STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. Classical Philology Volume XII April lOIf Number z KROLL ON THE INDEPENDENT LATIN SUBJUNCTIVE By Charles E. Bennett In a recent issue of Glotta Wilhelm Kroll has a lengthy and significant article on the Latin subjunctive. 1 Kroll's entrance into the field of syntactical discussion is sudden and unexpected — doubt- less to himself as well as to others. While well and favorably known in other fields, he had never before devoted attention — at least in his published work — to questions of Latin syntax. Even in the article to which I am inviting attention he confesses that his famil- iarity with certain standard syntactical works is recent. However, Kroll's general eminence as a Latinist and classical philologist, com- bined with the novelty and boldness of the opinions he enunciates, compels consideration of his views. For forty-five years — ever since Delbriick published his Kon- junktiv und Optativ im Sanskrit und Griechischen — opinion has grad- ually been crystallizing in favor of the views therein set forth. Delbriick's position is familiar and needs no restatement here. It has not passed without challenge — notably from Abel Bergaigne in France and from Morris in our own country. Morris denies " Grund- begriffe" for inflected forms; Bergaigne had not only done that, but had contested for Vedic the thoroughgoing application of the modal values set up for subjunctive and optative by Delbriick. There had been other dissenters. Most of these were not scholars whose names carried great weight, and their dissent, as a rule, was only partial. 1 "Der potentiate Konjunktiv im Lateinischen," Glotta, VII, 117-52. [Classical Philology XII, April, 1917] 121 122 Charles E. Bennett On the whole, it is fair to assert that a decided unanimity of opinion — especially in the last quarter of a century — had come to exist as to the values and functions of the Greek subjunctive and optative and of the Latin subjunctive, so far as the employment of these moods in independent sentences was involved. When I speak of unanimity of opinion, I mean, naturally, on the part of those whose researches and published works entitled them to be heard. The names of Delbrtick, Brugmann, Stahl, Schmalz, Blase, Landgraf, Stegmann, Riemann, Lindsay, Hale illustrate my meaning. All these scholars and hundreds of others who give adherence to their teachings have long recognized a pure future and a volitive function in the Greek subjunctive; a "wish" meaning and a " should-would " (or potential) function in the Greek optative. For the Latin, a voli- tive function, referred to the Indo-European subjunctive for its origin, has been recognized; likewise the "wish" and "potential" meanings, which have been regarded as lineal descendants of the Indo-European optative. Even those who have felt compelled to exercise reserve in connecting Greek or Latin syntactical phenomena with Indo-European forms have nevertheless as a rule recognized the actual usages of the Greek and Latin subjunctive and optative in substantially the terms I have stated. As regards these views, Kroll is frankly iconoclastic. He enters into no discussion of the Greek moods, but as to the Latin subjunc- tive in independent sentences he denies the true optative or wish use as distinct from the volitive; he likewise denies an original potential use for the subjunctive in Latin, while in a considerable number of cases he interprets as pure futures, Latin subjunctives which no one hitherto has had the hardihood to take in this sense. He naturally attaches great importance to the evidence of Early Latin and bases his conclusions largely upon Plautine usage. Barring certain individual observations of Kroll, I am unable to agree with any of his conclusions. He is unwilling to recognize in the Latin subjunctive any traces of the Indo-European optative. Optative forms, he frankly admits, appear in the Latin subjunctive, but optative functions, as recognized in Greek and Sanskrit, he refuses to accept. Evidently he starts out with the avowed inten- tion of recognizing in the Latin subjunctive only the functions Kroll on the Independent Latin Subjunctive 123 peculiar to the Indo-European subjunctive. As a general principle of philological method, this procedure cannot be too severely con- demned. Our science has already suffered enough from such attempts to make the interpretation of the text square with some preconceived conception as to the nature of a modal or case form. This tendency was most extreme, of course, in Gottfried Hermann, but for over a century we have gradually been growing away from it, and its revival in present-day investigation is deeply to be deplored. I feel that Kroll has ignored or jauntily dismissed from consideration certain facts of prime importance for the settlement of the problems of the Latin subjunctive. This is especially true in his attitude toward the "wish" and the "should-would" optatives. The facts are plain. All perfect subjunctives in Latin are optatives; and several presents — particularly presents in very common use, viz., sim (including possim), velim, nolim, malim, and in Early Latin (on whose testimony Kroll lays great stress) edim, duim perduim. Now the testimony of Greek, Sanskrit, Iranian, and, to some slight extent, of other Indo-European languages shows that the optative had the "wish" function. Why not, therefore, recog- nize the "wish" use of the Latin subjunctive as an inheritance of the Indo-European optative ? There are indeed special reasons for so doing in the Latin of Plautus, to which Kroll attaches such importance, for in Plautus it is a fact — first pointed out, I believe, by Morris 1 — that the majority of independent perfect sub- junctives (true optative formations) are used to express a wish. I find it impossible to reject the great cogency of this evidence. Kroll fails to reckon with it or even to allude to it. These "wish" subjunctives, he declares, are nothing but volitives. To his mind statim ad me redeas and viinam ad terram incolumis redeas are logically identical. To me this is merely a "Machtspruch." If any two things are logically distinct, it seems to me that the jussive and optative uses of the Latin subjunctive are distinct. The one is, " Drop that gun! " the other is, " Ihope you'll haveapleasant journey." The one implies authority and control on the part of the speaker; the other just as distinctly implies that the speaker's power does not exist and that he throws himself entirely on the outcome of fortune. i Amer. Jour. Phil., XVIII, 166. 124 Charles E. Bennett To me, therefore, it is clear that Kroll has not suceeded in his attempt to eliminate the optative subjunctive from Latin. The purpose of such elimination was to pave the way for the denial of the other of the two uses of the Indo-European optative in Latin, viz., the so-called "potential" or "should-would" use. Kroll dis- putes this use also, his object being to disprove, for the Latin subjunctive, the presence of any functions inherited from the Indo- European optative. There have been various assaults in recent years on the Latin potential; but these have been on the "may" potential and the "can" potential. Elmer questioned the existence of these in an elaborate discussion contained in Vol. VI of Cornell Studies. He dealt with expressions of the type aliquis Meat, quis- piam dixerit as illustrating the "may" potential; and videos, cernas, credas; videres, cerneres, crederes as types of the "can-could" poten- tial. So far as I know, no one has accepted his conclusions. Kroll not only rejects the "may" and "can-could" potentials, but, what is entirely new, rejects the "should-would" potential as well. Let us undertake to follow his reasoning. It is best to proceed at once to the consideration of concrete examples. Kroll confines himself mainly to Plautus as representing most nearly the primitive state of things for Latin. The illustrations cited by me in my Syntax of Early Latin (and by everybody else from time immemorial) Kroll declares to be spurious. Among the passages he specifically cites are Trin. 994, qui sis qui non sis, ftoccum non interduim. This is traditionally inter- preted as meaning, " I wouldn't give a straw to know who you are, or who you aren't." Kroll says it means, "Ich mag keinen Deut dafur geben, zu wissen wer du bist," etc. In other words, he rejects a natural and easily explained interpretation in favor of an unnatural and unaccountable one. How can interduim mean "Ich mag geben" ? Can amem mean "Ich mag lieben" ? Or earn, "Ich mag gehen"? The same applies to his interpretation of quod tibi sua- deam, suadeam meo patri (Capt. 237). In Bacch. 149 we have barathrum, ubi nunc es? ut ego te usurpem lubens! I interpret this as meaning, " O bottomless abyss, where are you now ? How I would gladly use you!" Kroll contents himself with the observation that my interpretation is "unverstandlich." I fancy he takes it as a purpose clause. This might be possible, were it not for lubens. Kkoll on the Independent Latin Subjunctive 125 With lubens it seems to me my interpretation is not only possible but necessary. Aul. 569 reads: potare ego tecum volo::non potem ego quidem hercle. Kroll says this means "I will not drink with you." In other instances hitherto ordinarily taken as potentials, Kroll finds a pure future meaning. No one hitherto has had the boldness to recognize the pure future meaning in the independent uses of the Latin subjunctive. Hale had recognized it in certain types of dependent clauses — his anticipatory subjunctive. But he has been followed by very few. Kroll, however, is ready to see it in inde- pendent sentences, e.g., True. 907, numquam hoc uno die efficiatur opus, which he takes as meaning "this job will not be finished in a single day" — precisely as though Plautus had written efficietur. So in Amph. 1060, nee me miserior feminast neque ulla videatur magis. Asin. 558 reads: edepol virtutes tuas qui nunc possis conlaudare, sicut ego possim ? " How would you be able to commend your merits, as J would ?" Kroll dismisses this with the remark that the "Mog- lichkeit" is in the meaning of possim, not in the subjunctive. If that is so, why doesn't Plautus use the indicative ? Other examples that assert plainly their potential ("should-would") character, but are rejected by Kroll, are Aul. 489, hoc quis non credat abs te esse ortumf Eun. 460, ex homine hunc natum dicas? Pseud. 290, egon patri subripere possim quicquam? If these are future indicatives in sense, I fail to see why Latin employed the subjunctive. A special class of cases is recognized by Kroll in velim, nolim, malim. In velim adeas, velim progrediri, etc., it is clearly impossible to take the velim, nolim malim as futures. Kroll therefore has recourse to the view advanced some twenty years ago by Morris, viz., that we have in velim veniat a species of attraction — not like what we ordinarily call attraction, where the subordinate verb is attracted to the main verb, but the reverse, where the main verb is attracted to the subordinate. Now the odd thing is that we have nothing else like this in Latin or — so far as I know — in any other language. Kroll, following Morris, would have us believe that volo veniat becomes velim veniat owing to the veniat, or, as he puts it, "Die ganze Vorstellung liegt in der Sphare des Wunsches und das fuhrt dazu dass man das den Wunsch ausdruckende Verbum in den Modus des Wunsches setzt." Very well. Why then in edico abeas, praecipio 126 Charles E. Bennett fatearis, moneo desinas, jvheo custodies, and literally hundreds of similar expressions in Plautus, do we never find an edicam abeas, praecipiam fatearis, moneam desinas, jubeam custodias t The situa- tion is precisely similar. There is also another serious objection to the Kroll-Morris view. In Plautus velim occurs with the simple subjunctive some 14 times. Against these 14 instances, we have 27 other instances of velim unaccompanied by the subjunctive, including 10 with the infinitive, 7 with participles and adjectives, 4 with a direct object, and 6 absolute uses. As for malim, that is found twice with the subjunctive. Against these 2 instances there are 21 others where there is no accompanying subjunctive, including 13 with the infinitive, 4 with participles and adjectives, 1 with a direct object, and 3 absolute uses. Nolim does not occur with the subjunctive, but is found twice with the infinitive and once with a participle. Out of 70 occurrences, therefore, of velim, nolim, malim, pervelim, 54 are not accompanied by the subjunctive. Granting now that attraction may account for the 16 cases of the subjunctive, how are we to account for the subjunctive in the 54 remaining cases? Kroll, following Morris, accounts for them by analogical extension, i.e., we are asked to believe that aquam velim, lapidem velim, malim vincere, emortuom me malim along with ut velim, malim quidem, istuc mavelim, all stand in the subjunctive by analogical extension. I can see no basis for such analogical extension; aquam, lapidem, istuc do not represent equivalents of dependent clauses. More- over, the whole theory is based on the assumption that velim with a dependent subjunctive is historically earlier than aquam velim, lapidem velim, ut velim, istuc mavelim, etc. There is nothing to show that this is true. In fact, it seems intrinsically improbable. Kroll is satisfied when he declares "velim ist hochstens Hofiichkeits- ausdruck fur volo." But I can see no "Hoflichkeit" in lapidem velim; or in ut velim, in Mam diruptam velim, me pervelim sepultam, malim amicos mersos, or dozens of other similar expressions that might be cited from Plautus. I see a distinct potential force — a "should-would" meaning: "I should like some water," "I'd like to have a rock," "I should like to be dead and buried," "I should like to have her knocked into a cocked hat," "I should prefer that." Keoll on the Independent Latin Subjunctive 127 At times, to be sure, some of these velim's, nolim's, malim's do become "Hoflichkeitsausdrucke," and it is simply because of the potential force that this "Hoflichkeitston" arises, precisely as in English our "I should say," or "No one would believe this," is more courteous, because more qualified and guarded, than a bold "I say," "No one believes." I pass to the examples of the imperfect subjunctive. Of this usage I have cited for Early Latin some 50 examples, all of which seem to me clear potentials. Some of these are of vellem, mavellem. In view of what he has already urged in connection with velim, nolim, malim, Kroll feels justified in dismissing these with a reference to his previous discussion, e.g., Asin. 589, vellem habere; Poen. 1107, magis quam vellem; Amph. 512, experiri istuc mavellem. Examples of potentials other than vellem, mavellem are more frequent, but these also are rejected by Kroll, e.g., Bacch. 313, ibidem piiblicitus serva- rent : : : occidistis me; nimio hie privatim rectius servaretur. Kroll says that this is a wish, despite the obvious absurdity of the interpretation and the well-known fact that utinam is always present with the imperfect subjunctive. Take again Poen. 1139, hodie earum muta- rentur nomina, which Kroll translates "Heute sollten ihre Namen geandert werden," where German "sollten" is the equivalent of "were on the point of being." Kroll says the mutarentur is volitive, but I submit that "sollten" is not volitive in his German rendering. Take again Men. 160, en edepol ne tu esses agitator probus. "You'd make a good charioteer," which, Kroll declares, is equivalent to si agitator esses, esses probus. Suppose it is: we shall have to account for the subjunctive in apodosis. But of that later. Kroll addresses himself also to the perfect subjunctive. Of this tense I have cited some 35 examples in independent potential uses for Early Latin. Kroll contests all of these, citing a few special examples, which he evidently feels are typical of all. He begins by declaring that formations like faxim, ausim, empsim, negassim, pro- hibessim are not perfect subjunctives, defending himself by reference to Liibbert's work of 1870. Lubbert did take this position, but he was no comparative philologist — not even in his own day — half a century ago. Today, I think, no person recognized as entitled to speak with authority denies that these formations are just as much 128 Charles E. Bennett perfect subjunctives as fuerim, dixerim, ceperim. So far as I can see, both types of formation are used in Plautus in precisely the same sense, and that sense is a true potential, e.g., Poen. 1091, malefaxim lubens, I render this, "I would gladly injure him"; Kroll renders it, "Ich mochte ihm tables tun," and adds that "es zeigt unver- kennbare voluntative" value. He uses "voluntative" to cover both volitive and optative uses, just as he identifies the two uses them- selves. But it isn't the faxim that means "Ich mochte tun," it is the faxim lubens together. If faxim is a voluntative use — using the word in the same sense in which Kroll employs it — then faxim lubens would mean "I should like to do him an injury gladly" — a patent absurdity. His interpretation of Aul. 494, ego faxim mult sint viliores Gallicis cantheriis, is either "Ich mochte bewirken" or "Ich werde bewirken." Trin. 221, pauci sint faxim qui seiunt quod nesciunt, he asserts, can mean only, "Ich werde bewirken." M.G. 316, non ego tuam empsim vitam vitiosa nuce, is, "Ich mochte nicht eine taube Nuss fur dein Leben geben"; while Asin. 503, haud negas- sim, is, "Ich mag es nicht leugnen." Note here that "mag" is equivalent to "kann," so that Kroll here adds a new type of potential to those already recognized. For myself, I feel that in all these cases the "should-would" potential meets every demand of the sense and that nothing else does. All the faxim's mean " I would bring it about," "I'd warrant that"; the empsim means "I wouldn't give (a rotten nut for your life) " ; the hand negassim, " I should not deny " ; and similarly with the 30 other instances recorded for the early period, including such examples as And. 203, ubivis facilius passus sim quam in hac re me deludier; Ad. 443, haud cito mali quid ortum ex hoc sit; M.G. 11, haud ausit dicere. To abandon the ordinary potential interpretation in these seems to me to reject the easy and natural in favor of the difficult and artificial. Kroll does not discuss the eight instances of the pluperfect sub- junctive in Early Latin. He seems to me disingenuous in ignoring them. Examples are Stich. 590, haud maligne vos invitassem; Eun. 667, ilium conclusissem neque commisissem ut, etc.; Ad. 397, non olfecissem ? Kroll's last type of usages embraces conditional sentences. Here, if anywhere, one would say that Kroll's irresponsible methods would Kroll on the Independent Latin Subjunctive 129 find the path closed. In a sentence like si sciat, suscenseat, or si plus perdiderim, minus aegre habeam', or si id esset mihi, pollicerer, the potential character is so clear that it seems impossible to deny it. Most persons go farther, of course. For most persons believe that the simple sentence is earlier than the compound, and hold that the poten- tial use found in the apodosis of conditional sentences existed in independent uses unaccompanied by any protasis. It is interesting to note how Kroll meets the difficulty in dealing with these subjunctive apodoses. The apodosis, he says, is simply attracted to the mood of the protasis — another instance of the main verb being attracted to the subordinate. When once attracted to the subjunctive pro- tasis, the subjunctive apodosis does, according to Kroll, become potential, and he expresses himself as content to allow the name "potential" to be given to subjunctive apodoses provided no claim is made that the subjunctive naturally had this value in apodoses. In other words, the Latin language, according to Kroll, did not naturally possess the mechanism for expressing the potential idea in the apodoses of conditional sentences. It evolved that mechanism by two steps: (1) by formal attraction of the mood of the apodosis to the mood of the protasis; (2) by arbitrarily attaching to this formally attracted subjunctive the meaning required. In other words, the Romans had the idea, but not the means, of expressing it. This is a novel theory as to language, and especially would it be strange if the Romans lacked a means of indicating an idea so universally prevalent among civilized peoples as that repre- sented by the " should-would " potential. I repeat that no such attraction as is assumed by Kroll is known to us elsewhere in Latin or Greek. Nor do I know of its occurrence in any modern language. It seems unnatural. If we find the sub- junctive in the apodosis of a conditional sentence, we naturally expect it to possess some function characteristic of the use of the subjunctive elsewhere in independent sentences. In other words, the function appearing in a subjunctive apodosis in Latin may fairly be taken as representing a normal and legitimate function of the sub- junctive mood. Moreover, we have subjunctive apodoses from the earliest period combined with indicative protases, e.g., Aul. 747, si istuc jus est, luci deripiamus; Rud. 744, tanta esset, si vivit; and 130 Charles E. Bennett elsewhere. Kroll conveniently ignores these. But they are fatal obstacles to his views, unless they be explained as analogical exten- sions and historically posterior to the type si venias, videos. Kroll's theory can hardly be called illuminating. He solves no problems. Nothing is made clearer; nothing is made easier. Instead he does violence to the language by the forced interpretation he puts on scores of passages — interpretations adopted for the sole purpose of maintaining his theory that the Latin subjunctive is solely "voluntative" and future. Besides this violence of interpre- tation — a violence which involves departure from the traditional interpretation of centuries — Kroll is guilty also of much arbitrari- ness, as exemplified by his identification of the " wish " use with the jussive, while his theory of attraction in his discussion of velim, nolim, malim, and, later, of the apodoses of conditional sentences, is posi- tively fantastic, involving among other things highly improbable sequences of historical development. In the case of velim, nolim, malim, it involves the employment of these with the subjunctive before they were employed in any other way. In the case of the apodoses of conditional sentences it involves a use in compound sentences which Kroll denies for simple sentences. Lastly, Kroll's whole theory is gratuitous. It is unnecessary. Why should we hesitate to recognize "wish" uses and potential uses in Latin and to recognize them as a legacy of the Indo-European optative ? The optative is there, clearly manifest by its morphology. What possible reason there can be for refusing to believe that it brought with it functions vindicated for Indo-European by San- skrit, Iranian, Greek, and Germanic, is obscure to me. This reluct- ance is particularly strange in the case of the perfect subjunctive, where the Indo-European optative forms naturally exhibit the "wish" and potential meanings. Plautus and Terence show some 35 instances of the potential use of the perfect, and at least 50 of the "wish" use. Kroll is also repeatedly guilty of a strange error in his under- standing of what is meant by the optative use. It is bad enough for him to identify the true optative or "wish" use with the volitive or "will" use, as exemplified by the hortatory and jussive. But he goes farther and confuses the "wish" use with the function of Kroll on the Independent Latin Subjunctive 131 desiderative verbs. He has no compunctions about interpreting earn as "Ich mochte gehen," and calling it voluntative (or optative). But that is not a voluntative or optative use; it is desiderative. If earn meant "Ich mochte gehen," eas would mean "Du mochtest gehen," and eat, "Er mochte gehen"; as a voluntative earn could mean only, "May I go"; eas only, "May you go," etc. Kroll's error is not new; it has often been pointed out. Cornell University