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 byJSTOR. 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.istor.org/participate-istor/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 WOED-GROUP ACCENT IN LATIN HEXAMETER By Albert Granger Harkness In the first part of this discussion [Class. Phil. II, pp. 51 fp.) I have attempted to show that in certain kinds of metre the first of two short syllables forming the integral part of a foot receives the accent. In Saturnian verse the rule is invariable. In comedy it presents comparatively few exceptions. It has been found to prevail in the second and third feet of hexameter in the types of words already examined. This paper will be devoted mainly to a consideration of the relation of the accent to the ictus in those cases which involve a dissyllable with the first syllable short, or a monosyllable, i. e., words which may lose their accent when they form a part of a word -group. In considering word-group accent, which comes under the more general head of sentence-accent, I wish to make two citations, the truth of which I trust will be illustrated by my conclusions. Brugmann ( Vergl. Gramm. V, p. 976) in speaking of the problem of the coincidence of accent and ictus says that it is not a question of the word-accent but of the sentence-accent. "Ber-flcksichtigt man diesen, so schwinden viele Discrepanzen, die man frflher nicht zu beseitigen vermochte." Munro ( Trans. Camb. Phil. Soc. X, pp. 64, 382) states that "the influence of the accent is more perceptible in the Augustan and later poets than in the earlier." A sentence is divided phonetically, not into words, syllables, and letters, but into stress-groups, syllables, and sounds. A stress- group in the wider sense may have one main accent and one or more secondary accents, or this larger stress-group may be con- ceived as consisting of shorter stress-groups, each having its own accent. Thus give me the hook may be considered as one stress- group with the main accent on give and a secondary accent on hook, or as two stress-groups give me the bdok. Thus in Latin we may have festinantem dnimam as one stress-group, or festi- ndntem dnimam as two, ox f6stindntem dnimam as three." Whether ' Of. supra, pp. 69 f . [Classical Philoloos HI, January, 1908] 39 40 Albert Geangeb Haekness a given phrase shall be treated as one stress-group or more depends on the thought, and a change of stress involves a change in the shade of thought.' Brugmann says {loc. oit., p. 875) "Nur im Zusammenhang der Kede hat das Wort sein wahres Leben," and the same may be said of accent. In verse the word-group accent must be taken into account as well as the word-accent. In regard to the accent of the word-group it is to be noted: 1. That there is not a definite line of demarkation between a compound, and a word-group united under one accent.^ The formation of a compound is often favored by the existence of a word-group united under one accent.' 2. That even the most stereotyped word-groups do not exactly correspond in their use to single words. The type et Sgo is more common in comedy than genSre.* Though these word-groups approach the unity of a single word they are subject to phonetic laws which do not apply to individual words. ^ 3. That there is not a definite line of demarkation between a word-group united under one accent, and a similar combination not so united. Thus in the middle of the verse in comedy we have matr6m meam and mdtrem m^am.^ In other words, these groups differ in regard to the unity to which they attain, and the same combination of words may be treated in one case as a word- group united under one accent, and in another, each word of the combination may retain its accent. This fact, which is sometimes overlooked by the modern metrician is recognized by Priscian. After referring to such words as iurisperitus, legislator, etc., he says (6fr. L. II. 183) : possunt tamen haec eadem etiam separata esse, si diuersos accentus his dederis uel interponas coniunctipnes. In some cases we can trace the change in accent in a combination of words, each of which has its independent accent, to a word- group united under one accent, and finally to a compound.' 1 Sievers Phonetik^ p. 235, §630. 2Gf. etenim, et ita Bursian's Jahresber. CXXX, p. 163; Stolz Hist. Oramm. d, lat. Sprache, p. 108. 3 Brugmann op. cit. W, p. 38. 5 Brugmann op. cit. 1 2, pp. 265, 877. * Bursian's Jahresber. CXXX, p. 163. « Bursian's Jahresber. OXXX, p. 167. 7 TAP A. XXXVII, p. 189, note; Corssen Ausspra^ihe 11^, p. 887. The Word-Accent in Latin Hexameter 41 There is an unbroken gradation in word-groups from those which are united under one accent and can with difficulty be dis- tinguished from compound words, to those which are more loosely connected and which appear almost as independent words. Some classes of these combinations appear as a rule united under one accent; other classes are on the border-line, and are as frequently united under one accent as they are accented as individual words ; others regularly retain their separate accents and are only in exceptional cases treated as word-groups.' To the class of word- groups which are almost as fixed in their accent as compound words belongs the preposition with its case. This combination is treated in verse as one word rather than as two. Thus a foot may end with inter followed by its case as inter se, though it could not be formed by a spondaic word which did not form a part of an unmistakable word-group. Word-groups are also freely formed by adjective and noun, subject or object and verb, especially the verb followed by a pronoun as object.'' It is how- ever to be noted that under the influence of the verse rhythm word-groups may sometimes be formed by combinations of words not intimately associated in thought. Even compounds may be formed by words which are not grammatically connected.' 1 This difEerence in word-groups may be easily illustrated by the English. In the groups good-looJdng, bill-of-fare, the union is more complete than in far-seeing, bill-of -goods. 2 Norden Aerwis VI, p. 425. 4 ; Meyer Beobachtung des Wortaccentes, p. 49. 'Brugmann op. cit. II*, p. 50, Professor Eadford, who has contributed so much to our knowledge of the recession of accent in word-groups, seems to me to limit the application of the principle more than the facts justify. He regards the recession as inseparable from the traditional word-order {Am. Jour. Phil. XXV, p. 258; TAP A. XXXVI, p. 1S>7, n. 1). He objects to the word-group uoluptfis mea because the more usual order is mea uoluptas, yet he admits the accent operSm datis and tuSm rem though the word-order is in these examples far from invariable ; Am. Jour. Phil. XXV, p. 258; TAPA. XXXVI, p. 191. In regard to the use of matr6m meam and the avoid- ance of patr6m meum at the close of the line see Bursian's Jahresber. OXXX, pp. 167 f . Again Radford admits. Am. Jour. Phil. XXV, p. 259, n. 2, that "the word-order sub- ject + verb must be recognized as sometimes causing recession in Latin." Of casual combinations involving the enclitic he says (ibid., p. 155), " Wholly exceptional, how- ever, is the peculiar Greek usage which is seen, for example, in the recessive accent of a casual combination like &v$pwir6s Tives; " but, ibid., p. 266, he says in reference to tribrach and anapaestic groups : " recession has usually occurred even in purely casual combinations." Thus we see that while Eadford is dealing mainly with combinations of words which have been stereotyped into fixed word-groups, he admits that there are combinations which we may call possible word-groups. 42 Albert Granger Bareness Such word-groups as matrem meam present the same ques- tion in regard to the accent, as arises in the case of the insepar- able enclitic. Is the accent mdterque or matSrque? Both meth- ods of accentuation find vigorous defenders, and a sufficient array of facts and arguments can be found to justify the recognition of either accent.' While Schrader shows in regard to -ne in Plautus, the type hdchie is far more common than hocine, the types memi- nistin, uiMn, datisne are employed with almost as much freedom as meministin, uiden, dpusne. The same characteristics are found in Plautus and Terence in compounds formed with the other inseparable enclitics ; and what is true for comedy is true for other forms of verse (cf. supra, p. 71). We must accordingly conclude that words compounded with an inseparable enclitic and having the penult long, have the hovering accent (schwebende Beton- ung).^ Without the recognition of this accent, modern verse would be unintelligible, and the fact that it has not been recog- nized in Latin verse has led to contradictory theories." As the accent of miitrdmque is hovering, or nearly equal on the two syllables, so there is hovering accent in md,trdm suam, and the verse is free to recognize either of these two accents. In modem languages words with hovering accent usually have a lighter accent than other words of like emphasis. The same may be assumed for Latin. Accordingly the fifth foot of the hexameter 1 Schrader De particularum -ne, anne, nonne apud Plautum prosodia (Triibner, 1885), pp. 5, 6, 10, 12, 20, 22, 26. 2 At the period when words of the type illius were passing from illius to ilUus the accent must have been hovering, i. e. nearly the same in stress on the first two syllables. In Terence's time the hovering accent is found in the type w w », is , it is sometimes i w w a; and sometimes w i w s . 8 In old English hovering accent was more common than in the English of today, but the present pronunciation of iidvirse will illustrate the principle. Definite rules have been formulated for the use of this class of words. For example Alex. Schmidt Shakespeare-Lexicon, p. 1413, says : " Dissyllabic oxytonical adjectives and participles become paroxytonical before nouns accented on the first syllable." Compare "Can pierce a complete bosom" with "He is complete in feature and mind." Bridges (Miltmi's Prosody, p. 54) says of this class of words, "They must of course have acquired their shifting pronunciation before the poet could have relied on their obey- ing the verse stress." Many words in Chaucer especially those of Romance origin, vary in accent, as hdnour, hondur. Almost any of the poets of Germany or England will illustrate the use of hovering accent. For references to the subject see, for example, " schwebende Betonung " in the indices of Paul Orundriss d. Germ. Phil. II. 2 2, and Schipper Bnglische Metrik. The Woed-Aocent in Latin Hexameter 43 is rarely formed by a final followed by a pyrrhic, just as the word type - ^ - _ a: is avoided because only a secondary word- accent would occur in the fifth arsis. When a pyrrhic word forms the fifth thesis in hexameter it is regularly preceded by a mono- syllable. The sentence -accent of a monosyllable is more marked than the final of a polysyllable. In the former the word-accent is reinforced, in the latter the place of the accent is changed; the accent is distributed over two syllables." The difference between the accent of the second and third feet, and the fifth and sixth feet of hexameter resembles the difference between the accent at the close of pentameter and of hexameter. The hexameter usually closes with a word of marked accent; sentence-enclitics are rare in this position. Of the 408 lines in Catullus Ixiv, 402 lines end with a noun, adjective, or verb (Nor- den op. cit, p. 391). The pentameter, on the other hand, fre- quently closes with sentence-enclitics, such as mens, fuit, etc. The close of the iambic verse also resembles in its structure the second and third feet of hexameter, and the end of the pentameter. The word-group accents which here form so prominent an element of the verse structure are based on natural tendencies of the language, tendencies which are brought out and are reinforced by the verse rhythm. It would be a strange and inexplicable phe- nomenon if iambic verse, which evidently strove to reproduce on the stage the living language of Kome, should in the most important and conspicuous part of the verse so frequently disregard this principle. I shall consider the following subjects: I, pyrrhic words in the second and third feet, (a) in Ovid, (h) in Lucan, (c) in Virgil; II, the relation of the accent of the fourth foot, when formed by a spondaic word, to the preceding monosyllable, (a) in Ovid, (6) in Virgil; III, the fourth foot when formed by two monosyl- lables, (a) in Ovid, (&) in Virgil; IV, monosyllables in the sec- ond and third feet, (a) in Ovid, (6) in Virgil; V, monosyllables in the sixth foot; VI, iambic words in the second and third feet, (a) in Lucan, (6) in Ovid and Virgil. 1 The accent of the word-group in Latin is a compromise between the word-aceent and the word-group accent, as in modern languages the accent of a phrase is a com- promise between the word-accent and the sentence-accent (Sievers op cit., p. 253, §682, 44 Albeet Granger Harkness I. I wish to show that pyrrhic words are used in such a way that they will not form a word-group with the preceding word. a. In Ovid' in 61 cases the pyrrhic word is preceded by a period, colon, semicolon, or interrogation. In 48 cases it is pre- ceded by a comma which in all cases corresponds to a sense-pause. In 42 cases it is followed by a monosyllabic form of esse, as humanum genus est. Here the sentence-enclitic est forms with genus a word-group of three syllables, so that the accent of genus cannot be thrown back on the final of humanum. In 20 cases it is nisi, modo, sine, simul, tamen, neque, ita, ubi. Here the pyrrhic is not closely connected with the preceding word, but with the following word or phrase. In 28 cases it is not separated by so definite a sense-pause as in the cases mentioned above, but it is closely connected with the following word.^ In 31 cases it is not closely connected with the preceding word, but it is not in all cases closely connected with the following word ; it has, however, a special emphasis' of its own which prevents it from losing its accent and forming a word-group with the preceding word.' iFor Ovid I take the first six books of the Metamorphoses as the basis of my sta- tistics, using Merkel ; for Lucan the editions of Franoken and Hosius, references to punctuation refer to Franckeu ; for Virgil, Bibbeck. I do not include In my examina- tion the last six books of the Aeneid. 2 In 13 cases the pyrrhic is a verb followed by its subject, or by an adjectiye in agreement with the subject ; i. 215, 316 ; ii. 8 ; ill. 229, etc. In 7 cases the pyrrhic is an adjective immediately followed by its noun, as ferns hostis. In 4 cages a word inter- venes between the adjective and its noun, as ii. 21 sua fert uestigia; i. 399, v. 232, vi. 197. In 2 cases the pyrrhic is a noun closely connected with the following noun, as pater Inicus, i. 651, ii. 251. In vi. 467, 584 it is an adverb modifying the following word. 3 On the importance of emphasis in verse rhythm cf. TAP A. XXXVI, pp. 163 £E.