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. SYLLABIFICATION IN LATIN INSCRIPTIONS By Walter Dbnnison Since the time when the "quantitative" method of pronoun- cing Latin was first adopted in America, much study has in this country been given to questions relating to the actual speech of the Romans. Thus such subjects as the correct reading of Latin poetry, the reconciliation of word accent and verse ictus, the mode of pronouncing elided syllables, and the admission of elision into prose, have commanded respectful, and even enthusiastic, atten- tion. It is with a question relating to the pronunciation of Latin that the present paper also has to deal. A rule of pronunciation, resulting from a misunderstanding of the Roman grammarians, a rule that has been perpetuated for eighteen centuries and finds expression still in nearly all of our grammars, is that in dividing the syllables of a word the Romans placed with a following vowel as many consonants as may stand at the beginning of a word in Latin (or in Greek) ; that, for example, the correct division in the following words was fru-ctus, ma-gnus, o-mnis, i-pse, nu-ptus, lu-scus, pro-spera, maie-stas, no-stri. Such groups of consonants are bd, cm, en, ct (net), dn, gn, mn, pn, ps (mps), pt (mpt), sb, scl, sm, sp, st (nst), str. The statements of the grammarians on this point are accepted by Seelmann, 1 who introduces as supposed confirmative proof the evidence also of the Romance languages and of a few inscriptions. In 1896 Professor William Gardner Hale 2 showed that the doc- trine of syllabification was based upon a misunderstanding of the grammarians, and that the evidence presented by Seelmann, especially the epigraphic evidence, was in reality damaging to the theory which Seelmann himself upheld. Professor Hale's posi- tion may be best defined by quoting his own language (pp. 251 f.) : My own conception is that the particular statements of the Roman grammarians on which the received doctrine is founded represented *Die Aussprache des Latein (Heilbronn, 1885), pp. 132-48. *Harv. Stud. VII (1896), pp. 249-71. [Classical Philology I, January, 1906] 47 48 Walter Dennison neither the facts of Greek pronunciation nor the facts of Roman pro- nunciation, but had their origin in a mere practical rule, — admirably simple and easy of application, — devised by some Greek grammarian for the division of words in writing, when one was near the end of the line and had room for a part of a word only; that this rule was adopted bodily by the Roman grammarians ; that the Roman grammarians mostly, in discussing the matter, were not thinking of pronunciation at all, though if asked how they pronounced, would doubtless have answered : " As we write;" that, nevertheless, a few of them did think of it, probably draw- ing the false inference, natural enough to anyone except a trained pho- netist, that, as writing and pronunciation conformed in most respects in Latin, so also they conformed in this. In support of this position Professor Hale discusses the state- ments of Roman writers from Quintilian through Caper, Caesel- lius, Servius, and Priscian to Bede, and dwells upon the point that the use of such phrases as in scribendo, in scriptura, si scribis, which these grammarians employ when they touch upon the syllabic division of the disputed combinations sp, st, pt, etc., shows that they were thinking merely of orthographic division ; in other words, that they were thinking of the division employed in writing the word down, not the division observed in actual pro- nunciation. Priscian and Bede, however, may, he thinks, have had phonetic division also in mind with more or less distinctness ; the former, in stating (translating) the doctrine of Herodian for the treatment of a preposition in composition, uses the phrase rationabilius esse sonoriusque (II, p. 45 K) ;' while the latter, who is too late to be a witness, expressly says, cum diets sive scribis (VII, p. 273 K; cf. p. 279). As additional proof of his thesis, Professor Hale adduces (1) the pronunciation of modern Italian, about which his observation is contrary to the verdict of the phonetists, 2 (2) fifty-seven instances of interpunctuation in inscriptions, e. g., in the w T ord VIC -TORI, (3) the general impression produced by the examination of some five hundred inscriptions in which divisions of words between lines occur, and 1 Professor Hale op. cit., p. 268, points out that in his Partitiones XII versuum Aeneidos III, pp. 478 ff. K), Priscian, according to the manuscripts, violates his own rule in two cases out of the three involved (om-nes, infrcw-tos, si-gnum), and thus probably betrays his own actual pronunciation. 2 Skutsch Krit. Jahresb. iib. d. Fortschr. d. roman. Philologie IV. 1. 89, expresses his agreement with Hale's observation. Syllabification in Latin Inscriptions 49 (4) the testimony of Latin poetry. At some later time he will present also evidence gathered from early manuscripts. At Professor Hale's request, as mentioned in his paper, I furnished a part of the evidence from interpunctuation and the general results of my examination of the five hundred inscrip- tions involving word-division. The present article attempts to set forth the evidence of these two kinds, that is, the entire epigraphic evidence bearing on the question, in a practically complete form. 1 This evidence is mainly statistical, but a presentation of it will perhaps be tolerated, since it is apparently conclusive. Two classes of inscriptions may be consulted for help. We have, in the first place, a small number which indicate the divi- sion of syllables by interpunctuation, as. e. g., OIL. VI. 26353: D M M ■ SERTO-RI-VS CRES-CES AE-LIE • TY-CE-NI COIV-GI • OP-TI-ME FECIT A separate study of this kind of evidence is made below (pp 64 ff.). Secondly, there are some thousands of inscriptions which divide words between lines, as in CIL. VI. 22105: D . M M • MARCIO • TERMI NALI • IVLIA MODES TA • FRATRI ■ DVLCIS SIMO • BENE MERENTI- FECIT The principle of division followed by the Roman stonecutter could be revealed on an examination of a large number of such 1 The epigraphic evidence has never extensively occupied the attention of scholars. Schneider Elementarbuch der lateinischen Sprache 1, p. 763, in a long discussion of syllabification (pp. 763-91) summarily dismisses the inscriptions as offering only care- less or accidental instances of word-division. Kiihner AusfUhrliche Grammatik I, pp. 155-58, also devotes a long section to this subject, but his statement regarding usage in inscriptions is incorrect, namely, that word-division in them is " ganz will- kuhrlich." Lindsay Latin Language, p. 125, Bennett Appendix, p. 31, and Sommer Handbuch der lateinischen Laut- und Formenlehre (Heidelberg, 1902), pp. 312, 313, refer briefly to the epigraphic testimony cited by Seelmann. The real significance of the inscriptional evidence was first pointed out in the article of Hale referred to above. 50 Walter Dennison inscriptions. But it is reasonable to suppose at the outset that the average Roman workman would make such line-divisions as seemed to him natural ones, not divisions which conformed with the technical rules laid down by the grammarians. In other words, these divisions would probably show the syllabification current in actual Roman speech, not the syllabification observed in formal writing. Now from literary sources it is perfectly clear that even in the time of Augustus words were not always spoken just as they were written, and that in some cases the grammarians taught a mode of writing that did not represent the true mode of speaking; 1 while there were also those who opposed this teaching, and believed that a Roman should write a word precisely as he spoke it. 2 In dealing with this question it was necessary at first to deter- mine whether any principle of division at all was observed in inscriptions, that is, to discover whether the divisions at the ends of lines were haphazard, or whether in the large majority of cases they followed some rule. At the outset, therefore, I tested inscriptions whose line-divisions involve the application of ac- cepted rules of syllabification as, for example, the rule that a single consonant between two vowels is placed with the latter vowel, or the rule that doubled consonants are separated. If I find these accepted rules commonly observed, it would seem justifiable to conclude that whatever divisions occur in the dis- puted combinations indicate the current and correct syllabification. The methods employed in collecting evidence and the extent of ground which this investigation covers may be briefly stated. I confined myself naturally to epigraphic evidence of Italian provenance. The volumes of the Corpus, upon which the follow- ing results are based, were therefore IV; V. 1. 2; VI. 1. 2. 3. 4,. 4 n ; IX; X. 1. 2; XI. 1. 2; XIV; including the parts of the Ephemeris Epigraphica supplementary to these volumes. The lOf. the writing, advocated by the grammarians, of 6s (as in urbs) and bt (as in obtineo) which does not represent, of course, the real pronunciation. 2 Suetonius Aug. 88: Orthographiam, id est formulam rationemque scribendi a grammaticis institutam, non adeo custodit (Augustus) ac videtur eorum potius sequi opinionem, qui perinde scribendum ac loquamur existiment. This remark follows the statement that in his own writing Augustus did not divide words at the ends of lines. Syllabification in Latin Inscriptions 51 only use made of the first volume of the Corpus was to obtain data with reference to earlier inscriptions (see p. 63). The fourth volume was admitted because of the value that attaches to results obtained from cursive writing, in which the writer was perfectly free to divide a word as his own sense of correctness dictated. The graffiti published in Bull, del Inst, and Notizie degli Scavi since the appearance of CIL. IV were also examined. On the other hand, no attention was given to inscriptions appear- ing in the fifteenth volume of the Corpus or to those published in other volumes under the head of " Instrumentum domesticum," since inscriptions of this class are likely to have a monogram- matic character, the letters being disposed according to a given space or form. Several long inscriptions, of presumably Italian provenance, though appearing in other volumes than those already mentioned, were also tested, e. g., the Aes Malacitanum (CIL. II. 1964), the diplomata militaria, etc. The whole number of inscriptions examined was about 80,000. Within the limits thus set the investigation included such inscriptions only as have been copied from the original (usually still extant) or from a squeeze or photograph, by an editor of the Corpus or by some other trusted epigraphist. It was obviously necessary to exclude all inscriptions whose arrangement by line is not absolutely known. 1 To have examined critically those known from manuscripts only, with a view to determining the line-division, would have been a task not only difficult and often fruitless in itself, but, since existing material is so plentiful, use- less as well. So, in the case of fragmentary inscriptions that have been restored either by conjecture or from early copies, in- stances of division are counted only when the restoration is made at but one end of the word and is practically certain. Other cases which I found it necessary to omit are words that have been 1 It is perhaps needless to call attention to such notes as the following frequently appended to inscriptions now lost: "Versuum divisio incerta," IX. 5407. 5922; VI. 502. 695 ; " Versus aliter divisit Don.," XI. 1615 ; " male divisit Maur.," X. 372. Differ- ences of versification in two copyists may be seen in IX. 410 (and p. 659) . Occasionally the lost originals of previously copied inscriptions turn up later and show wide differ- ences in versification, e. g., GIL. V. 1050 (and note p. 1025), 4896 (and note p. 1081) ; VI. 1662 (and Add. p. 854) ; X. 304 (and note p. 962). Copies made in modern times can not always be trusted ; cf. CIL. X. 441 (and note p. 964), 3607 (and note p. 974). 52 Walter Dennison misspelled or engraved incorrectly, as DVLCIS | MO ; enclitics, as POSTERIS | QVE, or divisions that are plainly etymological, as PER | MISIT, OIL. VI. 10241, OP | TVLERIT, XIV. 3014; and uncertain readings, as in CIL. IV. 396. In brief, then, the attempt is here made to discover how words are divided at the ends of lines in perfect, extant inscriptions where copies have been made or verified by expert epigraphists. As stated above it was necessary first to observe how the acknowledged rules of syllabification were treated by the Roman stonecutter. The results of this observation, based upon an examination of all eligible inscriptions found in Italy, are pre- sented in two tables. In the first table instances of correct division are arranged under the following heads: (1) Diph- thongs and vowels, e. g., i-ae, i-ei; (2) single consonants, e. g., o-no, i-mo; (3) mute and liquid, e. g., a-tra, e-brae; (4) two A. Following Acknowledged Rules Kinds of Division Perfect Cases Restored by Conjecture or from Early Copies Total 184 7,153 270 1,019 1,816 53 13 1,166 25 108 299 9 197 8,319 295 Two consonants — different . , 1,127 2,115 62 Total 10,495 1,620 12,115 B. Violating Acknowledged Rules Kinds of Division Perfect Cases Restored by Conjecture or from Early Copies Total Diphthongs and vowels 50 1,113 111 135 451 29 4 no 10 11 40 3 54 1,223 121 Two consonants — same Two consonants — different. . 146 491 32 Total 1,889 178 2,067 Syllabification in Latin Inscriptions 53 consonants, the same, e. g., s-s, m-m; (5) two consonants, differ- ent, e. g., m-p, n-c; (6) three consonants, e. g., n-tr, m-pl. In the second table, similarly classified, are instances of division which violate the acknowledged rules. We thus observe that in a large number of extant, well-authen- ticated inscriptions containing words divided at the ends of lines there are 12,115 instances where the division follows the accepted rules for syllabification, and 2,067 instances where it does not ; in other words, more than 84 per cent, of the sum total of these instances divide correctly. The percentages by volumes of the Corpus are: IV, 84 per cent. ; V, 80 per cent. ; VI, 87 per cent. ; IX, 82 per cent. ; X, 87 per cent., XI, 83 per cent. ; XIV, 83 per cent. Since, therefore, syllable-division in inscriptions conforms on the whole with acknowledged rules, we shall be justified in accepting the epigraphic evidence, whatever it is, which relates to syllabification in the consonant groups in question. The evidence follows. The cases in question are arranged alphabetically, first, those which divide according to the received theory (the theory upheld by Seelmann) ; secondly, the instances which divide the consonants of the disputed groups between the preceding and the following vowel. The illustrations of the grammarians will be found interesting for purposes of comparison. 1 CN 2 PHILOTE | CHNVS, VI. 19365. AC I NELLVS, X. 4519. CT 3 ADAV | CTVS, VI. 17221, IX. 152; ADLE | CTO, Eph. Epig. VIII, 368; au | CTOR, X. 5670; CATAFRa | CTARIORVM, V. 6784; DO | CTOR, VI. 10183; EPI- | CTE-TO, VI. 10681; EPI | CTESI, V. 4638; EPI | CTE-SIS, VI. 21278; EPI | CTETUS, VI. 11986; PA | CTAM, IX. 2827; FI | CTOSA, XIV. 2275; INDI | CTIONE, XI. 549; INVI | CTIS, V. 8032c; IREZE | CTESI, X. 2611; NO | CTIS, X. 7112; PERFE | CTOQVE, VI. 2078; PREFE | CTI, VI. 32006; STRV | CTO-RI, X. 1959; VI | CTORI, V. 5063; VI | CTORIA, X. 2482; VI | CTORIAE, X. 3076, XI. 2547; VI | CTORINAI, VI. 18703; BI | CTORINO, VI. 723. 1 Illustrations of the grammarians involving a group of consonants not in this list are : a-bdera, a-bdomen, Priscian I , p. 42 H ; ab-ditur, Servius in Don. IV. 427 K. 2 Cf. the grammarians' illustrations, Pyra-cmon, Al-cmene, dra-chma, Priscian I, p. 42 H ; Pyra-cmon, Menae-chmus, id. II, p. 112 H. 8 Illustrations cited by the grammarians are Eri-chthonius, Priscian I, p. 13 H ; a-ctus, le-ctus, id. I, p. 42 H; aspe-ctum, affe-ctum, affli-ctum, Bede VII, p. 263 K; conspe-ctus, id. VII, p. 268 K; fru-ctum, fa-ctum, fi-ctum, id. VII, p. 273 K. 54 Walter Dennison ac | TVS, VI. 1527; ADIEC | TO, X. 5796; ADLEO | TO, XI. 5215; AIEC | TE, VI. 19027; ATHIC | TVS, VI. 28562; AVC | TORI, X. 1126; BENEDIC | TA, XI. 3566; BENEDIC | TO, VI. 10192; CORREC | TORI, VI. 1768; DOC | TOR, VI. 533; AEPIC | TESINI, VI. 4565; EPIC | TESIS, X.2964; EPIC | TETI, XIV. 636; EPIC | TETO, VI. 18711a; EPIC | TETVS, VI. 17212, X. 4130, XIV. 683; EPIC | TE— | TVS, IX. 5547; EVTAC | TVS, VI. 17571; EXSTRVC | TVRVS, VI. 1585b; Fac | TIONIS, VI. 10076; fac \ TO, VI. 2075; FAC | TVM, XI. 1146; FRVC | TVOSO, VI. 7788; FRVC | tus, VI. 5040; FRVC | TVS, VI. 16199, 28808; GALAC | TINE, VI. 25912; Indie | TIONE, V. 5423; invie | TO, V. 4320; INVIC | TO, VI. 755, IX. 5997, X. 6816, 6925, XI. 6631; INVIC | TIS, V. 8030; LEC | toR, XI. 1704; LVC | TVM, VI. 18086; OC | TAVI, V. 6644; OC | TO, X. 4515; OC | TOB, V. 6730, Eph. Epig. VIII. 880; OC | TOBRES, IX. 5684; OC | TVB (sic), XI. 1408; PEC | TVM, XI. 3571; PRAEFEC | TO, XI. 3367; PROTEC | TO, VI. 11998; SENEC [ TAE, VI. 18086; SOC- | TES, VI. 18919; SPEC | TATO.X.2383; TRAC | TVS, IX. 334; VEC | TIQALI, XI. 1147; VIC- | TOR, VI. 31841; VIC* | TOR, VI. 16326; VIC | TOR, IX. 1291; VIC | TORI, V. 8030, VI. 1167, 13266; VIC | TORI, VI. 18693; VIC | TORIA, VI. 7977, 36011, X. 3455; VIC | TORIAE, VI. 8084, 20635; VIC | TORINA, XI. 6553a; VIC | TORINAE, XI. 2328; BIC | TORINAE, X. 1503; VIC" | TORINE, XI. 2775; VIC | TORINO, VI. 8875, XI. 2539; VIC | TORINVS, V. 5058, X. 3459; 1 VIC | toriNVS, XI. 6576; VIC I TORIS, X. 677. CTK 2 VIC | TRICI, VI. 531; VIC | TRICIS, XI. 5959. NOT 3 DEFVN | CTO, VI, 21342; FVN | CTI, V. 4055; fun | CTO, XI. 4214; FVN | CTus, VI. 33992; SAN | CTA, V. 1686; SAN | CTAE, 1 X. 2444, XIV. 6; SAN | cTE, 1 VI. 30738; SAN | cilSSIM, 1 XIV. 839; SAN | CTISSIMA, 1 X. 1909, XIV. 1509; SAN | CTISSIMAE, 1 VI. 7718, 16776, 20735, 22159, 28188 > 29737, Eph. Epig. VIII. 416; SAN | CTIS SIME, 1 VI. 35549; SAN | CTIS- SIMO, 1 VI. 13986, X. 5983; saw | CTISSImo, 1 VI. 18266; SAN | CTIS -SI -MO, 1 XIV. 440; un \ CTIS, VI. 2075. CVNC | TARVMQ., VI. 32051; FVNC- | TA, XI. 2702; SANC | TAE, XI. 4967; SANC | TAES, VI. 19062a; SANC | TE, IX. 6150; SANC- | TE, XI. 2702; SANC | TI, X. 4514; SANC | TISSI, IX. 3438; SANC | TISSIMAE, VI. 10761, 12537, 14289, 15641, 19173, 20327, 27945, XIV. 2202; SANC" | TISSI MAE, VI. 26510; SANC | TISSIME, VI. 24101; SANC | TISSIMO, V. 3791, VI. 8954, IX. 888; SANc | TISSIMO, VI. 24436. DN* ARIA | DNE, 5 VI. 28414. i Of. IN- VI-OITE, VI. 80, and note 4, p. 64. 2 Vi-ctrix is the illustration given by Priscian I, p. 43 H. 3 The illustration, defun-cttis, is offered by Bede VII, p. 270 K. 4 Grammarians' illustrations are Cy-dnus, Aria-dne, Priscian I, p. 42 H; cf. A-dmetus and Ae-tna, ibid. » Cf . ARIADINE, VI. 21398. Syllabification in Latin Insoeiptions 55 GN 1 BENI | GNITATE, VI. 15860; CO | GNATO, 2 VI. 23849; CO \ gnaTVS, X. 5899; INDI | GNE, VI. 28523; INSI | GNIA, IX. 5684; MA | GNO, IV. 2174; MA | GNV, VI. 34635a; SI | gnisque, VI. 2086. AG | NES, VI. 36476; ADSIG | NARETVR, XL 1924; ag | NAM, VI. 2068; BENIG | NVS, V. 6464; BENIG | NISSIME, X. 4088; BENIG | NITATE, VI. 32420; COG | NATVS, 2 VI. 2723; DESIG | NATVS, X. 6893; DIG | NA- TIONE, XI. 6335; DIG | NATVS, XIV. 2934; DIG | NISSIMAE, XI. 3368; DIG | NISSIME, V. 4850; DIG | niss IME, VI. 7142; DIG | NIT, Eph. Epig. VIII. 101; DIG | »iTA(EM, V. 532; EG | NATIVS, VI. 18374; eg | NATIVS, IX. 853; MAG | NA, VI. 15029; MAG | NI, X. 6893; MAG | NO, V. 3221; Ukg | NO, VI. 32808; MAG | NIFICO, IX. 2237; TESSIG | NIVS, V. 805; TIG | NVARIOR, XIV. 418. MN S ALV | mnae, VI. 11380; ALV | MNAE, VI. 12402, 23450, 25468; ALV | MNE, VI. 19219; ALV i MNI, VI. 26294; ALV | MNO, 4 VI. 3367, 8679, 13139, 17198, 21095, 35501, X. 2248; ALV | MNVS, XL 1854; DO | MNINA, X. 3452; DO | MNIONI, VI. 3572; HY | MNIDI, VI. 21561; HY | MNO, X. 3048; HY | MNVS, V. 5557, VI. 20334, 28041; LE | MNO, VI. 6702; O | MNIA, VI. 3495; o | MNIS, IV. Suppl. I. CLV; SCY | MNVS, VI. 19074; SE | MNE, VI. 15986; SO | MNO, IX. 1393. ALVM | NAE, VI. 21903, 22729; ALVJtf | NE, XIV. 530; ALVM- | ne, XIV. 1481; ALVM | NIS, VI. 28929; ALVM | NO, VI. 5499, 10903, 13457, 25085, 27629, Eph. Epig. VIII. 284; COLVM | NIS, IX. 2475, X. 7946; colum | NIS, VI. 207; OM | NES, VI. 21787, 32420; OM | NI, VI. 11252, X. 7112; OM | NIBVS, V. 4937, VI. 13208, X. 476, 5654,5657, XL 1146, 4097; om | NIVM, IX. 5071; CGM I NA, VI. 27246; SOLLEM | NITAS, XL 5265; SOM | NO,V. 6730, IX. 1082, 1386, X. 4497, 4519, 4546; som [ NO, X. 3302. PN 5 DA | PNEN, IX. 2574; HY | PNVS, VI. 12375. DAPH | NVS, 6 VI. 8855; THRECEDIP \ NVS, VI. 14306. 1 The illustrations of the grammarians are ma-gnus, di-gnus, a-gnus, si-gna, pu-gna, Terentianus Maur. VI, p. 352 K ; ma-gno, Dositheus VII, p. 385 K ; ma-gnus, Albinus VII, p. 305 K; prae-gnantem, pi-gnus, di-gnus, id. VII, p. 307 K; prae- gnantem, pi-gnus, pu-gna, Bede VII, p. 286 K ; cf . My-gdonides, Priscian I, p. 42 H ; a-gmen, id. I, p. 42 H, II, p. 112 H. a-gmine, Dositheus VII, p. 385 K. 2 Cf . note 4, p. 65. 3 Illustrations of the grammarians are a-mnis, Servius in Don. IV, p. 427 K, Priscian I, p. 41 H, Marius Victorinus VI, p. 29 K, Terentianus Maurus VI, p. 351 K, Dositheus VII, p. 387 K; o-mnis, Terentianus Maurus VI, p. 351 K; sole-mne, Albinus VII, p. 310 K ("secundum Priscianum") ; solle-mne, so-mnium, Bede VII, p. 289 K; calu-mnia, colu-mna, Bede VII, p. 268 K. < But cf. ALVM | MNO, VI. 18222, 29056; ALVM | MNVS, XIV. 222; ALVM | MNE, VI. 36234; ALVMPNVS, V. 6722. * The grammarians cite Thera-pne, Si-phnus, Priscian I, p. 42 H. <>Cf. DA | FINE, XIV. 2841, which suggests that there was a pause between/ and n. 56 Walter Dennison PS 1 COwLAB | SAs, X. 6922; CONLAP | SVM, X. 4858; IP | SIS, X. 1402; IP | sum, V. 532; LAP | SVM, VI. 1179; VIP- | SANIA, XIV. 1781. MPS COM | PSES, XI. 2288; content | PSI, VI. 18659. PT 2 ADO | PTIONEM, VI. 2041; COE | PTVS, VI. 462; CONSCRI | PTIS, X. 3903; corru | PTAM, X. 8031; HYPO | PTO, VI. 26960; NE | PTVNO, X. 3412; o | PTIMVS, X. 1811; O | PTIONI, XL 5646; PRAECE | PTORI, VI. 8979; SCRI | PTI, VI. 7582, Eph. Epig. VIII. 210; SE | PTIMIVS, VI. 20853; THRE | PTO, VI. 2334, IX. 6164; THRE | PTVS, XIV. 668. ACCEP | TORE, VI. 10050: ADOP | TATICIO, VI. 23673; CAP | TVS, VI. 10115; CORRVP | TA, X. 5348, 7996, 8009; CORRVP | TAM, X. 7997; coRRVP | TAM, Eph. Epig. VIII. 739; CONRVP | TVM, VI. 22120; COR- RVP | turn, XL 6040; GRAP | TO, VI. 15405, X. 7630; NEP | TVNALI, VI. 12515; OP | TATI, VI. 3257; OP | TA-TO, X. 4079; OP | TATO, X. 4367 OP | TIMAE, X. 1500; OP | TIMI, V. 6122a, IX. 4999; OP [ TIMIS, IX. 1478 OP | TIMO, X. 342, XL 3367; OP | ilMO, X. 4553; OP | TIO, VI. 30760 OP | TIO NIS, V. 6609; PRECEP | TV, V. 6503a; RECEP | TA, VI. 26028 SEP | teMBRIS, X. 1193; SEP | TIMA, VI. 14513; Sep \ TIMI, V. 5806 SEP | TIMI, X. 5052; SEP | TIMO, VI. 17744&; SVBSCRIP | TVM, VI. 2120; SVSCEP-|TO, XL 3287; THREP | TIANI, XIV. 348; VOLVP | TATES, IX. 3473; VOLVP | iATIBVS, X. 1251. MPT ADEM | PTA, VI. 21285; EM | PTO, VI. 28706; POM | PTINA, VI. 2466; REDEM | PTA, VI. 13055; REDEM | PTAE, X. 3677. CONSVMP | TAS, XL 4781; CONSVMP | TVM, VI. 831; EMP | TVM, VI. 17158; EMP | TVS, VI. 17051; POMP | TINA, XIV. 532; PROMP | TO, XIV. 2934; SVMP | Usque, VI. 2071. PTR 3 accep | TRICes, VI. 32328; DECEP | TRIX, X. 2601. SB 4 LES | BI, VI. 17135; LEZ | BIAE, VI. 27031; TTPEC | BEYTQN, VI. 31961; TTPEC I BITEPEC, IX. 6209, 6230. SC 5 ABA | SCANTO, VI. 25683; CRE | SCENI, VI. 22496; CRe | SCENS, IX. 5217; CRE | SCENTI, VI. 22055, 27872; CRE | S-CENTIANVS, VI. 1 The grammarians cite i-pse, nu-psi, scri-psi, Priscian I, p. 43 H ; i-pse, Priscian I, p. 461 H. 2 The grammarians' illustrations are a-ptus, di-phthongus, Priscian I, p. 42 H : nu- ptam, scri-ptum, id. I, p. 43 H ; a-ptus, id. I, p. 461 H ; ca-pto, Caesellius op. Cassiod. VII, p. 205 K ; Aegy | ptum, Bede VII, p. 263 K ; pro-pter, id. VII, p. 286 K. 3 sce-ptrum is cited by Priscian I, p. 43 H. i asbestus and A-sbustes are cited by Priscian I, p. 42 H. 6 Illustrations cited by grammarians are pa-scua, lu-scus, Priscian I, p. 50 H; ne-scio, Terentius Scaurus VII, p. 12 K. Syllabification in Latin Insceiptions 57 15719; CRE | SCENTINA, VI. 21095; CRE | SCES, XIV. 916; JDI | sceNTI- BVS, X. 7657; EV | SCHEMVS, VI. 19792; FI- | sco, VI. 29917; GRAVI | SCANORVM, VI. 1408; PRI | SCIANO, VI. 25390; PRI | SCILLAE, VI. 25039; PRI | SCO, X. 2711; PRI | SCV8, VI. 25048, X. 380; pri | SCVS, XI. 4394; QVIE | SCANT, V. 5078; SV | SCEPTO, IX. 4192. ABAS | CANTI, 1 VI. 4080; ABAS | CANTO, VI. 16922; ABAS | CANTVS, VI. 6710, 10854, 17812,18140,18933; AAEAEC | XOC, VI. 14672; ADQVIES | CIT, V. 4063; ADVLES | CENTI, XI. 1147; ARES | CONTIS, VI. 13608; ARES- | CVSA, VI. 13596; ARES | CVSA, VI. 21650, IX. 161, XIV. 1367; AS | CONIVS, V. 2856; BASILIS | CVS, VI. 36393; ORES | CENS, VI. 11294, 21347; CRES- | CENS, X. 21; ORES | CENTI, X. 2658, 3660; CREs | CENTIA, X. 2163; CREs | CENTIAE, IX. 4894; CRES | CENTIANO, VI. 19914; CRES | CENTILLA, V.2187; CRES | CENTINA, XIV 1323; CRES | CENTINI, X. 7239; CRES | CENTINO, VI. 6693; CRES | CENTIO, V. 6075; CRES | CENTIS, VI. 21871, X. 2421, 2658, 3363, XIV. 3632; CRES | cer-ET, V. 532; CRES | CES, Bull, del Inst. 1878, p. 89, XI. 2553; CRES | CETIANO, VI. 25837; DIOS | CORO, VI. 19939; DIS | eiplina, VI. 8619; DIS | CIS, V. 944; HPAC I KUUNAUJN, V. 8733; FALIS | corum, XI. 3091; FIS | CI, XIV. 3643; FVS | CO, VI. 36289; GIS | CINNA, VI. 10910; LARIS | CVS, VI. 9494; MAS | CVLIS, V. 5002; MOS- | CHIDI, VI. 23049; MOS | cHIS, VI. 34557; MOS | CHO, VI. 28414; PES | CENNIANVM, XI. 1147; PRIS| , X. 3570; PRIS | CAE, V. 4064; PRIS | CAE, IX. 5422, X. 8194; PRIs | CI, X. 3724; PRIS | ciaNO, VI. 25036; PRIS | CIANO, X. 8404; PRIS | CIANVS, X. 1208; PRIs | CILLA, IX. 5100; PRIS | CILLA, XI. 3521; PRIS | CIILLA (sic), VI. 22585; PRlS | CILLAE, VI. 20635, 34238; PRIS | CINVS, VI. 22841a; PRIS | CO, V. 8139, VI. 12079, X. 373, 4686; XI. 1147, 4533; PRlS | CVS, V. 2117; PRIS | CVS, VI. 22383, IX. 1506, X. 3452, XI. 2688; PRIS | CVS, VI. 24499; QVES | CET, X. 7550; QVES | QVAM, VI. 25667; QVIES | CIT, XI. 312; REQVIES | CET, V. 5710, 7531, X. 4519; REQVIES | CIT, IX. 1080, 1384, X. 4500, 7753, XI. 1412; REQVIes | CIT, IX. 1082; REQVIIS | CIT, IX. 1388; REQVIS | CET, V. 6251; ROS | CIAE, VI. 25486; SES | CVNCIAM, V. 4108; tus | CANICAS, VI. 2099; TVS | CIAE, VI. 1778. SCL 2 AS | CLEPIAS, VI. 6931. SM 3 CO I SMVM, IX. 2438. COS | MIAE, VI. 11061; COS | MIANVS, VI. 13517; COS | MO, V. 8151; COS | MVS, VI. 411; KOC I M0Y, VI. 33976; MYRIS | MI, VI. 7770; MYRIS | MVS, VI. 25827; THUS | MVS, IV. 2160. SP* BUI I CTTOPANOC, VI. 5207. 'Of. the suggestive spellings DISSCENTE, IV. 1278; AEESSOVSA, IX. 3446; PESSOENITI, IX. 4055. 2 An illustration cited by Priscian I, p. 43 H, is A-sclepiodotus. But cf . the spelling ASSCLEPIODOTE, VI. 12504. s co-smus is cited by Priscian I, p. 50 H. 4 Grammarians' illustrations are a-spice, Servius in Don. IV, p. 437 K ; pro-spera, Priscian I, p. 50 K ; pro-spere, Bede VII, p. 286 K. 58 Walter Dennison CRIS | pina, VI. 16593; CMS | PINI, X. 1784; CRIS | PINIANO, V. a31; CRIS | PINIANVS.V.8710; CRIS | PINILLA, VI. 14290; CRIS | PINO, IX. 1684; CRIS | PO, VI. 32694; CRIS | PVS, VI. 7276, X. 3431; HIS | PAN, IX. 1572; HOS | PITIVM, X. 7845; IS | PIRITO, X. 7551; POS | PHOR, VI. 27499; TELES | PHOR, VI. 12792; TELES | FORIANO, XIV. 1208; TELES | PHORIDI, VI. 12792; TELES | PHORVS, 1 VI. 290, 13246; TELES | i>7w>RVS, VI. 29385; TELES | PORVS, VI. 29605; TELIS" | PHORVS, VI. 5768; VES | PER, VI. 9977. ST 2 AERA | STVS, VI. 7513; AVGV | STA, V. 3555; AV-GV | S-TALES, X. 2194; AVGV | STAS, VI. 33865; AVGV | STI, VI. 30894; AVGV | STINA, V. 3555; Augu | STO, VI. 32340; AVGV | STORVM, VI. 36308; BLA | STI, VI. 18416; BLA | STO, VI. 36322; CALLI | STVS, V.8215; CE-LE | STI, VI. 80; CELE | STINO, X. 3550; CHRE | STE, V. 232, VI. 9477; chre | STE, VI. 33169; CHRE | STVS, V. 8356; cu | STODIA, VI. 1527; GUI I CTAM6N H, VI. 30966; EVRHO | stus, VI. 23178; PAV | STHILLA, VI. 29623; PAV | STE, VI. 3012; FAV | STI, XIV. 1745; FAV | stILLVS, X. 7841; fau | STINA, IX. 3627; FAV | sttNAE, IX. 4292; FAV | STINAE, VI. 35807, X. 3106; FAV | STI | NVS, VI. 35067; FAV | STO, XL 4333; FAV | STVS, VI. 27876, XIV. 676; FAV | sTVS, X. 5252; MNE | STERIS, VI. 18140; MODE | STE, IX. 3028; PO | STER, VI. 3372; PO | STERIS, VI. 14090; PO | STERISQ, VI. 24347; po | STERISQwe, VI. 14401; PO | STERISQVE, VI. 25740; PO | steRISQVE, VI. 35945; PO | STVMIVS, XIV. 3627: QVAE | stori, Eph. Epig. IV. 823; QVAE | STORIAS, VI. 2086; QVE | STORI, VI. 1803c; RV | STICI, V. 5557; se | STERtas, IV. Suppl. I, OLV; TE | STAMENTO, X. 3359; VE | STA, X. 3365; VILAGO | STI, V. 7837. ABES | TO, 3 VI. 19931; A | CHORIS | TVS, XL 3393; AMETHYS | TVS, VI. 13986; ANTIS | TIAE, VI. 6688; ANTIS | MARV, VI. 7650; APOLAVs | TO, VI. 19841; APOS | TOLI, VI. 33895; AVFVS | TIO, VI. 34585; AVGVS- | TALI, VI. 15876; AVGVs | TALIB, X. 451; AVGVS | TALES, XL 3805, 6005; AVGVs | TALITATIS, X. 114; AVGVS | TALIVM, IX. 4067, X. 114; AVC-VS|TI, VI. 10931; AVGVS | TI (2), VI. 17924, XI. 6623; AVGVS | TIANVS, VI. 25038; AVGVS | TEST, VI. 13715; AVGVS | TIS, V. 8030; AVGVS | TO, VI. 1181; AVGVS | TORVM, VI. 29294, 36507, X. 1586; AVGVS | TVS, V. 1784, VI. 2605; BIAS | TVS, VI. 16995; BLAS | TVS, VI. 9571; CAELES | TES.IX.4192; CALLIS | TO, VI. 20833; CALLJS | TO, VI. 20283; CAS | TA, VI. 6148, Eph. Epig. VIII. 219, X. 4763; CAS | TISSIMAE, X. 3356; CAS | TITATIS, V. 4187; CAS | TOR, IX. 3219; CAS | TORE, XL 53; CAS | TVLAE, XIV. 1488; CAS | TVS, Eph. Epig. IV. 933; CHRES | i Note TELES | SPORO, VI. 18222. 2 Grammarians' illustrations are maie-stas, Caper VII, p. 96 K ; maiestas Bede VII. p. 279 K. ; pote-stas, noster, Oaesellius ap. Cassiod. VII, p. 205 K ; testis, Priscian I, p. 50 H; Nestoris, Terentianus Maurus VI, p. 352 K; apostolus, Orthogr. Bour., p. 300 K, Suppl.; and cf. astla, *pestlum (= pestulum), Gaesellius ap. Cassiod. VII, p. 205 K, and is-thmos, Priscian I, p. 42 K. s Of. such spellings as CHRESSTINA, VI. 13712, IVSSTI, V. 6127, IVSSTA, IX. 1268, and VESSTA, VI. 3097, which seem to show an effort to isolate the s. Syllabification in Latin Inscriptions 59 TENI, VI. 24253; CHRBS | TVS, VI. 11707, 32997; CLVS | TVMINA, XI. 4489; COMAS [ iVS, VI. 3396; CONSIS | TIT, VI. 404; COS [ TINO, Eph. Epig. VIII. 788; CVS | TODIAE, VI. 10238; DIS | TINGVITVR, XI. 5265; EDIS | TE, VI. 18351; EPIS | TVLIS, VI. 8613; ES | TIVALIVM, X. 5348; ES | TO, XIV. 2112: EVARIS | TO, VI. 25287, 34403; EVARIS | TVS, XIV. 882; AEVARIS | TVS, XI. 4224; EVENVS | ti, VI. 24079; FAVS | TA, X. 8323; FAVS | TILLA, VI. 17051; FAVS | TINA, V. 3710; FAVS- | TINAE.VI. 24612; FAVS | TINO, IX. 1641, X. 4785; FAVS | TINVM, Not. d. Sc. 1880, p. 97; FAVS | TINVS, IX. 647; FAVS | TIO, X. 2919; FAVS | TO, VI. 5700, X. 8119; FAVS | TVM, VI. 2074; FAVS | TVS, VI. 11839, 23706; FAVS- | TVS, VI. 23560; FES | TA, V. 3730, 4994; FES | TAE, VI. 5345; FEs | TI, VI. 2696; FES | TIANVS, XI. 5785; FES | TO, VI. 14537; FES | TVS, V. 8247; HIS | TAR, VI. 5277; HIS | TONIESIVM, IX. 2860;! HONES | TA, X. 7890; HONES | TISSIMAE, IX. 2603; HONES | TVS, IX. 2613; HOS | TIAS, IX. 5565; HOS | TILIA, VI. 18773; HOS | TILIVS, VI. 25551; IADES | TINI, X. 1402; IS | TERCORIA, V. 1706; IS | TIPEN, VI. 32694; IVS | TAE, V. 3968, IX. 1478, X. 3751, VI. 33417; IVS | TE, VI. 21332; IVS | TINA, VI. 22493, X. 4514; IVS | ttnAE, IX. 1274; IVS | TINAE, Eph. Epig. VIII. 114, XI. 1857; IVS | TINE, XI. 6023; IVS | TINI, V. 694; ius | TINO, X. 4512; IVS ! TINVS, VI. 10221, IX. 1759; IVS | TITIA, IX. 1576; IVS | TITIAES, XI. 6204; IVS | TO, VI. 18121a; IVS | TO, VI. 20802, XI. 1891; IVS | TORVM, IX. 6400; IVS | TVS, V. 5020, VI. 24211, XI. 603, 1790, 6281, XIV. 634; MAGIS | TER, VI. 2078; MAGIS | TERIO, XI. 1947; MAIES | tatiQ, XI. 1429; MEDIAS | TINORVM, XIV. 1878; MEGIS | TE, VI. 34928; MERVis | TIS, V. 5701; MINIS | ffiATOR, VI. 33781; MNHCI0HCO- MENOY, X. 7176 (note); MODES | TA, VI. 17693, 22105; MODES | TAE, V. 2523; MODES | TO, VI. 13a; MODES | TVS, IX. 1895, VI. 33215; MODes | TVS, IX. 2223; ORES | TES, VI. 26432; OREs | TES, VI. 29800; OS | TENSE, XIV. 294; OS | TILIA, VI. 27977; OS | TITATIS, VI. 3454; PALES | TRICE, VI. 36393; PHILOS | TORGO, VI. 6788; TTIC I TH. VI. 11933; POS | TAE-RISQVE, VI. 34574; POS | TERI | QVE, VI. 25306; POS | TERIS, VI. 10677, IX. 3922; POS- | TERIS, VI. 27977; POS | TERISQ, VI. 7899, 11071, 15197, 15710, 22057, 34267, 35567, 35672, 36338, X. 2191, XIV. 508; POS- | TERISQ, VI. 18818; POS | TERISQ VAE, VI. 17213; POS | TERISQ VE, VI. 6783, 7457, 8931, 10235, 10238, 10498, 10536, 10989, 11824, 12739, 13205, 13418, 14513, 15423, 16126, 17158, 18052, 18935, 20227, 20596, 21852, 22275, 22685, 23111, 24630, 26159, 26271, 26313, 26796, 30072, 33076, 34241, 34890, 36581, IX. 4329, X. 3594, XIV. 667, 735, 864; POS | TERIS | QVE, VI. 25793; pos | TERISQ VE, VI. 16620, 19140, 22569, 35554, 36085, X. 7853, XIV. 1502; pos \ ffiRISQVE, VI. 23208; POS | feRISQVE, VI. 29527; POS | terisque, VI. 30226a, 35577, 36195; POS | TERISQVE, VI. 12417; POS | TERISQVE, VI. 22682; POS | TVMIawVS, VI. 32742; POS | TVMIO, VI. 24877; POS | TVMVS, VI. 2065, 2068; pos | TVMws, VI. 2074; POS | TVMms, VI. 2074; POTIIs [ TATII, Eph. Epig. VIII. 6; POTES | TATEM, VI. 1759; POTES | TATIS, X. 7996; PRAEPOS | TERAE, VI. 22203; PRAES | TANTIAM, X. 3764; PRAES | TANTIOR.X. 2936a; PRAES | TANTISSIMAE, IX. 1578; PRAES | TANTISSIMO, IX. 1685, X. 1126, 5919; PRAES | TARE, X. 5336; PRAES 1 TETVR, XI. 2596; PRIS | TINAM, VI. 34561; QVAEs | TORI, V. 4341; 60 Walter Dennison quaes | TORI, X. 408; REPOS | TO, XIV. 1688; RES | ftTVERIT, 1 X. 4860; RES | TITVJTT, X. 7494; RES | TITVIT, Eph. Epig. VIII. 798; RIIS | TI- TVIT, IV. 806; RES | TITVTA, XI. 5217; RES' | TITVTA, VI. 22205; Res | TITVTA, VI. 25559; RES | TITVTAE, VI. 26342; rES | TliVTAE, VI. 19886; RES | TITVTI, VI. 32873; RES | tituti, IX. 6333; RES | TITVTO, VI. 18200; RES | TITVTORI, XIV. 2917; RES | TITVTVS, VI. 23711, XIV. 1263; RES | TVTAE, V. 8856; RES | TVTO, VI. 2636; RES | TVTVS, VI. 18333; RVS | TICA (2), XI. 1147; RVS | TICO, Eph. Epig. VIII. 795, XIV. 58; SALLVS | TIA, VI. 8454; SALLVs | TIAE, X. 6109; SALLVS | TIAE, XIV. 912; SINis | TERIORI, VI. 26562; SVPER* | TIS, V. 6464; SYNIS | TOR, VI. 17580; TES | TACIO, VI. 10242; Tes | TAMENTI, VI. 33033; TES | TA- MENTO, VI. 2437; TES | TAMENTO, X. 7518; TES | TIMONIVM, XI. 4097; TETRAS | TYLO, VI. 2080; TETRAS | tylo, VI. 2086; THES | TYL- LIS, V. 7935; THAVMAS | TV ..... IV. 383; VENVS | TA, VI. 28454; VENVS | TAE, X. 7964, XIV. 1266; VETVS | TASQVE, VI. 1793; VETVS | TATI, Eph. Epig. VIII. 798; VETVS | TATE, IX. 3162; XYS | TVS, VI. 10531. NST IN | STEIVS, VI. 15641. CONS | TANTIAQVE, X. 4863; CONS | TANTINO, IX. 5955, X. 287, 6837, 6970; CONS | TANTIO, XI. 2697. STK 2 NICO | STRATO, VI. 10503; no | STRVM, VI. 1527; PALE | STRA, VI. 25473. CAMPES | TRI, VI. 32706, IX. 3445; CAS | TRA, IX. 795, XL 2606; CAS | TRENSE, 3 VI. 8547; cas | TRENSI, VI. 29932; CAS | TRENSIS, VI. 4444; CAS | TRICIAE, VI. 16652; CAS | TRICIVS, X. 6338; CAS | TRIS, VI. 3293; CAS | TRORVM, V. 29, VI. 31172; EQVES | TRE, VI. 32328; EQVES- | TREM, IX. 339; eqVes | TRIS, IX. 3158; EVPALES | TRO, XI. 1732; filias | TRO, VI. 19462; HIS | TRICILLA, X. 4036; INLVS | TRIA, X. 5651; INLVS | TRIS, VI. 32023; INLVS | TRE, XL 4118; INDVS | TRIAQVE, VI. 1793; INDVS | TRIENSIVM, V. 7468; INDVS | TRIVM, X. 478; MAGIS | TRATIBVS, XL 3943; MAGIS | TRATVS, XL 1420; magis | TRVM(2), VI. 2068; MES | TRIVS, VI. 2812; MINIS | TRANTIBVS, VI. 2060; MINIS | TRATORVM, VI. 8924; NICOS | trato, VI. 34887a; NICOS | TRATVS, VI. 29248; NOS | TRAE, VI. 10458, IX. 3429, XL 5265; NOS | traM, V. 532; NOS | TRAM, X. 477; NOS | Htl, V. 532; NOS | TRI, XL 1354; NOs | TRO, V. 532; NOS | TRO, VI. 9150, XL 6658; NOS | *RO, X. 6836; PROMAGIS | TRO, VI. 1422; PROS | TRAVIT, VI. 17166, 1716c; SILVES | TRI, XL 2650; SOS | TRATA, VI. 13684; VES | TRAM, VI. 15856, VOS I TRVM, XL 1823. NSTR DEMONS | TRATA, V. 5050. iCf. RESS | TITVTORI, X. 4553. 2 Grammarians' illustrations are nostrum, ve-strum, Caper VII, p. 96 K ; plo-strum, lu-strant, eapi-strum, clau-strum, ra-strum, campestre, Caesellius ap. Oassiod. VII. 205 K ; plau-strum, capi-strum, clau-strum, ve-strum, ra-strum, Albinus VII. 307 K. s Such spellings as CASSTRESE, VI. 8523 apparently suggest the separation of s from the following consonants. Syllabification in Latin Inscriptions 61 A. Summary op Cases in Question (by Consonant Groups) Consonant Groups era ct (ctr<) (net 1 ) dn gn mn pn ps imps') Pt (mpt 1 ) (ptr') sb sc scV sra sp st (nsf) .... nstr 2 ) str 1 Total Foe Received Theory Placing Consonants with the Following 'Vowel Perfect Cases 1 25 is 1 5 25 2 'i 13 5 15 i l 38 1 154 Restored r, . by Conjecture or from Early Copy 13 i 33 Total 1 26 24 1 8 27 2 "% 15 5 19 i 1 51 1 187 Against Received Theory Dividing Consonants between Preceding and Following Vowels Perfect Cases 1 64 2 21 19 30 2 5 33 6 1 5 83 1 8 20 241 6 1 41 590 Restored „ by Conjecture or from Early Copy 10 i 5 5 6 1 1 17 3 48 10 108 Total 1 74 2 22 24 35 2 6 39 7 2 5 100 1 8 23 289 6 1 51 B. Summary of Cases in Question (by Volumes of the Corpus) Volumes of the Corpus For Received Theory Placing Consonants with the Following Vowel Against Received Theory Dividing Consonants between Preceding and Following Vowels Perfect Cases Restored by Conjecture or from Early Copy Total Perfect Cases Restored by Conjecture or from Early Copy Total IV 1 15 92 11 23 4 8 2 i<5 3 6 4 2 3 15 108 14 29 8 10 5 50 289 49 98 67 32 is- 58 12 17 5 3 5 V 63 VI 347 IX X 61 115 XI 72 XIV 35 Total 154 33 187 I 590 108 698 1 In these groups of three consonants it is a question of dividing c-tr, nc-t, mps, mp-t, p-tr, s-cl, ns-t, s-tr, or of placing the three consonants with the following vowel. 2 It is a question here of dividing ns-tr or nstr. 62 Walter Dennison It is a justifiable statement, therefore, based upon the results of an elaborate count, that in the disputed groups of consonants Latin inscriptions nearly uniformly divide between the preceding and following vowels. Of 885 instances, representing a wide extent of territory, 698, or 79 per cent., divide in this way. The percentages by volumes of the Corpus are: IV, 63 per cent.; V, 81 per cent. ; VI, 76 per cent. ; IX, 81 per cent. ; X, 80 per cent. ; XI, 90 per cent. ; XIV, 78 per cent. The percentages for con- sonant groups numerously represented are : ct, 74 per cent. ; gn, 75 per cent. ; ran, 56 per cent. ; pt, 72 per cent. ; sc, 84 per cent. ; st, 85 per cent. ; str, 95 per cent. Noticeable is the small per- centage in the group ran. In divisions like n-ci, m-ps, m-pt, n-st, the syllabification may have been due to a perfectly natural desire to separate n and c, ra and p, n and s quite as much as to place ct, ps, pt, st with the following vowel. 1 Subtracting these cases from the total number, we have 818 cases, of which 663, or over 81 per cent., are against the received theory, sc, sp, st, and str are the only groups which may begin a word in Latin. There are 537 cases of these, of which 463, or more than 86 per cent., divide contrary to the grammarians' rule. Of the groups en, ct, dn, gn, mn, pn, ps, pt, st, scl, and sm, which may begin a word in Greek, there are 276 cases, of which 195, or 77 per cent., divide contrary to the grammarians' rule. 2 Inscriptions of a public character, inscriptions of a relatively early date, and those which are made with evident care, show a marked tendency to avoid breaking words at the ends of lines. It is avoided by the use of ligatures or abbreviations, by super- imposed and circumscribed letters, 3 by spreading words, and by extending the line out into the frame or molding surrounding the inscription. When a division does occur in these three classes i Such cases as SANO | TI and EMP | TVS, therefore, gain greater weight. 2 The figures in Table A (p. 61) would be convincing even if we knew that there was no grammarians' rule, but they are especially significant since they represent the usage that existed in the face of a firmly established, traditional rule of syllabification. 3 This method was employed also by Augustus ; Suetonius Aug. 87 : notavi et in chirographo eius [Augusti] ilia praecipue : non dividit verba nee ab extrema parte versuum abundantis litteras in alterum transf ert, sed ibidem statim subicit circum- ducitque. Syllabification in Latin Inscbiptions 63 of inscriptions, it is nearly always a correct one. On the other hand, private inscriptions of a late date, and carelessly made inscriptions of even an earlier period, show numerous line-divi- sions, and these often violate the regular rules of syllabification. Unfortunately, however, since there are relatively so few inscrip- tions that can be dated with even approximate exactness, it is impossible to make more than this general statement with refer- ence to the comparative frequency and correctness of division in different periods. A study of data earlier than Julius Caesar shows that inscriptions antedating the last half of the first century b. c. rarely admit divisions of word by line. 1 No division occurs in the Epistula consulum ad Teuranos de Bacchanalibus, 186 b. c. (OIL. I. 196), the Lex incerta rep. Bantiae, 133-118 b. o. [ibid. 197), the Sententia Minuciorum, 117 b. o. [ibid. 199), the Epistula praetoris ad Tiburtes, early part of the first century B. o. (ibid. 201), the Lex Cornelia de XX quaestori- bus, 81 b. c. (ibid. 202), or the Lex Antonia, 71 B. c. (ibid. 204). The Lex parieti faciendo Puteolana, 105 b. c. (ibid. 577), has one (etymological) division, ANTE | PAGMENTO, and the Lex Vicana Furfensis, 58 B.C. (ibid. 603), shows TABVLA | MENTA and DEDICATVM | QVE. The Lex Bubria, ca. 49 b. c. (ibid. 205), shows twenty-six cases of division, eighteen of which are correct following the acknowledged rules, five are enclitic, one is etymo- logical, and two involve disputed combinations, namely, FAC | TVM, II. 20, 21, and OM | NIBVS, II. 45, 46. The Lex Iulia munici- palise 45 b. c. (ibid. 206), has fifteen instances of division. Ten of these are correct according to rule, two are enclitic, two are ety- mological, and one is a case in question, PRAEFE | CTVRA, 11. 89 f. Occasionally a real desire is manifested to end a line with a complete word, as in CIL. XIV. 2 696. D • M C • BLOSSIO • FAV FAVSTO • FECE RVNT • HREDES sic PATBONO • B • M Cf. note 3, p. 62, with reference to Augustus' usage in his own handwriting. 2 A note in the Corpus says : " 2 fin. FAV litterae erasae sunt, sed adhuc leguntur." 64 Walter Dennison Similarly a wish to employ correct syllabification is apparent in CIL. XIV. 1134. 1 D M L • IVLI • ALBINI QUIETI- SEQ • L-ALBINIVS PVPVS FIL-PATRI PIISSIM MEMORIAE INSTIT- TVIT The other method (mentioned above, p. 49) by which syllabi- fication is indicated in inscriptions, and which betrays more clearly, perhaps, a conscious effort to divide words into their syl- lables, is that of interpunctuation. 2 Inscriptions of this class are not numerous, about 106 having been found in Italy. 3 In some cases nearly every word in the inscription is divided into its syl- lables by interpuncts, as in CIL. VI. 15546, IX, 4028; but more often this mode of division is employed only in a few words, as in XIV. 2202, or in one word merely, as in V. 5021. In these inscriptions there are 504 instances of syllabification that involve the application of accepted rules of division; of these eleven only violate the rules. The cases in question are: VI*CTORI, VI. 15221. DI • LEC • TVS, V. 6742; E • PIC • TE • SIS, VI. 16898; FAC • TIONIS, VI. 10074; FRVCTIANO, X. 366; FRVC • TI • ANVS, X. 366; OCTAVIA, VI 23357; OC-TABIA | NO, VI. 14560; VIC'TORI, VI. 28905; VICTORI- NVS, V. 5021; VIC • TORI • NVS, VI. 25928; VIC -TVS, VI. 33975." JThe Corpus explains: "6 fin. T erasa, sed legitur." 2 An illustration of this class of inscriptions is given on p. 49. SThey are CIL. V. 783, 1117, 4868, *5021, *6742, *7542; VI. *77, 80, 440, 1070, 1089, 1344, 2428, 2459, 2553, 2557, 2570, 2591, 2669, 2706, 3581, *3689, *5692, 6700, *10074, 10238, 10724, *11682, *12269, *12288, *12524, *12866, *13668, *14560, 14949, *15221, *15546, *15719, *15799, *15917, *16567, *16569, *16898, *17114, 17305, *17308, 17812, 17929, *18034, *19574, *20137 * 21617, *22627, *23357, *25928, *26146, *26345, *26353, *28093, *28427, *28905, *29737, 32460, *32475, 32746, 33206, *33975, *34246, *34714, *35572, *36284; IX. *394, *1520, 2892, *3437, 3721, *3997, *4028, *5546, 5906 ; X. *366, 633, *1344, *1585, *2002, *2166, *2194, *2349, *2723, *2758, *2966, 6354, *6382. *6729, *6792; XI. *3180, *3354, *3854, *3990, *5488; XIV, *34, *608, 1481, *1648, 2170, *2202. Starred numbers contain cases in question. The method of selection was the same as that set forth on p. 50. Such an example as CIL. XIV. 1731 was therefore excluded. "Of. IN -VI 'CITE, VI. 80, which seems to show a pause in pronunciation between O and T ; and VIC • CTOBJNVS, which occurs twice in XI. 5488. Cases like SANO • | TISSIMAE, VI. 26510, where the interpunctuation corresponds with the line division, are entered in the lists constituting the main body of this article, pp. 53 ff . Syllabification in Latin Inscriptions 65 SA | N-CTIS'SI-MAE, X. 2723. DEFVNC-Ta, VI. 36284; /VNC-TA, IX. 5546; SANC'TIS, V. 6742; SANOTIS- | SI -MAE, VI. 15917. 1 SI-GNO, XIV.34. COG -NA- TVS, 2 VI. 18034; EGNA-TPA, VI. 17114; MAG-NA-NIMVS, V 6742; SIG-NVM, XI. 3854. ALV-MNO, X. 2966. A | LVMNIS, X. 6792; OMNI, V. 6742; SEM-NE, VI. 26146. HE DYPNVS, VI. 77. IP-SE, VI. 15546. OPTI | MAE, VI. 15799. OP-TA-TVS, XIV 2202; OP-TIME, VI, 26353; OP TI- MI, XI. 3990; OP-TI-MO, VI. 12288, IX. 1520; OPTIMO, VI. 17308; RUP • TA, VI. 15546; THREPTE, VI. 12524. REDEMPTO, IX. 1520. CRESCENS, VI. 12269; CRIIS-CIINS, X. 2349; CRESCENTI, VI. 16569; CRESCENTI, VI. 16567; CRE | S-CENTIANVS, VI. 15719; ORES- CES, VI. 26353; NAS-CI, VI. 15546; PRIS-CIL- | LAE, XI. 3351; PRSCVS (sic), VI. 11682. COS-MAE, VI. 26345. AC | ES-TIA, XIV 2202; ADRAS-TVS, VI. 21617; A-MET | YS-TO IX. 4028; AV-GV | S-TA-LES, X. 2194; AVGVS • TALIS, VI. 13668; BLAS T | E-NLVI. 11682; CAELES-TI, VI. 77; CALLI S • TE, VI. 5692; CAS TI • NI • O, VI. 14560; CAS • TI • NI • VS, VI. 14560; CEL | ES • TINO, XIV. 608 CVS-TODIAM, V. 7542; EGES-TAS, VI. 15546; E-VHEL-PI | S-TO, IX 394; FAVS'TE, VI. 12866; FA-VS'TI, X. 1344; HOSTILIAE, VI. 19574 HOS-TILIO, IX. 3997; INIVS-TA, VI. 22627; INIVS-TAE, VI. 15546 IVS ■ TI • NA, IX. 4028; IVS • TVS, VI. 22627; *VS • TVS, VI. 34714; MODES TAE, X. 3180; MO-DESTE, X. 6382; MYSTAE, VI. 32475; PANCRIES TVS, X. 6729; PER • EGIS • TIS, VI. 15546; PISTE, VI. 15546; PIS-TES VI. 15546; POS-TERIQVE, VI. 20137; POSTE-RISQVE, VI. 15917, 29737 POS • TE • RIS | QVE, IX. 3437; POS ■ TER | QVE, 3 VI. 34246; PO • TVIS ■ TIS VI. 15546; RES TV -TAE, VI. 28427; SES'TV-LEI-VS, IX. 4028; SES-TV LE-IO, IX. 4028; STA • TV • IS • TIS, VI. 15546; VE-NVS-TA, XIV 1648 BE-NVS-TI, X. 2166.* The mode current in Italy of indicating syllables by inter- punctuation is thus seen with perfect clearness. Of the eighty- nine examples here cited, eighty-two — all but seven- — divide the consonants of the disputed groups between the preceding and the following vowels. The exceptions occur in the groups ct (one instance, net (one instance), gn (one instance), mn (one instance), >Cf. SANC-T-O, VI. 3689. 2Cf. CON-G-NA-TVS, X. 2758; CONGNATO, X. 1220; CON | GNATVS, VI. 25828; and CONGNAT, X. 3408. "Of. POST-TERIS-Q, VI. 35572. *Cf. ERAS-STO, X. 2002. 66 Walter Dennison Summary op Cases op Division by Inter- punctuation Consonant Groups Ct {net) gn TOW pn ps pt (mpt) se sm st Total According to Grammarians' Rule Contrary to Grammarians' Rule 11 4 4 3 9 1 41 82 pn (one instance), pt (one instance), and mpt (one instance ,, none of which may begin a word in Latin. At times a preference for division by interpunctuation to division by line is manifested, for cases like the following repeatedly occur: HERE | N • NIVS, V. 783; KAR ■ N | VN • TI • NO, D | VL • CIS • SI- MO, V. 1117; CO | N • IVGI, VI. 6700; AV • GV | S • TA • LES, X. 2194. Of some interest and value are cases of syllable-division occur- ring within lines: these are caused by (1) some sculptural or sacrificial detail, e. g., VI. 5660, 7930, 22142; (2) symmetrical arrangement, e. g., VI. 5314; (3) a blemish on the surface of the stone, e. g., VI. 25009. Cases of the first class are the most numerous by far, but divisions made for the sake of symmetry, that is, intentional divisions, more often conform, as we should expect, to the acknowledged rules. Additional proof, if it be needed, of the theory of pronuncia- tion advocated in this paper is afforded by certain inscriptions not discovered in Italy but presumably of Italian origin. These include several public documents found in Spain, and the well- known diplomata militaria. In the Lex Malacitani, 81-84 a. d. {CIL. II. 1964) there are 123 line-divisions which are made Syllabification in Latin Inscriptions 67 in accordance with acknowledged rules, and six divisions which violate them. 1 The following are cases in question: FAC | TA, LI; REC | TE, LXV; COG | NITORES, LXV; OM | NI- BVS, LI, OM | NES, LXIV; CIS | TAM, LV; POS | TEA, ZXZV,QVAES | TORVMVE, LX. In the Lex Salpensana, 81-84 A. d. (CIL. II. 1963), all ten cases of division involving acknowledged rules are correct. PRAEFEC | TO, XXV, and CO | GNITA, XXIX, are the only cases in question. The Lex Ursoniensis, 44 A. D. [CIL. II. 5439) contains 124 correct line divisions and two incorrect ones. The cases in question are: FAC | TA, LXXII, RELIC | TVS, XCIII; ST A | GNA, LXXIX; OM | NIA, CXXIX; IP | SE, CXXXII; REDEMP | TORIBVS, LXIX; ES | TO, XCI, TES | TIMONIO, XCV, TES | TIMONIVM, XCV. The S. C. de sumptis ludorum, 176-180 A. D. (CIL. II. 6278) shows seventeen cases of correct versification, with the following : SANC | TISSIMI, 1. 17; PRAESCRIP | TVM, 1. 46; SVMP | TV, 1. 29; FIS | CVM, 1. 13; QVAE | STVS, 1. 7. Of the 112 diplomata militaria known, eighty-seven were found outside of Italy. In dividing words at the ends of lines they follow the accepted rules of syllabification 433 times and violate them nine times. The cases in question are: fun | CTI, CIL. Ill, p. 2000. FVNC | TIS, p. 853, ip | SIS, p. 870, IP | SIS, p. 877; SVBSCRI | pta, p. 851, SVBSCRI | PTA, p. 852, SVBSCRIP | ta, p. 2001; HIS | PANORVM,pp.864,1962,VES | PASIAN,p.877; AS | TVRVM, p. 852; AVGVS | TVS, pp. 857, 1965, HONES | TA, pp. 856, 867, 893, 1973, POS | TEA, pp. 855, 1962, 1975, POS | tea, p. 2001, POS | TERISQVE, pp. 868, 1963, pos | TERISQ, pp. 860, 879. Of these twenty-five cases, three only place the consonant group with the following vowel. The results of the foregoing discussion may be summed up briefly as follows: Latin inscriptions frequently divide a word (1) by interpunctuation within a line, or (2) by placing the first part of the word at the end of one line and the last part at the beginning of the following line. These divisions are not hap- hazard, or due to caprice or individual fancy, but they are made 1 These are of such a uniform character that I quote them : QV | AE, QV | ISQVE, QV | OD (2), QV | EM, QV | OS. 68 "Walter Dennison in accordance with a nearly uniform rule. The results obtained from the study of a large number of such inscriptions should therefore be of value in determining the current mode of syllabic pronunciation. Now, the great majority of divisions involving the application of accepted rules of syllabification are made cor- rectly. Consistently with this we should expect the true mode of syllabic pronunciation to be reflected also in divisions involving disputed groups of consonants, as pt, sc, st. An examination shows that in these cases the first letter of the group is generally placed with the preceding vowel or diphthong, the second with the following vowel or diphthong. But, it may be said justly enough, we are attempting to deter- mine the spoken division of syllables, whereas the inscriptions should illustrate, if anything, the orthographic division. It is perfectly clear, however, from the statistics presented in this paper that the prevailing epigraphic division is not the division taught by the Roman grammarians, and the language of the grammarians is such as to justify the assertion that they were thinking of an orthographic division. It is more likely, therefore, that the average stonecutter divided his lines, not according to the set rules laid down by the technicians, but by his subconscious feel- ing, that is, according to the practice actually observed in his everyday speech. The exceptional cases which we find in the inscriptions are probably due to the carelessness or ignorance of the engraver, or possibly at times to a real desire to follow the formal rule. University op Michigan.