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. PREFERRED AND AVOIDED COMBINATIONS OF THE ENCLITIC QUE IN CICERO CONSIDERED IN RELATION TO QUESTIONS OF ACCENT AND PROSE RHYTHM By F. W. Shipley In two papers published last year, "The Heroic Clausula in Cicero and Quintilian," 1 and "The Treatment of Dactylic Words in the Rhythmic Prose of Cicero," 2 1 have stated briefly some of the statistical results of a study of the enclitic usage in Cicero. These two papers, though published before it, are in reality corollaries of certain portions of the present article, and had their beginning in clues furnished by Cicero's avoidance of certain combinations of the enclitic que as shown by the statistical table on p. 26. The study of this enclitic in Cicero's speeches was originally undertaken with a view to determining, by the evidence of prose, the accentuation for his day of combinations such as eaque, multaque, luminaque; but it was soon evident that not only this question but all aspects of his usage of the enclitics were so closely bound up with his style and his sense of rhythm that I was obliged to extend the scope of the investigation and to include statistics for all combinations of this enclitic. The laws of rhythmic prose, on the one hand, throw considerable light upon the accentuation of enclitic combinations, and, conversely, the enclitics throw light upon several questions of Ciceronian rhythm. In fact, the enclitic que serves as an important clue in checking the results of investigators of rhythm, and particularly of those who adhere to the mechanical scansion of longs and shorts. Itself a movable short syllable, with the power of shifting the accent upon the majority of the words to which it is appended, and of changing the quantity of the last syllable in the case of words ending in a short vowel and a consonant, it furnishes to an author desirous of producing certain effects a most convenient means for i Class. Phil. VI, No. 4. 'Trans. Am. Phil. Assoc. XLI, 1911. [Classical Philology VIII, January, 1913] 23 24 F. W. Shipley accentual and quantitative variation. Thus the enclitic que added to a trochaic word converts it into a dactyl if it ends in a vowel, e.g., multaque, and into an antibacchius if it ends in a consonant, e.g., multumque. A pyrrhic word ending in a vowel becomes a tribrach, e.g., eaque, and if it ends in a consonant it becomes an amphi- brach, e.g., eumque, etc. If the addition of the enclitic is likely to produce an awkward combination metrically or accentually, this effect may be avoided by selecting for the next word one which begins with a vowel, so that elision results. A high or a low rate of elision points to avoidance of, or preference for, certain forms. Elision is therefore an important factor which, so far as I know, has not been used in the study of the enclitic except by Harant in con- nection with the appending of the enclitic to words ending in short e. 1 The evidence for Cicero's preference for, or avoidance of, the various possible combinations with the enclitic que is set forth in the statistical table. In this table the columns marked A give the percentage of occurrence of the various combinations with que; the columns in heavy type, marked B, give the percentage of elision cases, or rather of all cases where the following word begins with a vowel; and the column marked X (the figures in brackets in the list of totals) shows the relative frequency in the language of the types of words (or word endings) to which the enclitic is attached. A comparison between these sets of figures yields some surpris- ing results, and discloses some details in Cicero's usage of the enclitic que which had escaped the writers and revisers of the larger gram- mars, of such works as Krebs Anlibarbarus, and the authors of various monographs on the enclitics. In some points they are in direct contradiction to the statements in these works. In others they furnish an explanation for phenomena which had been accounted for incorrectly. To avoid complicating issues, the statistics are confined to the enclitic que in Cicero's orations (figures for the Letters, and for other authors are given incidentally in discussing the various combinations). For the sake of tracing chronological development they are based upon the following speeches or groups of speeches, amounting in compass to two-thirds of the total bulk of the orations: (1) Pro iSeep. 39. The Enclitic que in Cicero 25 Quinct., Pro Rose. Amerino (81 and 80 b.c); (2) Divinatio in Conci- lium and the Verrines (70 and 69) ; (3) De Imperio Pompei and Pro Cluentio (67) ; (4) the orations against Catiline (63) ; (5) Pro Sestio (57) ; (6) Pro Plancio and Pro Milone (54 and 52) ; (7) the Philippics (44-43). These speeches (Mueller's text) form the basis of the statistical table. But in dealing with the rarer combinations, which show evidence of avoidance, I have given the examples from all the orations so that the records for these may be complete. In the case of these rarer combinations I have also given the statistics for the letters to Atticus in order to compare the data for the orations with the less formal correspondence. To facilitate comparison, the figures in the statistical table are reduced to a percentage basis. At the bottom of each column, how- ever, I have given the total number of occurrences of the enclitic que. In the left-hand margin of the table I have given the various types of words or word endings to which the enclitic is attached. The basis of classification is the metrical value of the last two (or three) syllables in the word. In final syllables ~ indicates a short vowel, s indicates syllables containing a long vowel, or a short vowel followed by a consonant which becomes long when the enclitic is added. For convenience of reference the types will be referred to by the number in the margin; thus: type 5 = ~ s eo(que), testimo- nium(que). In 1 and 7 (pyrrhic words ending in a vowel, and monosyllables) the type indicates words and not merely endings, for reasons which will appear later. The other five types indicate either whole words or (in the case of longer words) the word ending only. In the rarer types (2, 3, 4), where the enclitic is added to a word ending in a short vowel, the figures in column B are a real criterion only in the column of totals, since the number of examples is too small to furnish a rational percentage system. Thus, a single example in any of the groups of speeches, if elided, would give 100 per cent of elision, and if unelided, per cent. In the column of totals, however, the numbers are large enough to give reliable aver- ages in the elision column. All the figures in the table are gross figures. These reduce con- siderably in the rarer types by reason of questioned readings and by the elimination of those passages in which, as in the Har. Resp., 26 F. W. Shipley 9 2 z w « H B H cc h at £°® W la m J * So SOB n ■e C5 e 00 3 T-l 26.3 22. 62 fi cc c\ (eSsnS-j oo t^ 1 5? C? C? o 1 «H .-< in Or, B "3 n CI g co ce CO iH IO cc cc < to ■* CO CC OJ Tli K (M "5 i- T— & *4 m 0> IO IC th oo ce cr c M M M >-.o ,; n © IO 'Jl CN US o co i-i CO CM CC t^ «j il5 CD CO CC -* CO I> ill n ■o d ce> io ih ce ci o t-^ to en io ce Til O c CC tH a < r~ t^ ^ *h r- to »£ " ,-H - co •* i- »■■ o aS« BO oa ^H b S £° OS o 3 A H 1: .-5 -e ) "cc ) I -.2 = -.2 " ) | ) e > ''g| <) c CO & c > -p 5. w s: eo(que), testi- mon]ium(que) . . . 6. - s vi]dendo(que) vi]dendum (que) , etc 7. *■ monosyllables, etc., id(que), «4-l O o * -. Words ending in a trochee followed by a final word of the types confectus, cognoscite, constitute the favorite forms of two of his three most common clausulae, e.g., morte vicerunt and lustrisque confectus = ly, cessit audaciae and insolen- tQamque cognoscite =2 y. (The symbols here and elsewhere are those of Zielinski Das Clauselgesetz in Cicero's Beden.) The language normally furnishes about 25.5 per cent of words with trochaic endings (8.5 per cent ending in a vowel, 17 per cent ending in a consonant), and the enclitic, added to words of types 5 and 6, makes it possible to add materially to the ratio of trochaic endings. It may be said in passing that, where que occurs in the clausula, the elision cases fall far below the 27 . 8 per cent which is the average for enclitic combinations. For example, in the Pro Plancio and Pro Milone taken together que occurs in 36 clausulae (either in periods or at the end of members). Of these 36 examples there are but three elision cases, or 8 . 5 per cent. The preference for types 5 and 6 is thus readily explained, and we may pass now to the types which were avoided. Ha. Types 2, 3, 4 (avoided), and type 1 (first avoided and then cultivated). — These types have in common the final short vowel. They will be considered separately, however, as two elements are here involved, namely, accent and quantity. (1) Type 4> Vque; multaque, videndoque. — The rarity of this combination is attested by low percentage of occurrence, 1.3 The Enclitic que in Ciceeo 29 per cent (the ratio for the language is 8.5 per cent in column X). Furthermore, while the average ratio of elision to the total number of occurrences of the enclitic is 27 . 8 per cent, in this class the per- centage of elision rises to 57.1 per cent. In other words, even on the basis of gross figures, which are reduced by interpretation, the number of cases in which the addition of the enclitic to a trochaic word like multa forms a dactyl is but one-half of 1 per cent of the total number of occurrences of que. These figures are surprising in view of the fact that, at least to my knowledge, no attention has been called to the rarity of the combination, and modern writers of Latin do not hesitate to use it. Before attempting an explanation of the figures, I shall first give all the examples from Cicero's orations. The elision cases are marked by an asterisk (*). "CI." after an example indicates that it occurs in a period clausula. 1. Quinct. 32. In hoc summa iudicii causaque tota consistit. 2. Quinct. 47. Eo leviora infirmioraque* existimo. CI. 3. Rose. Am. 90. Multi observandi multaque servanda sunt. CI. 4. Rose. Com. Natura vitaque* est expressa. 5. Tull. 33. Pro portu stationeque* utuntur. 6. Div. 71. Nullaque* in hoc ipso iudicio expectatio futura sit. 7. Act. Pr. 8. Profligataque* esse arbitrantur. 8. Act. Pr. 47. Tantaque* eius vitia. 9. Verr. i. 48. Tantaque* eius auctoritas. 10. Verr. i. 59. Per scelus ablata ereptaque* essent. 11. Verr. ii. 7. Religiones .... omnium sacrorum fanorumque vio- latae, simulacraque deorum, quae, etc. 12. Verr. ii. 14. Mandataque* acceperant. 13. Verr. ii. 53. Spectantibus omnibus evecta exportataque* esse. CI. 14. Verr. ii. 55. At vero urbem totam templaque^ deorum. t (The MSS vary between templaque and templa.) 15. Verr. ii. 88. Multaque*, ut in primis Siculorum. 16. Verr. iii. 21. Composita comparataque* essent. 17. Verr. iv. 82. Qui cum res maximas gesserit monumentaque suarum rerum gestarum cum maxime constituet. 18. Verr. iv. 122. Sacra religiosaque* accepisset. 19. Verr. iv. 138. Cum a me auxilium communi omnium legationum consilio petebatur causaque totius provinciae ad me deferebatur. 20. Verr. iv. 1. Sacra profanaque* omnia. 21. Verr. v. 34. Votaque pro imperio nuncupasset. 30 F. W. Shipley 22. Verr. v. 83. Ubi tot tantaque* ornamenta. 23. Verr. v. 118. Cui ex omni gemitu doloreque certa merces compara- batur. 24. Agr. ii. 41. Alexandria cunctaque* Aegyptus. 25. Agr. ii. 86. Suo iure libertateque* utuntur. 26. Har. Resp. 37. Et video in haruspicum responsum haec esse sub- iuncta: sacra vetusta occultaque minus diligenter facta pollutaque (ritualistic diction). 27. Har. Resp. 57. Multaque* in eo. 28. Plane. 98. Confiteareque* aliquando. 29. Mil. 42. Quantaque* et quam sollicita. 30. Phil. v. 46. Legio Martia, quartaque^ summo studio .... (legal diction). 31. Phil. xi. 33. Summa ingenii indoles, summaque virtus f (summaque D, summa i). CI. 32. Phil. xiii. 46. Maioreque deorum immortalium beneficio (maiore i). 33. Phil. xiv. 26. Ipseque* in primis (legal diction). 34. Phil. xiv. 36. Ipseque C. Pansa consul (legal diction). We have thus 34 examples of - ~+qne in all the speeches, with 21 elision cases, leaving 13 with the value of a dactyl. It will be noted also that of these 34 cases no less than 23 occur in the speeches delivered before 68 B.C., which form but one-third of the total bulk of the speeches. In other words, the usage is rare throughout, and was avoided by elision; but it becomes practically negligible in the last two-thirds of the speeches. Of the 11 examples which occur in the speeches delivered after 68 B.C., No. 31 is corrupt, 1 Nos. 26, 27 occur in a quotation from the Har. Resp. written in religious phrase- ology, and Nos. 30, 33, 34 are from the words of Cicero's resolutions in the Senate in which he uses the conventional legal parlance. If we eliminate these, as not representative of Cicero's normal ora- torical diction, we have left in the speeches delivered after 68 b.c. but 5 examples, of which all but one (No. 32 maioreque, and here the MSS have a variant maiore) are elision cases. In the first third (estimated by pages) of his speeches we have 22 examples (No. 11 is omitted because of the uncertainty of the text), with 15 elision ■ The manuscripts vary between summaque and summa. The laws of the clausula point to summa. See Zielinski Clauselgesetz, p. 215, and my own article on the " Heroic Clausula in Cicero and Quintilian," Class. Phil. VI, No. 4, where I have shown that Cicero never, except here, uses the combination --»- s in the clausula except with a strong pause after the dactyl. The Enclitic que in Cicero 31 cases; in the remaining two-thirds we have 5 examples, with 4 elision cases. The exception is again No. 32, with the variant already re- ferred to. The increase in the elision cases and the decrease in the actual frequency point to a tendency to avoid the combination which in the later speeches amounts practically to elimination. As we are here dealing with rhythmic prose, the avoidance must be due to one of two things — accent or quantitative value. If one accepts the statement of the late grammarians that the accent was musdque dondque, multdque, this combination would fit beautifully into metrical forms involving a resolution of the first long of the cretic or molossus of the base, a resolution of the last long of a molos- sus in the base, or a resolution of the long syllables in the cadence of any of the 8 forms. 1 The avoidance of the combination is for this reason sufficient ground for denying that it ever had the accentua- tion multdque, dondque, simulacrdque, in Cicero's day at any rate. If, on the other hand, we dismiss the statement of the grammarians and assume that the accentuation was multaque, donaque, simuldcraque, according to the regular penultima law, these combinations would differ in no way from any pure dactyl, ■*■ ~ ~, of the type condere, cdrcere, crimina, dmnia, omnibus. This assumption is not valid, however, unless it can be shown that - ~ - is avoided also in the case of words of the type condere, crimina, etc. In reality these words are not used with the dactylic values by Cicero in his orations. Acting on the parallel suggested by the rarity of this enclitic combination, I made a special study of the dactylic words in several speeches of Cicero. My results for the De Imperio Pompeii have been published in the Transactions of the American Philological Association, XLI, under the title, "The Treatment of Dactylic Words in the Rhythmic Prose of Cicero." In summary they were as follows: Cicero avoids in his speeches words with dactylic values - <- ~. In practically every case dactylic words, when not elided, or when the last syllable is not long by position, are followed by pauses, either pauses of ordinary punctuation, or those caused by phrasing or emphasis, so that the pause lends to the final short the inane tempus and the words thereby take on the value of cretics - - £. The great rarity of combinations of the type videndaque, multaque 1 1 use here Zielinski's terminology. 32 F. W. Shipley is now explained. They are avoided because, inasmuch as the enclitic is for the most part a particle of continuation, they are rarely followed by punctuation or rhetorical pauses, and therefore rarely assume cretic values - - £ as in the case of words like condere, crimina. The value of multaque, etc., is therefore - — , and the accent is multaque, and not multaque. The figures for the letters to Atticus, though in a less formal style, are almost parallel to those for the orations. Examples of the type multaque, simulacraque form 1 . 9 per cent of the que combina- tions, of which no less than 70 per cent are elided and therefore remain trochees instead of forming dactyls. (2) Type 8, - ~ ~+que: omniaque, ceteraque, criminaque, indul- gentiaque, periculaque. — The enclitic is but rarely added to a dactylic word ending in a vowel when the number of syllables is increased by its addition. The total number of examples in the speeches covered by my statistics is 23, or 1 per cent. The ratio of elision, 34.4 per cent, in column B of the totals, is lower than in the case of type 4, but, as will be seen later, the increase in the number of syllables by the addition of que could in many cases be avoided by synizesis or syncope. Words of this type form 2 . 3 per cent of the words of the language. The relative rarity of the type cannot be due to quanti- tative values, since -««~ e.g., accipere, iudicia, occurs 242 times 1 in period clausulae, in a single type of resolution, viz., that of the first long of the cretic in the base (-] * ) in forms l 1 , 2 1 , 2 1 , etc. The reason for the rarity of the combination must lie in awkwardness of accentuation. But first let us consider the examples. Here again I include all the examples from the orations. Elision cases are marked with an asterisk (*). Period clausulae are indicated by "CI." 1. Rose. Am. 31. Periculaque* impendeant omnia. 2. Rose. Am. 91. Omniaque miscebant. 3. Tull. 8. Omnes leges omniaque iudicia, quae etc. 4. Div. 6. Cuius furta atque flagitia .... ante oculos omnium maxi- ma turpissimaque nossemus. 5. Verr. i. 49. Quid ego adventus eius, prandia, cenas, equos, munera- que commemorem? CI. (Asconius gives munera.) 6. Verr. i. 91. Vina ceteraque quae .... facillime comparantur. 1 See Zielinski's tables in Das Clauselgesetz in Cicero's Reden. The Enclitic que in Cicero 33 7. Verr. i. 103. Criminaque* ea, quae. 8. Verr. i. 112. In quo omnis nostra diligentia indulgentiaque con- sumatur. CI. 9. Verr. ii. 95. Omniaque* habuisset aequora. 10. Verr. iii. 183. Tabulae publicae periculaque magistratuum com- mittuntur. 11. Verr. iv. 4. Thespiadas .... ceteraque prof ana ex illo oppido signa. 12. Verr. iv. 72. Omniaque, quae ornamento urbi esse possent. 13. Verr. iv. 138. Cum .... mandata, legates, litteras, testimoniaque sumpsissem. 14. De Imp. 4. Causam rei publicae periculaque rerum suarum. 15. De Imp. 40. Signa et tabulas ceteraque* ornamenta. 16. De dom. 1. Commissa creditaque* esse videantur. 17. Har. 40. Ne caedes periculaque creentur (citation from the respon- sum). 18. Sest. 17. Scelera vulneraque* inusta rei p. 19. Vat. 6. Omniaque* ea. 20. Vat. 13. Omniaque mea tela in sic te coniicientur. 21. Plane. 41. Omniaque potius permiscuerunt. 22. Phil. v. 37. Galliaque, quae semper praesidet. 23. Phil. x. 19. Omniaque* ad eorum voluntatem. 24. Phil. x. 25. Consuetudineque rei p. 25. Phil. x. 26. Omniaque* eius consilia. 26. Phil. xi. 30. Navis nautas pecuniam ceteraque quae .... pertineant. We have in all 26 examples. These are reduced as follows: in No. 5 Asconius gives munera, which is no doubt correct, since the long series is more forceful with asyndeton. No. 17 is quoted from the haruspicum responsum, and is an example of legal phraseology. Nos. 24, 25, 26 are from Cicero's own sententiae expressed in the Senate, and are also examples of legal diction. This leaves 21 examples. Of these, 14 (with 4 elision cases) occur in the speeches delivered before the end of the year 67, and 7 (with 4 elision cases) in the speeches delivered after that date. There is therefore a marked decrease in the usage, with a considerable increase in elision, in the last half of the speeches. Both point to increased avoidance. We have already said that the avoidance cannot be due to quantity, and must be ascribed to awkward accentuation. Let us first assume the accentuation omnidque, ceterdque on the evidence of the late grammarians. The value of these combinations would 34 F. W. Shipley then be — * ~, a cretic with the last syllable resolved. Used in the clausula, the values -would be l 2 , 2 2 , 3 2 , etc. Now, when this form of base is used, Cicero is obliged to form his clausulae in the following fashion: esse videdtur, esse videdmini, etc. The resolved long is regularly formed by the first two syllables of a word, never by the last two. The reason is that under the ordinary accent law a short penultima cannot have the accent. But, according to the gram- marians' evidence, we have here an accented penultima, so that combinations with this accentuation should be very common in clausulae of the types mentioned, and particularly in the last, 3 2 , 8 where the typology of the normal clausula is - — | - — s. But, in fact, this resolution is exceedingly rare. The 14 examples given by Zielinski all have the 7 caesura, and the B caesura is not found at all. The rarity of criminaque, etc., is sufficent evidence that an accentuation so convenient for Cicero's clausulae, and particularly for the form 3 2 , did not exist in Cicero's time. Whatever the accent was, it was not crimindque, ceterdque. The avoidance is therefore unexplained. What, then, was the accent ? If we fall back upon the regular penultima law, the accent would be turpisslmaque, indulgentiaque, omniaque. With this accent these combinations would fit beauti- fully into clausulae of the types l 1 , 2 1 , 3 l , etc., with the first long of the cretic resolved, e.g., * - ~ - | - ^ , facer e conetur. Even including the end of members, there are but three examples which could be assigned to this type on the hypothesis of this accentuation: turpis- slmaque nossemus, testimoniaque sumpsissem, l l y, and indulgentiaque consumatur, S 2 1 . The examples are nearly as rare as on the other assumption. We have seen that in combinations of the type multaque the phe- nomena can be explained only on the assumption that the accentua- tion was multaque, and that the statements of the grammarians cannot apply to Cicero's time. Now, assuming the operation of the ordinary accent law, one can see the reason for the avoidance of a combination in which the accentuation would be testimoniaque as against testimonia. Such a shift of accent from the usual syllable to that which was ordinarily the weakest syllable in the word must The Enclitic que in Cicebo 35 have been felt as awkward, and this must be the reason for Cicero's avoidance. The shift of accent can of course be avoided by elision. But, if we consider the types of the unelided examples, it will be seen that Cicero was probably avoiding the shift of accent in them as well. In the 13 unelided cases (omitting those which have been eliminated, as couched in legal diction) the words to which the enclitic is attached have the following endings: •4a, 8 cases, omnia 5, Gallia, testimonia, indulgentia. -era, 2 cases, cetera. -via, 2 cases, pericula. -ima, 1 case, turpissima. Twelve examples are in cases of words capable of synizesis or syncope; one only, turpissima, is incapable of contraction. Syni- zesis in words ending in -4a needs no comment; periclum beside peri- culum is attested by the evidence of both poetry and prose; and for the syncope of -er- in ceteraque we have the evidence of altrinsecus beside atteri. 1 It is significant that the unelided examples from speeches written after the year 67 all end in -iaque. Cicero's usage may therefore be stated as follows: He avoids adding the enclitic que to words ending in a dactyl when the last syllable is formed by a short vowel, except where the shift of accent to the antepenult of the combination can be avoided either (1) by elision, or (2) by synizesis or syncope within the word to which it is appended. EEMARKS 1. That the avoidance of this combination is not a phenomenon of rhythmical prose is shown by the fact that there are only three examples in Cicero's letters to Atticus (7, 12, 2; 7, 25, 1; 9, 18, 3). They are confined to omniaque, and 2 of them are elision cases. In the philosophical and rhetori- cal works there are 12 unelided examples, as follows: ceteraque 8 cases; Tusc. Disp. 5. 100; delnv. 11; Top. 52; deFat. 33; de Nat. 3. 75; 1. 28; de Senect. 16; de Or. 1. 85; litoraque collucent, de Nat. 2. 99; cognominaque Statoris, Leg. 2. 28; reliquaque praeclare, Fin. 2. 106. One passage is corrupt, so that litoraque and reliquaque are the only examples, except the cases of ceteraque. In the other writers of the time of the Republic the unelided examples are still more rare, if we omit the cases of -iaque. Caesar B.G. vii reliquaque 1 Lindsay hat. hang., p. 172. 36 F. W. Shipley quae 19. 6; 81. 2; 84. (These three chapters of Book vii show other peculi- arities which differentiate them from the diction of Books i-vi. See "The Seventh Book of Caesar's Gallic War," Proceed. Am. Phil. Assoc, 1911.) B.C., plurimaque proelia, 1. 7; singulaque latera 1. 42. 1; tabernaculaque constitute, 1. 80. 3; proximaque respiciens 2. 39. 2; Auct. Bell. Afr., grandi- neque constituta, 47. 6; amplitudineque caesa 40. 6; oppidaque vexare 25. 4; notissimaque conciliavit; Lentulus Ad. Fam., ceteraque tot, 12. 15. 4. 1 2. A natural objection may be raised to my assumption of syncope and synizesis. If omniaque, ceteraque reduce to omnyaque, cet(e)raque, the result is the same as in the case of muliaque, videndaque, namely a dactyl, which Cicero avoids. But an examination of the 12 examples in which synizesis or syncope is postulated will show that the context demands a pause in the following instances: in Nos. 6, 12, and 22 the enclitic combination is followed by a relative clause; omniaque is emphatic in Nos. 2, 3, 20, and 21; and in No. 13 testimoniaque ends a series of four nouns. The value of the final syllables is therefore not - — but - - £. The matter is discussed in "The Treatment of Dactylic Words in Cicero," already cited. (3) Type 2, ■* ~ ~+que: positaque, mariaque, iudiciaque, etc. — The number of examples of this type is exceedingly small, two- tenths of 1 per cent in the statistical table, although words of the type posita, maria, etc., form 2.7 per cent of the words of the lan- guage. Elision cases form 40 per cent of the examples, but, as in type 3, the percentage of elision does not adequately represent the avoidance of the combination, since here also synizesis and syncope are operative. In the Latin introduction to a recent critical edition by an eminent German scholar I have found half as many examples of the combination on a single page as occur in the 1,700 pages of Cicero's orations. The writer was probably all unconscious of the fact that Cicero practically never uses the enclitic in this way. I give below the 7 examples to be found in all of Cicero's orations. 1. Rose. Am. 131. Cuius nutu et arbitrio caelum terra mariaque reguntur. 2. Div. 68. Leges iudiciaque* administrentur. CI. 3. Verr. ii. 73. Operaque daretur (quotation from a decree). 4. Verr. iii. 130. Post hominum memoriam iudiciaque de pecuniis repetundis constituta. 5. Arch. 5. Studiaque* haec in Latio vehementius turn colebantur. 6. Dam. 38. Auspiciaque populi Romani. 7. Phil. vi. 6. Merita beneficiaque* eorum (from a mandatum senatus). I I am indebted to Mr. W. H. Schlueter for the data for much of the pre-Ciceronian literature. The Enclitic que in Cicero 37 Of these, No. 1 is probably a religious formula, 1 and Nos. 3 and 7 are quotations from decrees. All three are examples of archaic religious or legal phraseology. 2 This leaves 4 examples in Cicero's own diction, of which 2 are elision cases, and the other 2, ending in ■4a, may have avoided the increase in the number of syllables, and also the shift of accent, by synizesis. In view of this, positive evi- dence for the accentuation of such combinations is lacking. 3 The absence, in Cicero's own diction, of any examples in which the number of syllables is surely increased by the addition of the enclitic shows in a negative way a reluctance to shift the accent from the syllable which held the accent before the addition of the enclitic. This points to the operation of the penultima law rather than to the accen- tuation indicated by the grammarians, since Cicero was clearly avoiding an awkward accentuation, while, if it were the rule to accentuate the syllable which precedes the enclitic, no such awkward- ness would in that case be felt. That Cicero's avoidance of combinations of this type is not confined to the speeches is shown by the fact that not a single example occurs in the letters to Atticus, which are colloquial and, for the most part, non-rhythmical. In the philosophical works there is one example (which, however, has a variant incolumem), incolumeque vellet, de Fin. 4. 19. In all the prose writers of the republican period (exclusive of Cicero) there are but 4 unelided examples which do not have the ending -iaque. Cato de Agr. 141. 2, pecuaque quae; ibid. 2. 1, operaque quae; Bell. Afr. 47. 2, opereque faciendo; Plancus ad Fam. 12. 24. 1, meritaque tua. (4) Type 1, *» ~+que: Combination with a pyrrhic word ending in a vowel: eaque, meaque, suaque, etc. — I have here omitted from consideration itaque (where it is often impossible to distinguish 1 It is possible, however, that Cicero wrote caelum terra maria, with asyndeton, as in other passages in which the same words occur. - Tusc. 5. 105, caelum terras maria; de Nat. i. 22, caelum ignes terrae parerent; ibid. ii. 77, caelum maria terrae regnantur; ibid. i. 100, caelum terras maria .... vidissent. ' In legal phraseology, in particular, que was commonly added to words of this type (see Elmer, op. cit.) . The practice may date back to a time when initial accent was still operative, and may have continued with the tenacity of legal diction after the accent laws had changed. » The evidence of the clausula does not help, since in it Cicero avoided words end- ing in four short syllables (Zielinski Clausdgesetz, p. 234). 38 F. W. Shipley between the conjunction "accordingly" and ita+que), and mihique, tibique, sibique (where the second i is probably long). 1 We have seen how, in the other types ending in a short vowel, Cicero's tendency toward avoidance grew in the later speeches, as is indicated by the decreasing ratio of occurrence and the increasing ratio of elision. With this combination the case is reversed. Rare at first, its frequency increases in the later speeches. In columns i-iv it occurs but nine times in 1,398 examples of the enclitic =0.7 per cent of occurrence, with 43 per cent of elision. In columns v-vii it occurs 20 times in 936 examples =2.1 per cent of occurrence (the percentage of occurrence in the language is 1.8), with only 15 per cent of elision cases. The combinations of this type are: eaque 17 times, elisions 5; utique 5 times (in decrees only), no elisions; suaque twice, no elision; meaque twice, no elision; viaque once, no elision; bonaque once, with elision; sineque once, with elision. Utique occurs only in the legal phraseology of Cicero's resolu- tions in the Senate in the Philippics, and all the other examples are sporadic except eaque. Cicero uses eaque with increasing fre- quency. Of the 17 examples 7 occur in the Philippics, of which but one is an elision case. Furthermore, while in the earlier speeches eaque occurs only at the beginning of a member, 4 of the 7 cases in the Philippics are in clausulae, 2 at the end of a period, and 2 at the end of a member. These four clausulae furnish the only evidence for the accentua- tion, but it is conclusive. They are : 1. Phil. i. 3. Multa praetereo eaque praeclara. V-y. 2. Phil. i. 34. Multa ex me eaque saepissime. 2'y. 3. Phil. ii. 73. Non multa eaque maculosa. I 12 '/. 4. Phil. vii. 2. Eaque defendunt. l'y. In all the accent is the same, 6aque. In other words it is the penultima law that is operative and not the rule given by the late grammarians. 1 The evidence of the clausula is in favor of mihi, tibi, sibi. See Bornecque la prose mitrique dans le correspondence de Ciceron, p. 212, and Zielinski Clauselgesetz, p. 183. The presumption is therefore in favor of mihique, tibique, sibique, although, since unelided cases in clausulae do not occur in my examples, positive evidence is lacking. The Enclitic que in Cicero 39 Let us briefly recapitulate our conclusions in regard to the types ending in a short vowel. Cicero's usage. — Cicero avoids adding the enclitic que to words ending in a short vowel (except pyrrhics) where the number of syllables would be increased by its addition. In the case of words ending in a tribrach or a dactyl the avoidance is due to reluctance to shift the accent from the usual syllable to the weakly accented syllable which follows it; in the case of trochaic words it is due to the tendency to avoid a dactyl except before a pause. The enclitic may be added, however, if the increase in the number of syllables can be avoided by elision, synizesis, or syncope. Accent. — In spite of the statements of the grammarians, the evidence for Cicero's usage, negative in types 2, 3, and 4, positive in type 1, points to the penultima law as the only accent law operative, 1 except in passages couched in legal or ritualistic diction which may have preserved the older initial accent. The enclitic "que" and words ending in e. — Harant (Revue de Phil. 1880, 25) called attention to the rarity of this combination in classic prose, stating that there was not a single example in the orations of Cicero. His statement for Cicero was subsequently modified by Thomas, 2 and by Lebreton, 3 who cite the examples of e" given in the above lists. They are: Type 1, sineque ferro, Div. 64;. type 3, consuetudineque rei p., Phil, x, 25; type 4, stationeque utuntur, Tull. 33, doloreque certa, Verr. v. 118, lihertateque utuntur, Agr. ii. 86,, confiteareque aliquando, Plane. 98, maioreque deorum, Phil. xiii. 46, ipseque in primis, ibid. xiv. 26, ipseque C. Pansa, ibid. 36. It may be questioned whether the avoidance was due to the I or simply to • Similar results were reached by C. Wagener (New phil. Rundschau 1904, p. 505) for type 4 (Mtisaque) on the evidence of the usage of the epic poets. However, in the case of polysyllabic words with the accent on the antepenult (types 2 and 3) he assumes the accentuation Uminague, on the evidence of liminaque laurusque dei, Aen. 3. 91. But all the examples of the lengthening of the que in these combinations are followed by a caesura, and are furthermore imitations of Homer. Besides, if this were the regular and normal accentuation, the rarity of the combinations in non-rhythmical prose would be hard to explain. It is very probable, however, that this was the accentuation in the few cases in which the combination is followed by a pause. See "Treatment of Dactylic Words," p. 153, where I have discussed the value of the final short before phrase and emphasis pauses. 'Rev. de Phil. 1884, 132. « La langue de Ciceron, Paris, 1901, p. 415. 40 F. W. Shipley the short vowel in the avoided types which we have been considering. In type 2 there are no examples of e, but all the examples with final a end either in -4a or -era and, where there is no elision, are subject to synizesis or syncope. In type 3 there is but one example of 8, but there is also but one example of & which is not subject to elision, synizesis, or syncope. In type 4, which contains the largest number of examples, both of I and &, and where synizesis and syncope are not operative, the evidence of elision is interesting. The ratio of elision in the case of & is 57 per cent; in the case of 8 it is 42 per cent. It is true that in the language words of this type ending in e occur more frequently than words ending in &. But, if posse and esse be left out of consideration — since, granting the tendency to avoid adding the enclitic to trochaic words ending in a short vowel, posse is generally associated with an infinitive in a word group, and esse with a noun, adjective, or participle, to which the enclitic may be attached — the words in & occur more frequently. In type 1 words in e are rare. 116. Type 7. Monosyllables. — What is the reason for the rela- tively low ratio of occurrence (8.6 per cent as compared with the ratio of frequency of monosyllables in the language, 23.5 per cent), and the high ratio of elision, 52 per cent ? In Caesar B.G. i-vi the ratio of occurrence, 7.4 per cent, is even less, and the ratio of elision is much higher, 77 per cent. The explanation cannot be either of those given for the avoidance of the other combinations. This combination forms a trochee, which should class it with the pre- ferred types 5 and 6, and there is here no awkwardness in the accen- tuation. Before attempting to answer this question I shall first give in Table II the figures for each word in the type (column A), with the number of elision cases (column B). In columns A 1 and B 1 the same figures are given in percentages for each of the parts of speech. These percentages are limited to the totals except where examples are sufficiently numerous to give an adequate percent- age basis. The figures in column X express the ratio of words in each group to the average occurrence of these monosyllables in the language, after deducting those words to which, owing to their meaning, the enclitic could not possibly be appended: vel, et, aut, nee. The Enclitic que in Cicbko 41 TABLE II Monosyllables Total No. of Examples Elisions A> Col. A in Percent- Monosyl- lables of This Type In Per- centages Bi Col. B In Percent- Conjunctions — quodque cumque 4 44 Relative pronouns — quique quaeque cuique quasque quodque Demonstrative pronouns — isque idque 48 17 4 32 Personal pronouns tuque meque teque seque nosque Verbs — datque . . Prepositions- inque exque . . . cumque . . deque . . . proque. . Adverbs — dumque . tumque. . quamque tamque. . plusque. . 36 1 9 13 6 10 39 16 1 1 2 14 1 19 1 30 31 10 11 11 19 13 1 1 1 6 1 10 70. 25. (25.) 66. 9. (11.) 59. 0. 34 19. (7.) 30. 21. .5 (14.) (11.) 48. 100. 8.4 (25.) 10. (7.) 81. 43. 53. 42 F. W. Shipley TABLE II— Continued Total No. of Examples B Elisions Ai Col. A in Percent- Monosyl- lables of This Type in Per- centages Bi Col. B in Percent- ages Nouns — iusque . . . spemque . duxque. . resque. . . remque. . disque. . . vique vimque . . 13 (3.) In the table the individual words to which que is added most frequently are cum (the conjunction), id, and tarn. The ratio of elision is very high in the case of cum; it is relatively low in the case of id, 34 per cent, and tarn, 43 per cent. If we consider the totals for the various parts of speech, the percentage figure in column A 1 is very much less than that in column X in the case of verbs — of which there is but one example and that elided, in spite of the very frequent occurrence of est, sunt, sit, sint — and in the case of preposi- tions, with 81 per cent of elisions. On the other hand, the figure in column A 1 is much higher than that in column X in the case of nouns and demonstrative pronouns. Here the ratio of elision is very low, 23 per cent and 30 per cent respectively. What is the reason for Cicero's preference for some combinations and the avoidance of others? It seems to be this: the majority of the monosyllabic words occur in word groups in which the main group accent falls either (1) upon the following word, as is the case with the conjunctions (e.g., quae cum ita sint), the prepositions (e.g., in primis), and to a large degree with the relative pronouns, or (2) upon the word which precedes the monosyllable, as, for instance, is the case with the monosyllabic forms of sum. In the former case the monosyllable is treated as a proclitic, in the latter as an enclitic. On the other hand, some monosyllables may carry the main accent in their groups, e.g., the nouns, the demonstrative pronouns, or a The Enclitic que in Cicero 43 demonstrative adverb such as tarn. In the case of the personal pro- nouns the accent would vary according to the emphasis placed upon them. Now the addition of an unelided que to an unaccented monosyllable in a word group gives to the monosyllable an accent and tends to destroy the balance of the group. This is the principle of avoidance. The enclitic is not ordinarily added, without elision, to a monosyllable, unless the monosyllable has an accent of its own. This explains the rarity, or the high ratio of elision, in the case of the forms of sum, in the case of the prepositions, and of the relative pronouns; 1 it also explains the relative frequency of occurrence, or the low rate of elision, in the case of the nouns, the demonstrative pronouns, and of the adverb tarn, which is frequently emphatic. EEMARKS ON THE MONOSYLLABLES 1. The conjunctions. — It is interesting to note that, of the conjunctions which most frequently introduce subordinate clauses, cum, ut, quod, the first should be the only one to which que is attached with any frequency, and here elision is the rule. These are four examples of quodque, but three of these occur in the legal diction of Cicero's motions in the Senate. Of utque there is no example. Draeger, Hist. Syntax II, 36, though noting the rarity of vtique, does not mention utque. TJtique occurs 5 times in the Philippics, in Cicero's resolutions in the Senate: 2. 25; 5. 53 (twice); 11. 30 (twice). 2. The prepositions. — In the statistical table que occurs with monosyllabic prepositions 16 times, with but three unelided examples. These three cases are repetitions of the same official formula: exque re publica. The 13 elided cases are interesting, since they furnish but a single type. The following word is invariably a demonstrative pronoun (is, id, 12 times, hie, once). The examples are : Inque followed by earn rem, iis sacris, eo exercitu. Proque followed by eo, iis rebus, hac .... diligentia. Exque followed by eo. Cumque followed by eis (twice), eis copiis, eorum exitio. Deque followed by ea societate, eius exercitu. Cicero's practice may be stated as follows: Except for the formula exque re publica, Cicero does not add the enclitic que to a monosyllabic prepo- sition except before the demonstrative pronouns 2 is and hie. Here elision takes place and the number of syllables is not increased by the addition of the enclitic. That que is not avoided with dissyllabic prepositions such as contra 1 Elmer, op. cit., p. 350, has pointed out that que is added to a relative pronoun only when the pronoun is repeated. * Landgraf had noted that que is only added to in and ex when a demonstrative pronoun follows, and Mayhoff (Archiv. XIII, 196) that this was also the case for cum. 44 F. W. Shipley 9 3 a) >> CM • i— i co ,. H GQ 23 O s i— t • t-H 9 i— t • i— t m ft. a B§!» !» co -co O..S c8 7 a ?>■ «. 3,'S.p" fsg a) m o w 2 «> . „ so» onT3 a" C 5S« ^2 §§gj •§m .2 o o > r&S'B o o »• QQ )l ; S )I — S ' ' 3 )i )Di » ) 1 OQ >l > * >l > >! > | 1 > •) > ' ' > >l .> i »> 1 1 ' • i — ._: 1 > ___ 1 ) ' T>' } § y > > > > 1 ■ i a ^ I'd a ,011 ) 1 1 ' J— ii*3 ? J.*- ;~ <- 1 - ~ ~+que. Type 3, - ~ - +que Type 4, - ~+que... Type 5, •" ~ +que \ Type 6, \-que ) Type 7, -+que. . . . A Frequency of occur- rence Ex- pressed in Percentages 100 B Elision Cases Expressed in Percentages 6.6 6.6 The table bears out previous statements as to the avoidance of types 2, 3, 4, either for reasons of accent or quantity. In type 7 the monosyllables are is (2 examples), id, quas, dux, quam. Is, id, and dux belong to the few favored types of monosyllables, and quam may be reckoned with tarn (see list of favored monosyllables). In type 1 the presence of iaque, with the value of a resolved trochee, shows that when both accent and quantity are favorable, Cicero does not hesitate to add que to a short vowel. In discussing the avoided forms for all the speeches, I have given all the examples which occur in clausulae. I have not included in the above statistics the ends of phrases, but there the results are prac- tically the same. Note. — The relative rarity of elision in the case of que at the end of mem- bers and of periods would seem to indicate that elision was in general more frequent in the more rapid enunciation at the beginning of phrases and clauses, and less frequent in the more distinct enunciation which marks the end. THE ENCLITIC WITH WORDS OF 5 OB MORE SYLLABLES It will be noted in Table III that, of the 106 examples which occur in clausulae, Cicero adds the enclitic que to words of 5 syllables 16 The Enclitic que in Cicero 47 times, and to words of 6 syllables 3 times, 19 examples in all, or 18 . 8 per cent. Draeger, Historische Syntax, II, 314, and Schmalz in Krebs Antibarbarus, 7th ed., make the statement that the enclitic is seldom added to words of 5 or more syllables. This statement does not apply to Cicero. The period clausula is, as I have said, the most carefully constructed portion of his period, and the large proportion of examples, and the entire absence of elision, show that Cicero cultivated the combination as giving a sonorous ending. These words of 5 or more syllables form but 6 per cent of Cicero's vocabulary in the orations, and the high rate of occurrence of que in combination with them in the clausula is an indication of distinct preference. Thus far I have given the figures for the clausula only. Even in the body of the period the figures both for occurrence and for elision still show a preference. In the Verrines and the Philippics, for instance, to take two large groups representing his earlier and his later speeches, there are 1,428 examples of que. Of these 1,428 examples, 199 are added to words of 5 or more syllables, or 14 per cent. The percentage of elision is 24 per cent, which is much higher than in the clausula but less than the average ratio of elision for enclitic combinations, which is 27 . 8 per cent. Naturally the 5-syllable words predominate. But, of the 199 examples just mentioned, que is added to words of 6 syllables 34 times, and 10 times to words of 7 syllables. In the present paper I have dealt only with the types of words and word endings to which que is added, and from the point of view of the bearing of Cicero's practice in this regard upon his sense of rhythm, and upon questions of accent. I have not attempted to treat of several questions, such as the relation of que, at, ac, and atque, which have been fully discussed by others. Washington University St. Louis, Mo.