfopevos tas SiatpiBas Trovjrar, ywpis 68 Svareyo- pevos, Kal dv piv t@ taitn te nal opddrAdr(gy Kal Goov av Suvntas, év 58 rH SvaréyerGar arovddky). If we lead a discussion with the object of arriving at the truth and deal fairly with our opponent, then he accuses only him- self and hates his errors, whereby he is led to philosophy, with a complete change of his former nature (168 4). That such a purely Platonic precept should be given as an exhortation of Protagoras to the Platonic Socrates, appears to be an expiation of earlier polemics and an announcement of that purely objective historical stand- point which we see in the dialectical dialogues. Also Rhetoric as an art of persuasion 1s here mentioned with irony but without the bitterness of the Gorgzas, and more in the indulgent mood of the Phaedrus. Plato recognises the power of Rhetoric to persuade withont knowledge, and sees herein an argument for the great distance separating right opinion from knowledge (201 4: ov dcda- oxovtes, aNAd So€dlev rotovvtes & Av BovAwvrat). This importance attached to a distinction between right opinion and knowledge might be better appreciated if we could guess with some certainty against whom the polemic is directed. Knowledge is emphatically affirmed developed and in few persons. Increas- ng serious- ness of philo- sophical purpose. be Prota- goras in- troduced as exhort- ing to impartial dialectical discus- sion. An implied confession of earlier partiality. Recog- nition of rhetoric, aS giving beliefs without know- ledge. Rhetoric still dis- tinguished from phi- losophy. Changes in the logical point of view not made explicitly. Karlier state- ments not revoked. But cate- gories take the place of ideas. 382 ORIGIN AND GROWTH OF PLATO’S LOGIC to be one of the highest aims in life (148cC: émiornun... Tav axporatwy), worthy to be explained (148D: mpo- OupnOnre wavti Tporm tov te Ad\wv qépt Kal aricTHuns AaBely AGryov TL Tote Tuyyaves Sv), and giving authority to those who possess it (170 A: 8p rye rots peyiocrots Kivdvvots . . . Borep mpos Geos Eye .. . cwrhpas spay tpocdoxwvtas, ouK GdAAW Tw Siadépovtas 7 THO eidévar. Cf. 1710, 1838 ¢). Althougk the ultimate distinction between knowledge and right opinion is not given, it results at least that there is an essential difference between them, and this consists in the systematic unity of knowledge founded on one highest principle, as has been postulated in Phaedo and Republic. It is exceedingly significant that no use of the theory of ideas as known from those dialogues has been made in the whole inquiry, and that the transition from self-existing ideas to categories of reason is made without a formal revocation of earlier views. But it must be recognised that these views are not entirely contradictory, and that ideas of moral notions might continue to exist along with the categories of percep- tions. Only in some special cases the conflict becomes evident, as for instance if we compare some passages of the Phaedo and Theaetetus referring to a problem which was one of the starting points of the theory of ideas and which again returns here as requiring a new explanation : Phaedo 100 &: ovd€ av dp’ av Theaet. 1540: optxpdv aBe dmodéxouo, ef ris twa dain érepovy mapddetypa, Kal mdvra ceive a érépov th xeadrg peitw elvar, cat Bovdopat. dorpayddouvs ydp mov tov é\drra TH ai’t@ Tour éAdrru, €f, dv peév rérrapas avrois mpoc- 101 a: dAXd Sapapripoto av dre evéyxns, mAelovs ghapev elvar radv ov pév ovdey GdAo Ayers Gre rd = Terrdpwy Kai nutodiovs, dav be HeiCoy may érepov érépov ovdeviAdp 8adexa, éAdtrovs Kai juices. peifov corw h peyéber . . . pn tis 1554: drra mor’ éori ravra rd cot évavrlos Xéyos dnavrnan, cay tH «= Po para ev nyiv; by rporov... neparj peitovd rwa gis elvat wal pndérore undev dv peifov pnde A\drra, mparov pev rq atrg rd darrov yevérOar pyre yxy pire peifov pettov elvat cai rd Edarrov apiOuq, ews iow ety adrd daurg.... Garrov, drera rq xehadg oyipa devrepov 8€ ye, d unre mpooriboiro oton rov peifo peifw elvat. unre adatpoiro, rovro pyre aigd- REFORM OF PLATO’S B:... 7a Sexa trav oxra dvoiv ‘ ? ‘ ‘ , a ; ad mieiw etvat, kat dud ravrny ry airiav UrepBadrAewv, poBoto dv Néyewv, dAda ¢ py mwdnOe ... évt évos mpoore- ’ ? Oévros rnv mpcobecw airiav eivat rY id é “a bs rou dvo yeverba fh SvacxyicGevros 4 ‘ 3 ? a 7 Thy oxiow ov« evrAaBoio dv déyetv ; . ‘ 4 s Lig ? C: Kal péya av Bogns Gre. . . ovk , ~ éxets GAAnv riva airiay rov 8vo yevéo Gat add’ fH rnv rns budBos pera- axeow ... tas b€ cyiges Tavras ry 4 A A Mw 4 cai mpooOéces Kat ras adAas ras o , »,¢ UU Tovavras Kopifeias éans av xaiper, ? U “ a mapets drokptvarOa ois éavrov LOGIC: THEAETETUS 383 Ul veo Gai rrore pyre Pbivey, det 8é Crov elvat. B: rat rpirov, & py mpdrepoy fy, 3 A a ~ ‘2 a ~ adda votepoy rovra elvac dvev Tov a’ yeverOar Kal yiyverOa: advvaroy - Tavta dpoAdoynpara rpia pa- Xerat avra avrois év ri npetépa Wuxy, orav ra wept trav dorpu- yaAov Néywpev. C: cat G@dAa 87 pupia ent pu- hd ” ~ plots ovtws exe. . . Soxets your ovUK Gretipos TOY ToLOVTwWY eivat ;— € a“ € U , > > A) Umrephuas ws Oavpalw ri mor’ éort ~ a > # € > ~ id Taura kat eviore ws adnOas BAerwv o 9% > 8 ~ coperepars. eis avUTa OKOTOOLYLO. We see here” that in the earlier dialogue the diffi- culty is stated and left ironically to wiser men for solu- tion. In the Theaetetus the statement of the difficulty is no longer particular as in the Phaedo, but is expressly generalised, and shown to be applicable to innumerable instances, out of which one had been selected as ex- ample. Then also the form of the statement 1s much sharper in the later work, where the problem 1s reduced to three axioms (@dcputa), two of which are in contradiction with the third. The axioms are here said to be in the soul, whereby it becomes clear that we are no longer dealing with transcendental ideas, as in the Phaedo, but with sub- jective notions. While in the Phaedo only the fixity of notions is insisted upon, here we see activity as a condi- tion of change, which corresponds to the increasing interest in physical science, and to the constant applica- 217 H. Jackson (‘ Plato’s later theory of ideas: iv.’ Journ. of Philol. vol. xiii. pp. 267-268) infers from this passage of the Theaetetus that ‘ the inter- vention of the idea is wholly unnecessary for a change of relations,’ while in the Phaedo this intervention was held to be necessary. But really in the Phaedo there was no question of change, and only fixity of relations was sought. The notion of change and movement belongs to a later stage, prepared in the Republic, beginning with the Phaedrus, and growing in the Theaetetus and Parmenides. The problem now con- sidered with more apprecia- tion of its logical nature and its relation to other instances. Form of the statement sharper. Import- ance of the soul increased. Specula- tions as to the possi- bility of error do not lead to definitive conclu- sions. No solution possible until knowledge is defined. Theae- tetus not a Socratic dialogue. The in- conclusive ending marks & new departure. 384 ORIGIN AND GROWTH OF PLATO’S LOGIC tion of the opposition between voy and mdoxeur, common to the Theaetetus with the Phaedrus. In connection with this we find in the Theaetetus a general investigation into the possible conditions of error, which does not lead to a definitive conclusion, but contains very subtle distinctions and deserves our closest attention. It appears first that errors are only possible when one perception is taken for another (193 Bc D) under the influence of an imperfection of sense activity (194 B: qept ov iopev Te Kai aicbavoweda, év avtois tovTots oTpéheras kai éXirrerat 4 dd£a ypevdns Kat adnOns yeyvoyeévn) combined with thought (195 cD: nipnxas 8 wevdh Sofav, Sti obre év tais aicOnceciv got. mpos add HAAS oT’ dy Tais Siavoiats, GN’ év 1H avvarvat aicOnoews pos Sidvorav). But then an instance is adduced of errors possible without the partici- pation of the senses (196 A B), and the difficulty is left unsettled. It results that without a definition of know- ledge no definition of error can be given (200 D) and know- ledge remains undefined, though Socrates remembers that in the whole discussion it had been dealt with as already known (196 E: wupidecs yap eipnkapev TO yuyv@cKkopev Kal ov yiyvwoxopev, Kat érictapela Kai ovK émioTdpeba, ads Te cuvidvtes GAANAWY bv @ ETL emioTNuny ayvootpev) because dialectical discussion would be impossible without a notion of knowledge (196 E: tiva tporov SiaddEer tovTwy ameyo- pavos ;—ovdeva wy ye bs eipt). These fundamental problems were not yet appreciated in their whole importance in the earlier works, and their appearance in the Theaetetus brings us back in one respect to the Socratic stage, namely in so far as no definitive conclusion is apparently reached. But the above significant logical contents involve subtle distinc- tions which would be looked for in vain in the Socratic dialogues. The similarity consists only in the circum- stance that here as well as there a new development of thought was beginning. This new development beginning here—with the substitution of categories for ideas, of REFORM OF PLATO'S LOGIC: THEAETETUS 385 the individual soul for the supercelestial space, of analysis and synthesis for poetical vision, of activity and passivity for immutable identity, of critical cautiousness for poe- tical eloquence—is a momentous step in the history of human thought and would have required another thinker than the author of the Republic and Phaedrus, were he not of such an immense intellectual power and had he not lived so long as to initiate a new philosophical movement after the age of fifty. Thus considered, the question of the date of the Theaetetus acquires an exceptional importance, and no consideration of evidence will be wasted, if it helps to decide the question, whether we are right in placing this dialogue after the Republic and Phaedrus. Up to the present time some of the most competent scholars agree with Zeller in believing that the Theaetetus must have been written within the first ten years after the death of Socrates, or about the same time as the Huthydemus. We have seen that this position is contradicted by the style as well as by the logical theories of our dialogue. But in view of the paramount importance of the question and of the great authority of those who are supporting an early date for the Theaetetus we are obliged to consider in detail the arguments in support of this opinion, which has been unanimously sustained by the chief writers on Plato from Tennemann, Schleiermacher, Ast, Socher, Stallbaum, Hermann, Steinhart, Susemihl up to the last editions of Zeller’s Philosophie der Griechen (1889) and of Ueberweg’s Grundriss der Geschichte der Philosophie (1894) besides many special dissertations.“* The most eminent supporter of an early date of the Theaetetus is 48 Among these are conspicuous Natorp's Forschungen zur Geschichie des Eirkenntnissproblems im Alterthum (Berlin 1884) and his paper on the Phaedrus (Philologus, 48°" Band, pp. 428-449, 583-628, Gdttingen 1889), wherein he looks upon the Theaetetus as preparatory to the theory of ideas. In favour of the opposite view we have, besides all those who have written on the style of Plato, also some authors who admitted a late date for the Theae- tetus for other reasons, as for instance Munk (see note 89), Berkuski (Platons CC Conse- quent im- portance of the date of compo- sition, which by Zeller and others is assumed to be very early. Zeller, in common with many previous critics, still up- holds an early date. Allusion to an encamp- ment near Corinth. Assump- tion of the identity of the date of composi- tion with the sup- posed date of the intro- ductory dialogue 886 ORIGIN AND GROWTH OF PLATO’S LOGIC Zeller, and he has not yet been thoroughly refuted. Though polemic enters to no extent into the plan of the present investigation, it seems to be in this special case our duty to consider Zeller’s arguments, and to prove that they are insufficient to establish his claim. 1. The first chronological indication is seen by Zeller in the allusion to an encampment near Corinth (Theaet. 142 a). He refers it to the war which is known in history as the Corinthian war and lasted about seven years 394-387. Even if we admit this reference as possible—instead of accepting the very convincing argu- ments of Ueberweg, Teichmiiller, Bergk, and Rohde, according to which the allusion refers to a battle of 368 B.c. mentioned by Xenophon (Hellen. vii, 1,15) and other historians— Zeller’s inference as to the identity of the date of composition and the presumed date of the con- versation is not cogent. The more striking the campaign the more probable becomes a later allusion to it. All that is really proved is that the date of composition 1s subse- quent to 392; there is no reason to identify both dates, as has frequently been done in the case of the Phaedo and Phaedrus. The association of ideas between Corinthian war and ‘encampment near Corinth’ is more immediate for us than for the first readers of Plato. But we see in the dialogue the mention of an encampment not of a battle. A soldier might have been wounded in some insignificant attack on his encampment, without having taken part in an historical battle. If we take the mere fact of an en- Thedtetos und dessen Stellung in der Reihe seiner Dialoge, Inaugural-disser- tation, Jena 1878), H. Schmidt (Hxegetischer Commentar 2u Platos Theitet, Leipzig 1880), H. Jackson, E. Rohde, W. Christ (‘ Platonische Studien,’ in vol. xvii. of Abhandlungen der philosophisch-philologischen Classe der kiniglch bayerischen Akademie der Wissenschaften), Teichmiiller, Siebeck, Archer Hind (Introduction to the Timaeus, p. 21), M. Jezienicki (Ueber die Abfassungsseit der platonischen Dialoge Theaitet und Sophistes, Lemberg 1887). Zeller did not consider all the above authors and their arguments when he declared repeatedly the discussion as definitively settled (Archiv fiir Geschichte der Philosophie, vol. iv. p. 189, vol. v. p. 289, vol. viii. p. 124, and on many other occasions). REFORM OF PLATO’S LOGIC: THEAETETUS 387 campment in which dysentery is reigning, we have no reason whatever to refer it to 392 rather than to 368 unless some independent testimony is forthcoming about an epidemic of dysentery occurring at one of these dates alone. In both cases a fight near Corinth took place. It has been argued that Theaetetus, who was a boy according to the dialogue at the time of Socrates’ death, could not already be famous seven years later. Here, as in the Phaedrus, we have a prophecy put in the mouth of Socrates realised at the time of writing. If in the Phaedrus the prophecy refers to the Panegyricus written 26 years later, the prophecy about Theaetetus might well have been realised in a length of time almost equal. Zeller believes that the mention must refer to a recent fact. The notion of recent facts is often abused. Anybody might speak to-day of the Russo-Turkish war as recent if compared with the conquest of Constantinople by the Turks. There is no reason to believe that for Plato current events ceased to be recent sooner than for us, at a time when he spoke of twenty-five generations as a ridiculously small period. 2. If historians are right in saying that Iphicrates in this very Corinthian war introduced the peculiar force of light-armed infantry known as zreAraotai, the allusion to them on the part of Socrates (165 D) certainly involves an anachronism. But if the use of peltasts began at that time, there is no reason to think that it ceased twenty years later. It would be more reasonable to argue from a similar mention of peltasts in the Protagoras (350 a) that the Protagoras cannot have been written earlier than 393; and any one who compares the Protagoras with the Theaetetus will find such differences of style, of method, of literary perfection, and of philosophical theory, that it is impossible to ascribe both to the same period. But the truth is that, whatever may have been the device of Iphicrates, the word weAracrys occurs in several earlier writers, Euripides, Thucydides, Lysias, Xenophon, and cc 2 Compari- son with the Phaedrus prophecy. Notion of a@ recent fact. Mention of the Peltasts common to the Theae- tetus wit Prota- goras and Laws. Twenty- five ancestors. Genealogy less interest- ing for Plato than for some modern historians of philo- sophy. 388 ORIGIN AND GROWTH OF PLATO'S LOGIC is common to the Laws with Theaetetus and Protagoras, so that it has no chronological value whatever in Plato. To infer anything from it means almost as much as to refer any work in which a mention of potatoes occurs to the next time after the first introduction of this vegetable in Europe in 1584 a.p. It seems astonishing that Zeller should have followed Teichmuller in such inferences from an accidental mention of an object familiar to Greek readers before Plato began to write. 3. A third indication of the date of the Theaetetus is seen by Zeller in the allusion (175 a) to those who are proud of twenty-five ancestors, and of their descent from Heracles sonof Amphitryon. This allusion has also been treated as a mark of date by Bergk and Rohde, but each assumes a different descendant of Heracles. And even if we take Plato to be referring to a contemporary, who is to decide whether among the twenty-five ancestors Amphitryon’s father Alcaeus or his grandfather Perseus are to be counted or not? Im any case Heracles need not be the twenty-fifth. The discussion whether Agesi- polis (Zeller), Euagoras (Rohde), Dionysius of Syracuse (Teichmiiller), Agesilaos (Bergk) or anybody else is meant by Plato is a curious example of the abuse of erudition leading to misunderstanding of the text on which the erudition is spent. Plato speaks of twenty-five genera- tions as he does of ten thousand plethra of land, probably without any intentional allusion to any one in particular. The pride of counting Heracles among one’s ancestors, and even a catalogue of twenty-five or more of them, cannot have been uncommon in Plato’s time, if after so many centuries four historians are able to quote four different descendants of Heracles with twenty-five or more ancestors a-piece (175 A: ceuvuvoptyoy Kal avadepoyTwr is a plural that might be taken literally). But it is by no means certain that Plato was as skilled in genealogy as his modern interpreters. He regards the whole question as contemptible, a monstrously small way of reckoning REFORM OF PLATO'S LOGIC: THEAETETUS 389 (cpixporoyla). Those acute critics who perceive in each round number quoted a statistical datum incur the danger of being accused of a cpusxporoyia more blameworthy than that complained of by Plato. 4. A fourth argument of Zeller is more serious than the preceding. He says that the critical character of the Theaetetus does not agree with the positive constructive exposition of the Republic. Zeller means that such elementary inquiry into the foundations of knowledge was most probable in a time when Plato began the build- ing of his philosophy. We quite agree with Zeller, but if we add that Plato in his exceptionally long and active life had time to build more than one philosophy, we are at liberty to place the Theaetetus at the opening of Plato’s second voyage for the discovery of truth, after the Republic. In two passages we notice allusions which may with some probability be referred to Republic (177 EB: wapadevypatwv dy T@ byte sotwrwy, Tod pev Oslov evdatpoveordtov, Tod dé abéov aOA\wwtdtov, ovy Spavres StL ovTws yet, UTS HALOLO- TyTos TE Kal éoyadtns dvoias NavOdvovet ...175 C: one autis Stxavoovvns Te wal adicias . . . Baoidrelas trépe Kal avOpwrrivns ddws evdaipovias nai dOdudtyTos . . . Troiw Té tive éorov Kal Tiva TpoTrov avOpw@rroy hice. mpoonKer TO mev xktnoacbar avtoiv, To 5@ amropuyeiv), and to the Phaedrus (175 E: dppoviay Noywr AaBdvTos GpOds tpvicas Oedv re Kal avOpwrwv svdaiovev Blov)—while Zeller could not find in the whole Republic an equally probable allusion to the Theaetetus. If we compare the critical tendency of the Theaetetus with the critical and elementary character of the works belonging to the Socratic stage, we shall easily notice the difference between those youthful personal criticisms and the fundamental criticisms of the Theae- tetus similar to those of the Parmenides and Sophist. 5. Geller finds an argument for the early date of the The incon- clusive form indicates new de- parture, or second beginning. Allusions to Republic and Phaedrus. Zeller’s Theaetetus in his belief that the Politicus is earlier than view of Symposium and Phaedo. But he has not furnished any proof of this assumption, which contradicts everything we early date of Politicus clearly wrong. Relation to Euclides and Anti- sthenes uncertain. Zeller’s view that a late date for the Theae- telus leaves no room for the dia- logues which follow it. Which are these ? Relation of size between dialogues earlier and later than the Theae- tetus. 390 ORIGIN AND GROWTH OF PLATO’S LOGIC know about the development of Plato’s style and his logical doctrines. 6. The relations between Plato, Antisthenes, and Euclides, which Zeller also invokes in favour of an early date of the Theaetetus, are too little known for any chrono- logical inferences, and they could never prove anything about the date of composition, because Antisthenes is not named in the dialogue, and Euclides appears at the beginning without any mention which would allow infer- ences about his relations to Plato. 7. Zeller enumerates the dialogues which in his opinion followed the Theaetetus, and finds it improbable that they could have been written in the last twenty years of Plato’s life. But he includes the Republic in this enumeration, on the ground that he holds the Republic to be later than the Philebus, and the Philebus than Parmenides and Theaetetus. We quite agree that the Parmenides and Philebus follow the Theaetetus, but we see no sufficient reason for placing the Republic after the Philebus. Zeller relies on some parallel passages which are too general to prove anything, and even rather confirm the priority of the Republic.” Such parallels are rarely decisive, and have only then a certain value, if many concomitant variations point in the same direction. The seven dialogues which, according to our exposition, precede the Theaetetus (Huthydemus, Gorgias, Cratylus, Sym- postum, Phaedo, Republic, Phaedrus) are in their total size (453 pp. ed. Did.) almost equal to the seven dialogues which we suppose to be later than the Theaetetus (Parme- nides, Sophist, Politicus, Philebus, Timaeus, Critias, Laws, 457 pp. ed. Did.). If we are right in supposing that the seven earlier dialogues were written in the years 390-379, there is no difficulty whatever in admitting that the seven later works fall within the last twenty years of ® This question has been recently dealt with by Jackson (‘ Plato’s later theory of ideas VII. The supposed priority of the Philebus to the Republic,’ in the Journal of Philology for 1897, N. 49, pp. 65-82). REFORM OF PLATO'S LOGIC: THEAETETUS 391 Plato’s life (867-347), or even within the time after his third voyage to Sicily (861 B.c.). If this were proved, then the mention of the superiority of oral teaching at the end of the Phaedrus would mean nothing less than an interruption of about twelve years in Plato’s literary labours. But of course such a conclusion requires more serious arguments than those on which Zeller founded his conviction about a very early date of the dialectical works. Here it is only put forth as a possibility which may be made probable by further investigations. At all events, the above reasoning shows that Zeller’s arguments prove only that the Theaetetus is later than 392 B.c., without any determination of the distance be- tween this terminus a quo and the date of composition. All the allusions found out by Zeller with such acuteness and erudition, even if we admit the interpretation he gives them, would remain quite as natural twenty-five years after the Corinthian war as immediately afterwards. In such things we have not the right to look at Plato from the point of view of a newspaper editor, who wishes to give to his readers the most recent information. Plato was free to choose from his large stores of experience at any time any example proper for an illustration of his views, without considering whether it occurred long ago or yesterday. Nosuch immediate allusion as the Ssocncopos of the Symposium has been found as yet in the Theaetetus. On the contrary we have several reasons to believe that the Theaetetus is a late dialogue, written by Plato after fifty and possibly after sixty. These reasons have been collected since Munk and Ueberweg by many investi- gators and can easily be supplemented by considerations of style and logical comparisons. We find in the Theaetetus clear allusions to Plato’s school. The person of the younger Socrates, introduced here, is also known from the Metaphysics of Aristotle (1036 b 25), where he is quoted in the manner in which Aristotle quotes oral reminiscences. This led Ueberweg to Probable interrup- tion in Plato’s literary activity. Zeller’s argumen- tation not con- vincing. Prevailing reasons for a late date of the Theae- tetus. Allusions to Plato’s school. The younger Socrates and Aristotle. Allusions to travels. Theo- dorus of Cyrene. Teich- miiller’s argument from the dramatic form. How far defensible. 892 ORIGIN AND GROWTH OF PLATO'S LOGIC the conclusion that this younger Socrates belonged to the Academy at the same time with Aristotle, or after 367 B.c. If we consider that he does not take an active part in the conversation, it becomes probable that Plato introduced him as a witness out of some personal sympathy at the time when he already had known him for some time past. This argument is not decisive, because the younger Socrates may have belonged to the Academy a long time before Aristotle and still have continued in it afterwards. The Academy was not similar to our universities as to the limits of time fixed for the studies, and Plato’s pupils probably remained in touch with him for life. But a more important observation has been made by Ueberweg as to the picture drawn of the philo- sopher, that it can best be explained if we refer it to Plato’s experience in Syracuse, where he may have found many parasites ready for all kinds of slavish services to please the tyrant. It may also be argued that the insist- ence with which Theodorus of Cyrene is asked to take an active part in the discussion is most natural after Plato’s visit to Cyrene. Such allusions to external events are always open to doubts, and are here quoted without attaching to them any special importance. There is another chronological indication of a more serious character, noticed already by Schleiermacher and brought forward afresh with strong conviction by Teichmiiller. This is the statement at the beginning of the dialogue that it has been wntten down in the dramatic form to avoid frequent repetitions of such formulas as cai dyw pny, xai dym elrrov, cuvidn, ovy dpordyet (148 c). Teichmiller infers from this passage that Plato began only with the Theaetetus to write his dialogues in a dramatic form. But the dramatic form is the primitive form for a dialogue, and needs no apology. The narrated form of a philosophical dialogue is a much more com- plicated mode, and was perhaps introduced into Greek literature by Plato. After trying its different variations, REFORM OF PLATO’S LOGIC: THEAETETUS 393 he returns to the dramatic form and apologises for the change. In point of fact the narrated form has been tried by Plato only in a few of his works, and almost in every case with some difference, as the following classification of the form of Plato’s dialogues shows : 1. A continuous speech, including questions and answers. This is the character of the Apology, in which some passages refer to con- versations held by the speaker (20 a), and others introduce an imagined conversation with the accuser (24 DE, 27 Bc, &e.). 2. Dramatic dialogues in which Socrates acts as leader of a con- versation. This is the most numerous class, including Euthyphro, Crito, Laches, Io, Meno, Gorgias, Cratylus, Philebus—and among the doubtful dialogues Alcibiades I. II., Hipparchus, Theages, Hippias maior and mimor. A slight variation appears when the dramatic conversation includes long speeches of Socrates or others : Menexenus, Phaedrus. 3. In a narration in which Socrates gives an account of some earlier conversation, the chief part is a narrated dialogue. This is the form of the Republic, and besides only of Lysis and Charmides (among the spurious dialogues: Erastae). In this form the re- petition of the formulas complained of at the beginning of the Theaetetus is most conspicuous. 4, After a dramatic introduction, in which Socrates appears as one of the persons of the dialogue, he begins to narrate an earlier conversation, and this narration follows up to the end. This form is found only in the Protagoras. 5. Different from the above is a narration interrupted by dramatic portions in which other persons speak with Socrates about his narration, and such a conversation forms the conclusion of the whole. This occurs only in the Huthydemus. 6. After a dramatic introduction another person than Socrates narrates a dialogue in which Socrates played the chief part. This is limited to the Symposium. 7. The above form is improved by dramatic interruptions in which some opinions are expressed by the hearer about the narrated dialogue. This occurs only in the Phaedo. 8. After a dramatic introduction follows a reading of a dramatic dialogue, excused by a censure of the narrated dialogues generally. This is the case of the Theaetetue alone. 9. After a short narration designed to explain the circumstances of a conversation, follows the dialectical conversation without the interruptions complained of in the Theaetetus. This distinguishes the Parmenides from all other narrated dialogues, and makes it possible that this work was written after the Theaetetus, though in Twelve different modes of dialogue in Plato. Pure narration least common of all, and occurs in three works. 394 ORIGIN AND GROWTH OF PLATO'S LOGIC its general form it is a narrated dialogue, and even a narration of & narration, the dialogue being represented as first narrated by Pythodoros, then from him learned by Antiphon, and from Antiphon’s narration repeated by the actual narrator. But formulas peculiar to the narrated form occur only on pp. 126 a-137 o, here being also often omitted, while they are altogether missed on pp. 187-166. Those occurring in the introduction are different from those condemned in the Theaetetws and from the use of other works : én davat, eireiv being chiefly used. 10. Dramatic conversations in which Socrates proposes a subject, which is then dealt with by another philosopher: Sophist, Politious. 11. After a short dramatic conversation in which Socrates proposes a subject, follows a much longer speech by another person. This long speech may be interrupted by some words of recognition from Socrates (Tvmaeus) or not at all interrupted (Critvas). 12. Dramatic dialogue in which Socrates no longer appears even as hearer: Lawa. It results from the above distinctions that what Teichmiller calls the narrated dialogue includes seven kinds (No. 1, 3, 4, 5, 6, 7, 9), which represent different attempts towards a more perfect form. Only the Lysts, Charmides, and Republic take the form of a contmuous narration. The nearest mode to this is a narration with dramatic introduction, as in the Protagoras. From the Protagoras the Huthydemus differs by dramatic interrup- tions and conclusion, the Symposium by the absence of Socrates in the Introduction, the Phaedo in addition to this by its dramatic interruptions. At last, in the second part of the Parmenides narration is abandoned altogether without any explanation, and the whole dialectical dis- cussion follows dramatically. Teichmiiller’s inference, if limited to the supposition that Plato did not return after the Theaetetus to the form criticised in this dialogue, appears very probable, and 7 An attempt at such a classification has already been made by Stein (Sieben Bucher sur Geschichte des Platonismus, Gdttingen 1864), who divided all the works of Plato into five classes, in a somewhat different manner from the above. It is noteworthy that all the spurious dialogues have the form 2 or 8, while the ten other kinds of dialogues used by Plato have not been imitated. REFORM OF PLATO'S LOGIC: THEAETETUS 395 furnishes us with valuable chronological information, giving additional strength to other reasons, according to which the Theaetetus is later than the Phaedo and Republic. It is not contradicted by any well-established fact, that Plato in his later age used the dramatic form exclusively. Allthe dialogues known to be the latest are dramatic, and the narrated form of the Republic compared with the dramatic form of the Timaeus, its professed con- tinuation, confirms again the supposition that Plato re- linquished the narrated form in order to adopt the dramatic. But it does not follow that he should never have used the dramatic form before he started with narra- tions, nor even in intervals between narrated dialogues. The small dialogues, as to which there is great probability that they were written early, are dramatic, and it is most natural for anybody who writes philosophical dialogues to begin with this form. Esthetical reasons, and the desire to give a greater poetical plasticity or historical probability to an imagined conversation, led later to the more difficult form of narration, which, after different variations, had to be finally abandoned in the Theaetetws and Parmenides. The inconvenience of narration could nowhere be felt more clearly than in the composition of the Republic, and thus one of the most probable inferences from the explana- tion given in the Theaetetus is the priority of the Republic. This is further confirmed by a parallel passage in the Re- public, where the dramatic form is condemned, after a long explanation of the difference between narration and dramatic representation (Rep. 392 p-396 c) on the ground that the dra- matic form is less immediate and sincere than the narrative (896 o: 6 pév pos Sone? pérptos avip, sreddav adlenrar dy tH Senynoe dort AdEw rid 4} wpakw dvbpos ayabod, 2ed- nosy ws autos dv éxeivos amayyéNnrew Kal ov« ais yuveicbar dh rH TowavTy pinoe .. . B: Sinynoes xphoeras olg resis OAfyov aporepoy SundOopev . .. Kai sora adrod 4 rékis wetéyovca piv dudorépwv, puunoeds te Kal Ths &dAns Sinyjoews, opixpov 54 Te pspos dy TOAND AOYH THY sunoews). It is true, however, that all the latest dialogues are dra- matic in form. Theae- tetus after Republic. Possible motive for the preface to Theae- tetus. Notions ‘ familiar in Theae- tetus, but carefully explained in the Republic. Relation of both dialogues. 396 ORIGIN AND GROWTH OF PLATO'S LOGIC This recommendation of narrations is given in a narrated dialogue, and we know that Plato wrote afterwards dramatic dialogues, as, for instance, his Laws, Timaeus, Critias. If now we meet in the Theaetetus an apology for avoiding the form of a narrative when it might be expected, it appears very natural that this apology is later than the condemnation of the dramatic form enunciated in the Republic. This conclusion is the more justifiable, as Plato warns us that his condemnation of the dramatic form is not limited to tragedy and comedy (394D). The above is only an indication, but seems to be more significant than the inferences drawn from the genealogy of various descendants from Heracles. The formulas objected to in the Theaetetus can occur only in a dialogue narrated by Socrates, and therefore the whole objection, if taken literally, refers solely to Lysis, Charmides, Protagoras, Euthydemus, and Republic. Besides the formulas expressly named other answers are used, and it would be an interesting investiga- tion to find out in which of these five dialogues the expressions rejected in the Theaetetws are most frequent. There can be scarcely any doubt that the greatest number of them is to be found in the Republic. The priority of the Republic to the Theaetetus 1s confirmed also by other allusions and comparisons already mentioned which may be here briefly recapitulated : 1. 8vvauss is first explained in Rep. 477 c as a new notion. It is used currently as familiar in the Theaetetws: 158 £, 185 c, &e. 2. The eternal models of the happiest and unhappiest life (176 2) as well as the mention that the philosopher investigates the nature of justice (175 0) are best explained if the reader is supposed to be familiar with the Republic. 8. The short and matter-of-fact enumeration of mathematics, music, astronomy, geometry, and stereometry (1454 c, 148 8B), as preparatory to philosophical problems, seems also to be a remin- iscence of the Republic. 4, The poets are placed on the same footing with Protagoras in their error of denying permanent substance (1528). This is best explainable after the Republic, as in the Sympostwm and even in the Phaedo (95 a) Homer was praised without irony. REFORM OF PLATO’S LOGIC: THEAETETUS 397 5. The notion of movement as distinguished into change of quality and change of place, common to the Theaetetus with Parmenides and Laws, could not easily be ignored in Republic and Phaedrus if already familiar to Plato. 6. The idea of innumerable periods of ten thousand generations (Theaet. 175 a, cf. Legg. 676 Bc, 677 D: pupidxes pupia rn) implies an advance beyond the Republic and Phaedrus, where large periods of generations first appeared, and were specially justified. The long duration of life on the earth is here assumed as known to every educated man, and this was first explained in the Republic. 7. The logical standpoint goes very much beyond the theory of ideas as known from the Republic and Phaedrus. This results from our whole exposition. Some of the above points apply equally to the priority of the Phaedrus, and there is besides one special point of comparison which places the Phaedrus before the Theae- tetus, namely the calm recognition of rhetoric (201 4), which seems to imply what has been said on this subject in the Phaedrus. But the strongest reason why the Theaetetus must be looked upon as later than the Phaedrus lies in the affinities of both dialogues to different groups of other dialogues. The Theaetetus is in style and contents nearest to the Sophist and Poltticus, which are proved to be very late. The Phaedrus shows in style and contents the greatest affinity with the Republic, which is proved to be earlier than the Sophist. The poetical imagination displayed in the Phaedrus and Republic is radically different from the dialectical imagin- ation of the Theaetetus and Sophist. The retirement of the philosopher from the world, which we see in the Theaetetus, remains throughout all later dialogues, and also the complaint that life on earth 1s too imperfect for the realisation of a philosopher’s dreams. This complaint, quite opposed to the optimism of the Republic and Phaedrus, betrays an interval not only of time but also of bitter experience between the poetical and the dialectical group. We know in Plato’s life, after the foundation of the Academy, only one great disenchantment which could Priority of Phaedrus to Theae- tetus. Affinity of Phaedrus to the Republic, and of Theae- tetus to the Sophist. The Theae- tetus probably subse- quent to the second voyage to Sicily. This agrees with other argu- ments. Probable interval between Phaedrus and Theae- telus. This would explain peculiar style of the Theae- tetus. 398 ORIGIN AND GROWTH OF PLATO'S LOGIC justify that change of attitude on the part of the great thinker. This was his second voyage to Sicily in 367 B.c. which he undertook in the hope of realising his ideal schemes, and which ended unsuccessfully. It appears most probable that the new departure, beginning with the Theaetetus, coincides with his return from this voyage. This cannot be proved, but may be suggested as a plausible hypothesis, well adapted to explain many things otherwise unexplained. Those who believe that the battle near Corinth, mentioned at the beginning of the dialogue, must have been quite recent when Plato wrote the Theaetetus are then at liberty to accept Ueber- weg’s supposition that a battle in 368 B.c. is meant here, and they can seek additional evidence in inscriptions and literary monuments in order to prove that dysentery was reigning then in the encampment. The lovers of genea- logies will have a greater choice to select from, and may find in some contemporary encomium, as Diimmler expects, a clear statement about twenty-five ancestors descending from Heracles, thus removing the improbable supposition that Plato himself counted somebody’s ancestors. These are trifling advantages, compared with other con- siderations. If, as we suppose, the Phaedrus was written about 379 B.c., and the Theaetetus after 367, then the passage at the end of the Phaedrus, in which oral teaching is extolled over writing, would obtain a new and original interpretation: it was a farewell to literary activity for about twelve years. And also one strange peculiarity of the style of the Theaetetus is psychologi- cally explained. The Theaetetus, having according to our calculations a slightly later style than the Phaedrus, is distinguished by the entire absence of very important or very frequent stylistic peculiarities. This is natural if that dialogue is written after a long interruption of literary-activity. Plato was then to a certain extent free from acquired habits, and he did not at once fall into new idioms which might become very familiar in later works. REFORM OF PLATO’S LOGIC: THEAETETUS 399 He used freely the richness of his old vocabulary and style, recurring less than usual to new formations and new idioms. Out of 500 peculiarities observed only four accidental words or locutions (11: puewmrrds, 208: évredOev Hon, 399 : wept dH with genitive, 467: yuuvacia) are new, being missed in earlier works. All other peculiarities of later style occurring in the Theaetetus (58 accidental, 41 repeated, 31 important) have been also found in dialogues which we have placed earlier. While the number of accidental, repeated, and important peculiari- ties is much greater than in the Phaedrus (130 against 112) there is not one very important peculiarity in the Theaetetus though seven are found in the Phaedrus. But none of these seven is missed in the Theaetetus, only their frequency is smaller, so that they are counted only as important or repeated in the Theaetetus, while they are more important in the Phaedrus (23, 231, 376, 377, 390, 412, 451). The difference between both dialogues is just what might be expected if we place the Phaedrus at the end of a period of extraordinarily intense literary activity, and the Theaetetus at the beginning of another period, after a long interruption. Nor is the time of twenty years from 367-347 B.c. too short for the composition of the Theae- tetus and the seven dialogues which are left, as their total size is inferior to the total size of the nine dialogues preceding the Theaetetus (Protagoras—Phaedrus) written according to our view between 393-379 B.c. or in about fourteen years. Whether a writer like Plato writes more at forty than after sixty is a question that cannot be decided on general grounds, and we make a due allow- ance for the diminution of activity in old age, down to an average of only four lines (ed. Didot) every day if the last eight dialogues (Theaetetus—Laws) were written in about nineteen years. What is here proposed as a plausible hypothesis is susceptible of proof by further investigation of style. At Absence of very important pecu- liarities. Amount of text written after the Theae- tetus inferior to the pre- ceding nine dia- logues. The interval between Phaedrus and Theae- tetus might be confirmed by further research. The Theae- tetus is certainly later than the Republic, Phaedrus, and Sym- posium. Authen- ticity doubted but with- out cause. 400 ORIGIN AND GROWTH OF PLATO'S LOGIC present the stylistic difference between Phaedrus and Theaetetus is only just sufficient to confirm the later date of the second. But if we remember that thirty years ago the style of the Theaetetus so far as it could then be ascertained appeared as early as that of the Protagoras, and that Campbell resisted the temptation to trust that appearance and judged the Theaetetus to be later than the Phaedrus, which has been fully confirmed by later research—then we are entitled to hope that also our present supposition, that the Theaetetus is about twelve years later than the Phaedrus, may be confirmed by further research. It may also be contradicted, but one thing results as certain from the whole above investiga- tion: the Theaetetus is certainly later than the Republic, Phaedrus, Phaedo, and Symposium.“ This relation will be still better confirmed if we study the next dialogue, the Parmenides, which in many respects shows a greater affinity with the Theaetetus than its acknowledged con- tinuation, the Sophist. II. The Parmenides. (Relative affinity with the latest group, measured on the Laws as unity, = 0°34 ; see above, p. 177.) Among the greater works of Plato none has raised so many suspicions as to its authenticity as the Parmenides, since Socher (1820) had the courage to confess that he felt unable to share the traditional admiration for the antinomies forming its second part. Many doubts ex- pressed by Ueberweg and Schaarschmidt have been removed by the subsequent studies on Plato’s style. This dialogue presents such numerous Platonic peculiarities, déspite its abstract contents, as never occur in spurious ! The relation between Theaet. and Symp. can also be judged from @ comparison of what in hoth dialogues is said about intellectual pregnancy, which is first introduced in the Symposium (206 8), and here supplemented by the notion of intellectual midwifery (Theaet. 148 z-149 3). REFORM OF PLATO’S LOGIC: PARMENIDES 401 works. Whatever may be thought of the philosophical value of antinomies, we find them here presented with great skill, and the conclusions are not more puzzling than those found in a similar treatment of philosophical problems by modern thinkers. The great originality of form and contents can raise suspicion only in critics who are unaware of Plato’s originality in other works. The Parmenides 1s not like other dialogues, but the Phaedrus and the Timaeus also differ widely from the Phaedo and Symposium. It has been thought that Plato could not have invented such objections to his own theory as those with which he credits Parmenides in this dialogue. Thus Teichmiiller and Siebeck *? have been led to the supposi- tion that Plato wrote the Parmenides against Aristotle, and that the second part is intended to refute the objec- tions raised in the first part against the theory of ideas. Even if we admit that the Aristoteles of the dialogue is introduced here with reference to the philosopher Aristotle, there are serious difficulties in the way of crediting him with the objections expressed by Parmenides. Aristotle came to the Academy in 367 B.c. at the age of seventeen, and in view of the extent of the six dialogues which are later the Parmenides cannot have been written long after this. We have seen in the Theaetetus how Plato pro- ceeds when he seriously wishes to refute an objection, and according to this standard we cannot accept the second part of the Parmenides as a refutation of objec- tions raised in the first part. It leads, like the Theaetetus, beyond the primitive theory of ideas to a system of categories, among which unity and variety are discussed by a peculiar method, and shown to supplement each other. Every exclusive hypothesis leading to contradictions, it follows that neither the one alone nor the many %2 ¢Plato als Kritiker aristotelischer Ansichten,’ in Zeitschrift fur Philosophie und philosophische Kritik, vol. 107, pp. 1-28, Leipzig 1895. DD Grounds urged by Ueberweg and Schaar- schmidt removed by sub- sequent research. The Parme- nides not written against Aris- totle. One and many. The terms are used with absolute generality. Plato himeelf discovered the ob- jections : whether to his own theory, or that of some follower. The Tplros &vOpawros argument. 402 ORIGIN AND GROWTH OF PLATO'S LOGIC explain existence altogether, and that therefore we have to seek everywhere the one and the many, as is done in the subsequent dialectical dialogues. It has been asked whether the one means the Platonic idea, or God, or anything else. This question is out of place here. The whole discussion is kept in the most general terms, and may apply to many particular cases. We notice the same tendency as in the Theaetetus to substitute abstract notions for the primitive conception of the ideas, and we need not deprive Plato of the merit of having discovered his objections for himself, the more so as these objections do not necessarily refer to his own earlier views, but to certain special determinations of these views, which may be ascribed to his pupils. In no earlier dialogue had the different conceptions of the relation between ideas and the particulars been stated with such clearness. It remains uncertain whether these different conceptions are Plato’s own, because his theory of ideas so far as it was expressed in earlier dialogues admitted different interpretations. It might be supposed that these interpretations had been attempted by some of his pupils and that he wrote the Parmenides with the purpose of showing the difficulty of such very concrete and special interpretations. The chief point which had been always insisted upon with sufficient clearness, the essential difference between idea and par- ticulars, remains untouched by all objections, and for the first time we find it clearly stated that an idea may vary according to the conception of the conceiving mind. The chief objection, known as the ‘third man,’ consists in the representation of an infinite number of identical ideas (182 A: avro Té wéya Kal TANG TA pEydra, day doavtws TH Yruxy él mavra ins... & Te ad wou peya gaveiratl, @ Tatta mavta avdyen peydda dalvecOa. Gro dpa eldos peyéOovs dvadavyceta, rap’ ai’téd te TO pbyeOos yeryovds Kal ta patéyovta avrod * xal dri rovrots ad maou frepov, @ Taira wdvta peyada bora kal obkérs Sy by REFORM OF PLATO'S LOGIC: PARMENIDES 403 éxacrov cot trav sd@v Borat, adXrN arreipa TO TAHOOs). This objection is by no means peculiar to the Parmenides. It had occurred in the Theaetetus (200 B: 7 mddw ad poe épetre OTe Tay éricTnua@v Kal averriaTnuoouveV sialy av ériucthpar ... Kat ottw &) dvayxacOnoecbe cis Travrov mepirpéyery wupiaces ovdav amAgov ‘rovovyres) applied to knowledge, and in the Republic (597 Bc) to the idea of a chair. There Plato indicated the logical necessity of stopping in this infinite progress. A certain analogy to this is found also in the Timaeus (31 a) where the question is raised, whether besides our world there is not an infinity of worlds containing it, and this is denied. This argument has been attributed to Polyxenos whom Plato met in Syracuse, and is here for the first time answered by the supposition that each idea might be a thought and exist only in our soul (182 B: yun tev edav Eeactoy 7 TOv- TOY vonpa, Kat ovdapmod avT@ mpoanKkn eyyiyverOar AArAOM 4 év ruyais). This explanation is not contradicted by what follows. Parmenides says that if each idea is thought of as unity (1382 c: eldos soras rovro TO voovmevoy Sp elvat, ast dy To avTo él maotv) the primitive theory of péeE:s could not be maintained (182C: ef radXa ys Tor eidov weréyery ... ovn #yet AOyov). Then Socrates proposes, not as a differ- ent solution, but only as an additional explanation, a view of the ideas as models of natural kinds, to which the particulars are similar (132 D: Td pév eidn tadta worep TwTapadstypata éordvatév tH puicet, ra bs AAXa TOUTOLS soLKsvaL Kat elvat Gnoumpara’ kat 4 pweOskis aitn toils GNAOUS yiy- veoOar Tov sid@v ovK ArAXrAN TLIs 7H EixacOHvat avrois). This view is consistent with the psychological character of ideas as notions, and the further objections refer to edn avta xa?’ avra (133 A), not to general notions. The one and the many, to which the antinomies of the second part refer, are also notions, not ideas existing outside thehuman mind. This is perfectly consistent with what has been said in the Theaetetus about the activity of the soul. It is one of the aspects of later Platonism : the soul as the Dp 2 Not peculiar to the Parme- nides. Ideas as notions. Ideas as models. All centres in the soul. Extension of know- ledge to imperfect things. Relativity and fixity. Dialee- tical exercise after a 404 ORIGIN AND GROWTH OF PLATO'S LOGIC source of movement acquires an increasing importance and considers 1ts own notions as objects of knowledge. In the Parmenides the link is given which makes it possible to use the terminology of ideas for general kinds or notions. One of the objections of Parmenides against the universal application of transcendental ideas is at once admitted by Socrates and gives the explanation of the subsequent discussion. The idea in its former shape had to be perfect, and at that earlier stage Plato cared only for the knowledge of what could attain perfection. Now his desire of knowledge extends to everything existing, and there are things imperfect by their very nature (130 c: OpiE nal wnrdos Kai piTros 4} GdXO G TL aTimdoraToY TE Kal gavAétatov) of which we conceive notions, but not trans- cendental ideas, under the penalty of falling into an abyss of absurdity (130 D: deécas yun rote eis rw’ dBvOov drvapiav éurrecwy SiapOape). Rising from particulars to more general kinds, human notions are susceptible of improvement up to the ideal standard of the divinity. Thus perfect ideas appear to be out of the reach of human reason (135 a: ToAN) avayKn auta elvac tH avOpwrivn dice ayvwoTa). If anybody denies their existence, it is difficult to prove his error: 1t requires an exceptional intelligence to show that each thing has its own substance (1354 B: dvépos mdavu pav evdpvods tod Suvncopdévou pabeiv ws %ots yévos te éxdorou Kat ovola airy Kal? abthy, tre 58 Oavyactorépou Tov supjoovtos Kal adXov Suvncoudvou SiddFar tadta Twravra ixavos Svevxpwvnoduevoy). What Parmenides says, that without fixed ideas neither dialectic nor philosophy is possible, refers to the general kinds of Being as they have been presented in the Theaetetus, and does not necessarily imply their separate existence. He then recommends dialectical exercise as the best way of advancing know- ledge, and proceeds to give a sample of such an exercise, which is here called a laborious pastime (137 B: mpay- paraodn madidy raifev), convenient only in a limited REFORM OF PLATO'S LOGIC: PARMENIDES 405 circle of friends and pupils (137 A: avrot éopev), not before a larger public (1386 D: dmperq yap ta TovadTa modnN@y évavtioy eye . ... ayvoovar yap ot wool STL avev TautTns THs Sta tmavrwv Sdiekodov re Kal mwravns advyaToy éytuydvTa T@ adnOei vodv yew). The method is supplementary to the method which had been proposed in the Phaedo. There it was the philoso- pher’s aim to explain each hypothesis by another up to the highest hypothesis which might be confidently accepted. Here Parmenides wants us to follow out the consequences of each hypothesis affirmed or denied, and its relation to the whole of our knowledge (136 B: évi Ady, sept Grou dv del tir087 ws Syros nal ws ovK Sytos Kal orLovy adXo Tab0s Tacxo- ytos, de oxotreiy ra Evp,Baivovra mpos avTo Kal mpos bv ExacTov TaY GAXrwv, & te av mposdyn, kal mpos mAiw Kal mpos Evp- Tavta MoavTws * Kal Tada ad pos abTa Te Kal pos Addo 5 te dv mpoatpy adel, édv te ws dv wrobq 5 drreribeco, édy te ws pn ov, ei pédrets TEAfws yUuVYacduEVos KUpiws SioyrecOar +6 adnOés). This method implies the recog- nition of a mutual relation and interdependence of all things that exist, and we need not expect in the following large sample of antinomies about the one and the many a full realisation of the proposed problem. The idea of relation occupied Plato’s mind with increas- ing fascination, as 1s shown not only in the antinomies of the Parmenides, but also in the surprising conception according to which our notions are in the first instance related only among themselves, and could be out of relation with more perfect notions or ideas of the Divinity. The example chosen to illustrate this relativity 1s the relation between a slave and his master. This relation is a relation of two men, says Parmenides, and not of the ideas of slavery and mastership (133 £). Although this view is here introduced as an objection to transcendental ideas generally, it agrees very well with the tendency of the dialectical dialogues which follow, in which we shall find frequently a complaint about the relativity of human model. Disjuno- tive in- ference. Mutual relation of all existing things. Remote- ness of the per- fect idea. Platonic and Kantian anti- nomies. Know- ledge more clearly conceived. Univer- sality of the philo- sopher and his high training. 406 ORIGIN AND GROWTH OF PLATO'S LOGIC knowledge. The distinction between a subjective notion and its objective counterpart is nowhere so clearly stated as here; this is not the only feature in which the Parmenides approaches Kant’s Kritik. Also the dis- covery that abstract notions, if applied without restriction, lead to antinomies of reason, is common to Plato and Kant, although they have treated the subject differently. These antinomies are the further consequence of the dualistic tendency already visible in the Theaetetus and increasing in the Laws, where even the unity of soul throughout the universe is denied, since evil cannot be ascribed to God. On the other side we find here a partial answer to the question ‘ what is knowledge?’ which was raised in the Theaetetus and left unanswered. Knowledge is a system of notions from the highest down to the lowest, brought into manifold mutual relations. Only uneducated people look upon logical exercise as idle talk (135 p). Such exercise leads us from the visible world to the ideas which are an object of reason (135 E: yyaoOnv, Ste ovx elas év trois opwpévors ovd8 TEDL TadTA THY TWAaYHY sricKoTEty, ANAG TEpi écsiva & pwddtora tis dv XOYH AdBoat kal eldn av Hynoasto sivat). The true philosopher neglects nothing, however insignificant it may appear, if it has a bearing upon his general theories, and 1s not influenced by the unscientific opinions of the many (130 E: véos yap ef tri, Kai obra gov avreiAnrrat pihocodia, ds kre avtiniyperat Kat’ dunp dofav, Ste ovdav avtav arysacas * vov S2 Fre mrpos avOparTrwv amoPrsrets Sdkas 51d THhv Hrxiav). This attitude is pre- served also in the Sophist and Politicus, m which the dialectical pastime is continued. We have seen in the preceding works the theory that the highest standard of knowledge is attainable only through the highest intellectual training. The training proposed in the Republic was in mathematical, astronomical, and musical studies as preparatory to Dialectic. Dialectic was there only the knowledge of REFORM OF PLATO’S LOGIC: PARMENIDES 407 the highest idea of Good. In the Phaedrus it was defined as the art of analysis and synthesis of concepts, and this programme was probably followed out in many particulars in the oral teaching of Plato. The result was an essential change of the former views about ideas. The occupation with particulars of nature brought the concept of movement into prominence, and movement was in some way brought into the fixed and unalterable world of the ideas as we know them from the Phaedo and Symposium. This movement consisted first in the universal mutual relations among ideas, and then in the progress of each idea, according to the individual perfec- tion of the thinker. Plato’s love of ideal perfection is not on the decrease, and the ideas of the perfect Being or God remain as perfect as they were seen in the space above heaven of the Phaedrus. But they are not out of all relation to a living consciousness, and each of those unities has infinite approximations in the minds of the whole hierarchy of beings, and in the variety of appear- ances. No doubt the philosopher is able to bring his ideas to divine perfection, but only through dialectical exercise. In agreement with the importance acquired by general concepts, we find in the Parmenides some new notions. Besides évvaws (183 E, 135 Cc), Kivnots (188 B), dAAoiwots, dopa (1388 c, 162 DE), py dv (142 a) and other categories used already before, we meet here for the first time rd cupBeRnxes as a logical term (128 c), otépecOar (157 Cc, 1598), ro &aidyns (156 D), which are clear as general notions but scarcely fit for representa- tion as transcendental ideas. If our interpretation of the logical meaning of the Parmenides is right, it becomes exceptionally important to determine the place of this work among Plato’s dia- logues, as it begins together with the Theaetetus a new philosophy of Plato. That the Parmenides is not an early dialogue, results from many hints. Beginning of move- ment among ideas : increased interest in he- coming. Evolution of ideas according to the perfection of the thinker. New terms and notions. Parme- What is here repeatedly said of ‘ides not early: as appears from the way in which youth is regarded. Socrates repre- sented as very young and sub- ordinated to another master. Plato’s conscious- ness of his own superi- ority. The con- ception of ideas as patterns has been antici- pated ; 408 ORIGIN AND GROWTH OF PLATO'S LOGIC youth (130 z, 135 p, 137 B), that young men are inconse- quent, that one must learn while young, and that youth is pleasing and compliant, is only explainable if the author was comparatively speaking an old man when writing. If we consider that the limits of youth were wider with the Greeks than with us, that youth must be already at some distance to be thus treated, and that we find in the Laws and in the other works of Plato’s old age similar remarks on youth, we are justified in admitting that Plato must have passed middle life when he wrote the Parmenides. Another general argument in favour of a late date is the characterisation of Socrates as a young man, receiv- ing instruction from Parmenides. There is nothing dis- paraging for Socrates in this position, as Schaarschmidt thought. He is here clearly admired by Parmenides and Zeno, and his philosophical aptitude is extolled. In all preceding dialogues we have seen Socrates as the ideal teacher, only in the Symposium subordinated to the ideal Diotima, but even there supposed to be the true author of all that he attributes to her. If now we meet for the first time a Socrates who 1s truly subordinated to another Master, and if we know that in all remaining works of Plato, except the Philebus, Socrates is only a hearer, it becomes very natural to suppose that the Parmenides was written at a time when the living picture of Plato’s Master was fading away in a distant past, under the influence of a consciousness of his own superiority. That Socrates appears here as a young man, is a consequence of the plan of the dialogue, in which a theory formerly attributed to Socrates had to be corrected and abandoned. It has been thought that the view of paradeigmatic ideas or eternal models (182D: sapadelypara iv tH pice, cf. Theaet. 176) appears here for the first time, but this cannot be maintained in view of the fact that we had already in the tenth book of the Republic paradeigmatic ideas, and that such are also implied in the allegory of the REFORM OF PLATO'S LOGIC: PARMENIDES 409 Cave. The only view which is really expressed for the first time is the identification of the ideas with notions in the soul. This view, which we shall see recurring in later works, cannot belong to an early time in Plato’s life, at least in connection with a criticism of self-existing ideas. The meeting of Parmenides with Socrates, whether historic or not, is mentioned besides this dialogue also in the Theaetetus and Sophist. If we compare ** both men- tions, it is obvious that the Sophist refers to our dialogue, while in the Theaetetus the mention is more general : Theaet. 188 ©: rovs GAXovs, of év €ords A€youot TO wav . . . Hrrov aioxvvopa f €va dvra Tappevidny . cupmpoaeptta yap 37 r@ avdpi mavu veos mavu mpeaBitn, Kai pot épaym Bdbos te Exew ravrdract yevvaiov. 184.4: goBotpat ody py) otre ra Neydpeva Evvidpev, ri re Stavoowpevos ele roAU mAéov Aet- Soph. 217 c: morepov ectwOas Lf > A > 4 “~ a ’ WOotov avros emi cavrov paKkp@ royy , U ‘ Stefcevas . . . 7) Ov epwrnoewy, otoy mote kat Tlappevidn ypopév@ kai SieEcovre Aoyous mayxdAous trape- yevdunv eyo veos dv, éxeivov pada . oy , , ~ 4 67) Tore dvros mpeoBurov ;—To pev > Ld A > , adures Te Kai evnviws mooodiadeyo~ Z ca ” a a pév@ paov ovTw, Td mpos GAAoy. mopeba... We see that Plato in the Theaetetus mentions in general terms his admiration for Parmenides, and an interview which might be historical without necessarily implying a special reference to the dialogue, while in the Sophist¢ an allusion is made to the short generally affirma- tive answers which characterise both the Parmenides and Sophist, not the Theaetetus. These three dialogues con- tain very frequent mentions of Parmenides, who is besides quoted only in the Symposiwm (178 8B, 195 c) on an in- significant matter and without great esteem. In the Theaetetus the examination of the philosophy of Parme- nidés 1s declined and adjourned ; in the Parmenides the 258 This comparison has been specially insisted upon by P. Natorp in his review of O. Apelt’s Betirdge zur Geschichte der Philosophie, Leipzig 1891, in the Philosophische Monatshefte, vol. xxx. pp. 63-70, but 1n con- nection with a very early date of the Theaetetus. Natorp’s own argumenta- tion gains in strength if the Theaetetus immediately preceded the Par- menides. ideas as notions appear for the first ime. Other allusions to the meeting of Socrates with Parme- nides. Elea- ticism for the first time seriously con- fronted. Possible occasions for this. Categories more differen- tiated. Remote- ness of the imaginary dialogue. Stylistic relation to the 410 ORIGIN AND GROWTH OF PLATO’S LOGIC philosopher is introduced as criticising earlier Platonism and explaining the consequences of his own hypothesis in @ manner which might lead the hearer to some doubts ; in the Sophist he is criticised by the anonymous guest from Elea, introduced as a friend of Parmenides and Zeno. If these three dialogues, in which the influence of the Eleatic philosophy is first noticed, are written after a sojourn of Plato in Sicily, then it might appear probable that on this voyage he came into closer rela- tions with the Eleatics, just as in the period of middle Platonism the influence of Pythagoras’ school is notice- able. So long as we have no more detailed testimonies about these voyages, we must limit our inferences to the observation that Plato at a later stage of his life con- ceived a special interest in the Eleatic philosophy, either in consequence of personal acquaintance with the repre- sentatives of this school abroad, or perhaps under the influence of his own pupils in the Academy, some of whom might have arrived from Italy. An important argument for the priority of the Theae- tetus to the Parmenides is the different manner in which the categories and the subdivision of x«lynois into GAXolwors and gopd appear, being in the earlier dialogue distinctly meant as something new, while in the later both theories are supposed to be known. Both the Theaetetus and Parmenides are distinguished from other dialogues by the introductory information calculated to make on the reader the impression of things of a remote time: in the Theaetetus this is done by the fiction of a written account repeatedly corrected ; in the Parmenides the source appears more distant, as the dia- logue has been first narrated by Pythodorus to Antiphon, and by Antiphon to Kephalos, who narrates it to the reader. Some reason for placing the Theaetetus before the Parmenides 1s given by stylistic comparisons. The total stylistic affinity of the Parmenides with the latest REFORM OF PLATO'S LOGIC: PARMENIDES 411 group (equivalent to 243 accidental peculiarities) exceeds only slightly that of the Theaetetus (equivalent to 233 accidental peculiarities), and this alone would not yet justify a conclusion, were there not a great difference of size between the two dialogues, the Theaetetus being one of the largest (53 pp. ed. Did.), and the Parmenides one of the shorter (31 pp. ed. Did.) dialogues. Under these cir- cumstances the priority of the Theaetetus appears to be very probable, so much more as the Parmenides has a much greater number of peculiarities of later style which are absent from the Theaetetus, than vice versa, as can be seen from the following comparison : Peculiarities of later style not occurring in works earlier than the Republic and found: in Theaet., not in Parm., acci- in Parm., not in Theaet., acci- dental: 218, 387, 348, 395, 404, dental: 486, 487, 488, 189, 216, 836, 190, 835, 341, 324, 11, 208, 224, 331, 485, 470, 492, 483, 490, 899; repeated: 192, 227; im- 478, 328, 476, 25, 28, 225, 822, portant: 247, 12, 452. 458, 459, 461, 462, 464, 466; re- peated : 481, 477, 489, 382, 480, 475, 24, 468, 26, 460, 468, 465; important: 479, 818, 27; very important: 14, 15. This relation of style between Parmenides and Theae- tetus was less evident as long as smaller numbers of peculiarities were compared. Thus, according to Camp- bell’s table, the Parmenides appeared to have less affinity with the latest group than nearly all Socratic dialogues, and C. Ritter was led even to doubt the authenticity, because he found fewer peculiarities of later style than he expected in a work which betrayed by some very characteristic marks its late origin. Now we have just enough stylistic evidence to confirm the place assigned to the Parmenides between Theaetetus and Sophist, and further stylistic investigations may very possibly increase such evidence in this case, as they have done in the case of the Theaetetus. Both Parmenides and Theae- tetus are stylistically more difficult to class than most T heae- tetus shows that the longer dialogue is earlier.. Proba- bility of an interval after Republic and Phaedrus. Supposed allusion to Aris- totle difficult to verify. The younger Socrates. The Phaedrus affords a point of transition towards the new dialectic. 412 ORIGIN AND GROWTH OF PLATO'S LOGIC other works of Plato. The supposition that both followed after a longer or shorter interval of literary inactivity accounts best for this circumstance. An author who returns to literary labours after an interval does not reach at once a certain fixity of expression and is less likely to introduce many new peculiarities of a permanent character. Thus, however original may be his style in such works, they will contain fewer peculiarities recurring later than the following dialogues, and this produces a diminution of the stylistic affinity with the latest group. The close relation between Theaetetus and Parmenides as critical dialogues has been illustrated by Campbell through a number of analogies (‘On the place of the Parmenides,’ pp. 6-7, see note 145) which are the more striking as the subject of both dialogues is not identical. There is no definitive indication which could help to fix the date of the Parmenides with exactness, except the supposed allusion to the philosopher Aristotle contained in the mention as a person of the dialogue of another Aristotle, one of the thirty tyrants. This allusion is plausible, and has been brought into relation with Aristotle’s criticism of the Platonic ideas. It acquires some additional plausibility if compared with the intro- duction of the younger Socrates in the Theaetetus. But these conjectures require some independent testimonies before they can be accepted as certain. If we accepted them, then the Parmenides would have been written after 367 B.c., and shortly after the Theaetetus. Without reject- ing this hypothesis, it remains still possible that both dialogues were composed earlier, but not before the Phaedrus, and not in the next time after the Republic, as the elaboration of the new point of view required a certain length of time. The nearest approach to this new point of view was the recommendation of analysis and synthesis given in the Phaedrus. REFORM OF PLATO’S LOGIC 413 Plato’s critical philosophy. To resume the results of the above inquiry on the Theaetetus and Parmenides, we see in these two works the trace of a new logical departure, which does not quite amount to a brusque negation of earlier views, but changes the aims of science. While Plato in the works of his middle lifetime had a conception of truth eternally fixed, which can be perceived by a well-trained mind exactly as it is, he became later aware of the subjectivity of knowledge, of its existence in an ascending scale of souls up to divine perfection. The ideal was thus further removed from the present life, while losing nothing in its perfection. The aim of science is now not the immediate contemplation of truth eternally pre-existent, but the perfecting of our own ideas so as to form a system built on the mutual corre- lation of all particulars. The particulars of sensible ex- perience are no longer rejected as useless or perturbing, but they have to be brought into relation with the general stock of knowledge. In the physical world movement is acknowledged as the chief factor, and the origin of move- ment attributed to the soul. The causes of error are in- vestigated with greater accuracy and found chiefly in the imperfection of our perceptions. The notions are paired with their opposites, and the preference for dichotomy is manifest, but is not suffered to stiffen into a conventional rule. Plato remains in this period faithful to his custom of fixing in a literary form only certain aspects of his thoughts, obliging us to supplement by inferences what he omits to mention. Neither the Theaetetus nor the Parmenides are systematic accounts of any part of the doctrine which probably was imparted to Plato’s pupils according to the precepts of the Phaedrus. The centre of gravity of the Platonic system has been changed without recapitulating all the details it carried with it, and the dialogues written Beginning of a re- form in dialectic. Remote- ness of the ideal. Syn- thetic en- deavour. Ideas correlated with one another and with particular things. The soul 88 source of move- ment acknow- ledged as chief factor. Preference for dicho- tomy. The change is not explicit. Plato’s dialogues are still works of art. Their protreptic and edu- cational character. The ideal recedes, and becomes more divine, but is ap- proached continu- ally. 414 ORIGIN AND GROWTH OF PLATO’S LOGIC after the change continue to be works of art rather than expositions of doctrine. They are only ideal samples of conversations held in the Academy, and the artistic pur- pose of harmonious proportion is quite as evident in these conversations on abstract subjects as in the more poetical Symposium. In these works, as in the preceding, from the Symposium onwards, we have didactic conversations between pupil and master, not as in earlier works like the Gorgias, discussions between men of opposed convictions. The pupil is led by an ascending way so that at each turning point he believes himself to reach the summit, when a new horizon is opened, leading higher, and at the end the infinite ideal of knowledge remains still high above the highest summits hitherto described. This protreptic character is maintained in the critical dialogues no less than in the constructive works. In the Re- public the idea of the Good remained beyond the reach of Adeimantos and Glaucon ; in the Phaedrus the ideal rhetoric appeared as a powerful ideal beyond the under- standing and ability of the greatest orators of the time; in the Theaetetus knowledge appeared at a height much above all human opinions, even those which guess the truth correctly. In the Parmenides the objects of know- ledge are shown not to correspond to poetic metaphors, and to be attaimable only by a difficult exercise of reason. In all these cases the rising soul of a lover of philosophy is the chief object of literary exposition. The contents of philosophy are mentioned occasion- ally and never exhaustively. The distance between the philosopher and vulgar humanity is increasing while the philosopher’s constant aim is to approach his ideal of the divinity. The occasional glimpses of theory show us a great wealth of intellectual life, and a consciousness of some cardinal conditions of truth. The chief results arrived at by Plato at this stage appear to be: the subjectivity of sensations, the unity of consciousness in the act of judg- REFORM OF PLATO'S LOGIC 415 ment, the plurality and mutual relation of the highest kinds of Being, the universal analogy between great and small things which must be considered all with equal care in order to increase our knowledge. The method pro- posed leads to a general system of science, some aspects of which are developed in the three following dialectical dialogues. 416 ORIGIN AND GROWTH OF PLATO'S LOGIC CHAPTER VIII NEW THEORY OF SCIENCE Asthe | WE have seen Plato begin his literary career with small Socratic critical dialogues, culminating in Protagoras, Meno, and stage was Huthydemus, and progressing from this first critical stage ee to the positive exposition of some of his moral, political, ace 'V° and educational theories in the Gorgias and later works up exposi. $0 the Phaedrus. In like manner the second critical stage, tion,so manifest in the Theaetetus and Parmentdes, was followed thesecond by some dialogues full of positive metaphysical and logical critical theories, skilfully treated with regard to questions of stage purely formal importance. This indirect manner of expo- was fol- sition is prominent in the three dialectical dialogues which anova follow the Parmenides, namely the Sophist, Politicus, logical 90d Philebus. Here, as in the preceding works, we do and meta- not find a systematic exposition of doctrine, but occasional physical glimpses which betray studies very remote from those of teaching. middle Platonism, and show usa part of that ‘longer way ’ alluded to in the Republic as leading to the knowledge of truth. I. The Sophist. The aim is In this dialogue the definition of the Sophist is only a formally, pretext for the exposition of Plato’s views on scientific todefine method, on the origin of error, and on the nature of true me Being. These views are presented in a form which i leaves no doubt as to the author’s own convictions and soe his judgments about other philosophers. The historical Plato's method of comparing existing theories and contradictions viewson is here maintained, as in the Theaetetus and Parmenides, NEW THEORY OF SCIENCE: THE SOPHIST 417 but with greater maturity of treatment. In this respect, as well as in the manner of the didactic proceeding accompanied with frequent quotations of results obtained before, and with recapitulations after each progress of the argument, the Sophist approaches more nearly to the writings of Aristotle than any earlier dialogue of Plato. The dialogical form is still preserved, but the answers for the most part only confirm opinions expressed in the question, so that they could easily be omitted. While in the Parmenides it was still assumed as natural and necessary that a dialectical exposition must be given in the form of a conversation (137 B: ris ovy pol atroxptvetrat; 7 0 vewrartos ;), we see in the Sophist for the first time a clear admission that philosophical teach- ing may be given in the form of a continuous lecture (217 C: worepov eiwOas Hdv0v avtos él cavToD paKp@ AOY SueEvgvae Afywv Toit, 6 Av évbdeiEacOai tw BovrnOFs, H Sv’ dpwricewy, olov wore xai Tlappevidn ypwpévp . . . wapeye- vounv). If we take into account that this form of con- tinuous lecture prevails in the Tamaeus and Critias and some parts of the Laws, which are acknowledged to be late works, it becomes evident that the Sophist is in this respect intermediate between Parmenides and Timaeus. This inference is strengthened by the obser- vation that in an admittedly early work, the Prota- goras, lecturing is condemned and dialogical discussion required (Prot. 384D: éym ruyydvw simidnopwy tis av GvOpwrros, nal dav ris por papa Aéyn, SriravOavopuar rept ov dv 70 AOyos .. . auvTeuvd Lor Tas atroxpicets Kal Bpa- xutépas role, et uéAXw oot éreaOai). Thus we see how Plato advanced from the form of philosophical conversations to that form of a _philo- sophical lecture or dissertation which has been adopted by his pupil Aristotle and by the majority of later philo- sophers. This fact is not without logical importance. In conversation at least two persons are wanted to elaborate the truth. This implies a stage of personal EE scientific method. Use of historical compari- sons. Approach to the manner of Aris- totle. Con- tinuous exposition admitted aS pos- sible. Not as in Prota- goras. Logical signifi- cance of the change. Form of dialogue gradually relin- quished. Conscious- ness of method. 418 ORIGIN AND GROWTH OF PLATO'S LOGIC uncertainty or at least the absence of a recognised authority. The thinker who has arrived at the highest degree of certainty needs only receptive hearers to whom he may communicate his knowledge, and looks upon discussion as useless and tiresome. The earliest works of Plato were discussions; even later, despite the in- creasing authority of Socrates, the persons represented as partners in his conversation still enjoyed the freedom of expressing other views. In the Theaetetus Socrates 1s represented as desiring to discuss freely philosophical difficulties with Theodoros rather than with a young man who dares not go against his authority. It is only in the Parmenides that discussion (oAvipaypovely) 18 declared useless. This is a logical mode of regarding the matter and amounts to this: whoever is in possession of truth can impart it to others without expecting an advance of knowledge from the conflict of opinions. Or, truth is the result of the activity of one soul, not of the co-operation of many. In all the six latest dialogues Plato remained faithful to this principle, which he adopted definitively in the Parmenides. There is no discussion in the Sophist and Politicus, nor in the Philebus and the Laws. Inthe Timaeus and Critias even the dialogical form is extinct. Plato appears to have abandoned con- versational equality between investigating friends, he prefers now a didactic authority of one Master of wisdom. The consciousness of method is also increasing. The art of reasoning, postulated already in the Phaedo (90 B: ) tept rovs Noyous Téyvn), is now a reality and bears the name of a logical method (Soph. 2274: trav royov uéO0d0s), which remained in the highest esteem among all later philosophers. Many translators of Plato refrained from the identification of pwéodos with the modern term method, as if they were afraid to credit an ancient Greek philosopher with a consciousness of regulated proceed- ing which seems to be a privilege of recent science. Thus, for instance, Schleiermacher renders péOodos by NEW THEORY OF SCIENCE: THE SOPHIST 419 ‘das erklérende Verfahren,’ Deuschle by ‘der Gang der Untersuchung,’ Miller by ‘der Fortgang unserer Erérterung.’ This is really a wrong cautiousness, and Jowett and Campbell were perfectly right in translating pe0o60s here by ‘method.’ In earlier dialogues, as Phaedo (79 B, 97 B) and Republic (435 p, 510 B oc, 531 c, 533 Bo, 596 a), this word had not yet a fixed meaning and was equivalent to ‘ argument,’ ‘study,’ or ‘way of reasoning.’ In the Phaedrus pé0odos (269 D, 270 D) is used in the same primitive meaning of ‘ way of reasoning.’ In the Theaetetus (183 c) it means ‘hypothesis’ or ‘theory.’ But in the Sophist there appears for the first time a ‘ logical method,’ essentially different in form and contents from the darexrexy wéOodos of the Republic (533 c), which meant no more than the study of dialectic, or vision of the idea of Good. Here the ‘logical method’ means what up to the present time is known as the method of classification, or scientific method generally. This method neglects nothing however insignificant it may appear to be, and seeks truth quite independently of all practical applications or advantages (227 A: ty Tov Aoyor we00d@ orroyytotinns pappaxotocias ovdiy Hrrov oudé TL MANNY TUYyavEL pEoV, & TO pev TuiKpd, TO Oe weydra nas @perel xaBaipov). Its aim is pure knowledge, which depends upon the distinction of natural affinities and similitudes between different things, without any prejudice in favour of one subject or another (2278). Of this dis- interested impartiality of pure science Plato gives curious examples which show his tendency to free himself from every authority or reigning opinion. The art of human war, he says, belongs to the general kind of hunting, no less surely than the art of vermin-destroying, despite the greater vanity of man-killers (227 B, cf. Theaet. 174 D). The philosopher finds out the true similarities and differences which allow an exact definition of each kind of beings as belonging to a more general class (235c: mdvrws obra otros ore GAXNO yévos ovdSay ph TroTe deuryov EE 2 Meaning of uéGo8os more definite. Disin- terested- ness of science. Similarity and dif- ference impar- tially surveyed, without trusting appear- ances or following arbitrary lines. Definition of primary notions too much neglected. Scientific truth the philo- sopher’s single aim. Generali- sation and division proceed- ing from the simple to the complex. 420 ORIGIN AND GROWTH OF PLATO’S LOGIC drrevEnrat tiv Tav obtw Suvapévev petievar nal’ Exacta Te xal dri mdvra pébodov). The greatest care must be taken about apparent similarities (231 a: tov 8 dodady Sei TavTov paddiota wept Tas opotoTnTas adel qroleicOar Tv duvanjy * dducOnporatoy yap To yévos). The temptation to mix all things and to make the great appear as small and the like as unlike is the sign of a man who is only beginning to approach the problem of being, and delights in contradictions (259 D: ro 8% ravrov érepov drrodaivey dun yé wy nal To Odrepov Tavtov Kal TO péya ouLKpoy Kai TO Swovov dvopuaov, xal yaipaiv otw Tavayria asl mpopépovta éy Tots Noyols, ovTE Tis EXeyyxos ovTos GAnOwos apTt Te TAY dvrwv Twos éparrouévou Shros veoysvyns ov). Many notions as to which apparently there is no disagreement among disputants are insufficiently defined, and ought to be in- vestigated again, however clear and simple they appear at first sight (242 c: ra Soxodyra viv évapyas tyew éemioxé- aca, Tparov, uy Ty TEeTapaypdvor pev wpev TEepl TadTa, padies 8 aGdAnXroLs Oporoyapev ws svxpiwas ¥yovtes). The true logician follows his opponents on their own ground and refutes them according to their own principles (259 C D: yarerov dua Kai Kadov . . . Tos NEyousvots oldy 7 elvat Kal’ Exactoy édéyxovr’ erraxonovGeiv, Bray te ris Etepov ov wy tavroy élvat hy Kal drav ravtov dv Erepov, éxeivn nat kat’ éxeivo 6 hynoe rovTwv merovOsvat qorepov). He seeks the truth first for himself and then for those who are able to partake of such investigations (264 E: érdei- Eowev padiora wey iy avtois, erecta 8&8 Kai Trois &yyutaTw yéver THS ToLavTns weO0Sou TreduKccw). This aim is reached by the subdivision of notions into indivisible ultimate kinds (229 D: xal rodro oxerréov, et dropov Hon dori wav, 4 Tia Byov Sialpeow aklav érwvuplas), and by a training which consists in a consecutive selection of examples, beginning with those which present less difficulty and rising progressively to the most difficult problems (218 0: dca 8 ad trav peydrov Sei Siarroveiobar Kaos, TWEpt THY TowovTwY SéSoxTat Tact Kal wddaL TO Tpd- NEW THEORY OF SCIENCE: THE SOPHIST 421] Tepoy éy ouuskpois Kai pdoow aura Seiv wereTav, mpiv ev avrois tots peylorous). This notion of logical exercise is here new, and did not occur in any earlier work of Plato. When Parmenides recommended dialectical exercise, he took as subject of this ‘play’ at once the highest notions of the one and the many; also the illustration of the nature of justice in the Republic through the idea of the state was not a vulgar example. Now we see that any insignificant object is admitted to be a convenient model for logical exercise (218 D: Bowne Sita epi Tivos TeV gavrov periovres trepabapuev trapdderypa avTo Oécbat rod pecCovos). Here we are at a considerable distance from the time when observation of stars appeared to be a use- less and even pernicious occupation if not immediately connected with a knowledge of the general laws of astro- nomy. Now not only stars, but all animals and plants come within the range of observation and investigation. When Plato in the Republic described the philosopher as desiring intensely every kind of knowledge, he had not yet drawn all the consequences from this universal desire, and he despised many kinds of knowledge which in the Sophist are gravely included in the system of scienée. Newly discovered kinds are named by means of new words, with the observation that we ought not to pay too much attention to the existing names, which are often understood in different ways by different men (218 c: de del TWavTos Tépt TO Wpaypa aVTO padrov 81d Adywv 7) ToUVOLA povoy cuvoporoyncacba ywpls Noyou). Dialectic is no longer, as in the Republic, the knowledge of the Good, but the science of division of notions, as in the Phaedrus. This important coincidence between the Phaedrus and the Sophist (253 0 D: émriotypns .. . tows THs peylotns ... Tay sdevdlpwv ... 76 KaTa yévn SiarpeicBat nal pnte tabrov by sidos srepov ayjoacGa pte Erepov dv tavTov ... Ths Svadrextixns dycomev arriotyuns elvar) i8 difficult to account for by those who place the Phaedrus before the Republic. In earlier dialogues dialectic was Dialec- tical exercise to be first used on obvious examples. No object of know- ledge to be despised. The logician is not to be misled by common language. Division of con- cepts a link be- tween the Phaedrus and the Sophist. But the process is here more elabor- ately described. Commu- nion of ideas not transcen- dental. Propae- deutic through playful defini- tions. 422 ORIGIN AND GROWTH OF PLATO’S LOGIC merely the art of asking and answering questions (Crat. 390 c), as it was for Xenophon. Now the dialectician follows each idea through its manifold appearances, and distinguishes within each notion many differences, uniting again one notion with many others into one higher kind (253 D: piav iSéav dia trodrwv, &vds Exdorou KEipévou yupis, mavrTn Siarerapévnv ixavos StacoOdvetat, Kal troAXNas étépas GAnAwV bird wads wey repteyopsvas, cal piav ad de’ Odov moAnay av évi Evynupevnv, cal wodrAds ywpls wavtn Siwpic- pévas...). The ideas here mentioned can evidently only be notions of the human mind, never the self-existent ideas of a space above heaven. The aim of dialectical operations is precisely to learn the relation between ideas (253 E: todro 8 égotw, 4 Te Kowwvelv Exacta Svvatat Kal On py, Svaxpivew Kata yévos ériarac@a.). Many definitions of notions are given, and we are asked to determine the specific difference which distinguishes each notion from others of the same kind (232 a). Here again, as in the Theaetetus, Plato insists upon the difference between an enumeration of examples and the definition of the class to which these objects belong (240 A: ro 81a rdvtwy TovTwr, & TOANA sitr@v HEiwoas évl mpocerety ovopat, POeyEdpevos eidwXov aari Tacw ws dv év). We have here the teacher who warns his pupils repeatedly against familiar logical errors. His own defi- nitions are not always serious, as, for instance, when he calls the sophist a paid hunter after wealth and youth (223 B), a merchant in the goods of the soul (224 6, cf. Prot. 313 c), a retailer of the same sort of wares (224 p), a manufacturer of the learned wares he sells (224 &), a money-maker of the eristic kind (226 a), a purger of souls who clears away notions obstructive to knowledge (231 £), @ Magician and imitator of true being (235 a), and a dis- sembler who in private and in short speeches compels the person who is conversing with him to contradict himself (268 c). This is intended to show the various relations of notions apparently very distant from each NEW THEORY OF SCIENCE: THE SOPHIST 423 other, and can only be taken as a sample of dialectical exercise. There are definitions of other notions to which a serious importance seems to have been attached, and one of these generalises a view already enunciated in the Symposium : Soph. 265 B: mouriny racay éhaper eivac Svvauiy, 7 tis dy alria ytyyynrat Trois py mpdrepoy ovow Vorepoy yiyver Oat. 219 B: . . . wav Grep dv py mpo- repov Tis by Vorepoy eis ovoiay ayn, TOV pev Gyovra motety, TS O€ dydpevoy motecc ai rrov apev. Symp. 205 B: moinais éoriv te TOAY + H yap Tot éx Tov py Svros eis TO by lovre dr@ody airia wacd éore moinots, Gore kat ai trd macats rais réxvats épyaciat motnoets eiot Kat oi rourwy Snusovpyoi mavres moinrai. This definition of creation as the power of bringing into Being anything not existing before presents in both dialogues a characteristic difference resulting from the increasing importance attributed to the personal agent. In the Symposiwm Plato spoke of an impersonal cause of new existence and named it for the purpose of his argument ‘poetry,’ thus extending the notion of poetry to all kinds of making. In the Sophist the formulation is sharper, and the opposition between the agent and the object of activity is introduced, with the use of the favourite term dvvayis, familiar since the Republic. More important is the definition of true Being as any- thing that has the power of activity or passivity, to act or to undergo an influence from anything else, be it even only once (247 D: Adyw 87) 7d Kal drovavody Kexrnpévov Sivauw er’ eis TO moteiv Erepov ottovy Teduxos elt’ eis rd wabeiy xa ocpuixpotatov ind Tov havdoTdTov, Kav a povov eiadrak, wav todro dvrws elvar* TiMewas yap Spoy dplteyv ra dvra, ws kori ox Aro Te wAHV Sivas). This is here proposed after the complaint that none among the earlier philosophers has given a definition of Being, and that many would be unable to do it (247 p). Thus we must accept it as Plato’s own view at the time of writing the Sophist. This definition does not correspond to the Others more serious. ‘ Making ] includes more than poetry. Definition of Being in reply to the Mate- rialists: as capability of acting or being acted on. Dynamic notion of existence, Not, like the old ideas, un- alterably fixed Know- ledge an activity. Ideas are no longer true Being. 424 ORIGIN AND GROWTH OF PLATO'S LOGIC primitive ideas, which according to the Symposium re- main unaffected by the changes occurring in the world. If we compare it with the definition of the soul as the first cause of movement, it becomes very probable that Plato attributed true Being to souls more than to anything else, and this is confirmed by the following argumentation in which the author states clearly that the soul acts in acquiring knowledge, while the substance of things under- goes the influence of the soul’s activity (248D: riy yuyqny yuyveocka, thy 8 ovoiav yuyvooxecOar .. . . E: THY ovciay 57) . . yuyvooxopéyny U70 THs yuorews, Kal’ dcov yuyvMoKETal, Kata tocoirov KiveicOa did 7d waoyetv, 5 bn pauev ov dv yevéoOae rept To npEpodr). If this view is maintained, the objects of knowledge are here not unchanging and unaffected ideas, but our own notions, which undergo some changes under the influence of our intellectual activity. This agrees well with the view put forth in the Parmenides, and we may accept it as Plato’s conviction with the restrictions which are made by himself in connection with this passage. He says that true Being must have movement, life, soul, and reason (248 BE: as ddnOas kivnow Kai bwiy Kal yuynv Kai dpdvnow 7 padlws racOnodpeOa tH Tavredds SvTe pn mapsivas, unde Shy avro pnde hpovely, dAAA cEepvoy Kal aytov, vouv ovx syov, axlynroy éords elvat;). Students of the Sophist who read this dialogue with the prejudice that true Being can never mean anything for Plato besides the ideas, have drawn the curious inference from this passage that Plato credits here the ideas with life and a soul— why not with a body also? Such ideas, if still named ideas, could evidently be nothing else than individual beings, very similar to human persons. Any unprejudiced reader who remembers what is said in the Phaedrus about the soul as origin of movement, and in the Laws about the stars as bodies of individual gods (967 a-E), must infer from this passage that here true Being means no longer ideas but souls, including NEW THEORY OF SCIENCE: THE SOPHIST 420 human souls. This view is well prepared by the theory of unity of consciousness in the Theaetetus and by the contradictions shown in the Parmenides as resulting from self-existing ideas. Only the circumstance that the dialectical dialogues, being more difficult, were less read, could lead to the reigning conception of Platonism as a mere theory of ideas. We have seen that the ideas ap- peared first in the Sympostwm and were maintained only in three other dialogues (Phaedo, Republic, Phaedrus), undergoing a change from immanence to transcendence, and becoming at last ideal models of things, which apart from their copies retain their own existence. After the Parmenides we have no reason to identify true Being with ideas in this sense. We shall see in later works of Plato that he more and more dedicated himself to the investigation of notions of his own soul and of the par- ticulars of experience. He says unmistakably that reason and life are possible only in a soul (249 A: vobv pav eyeu, Cony d2 un, POwev ;—Kai Tas ;—addAA TadTa pev auporEepa évovr’ avT@ Asyouev, Ov nv dv Yruyyn ye dynoopev avo Eyew avra ;—k«al tiv’ dv érepov éyot TpoTroy ;—adXa O7Ta voby pav cal Swnv Kal yuynv, axlvnrov pévrow TO Trapatray, Euyruyov bv, sordvat ;—mdvta Eyorye didoya tadr selva daivetas. —xal 70 xivodpevov 87 kal Kiynow ovyywpntéov ws ovTa). We see here movement recognised as true Being. In the Phaedrus and Laws the cause of movement is the soul. Here equally in the whole passage the soul is identified with true Being. The only difficulty of interpretation might be seen in the ambiguity of the term ‘soul,’ as it is not always the individual soul. But we have seen that in the Phaedrus the individual soul was meant, as results from the avowed purpose of the exposition there given. Equally in the Laws the priority of soul has a practical application to the individual life of each citizen, and unity of soul in the universe is even denied. Thus we must admit as Plato’s view a plurality of souls, and this agrees with the myth of the Timaeus. In the Timaeus these Thus the theory of ideas has been modified : first pass- ing from imma- nence to transcen- dence, then be- coming models of things. Now they are notions inherent in a soul. Plurality of souls acting and being acted upon. A certain fixity or stability still required in the objects of know- ledge. Existence implies unity and totality. The com- munion of kinds. 426 ORIGIN AND GROWTH OF PLATO’S LOGIC souls are said to be created by one universal creator. But this is a mythical allegory which means only the sub- stantial similarity of all souls. Whatever Plato’s opinion about the relation of the individual human soul to the Divinity could have been, so much isclear from the above comparisons, that he credited the individual soul with true existence, the power of acting and being acted upon. The movement of the objects of knowledge is limited by Plato in so far as without the fixity of notions knowledge appeared impossible (249 c: ro cata tavTa Kal daavtws Kai tept TO ato Soxsi cor ywpis ctdcEews yevéoOat trot’ av ;— ovdapas *—ri 8 ; dvev tovtwy votv xaBopas dvta 7h yevouevov dy nai orrovoiv ;—ixtora). The object of philosophy is the divine substance of Being, which is not attainable to vulgar minds (254 A: o ¢tAdaogos, tH Tod Svtos det dia Aoyicu@v mpockeimevos idéa, Sia TO Aapmpov av THs ywpas ovdapas evrretns opOjvat * Ta yap THs TOV TOM@Y Wuyis dupata Kaptepeiy mpos TO Oaiov abopavra advvata). But this does not mean that we have to imagine this substance as the idea of Good in the Republic. The notion of Being extends to all individual things (237 D: Kat tovTo piv wrov pavepov, ws Kal TO TL TOUTO pha at’ Ovte Neyouev sxadoroTe* povoy yap avTo Aéyew, @oTrEp yuLVvov Kal aIrnpnuwpévoy atro TaV dvTwY atdyTwr, aévvatov) which constitute unities of thought (237 D: avdyen Tov Tt Asyovta sy yd te Afyewv), each of them an existing whole (245 D: obte ovclav ovre yéveow ws ovoav Set rpocayopevew To by } TO Gov av Tos overt un) TLOEVTA). The theory of the mutual relation («owywvia) of notions among each other is proposed after the refutation of two contradictory suppositions. That all notions cannot be predicated of each other (252 D: mavra dddAj dos sdpev Suvapiy syew srixowwvias; ... TovTO yé Trou Tals peyloTats avdykxats ddvvarov) is seen from the impossibility of join- ing in one judgment contradictory ideas, as, for instance, immobility and movement. On the other hand, if each idea stands apart from all others (251 E: pndev) pndev NEW THEORY OF SCIENCE: THE SOPHIST 427 pndepiay Svvayw Eyew xowwvias sis pndév), all reasoning becomes impossible. It remains only to admit that some notions agree and others not; a special investigation is needed to find which is the case in each instance (253 A). This is illustrated by the example of letters, which form syllables and words only in certain combinations, deter- mined by the science of grammar. Similarly the combi- nations of ideas are the object of dialectic. But Plato warns us against the illusions of thought which can be produced by the charm of skilful eloquence (234 c). The recognition of the power of a perverse rhetoric goes so far that it implies a certain opposition between pure thought and acquired experience, conceding to the latter the power of correcting the illusions of thought. Such a view is far removed from the triumphant idealism of the Republic and Phaedrus, and cannot be interpreted other- wise than by an increasing esteem of outward experience, which is common to the Sophist and the Laws: Soph. 284 D: rovs modAovs trav rére dxovévrwv dp” ov« dvdyxn, 3 > nm € a“ Xpdvou re éreAOdvros avrois ixavod kal mpotovaons nAtKkias, Tois Te ovat o > , ‘ A mpoorminrrovras eyyvbev Kai dua mran- pdrev dvayxalopuévous évapyas épa- nrecOat Tv dvray, peraBdaArdecy ras rére yevonevas Sdéus, Sore opicpa pev atverOar ta peydAa, xaderd 8€ ra padia, kat wavra dvr 3 o 4 3 ~ 4 avarerpad@at ra é€v rois Ad- yots phavracpara ind rav év rais mpdgectv epywy rapayevo- pevav. Legg. 769 D: rovovroy Tov vopo- Oérov BotAnpa - mparov pev ypayra ToUs vdpous mpos THY axpiBeay Kara dvvapey ixavas : éretra mpotdvros rou Xpdvov kal rav dokavrav epye metp@pevov . . . maumoANa avayxn mapaneirec Oa roavra, A det rivd Evverrduevoy émavopGovv . . . 888 A: véos ef > mpotdy 8¢ ce 5 Xpévos tromoe mod\dka by viv 8ofalecs peraBadrdvra én rap- avria ribecOat * mepipevov ovv eis Tore kpiTns wept Tay peyiorwy yiyve- vbat. It was a natural consequence of the extension of detailed investigations that Plato began to think more highly of experience than he did at the time when he was still inebriated with his discovery of absolute ideas. For the same reason it is impossible to explain the above passage without the admission that the writer is an aged man. He knows that truth is reached through bitter Only certain combina- tions possible. Increasing recogni- tion of the value of ex- perience. The highest kinds, or categories. Being and Not- Being. Not-Being always relative. Not-Being is dif- ference. Prepara- tion in previous dialogues for this concep- 428 ORIGIN AND GROWTH OF PLATO'S LOGIC experience, and that experience can prevent the pain to which youth without a guide is often exposed (234 E: nets oe olde mdvres tepacdueOa Kal viv metpwueda ws dyyutara dvev Tov TaOnudtwy mpoodyew). This concession to practical experience, which led to the substitution of a second best state for the ideal Republic, did not change the fundamental postulate of earlier Platonic logic, namely the fixity of ideas, without which knowledge and reason would become impossible (249 c: apds ye Tobroy travti Ady@ wayeTéor, ds av éemiaTHunv 4 ppovncey % vodv adavifwy icyupl{nrat mepi Tivos orryody). The ideas exist in the soul and are quite as invisible and intangible as the soul in which they abide (24748). There are certain highest kinds (254 D: péyiota Trav yevov), which Plato enumerates as Being, rest, motion, identity, and difference (dv, ardovs, civnots, TavTov, OadtEpov, 254 D E). The idea of difference explains the notion of Not-Being which presented such difficulties to Plato’s predecessors (237 c-238 D). Being is absolute or relative (255 c: olua: a8 ovyywpey Tov dyTwy TA “ev avTA Kad’ avTa, Ta 52 pos GAnXra ae AéyerOar), while Not-Being is always relative. It is impossible to affirm that something contradictory to Being exists (257 B, 258 £). But Not-Being means only different Being, and denotes the relation of notions which do not agree with each other (256 D). Of each thing an infinity of negations can be predicated, because we can compare with each Being all different Beings which are not what the chosen Being is (256 B: qwepl txaoroy dpa Tay sidav Tord pév sore TO Sv, Arerpov 88 TAN To yw by ... 2574: wal to dy... dca mép dott Ta adda, Kata tocaita ovx tori: axsiva yap ove by by piv ado sori, atrépavra 88 tov apiOpdv Tada OvK Eotiy ad). This logical solution of the riddle which caused so much difficulty to Parmenides has been prepared already by the mention of a perception of opposites in the Theaetetus (186 8B), and by the antinomies of the Parmenides. Such antinomies would have no meaning after a definition of Not-Being as NEW THEORY OF SCIENCE: THE SUPHIST 429 different Being, and after the transition from a meta- physical idea of Not-Being to the logical conception of Other-Being. The term Not-Being had been used already in the Republic (478 c), where, as in the Parmenides, it was declared impossible to be a subject of thought or opinion. Such a declaration coming after the inquiry of the Sophist would be unaccountable, and has never been accounted for by those who believe the Republic to be later than the Sophist. For the explanation of Not-Being in the Sophist is not a passing fancy like the creation of an object of ignorance in the Republic. It is an important step in the history of Philosophy, and brings Not-Being from the region of metaphysical speculation into the dry light of formal logic. It is a consequence of the recog- nition of Relation as a chief factor of knowledge, without which error in pure thought is inconceivable (237 a: Teropinney 6 Adyos oTOS UTrodecOar To wy dv elvar* reddos yap ove dv adrws ayiyveto Sv). : If the ideas were always perceived as they are, the participation of concrete things in them would allow of a determination free from error. But as the relations of ideas between each other are not evident to our intuition, we commit errors by supposing relations which are not. The question of error was left unsettled in the Cratylus (429 p), and in the Theaetetus (187 pb, cf. 200 pd). It is only here that Plato explains error as a judgment about Not-Being, while in all earlier works the possibility of thinking or judging Not-Being was denied in agreement with Plato’s philosophical predecessors. Not-Being is recognised as a notion in one line with Being (260 B: +o pav Sh wr) dv quty gv te tov Gddwv yévos bv avedavn, cata ravrTa Ta OvtTa Siwctrapuévov), from which it differs by its relativity. While the elements of earlier Platonic logic were single ideas, the importance of judgment is here asserted as a first element of knowledge. Judgment is analysed into tion of Other- Being. These imperfect views could not be later than the Sophist. First clear concep- tion of Relation aS & con- dition of thought. Conse- quent possibility of errors. Judgment a first element of know- ledge. Subject and pre- dicate. The terms here first accurately defined. The Cratylus compared. 430 ORIGIN AND GROWTH OF PLATO'S LOGIC its essential parts, and for the first time Plato establishes the distinction between the subject and predicate of a pro- position (261). He divides the signs used in language into dvéuata and pyjyara and states expressly that pjya means the sign of an action (262 A: ro él rais wpakeow by Syrwpa pHud mov Azyouev). Thisis a new term, because in earlier dialogues piua, even if used along with dvopya, meant a phrase or expression. Thus, for instance, in the Protagoras (341 8, 3438) the term pjya is used for sayings of Simonides and Pittacos. In the same meaning of a saying or phrase pjya is often used (Prot. 342 8, Crat. 399 B, 421 B, E, Rep. 336 a, 463 £, 498 B, 562 c, Phaedr. 269 B, Theaet. 190 c, Legg. 660 a, 669 c, 839 B, 840 oc), also in such expressions as pyya Kxal Noyov (Rep. 473 &), ovopara Kai pnwata (Apol. 17 B, Crat. 425 a, Symp. 198 B, 221 B, Rep. 601 a, Theaet. 168 B, 184 c, 206 D), pjua nat Sdyya (Rep. 464 a, Soph. 265 c, Legg. 797 c). In other cases piua means a single word (Rep. 462 c, Theaet. 165 a, 183 B, Soph. 237 p, Tam. 49 £, Legg. 627 D, 656 c, 669 Ez, 783 c, 800 D, 906 c) or textual expression (Huthyd. 305 a, Gorg. 450 E, 489 B, Phaed. 102 8B, Rep. 340 p, Phaedr. 228 p, 271 c, Theaet. 166 v, 190 c, Soph. 257 B). It is quite another thing in the above passage of the Sophist in which évoua and pjya have each an unmistakable technical meaning, as subject and predicate, clearly intro- duced for the first time. The term fjjya is used in this meaning of predicate also in some later instances (Polvt. 303 c, Legg. 838 B). If we compare Cratylus and Sophist on the connection between évoua, siya, and Noyos, it might at first sight appear that the later dialogue repeats only a definition given in the earlier : Crat. 425 4: éx rav évoparoy kal pypdrev péya fon rs Kai Kaddy kal 6d\ov ovoTnropey .. . TOY Adyov rH svopacring } pyropiKy # qris €oriy 7 TéEXYN. 481 B: el ors py dp0ds diavépew ra évdpara .. . ein dy xal pyyara Soph. 262 a: €& évopdray povwy auvex@s eyoudvov ovx gore sore Adyos, 008 at pnudrov yupis dvopdray dex Oévrav. C: ovdepiay yap ofre ovras ofr’ éxelvos mpakw ov8’ drpagiay ovde ovaiay dvros ov8é pi) Svros dndoi ra NEW THEORY OF SCIENCE: THE SOPHIST 43] Tavrov rovro motety. ef S€ pnuata kal gwrnbévra, mpiv av ris rois dvopace évopara gorw ovrw riWévat, dvdyen ra prpara Kepdon * Tore 8’ fppoce kai Aoyous* Adyot ydp mov... Te Kat Adyos eyévero evs 7 mpaTN rouray ~vvbecis dori. cuprAokn, axeddv trav oywr 6 Mp@ros kai opixpéraros. Many translators understood pjya in the above passage of the Cratylus as‘ verb’ or ‘ predicate,’ but if we compare other passages of the same dialogue 1t becomes evident that here also pjua means ‘phrase.’ Plato deals with a suc- cession of increasing units, beginning with a single letter, progressing to a syllable, a word, a phrase, and a speech. The parallelism of dvopactien and dvopa, pnropixyn and pyua confirms this, and Adyos means here not a sentence but a speech, or language generally; also in the second passage the progress from a wrong distribution of words to a wrong distribution of phrases is a plausible induction, while it would be unjustifiable to apply to the Cratylus a definition given only in the Sophist, and received first by the pupil as requiring nearer explanation (2624: tar’ ovx guabov, C: was ap’ woe reyes ;). Even if we had not many other reasons to admit the priority of the Cratylus to the Sophist, this comparison would show that the distinction of subject and predicate, made in the Sophist, must be later than the opposition of words and phrases, which in the Cratylus is already familiar at a time when the need of a theory of predication was not yet felt. A judgment, says Plato here, refers to things present, past, or future, and connects a predicate with a subject (262D: Snroi yap 75n wou rére TEpt TOV dvTwV F yuyvomevwn } yeyovdrwv 7) wedddyrav, cai ove dvonater wdvor, ANG Tt Kal Tepalver, cuptTrAdcwy Ta pyyata rots dvopact). This con- nection is not, as some logicians even now suppose, limited to an identity of subject and predicate, but presents a great variety of aspects (251 A: Adéyousv dvOpwrrov 81 srov mwOAN arta srovondlovtes, Ta Te ypwpata smihepovTes avT@ nal taoxnpata Kat peytOn Kai xaxias Kai aperds, dv ols ract xal érépors puplors ob psvov avOpwrov avror elvat dapdy, Gdrd Predica- tion does not imply identity ; variety of predi- cates. Negation not con- tradiction. Refuta- tion an instru- ment of moral training. Import- ance of the new theory. 4382 ORIGIN AND GROWTH OF PLATO'S LOGIC kal aya0ov Kal Frepa area). This is misunderstood by those who are unable to grasp the relation between the one and the many, and believe that each judgment implies an identity (251B: Oey. . . Trois Te véots nal Tov yEepovTov Tots oyriyuadéor Galyny mapecKevdxapev’ evids yap avtiNa- BécOat trayti wpdyeipov ws aduvaroy td Te TrodAa by Kai TO § moda eva, nal dn mov yaipovow ovx éavres ayalov Afysiy dvOpwirov, adAd To pav ayadov dyabov, tov 82 dvOpwirov dvOpwirov). What is here explained about the nature of the sentence applies also to the unspoken judg- ment (263 EB: Sidvora Kai Noyos TavToy’ mAHV O pap evTos THs Wuyns wpos abriy Suadoyos dvev havas yryvepevos rodr’ avro Huiv stavopdacbn, Sidvoia ... Td 54 y am’ sxelvns psdpa Sia rob orduaros iov pera POdyyou KéKAnTat Adyos). The negative judgment is not contradictory to its positive counterpart, and the negation means only a difference, leaving open an infinity of possibilities (257B: ov« dp’, dvayriov Srav amddacis NéynTaL cnpaivew, cuyywpnooueia, Togovroy 52 wovov, Ste THY ANNwv TL uNvUEL TO fy Kal TO ob mporisneva Tay ariovTwY Gvopdtav, padrdrov 88 Tov mpayudrov wept arr’ dy xénras ta dribOeyyopeva tortepov Ths atopacews ovéuata). Wrong judgments are refuted by showing the contradictions they imply (2308). Such refutations are extolled as being not only of logical but also of moral importance (230 D: roy edeyyov AexTEov ws dpa peyiorn Kat Kxuptwrdtn trav Kabdpoewy tots, Kal Tov avineyxtoy ad voymioréoy, ... Ta p&ytota axdOaproy dvra, drraldeurdv te Kal aioypoy yeyovsvat). Plato presents his theory of negation and of predica- tion as a truth which alone can account for the existence of error, and could only be denied under the penalty of being involved in constant contradictions (241 £). Ignor- ance, named here an ugliness of the soul (228 4), is always involuntary (228 0: yuyny ye louev dxovoay Tracay Tiév ayvootcay), being worst 1f he who is ignorant is under the illusion that he knows (229c: dyvoias ... péya cal NEW THEORY OF SCIENCE: THE SOPHIST 433 yareTrrov apwpicpévov eldos . . . TO oy KaTedota TL Soxeiv giddvat . . . TOUT@ LOVy THs ayvolas duabia Tovvoma). The most impressive passage of the Sophist (242 c- 251 A) represents the metaphysical and logical conflict between materialism and idealism, wherein Plato chooses a middle solution, thus confirming his criticism in the Parmenides of the primitive theory of ideas. The im- proved materialism here represented has, with some plausibility, been attributed by Siebeck * to Aristotle ; the idealism here represented bears some relation to Plato’s own views as expressed in Phaedo, Republic, and Phaedrus. The third or middle view proposed is the true existence of souls, not of animated ideas as some critics thought. Here, exactly as in the tenth book of the Re- public and the Phaedrus, Plato, at the end of an argument on another notion, suddenly introduces the soul as corre- sponding best to the general notion first explained. There it was the notion of a self-moving principle—here it is the notion of true Being (248 E: 16 wavredds Sv) which, besides movement, as postulated in the Phaedrus, must have reason, and if reason, necessarily life (2494: vodv pav tye, Conv S¢ uy, G@psv ; xalards;). But reason and life are found only in a soul (249 A: radra pav audorepa dvovt’ avT@ (T@ TravTedas Svtt) Aéyouev, ov unv sv Yruy7 ye pyjooper avo Bye avira ;—xai tly’ dv srepov you tpéTrov). It results that the soul or souls correspond best to the idea of true existence, though Plato at the end does not insist on this conclusion, because his aim was only to show that both materialists and idealists have a too narrow con- ception of Being (246 4): earlier philosophers have taken it lightly, and spoke of quality and quantity of Being without a definition of their starting point (2420: edxorws pot Soxet Llappevldns jyiy Siatrdy Oar nal was Soris morore art kplowy dpynoe tod ra byta Sioplcacbas mréca 34 H, Siebeck, ‘Platon als Kritiker aristotelischer Ansichten : III. Der Sophista,’ in Zettschrift fir Philosophie und philosophische Kritik, Band 108, pp. 1-18, Leipzig 1896. FF Conflict between material- ism and idealism. Plato’s mediating view. Dynamic aspect of Being identifies Being with Soul. Narrow- ness of earlier concep- tions. Being becomes the chief object of research, not the Idea of Good. 434 ORIGIN AND GROWTH OF PLATO'S LOGIC Te kal trota ori). They invent fables as if they were speaking to children (242c: piOov twa Exacros paivetai por SinysicOar twatciv ws ovo nuiv) instead of analysing the chief concept of philosophy, the idea of Being. This substitution of Being as the ultimate aim of Dialectic instead of the earlier hegemony of the Good is one of the signs of the change which occurred in Plato’s thoughts, from absolute ideas to the ideas of the human mind. At the same time the bold review of philosophical doc- trines betrays a Master in metaphysics who could be nobody else than Plato alone, so that all doubts as to the authenticity of the Sophist must be dismissed. Those who up to quite recent times ascribed the Sophist to another writer *> had not considered the close 28 Firnst Appel (‘ Zur Echtheitsfrage des Dialogs Sophistes,’ in vol. v. pp. 55-60 of the Archiv ftir Geschichte der Philosophie) and Huit (in vol. xviii. pp. 48-69, 169-188 of the Ammales de philosophie chrétienne, Paris 1888) have added nothing to Schaarschmidt’s arguments, which have been abundantly refuted by R. Pilger (Ueber die Athetese des platonischen Sophistes, Berlin 1871) and many others. Huit adds only a very strange objection (p. 175): he believes that the historical character of the Sophist is unplatonic, and that Plato never reviews his predecessors. This needs no refutation for anybody who knows the Theaetetus, the Phaedrus, or the Phaedo. Fouillée (La Philosophie de Platon, Paris 1888) was right in saying (p. xii. Préface) that to deny the authenticity of the Sophist and Politicus ‘il faut étre myope intellectuellement.’ The logical importance of the Sophist has been recognised among other authors by: Bertinj (Nuova interpretazione delle idee Platoniche, Torino 1876, p. 23 sqq.), Achelis (‘ Kritische Darstellung der platonischen Ideenlehre,’ pp. 90-103 in vol. 79 of the Zeitschrift fiir Philosophie und philosophische Kritik, Halle 1881), Benn (The Greck philosophers, London 1882), Peipers (Oniologia Platonica, Lipsiae 1883, pp. 319-346), Lukas (Die Methode der Eintheilung bei Platon, Halle 1888), Apelt (Bettrdge sur Geschichte der griechischen Philosophie, Leipzig 1891, pp. 67-99, also pp. 529-540 of vol. 145 of Neue Jahrbticher fir Philologie und Pddagogik). Very peculiar are the views of Wolff (Dre Platonische Dialektik, Halle 1875), who thinks that Plato ignored ‘den Unterschied zwischen Gattung und Art,’ and Uphues (Das Wesen des Denkens bei Plato, Landsberg 1881), who credits Plato with such opinions as: ‘das Denken ist eine Verbindung der die Vorstellungen verbindenden Worter zu Siatzen,’ and resumes his opinion on Plato’s logic thus: ‘das Versténdniss des Satzes wird uns nicht durch ihn selbst sondern durch ein Anderes gegeben; als Quelle unserer Erkenntniss der Wahrheit kann nicht der Satz sondern muss eben dies Andere gelten. Dieses Andere ist die NEW THEORY OF SCIENCE: THE SOPHIST 435 stylistic relations between the Sophist and the Laws. The strangest of all objections to the authenticity of the Sophist rests on a very low estimate of Plato’s sincerity. It has been said that Plato would not have criticised his own theories as the author of the Sophist criticises the Platonic ideas. Such critics seem to measure Plato’s ambition according to the standard of a vulgar school- master. The dialogical form of Plato’s works left him a great liberty for introducing new theories, attributing them to new speakers. In the Laws many political theories of the Republic are abandoned, and thus also the Parmenides and Sophist take leave of the theory of ideas as expounded in the Phaedo or Phaedrus. The Sophist appears to be in every respect a continua- tion of the Parmenides and a fulfilment of a part of the programme there proposed. There are at least two passages in which the Parmenides is alluded to in the later dialogue : at the beginning (217 c), where the form of the dialectical discussion of the Parmenides is men- tioned in an unmistakable manner, and at a further stage, where an equally clear allusion is made to the contents of the antinomies (244¢: 16 tavrny tH Urobeow tbrodEnévp (ro &y eivat), mpos TO viv EpwrnOdv, Kal mpos Addo 4é OTLOdy, ov Tdavtwy padotoy atroxpivacOa: cf. 245 E: cal dd\Xa pupla dmepavrous atroplas &xacroy siknpos paveirat TH TO dv elite Svo tive cite 8v wovoy elvyat Asyovte™®). What Zeller says in order to invert the relation and to place the Parme- nides after the Sophist is by no means convincing. After the determination of negation in the Sophist a great part of the antinomies of the Parmenides would be superfluous, as can be seen from what is said in the Parmenides about Not-Being (Parm. 142 A: 7@ pn Svte od8 Svopa od Novos christliche Trinitétslehre.’ This touching simplicity is equalled only by Pfleiderer, who sees in the Sophist ‘die Ehrenrettung des richtigverstan- denen Nichtseins ’ (p. 847). 26 This passage, in which True Being appears neither as only one, nor as Two opposite, seems also to imply a plurality of Beings, or souls, as the ultimate solution of the metaphysical problem. FF 2 Stylistic relations. Criticism of the cruder theory of ideas. The Sophist subse- quent to the Parme- nides. Zeller’s parallels uncon- vincing. 436 ORIGIN AND GROWTH OF PLATO'S LOGIC ovdé tis émriotnun ovd2 aicOnots ovds Soka: cf. 164 B). Zeller quotes several parallel passages of both dialogues which either prove nothing about the chronological order or even confirm the priority of the Parmenides : Parm. 128 £-129 c: the parti- culars are said to participate in ideas and even in opposite ideas, and Socrates adds: ¢«f & gorw év avro Touro 7oAAd arobei~et, Kai ad ra mwodAd 87 €v, rovro dn Oavpd- wopat.... et pey aura ra yérn Te Kat etdn év abrots arodaivos ravavria. Soph. 251 a-c: the multiplicity of predicates referring to one sub- ject is denied only by persons tré mevias ths wept pdvnow xrh- ews Ta ToLaUTa TeOavpakdot .. . 258 p: the dialectician distin- guishes ideas and their rela- tions. In the above two passages the Hleatic stranger takes for granted what Socrates in the Parmenides represents as a great and unsettled difficulty. If any chronological inference is allowed from such general coincidences, the later date of the Sophist is the most probable conclusion. Other passages compared by Zeller are quite as incon- clusive : Parm. 188 c: ofpa ay xal oeé cat Gdov, Goris airnv riva Kal? avriy exdorov ovgiay ridera eivat, époroyjoa dy mporoy pev pyndepiav atrav elvar ¢v nyuiv. mas yap ay avr cab? atrny ert ein; ... doas trav ideav mpds addAndas eiolv at elgiv, aital mpds atras tiv otciay €xovowv, GAA’ ov mpos Ta tap’ Huivy dpowpara. Parm. 148 a: ovcias gapev peréxey rd €v, Od Corw ... xa da ravra 3) rd év dy rodAa edn. .. avrd rd év, & 89 hapev ovcias perexesy, day adrd ry Siavoig pdvoy xa’ atrd AdBwpev dvev rovrov ov hapev perdxew, God ye éy pdvov ha- ynoerat # wai roANa 7d abrd rovro; —év, oluas fywye ... B: Go re Erepoy yey avdynn Thy ovciay atrov elvat, Erepov 8€ adr; elrep pn otoia rd dy, GX’ ds dy obcias perioxer ... el €repoy peév 7 ovcia, érepow Se Soph. 255 p : oiuai oe cvyxwperv Tov Gvrav ra pev avira xa?’ adrd, ra S€ mpos GAAnda aet Adyea ba... ro 8’ érepov det mpos érepov ... etrep Odrepov duoty peretye roi eldoiv Sowep ro dv, Hv ay more re kal tov érépwv Erepoy ot mpos érepov: viv 8 drexvas nuiv 6 ri wep Ay érepov 7, cuuBéBnxew €€ dvdynns éréyou Tovro 6 mép éarty eivat. Soph. 244 B: &v wov dare pdvov eivat ;—apev ydp—bv xadeiré re; —vai—norepov érep év, evi rq attra mporxpapyevos Svoiw dvdpaciv, ff wos ;—the answer to this ques- tion is stated to be difficult, with a very probable reference to the Parmenides, in which precisely the same question led to contra- dictory conclusions. The theory of communion of kinds as set forth in the Sophist may be regarded as an attempt to solve NEW THEORY OF SCIENCE: THE SOPHIST 437 the riddles of the Parmenides, and to progress beyond the rigid Eleatic unity of Being. 7d év, ore ra év rd ev ris ovcius érepov obre tq otcia elvar 4 ovcia rov €vds GAXo, GAAa T@ Erepo re xal GA@ Erepa GrAnrov. The above comparisons speak rather against Zeller’s conclusions, though the chronological value of these passages is much less evident than the above quoted references to the Parmenides in the Sophist (217 o, 244 c). The general contents of both dialogues are best explained by the priority of the Parmenides. The Parmenides prepares the ground for the theories of the Sophist, and is as we have seen intermediate between Theaetetus and Sophist in its list of categories as well as in its antinomies. This is confirmed also by numerous stylistic observations. The vocabulary of the Sophist, despite the difference of contents, shows such a surpris- ing number of coincidences with Timaeus, Critias, and Laws, as no earlier dialogue except the Phaedrus. We have already seen what reasons account for the excep- tional style of the Phaedrus, and there are many stylistic peculiarities in which the Sophist is much nearer to the Laws than the Phaedrus and Parmenides. To these belongs first of all the avoidance of hiatus, which cannot be accidental, and is common to the Sophist with the five latest works. Other important peculiarities absent from the Parmenides, Theaetetus, Phaedrus and all earlier dialogues appear for the first time in the Sophist and remain in the style of all the latest works of Plato: the prevalence of xaQadrrep over @orep, the very great frequency of roivun, mwas, Edutras, and the scarcity of pévro.. Besides these important peculiarities, others of less importance appear for the first time in the Sophist and are common to this dialogue with the latest works of Plato : rw dun, raya ious, Tovyapodyv, wav ovv, pov ov, Inversion Of Aédyes, Evya7ras are found repeatedly in our dialogue, and the number of accidental peculiarities of later style is much greater than in the Parmenides; thus the stylistic affinity of the The list of cate- gories in the Sophist more mature. Vocabu- lary. Hiatus avoided. Other pecu- liarities of later style. Zeller is singular in pla- cing the Sophist before the Republic. Important confirma- tion of the later date by Hirzel and Bruns. R. Hirzel on Dia- logue in Litera- ture. 438 ORIGIN AND GROWTH OF PLATO'S LOGIC Sophist with the group of the Laws amounts to 468 units of affinity against only 243 of the Parmenides. This is quite sufficient to render the later date of the Sophist as probable as anything can be in Platonic chronology, and it has been recognised since Campbell by all investigators of Plato’s style, as well as by many other critics, as for instance Jowett, Tocco, Teichmiiller, on independent internal grounds. Against all this evidence Zeller continues to place the Sophist before the Republic and the Symposium. This is chiefly due to the circum- stance that he is evidently unaware of the existence of so many investigations on the style of Plato, and that he has not given a special attention to Plato’s logical theories. On the other side the late date of the Sophist has been recently confirmed in a most decisive manner by two different lines of inquiry, which enabled two authors, who knew nothing of Campbell, to find out that in two different ways the Sophist and Politicus belong to the same group as the Tiwmaeus and Critias. These con- firmations acquire an increased importance through the fact that they touch upon our problem from a standpoint not yet applied specially to Plato. R. Hirzel *’ dedicated two volumes to a general investigation of the form of literary dialogue from Plato to the present time. This he did with remarkable acuteness, at least so far as Plato is concerned, and he made 1¢ still more evident than Ueberweg and Campbell had done that the form of the dialogue in the Sophist and the dialectical dialogues cor- responds necessarily to a later stage of literary activity than that evinced in the Republic and Phaedrus. In view of the special attention paid by Hirzel to the dia- logical form in the literature of all ages and nations, we are bound to accept his testimony as a valuable confirma- tion of the results obtained by comparison of style and logical theories. Hirzel observes that the change in the 7 R. Hirzel, Der Dialog, ein literarhistorischer Versuch, 2 vols. Leipzig 1895. NEW THEORY OF SCIENCE: THE SOPHIST 439 form of the dialogue consists in many peculiarities, by which the dialectical dialogues are distinguished from earlier more poetical works. The characterisation of persons and of the place of conversation is less elaborate, and the leader of the conversation becomes the impersonal representative of abstract reason, not only without personal character but even without name in the Sophist, Politicus, and Laws. There is nothing in these dialogues to remind us that they are represented as held in Attica or even Greece: they could be imagined anywhere in the universe. Throughout these works we move in a spiritual atmosphere apart from the material world (vol. 1. p. 252: Farb und gestaltlos legt die Welt um uns, Platons Dichtergeist entziindet kein sinnliches Leben mehr in thr, wir befinden uns in einer Geisteratmosphire, die erhaben ist tber Zeit und Raum). The connection of several dialogues into one larger whole is also indicated by Hirzel as a peculiarity of Plato’s latest manner, and he agrees with Christ in the supposition that the term trilogy and tetralogy had been used for Platonic dialogues before 1t came into use for dramatic poetry. The progress from single dialogues to trilogies or tetralogies appears to Hirzel a psychological evolution similar to that which is notice- able in epic and dramatic poetry. Plato saw after the Republic the difficulty of representing very complex sys- tematic expositions in a single dialogue, and he was also led to simplify introductory matters by the connection of dialogues in series. These observations of Hirzel, made in a work of more general aims and not limited to Plato, deserve the most serious attention of all who still have any doubts as to the authenticity and late date of the Sophist and Poltticus. They were unexpectedly confirmed in a most satisfactory manner by another author, who also referred to Plato only in connection with an investigation into another general aspect of literary composition. Ivo Bruns wrote Less marked charac- terisation. Scene indefinite. Dialogues connected in series. ‘ Trilogy’ and ‘ tetra- logy.’ Ivo Bruns, on Literary Portrai- ture, also observes the de- cline of charac- terisation. Plato had retired from life to the School. Less of realism, more of system : hence connected series. This con- nection begins with the Sophist. Neither Republic nor 440 ORIGIN AND GROWTH OF PLATO'S LOGIC a very interesting volume ™* on the literary portraits in Greek literature, from Thucydides down to Demosthenes. This work deals also at some length with Plato as a great artist in skilful personal characterisation. Bruns found this art no longer present in the Platonic trilogies, namely in the Sophist, Politicus, Timaeus, and Critvas. While in the Republic, Phaedrus, Theaetetus, and in earlier dialogues a careful characterisation of each speaker is given, and the philosophical conversation comes as if by accident, Bruns observes that in the Sophist and later dialogues the speakers are not characterised indivi- dually, except that they are spoken of as competent and well prepared for philosophical conversation. This, says Bruns, is a sign that Plato, when he wrote the dialectical works, had retired from life to the School (p. 272: der Unterschied scheint gering, ist aber in Wirklichkeit ein tiefgreifender : er bedeutet den Schrift des platonischen Dialogs von dem Leben in die Schule; er bedeutet das Aufgeben des kiinstlerischen Princips, mit dem der friihere platonische Dialog untrennbar verbunden ist). This is called by Bruns a new style, essentially different from the ‘realistic’ style of the Republic and earlier works, in which each conversation was accidental and ended naturally after a single problem had been ex- hausted. In the trilogies the subject of the conversation is not accidental, but well planned, and this produces the systematic connection of several works into larger wholes. The Sophist and Timaeus are only apparent continuations of earlier dialogues: really each of them begins a new trilogy, and their connection with a dialogue of the old style 1s only employed to avoid an introductory exposition of the circumstances in which the dialogue was started. Neither in writing the Theaetetws had the Sophist been planned, nor in writing the Republic had Plato already formed the plan of the Timaeus; but with the Sophist 8 Ivo Bruns, Das literarische Portrdt der Griechen im fiinften und vierten Jahrhundert vor Christt Geburt, Berlin 1896. 44] and Timaeus begins the plan of two trilogies. The difference between the trilogies and the dialogues of the old style which are supposed to be introductory to the trilogies consists, as Bruns very judiciously observes, in the great authority given to the new leaders of philo- sophical conversation. The stranger of Elea, who leads the dialogue in the Sophist and Poltticus, is expressly recommended at the beginning of the dialogue as a remarkable philosopher (Soph. 216 a: Edvov dyopev . . && "Enéas, éraipov 68 tay audi LappevlSny xai Zyveva, para d2 dvipa dirocopov ... C: Soxet Osos piv ... ovdapas elvat, Ocios nv... 217 B: Staxnxodvar yé now ixavas Kai ovK dpvnpovetv), Similar is also what is said in the Timaeus and Critias about the special authority and preparation of the speakers. We see in all these dialogues perfect teachers, accustomed to repeat their lessons, and well prepared for what they are to say, and hearers equally prepared to receive the instruction. What Bruns says about the psychological motives of this change in Plato’s later style coincides with the similar observations of Ueberweg, Campbell, and also of Hirzel, though Bruns seems not to be aware of this coincidence, or, at least, does not quote his predecessors. His testimony, coming thus quite independently, in- creases our confidence as to the absolute certainty of our conclusions about the date of the Sophist. This dialogue belongs evidently to Plato’s old age, and is much later than the Republic and Phaedrus ; it may even have been written after the third voyage to Sicily. In style and contents there is a progress beyond the Theaetetus which prevents us from seeing in the Sophist an immediate continuation of the former. The external relation between Theaetetus and Sophisé is no sign of a continuity of com- position, just as, in despite of a similar connection, the Timaeus 1s much later than the Republic. NEW THEORY OF SCIENCE: THE SOPHIST Theae- tetus contem- plated other dialogues to follow. The leader of the conver- sation is invested with far more authority. Bruns’ testimony evidently inde- pendent. The Sophist belongs already to Plato’s old age. Continua- tion of the Sophist. The scientific method still in use. Logical method, especially classi- fication regarded 88 & pre- paratory exercise, Defence of lengthy arguments against objectors, who remain unknown. 442 ORIGIN AND GROWTH OF PLATO'S LOGIC Il. The Polsticus. The Politicus is a genuine continuation of the Sophist much more than the Sophist can be esteemed as a con- tinuation of the Theaetetus. Here we have a close and mutual connection: in the Sophist (217 a) the Politicus is announced, and in the Politicus the Sophist is expressly quoted (257 a, 266 D, 284 8, 286 B). This close literary connection of the two companion dialogues corresponds to the near relation of their con- tents and method. The scientific method is here equally praised as leading to truth against every prejudice, and neglecting nothing, however insignificant 16 may appear (266 D: 7H roads pe00d@ Trav Aoywy ote TEevoTépov paNdov sutrnoev pn, TOY Te OMIKpoTEpoy ovdeY HTi“aKEe TPO TOD petCovos, ael Se xa? avrnv mepaiver tadnOdoraroy). This method consists here, as in the Sophist, in the classifica- tion of particulars according to their natural kinds (286 D : 0 Aoyos rrapayyéANel TOAD pddAtoTa Kal mpaToy THv wéodov avThHy Tiav Tov Kat’ edn Suvartov atvar Siacpeiv). The aim of logical exercise is to become better prepared for more diffi- cult problems, and the impatient pupils are warned that the way may be long or short according to the subject (286 E : Aoyor, dv Te Taupnnns rexOels Tov axovoavTa eipeTiKa@TEpoY amepyaCntat, Tovroy omrovoaley Kal TO pnKer wndev ayavaxreiv, av 7’ ad Bpaxvtepos, wcavrws). It seems that the form of the Sophist had been criticised as too lengthy, and as winding around the subject with which it deals. Plato answers here that such critics ought to have shown how the same results could have been reached by a shorter way, and whether the shorter way would have been equally useful for the purpose of developing dialectical power (287 4). This is clearly a polemic reference, and if in a contemporary writing we could discover some censure of the Sophist of Plato, the relation between this writing and the Politicus would be established beyond every doubt. Unluckily, no NEW THEORY OF SCIENCE: THE POLITICUS 443 such writing seems to be known—at least, Teichmiller and Dimmler, who believe themselves to have found so many other ‘ literary feuds,’ have made no use of this in- teresting passage. The ideal of logical training occupies Plato’s mind with increasing fascination, and he insists on its im- portance at every step. He quotes manifold instances of the way in which higher aims are furthered by elementary exercise. One who learns reading, and is asked of what letters a word consists, does not aim only at answering that particular question, but at becoming more proficient in grammar (285 c). Thus also the investigation of the nature of the statesman is only a lesson in dialectic (285 p, cf. Soph. 227 B). For the aim of life is to become better and wiser by means of science and justice (293 pb). True and well-founded opinions on these things are divine, and to be seen only in divine souls (809 c: tHv Tév Kadov kat Sixaiwy trépt cai ayaa Kai Tov TovTois évavtiov svTws ovaav adnOn Sokav pera BeBawwoews, ororav ev rpuyais yyiyvntar, Oeiav dnt dv Sacpovim yiyverOat yéver). Their greatest happiness is knowledge (272 B), and Plato invents here a new word never used before him to signify the trea- sury of human knowledge (272 c: cuvayuppos dpovncews) as an ideal totality of individual endeavours, eternally in- creasing and transmitted from generation to generation. Such a conception he could not have had when he wrote the Socratic dialogues, and it 1s really difficult to under- stand how so many distinguished Platonists could believe in an early date of the Polsticus. The use of dvvapis alone in this passage is a sufficient sign that the Politicus is written after the Republic, and many other signs are here available for the determination of this relation between the two dialogues. Here even the notion of desire 1s subtilised to such a height that it is applied to logical training (272 D: ras éruBupias mept te ériotnpov Kal THs Tay Adywy xpelas). The chief instrument of this training is the same here as at an earlier stage, the power Tlustra- tion from learning to read. Know- ledge of truth and justice is the ultimate aim. Compre- hensive- ness of knowledge as now conceived. Maturity of this concep- tion a mark of lateness. Simi- larity and difference again. Scientific construc- tion. Reality only attained by reason. True con- ceptions indepen- dent of senge and of language. Platonic absolu- tism, 444 ORIGIN AND GROWTH OF PLATO'S LOGIC to ask and answer questions (286 A: dei pererav Aoyor éxdaorou Suvarov elvat Sodvat nat SeEacba, cf. Phaedo 76 B, Crat. 426 a, &c.). The notion of similarity and difference retains the im- portance which it acquired in the Parmenides. The true dialectician is asked first to find out all the differences in a group of things, and then to discern all common pecu- liarities which unite them into various logical units (285 B). Each science is built up by a skilful selection of appropriate elements, and the right union of similar particulars into one, while useless observations and notions are rejected (308 C: maca émurtnun Tavtayod Ta per woyOnpa sis Sivapwy adroBarnra, Ta 8 arurndeva Kal ypnora EdaBev, 2x tovtwv o2 Kal opolwy xal avopolwv dvtwy, rdvta eis tv abta Evvayovaa, piav twa Sivaymw nai ideav Snusovpye, cf. Crat. 438 8). For this an exact definition of each notion is required, based on reasoning not on sense perception (277 c: ypadhs 52 xal cupracns yetpoupyias rset nal roy Snrodv qav Ssov padXrov mpére tots Suvauévos ErecOat). No figure or drawing can correspond to the true substance of things, which is conceived only by pure reason (286 A: tots 8’ ab peyiorots ovot Kat Timiwrdtos ouK zoTw eldwrov ovdev mrpos Tous avOpwrous sipyacpuevon évapyas, od SevyOévtos Thy Tod TuvOavopevov ruyny o BovdAduevos atroTAypaoat, mpos TV aidOyjcewy TivAa TpOcAapLOTTMY, ikavas TANPwCEL). This relation of truth to reason is here insisted upon (286 A: Ta yap dowpata, KddAdora Sévta Kal péeyiota, OY povoy, GAdrw b8 ovdevl cadas Selxvutar, cf. Phaedo 65 D). The ideas must be understood independently of the use of language and without attaching any exceptional import- ance to words (261 E: «av d:advaratns ro pH orovddfew deri Trois dvdpact, tovoiwTEpos sis 7 yhpas avahavncet dpovncews, cf. Crat. 439 A). The greatest differences of opinion, which divide men into opposite camps, refer to moral convictions, and the philosopher appears here possessed with that Platonic absolutism which in a later age produced the Christian NEW THEORY OF SCIENCE: THE POLITICUS 445 Inquisition. Such an intolerance is a distinct peculiarity of Plato’s later years, and distinguishes the Laws from the Republic, forming at the same time a link between the Politicus and the Laws. That ethical questions divide men more than purely theoretical discussion was assumed already in the Socratic dialogues, but here it is asserted with much greater strength : Euthyphro 7 ©: epi rivos 8 bn SuevexOévres kai emi riva Kpioww 9 6 > se 3 é ov Suvdpevoe adicéoOat éxOpoi ye dv aGAAnAows eipev Kai dpytfoipeda; . oxorer ef rade e€grt TO TE Sixatov nat rd adixov Kai KaAddv Kat Polit. 808 &: 7» Baowlun ... rovs pn Suvapevous Korvwveiv FOovs av8peiov cat cadpovos daa re dAda > oN ’ \ ’ , 7. \. 2? €oTt Telvovra mpos aperny, adr’ eis abeornra xai UBpey xai adixlay ord xaxns Bia dicews arwbovpevors, ? 4 a > A A 4 alaxpov kat ayaOov kai Kaxov. Crito 49D: oida .. . dre dXLyos 4 a ay ”~ \ , tigi ravra xat Soxet nai dd€ee - Oavdras re éxBdddec Kal duyais kat rais peyiorais xoAd{ovea ari- pias. Cf. Legg. 909 a: where those who disagree with the law- giver on religious matters are condemned to death. z ols ? 9 . , ff U ’ oy ovrw Se8oxrut xal ols pn, Tovrots ovx eatt Kown BovAn, adAd avdyxn rovrous dAAnAwy Karadpoveiy, dpav- ras Ta GAAnAwy Bovdevpara. Thus we see that Plato admitted the impossibility of proof in moral questions, otherwise he had no reason to propose the penalty of death for moral dissenters, and specially for atheism. He recognised here a power of individual nature, resisting the charm even of the highest philosophical rhetoric, which produces conviction only in purely theoretical matters of science, not in practical tendencies of life. The unity of universal science, already affirmed in the Sophist (257 c), 1s here taken for granted, and a division of the whole is attempted into theoretical and practical, or pure and applied science (258 E: tavry Toivuy cuprracas driatipas Siaipa, tTHv wey Tpaxtinyy mpoceimayv, thy 82 pOvov yooriny—torw cor rad’ os pias émiothns rhs 8Ans ef5n Sto). Pure science is again divided into critical and epitactic, of which the former teaches what is, and the latter what ought to be (260 B: «plo 58 wal amirdée Staddperov GAAHAOW ToUTW TW ydvee; TUUTAENS THS yrwoTLKS TO piv dmiraxrixoy pépos, TO 52 Kpirucdy ...). To the the source of later Christian Inquisi- tion. Division of uni- versal science into pure and applied, critical and epitactio. Even casual observa- tions here take a scientific form. Produc- tions classified. Rules of classifica- tion. Dicho- tomy preferred, but natural unites always to be pre- served. 446 ORIGIN AND GROWTH OF PLATO'S LOGIC epitactic sciences belong not only ethics and politics but also all practical pursuits which require helpers to execute the leader’s instructions, as, for instance, the art of archi- tecture. The further subdivisions are only playful, and cannot be taken seriously as a permanent contribution to the classification of sciences.” Also other samples of classification given at some length in the Politicus, as, for instance, the classification of living beings (262-267), have no permanent value, and offer only an opportunity for the application of logical rules. For instance, Plato criticises the division of men into Hellenes and Barbarians (262 D), and compares it with a division of all numbers into ten thousand and other numbers than ten thousand. For the purpose of showing his independence of every prejudice he finds a similarity between swine and men, so much that both kinds of animals are distinguished only by the number of their feet, man being a gregarious tame hornless animal walking on earth by the power of two divided feet, while the swine use twice that number, thus appearing to have even an advantage over men (266 c). Such a fanciful definition is meant as a protest against the undeserved exaltation of vulgar mankind over other animals. Also the subdivision of productions and pos- sessions (279 D) 1s mainly an example by which the rules of classification are illustrated. These rules were then first expressed by Plato, and appeared to his mind as very important logical laws. The subdivisions ought to be nearly equal to each other, and form natural units, not artificial parts (2624 B: wn outxpov poprov 8v mpos peyada Kat TONAA dhaipapev, wnde eldous ywpis’ GANA TO pépos Aya eldos dydrw . . . Sid péowy 88 dogadéorepop iévat téuvovtas). Ideas, as here conceived, are to be found by classification of notions, or are ideal notions in the same meaning as 29 The various classifications of the Politicus have been specially repre- sented by Lukas (Methode der Einthétlung), and also recently by C. Ritter (Platos Politicus: Beitrdge zu seer Erkidrung, Programm des Gymna- siums zu Ellwangen 1896). NEW THEORY OF SCIENCE: THE POLITICUS 447 the ideas were for Leibniz or Kant. Any attentive reader of the dialectical dialogues will at once observe that in this and similar passages eléos and idea are identical in meaning, and that they cannot mean transcendental beings, but natural kinds or ideas in the same sense as the term is used in modern philosophy, that is, perfect and universal notions of the human or any higher mind. A notion as first formed might be imperfect and subjective. As soon as by dialectical thought it attains perfection and objectivity, it deserves the name of an idea. Objectivity is not separate existence outside any mind, but uniform existence in all possible souls. The separate existence of ideas outside any mind is a poetical absurdity which could subsist only for a very limited time in the imagination of a thinker like Plato, and which has never been expressly affirmed in clear words by him—because the poetical metaphors of the Phaedrus, Republic, Phaedo and Symposium cannot be taken as literal expressions of abstract truth. They only supply an indication that Plato, when he first discovered the objectivity of notions, hesitated how to explain this objectivity and felt some inclination to a worship of ideas in anideal world, whence they could influence our 1m- perfect minds. This conception may have been developed by his pupils to such extremes that he undertook to demonstrate its absurdity in the Parmenides. Since that time he continues to use the terms eléos and (déa, but no longer suggests the separate existence of abstractions, as this would contradict the increasing importance attached to the priority of soul in the universe. The ideas can only exist in a soul, as has been clearly said in the Sophist: they are notions, but not every notion is an idea. The idea is a notion of a perfect soul, free from error, and we must carefully distinguish among our own notions the ideas from other imperfect notions. This is the only consistent interpretation of later Platonic logic, and might be confirmed by a long enumeration of Objec- tivity at first mis- under- stood. The ex- aggeration of Plato’s followers may have unde- ceived him. The idea is a notion of a per- fect soul. C. Ritter’s examina- tion of passages in the Politicus showsthat none of them imply the current ‘ doctrine of ideas.’ This throws the burden of proof on those who maintain the older view. Logical division compared to the breaking up of a 448 ORIGIN AND GROWTH OF PLATO'S LOGIC the passages in the six latest dialogues where eldos, idéa, yévos, pépos, popiov, Tuna, pvots, Sivapss occur. Such a full enumeration, however, with a sufficient interpreta- tion of each passage, exceeds the limits of the present work. But C. Ritter in his very interesting programme on the Poltticus *® enumerates the corresponding pas- sages of this dialogue and arrives at the conclusion that not one of these passages confirms the ‘herkémmliche, durch Aristoteles eingefuhrte Auffassung der platonischen Idee.’ This is also Campbell’s opinion in his Intro- ductions to the Sophist and Poltticus. Here it will be sufficient to re-assert as the result of a careful reading of the six last works of Plato the conviction that the philo- sopher at this stage of his thought no longer admitted the conception of ideas as existing outside every soul. If anybody sustains the opposite view, he must always recur to the very improbable hypothesis that the second part of the Parmenides is a refutation of the objections raised in the first part, and to the serious blunder of interpreting wavredos Sv in the Sophist (249 a) as ideas,. possessing each of them soul, life, movement, and reason. Even this absurdity is insufficient to prove the separate existence of ideas in later Platonism: we challenge our readers and critics to point out in works written after the Parmenides a single passage supporting the assump- tion that ideas exist outside every soul, or contradicting our view that ideas are perfect notions of a perfect Being, natural kinds of particular things in agreement with the thoughts and aims of their Creator. All the rules given for the finding of ideas by classifi- cation become useless if we understand ‘ideas’ to mean anything else than this. One of these rules compares the division of an idea with the cutting into parts of a sacrificial animal, and recommends dichotomy as the best way of division, leaving open the recourse to a partition in three or more parts only when for some reason dichotomy is impossible (2870: xara pédn tolvuy adtas olov lepsiov NEW THEORY OF SCIENCE: THE POLITICUS 449 StarpapeOa, arevdy Siya dduvaroduev. Set yap ets tov ayyv- rata 6 tt padora Téuvewy aptOuov asi). In these divisions the parts should be always natural kinds (2634: yévos dua cai pépos evpioxev .. . 262E: padaArov nat’ edn Kat diva . 285A: Kar’ sidn cuvetBicOar oxoretiy S:atpov- pévous .. . 285B: Svadopas omdcairep ev eideor xeivras .... 262B: 70 pépos aya sidos dyérw, &c.). Constantly eldos and iééa are used in the same meaning as yévos, coinciding with the conception of parts of a class of objects. This process of classification enables us to find the principles or elements of Being, unknown to those who cannot recognise the essential identity of things apparently different, but really belonging to the same class (278 c D). In order to attain a greater dialectical power, it is necessary to recur to exercise on familiar examples and to observe the analogies between such examples and the highest metaphysical problems (277 D: yaXerov py trapa- Seiypwace ypwpevov ixavds évdeixvucbar tT. Tov pefover. xevOuvever yap Nuay Exaotos oloy dvap eidws Amavra TavT’ av’ wddw @orep trap dyvoeiv). Sometimes we believe ourselves to have seen things distinctly in dreams, while we are unable to describe them after we are awake. Thus untrained people often are unable to distinguish notions which on other occasions they had distinguished. We are best led to the knowledge of truth by the skilful selection of convenient examples which bring us gradually nearer to the aim of our inquiry. Plato applies this rule immediately by giving an example of the use of example (277D: mapabdelyywatos kal ro Trapddevryua adto Sebdénxev . . . 2788). Children when they learn reading recognise a letter more easily in short than in long and difficult syllables. They will learn with the greatest facility if they are first shown short and easy syllables, as examples of the use of letters which recur in long and difficult syllables. Then they will without effort develope their faculty of recognising the same letter wherever they see GG sacrificial victim. Natural kinds in- differently spoken of as yévn or e%3y. Deeper and fuller concep- tion of know- ledge. Use of the argument from example. Example of example. Oom- plexity of nature and of Life compared with the variety of words and phrases. The weaver compared with the politician. Example of the physician. Oppor- tunism as in the Laws. Tlus- tration from an 450 ORIGIN AND GROWTH OF PLATO'S LOGIC it, be it in a short or long syllable, as they will notice that each letter remains identical in all the combinations it enters into with other letters, so that it is sufficient to know well the small number of existing letters in order to be able to read the most difficult words and phrases in their innumerable combinations (277 E-278c). Though Plato had already in his earlier works made use of examples and recommended them (Phaedr. 262¢: vWiurds mws éyouev, OVK FyovTEes ikava apadeiyuata, cf. Soph. 218 D: rept twos TaY Pavrwy peTtlovTes TrEtpaOapmeV Trapd- Sevypa avro Oécar tov peifovos), he had never given such direct attention to the theory of analogyas he hasdonehere. The use of an example has the purpose of inducing a pupil to recognise an idea in a less familiar application, by comparing it with a familiar instance of the same idea (278 ©: Tobro . . ixavas ouverndapev, Ore wapadetypatos y’ sort rote yéveots, oToTayv by tavTov dy érépw Sieotrac- ptvo Sokalopnevov dpOas Kal aovvayOév aepi éxdtepov ws cuvappw piav arnOy 8dEav arroredyj). It is very characteristic that this practice 1s here represented as leading in the first place to true opinion, not to absolute knowledge, which cannot rest on mere analogy. The use of examples in the Polsticus is very frequent: thus for instance the long explanation of the art of weaving is an example which is given only for the purpose of ex- plaining the political art (287 B). The politician is also compared to a physician who prescribes drugs according to the state of the patient and changes them when he finds it convenient. Thus also the true poli- tician will change the laws if new experience requires it (295 o-296 a. Cf. Legg. 769 D). The political opportunism here proposed agrees well with the Laws, and is very different from the absolutism of the Republic, and for this reason alone it would be impossible to admit that the Republic could have been written between the Poltticus and Laws, as Zeller sup- poses. A very remarkable example is given to illustrate NEW THEORY OF SCIENCE: THE POLITICUS 451 the incompetence of the majority in political affairs. Supposing that the rules of medicine and navigation were entrusted not to physicians and seamen, but to a majority of citizens, the consequences of this arrangement would certainly be disastrous for all. Not less disastrous are the consequences of the political power of a blind majority (298-299, 300 E: @poroynpévov nuiv xetrar pundév mdrnO0s pnd jvrivodv Suvarov elvat AaBeiv téyvny). In the Politecws even more than in the Sophist, the idea of method acquires a prevailing power over the mind of Plato. At every step reflections on thought arise, thus giving to everything a logical aspect, and showing a tendency to an impartial consideration of all the condi- tions of each branch of knowledge. For instance we find here a digression on the difference between absolute and relative measure (288 E: Surtas.. ovoias xal Kpioes Tov meyadou Kat TOU apixpod Oeréov . . . THv pev pos adAAnAa...17 8 avd mpos To wétpiov). We judge about quan- tities by comparing them either with each other or with an absolute standard of what ought to be, in thought or action (283 E). The absolute standard named ro pétptov is the principle of every art and also of politics and morality (284 a). This absolute standard (284 £: apos TO petplov Kal To MpéTov Kal Tov Katpov Kat TO Séov Kal avd omdoa eis TO pecov um@KxicOn tav eéoxdTrwr) 18 equally distant from two extremes and is here indicated as an important new discovery (284 D: Sejce. rod vov Asx Odvros Tpos THY Tept avTo TaxptBes aTrodekw ... Hyntéov opoiws tas téyvas tmdoas elyat cai peitov te aya xal #datrov petpeicOar pt) mpds GAANAG povoy adAA Kal mpos THY TOD Hetpiou yévectv). Only those who are not accustomed to dialectical distinction are unable to see the difference between absolute and relative measure (285 A). This theory, later applied by Aristotle in his Ethics, 1s here re- peated several times with great insistence, and is evidently felt to be expressed for the first time (285 Cc: $uAarrapev 5s povov, drt Svo rydvn aEevpnras rhs perpnruchs). It corre- Ga2 imagined rule of the majority in medi- cine and naviga- tion. Increasing preva- lence of the idea of method. Measure, absolute and rela- tive: rd pérptoy. To be compared with Aris- totle’s meodrns. Causes and con- ditions. alria and tuvalria. Final and efficient causes. 452 ORIGIN AND GROWTH OF PLATO'S LOGIC sponds very well to the new conception of ideas and could not easily apply to the primitive transcendental ideas, which were out of relation with particular extremes. Another logical distinction, which had been already prepared in the Phaedo, is here developed as it were casually in the progress of the inquiry : Phaedo 99 A: airw ra rovadra (the physical conditions) xadeiv Aiay dromov - ef Se tis A€you Ore a@vev Tov ra rotavra exe, Kal 3 ”~ Q ” x. oa ” Gora kal vevpa xai doa GAda eyo, ? Ts 9 “ \ ‘ ovK dy olds f° Av rrotety Ta Sofavra pot, adnbn dv A€you- ws pévroe bia Tavta mow a mow Kal Ta’ra v@ mpatrwv, aN’ ov TH Tov BedTioroU @ 7 ‘4 .' x ¢@ a aipecet, moAAn Kat paxpa pabvyia dv ein rod Adyov. B: Td yap pn dted€- oa ofdv rt’ elvat Gre dAdo pev ri €ort TO aitioy T@ Sytt, GAAo Se exeivo dvev o¥ TO alriov ovK ay :.*” ” , ’ mor’ ein airtoy: 6 dn por haivovra Wnrahavres of modAol domep év Polit, 281 c: the production of the weaver’s tools is designated cuvaria of the art of weaving. This term has been used only once before, according to Ast, in Gorg. 519 B: ovx airiwy dvrav Tov kax@v GAN’ lows cuvariov. This use of the word is similar to that in Aeschylos and Isocrates. A more technical use of the same term is found Polit. 281 p: dvo réxvas ovcasmept mdvra Ta Spapeva, .. THY pev THs yeveoews ovgay Evvairiov, ray 8 avriy airiay . . doa. . . 6pyava mapackeuvd(ovaw .« . rauras peév ~vvacrious, ras be g > A “a 3 oxdre, dAdAorpi@ Gyépart mporxpe- al’ro To mpaypa amepyafopevas ” > A > 7s pevol, @s airwoy avTo mpoga- airias... yopeverv. cf. 287 B: §uvariwy kai ray airiwy, 287 D: doat yap opexpov f péya re Onpuovpyovos Kara mow cpyavoy, Oeréov dmdcas ravras ws odcas ouvarious. dvev yaprouTa@y our dy more yévorro médXts ovde roX«rixn, rourwy 8 at Bamwtkans Epyov réxyns ovdev rrov Onooper. Cf. Tim. 46 pv: 80€dlerat ims raév mreicroy ov Evyairia GAN’ alria elvat Trav sravrev (namely material causes as compared with final causes). In the same meaning 76D: rd... d€éppa, rois pev Evvatrioss rovrats Snpuovpynbév, ty 8€ airwrdry Siavoig trav emeta é€oopévov evexa elpyacpeévov. We see that the distinction between final and efficient cause, which remained the same from the Phaedo to the Timaeus, acquired its proper terminology only in the Politicus. What in the Phaedo is called ‘ éxeivo dvev ob 70 alruoy ovx dv Tor’ en atrcov’ becomes in the Politicus Evvairioy and is again designated by this term in the Timaeus. The special application of the term in the NEW THEORY OF SCIENCE: THE POLITICUS 493 Politicus changes nothing in its general meaning, and the word £vvairiov 1s a peculiarity of later style, limited to Politicus, Timaeus, Laws in its technical use, while it is used only once in an earlier dialogue in the earlier meaning, taken from tragic poetry. It will be difficult for the believers in the Megaric period to explain why Plato should use in the Phaedo a lengthy circumlocution if he had already introduced a short and convenient term with the same meaning. For us it is quite clear that Euvairiov as a metaphysical term was not yet in Platonic use when he wrote the Phaedo, as can be seen by the comparison of the above passages. Another interesting reference to earlier theories is here found in the familiar mention of the soul as the first principle of movement, which could scarcely be understood in the form in which it occurs here, if the demonstration of the Phaedrus were not presupposed (269 E: avro 52 gauto orpéhew det oyeddv ovdevt Suvaroy mANV TO TOY KWwoupévwoy ad TdvTwV nryounévm). In connection with the cyclic revolutions of the heaven Plato speaks here again of immense periods of time, which he never had mentioned nor imagined before the Republic. The universe 1s supposed to be subject to periodical revolutions which last millions of years ™ (270 A: avdmadw mopevecOar moddas Tepiodwy fupidoas). If we look at the logical character of the Polzticus and at the biting humour displayed in this dialogue as in few other works of Plato, it appears incredible that critics were found who doubted the authenticity of this dialogue. What Socher (1820) and Suckow (1855) said in favour of such doubts has been repeatedly refuted by Grote, Campbell and Jowett. But Schaarschmidt’s plea for the spuriousness of the Politicus seems not yet to 200 The meaning of weplodos is not quite certain; Campbell translates ‘days,’ but in view of the similar passages of Theaetetus and Phaedrus and of the astronomical studies which appear to have occupied Plato in his later years, it is quite as probable that he meant years, each year being the smallest period in which the heaven returns to the same relative position. Fresh proof of the priority of the Phaedo. Demon- stration of the Phaedrus also pre- supposed. Tlimit- able periods of time. Biting humour of the Polsticus. Objections of Schaar- schmidt to the authen- ticity of this dialogue easily refuted. 454 ORIGIN AND GROWTIL OF PLATO'S LOGIC have been specially considered, though nearly all com- petent authorities recognise the authenticity of this dialogue as established beyond every doubt. It may not be superfluous to consider these arguments, as Schaar- schmidt, living still, has not found it necessary to revoke them in the course of thirty years, and as he has followers among quite recent historians of philosophy.*' Schaar- schmidt thinks that such tedious divisions of notions as are found in the Politicus are unworthy of Plato. If we remember that divisions of notions have been recom- mended in the Phaedrus, and very much used in the Republic, there is no reason whatever to doubt that Plato at a later period of his literary activity gave a special attention to this logical exercise. What Schaarschmidt says about the use of example and analogy as contrary to Plato’s custom is equally contradicted by the Phaedrus (262cCD), where examples are as strongly recommended as in the Politicus. The myth of the Poltticus, like the myth of the Phaedrus, is used to help the progress of the philosophical argument, and Schaarschmidt has no right on this account to doubt the authenticity of the Politicus if he admits, as he does, the authenticity of the Phaedrus. The difference between the myth in the Politicus (271 D-274) and a short mention of the same legend in the Laws (713cp#8) has further excited Schaarschmidt’s suspicion. But Plato never attempted a painful identity of myths, and anybody can see how freely his imagination worked in the different versions of the eschatological myths. Schaarschmidt contradicts himself, because he holds #1 W. Windelband, Geschichte der alten Philosophie, 2° Aufl. Miinchen 1894, p. 114, says: ‘es ist nicht wahrscheinlich, dass der Philosoph neben der Republik denselben Gegenstand in einem andern Werke behandelt heben sollte, zumal da das letztere in wichtigen Punkten erheblich andere Lehren aufstellt.’ In France Huit (‘Etudes sur le politique attribué 4 Platon,’ in Séances et travaux de l’ Académie des sciences morales et poli- tiques, vol. 128, p. 569; vol. 129, p. 169, Paris 1887) popularised Schaar- schmidt’s views. On the relation between Rep. and Polit. see Nusser’s article, in Philologus for 1894, vol. liii. pp. 18-37. NEW THEORY OF SCIENCE: THE POLITICUS 450 every difference between two dialogues to be a reason for suspicion, while on the other hand every similarity also appears to him as an indication of the work of an imitator. Thus he wonders why in the Poltticus an ideal ruler is placed above the laws. He forgets that here, as in the Laws, the ideal is already admitted to be almost impossible to realise, and the idea of a second best state based on fixed laws, here introduced (297 £), is later developed in the Laws. On the other side, the view of a state without fixed laws is already prepared in the Republic (425D: ov« a&ov avdpact KaXots Karyabots érirarrev), as also the comparison of politics and medicine (Rep. 426 A, cf. Polit. 298 a—300 D). The usefulness of laws is fully recognised in the Polvticus in the case when a perfect ruler cannot be found. Ordinary governments do best to keep the law (801 4). The same doctrine occurs in the Laws, only there it is recognised as impossible that an ideal ruler should be born on this imperfect earth, so that the laws acquire an increased importance, though the notion of an ideal state is not altogether abandoned, except for transient practical reasons (Legg. 739D: % pév tocavTn mods (as proposed in the Republic), eire wou Oeot 4} traides Oeav avtiv otKodce WAEtous évos, oUTw Sialavres evdppavouevor Katotxovar, cf. 746 4B). In the Politicus as in the Laws (874E: vouous avOpwrros avayxaiov t1Oec0at xa SRv Kata vopovs, h undev Siadipey THY TavTn aypiwrdtwy Onpiwv) the fixed rules become necessary only in consequence of human ignorance and imperfection. This conviction led Plato equally in the Politicus as in the Laws (6848C) to recommend coercion in order to maintain the fixed legislation. In political theories it becomes especially evident that the Poltticus is intermediate between Republic and Laws, so that there is no reason to raise any suspicion from that standpoint against the authenticity of our dialogue. Schaarschmidt wonders why the ideal ruler in the Polt- ticus 1s not a philosopher as in the Republic, and thinks that this ideal ruler has no other aim than to satisfy the Relation to the Laws. Supposed silence of Aris- totle. Denied by Ueberweg, and shown to be unten- able by Bonitz. 456 ORIGIN AND GROWTH OF PLATO'S LOGIC personal needs and aspirations of the governed. This by no means agrees with what we really read in the Politicus. Politics is here counted among the theoretical sciences (259 D) opposed to the practical arts, and the politician’s aim is to produce divine and true opinions about justice in his subjects (3809 c, cf.293D). It is very natural that only Opinions are to be expected in the blind majority of men. Knowledge is also in the Republic a privilege of the rulers. Schaarschmidt’s inferences from the silence of Aris- totle about the differences between the Polsticus and Laws are sufficiently refuted if we consider the acci- dental nature of all allusions to Platonic dialogues in the works of Aristotle. There was no necessity for him expressly to quote the Politicus, and we must not apply our standard of literary erudition to Aristotle. His works have come to us in a state which does not guarantee that we possess all the quotations he might have made from Plato’s works. And the quotations preserved could in most cases be omitted without any prejudice to the argument of the passages where they occur. The Poltticus is not, as its title might suggest, a political treatise, and therefore there was no opportunity to quote it in Aristotle’s Politics, where the Republic and Laws are dealt with. Here we find more logical than political theories, and the definition of the states- man or politician is only a pretext for many digressions on the method of scientific investigation generally, as can be seen from the above exposition. Ueberweg has suffi- ciently proved that the Sophist and Politicus were known to Aristotle, and although he afterwards believed that some pupil of Plato might have written these dialogues, the references he collected show clearly that Aristotle knew them. It is difficult to admit that Aristotle would have named a pupil of Plato an ‘earlier writer.’ This, as the name is not specified, refers to Plato with greater probability than to anybody else. Bonitz quotes thirteen references to the Poltticus of Plato in the works of NEW THEORY OF SCIENCE: THE POLITICUS 457 Aristotle (Index Aristotelicus, p. 598). The feeblest of all Schaarschmidt’s arguments against the authenticity of the Politicus is based on a misconception of Plato’s style. He did not know stylistic peculiarities as they are known now, and was therefore entirely unaware of the fact that the Politicus is as near in style to the Laws as the Timaeus, and this despite the great difference of contents. After Schaarschmidt no really new argument against the authenticity of the Politicus has been advanced, for what Huit says on the subject demonstrates only the strange ignorance of this author. He is, for instance, astonished that the title is not a proper name, as if he had never heard of the Banquet, Republic, Laws, which he still holds to be authentic. He complains of the absence of well-characterised persons, which is common to the Poltticus with all later works. He objects to the person of the younger Socrates as unplatonic, and he does not notice that the individual characterisation of all persons in later dialogues is equally deficient. He wonders why the Politicus has been so little quoted by later authors, and asserts that only Proclus, Plotinus, Plutarch, Theo- doretus, and Simplicius quoted it, while Fischer in his edition (1774) without attempting completeness of enu- meration gives a list of a dozen classical authors who had read this dialogue. Such tests are generally of little value, because most of these quotations are accidental. But it is quite unjustifiable to ask for better authorities than Proclus and Plotinus when corroborating Aristotle as to the authenticity of a Platonic dialogue. Huit also professes indignation over the fact that in the Polsticus the Sophist is quoted, and he seems to be unaware that in the Timaeus and Laws the Republic is clearly referred to, and in the Critias the Tumaeus. The only argument of Huit which might claim some importance is based on a misinterpretation of texts. He thinks that Plato in this dialogue does not distinguish d6£a from érictnun. If this were true, we should have reason Huit’s arguments still feebler. Import- ance of the Philebus. Schaar- schmidt’s doubts refuted by Huit. 458 ORIGIN AND GROWTH OF PLATO'S LOGIC to consider it seriously, because the above distinction is fundamental in Platonic philosophy. But really, as has been shown above, Plato speaks of 50a as mere opinions, not as knowledge, as it cannot be expected from all common citizens that they should rise to the level of knowledge, and the ideal ruler must be satisfied if he is able to produce in their minds true opinions. Ail these arguments of Schaarschmidt and Huit prove nothing, and the authenticity of the Politicus is established beyond reasonable doubt by the similarity of its style to the latest works of Plato. Until somebody can show in a work written by another author two hundred and forty stylistic peculiarities recurring in the Laws, we must accept the Politicus as authentic. It is a work of rare literary and logical excellence, and could not easily be written by anybody else than the author of the Phaedrus. As to the date of the Poltticus, it is certain that this dialogue must have been written after the Sophist. This, taken together with the order of the preceding works, gives to the Politicus a place among the productions of Plato’s old age, in so far as only the Timaeus, Critvas, and Laws are distinctly later. It remains difficult to decide whether the Philebus followed or preceded the Politicus. Hirzel ** has already clearly demonstrated that the Polv- ticus 18 very nearly related to the Laws. Ill. The Philebus. This dialogue is one of the most important writings not only of Plato but of ancient philosophy in general. Yet it has not escaped unjustifiable suspicions as to its authenticity. Schaarschmidt’s attempts in this respect have been already refuted by Tocco, and even his faithful follower Huit feels obliged to dissent in this point from his master. It is delightful to read this refutation of Schaarschmidt by Huit (vol. ii. pp. 171-181), because %2 Hirzel, ‘Zu Platons Politicus,’ in vol. vii. p. 127 of Hermes for 187 4. NEW THEORY OF SCIENCE: THE PHILEBUS 459 nearly every word of it applies equally to the Sophist and Politicus, while Huit is very careful to produce all argu- ments of Schaarschmidt as his own when he triumphantly asserts %* the spuriousness of the Parmenides, Sophist, and Politicus. We have limited our previous discussion on authenticity to the Politicus, because the authenticity of the Politicus implies necessarily the authenticity of the Sophist and also of the Parmenides. Here it may not be superfluous to say something about a more recent attempt, undertaken by F. Horn,™ to strike the Philebus out of the list of Plato’s writings. Itis significant that Horn does not deny Aristotle’s testimony in favour of the Platonic origin of the Philebus. Hence he is obliged to recur to the strange supposition that Aristotle could be mistaken in such a question as the authenticity of a Platonic dialogue, whereby the whole of Ueberweg’s investigation on the authenticity of Platonic dialogues is brought into question. Not on this point only Horn seems to be unaware of the method which alone can lead to valid conclusions in such matters. He reasons continually thus: some arguments of the Philebus do not agree with enunciations on the same 263 A curious proof of the incomparable ingenuousness of that French author, whom, by a regrettable mistake, the Académte des sciences morales crowned, is given by the fact that in his whole argument on the spuriousness of the Parmenides, Sophist, and Politicus (pp. 269-311, vol. ii. of La vie et Vaewvre de Plaion) he quotes Schaarschmidt only once, and this in a note (p. 309) in which he disagrees with him as to the pretended stoic origin of the Sophist. This cautious silence about an author from whom nearly all arguments of the text are taken, and who, in the chapter on the Philebus, is often quoted with a humorous contempt, is an interesting sample of apparent erudition paired with real ignorance of the subject, dis- played for the competent reader at every step, despite all the numerous quotations. Thus Campbell is also quoted in irrelevant matters, and appears to the candid reader either as an authority for the spuriousness of the Sophist (vol. ii. pp. 282, 286), or even further from the truth, as a mere critic, populariser or supporter of the views of Dittenberger ! (p. 341). 464 F, Horn, Platonstudien, Wien 1898; see against this: Dr. Apelt, ‘Die neueste Athetese des Philebos,’ in Archiv fur Geschichte der Philo- sophie, vol. ix. pp. 1-28, Berlin 1895, and again the reply of Horn: ‘ Zur Philebosfrage’ in Archiv fur Geschichte der Philosophie, vol. ix. pp. 271-297, Berlin 1896. Objections of Horn, in spite of Aris- totle’s testimony. Plato’s later views not always consistent with earlier state- ments. Philebus, perhaps twenty years later than Republic. A master- piece of Plato's old age. Progress of thought. 460 ORIGIN AND GROWTH OF PLATO'S LOGIC subject in the Republic, therefore the Philebws cannot have been written by Plato. We have seen in the above exposition of Plato’s logical theories that even such a thinker as Plato could not be in every particular free from errors, which he corrected later. In the course of his long activity he changed his opinion on several important points, which become specially manifest to anybody who cares to compare the Republic with the Laws. Also the Philebus is according to stylistic observations very distant from the Republic—probably about twenty years later. This sufficiently accounts for some divergencies. Horn’s general view of the Philebus is extremely subjective. For him this interesting dialogue is ‘ein mit véllig unzuliang- lichen Mitteln unternommener und héchst schiilerhaft gerathener Versuch einer Vermuittlung zwischen den ethischen Hauptrichtungen der Zeit.’ Other scholars, as, for instance, G. Schneider, who devoted very special attention to the Philebus,™ are of an entirely different opinion and see in the Philebus a masterpiece of Plato’s old age. The differences between the Philebus and Republic are all of such a character that they are perfectly well explained by the length of time and the progress of thought from the earlier to the later dialogue. Such differences ought never to be esteemed as an argument against the authenticity of any work of Plato, because 7 G. Schneider, Die Platonische Metaphysik, auf Grund der in Phile- bus gegebenen Principien in ihren wesentlichsten Ziigen dargestellt, Leipzig 1884 ; also: ‘ Die Ideenlehre in Platos Philebus’ in Philosophische Monats- hefte, vol. x. p. 198, 1874: ‘Das Princip des Masses in der Platonischen Philosophie,’ Verhandlungen der 33 Philologenversammlung, Gera 1878 ; Das materiale Princip der Platonischen Metaphysik, Gera 1872. 26 The relation of the Philebus to the Republic has been specially investigated by F. Schmitt (Die Verschiedenheit der Ideenlehre in Platos Republik wnd Philebus, Giessen 1891) and Siebeck (‘ Platon als Kritiker aristotelischer Ansichten : II. Der Philebus,’ in Zeitschrift fiir Philosopme und philosophische Kritik, vol. 107, pp. 161-176, Leipzig 1896). They both agree as to the later date of Philebius; see also note 249. NEW THEORY OF SCIENCE: THE PHILEBUS 461 in this way we might find suspicious almost every- thing Plato has written. The method which Horn uses consists in exaggerating every difference up to the point at which it appears to be an irreconcilable contradiction ; on the other hand, if one dialogue agrees in some parti- cular with another, he takes it as a sign that the author imitated Plato; finally if some opinion measured by the standard of our own time appears wrong, he finds herein an opportunity for representing it as unworthy of Plato. Such reasonings are built on three wrong suppositions : 1, that Plato never erred ; 2, that he never recognised his errors nor changed his opinion ; 3, that he never repeated what he said in another work. Any reader of Plato can easily see in hundreds of instances that all these three suppositions are inadmissible, and, we may add, any philosopher will recognise them to be psychologically impossible. But they are the constant basis of nearly everything that has been said against the authenticity of the dialectical dialogues. The only argument of Horn which at all deserves our attention is advanced without evidence and rests on no quotation from the Philebus. Horn finds in the Philebus ‘Geringschatzung der Dialektik’ (Zur Philebosfrage, p. 292). The high esteem of dialectic is such a permanent Platonic peculiarity that any work in which dialectic is despised must excite serious doubts against its Platonic origin. But nothing of that sort occurs in the Philebus. It is strange and unjustifiable that Horn was not more explicit on that point, and that he did not quote the passages from which he has drawn his inference. It is evident that he misunderstands Plato and takes for irony what is either solemnity of tone or Platonic humour. It does not follow that Plato despised dialectic, when he required that the philosopher should also have other know- ledge. This is not even a difference between Philebus and Republic, because there also dialectic was only the crown of all sciences, and did not render them superfluous. Horn’s view is based on mistaken presump- tions. He strangely speaks of a disparage ment of dialectic in the Philebus. Union of practical with specula- tive know- ledge. But reason must be supreme. 462 ORIGIN AND GROWTH OF PLATO'S LOGIC There is a progress in the dialectical dialogues as com- pared with the Republic only in the increasing appreciation of concrete facts and details, which less attracted his attention in the period of self-existing ideas. Here we meet the same enthusiasm for the power of reason as in the Sophist and Politicus. He who has elected the life of a thinker 1s more divine than other men and re- mains, like a god, free from exuberant pleasures, as well as from the sorrow which usually follows such pleasures (33 AB: Td Tov Tod hpovelv EXopévp Blov ola ws todTov Tov Tpotrov ovdev atroxwrver Cv... Tov Too wh Yalpew pnde AuTetoOat . . . Kal lows ovdéy AtToTov ei TdyTwv Tav Biwv dori Oaotatos . . . oUKouv sixos ye ovTE yalpew Tovs Osos ovte To évavtiov). The satisfaction given by knowledge is the purest pleasure in human life, free from the pain which mostly accompanies physiological pleasures (52 AB,: »aOn- pdtev wAnpwOeiow ée@av votepoy atroBoral Sia tis AjOns yeyvortas . . . ywpis ANUTNs . . . ANON yiyverar sKxdoToTe). These pleasures of science are the privilege of a very small circle of men (52B: tas trav paOnpatwy jdovas . . . pytgov . ovdayas Tav ToA\AwY dvOpwreav GANA THY shddpa orXywv). Every manifestation of intellectual life is better than sensual gratification for all those who are able to partake of it (11B: 16 dpoveiv cal 76 voety Kal To peuvna Bat Kal ta TovTwy av Euyyevy, Sokay tre opOnv Kai adnOeis rox- Lomous, THS ye HodovyAs dpeivw Kal Adw yiyvecOar Evurracw, Scatrep avroy Suvata petadaBeiv). All sages are agreed that reason reigns on earth and in heaven (280: tavres Evpdwvotew ot codol, éavrovs dvtws cEuvivortes, ws vods éoTti Bactrevs huiv ovpavod Te Kai yns* Kai lows ed Asyovor). This rule of reason becomes manifest by the finality appearing in the magni- ficence of the universe (28 D: ta Evpravta cai rode 76 Kadovpevoy Srov ... vodv xal dpdovnoly twa Bavpaorny cuvtartovoay SvaxuBepvav .. . pdva nal ris Syews rod xoopou Kal HArAlov Kai cednvns Kal aorépwy Kxal mdons Tis mepipopas aftov). The ultimate goal of this finality is a NEW THEORY OF SCIENCE: THE PHILEBUS 463 self-sufficing aim, the Good (54c: To pv ov svexa 0 ivexad Tov yeyvopevoy dei ylyvorr’ av, av TH Tov dryaOov poipa éxeivo dove... cf. 60c). This is the union of beauty, measure, and truth (654). We see here a development of what had been said in the Republic on the idea of Good. There it was one and the highest idea, here the union of three ideas, one of which, the ideal measure (fuuperpla = perptorns 64 £), has been introduced only in the Politicus (uérptoy Polit. 283%, Phileb. 66 a corresponds to perpiotns which in that sense occurs besides Philebus 64, 65 B only in the Laws 701 £, 7365, while in Rep. 560 Dit has another meaning). Truth is the aim of each inquiry, and it must be found in agreement among investigators (14 B), not in their ambition to be each of them right against all others. This ambition is peculiar to youth, and is here described with incomparable humour and a certain benevolence which denotes an experienced teacher, accustomed to see many useless discussions among his pupils (15D: gor 76 TovovToy TAY AoYwy avTaY aGOdvarov Tt Kal dynpwv TAaBos bv neiv’ Oo S8 ITP@TOY AUTO yEvodpeEvos ExdoTOTE TOY véewV, HaOels as Twa copias sipnxas Onoavpov, Ud’ Hdovis évOovora TE Kai advrTa Kiel NOyoy aGopmevos ... Els ATropiav avTOV pev TPwToV kal padiota cataBddrwp, Sevrepov 8 del tov syopevov . . .). True wisdom consists in defining ideas and their relations, until we obtain a continuous system of notions from the highest ‘one’ down to the ‘many’ through measured degrees, subdividing each idea into the smallest number in order to give the detailed specification of each subdivision of the one (16D: Sey ody Huds rovTwy obtw dStaxexoopun- pdvov det wiav id¢av epi tmavtos éxaoTtote Ospévous Cnreiy’ eipnoew yap évodcar' dy oby peradaBapev, peta pia Svo, ei wws siot, sxotreiv, et S& un, Tpets } Twa dAXov GpiOuov, kal tay bv sxeivwv Exactov mddkw doavTos, péypirep ay ro Kat’ apxas Sv wn Ste Sv Kal worded Kat dmepa tor. povoy idy tis, adda Kal onda). This system of ideas is to be found in nature, as everything The final aim of Reason is the union of beauty, measure, truth, wherein dwells the Good. Humorous descrip- tion of juvenile logic. A more complex ideal of definition. Natural kinds must be numbered and co- ordinated. Inter- mediate kinds: ‘middle terms.’ Prepara- tion for the theory of syllo- gism. Difficulty of the true method. 464 ORIGIN AND GROWTH OF PLATO’S LOGIC has been arranged by reason, and our ideas are copies of the world’s finality, existing in the maker’s mind. There is no longer any trace of an existence of ideas apart from souls and from particular things. Hach ideais the result of the impression which the natural unity of a group of particulars produces in an observing soul. The faculty of thinking ideas is here a divine gift (16.c: dewy sis avOpétrovs Sécts) and human imperfection consists in errors as to the number of the subdivisions which connect the one with the infinite many (16D: ryv 82 rod arreipou iSéav mpos To TAHOos 4H) poe hépscy, mrplv av Tes Tov apLOpov avtod Tavra Katién Tov werakd Tov atreipov Te Kai TOU évos). Only gods herein attain perfection (16 E : of wiv ody Geol ovTws juiv tTapébocay oKoTety Kal pavOdvew Kal SiddoxKeuv d\AnAovs), while even the wisest among men are liable to pass too rapidly or too slowly from the one to the infinity of particulars, through the ignorance of convenient middle terms (174: of 82 viv rdv avOparrav codgol av pév, drs av TUYwOoL, TA TONG OAatrov Kai Bpadvrepov trovodat Tob dSéovros, pera S2@ To by drreipa evOus' ta 58 péaoa avtovs exdhevryer, ols Svaxeywpiorat To Te Stadextixds madw Kal Td éprotinas Huds rovetaOat pds GAAnAOUS TOvs AGyous). We see here for the first time the term pécoy used in its technical meaning as later accepted by Aristotle in his theory of syllogism. If we take into consideration that it would be entirely against Plato’s view of literary compo- sition to enumerate all possible figures of syllogism in a dialogue, as is done in Aristotle’s treatise, it becomes quite possible and even probable that Aristotle’s theory of syl- logism was more than prepared by Plato. This point must remain unsettled so long as we have no independent testimonies of contemporaries. At all events, we see in the Philebus the same striving as in the preceding dia- logues towards an universal system of sciences, and we are warned that the classification of ideas, being the most beautiful method and leading to all discoveries which have ever been made, is exceedingly difficult and full of per- NEW THEORY OF SCIENCE: THE PHILEBUS 469 plexities (168). Therefore a complete realisation of the proposed method, and the consequent reduction of all ideas to their highest aim, appears here as a tale heard in a sort of dream (20B: Adywv sroré tiwewy dda dkovoas Svap 4 wal dypnyopds viv dvvod epi te HSovis Kal dpovicews @s ovdérepov avrow éorl tayabov, GAN aGAXo TL TpiTov. This is perhaps a reference to Rep. 505 B). Careful distinction between truth and _ probability is recommended, and those arts and sciences in which certainty appears unattainable are deemed inferior (564: TONY pEplywevov Fev TO wn cadés, outxpov b& To BéBarov— is said of music, medicine, agriculture, strategy, naviga- tion, &c.). The highest perfection here as in the Republic is attributed to dialectic and mathematics (57 EB: ray Tod StaddyerOas Siva . . . 58A: wepl rd bv Kal to dvTws Kat TO Kata TavTOY del TeduKOS TaVTwWSs Hyorye olwat HyetoOas Evutravras, Scots vod Kal opixpdov mpoonprntat, paxp@ arnberrarnv elvar yvoow—this after the recognition of mathematical sciences). The priority of dialectic or metaphysics as compared with all other sciences is so insisted upon, that it 1s difficult to guess on what possible misinterpretation of texts Horn built his contention that dialectic is despised in the Philebus. Plato repeats clearly that only dialectical objects or eternal ideas lead us to absolute certainty (590: yp)... 10de Ssapapripacbar TO AOyP, ws 4 TEpt exeiva Ech’ tyiv Td Te BéBavov Kat Td KaOapov Kat To adnOas kal 5 81 Adyomev eidcxpwves, Trepi Td ae Kata Ta alta doavrws auixtérara kyovra, 7H éxelvov 6 Ts padmora tore Evyyevés). It is very important to observe that eternal ideas (ae) xara Ta avrd) are not now separate, self-existing, or independent existences (avTo xa6’ aid) as they were in earlier dialogues. They are simply eternal, or always the same, because the true thoughts of a perfect being are not liable to change, and ideas are nothing else than ideal notions. Natural science is represented as deficient in exactness, because it does not refer to eternal ideas, but to changing H H Dialectic is exalted, not despised. Ideas 41810 but not xwpiord, Imper- fection 466 ORIGIN AND GROWTH OF PLATO’S LOGIC attributed appearances (59 A) which are in time, not in eternity, and tophysical can never become an object of absolute knowledge (59 B: solence. § = rovTwy ody Tt cadas dv haipev TH axpiBectaty AdnOsia ylyve- oOat, dv pnre Foxe undev rretrote Kata TadTaA mi Be pyre eis Td viv mapov Eye; . . . ovd’ dpa vous ovdé Tis sriotnun wept avtd gore TO GdnOéoratoy tyovca). This view agrees perfectly well with what has been said on that subject in the Republic, and also with the theoretical views of the This view dialogue which deals chiefly with natural science, the, confirmed Tymaeus. It was Plato’s permanent conviction that scoliges the immense variety of the physical world did not admit Timaeus. of perfect knowledge. The distinction between theo- Science retical and practical or pure and applied science is also pureand here maintained (57 A-E), and illustrated by the example applied. of mathematical units, which are absolutely equal to each other, while for technical purposes units really unequal are counted as equivalent (56 DE: of péy yap mov povdadas avicovs xkatapOpotvras tay tmepl apiOpor, olovy orpatoreda S00 nal Bods S00 Kai dv0 ta opixpotata 7} Kal ta TayTwY péyvora* ot & obx adv mote avtois cuvaxoNovOyjceay, € 11) povada povados éxdorns Tay pupiwy pndepiay GAdAnV addAns ete Staddpoucdy tis Onjoet). This idea of unity in variety annie: haunts Plato’s mind here as in all the dialectical dia- tise: logues. He goes so far as to say that one who is not able to distinguish the quality and quantity of each kind and its opposite deserves no consideration whatever (19 B: alin . . . elt’ dot etre yn, Kal ordca doti Kal oroia . . « pi) Suvdwsvot xatd tavros éyos nal ouolov Kal ravrov Kal Tov dvavtiov (6nXadcat) . . . ovdsis eis ovdav ovdevos dy huav ovdétrora yévorro dftos). Qualita- Genus The difference of genus and species is illustrated aan through many examples, and the species shown to be different and sometimes opposed within one genus (12 E: yévec wey dors tray sy, ra 58 pdpn Tots pépecuy avTov Ta pap dvayri@tata Gros, Ta 58 Staghopétyta Fyovta puplay arov tuyydve). All the difficulties implied in the relation be- tween the idea and particulars are repeated as it seems NEW THEORY OF SCIENCE: THE PHILEBUS 467 with a clear reference to a similar exposition of these difficulties in the Parmenides (15 A: drav 5é tis gva avOpwrov sruyeipy tiOecOar . . . mwept rovTwY TaV évadwy Kal TOV TOLOVTWY 4 WoNAH apdicRnTHOLS yiyverar .. . mpatov pev el tivas Set rora’tas elvar povddas Urodap- Bavev adrnOas obcas* era Tas ad tavras, play sedorny oveoay dei Thy avTny Kat pynte yéverw ponte SdrsOpov mpocde- xouévnv, Suws elvac BeBaioratra wiav travrnv). Here is a very clear indication that a separate existence of ideas is deemed impossible (15 B: pera 88 robr’ év Tots yiyvopévois av Kat dweipos eire Stectacuévny Kai qodAa yeyovuiay Geréov, si’ SrAnvavrny aitis ywpis, 6 bn ravtwv advvato- Tarov dalvor’ av, ravrov Kal by dua év evi re wal trodXois ylyverOar). As in the Sophist, the theory of ideas is introduced objectively, and not directly supported by the leader of the conversation, at least nothing is decided about the difficulties referred to. Throughout the dia- logue the terms used for ideas have no other meaning than ideal notions, as is the case everywhere after the Parmenides. The nature of thought requires the union of notions into higher units, and this constitutes an eternal necessity of the human mind (15 D). The absolute unity of knowledge is not prevented by many differences and even partial oppositions between sciences (13 E: aoAXaé Te ai Evvdrracat ériotipar Sofovow elvat Kai avopovot tives avTa@v aGAdAnAats* ei 52 Kai évavTias 1y ylyvovrat Twes, dpa aktos Av einv tov SiarsyecOar viv, ei poByOels todro avro pndeplay avopovoy dainy arriornpny arrvotun yiyvec Oat ;). On the other side, we need not attempt a reconciliation of all contradictions (13 A: rovtT@ Te AOyp py wlaTEVE, TO TavTa Ta évavtiworata sy TroLouvTt). Sense perception is explained as a motion common to body and soul, whereby the theory presented in the Theaetetus is repeated and accepted (84 A: ro 8 éy éyl wad0e. thy wuyny cal TO odpa Koy yryvopevoy Kow7 Kal xiveioOar, tavrny § av thy Kivnow ovopalwy aic@now ov« a1ré tTpoTrou POsyyou’ dv). But the soul can become indif- HH 2 Abstract and concrete unity. Ideally but not really separable. The unity of know- ledge embraces sciences diverse and even opposed, The theory of sensation is further developed. Sensation, memory, remi- niscence. Compari- son and judgment. Thought indepen - dent of language. The human soul similar to the Divine. Judg- ments are in- scribed on the soul. 468 ORIGIN AND GROWTH OF PLATO’S LOGIC ferent to the action of the body, and then receives no sensations (83 E: éray () wWuyn) amabns ylynta tov CEOUOY TOY ToD gwHpaTos, avatcOnciay érovopacov). In that state the soul retains the faculty of repeating by its own power the qualitative changes it had undergone on former occasions through the action of the body’s move- ments, and thus reminiscence needs not the co-operation of the body (34 B: drav & peta Tod c@paros sracyé trol” 7) aux, Tad?’ avev To cwmaros avTn av tavTn 6 TL uadtoTa avarapBdvy, tote dvaptpvnoKxecOai rrov Asyouev). Sensa- tion and memory are the faculties on which opinions rest (38 B: é& pvnuns te nal aicOnoews Sofa hyiv Kat ro Svado- Edlew ayyepetv ylyvel’ éxdotore) ; Our judgments exist in our thought before we give them an expression in words (38 E, cf. Soph. 264 a). Plato insists on the faculty of thought as independent of words and of sense percep- tions. We are able to represent to ourselves all past perceptions at our will, and such images do not require the co-operation of the organs of sense (39 B). The origin of error is here, as in the Theaetetus, attributed to indis- tinct sensations (88 C: oAddKis iSdvte Tvl moppwOev wn wavy cadas Ta KaSopwpeva EvpBalvav Bovrgtec0ar xpivew gains dv tab@’ drep spa). The possibility of knowledge is founded on the funda- mental similarity between each individual soul and the world’s soul from which all individual souls are derived, and in which all our notions exist in far greater perfection (30 A: 7d wap’ Hpiv odpa ap’ ov spuyny djoouev EXE ;— Sijrov Ste Gyocopev.—roGev AaBov, elrrep ur) TO ye TOU TWavTos odpa suxpuyoy bv ariyyave tadtd ye Eyov Tour Kai ere advrn xaddlova ;—Sjrov.ds ovdauobev addoOev). Our soul is compared to a book, in which memory and sense-per- ception inscribe opinions and judgments (39 A: 4) pry tats aicOnoeoe Evuprinrovea sis tavrov, kaxsiva & mepl radTa dors ra WaOjpara, daivoyvral wor ayeddv olov ypddeav huav dv rais yruyais réte déyous). It may be taken for granted that these opinions remain for Plato essentially different NEW THEORY OF SCIENCE: THE PHILEBUS 469 from dialectical knowledge. The ideal of knowledge is in no way lowered, as can be seen from the above quota- tions on dialectic, but it 1s more and more looked upon as divine and opposed to mere human opinions. Our souls are copies of the world’s soul, our notions repeat God’s ideas, our knowledge finds out the Creator’s final aim. Thus ideas remain eternal, though they have lost their supramundane existence, and are to be sought and found only in souls. There is no substantial difference of doctrine between the Philebus and Politicus, and both belong most probably to the same time. Only in some points the Philebus appears to refer more or less clearly to the Politicus : 1. The division of sciences into theoretical and practical (57 A—-E) appears here more familiar than in the Politicus (258 E). 2. The world’s soul has been introduced in the Poltticus (270 A), and is mentioned in the Philebus (30 A) as a matter of course. 3. The absolute measure (rd pétptov) 18 explained as a new notion in the Poltticus (2844), while it is here applied (66 A). These tests are, however, not decisive, and only further stylistic research can lead us to settle the ques- tion of priority between these two dialogues, a question which appears not to have a great importance for the understanding of Plato’s philosophy so long as both are admitted to be later than the Sophist. The difficulty implied in the union of many different predicates with one subject, which has been specially investigated in the Sophist, is here mentioned (14D) in a manner which seems to point to the discussions of the Sophist. The relation of the Philebus to the Parmenides need not be insisted on here, as it is recognised even by Zeller, as well as by many other investigators, including those who have denied the authenticity of the Philebus, as for instance Schaarschmidt. Also the question of the relation be- Human knowledge may approxi- mate to the Divine ideas. The Politicus and Philebus represent the same stage of Plato’s thought. Priority between Phalebus and Politicus not yet deter- mined with certainty. Relation to the Parme- mides and to the Republic. Zeller thought that the Republic quoted the Phalebus, but the Philebus makes no reference to the ppovnors Tov ayabov 50 pointedly mentioned in Rep. 505 c. The Plulebus probably the latest of the dialectical dialogues. Charac- teristics of the dialectica] dialogues. Classi- fication. The ideas not self- existent. 470 ORIGIN AND GROWTH OF PLATO'S LOGIC tween the Republic and Philebus need hardly be discussed, although Zeller persists in arguing that the Republic is later. This view has been refuted recently by Siebeck,”® and Zeller has never accounted for the fact that the chief reference in the Republic to the question whether pleasure or reason is the good mentions a difficulty which is not found at all in the Philebus, namely, that the defenders of reason are obliged to confess at last that the sought-for dpdvncts is dpdvnats Tod ayabod (Rep. 505 B), a position which is declared to be ridiculous (yéAolws). This looks like a criticism of some philosophical opponent who 1s difficult to identify, and not like a quotation from the Philebus. We have seen above that the Philebus shows in every respect a later stage of thought than the Republic, and stylistic comparisons have made it so evident, that no reasonable doubts remain for those who know Plato’s style. Stylistically the Philebus is nearest to the Politicus, the collected evidence being insufficient to decide with certainty which of the two is later. If we take into consideration the close connection of the Poltticus with the Sophist, and the few points on which the contents of the Polsticus allow a comparison with the Philebus, it becomes very probable that this dialogue is the latest in the dialectical group, to which it naturally belongs. New dtalectic. If now we resume the logical theories of the three dialogues following the Parmenides, we observe that their chief peculiarity is the great importance given to division and classification. These logical operations apply here to notions of the human mind, which are similar to the divine ideas. Ideas are no longer self-existing, but exist in the divine mind, and from thence pass to our souls through the observation of concrete particulars. The material world is built up according to God’s ideas, and we have to find them out by comparisons and distinctions NEW DIALECTIC 471 of particulars. This credits the external world with an existence which is other than the existence of ideas and of souls. Plato first recognised this double meaning of existence, whereby he advanced far beyond his prede- cessors. The new dialectic is distinguished from middle Platonism and the earlier theory of ideas by the greater importance attached to particulars. No explanation of the universe is accepted as sufficient, unless it accounts for the smallest and most insignificant detail as well as for the greatest ideas. The unity of all existence is no longer an abstract unity, but a summit built up on the widest basis of the universal experience of mankind, to which each investigator has power to contribute ac- cording to his own aptitudes. The attempt made first in the Theaetetus to enumerate the highest categories is maintained and carried further throughout the dialectical dialogues. A worship of method, unknown in earlier Platonism, is here predominant over all particular subjects of inquiry. The influence of a long and _ successful educational activity is visible at every step of the argu- ment, and the final aim of an universal knowledge of all reality is sought through constant logical training and reflections on the method of inquiry. Reality of the external world, ordered according to God’s ideas. Import- ance of particu- lars. Know- ledge based on universal experi- ence. Enume- ration of categories. Import- ance of method. 472 ORIGIN AND GROWTH OF PLATO'S LOGIC CHAPTER IX LATEST DEVELOPMENTS OF PLATO’S THOUGHT WHILE our view of the way leading Plato from the Protagoras to the Philebus appeared as the result of difficult chronological investigations, and needs still many confirmations before it is generally accepted, there is an almost universal agreement as to the final stage of The latest Plato’s philosophy. All critics have unanimously re- group: += =cognised the Timaeus, Critias, and Laws as the latest Timaeus, works of Plato, and there is in this respect scarcely any Crittas, difference between the representatives of the most diver- Laws. ; : : ging views on Plato’s chronology. The reasons of this unanimity are found in the peculiar contents of these works, their relation to other earlier writings of Plato, and also in direct testimonies of Aristotle and other witnesses as to the very late date of the Laws, to which the Timaeus and the Crittas fragment are nearest in style and doctrine. Campbell has found in Timaeus and Coinci- Critias eighty-one words which besides these dialogues dences of gre found only in the Laws and form so many peculi- styleand grities of latest style, absent even from the dialectical contents. dialogues. There are also some important points of theory in which the Timaeus coincides with the Laws only. Thus for instance the representation of the stars as bodies belonging to perfect divine souls (Tim. 41D, Legg.8998), and the enumeration of more than two kinds of motion (Tum. 438, Legg.894D) are important views not found anywhere else in Plato. It is, perhaps, not quite accidental that both in the Timaeus (20 4) and in the Laws (6388) Locris, not mentioned elsewhere LATEST WORKS: TIMAEUS 473 by Plato, is highly praised, and it may well have a personal explanation in connection with the third voyage to Sicily, or with the tyrannis of Dionysius in Locri (356 B.c.) ; also Tyrrhenia is mentioned only in the three latest dialogues (Tim. 253, Critias 114c, Legg. 738 ¢). The number of stylistic peculiarities common to Timaeus and Laws, and peculiar to this group, is very much more considerable than is shown in our reduced list of 500 marks of style, because with a few exceptions such pecu- liarities as are common only to a very few dialogues have been excluded. Still it is easy to see that some peculi- arities of the Laws are found nowhere else than in the Timaeus or Critias. Such are mpérov dv sin (312), nal? irrvov (432), Oopu8wdns (278), otatpwdns (268), and some peculiar uses of re (230, 233, 235). It would be easy to increase this list of peculiarities of the latest style of Plato to any extent, but in view of the universal agreement as to the very late date of Timaeus, Critias, and Laws, it 1s not necessary to insist any more on this subject. The Critias being a fragment anda close continuation of the Timaeus, we have really only two works to consider in this group: Timaeus and Laws. I. The Tvmaeus. There are very few logical elements in the Tumaeus. Here true and probable opinion takes a larger place than in the dialectical dialogues, but the decisive and irredu- cible difference between such beliefs and perfect know- ledge is emphatically maintained (51D: vods xal S0fa GAnOijs . . . Svo Nexrdov, Sidts ywpls yeysvatov avouoiws Te Byerov). Knowledge is imparted by teaching, opinion by rhetoric, knowledge is unchangeable, opinion easily over- thrown, knowledge is a divine privilege of a few philo- sophers, opinion a common faculty of all men (51 E: 70 ev yap avrav dia Siwayijs, 768 irre reOods piv dyylyveras’ Kai TO pay dal wer AXnOods Adyou, To 58 dAoyov’ Kal TO wav anlynrov meOol, To 58 wetatreoToy Kal ToD wey madvTa dvbpa peréyew Natural science & work of human opinion, imper- fectly approxi- mating to Divine know- ledge. Priority of soul. 474 ORIGIN AND GROWTH OF PLATO'S LOGIC dardov, voi dk Oeovs, avOpwrrwv 53 yévos Bpayv 71). The ideas which exist in reason eternally are the object of knowledge (27 EB: ro dv dei, yéveow Sz ovK Eyov . . . vonoEeL peTa AOryou TepiAntrov, aél Kata TavTa Sv), and reason occurs only in souls (30B: votv . . . ywpls Wuyis advvarov wapayevec Bae tw). This close relation between soul and knowledge is here insisted upon as in the Sophist, and makes it evident that Plato no longer dreamed of separate ideas. He says repeatedly that knowledge and reason cannot exist out of a soul (87 C: vods ériotnun te... dv @ TeV dvTMD eyyiyvecOov, dv org tis avTd GAXO TAY Wuyhy Ein, av paddov 7 TadrnOés gost. . . 46D: Trav yap dvTwy @ vobv povm KTac8ar mpoonkel, NEKTEOV Yruy7y). If the ideas could exist independently, then they would form also an objective system of knowledge, and Plato need hot have credited the soul with such an importance in relation to the progress of science. For him the ob- jectivity of knowledge has its only basis in the common origin and similar power of all existing souls. Every soul is anterior to the body, and rules over it (34c: yevéeoes Kal apeth mpotépay Kal mpecButépay Wuynv copaTos @s Secrorw Kai apf~ovcay apfoudvou Evvertyjcato ... ). There is an apparent contradiction in the explanation of the relation between soul and space. Once the soul is said to be in the body (808: yuyny év capari Evyiotas To Trav Euverextaivero) and then to contain the body and to extend through space or to include it (36DE: érel cata voby T@ Euvotdyts mace 4 Ths Yruyns Evotacis syeyévnto, peta TOvTO may TO owpaTosdes ayTos auras étextaiveto Kal péoov uton Evvayayav mpoonpporrev). The latter view refers clearly to the world’s soul, which animates the universe just as each individual soul animates our individual bodies (3808 c: tov Kocpov Caov Euruyxov évvouy te TH ddAnOeia Sia Thy rod G20 yevéoOa: wpdvorav). But it must be remembered that the whole creation of the world-soul and of individual souls is here given as a myth (80 B: «ata Noyor Tov sixdta Set Asyeev), and admits of various interpretation. LATEST WORKS: TIMAEUS 475 The great philosophical thought here illustrated is the perfect unity of the universe, which 1s represented as the Unity result of an ordering and over-mastering power of a God of the over a primitive indefinite andchaotic matter. Therecannot ¥°"4- be a multiplicity of worlds, argues Plato, because the true world is only that which contains everything according to God’s conception of a perfect whole (31.4: qorepoy ody op0as Eva ovpavor mpocetpyKaper, 7) TroddOVs Kal arréelpous Aéyew hy opOorepov ; Eva, elep Kata TO wapdberypa Se- Snucoupynpévos zara). If several worlds were imagined, one idea of a universe containing them all would still be needed, and thus the whole forms one unique universe (31 B: wa ody 10de cata THY povwctw Spotoy 7 TO TavTEre Cow, Sua radta ovre Svo ovr’ arreipous eroincey o TOL@Y KOopous, GAN’ els Ode wovoryev)s ovpavos yeyovws Zot. Te Kal ET #otat, cf. 928). Out of the earlier world of ideas existing by themselves and influencing all appearances, there is only one left now, and so transformed that it is scarcely recog- nisable. It is the ideaof the Good transformed into the The good Demiurge, whom we ought not to call really a Demiurge Creator, because he orders the world only out of a pre- 924 the existing chaos, without calling into existence anything oe that was not before. This Demiurge is outside the world, and different from the world’s soul. He imparts to the world its shape and present nature, abiding thereafter in his own eternal peace (42 E: o pév 87 drravta Tatra diaragas Euevev dv TH éavtov Kata TpoTov HOE). This conception of a God, who dwells at a height above the world ordered by him, is common to the Timaeus with the Politicus (272: rod wavtos 0 wav KuBepvntns, olov andariwy olaxos adéuevos, eis Thy abrov mepiwmny avréatn, Tov 52 57) Koopov Tdadw avéorpepev eimappévyn Te Kal Evudutos ériOupia), where also the contrast between blind necessity and divine rule has been first recognised. God's self-contemplation of the Politicus has been here deve- loped into an invariable peace. The difference between Time and Politicus and Timaeus consists in a more complete sepa- Eternity. God is unenvious and rules through free con- viction. The whole dis- course is mythical, ‘ similar to truth’: 476 ORIGIN AND GROWTH OF PLATO'S LOGIC ration of time and eternity. While in the Polsticus divine rule and the power of necessity alternated in time, and thus formed consecutive periods, we have here an eternal permanent influence of divine rule opposed to the working of necessity in time. The Demiurge of the Timaeus (cf. Rep. 5804: rod ovpavod Snusoupyos, also 597BC) is good and free from envy, desiring to make everything as like himself as possible (29E: ayabos Hv, aya0@ 52 ovdels rrepi ovdevos ovdérrota ayyryveTat @Oovos * tovtov 8 éetos Ov TavTa 6 TL padrdLoTa yevérOat §BovrA7On mapamdjoia égavto). He is the best of all causes (294: dpioros tv aitiwv), and it is difficult for us to find him out, more difficult still to explain him to others if they are unable to find him by their own reason (280: roy ue ovv TrounTny Kal Tatépa rovde Tov maytTos eipely Te Epyov Kai eipovta sis mavtas advvaTtov Aye). His influence is compared to the free convic- tion of one soul by another, not to the necessary action of one body on another (484: vod 82 dvdyxns adpyovros To welOey avTny TeV Yyuyvonevwy Ta TrEioTta ati TO Bértiotov ayew, Tabtn Kata Tatra te bv avdyKnns nrrwpévns urd wetOods eudpovos obtw Kat’ apyas Evviorato Tobe 76 wav... 56C: Sirnmep nh ths avayxns sxotdoa wmetcbeiad te duos urreixe). He is supposed to have found already in existence matter with its necessary forces and move- ments (80 A: may dcov jv opatov tapadaBov oby jovxlav dryov GANA Kivovpevov TANmpEX@s Kal aTaxTws, eis Tae auto ipyayev &e Ths atakias, Hynodpevos éxEivo TOVTOU TavTas duewvov), but this is not necessarily to be understood as occurring in time, otherwise it would contradict one of the most important axioms of latest Platonism, the priority of soul. For a correct interpretation of the story of creation as told in the Timaeus we must constantly keep in mind that it is a mythical and allegorical exposition, which from the outset does not pretend to be true, but only similar to truth. Plato having reached his LATEST WORKS: TIMAEUS ATT view of an eternal existence out of time, can scarcely have believed in a beginning of the world in time. If he represents the divine reason as introducing order in the chaotic world of matter, he does not mean that this chaotic disorder had a temporal priority. He only wishes to impress upon the reader’s mind the truth that wherever order and reason are found, they ought to be ascribed to divine influence, the origin of all order and thought (30 AB: Oéues 88 ob’ Hy ob?’ Batt TO apiot@m Spav dddo TAY TO KGAMOTOV * Noyiodpevos . . . vouv mev av Yuyy, yuyny d8 dv owpats Evmotas to wav Evvetextaiveto, dws 8 Th KadAoTov ely KaTA pow apiordv Te Epyou atretpryacpévos). Taking this for granted, we shall easily recognise that the ideas were nothing else for Plato when he wrote the Timaeus than God’s thoughts. We see that he re- repeatedly represents the ideas as included in thought (28A: vonoe, pera AGyou Twepidnrrdv . . . 2DA: Ady xal dpovnce trepirnmrév). They are always the same (284: ael xara tavrd, cf. 294, 384, &c.) and unchangeable, because they have no beginning nor end, nor existence in time (88 A: 10 del Kata TavTa Eyov axiwyntws ore pec BUTE- pov ore vewTEpoy mpoarxes yiyverOa 51a ypovou ovdz yevé- o8at tote ovdé yeyovdvar viv ovd’ sicatOis goecPat), nor participation of any kind in particulars, being inaccessible to the senses, but evident to reason (52A: 10 xara tavra eldos dyov, dyévuntov Kat av@deOpor, ovTE sis savTo siadexd- mevoy GAXa AdNoGev ore adTo eis GAXO Trou Lov, adpaToy b8 kal ddAws avaicOntov, Todro b 8) vénows eidnyev ericKo- aetv). The eternal nature of ideas is expressed by various terms: they are aféia (29 a, 37 D) or have a vars aiwuos (87 D) whereby their separate existence in time is recog- nised to be impossible. Their function is to be eternal models of thought, first existent in God’s mind, then reproduced in the investigating souls of men. The term Trapdaderyna is now constantly applied to ideas (28 a, 37 ¢, 39 n, 48 z, &c.) ; they are the models according to which the Demiurge has brought order in the world, and we are the action is not to be under- stood as happening in time. The ideas of the Tymaeus are God’s thoughts, and are out of Time. They are models or patterns of our best thoughts and of natural kinds. Animated Beings. Condi- tional im- mortality. The mortal soul of the Timaeus. Antici- pated in Polit. 309 c. 478 ORIGIN AND GROWTH OF PLATO'S LOGIC able to recognise these models by our own soul’s activity. This applies more especially to the natural types which form the limits and definition of each kind of animals. The conception of an animal or animated body becomes more prominent here than ever before. Not only the whole world is an animal, but also each star is the body of a divine animal or a god, distinguished from all other animals by its subtle matter (fire), by the perfection of its spherical shape, and by the great regularity of its move- ments (40 4: Tod pay ody Oeiov thy wrsiorny iStay a rupos atreipydtero ... veluas Tept mavtTa KUKA@ TOV ovpavoV . cunoes S& Svo tpoonev Exdotm, Thy piv dv TabT@ KATG TAUTA TEpt TOV avTav aél Ta avTa savT@ Stavooupéeva, Thy de eis To MpocOev bro THS TavTOD Kal opolov TrEepipopas xpaTtoupnsva). It is exceedingly important for the understanding of latest Platonism that even these perfect gods with perfect souls are no longer conceived as absolutely im- mortal by their own nature; they owe their permanence to the Demiurge’s personal will (414: the Demiurge speaks : Ozol Gedy, dy sya Snutovpyss maryp te Epywv, & 81’ dod yevopeva Aruta suov ye wn 2édrovtos). We see here a theory which to some extent was already implied in the Phaedrus: only the simple substance is indestructible, all compounds being reducible to their elements, and subsistiny only through the divine influence. The last consequence of this view had not been drawn in the Phaedrus: there the human soul, with its three parts, existed indefinitely ; here a mortal part of the soul is distinguished from its immortal part. To this mortal part belongs nearly everything that constitutes personal character: pleasure and pain, courage and fear, anger and hope, perception and love (690: dAdo re sldos wuyis mpocwKodonouv Td Oynrov, Sewa Kat dvayxata dv saute mwabnpara #yov). A vague distinction of a divine and a human or animal part of the soul was occasionally touched upon in the Politicus (809: 16 devyevas dy rijs LATEST WORKS: TIMAEUS 479 apuyns avtav pépos Oeiw Evvappocapuévn Seou@, peta 52 76 Osiov 76 Swoyeves avrav avOts avOpwrivos), but only here we find a nearer explanation of this difference which exceeds in importance all previous partitions of the soul. The mortal part corresponds to the two earlier inferior divisions, with the difference that alc@nor1s combined with %ows takes the place of émriOuyia, and that @uyos holds a lower rank than formerly (424A: mpwtov pév aicOnow avayxatov ein piavy macw 2x Biatwy traOnpdtwv Evadurov ylyverOat, Sevtepov 62 Hdov7 Kal NUIT pEpvypevoy Epwra, mpos 58 TovTots PoBov Kal Oupov boa TE érropeva avtois Kal ondca évavtiws mépuxe Steotnxora). The lower part of the soul occupies the lowest part of the body, and is common to men with other animals and plants (77B: peréye ... TovTO . . . TOU TpiToUrpuyns Eldous, 6 peTakd doevav dupadod Te ispvcOa Aodyos, @ So—ns wey Aoyiopou TE Kal vod péTECTL TO pndev, alcOnoews 52 Hdelas Kal adyewns peta erOupior). It is very remarkable that the successive incarnations of the immortal part of the soul are maintained, with the supposition that in each incarnation the lower activities grow with the body. Thus it is admitted that not only the same soul is repeatedly incarnated on earth in the shape of men or women, but also the possibility of a descent into the form of lower animals is left (42 B¢). Under these circumstances nothing remains for the im- mortal part of the soul except the abstract conception of a principle (4py7 42 £), as already formed in the Phaedrus with special reference to motion. A place in the body is assigned to this immortal soul in the head (69 £) in order to keep it apart from lower tendencies. Knowledge is the only activity of this immortal principle, which is the divine element in man (90D: T@ 8 év nyiv Oem Evyyeveis iol xuvnoais ai roo Tmavros Siavonoes Kal wepidopal). The ultimate aim is here as in the Phaedo and Theaetetus to become as like God as possible, only here feelings and even virtues appear to be excluded from the divine perfection, for which only pure knowledge is left. This Re-ineat- nation. The immortal principle of thought located in the head. Know- ledge the supreme aim. Categories of Reason. Judgment and sentence. Control of bodily conditions necessary to thought. Ante- natal vision of 480 ORIGIN AND GROWTH OF PLATO’S LOGIC is the result of the dialectical construction of universal science. As knowledge was the starting point of Socratic Platonism, it becomes the final aim of Plato’s life. The same idea of intellectual exercise which had such im- portance from the Parmenides onwards is also here the predominant factor in reaching the aims of knowledge, being identified with a kind of movement which corre- sponds to the best part of the soul (89 E,cf. 908). These movements produce the various categories of reason, which are here more fully enumerated than in the Sophist, being very closely similar to the Aristotelian list of cate- gories : Tum. 87 AB: (7 ux) . . . Aéyes xivoupévn ia mdons éautans, 6r@ tr dy reravrov 7 Kat Grov av Erepor, mpos 6 ri re padtora Kai Grp Kai Sraws kat dwdre EupBaive xara ra ytyvdpeva re mpds éxacrov éxacra elvat cal madaxety nal mpds ra kara raurda é€xovra ae. Arist. Categoriae 1 b 25: rav kara pndepiay oupmdokny Aeyouevoy €xagtov rot ovciay onyuaive 7} mwoaov h Trotdv ) mpds Teh mov H Tore i cet Oat fj €xewy 7) mocetv h mdayey : cf. Topic. 108b 20, and above, pp. 868-369, on the categories in Theaet. For Plato these are the highest kinds of ideas, while Aristotle looks upon the categories as chief kinds of words. But the conception of categories, as later under- stood in the history of philosophy, we owe to Plato. He explains in the Timaeus our faculty of judging all existence through the recognised identity of substance in all souls. The familiar distinction between judg- ment as an act of the soul and the sentence as an expression of judgment recurs here also (87B: XAoyos . . dv TH Kivouptvp Up’ avbrod hepduevos dvev POoyyou Kai ips) and judgment includes opinion as well as know- ledge (870: Sofas xal wioras ... voids émornun Te). The reason acts by distinctions, and requires for the full development of its activity a certain limitation of the stream of bodily changes (44 B). Each soul is supposed to have seen once the nature of the whole universe and the moral laws which thus are an innate possession of each individual (41 E: Suetrs LATEST WORKS: TIMAEUS 481 Yuxas icapiOuous tois aotpots, Evewé 6 éxdotny mpos Exaotov, Kal éuBi8dcas ws és ynua THY Tov TavTos hvow Here, vomous TE TOUS eipapmevovs eimev avtais). It is highly characteristic of Platonic logic that such knowledge does not refer to the physical occurrences in the world, as to which Plato has only to offer uncertain opinions and probabilities which do not even pretend to be consistent or to attain any exactness (29 C: gay ovv moAXd TrOAA@V elrrovrewy rept Oe@y Kal THs TOD TavTOs yEevécews, un SuvaToL yiyvoueda mwdavTn TavtTws avtovs éavTois opodoryoupévous Aoyous kal arnxpiBwpévous atrododvat, wn Oavyaons). Full knowledge as to these things must be left to God, while men must be satisfied with probable myths and ought not to search further (29 p). Physical investigation is held to be only a convenient pastime in moments when we are tired of metaphysical inquiry (59 c). The same uncer- tainty refers to empirical psychology as well as to general physics (72 D: rad pay ovv trepi yuyns, dcov Ovntov Eyer kat Scov Osiov, kai Orn, Kat ped’ dv, Kat SC & ywpis @xicOn, TO pev adnOés, ws slpnrat, Oeov Evpdprycavtos tor’ adv ovTw povws SucyupiCoiucba). The sensible world consists of appearances which are becoming and changing without true permanent existence (28 A: d6&n pst’ aicOnoews adoyou Sofacrov, yuyvopuevoy Kai atroNdvpevov, Svtws 52 ovdérrote Gv). The physical universe, like everything that is material, had a beginning (28 B: OKENTEOY . . . TOTEPOY HV GEl, yEvérEWs AapYnY EXwV ovdEpiay, 4) yéyovev, dx’ dpyijs tivds apEdpuevos. yéyovev* opatos yap dros Té ott Kal oGpa eywv, wavta 62 Ta ToadTa aicOnra, ta 8 aicOnrd, dé€y mepikntta pet’ atoOnoews, yuyvomeva Kal yevynra épdvn). But if Plato adds later that the world did not begin in time but together with time, leaving it open whether both will have an end (38 B: ypévos & ovv per’ ovpavod yéyovev, wa dua yevvnbédvtes dua nal AvOdow, av Tore AVols TIS avTaY yiyvnta), he authorises us to interpret the beginning of the world not as a temporal beginning, but a relation of dependence of II Truth and Good. Physical occur- rences into which human beings must not inquire too curiously are matters of opinion. Every thing that is material had a be- ginning ; but only in the sense that it is dependent on Divine Power. The material world is made in the like- ness of an eternal pattern. General physical notions : Time, Space, Matter, Causality. Concep- tion of Space. 482 ORIGIN AND GROWTH OF PLATO'S LOGIC the material world on a divine power. Physical appear- ances are represented as an imitation or an image of the eternal ideas (29 B: waca davaykn rovde tov Kdcpov sixova tivos elvat, cf. 49 A: pipnua mrapadeiypatos, yévecww ¥yov Kal opartov). The ideas seem to be chiefly limited to natural kinds, and do not include some of the most general physical notions which are investigated in the Timaeus with special care, namely time, space, matter, and causality. The analysis of these notions is not without logical importance and therefore deserves our attention. Time is a@ moving image of the eternal nature of ideas, and is placed into close relation with the movements of stars (87 D: exw 8 darivost Kwnrov tia ai@vos Totnoal, Kal Staxocpa@v dua ovpavoy troet pévovros aiavos av évi Kat’ apiOuov iotcav aiwwoy sixova, rovrov by 8) ypovoy wvopd- xayev). This acts in obedience to a very primitive con- sideration: days, months, and years are made by the celestial movements, and as they are parts of Time, Plato infers that Time itself is a product of those movements (837 E: npepas xal vuctas nal phvas Kal énavrovs, ov« Ovras mplv ovpavoy yevécOat, TOTe Gua éxsivm Evricrapéve Thy yéverw avTay pnyavarat * tadta 88 wdvra pépn ypovov, wal to 7 hw TO 7’ Eorat ypovou yeyovdra edn). Thus Plato did not reach the abstract conception of time, and knew only concrete durations measured by physical movements. More elaborate is the conception of space, which Plato introduces as co-ordinated to ideas and their images at a later stage of the inquiry, expressly avowing that he had omitted it at the beginning (49 A: rpérov 58 tore piv ov SustdopeOa, vopicavres ta Svo sav ixavas). He finds this notion very difficult to explain, and unlike Time, Space is admitted to exist before matter, being necessary for the reception of matter into being (494: mdons slvar yevécews rodoxhy avtTyny olov r:Onvnv). The imagined identity of all matter is here the starting point (4938), based on the observation that water becomes ice as well as LATEST WORKS: TIMAEUS 483 steam, or air, which is supposed to be susceptible of a change into fire (49 c), thus forming a circle of transfor- mations. From the fact of transformations the unreality of appearances 1s inferred, and the reality underlying them is found in the notion of space, free from any determining quality (49 ©). Plato’s fondness for proportions and analogies finds here also an opportunity for display. Space has the same relation to matter as matter to form (50 4). Thus, if various forms were given to one and the same matter, for instance gold, each particular object could best be named gold, and not according to its special chan- ging form. Ina similar manner space remains always the same, however different qualities of matter might fill it. Here we see Plato advancing to a more abstract notion than he had of time, and he feels the difficulty of explain- ing it. He calls it by different names, beginning with Svvapis (49 A, 50 B) and gvcrs (50 B), and ending with yoépa (524). Itis ta wavra Seyouévn copara pvors (50 B), raons yevécews Urrodoy7 olov TuOnvy (49 A), TO ev @ yiyveras (500). Space has no shape, but appears differently according to the phenomena occurring in it (50c: déyeral te ael ra TavTa, Kal wopdyv ovdewiay rote ovdevi TAY EioloVTwY Opotay elandev ovdayy ovdauas). It is not an idea, nor imitates any idea (50 E: dwopdov dv éxetvwy atracay TaV iOeav, daoas wérrot SeyecOai Tobey . . . mdvrwy éxtos eiddv elvat ypewy TO Ta Tara éKxdekopevov gv avT@ yévn). Its relation to the ideas is recognised to be most difficult to explain (514: Gvoparoy édds tu Kal duopdov, ravdexyés, peradapBavov oe amopwrTatd wy tod vonTod Kal SucadwrToTaTov avTo AéyorTes ov yrevodueOa). Space is conceived not by the senses, nor by purereason, but by a kind of fictitious inference which has however a necessary character (52AB: yavos dy To Tis y@pas asi, Pbopdv ov mpocdexcpevov, Epav dé mapeyov Soa Byer yévecw waow, ato $8 per avacOnolas amrrov oyioug rivt vdbw, woyss miarov, pos & Sy Kal Gvecporro- Nodwev BArsrrovres xal hayev avayKaioy elvai mov 76 bv Gray Space void of qualities. Notion of Space more abstract than of Time. dy tive rom@ Kal Karéyov yopav tid). Space has been thus Matter. 112 Quality an appear- ance resulting from invisible motions. The infinitely little. Flux of particles in every organism. Lavoi- sier’s analysis of water antici- pated. Causality : final and efficient causes. 484 ORIGIN AND GROWTH OF PLATO'S LOGIC represented as an indispensable condition of matter, though essentially differing from matter. The difference consists in the entire absence of quality in space, while matter is held to be always qualified. To illustrate the relation of space to matter, Plato uses many metaphors taken from the impregnation of one matter by another. Thus various perfumes can be com- municated to an oil which has no smell by itself, various shapes to a shapeless clay (508). The chief kinds of matter, as earth, water, air, and fire, can be changed into one another (49 B), and existed in space before the Demiurge set to work (52D). The different qualities of matter are only appearances resulting from a variety of movements (52D). There are amid all the mythical fictions of the Timaeus some wonderful glimpses of deep insight which betray Plato’s genius. Thus he speaks about invisible matter and its infinitely small. elements (560: dia opixporynta ovdév opapevov Ud’ judv), about the stream of matter passing through our body (434: éip- putov capa Kat amdpputov), about the spermatozoa, which he seems to have divined many centuries before their actual discovery (91C: péypirep dv éxatépwv 7 eriOuuia xai o Epws Evydiaryayovtes . . . ws els Gpovpay THY pHTpav dopata vo apikpotntos Kal adidmAacta boa KaTacTEl- pavres Kal wadw Siaxpwavtes peydda évros éxOpépwvrar Kal peta TovTo sis das ayayovres Cowy amoTedéowor yéveotv). By a strange divination he calmly teaches us, in agreement with our modern discoveries, that each particle of water consists of three atoms, two of one gas and one of another (56 D), thus anticipating the results of Lavoisier’s experiments. But he declares the analysis of colours to transcend human ability, and to be a divine privilege (68 D). On the contrary, stereometry is here advanced (53 c—55 c) beyond the stage complained of in the Republic. One of the distinctions made by Plato already in the Phaedo is here developed and fully explained. The dif- ference between final and efficient causality has a consi- LATEST WORKS: TIMAEUS 485 derable place in Platonism ; while at an earlier stage only the final cause had been recognised as a true cause and opposed to the current notion of causality as employed by Anaxagoras, Plato later admitted efficient causes, and this change had a close relation to the increasing importance of the notion of movement in his system. (See above, p. 452.) The terminology established in the Politicus. is here maintained. The final cause is named airiov, and acts everywhere, nothing being possible without an aim (28 A: mav d& av TO yLyvopevov Um’ aitiou Tivos && avaykns yiyveoOat’ TavTt yap ddvvatov xwpis airlov yéverw oyeiv). Vulgar people call airvov what is only €vvairsoy, namely, material causes, used by God only as means for the realisation of the best which is his aim (46c). The reason of the superiority of final causes over mechanical causation lies in the absence of reason and design from physical causa- tion, if considered alone and apart from aims which can be conceived only by a soul (46D: ta&v yap dvTw@v & vodv Hove Krado0as mpoonjKe, Nextéov yrvynv). The philosopher, as lover of reason and knowledge, thinks more highly of aims of the mind than of necessities of matter (468). The final cause is here identified with a first source of movement and contrasted with the physical cause which is a movement caused from without. It is also called divine or free, as opposed to what is necessary (68 E: do 59) xpy Sv’ airias esidn StopifecOar, TO piv avayKaiov, To Se Ociov). The mechanical cause, here as in the Poltticus called £vvairiov and identified with the necessary condi- tion without which, as stated in the Phaedo, the aim could not be reached, is blind Necessity (avdyxn, 48 A), opposed to Reason. Reason acts on Necessity, leading it to the best aim, and Necessity yields to Reason (484: vov O68 dvdyxns dpyovtos TO TreiGew avTnv TOY yuyvopevey Ta wrelota ert to BéAtictov ayew, tavTn Kata Tabta te Se avayKns httwpévns Uro Teovs Eudpovos ottw Kat’ apyas Euvicrato rode 7d Tray). Con- current causes. Mechani- cal causes are tuvalria, Reason and Necessity. Being and Becoming. Place of the Timaeus. Stylistic affinity to the Lau 8 Implied references to previous dialogues. 1. Effi- cient causes. 2. The visible world proceed- ing apart from God. 486 ORIGIN AND GROWTH OF PLATO'S LOGIC This general view of Being and Becoming, as given in the Timaeus, is not essentially different from the theories contained in the dialectical dialogues, and some theories of the Timaeus are already prepared for in the Phaedrus and Theaetetus. This leaves very few points for confirm- ing by detailed comparison the place generally assigned to the Timaeus towards the end of Plato’s literary activity. We are justified in accepting in this case the agreement among all investigators (with insignificant exceptions, as, for instance, Munk and Schéne), be- cause the theories here observed agree exceedingly well with the Politicus and Philebus, two very late dialogues, and because the style of the Timaeus is nearer to the style of the Laws than the style of any other dialogue. This is here affirmed as the result of the personal im- pression produced by the perusal of many thousands of stylistic observations; it would take too much space to enumerate here all the peculiarities of style common to the Laws with the Tumaeus only, and we refer the reader to Ast’s Lexicon and to the authors quoted in Chapter III. In all these works he will find sufficient evidence for the great stylistic affinity of the Timaeus with the Laws, an affinity far exceeding the numbers of our own table of affinity, based only on 500 peculiarities. Besides the style there are still the following arguments in favour of placing the Timaeus after the dialectical dialogues : 1. The relation to the Politicus in the theory of final and efficient causes, as explained above. 2. The relation to the Poltticus in the theory of God’s retirement from the world. 3. The transition from the form of a dialogue to a continuous exposition, recurring in this form only in certain parts of the Laws. Longer speeches in earlier dialogues were either of no didactic character (Apology, Protagoras, Menexenus) or interrupted by many questions and answers (Gorgias, Symposium). Such a _ purely didactic exposition in a longer speech without interrup- LATEST WORKS: TIMAEUS 487 tion is a peculiarity of the Tumaeus, Critias, and Laws, found nowhere earlier. 4. The view that the same elements are common to man with the universe is found in both the Philebus and the Timaeus, but presents in the Timaeus a more advanced stage. In the Philebus this view is introduced as new and as a daring feat (29 A: cuyxivduvedwpev xal peTexopmer Tou poyou, Stay avnp Sewos ph Tavta sn ovTwWs AA’ aTaKTos #yeuv—this refers to the preceding axiom that reason has ordered everything, and also to the following hypothesis : Stu ouixpov TovTwy éxacrov Trap’ Huiv Eveoti Kal davdAov Kal ovdaun ovdamas eiducpwves Ov Kai thy Sivamty ovn akiay rhs g@ucews Zyov). An attempt is made to prove it by induc- tion (29 B: év évl 88 NaBwv wepl wavTwy vos TavTov. olov Top pév ote Tou Tap’ Hiv, Bott & éy TO WavTl. . . . TMIKpOV pév 7. TO Tap’ Hpiv Kal dobevis Kai hadrov, To 8 év TO ravi awrnOse Te Oavpacroy Kal adr Kal wdon Suvaper TH Tepl 70 tip oon). All this is supposed to be known in the Timaeus, and needs no further demonstration. 5. The world’s soul as the origin of individual souls is also first introduced in the Philebus (30 A), and appears there as something new, while the same view is the basis of the mythus in the Timaeus. The relation in this respect of Timaeus and Philebus 1s similar to the rela- tion between Phaedrus and Republic in the question of the threefold partition of the soul: the mythical exposi- tion uses truths previously reached by reasoning. This is not necessarily a general rule, as at an earlier stage the intuition of ideas was first mythically given in the Sym- postum, then reasoned out in the Phaedo. But as Plato later had an increasing liking for myths, it is natural that he should represent also mythically truths which had been earlier set forth us based on reasoning. 6. Philosophy is represented as gift of Gods in Phile- bus and Timaeus. Though this is a commonplace of Platonism, there is in the form in which the mention is 3. Con- tinuous exposi- tion. 4. The elements in Man and Nature. 5. Human souls derived from the world’s soul, as hinted in Philebus. 6. Philo- sophy a gift from God. 7. Doc- trine of sense- percep- tion. 8. God free from pleasure and pain. Relation to the Republic. Temaeus separated oy a long 488 ORIGIN AND GROWTH OF PLATO'S LOGIC made in the Timaeus something that may well be inter- preted as a reminiscence of the Philebus : Phil.16 0: Oedv per eis avOpoa- mous Sdats, Ss ye xarapaiverat €pol, robev éx Seay eppidy dia Twos Hpopnbéws dua pavordre tivi up. Tim. 47 AB: ... éroptodpeba hiroaodias yévos, ob peifov ayadoy odr’ AAGev ovre HEE wore TO OvnTe@ yéver dwpnbev &x bear. 7. The explanation of sense-perception in the Philebus (33 D) as a movement communicated to the soul through the body is more elementary than the corresponding mention in the Timaeus (48 c). 8. The state of God as free from pleasure or pain is announced in the Philebus to be the subject of a future inquiry (83 B: rodto ts Kal sicaiOis émicxeipopeba, bday mpos Aoyov rT 4), and the Tvmaeus more than any other work seems to correspond to that announcement. In the above statement we have made no use of the peculiar relation of the Timaeus to the Republic. The reference to the Republic at the beginning of the Timaeus is unmistakable, but the relation of the two dialogues is not quite the same as the relation of the Sophist to the Theaetetus. In the Sophist we have a direct continuation of the Theaetetus, and the persons of the dialogue are the same, with the single addition of the Eleatic guest. In the Timaeus the scene is different from that of the Republic, and Plato recurs to the fiction that the substance of the Republic dialogue has been narrated on the previous day to the persons first appearing in the Tumaeus. This fiction is deemed insufficient and improbable by the author himself, and he puts in the mouth of Socrates a recapitu- lation of the preceding dialogue. In that recapitulation not the whole of the Republic is included, and no mention is made of the four last books. Far-reaching inferences have been made from this omission, about the structure of the Republic as well as about the date of the Timaeus. The most obvious conclusion would be to allow a longer distance of time between Republic and Timaeus than between Theaetetus and Sophist. This conclusion is con- LATEST WORKS: TIMAEUS 489 firmed by our whole inquiry and best explains the great change of style and of the literary manner. Other conclu- sions, at first sight plausible, are refuted by the considera- tion of style. The recapitulation of the Timaeus seems to refer only to the first five books of the Repwblic, and thus we might be tempted to suppose that it was written before the following five books. But in view of the great unity of composition of the Republic and of the great homogeneity of its style, it is impossible to separate the fifth book from the following by any other work. On the other hand, the close relation of the Timaeus to the Laws makes it very probable that some twenty years have come between that apparent sequel of the Republic and the work which it presupposes. Under these circumstances it is very natural that Plato should omit some details from his recapitulation, and should limit it to the most general results, which happen to be concentrated in the first five books. There is also another psychological reason why he should not now insist on the rule of the philosophers, which is the chief subject of the sixth and seventh books of the Republic. We must assume that the Timaeus at all events is written after the third voyage to Sicily (361 B.c.), and after Plato’s great and definitive failure to obtain political ascendency. His explanation of that failure is given in the Politicus, where he says that the ideal state is too perfect for mankind, and that the philosopher who could bring it into existence ought to be a god. Now in the Timaeus he plans a practical representation of the conflict between a perfect state and its neighbours. This conflict has not been represented by Plato, because he left the Crittas unfinished. But we have every reason to assume that he did not intend to identify in everything the historical state of primitive Attica with the ideal state of the Republic. The outline given at the beginning of the Critias confirms that assumption. Thus it is natural that fixed laws had to play a greater part in the old state of Athens than in the Republic. Still the identification of interval from the Republic. Rule of philo- sophers, why omitted in the Timaeus. The Critias un- finished : why ? 490 ORIGIN AND GROWTH OF PLATO'S LOGIC the political ruler with the philosopher is here also alluded to (19 EB: dirccédwv avdpav kal todtiKGv) In & similar manner to that which appears in the Politicus. Philo- sophy remains the greatest gift of gods to men (47 B), and ignorance an illness of the soul (86 B). Also the low estimate of poets is maintained (19 D), with a similar ironical compliment to that in the Republic. Generally, whoever considers impartially the relation of these two dialogues must recognise not only that the Tivmaeus pre- supposes the whole of the Republic, but that it appears to be very much later. Thus the place of the Timaews as nearest to the Laws is confirmed by every consideration, and no valid objection can be raised against this conclusion. It remains, however, desirable that the great number of stylistic peculiarities of this latest group should be col- lected and classified in order to confirm the common verdict of all competent authorities. Il. The Critias. There is little to say in the present connection about this small fragment of a dialogue left unfinished by Plato for some reason unknown. If we consider the great interest manifested in its introduction for the intended subject of this work, and the circumstance that it is the only fragmentary work of Plato, the most natural suppo- sition 1s that only death could have prevented him from carrying out such a cherished plan as that of the Timaeus trilogy. This supposition is also confirmed by the very late style of the Critias, but a definitive proof could be given only through stylometric comparison of the Critias with the latest parts of the Laws. In view of the small size of the Critias (11 pp. ed. Did.) a very great number of stylistic observations is required, and they ought to refer not only to the vocabulary but also to the construc- tion of phrases, inversion of words, phonetic effects, and LATEST WORKS: CRITIAS 491 all details which constitute the less accidental peculiarities of style. Only then will it be possible to decide with absolute certainty, on a basis of some thousand peculi- arities resuming some hundred thousand observations, whether the Critias is contemporaneous with the latest parts of the Laws or not. Such a special investigation exceeds the limits of the present inquiry: the more so as the whole question has no philosophical importance, and claims only a purely literary interest. The Critias contains no contribution to Plato’s logic besides the incidental mention of the familiar view that knowledge has a divine origin (106B: avrov (Oedv) tedkewratov Kal dpistov dappdxwv emiotiunv evyoueda didovac), and the curious appreciation of the cosmogony of the Timaeus as acceptable only on account of human ignorance about the gods and everything there expounded (107 AB: wept Oe@v . . . AéyovTd TL pos avOpwrovus SoKeiv ixaves Agyewy paov 4 wept Ovntav wpos nuas. 1 yap arretpia Kal opedpa ayvowa tay daKovovtwy aepi ov av ovTws Eywor TOAAHY Evtropiay traptxecOov Te pédrOvTe Aéyew Te rept auTov). Ill. The Laws. The question has been seriously discussed whether the theory of ideas is alluded to or maintained in the Laws. The question put in this indefinite manner is entirely out of place. Anybody who reads the Laws must notice the entire absence of the earlier theory of ideas as known from Phaedo and Republic. This has been recognised by all students of Plato, and Ribbing,”” who made a special study of the theory of ideas, went so far as to deny the authenticity of the Laws chiefly because he did not find there any trace of the Platonic ideas. Equally Ueberweg (Untersuchungen, p. 100) recog- nised that in the Laws the theory of ideas is nowhere 67 §. Ribbing, Genetische Darstellung der Platonischen Ideenlehre, Leipzig 1868-64, vol. ii. pp. 150-190. Divine nature of know- ledge. The Timaeus gave only a probable account of Divine things. Supposed absence of the ideal theory. Generali- sation. The Laws not written for philo- sophers. But Plato never 492 ORIGIN AND GROWTH OF PLATO'S LOGIC mentioned. The same has been the impression of English scholars. Grote (vol. iv. p. 275) and Jowett (vol. 11. p. 18; vol. v. p. ccxxxvi) agree that the theory of ideas is left out in the Laws. The same conclusion 1s reached by C. Ritter, in his recent commentary to the Laws, and Zeller also finds only one passage which could be interpreted as an allusion to the theory of ideas (Philo- sophie der Griechen, II. i. p. 953). This passage (965 Cc: mpos piav idtay & TOV TOAA@Y Kai avopuoiwy Suvarov elvat Brérrecv) is really as insufficient as evidence in favour of the old theory of ideas as similar passages from the earliest Socratic dialogues (Huthyph. 6D: wid ida ra Te avooa avoota elyat). AlsoSusemihl, who maintained against Zeller that Plato remained faithful to his theory of ideas up to his latest age, acknowledged (Genetische Entwickelung, vol. 11. p. 577) that the idea of the good can be only guessed at in the passage concluding the dialogue. It is very strange that in the whole discussion about the traces of the theory of ideas in the Laws nobody cared to distinguish between the earlier self-existing ideas and the ideas as known from the dialectical dialogues, where they appear as existing only in souls. Such ideas, equivalent to perfect notions, cannot have been abandoned by Plato, though he had no opportunity to mention them in the Laws, because the whole dialogue, like the Timaeus, rests on right and probable opinion, not on dialectical knowledge. Plato was not obliged to write always for philosophers alone, and he seems to have dedicated his latest years to a popular exposition of his political doctrines adapted to the actual level of mankind, very much below his own ideal standard. If somebody, like Grote, believes that Plato could become untrue to philosophy, he betrays only his own incapacity to judge a philosopher. Plato remained a philosopher up to his latest age, and the very last pages of the Laws prove it to 2 ©, Ritter, Platos Gesetse, Kommentar sum griechischen Text, Leipzig 1896, p. 855. LATEST WORKS: LAWS 493 any unprejudiced reader. The distinction between know- ledge and opinion is one of those logical distinctions which, once reached, cannot be afterwards neglected by a true philosopher, and if Plato could be reasonably suspected of such a desertion, no hope is left for any one of a permanence of knowledge. Plato remains in all ages the ideal type of a philosopher, and philosophy which is not knowledge is nothing. Thus it is from the outset a psychological impossibility to accept Grote’s interpreta- tion of Plato’s silence about ideas in the Laws, according to which Plato contracted ‘a comparative mistrust of any practical good to come from philosophy,’ ‘ eliminating or reducing to a minimum that ascendency of the philo- sophical mind which he had once held to be omnipotent and indispensable.’ Such extravagant conclusions are the result of a widespread error about Plato’s philosophy, consisting in identifying the so-called ‘theory of ideas’ with Platonic philosophy and with his philosophical knowledge. We have seen above that no such identification results from a chronological survey of the development of Plato’s logic. Even in the Republic the transcendental ideas do not include all the philosophy of Plato, and some of his logical doctrines have little to do with the world of ideas. The last appearance of such a world is in the Phaedrus. Already in the Theaetetus the categories occupy the place of ideas, which in the Parmenides also are supplanted by logical exercise in the analysis of notions. In the Sophist Plato speaks of his own earlier doctrine of ideas as belonging to the history of thought, and after the Sophist he never uses the terms eldos and idéa in the meaning which they had in Phaedo, Republic, and Phaedrus. It becomes for him a cardinal truth of philosophy that ideas and reason exist only in souls, so that they cannot any longer be looked upon as inde- pendent substances, though they are always called True Being. Ideas are perfect notions and refer more espe- aban- doned philo- sophy. Nor did he mistrust philo- sophy, as Grote imagined. Plato’s philo- sophy has been too much identified with the so-called ‘ theory of ideas.’ The Soul is the centre of Plato’s later theory of know- ledge. 494 ORIGIN AND GROWTH OF PLATO’S LOGIC cially to the natural kinds of animals in the largest meaning of this word (including plants) in all dialectical dialogues and also in the Timaeus. In this meaning alone we can expect to find them in the Laws, and in the only passage in which Plato mentions an idea in this dialogue (965 c), this is the only interpretation acceptable. A very strange prejudice is needed if we are to find every- where the old supramundane ideas, where Plato speaks of an idea in a meaning which exactly corresponds to the use of this word in modern philosophy. It can only be recommended to all those who still have any doubts on this subject to read with the greatest attention what Campbell (Rep. II. pp. 294-321) wrote about the use of metaphysical terms by Plato. They will then see at once that no conclusion can be drawn from the use or absence of terms like edos or (64a which Plato borrowed from earlier writers and used himself in many different mean- ings, ‘very seldom with a pronounced metaphysical intention’ (p. 294). Plato’s philosophy is not a mere theory of knowledge, and his theory of knowledge is not limited to the conception of ideas. The soul is not an idea, and acts a more important part in later Platonism than all ideas of Middle Platonism. It is the soul, and not the ideas, which is the central point of Plato’s later theory of knowledge. Here it is expressly acknowledged that dialectical questions exceed the scope of the dialogue and the understanding of the hearers (892E: viv o pér\Xwv dori Novos opodporepos Kai oysdov lows GBaros ws Th odev poun * wy bn cxorodiwiav truyydv re bin sutroujon mapadepouevds re xal pwrdy ajOes dvtas dtroxpicewr) and even a simple classification of psychical movement is followed immediately by the confession of both Kleinias and Megillus that they are unable to follow (644pD: poyis pdy mos &ptropat, \eye phy TO peta Tadta ws ésrope- vou—«al év guot pv tavrd Todt dos ¥x). Still, despite these limitations, we see here the theory of the soul made the object of a longer explanation, given in a more LATEST WORKS: LAWS 495 popular tone than in the Phaedrus and Timaeus, where the same doctrines were already set forth, and resuming the conclusions reached in both these dialogues. The old distinction made in the Phaedrus between the self-moving principle and all other moving and moved objects of the universe is here again stated with great rhetorical strength and with all the absolute certainty that Plato attached to metaphysical truths. Once entered upon this argument the reader must notice at once that he is outside the realm of probable opinions and plausible myths in the calm region of absolute know- ledge which never changes. Material things move in space (893 c: é& ywpa til Ta Te goTMTA EgTHKE Kai TA