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. THE ACCENT IN VULGAK AND FOKMAL LATIN By Fkank Frost Abbott The publication of Weil and Benloew's Thiorie g6n6rale de V accentuation latine in 1855 marks the beginning of a scientific study of the nature of the Latin accent and of the laws which govern its incidence, and at the same time it seems irrevocably to have divided Latin scholars into two camps. So far as the nature of the accent was concerned, they laid down the dictum that "l'intensite' caract6rise l'accent moderne, l'acuite" l'accent antique" (cf. op. cit. p. 5), and this proposition Corssen accepted in the main in his Aussprache three years later, as did Langen in the Jahrb. /.' class. Phil. LXXIX (1859), pp. 44 ff., but Langen questioned the trustworthiness of the statements which the Latin grammarians make concerning the circumflex, a point which he had already raised in his monograph, De grammaticorum Lati- norum praeceptis quae ad accentum spectant (Bonn, 1857). This skeptical attitude toward the utterances of the grammarians led him in 1872 (cf. Phil. XXXI, pp. 98 ff.) to doubt their accuracy in describing the Latin accent as solely musical, and to maintain that, while it was primarily a pitch-accent, the stress-element existed in it (cf. op. cit., pp. 102, 103). To this conclusion he was also led by his theory that Latin verse furnished evidence of a desire on the part of the poets to harmonize the word-accent and the verse-ictus. Not long afterward Scholl in the Act. Soc. Phil. Lips. VI, and'Seelmann in his Aussprache des Latein, came out boldly in support of the view that the Latin accent was primarily one of stress, a theory which has been adopted by Ger- man and English-speaking scholars, almost without exception, up to the present time. Now and then an American scholar has for- sworn his allegiance to the current theory, as C. W. L. Johnson has done in the Trans. Am. Phil. Assoc. XXXV (1904), pp. 65- 76, or Schlicher in A.J. P. XXIII (1902), p. 55 n., but scientific treatises and textbooks in Germany, England, and America have [Classical Philology II, October, 1S07] 444 The Accent in Vulgar and Formal Latin 445 never wavered in their acceptance of the doctrine of the Ritschl school upon this point. 1 Equally astonishing has been the unani- mity with which French Latinists have followed Weil and Ben- loew. Even those who, like Vendryes in his Recherches stir Fhistoire et les effets de V intensity initiate en Latin, think that the prehistoric initial accent involved a stress, hold that the accent of the classical period was one of pitch only. It is easy to see the reason for this strange state of affairs. There is no lack of evidence, as there is, for instance, in the case of the Saturnian verse, but one class of evidence is fatal, as it seems to me, to the stress-theory, and another group of facts is equally disastrous to the pitch-theory, as the two views mentioned are commonly held. There seems to be no way out of the dilemma, and, consequently, scholars accept the evidence which appeals to them the more strongly, and they explain the facts of the language or the state- ments made by the Latin grammarians which are at variance with their conclusions as best they can. If the question were not in the field of pure science one might almost say that scholars accepted the national tradition, or inherited their theory of the Latin accent. One can scarcely hope to make any material additions to the evidence which has accumulated during the last fifty years, and it is my purpose not to rediscuss it in detail, but to suggest a view of the case which, it seems to me, is in harmony with the im- portant facts, and does not involve, I hope, such inherent improb- ability as to warrant its summary rejection. The evidence which bears upon the nature of the Latin accent comes from two sources. It is either drawn from the Latin gram- marians, or it is obtained from a study of the phenomena of Latin and of the cognate and derived languages. The adherents of the pitch-theory are influenced largely by the statements made by the Latin writers concerning the accent, while those who believe in the stress-theory find support for it mainly in evidence of the second sort. It will therefore be convenient at this point to state very briefly the main considerations which may be urged in sup- 1 In his Latin Language, which represents a revision of the Appendix to his Latin Grammar, and which has just appeared as this article goes to press, Bennett seems inclined to regard the arguments of Vendryes as convincing. 446 Frank Frost Abbott port of these two views. This presentation of the facts will show, I hope, that it is very difficult, if not impossible, to accept either theory in its present form, and that some via media is the only way out of the difficulty. The passages in Latin writers which deal with the matter in dispute run from the first century b. c. to the fifth or sixth century A. D. Scholl has collected one or two hundred of these, of which a large number either describe the Latin accent or permit us to draw inferences concerning it. These writers speak, for instance, of the acute, grave, and circumflex, give us rules governing their incidence in the case of many words, and all of them, with four exceptions to be discussed later, use words and phrases which are applicable to a musical accent only. But the grammarians are freely charged by the German-English school with blindly apply- ing to the Latin accent terminology which their Greek predeces- sors had used of the Greek language. Even if we bring ourselves to accept this theory of slavish imitation for a long line of Latin grammarians, how can we account for their precepts when those precepts are at variance with the Greek laws of accent ? A striking case of the sort is to be found in Donatus (K. IV, p. 371. 8 = Sch. LXIXc), who says: ergo monosyllaba, quae correptam vocalem habebunt, acuto accentu pronuntiabimus, ut falx, pix, ntlx; quae productam vocalem habebunt, circumflexo accentu pronuntiabimus, ut re's, d&s, sp6s. This same rule for the accentuation of monosyl- lables with a long vowel is given by Diomedes (K. I, p. 431.15= Sch. LXIXa) and others. In Greek, of course, some long monosyllables take the circumflex, others the acute, and the principle stated by Donatus and Diomedes must either correspond to the rules of Latin accentuation or be a pure invention of the Latin grammarians. In neither case can it be borrowed from Greek writers on accent. Obviously also the comments upon the accentuation of such word- groups as inter ea loci (cf. Sch. LXXIV6) and magister equitum (Sch LXXVIIIc) can scarcely be borrowed or adapted from Greek writers. But let us leave the professional grammarians out of account, and confine our attention to the statements which Varro, Cicero, and Quintilian make with reference to the nature of the accent. The Accent in Vulgar and Formal Latin 447 It is wise to limit ourselves to these three men because they belong to the period in which we are mainly interested, and because an examination of their remarks will bring out the strong and the weak points in the theory of those who hold that Latin writers in describing their accent have followed the Greeks. No Latin writer dealing with the subject in question shows Greek influences in a more marked way than Varro does, and the strength of the argument which the German-English school makes in this connec- tion cannot be better illustrated than by giving here his comments. He is quoted by Sergius (K. IV, p. 525.18 = Sch. Ill, Ho) as saying: in accentu materia, locus et natura prosodiae brevissime comprehensa sunt, nam materia esse ostenditur vox, et ea quidem qua verba possunt sonare, id est scriptilis; locus autem syllaba, quoniam haec propria verbi pars est, quae recipit accentum. natura vero prosodiae in eo est quod aut sursum est aut deorsum; nam in vocis altitudine omnino spectatur adeo ut, si omnes syllabae pari fastigio vocis enuntientur, prosodia sit nulla, scire autem oportet vocem, sicut omne corpus, tres habere distan- tias: longitudinem, altitudinem, crassitudinem. longitudinem tempore ac syllabis metimur; nam et quantum temporis enuntiandis verbis teratur, et quanto tempore modoque syllabarum unum quodque sit verbum, pluri- mum refert. altitudinem discernit accentus cum pars verbi aut in grave deprimitur aut sublimatur in acutum. crassitudo autem in spiritu est, unde etiam Graeci adspirationem appellant [Saa-uav et tj/iXrjv]; nam omnes voces aut aspirando facimus pinguiores aut sine aspiratu pronuntiando tenuiores. Varro's words here and elsewhere (K. IV, p. 531.10 = Sch. XXXVI; p. 532.16= Sch. XL VII) are clearly applicable to a musical accent only, and it seems hardly probable that a man of his acumen would have brought over from Greek a characterization of the Latin accent which had no applicability to it, nor does it seem likely that he, of all the Romans, would have failed to appre- ciate its stress-character, but his remarks about "crassitudo" point to a Greek source, and therefore what he says concerning accent is not entirely free from suspicion. No such suspicion can attach to Cicero's statement in the Orator 57. It is so simple and inci- dental in its character that it cannot be thought of as reflecting the system of any Greek grammarian. It runs as follows: 448 Frank Frost Abbott est autem etiam in dicendo quidam cantus obscurior .... sed ille, quern significat Demosthenes et Aeschines, cum alter alteri obicit vocis flexiones, in quo illud etiam notandum mihi videtur ad studium perse- quendae suavitatis in vocibus; ipsa enim natura, quasi modularetur homi- num orationem, in omni verbo posuit acutam vocem nee una plus nee a postrema syllaba citra tertiam; quo magis naturamducem ad aurium voluptatem sequatur industria. In the early part of this passage Cicero, as Lejay correctly observes [Rev. crit. XLIII, p. 293), is talking of the sentence- accent in oratory, but at in quo he digresses for a moment to speak of the word-accent, whose musical character is clearly indicated by the words quasi modularetur. As Lejay remarks, "On ne peut imaginer une description plus claire et en meme temps une allusion plus naturelle et plus degag^e de tout esprit de systeme comme de toute reminiscence d'une doctrine 6trangere." In a well-known passage (Inst. or. xii. 10. 33) Quintilian is discussing the Latin accent and remarks: sed accentus quoque cum rigore quodam, turn similitudine ipsa minus suaves habemus, quia ultima syllaba nee acuta unquam excitatur, nee flexa circumducitur, sed in gravem vel duas graves cadit semper itaque tanto est sermo Graecus Latino iucundior, ut nostri poetae, quo- tiens dulce carmen esse voluerunt, illorum id nominibus exornent. If in using the word rigor Quintilian had had in mind the harsh effect of a stress, as compared with a pitch-accent, an interpretation of the word which some people give, he would have employed, as Lejay observes, some such term as asperitas or impetus. He has been talking of the incidence of the accent in Latin, and rigor aptly indicates the inflexibility of the laws which govern it — an inflexibility which results in monotony, similitudo. 1 If this interpretation is correct, another significant thing about the pas- sage is the fact that Quintilian, while telling us how the Greek and Latin accents differed, so far as the laws governing their fall is concerned, says nothing of any difference in their quality, a difference far more distinctive, had it existed, than the one which he notes. We have looked at the testimony of the grammarians. Let us now examine briefly such phenomena of Latin and the cognate 1 Radford (A. J. P. XXV, p. 265 n.) understands rigore in the sense of "(compara- tive) inflexibility (of pitch)." The Accent in Vulgar and Formal Latin 449 languages as throw light upon the nature of the Latin accent. Among these facts three important points go to support the pitch - theory. These are: (1) the fact that the Indo-European accent was presumably musical, at least in the last period of the parent-speech, as in the earliest Greek and Sanskrit; (2) the treatment of quantity and word-accent in classical poetry; and (3) the presence of the ictus-stress in verse. The first point lends little support to this theory, unless with Pedersen (Kukri's Zeit. XXXVIII, pp. 338 f.; XXXIX, pp. 232 f.) we deny the existence of a prehistoric stress-accent. The uniform preservation of long vowels in classical poetry, however, must be taken into considera- tion. While we are not justified in asserting that the coexistence of a stress-accent and long unaccented vowels is impossible or without parallel, it is nevertheless true that in modern languages, like English, German, French, and Modern Greek, the stress- accent has led to the disappearance of quantitative differences in the unaccented vowels. Even on the assumption that the stress- element in the Latin accent was a slight one and that professional literary men made a conscious effort to preserve the purity of the language, the fixed character of the vowel quantities and their strict observance in the poets is difficult to explain. In classical poetry the existence of the word- (or sentence-) accent and the verse-ictus side by side raises another perplexing question, if the accent was mainly a stress. Some of those who hold the stress- theory believe that the word-accent was not heard in verse, or, what amounts to the same thing, that it was not heard when it failed to coincide with the verse-ictus; others, that it was heard, but that a certain degree of harmony between them was aimed at, while still others maintain that words had their usual prose-accent in poetry, but that no attempt was made by the poets to bring about a coincidence of the accent and the ictus on the same syllable. This difficulty disappears if the word-accent was primarily a pitch. If we turn now to the stress-theory we find only four Latin grammarians, all of them belonging to the fifth or sixth century A. D., referring to the stress element in the accent, but the facts of the language lend a strong support to this view of the case, and it is clear that if the changes which the Latin language underwent 450 Frank Fkost Abbott presuppose a stress-accent, no statements made by Latin writers can stand against this evidence. Almost all scholars, even those who maintain the pitch theory, 1 believe that the Latin language in the prehistoric period had an initial stress-accent. The Romance languages have a stress-accent. Consequently, to hold that the accent of the classical period was primarily musical makes it neces- sary for us to assume a change from stress to pitch and back to stress again. Furthermore, the change in the incidence of the accent from the earlier to the later period can be easily under- stood under the stress-theory, because it would mean nothing more than "a usurpation by the secondary accent of the prominence of the main accent" (cf. Lindsay Lat. Lang., p. 159), as can readily be seen in the case of such words as sdpidntia, sapiintia or tSmpestdtibus, tdmpestdtibus? But the evidence which the Latin language furnishes that the accent was primarily one of stress is found in the historic, as well as in the prehistoric, period. It consists mainly in vowel-shorten- ing, especially in the iambic combination, in the reduction of vowels, e. g., sinatus for senatus (App. Probi IV. 198 K.) and in such cases of syncope as vernaclus (loc. cit., p. 197). The facts have been brought together in a convenient form by Lindsay, Sommer, and Stolz, and need not be repeated here. Pedersen, it is true, maintains (Kuhn's Zeii. XXXIX, pp. 232 ff.), in speaking of vowel-reduction, that "es ist sofort klar, dass fast alles, was man als zeugnis fur einen intensitats-akzent des Lateinischen angefuhrt hat, im gegentheil fur die musikalische art der betonung spricht," but as Ahlberg acutely observes (Studia de accentu Latino, p. 13) the musically accented Greek language does not show the vowel changes which we observe in Latin. Whatever may be thought of Vendryes' brilliant explanation of syncope and vowel-reduction, vowel-shortening in Plautus presents a problem which he has not been able to solve. 1 Cf. Vendryes, Becherches sur Vhistoire et les effete de I'intensitS initiale, pp. 41, 42. 2 Vendryes believes (pp. cit., p. 100) that in the prehistoric period a word had both the initial stress-accent and the pitch, inherited from Indo-European, that stress and quantity being brought into conflict, the stress-accent disappeared, but as Solmsen remarks (Ar.f. lat. Lex. XIII, p. 137), "es halt schwer, sich vorzustellen, dass eine sehr intensive Betonung der ersten Silbe vollig aufgegeben worden sei zu Gunsten einer Betonung anderer Silben, der keine Spur von Intensitat innegewohnt habe." The Accent in Vulgar and Formal Latin 451 This brief r6sum6 of the evidence has, I hope, brought out the fact that each of our current views concerning the Latin accent is highly unsatisfactory, if not quite untenable, and that some via media must be sought. The solution may be found, I think, in the theory that in the sermo plebeius the accent was primarily a stress from the earliest period to the latest, but that in the literary language during the classical period the stress-element became secondary to the pitch. To this conclusion both the statements of the grammarians and the facts of the language point. Let us examine the theory just mentioned in the light of these two classes of evidence, taking up first the direct evidence coming from the grammarians. It is a noticeable thing that Latin writers are mainly interested in formal literature, 1 and that when the grammarians discuss the Latin language they have in mind not the speech of everyday life, 2 but the form which it takes in literature ; and when they speak of an inflectional form, an idiom, a syntactical construction, or a matter of pronunciation, instead of coining a phrase, or borrowing one from ordinary speech for the purpose in hand, they quote from a writer of prose or of verse. 3 This is equally true of their remarks concerning the accent, as could be shown by many illustrations. One or two will suffice. Thus Quin- tilian says (Inst. or. i. 5. 27): nam cum dico circum litora, tamquam unum enuntio dissimulata distinctione, itaque tamquam in una voce una est acuta, quod idem accidit in illo "Troiae qui primus ab oris." 1 This contempt for the more popular forms of light literature is well illustrated by the failure of Tacitus to mention the Satirae of Petronius when he summarizes the life of Petronius (Ann. rvi. 18). In fact there is no reference to Petronius' novel before the fourth century a. d. 2 Those grammarians who, like Probus in his Appendix or Consentius, mention vulgar words or phrases do so only to warn their readers against using them. None of them shows either a scientific or a curious interest in any usage but that of formal literature. Of. also on the practices of the grammarians in such matters Consentius V. 391. 16 f. K. sWhen Zielinski ("Das Clauselgesetz in Ciceros Reden," Phil., Suppl. Bd. IX), in attempting to prove that the fall of the accent in ordinary speech differed from that in literature, is forced to assume (pp. 829 f.) that the grammarians in this matter are giving rules for the sermo cotidianus and not for the sermo urbanus, he makes an assumption which those who are familiar with the practices of the grammarians can scarcely accept. 452 Frank Feost Abbott Priscian similarly illustrates a statement about the accent of certain interrogative and relative words by citations from the poets (XV. 30, p. 83, 11 H = Sch. CXXXVIIh) : ubi quod interrogativum paenultimam acuit, ut si dicam tibi est Pam- philusf relativum gravatur, ut Vergilius in I Aeneidos: saevus ubi Aeacidae telo iacit Hector, ubi ingens Sarpedon, ubi tot Simois correpta sub undis, quomodo et unde, ut Horatius in II sermonum unde et quo Catiusf interrogativum acutum paenultimum habet, relativum gravatur; Vergilius in I genus unde Latinum. We have seen that in their comments the grammarians always have literary Latin in mind. Consequently, their description of the Latin accent as one of pitch is a description of the accent in the literary form of speech, and if the literary and vulgar accent differed from each other, these writers do not help us to understand the nature of the vulgar accent, because they do not concern themselves with the sermo plebeius. The facts of the Latin language, with two exceptions already noted (p. 449), indicate the prominence of the stress-element in the accent. These exceptions are the strict observance of quantity and the existence of the word- accent and the ictus side by side in verse. In the case of both these exceptions it is significant that we are dealing with formal literature again, and with the assumption of a pitch -accent in literary Latin both difficulties disappear. The other cases in which the language was modified by the word-accent, as we have already noticed, are those of iambic shortening, of vowel-reduction, and of syncope. All of these changes point to a stress-accent, and it is noteworthy that all of them are characteristic of, if not peculiar to, popular Latin. The reduction of an iambus to a pyrrhic under the operation of the law of brevis brevians, so far as professional poetry is concerned, is practically restricted to the early dramatic poetry which reflects the pronunciation of ordinary speech. In the hexameter, and in all the other verses used by professional poets outside the drama from the time of Ennius on, the shortening occurs in two well- defined situations only, viz.: (1) in the case of words like bene, male, modo, and puta, which formed the small change of conver- sation, where the short quantity of the final vowel had become The Accent in Vulgak and Foemal Latin 453 permanently fixed; or, (2) where poets, as Lindsay puts it (Cap- tivi, p. 32), "give us a glimpse of the real everyday pronunciation by their occasional departures from the rules of classical poetry." Instances in point are abest, Lucil. 9. 29 M. ; veto, Hor. S. i. 1. 104, and rogas, Pers. 5. 134. Amateurs strove to follow the rules of conventional verse in their productions, but, in spite of their efforts, the influence of the stress-accent of everyday speech made itself felt even in their hexameters. 1 Vowel-reduction (or reduction in the quantity of a syllable) and syncope are just as clearly con- fined to the sermo plebeius. In his Latin Language, pp. 157 ff ., Lindsay has given a list of the words which have undergone one or the other of these two changes under the influence of the accent. An examination of these words shows that they fall in one of three categories. Either the change took place during the period when the law of initial stress prevailed, or in the preliterary period under the three-syllable law (this conclusion follows from the fact that the strong form does not appear in literature), or else, if the weakened form occurs in literature, it is confined to vulgar Latin. In the first group belong, for instance, efficio and biennis; in the second, curulis (from *cnrrulis), and perhaps lavare(ci. *lovdre), and in the third, such forms as noem(bres), GIL. I. 831, pinaria, Caper VII. 93. 100 K., and solia App. Probi IV. 198 K. So far as I am aware, there is not a single case where both a strong and a weak form of a word have been attested for formal literature of the classical period. This seems a remarkable state of affairs if the accent in literary Latin was primarily a stress. In contrast to literary Latin the weakening of the posttonic, and even of the pretonic, vowel in vulgar Latin during the period of the three-syl- lable law was very common, as one sees from such forms in the Pompeian graffiti as adias, habias, periat, and from baltius, brat- tia, calcius, cavia, lancia, lintium, tinia, vinia, citera, and sinatus, against which Probus warns his readers (App. Probi IV. 198 K) . 2 In the very thorough discussion of syncope which he has given in iCf. Bucheler Carm. Ep., no. 331, and Seitz Ueber die Prosodik d. Carrn. Ep. 2 Other cases occurring during the early centuries of our era may be seen in the notes to the App. Probi, by Heraeus (Ar. f. lat. Lex. XI, pp. 302 ff.), by Forster ( Wiener Stud. XIV, pp. 294 ff.), and by Ullmann Roman. Forsch. VII, pp. 188, 189. Cf. also Schuchardt Vokalismus I, pp. 424, 442, Seelmann, pp. 236, 237, and Carnoy Le Latin d'Espagne d'apres les inscriptions, p. 39. 454 Fkank Frost Abbott Bezz. Beitr. XXVI, pp. 188 ff. ("Zur Geschichte der lat. Vocal- synkope"), Ciardi-Dupre has shown (p. 199) that "wenn synko- pierte u. nichtsynkopierte formen nebeneinander stehen, letztere iiberhaupt der hochsprache, erstere der umgangs-bezw. der volks- sprache angehoren." This conclusion, which he reaches from an examination of the use of syncopated and unsyncopated forms of the same word, he finds confirmed by the sentiment which Quin- tilian attributes {Inst. or. i. 6. 19) to Augustus: sed Augustus quoque in epistolis ad 0. Caesarem scriptis emendat, quod is calidum dicere quam caldum 1 malit, non quia id non sit Latinum, sed quia odiosum sit et, ut ipse Graeco verbo significavit, ■jrep(eppovo$. It is true that the masses in listening to one of Cicero's ora- tions from the rostra would notice differences between his manner of speech and theirs, but they would have no difficulty in under- standing him, and his use of the pitch-accent, instead of prejudicing them against him, would win respect for him as a representative of the chosen class. 1 It may not be out of place to restate briefly here the theory which has been suggested in the foregoing pages and the evidence which has been offered in support of it. The main accent in the prehistoric period was one of stress and fell upon the initial syl- lable, with a secondary stress-accent upon one of the following syllables. Before literature begins the main initial stress became secondary in most words, and the primary stress fell on the penult or antepenult, as the case might be. In popular Latin the accent always continued to be one primarily of stress. Such facts of the language as iambic shortening, vowel-weakening, the reduction in the quantity of syllables, and syncope, as well as the nature of the accent in the Romance languages all point to this conclusion. In formal Latin, however, the phenomena just mentioned do not occur. On the contrary the careful observance of quantity in verse, and the presence of the ictus-stress harmonize far better with the assumption that the accent was a pitch than with the theory that it was a stress. Furthermore, Latin writers without exception in numerous passages from the first century b. c. to the fifth or sixth century a. d. in describing their accent speak of it as one of pitch. This evidence cannot be explained away on the assumption that these writers were blindly imposing upon the Latin language a Greek doctrine which did not conform to the facts, because among them are such men as Cicero and Quintilian, because some of their statements are made in quite an incidental 1 It would be quite possible for Oicero to modify his accent slightly in addressing the people, just as he makes his vocabulary and his phrases suit the topics which he discusses and the people whom he addresses in his letters. The Accent in Vulgar and Formal Latin 459 way, and because they give rules for the incidence of the circum- flex and acute accents which find no counterpart in Greek. Now when the grammarians discuss Latin they invariably have literary usage in mind, so that when they describe the Latin accent as one of pitch they are speaking of literary Latin. The development of the pitch-element in the literary accent is a natural result of the imitation of Greek, under whose predominating influence Latin literature and literary Latin developed. The difference which resulted in this way between popular Latin and formal Latin is no greater than the differences in vocabulary, syntax, and phraseology which we know existed between these two forms of speech, and is paralleled not only in modern times, but by the conditions in Greek just before the beginning of our era. It remains for us to fix the date, if possible, when the accent in literary Latin became primarily one of stress again. We may say at once that this point was reached when popular Latin sup- planted formal Latin in literature. Grober has shown (Ar. f. lat. Lex. I, pp. 35 ff.) that literary Latin may be said to come to an end in the sixth century, 1 although the date varies somewhat for different parts of the Roman world. As early, however, as the close of the fourth or the beginning of the fifth century we find clear evidence in the verses of Prudentius of the prominence of the stress element. It comes out in the quantitative treatment of such Greek words as eremus-—epr}fiopevr)