LANGUAGE MONOGRAPHS LINGUISTIC SOCIETY OF AMERICA EDITED BY GEORGE MELVILLE BOLLING Ohio State University AURELIO M. ESPINOSA SAMUEL MOORE Stanford University University of Michigan DANIEL B. SHUMWAY University of Pennsylvania i NUMBER 2 DECEMBER, 1926 eee] THE TEXTUAL CRITICISM OF INSCRIPTIONS BY ROLAND G. KENT Professor of Comparative Philology, University of Pennsylvania LINGUISTIC SOCIETY OF AMERICA 204 St. Mark’s SQUARE, PHILADELPHIA 1926 “LINGUISTIC § SOCIETY. OF "AMERICA Founded d 924 a eae OFFICERS. FOR 1926 President, PROFESSOR MAURICE BLOOMFIELD, Johns Hopkins University, Blab Md. Vice-President, PROFESSOR OLIVER FARRAR EMERSON, V ube sae Reserve. ENS Cleveland, Ohio. er Secretary and Treasurer, PROFESSOR. ROLAND 6. Kexr, Universiyy of Pennsylvania, Philadelphia, Pa. Ait EPSON C ss ah ie eA ft ay Executive Committee, the preceding, ane By pets PROFESSOR LEONARD BLOOMFIELD, Ohio State University, Columbus, Ohio, Proressor EDWARD SaPiR, University of Chicago, Chicago, Ill. _ s PROFESSOR EDGAR HOWARD, STURTEVANT oe eevee Ne Haven, Conn. Committee on Publications : Ne; Aa Editor and Charen, PROFESSOR Gronce Menvme Boling, Ohio Sia Uni versity, Columbus, Ohio. a To serve through 1926 : PROFESSOR SawvEL Moore, University of Michigan, Ann Arbor, Michigan. To serve through 1927 : PROFESSOR Danie. B. SHUMWAY, University of Penn- 2 sylvania, Philadelphia, Pa. To serve through 1928 : _PROFESSOR. AuREIIO M. Espinosa, Stanford Univers sity, oa we op ote: eaae : e The iiepuiste! Soe of America was ‘founded’ in Deceiate: 1924, ink ae a advancement of the scientific study of language. The Society plans to promote this aim by bringing students of language together in its meetings, and by publishing the — fruits of research. It has established both a quarterly journal anda series of mono- graphs ; the latter will appear at irregular intervals, according to the material offered to the Committee on Publications and the funds available for the PUPS Members : » will receive both in return for the annual dués of Five Dollars. — Membership in the Society is. not restricted to professed scholars in linguistics. All persons, whether men or women, who are in sympathy with the objects of the Society, are invited to give it their assistance in furthering its work. Application for — membership“should be made to the Secretary, Professor Roland G. Kent, University of Pennsylvania, cet arte | Pa, ar to whom all. other business communications bs should be addressed. és Manuscripts for publication, Ghee and eons for review aioite ge sent to Professor George Melville Bolling. Ohio State, ny Columbus, Ohio, eh LANGUAGE MONOGRAPHS LINGUISTIC SOCIETY OF AMERICA EDITED BY GEORGE MELVILLE BOLLING Ohio State University AURELIO M. ESPINOSA SAMUEL MOORE Stanford University University of Michigan DANIEL B. SHUMWAY University of Pennsylvania NUMBER 2 DECEMBER, 1926 THE TEXTUAL CRITICISM OF INSCRIPTIONS BY ROLAND G. KENT Professor of Comparative Philology, University of Pennsylvania LINGUISTIC SOCIETY OF AMERICA 204 St. Mark’s SQUARE, PHILADELPHIA 1926 ” i902 f un & ae lala dehy rine ents em hidden Apianaits rset ‘ aaseeerpsinbayae imaliey gen #4 Teamnnient a | alae taal r Z ey: 4 { “f peta ‘HD * A $y rere ee av ce 4 a j tT i re \ as ae Pe . Jags) ay Mae - HEM eiiey sa z ont ania i An A Ain fi PREFACE The first two chapters of this monograph appeared in the Journal of the American Oriental Society, 40. 289-99 (1920), and are here reprinted by permission, with some changes and considerable rearrangement of the material. The other chap- ters have not appeared elsewhere. If the plan of this series included a dedication, the present study would be dedicated to George Melville Bolling, in grate- ful recognition of his continued interest in its partial treat- ment in the /4OS, and of his exhortations that I should bring the subject to finished form. Ry Gr ke Digitized by the Internet Archive in 2022 with funding from Princeton Theological Seminary Library https://archive.org/details/textualcriticism00kent Chapter I. TABLE OF CONTENTS Tin treo tctio tyme tiie a vser etme NS es aria) ial 5 Old Persian : The Inscription of Darius the Great. 8 Greek. eine. Lochians Leeaties i otc ale ais vutaate | 17 scat h hen abet ACIS coed Patan. MeN auras 22 Umbrian : The Bronze Tables of Iguvium..... 26 Latin : The Lex Julia Municipalis............ 47 Latin: The Edict of Diogletiany Poi ve 57 RE ELITE TG Woh magn i 2°88 aie Bo Raina LS A 67 a ae ) 71, Lia , | a A WN Te AU tad ch ae anne a AM ry A OE REO g ou “ye wD us agin voile ge rah Un erm coe Rm dieisint tT: hes a . con) ie ie PNR : paioneneh gd’ E 400 7 oe . } i te Ray be nates: 6 oye wnortl ot T axils F Me “al CN a th eae £toiiaghlegl cit ae citi, ; 7 1 yeaa Panay sn oo pitt ott i aan : aka ai! f aa ’ ; p ite ay er ht ao, [hoon Ae eae CHAPTER I INTRODUCTION Scholars are now well equipped with treatises upon the corruptions which are found in manuscripts, and upon the manner in which edi- tors must proceed as they make up a corrected text. We may mention, in this connection, the following selected authorities, most of which contain references to earlier works : James Gow, A Companion to School Classics }, 47-66 (1891). W. M. Lindsay, An Introduction to Latin Textual Emendation based on the Text of Plautus (1896). Harold W. Johnston, Latin Manuscripts, 79-99 (1897). F. W. Shipley, Certain Sources of Corruption in Latin Manuscripts : a study based upon two manuscripts of Livy : Codex Puteanus (fifth cen- tury), and its copy Codex Reginensis 762 (ninth century), in Amer. Journ. Archaeology, 7. 1-25, 157-97, 405-28 (1903). J. P. Postgate, Flaws in Classical Research, in Proceedings of the Brit- ish Academy 1907-08, 161-208. J. P. Postgate, Textual Criticism, article in The Encyclopaedia Britan- nica, 11th edition, 26.708-15 (1911). Louis Havet, Manuel de Critique Verbale appliquée aux textes latins (1911). Alfred Gercke, Methodik, in A. Gercke and E. Norden’s Finleitung in die Altertumswissenschaft, 1.1-130 (ed. 1, 1910; ed. 2, 1912). Theodor Birt, Krittk und Hermeneutik, in Ivan von Miiller’s Hand- buch d. klassischen Altertumswiss., 3d edition, vol. I, pt. 3 (1913). F. W. Hall, Companion to Classical Texts (1913). J. P. Postgate, in J. E. Sandys’ Companion to Latin Studies, 791- 805 (1913). R. C. Jebb, in Leonard Whibley’s Companion to Greek Studies, 720- 33 (1916). Well adapted as these are for their purpose, which is to acquaint the scholar with the ‘rules of the game’ in the criticism and the emendation of manuscript texts, as he edits or elucidates them, they do not so well serve for the handling of inscriptional texts. For the manuscripts may be the results of one copying after another, each » 6 LANGUAGE MONOGRAPH NO. 2, 1926 new copy suffering perhaps additional corruption at points which are already corrupt ; but an inscription is in practically all instances merely transferred from a manuscript draft to its permanent position on stone or bronze, and therefore less subject to complicated corrup- tion. At the same time, the speed with which a copyist transcribes with pen upon paper or upon parchment, is a factor leading likewise to greater error than the slowness with which the engraver transfers his text, letter by letter (not word by word), to its place of permanent record. On the other hand, the inscription may be copied in an alpha- bet differing from that in which the original draft stands, and this will produce a series of corruptions to which manuscript copies rarely afford parallels, except that we may compare the manner in which Greek words in Latin texts have been miscopied by the scribes; or unless we include within our field the manuscripts of India and of the Avesta. For these reasons, it is my intention toexamine critically the accept- ed or suspected errors in certain inscriptions. The inscriptions which have been chosen for this examination are inscriptions of formal char- acter; these are normally written with a considerable degree of care, and should therefore not contain many errors of a haphazard nature. In this manner we may determine precisely the kinds of errors which actually do occur in inscriptions. The results and the principles thereby reached, even if not revolutionary, will be a firm basis on which philologists may found their utilization of the linguistic evidence furnished by inscriptional forms — evidence which, for ancient lan- guages, has no rival for validity, excepting only the remarks of con- temporary writers upon points of grammar and pronunciation. Variations from an original copy may be classified in several ways. Johnston (pp. 80 ff.) prefers a scheme based chiefly upon the causes : (1) Unavoidable changes ; (2) Intentional changes ; (3) Accidental changes, including (a) those of the ear, (b) those ‘of the eye, (c) those of the memory, (d) those of the judgment. Lindsay (p. 10) groups them mainly by their results : (1) Emendation, (2) Transposition, (3) Omission, (4) Insertion, (5) Substitution, (6) Confusion of Letters, (7) Confusion of Contractions. Neither of these classifications, how- ever, is free from its disadvantages, since the divisions and subdinisinge prove not to be mutually exclusive in practice; and for dealing with inscriptions, where the corruptions are not of such complicated nature as those in manuscripts, it seems better to revert to the old and simple classification of (1) Loss, (2) Addition, (3) Change, with subdivisions which will be developed as met with. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS a It must be understood that it is not within the province of the pre- sent investigation to include phenomena which rest upon a conven- tionalized orthography or upon confusion in pronunciation. In Latin inscriptions of the older period, the failure to double the consonants in writing would not here be handled, since that is a convention of the alphabet in use ; but a doubling of a consonant which should not be doubled would be taken into account. Similarly, in a Latin inscript- ion of the later period, the variation between e and ae results from confusion in the pronunciation, and is valuable as evidence for the pro- nunciation of the time; it isnot the purpose here to deal with such matters. The editor of a text must, it is true, eliminate such corrupt- ions as well as the grosser errors (e. g., Plaut. Epid. 231 crutulam BJ, for crocotulam, found in A) ; but errors or orthographic variations which rest merely upon conventions in spelling and confusion in the sounds, must in inscriptions be left as precious evidence for the stu- dent of philology. Our purpose is, then to prepare the text of certain inscriptions in such a way that the philologist may use it with confi- dence in reconstructing the history of the language ; and to fix the rules and principles for handling other inscriptions. Again, we are not do deal with restorations of missing characters, which, so far as no traces remain, are entirely conjectural ; nor may we accept such conjectures in poorly preserved portions and then seek to find errors in the few characters which are to be read ; such a pro- cedure would be quite unscientific. Our attention is to be directed to those words and characters which are legible, and our field overlaps that of conjectural restoration only when characters are preserved in part, so that they may be read in more than one way ; in this situa- tion we can hardly draw a definite line of demarcation between resto- ration and textual criticism. For this purpose the following inscriptions have been selected : I. Old Persian : the Inscription of Darius the Great, at Behistan. II. Greek : the Bronze Tablets with the treaties between Naupactus and the Hypocnemidian Locrians, and between the Oeantheans and the Chaleians. Ill. Oscan : the Tabula Bantina. IV. Umbrian : the Bronze Tables of Iguvium. V. Latin : the Lex Julia Municipalis. VI. Latin: the preamble to the Edict of Diocletian fixing maximum prices. CHAPTER II OLD PERSIAN : THE INSCRIPTION OF DARIUS THE GREAT The Inscription of Darius the Great, cut high up on the face of the cliff at Behistan in Western Persia, records the accession of Darius to the throne of Persia and his successful suppression of a number of revolts against his power. It is engraved in a cuneiform syllabary, the conventions of which are well determined and familiar to scholars (cf., for example, E. L. Johnson, Historical Grammar of the Ancient Persian Language, 29-35 ; also R. G. Kent, JAOS 35.325-9, 332, on spe- cial points). The text is presented in the cuneiform syllabary, with transliteration, translation, and critical annotations, by L. W. King and R. C. Thompson, The Sculptures and Inscription of Darius the Great on the Rock of Behistiin in Persia, 1-91 (1907), a publication of the British Museum embodying the results of their reexamination of the rock and its inscription ; this is the definitive text. A translitera- tion and translation, with critical notes and vocabulary, is contained in H.C. Tolman, Ancient Persian Lexicon and Texts (1908) ; and the same scholar’s Cuneiform Supplement (1910) contains an autographed copy of the text in the cuneiform, and as an appendix E. L. Johnson’s Index Verborum to the Old Persian Inscriptions, which is a complete word concordance : these two volumes are Nos. VI and VII in the Vanderbilt Oriental Series. These will be referred to hereafter by easily recognizable abbreviations. The most striking feature of the inscription is the extreme care with which it is engraved, demonstrable errors being very few, now that the text has been definitively recorded in KT. But this care is not to be wondered at; for without it the record would have become a hodge- podge, since 23 of the 36 characters of the syllabary are transfor- mable into other characters by the addition or the subtraction of a single stroke, and eleven of the remaining thirteen are convertible by subtracting one stroke and adding another — in some cases this being merely a placing of the same stroke in a new position. Besides 8 KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 9 ary Weowaaed cos UEL bk arama Lauria ss 1 u ke ku ou er ce r y’ t 11 Gomeiaaaaain 69 epee OEE [Void Ceey i ‘i Q4 d* di qd p? K« | ss oad | Ke jie be nt n® m* m* gate atl Goria wom Eabe autl ={Kireqar pen bon tee m" y" yr y" B ye Sh vl Nee 9 ACA EAL MIEN Ce MER Ca yi 54 guard >’ for the two words and the following pari[ba|ra leaves the final verb without its pronominal object, which is unusual in the inscription, and makes the uncompounded sia assume the s which would be proper only after prefixes ending in 7 or u and after the reduplication in 7. Yet as the 5 is found in aiitata and extended in niyastayam niyastaya, such an extension to Sta is not too unlikely. Tolman’s emendation, making the }" a miswriting for the word divider, and ¢a the pronominal object of the following verb, is improb- able, since the demonstrative stem fa- is not found as a separate word elsewhere in the Old Persian inscriptions, and the addition of 14 LANGUAGE MONOGRAPH NO. 2, 1926 the two strokes to the divider so as to make the 5* is an unlikely error. 15. ult*a[n* << n*Jam* 4. 83 = U]ta[na n]ama, is the probable res- toration of the passage, but KT 76 ftn. 2 state that the gap has room for only two characters, not three. It is likely that either the first or the second 2* was omitted ; an omission which may be termed Tele- Haplography, and is to be defined as the failure to write one of two identical characters or groups of characters which are not contiguous, though the intervening character or characters remain. There is a pos- sible alternative, that it was the divider which was omitted ; since the symbol * consists of two horizontal strokes followed by the divider, the omission of the divider at this point would be an instance of Pseudo-Haplography. 16. n*buk"urt{etr® 1. 5-6 for n*b*uk"ud*r"cr’ = Nabukudracara, has lost the d*. This is an easy example of Pseudo-Haplography, since u is the same as a? with a prefixed divider : thus ud* == < d‘d*. 17. bab*rtuv’ I. 11 for btabtir*uv’ == Babirauv. The i of the second syllable is omitted, although the preceding consonant has inherent a, and neither the preceding nor the following character Lata resembles 1. This must be classed as simple Omission. I hesitate to list further possible errors from the text of the Behis- tan inscription. Scholars have made many conjectures, as may be seen by examining the critical apparatus in Tolman, Lex., but most of the conjectures do not deserve consideration since the minute collation by KT. The following might, however, be listed, even if only to support the actual text : I. 22, 4. 66-7 ufrastam; 4.38 ufrastam; 4. 69 ufrasta- (cf. OP Stud. § 64-§ 69.) The variation between s and 5 is merely the result of level- ing (OP Stud. 351, ftn. 4). 30 hamata for *hamamata almost certainly represents the actual pronunciation, and is therefore not an example of Haplography, but an example of Haplology (OP Stud. § 46). 1. 86-7 u5*/b¢ar*im® is by many scholars supposed to lack two signs at the end of the prior line : u5¢r*/b¢artim* == ustrabarim ‘ camel- borne, ’ cf. Avestan w5tra ‘camel’. But usabarim may be correct, if usa was a doublet form of ustra as asa was of aspa ‘horse’ (OP Stud. § 47-§ 51). 1. 87 as*m® = asam ; 2. 2,71, 3. 41, 72 as*b*arib%is¢ == asabari- bis. The establishment of asa as a doublet of aspa makes emendation of these forms superfluous (cf. OP Stud. § 50). 74 berth’an'm* = harbinam ‘tongue’. KT 36 ftn. 4 explain it as KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 15 from the root in Latin sorbed ; this eliminates the need of correction (cf. Tolman Lex. 134). 2.75 and 89 ucasma ‘eye’ may be correct, though somewhat indis- tinct on the Rock (cf. Weissbach ZDMG 61. 726, quoted by Tolman LxS75): 3. 8 Oakatam is the correct singular form, and not an error for Qakata, which is the correct plural form, required in the other eigh- teen passages where the word is used (cf. Bartholomae, as quoted by Tolman Lex. 95). 4. 6 adamsim : the explanation of the difficult enclitic is given OP Stud. § 52-§ 63, especially § 63; cf. also R. G. Kent, JAOS 41. 74-5. 4. 65 ++ m°n"uvttm® or ++ tn*- or++t“un"- : the reading is too uncertain for the passage to be used here. 4. 89 i[ ya] dipi (the illegible gap has space for but one character, according to KT 77 ftn. 5); 4. 90 iya[d|ipi. This zy* is not to be emended to iy"m* == iyam, but is to be read iy, from Indo-European *i (OP Stud. 348, ftn. 2). 5. II uta < daiy < marda ‘and he annihilated them ’. Objection has been taken to daiy as an orthotone and as an accusative. But the change of enclitics to orthotones and vice versa can be parallel else- where, and the form of the accusative plural in Old Persian, outside the enclitic pronouns (which can have no nominative), is invariably that of the nominative plural (OP Stud. 336, ftn. 2), notably in the third person pronouns (avaiy, imaiy, tyaiy). The orthotone value and the nominative form as accusative therefore go hand in hand, and mutually confirm the reading of the text rather than make it suspi- cious. In the passages of the Behistan Inscription which are surely or prob- ably miswritten, therefore, we have found errors of the following kinds, which have been defined as they were met ; italicized numerals indicate examples which are of uncertain validity or may be differently interpreted and classified under another heading also : I. Errors of Omission : Omission, unmotivated : 10, 17. Omission for Phonetic Accuracy : 6. Haplography : 5. Haplography with Skipping : 2. Tele-Haplography : 15. Pseudo-Haplography : 3, 4, 14, 16. * 16 LANGUAGE MONOGRAPH NO. 2, 1926 II. Errors of Addition : Dittography : 11. Tele-Dittography : 1. Ill. Errors of Change : Change by Addition : 7, 8, 9. Change by Subtraction, including Semi-Haplography : L2, 13 CHAPTER III GREEK : THE LOCRIAN TREATIES As examples of Greek inscriptions, two bronze tablets have been selected, of which one bears a treaty between Naupactus and the Hypocnemidian Locrians, and the other a treaty between the Oeantheans and the Chaleians. IG IX 1.334 THe TREATY OF NAUPACTUS AND THE HYPOCNEMIDIAN LOCRIANS. The treaty between the Naupactians and the Hypocnemidian Locrians is inscribed on the two sides of a small bronze plate. It is No. 55 in the collection of inscriptions in C. D. Buck’s Introduction to the Study of the Greek Dialects, pp. 214-18 ; where there is a brief bibliography. For the present purpose, it is necessary to use also a facsimile, such as that given by E.S. Roberts, An Introduction to Greek Epigraphy, pp. 236- 7. This treaty is composed in the Locrian dialect, of which it is the most important extant document. The alphabet of this inscription lacks the characters » and w, but has f, b, and ?. Of these, - is fouud regularly as an initial where it belongs, but with less consistency in the interior of words. H is found only in the initial position, but even there it is not used in combina- tion with F, nor in the words co1x 2 == dora, rotiat 7, 16 = tortie, vdpray 45 == d8ptav ; amodts 26 == % wdAtg and hantpowne I == & extfouxia show the article with and without the rough breathing. As the Attic word hyéea has no hereditary justification tor the rough breathing, it is pro- per to find the Locrian equivalent without the ) : aitapagsyv 33, tord- Povt’ auspars 42, tTetkPovt’ ducer 42. So also xattPouevov 31 shows no aspirate beginning the second element ; but on this subject of the breathing in Locrian, see F. Bechtel, Die griechischen Dialekte, 2. 5-6, Chiilyer37. The sign ? is regularly used before o and po. N is used for the nasal before labial and palatal stops as well as before dentals. The preposi- tions xatx and xot: lose the final vowel betore initial +, and the doub- Y, 18 LANGUAGE MONOGRAPH NO. 2, 1926 Ae CP RB OA en a ene eon PON AP Oe en ee i de ve ee ead ASAT eo KE DE HirO. Oahu A Tey iv) ily A | eae ei eV ae ae A Ai Bog hl Ripa 4 Fado ddeth » On onlin Reet I ANY eg Oris hh EDEN OE a FIGURE 2 THE ALPHABETS OF THE LOCRIAN TREATIES The first alphabet is that used in the treaty between the Naupactians and the Hypocnemidian Locrians; the second and the third are those used respectively in the two parts of the treaty between the Oeantheans and the Chaleians. Some letters do not appear in all of the documents ; others are found in more than one form, and are so given above. led + is then written single : xatov3e I == x&(t) tévde, otovg 32 = ni(+) sobs. The preposition éx or 2§ assimilates the final consonant to a following initial consonant, and the geminate is then written single, except in one instance, where the assimilation is not complete : eyvav- Tanto 15 gy Navzdgxtov ; but the preposition év keeps its v before an initial vy : evvaunaxtor 17 etc. == év Navnaxtw, but evauraxzo 8 etc. = i(x) Navnéztov. A sign of word division, consisting of three dots in a vertical line, is freely used at the end of words, usually where there is a pause between phrases ; it is not always set at the end of a phrase, as in 13, nor at the end of aclause, as in 7, nor at the end of a sentence, as in 12, but yet may be set even between the article and the noun, as in 5, and before enclitics, as in 38 and 39. Elision, crasis, and apocope are not infrequent. The words and phrases in this inscription, which are certainly or possibly miswritten, may now be examined. 1. homobevov 2 = hdnw(s) Eévov. As the final ¢ has the same sound as the second part of the initial §, its omission might be regarded as by dissimilation, with partial phonetic justification ; but it is hardly to be regarded as a mere orthographic simplification of a repeated let- KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 19 ter, since the two letters ¢ and & are of quite different appearance in the Locrian alphabet. On the other hand, it is entirely possible that szw has in this dialect begun to encroach upon énw¢ ; for ézw, written hore, occurs several times in this inscription, though always introduc- ing a subordinate verb. Either then this writing is correct, or the loss of the final consonant may be regarded as Omission, with phonetic dissimilatton asa contributory factor. 2. atnadethetar 3 == at xa SelAnrar. This clause, immediately preced- ed and followed by the interpuncts, is repeated, and is a possible in- stance of Dittography ; but as there are numerous formulaic repetitions in the inscription, several of which are listed by Buck in his note to the passage, it cannot be considered as more than a possible instance. 3. xataAcimovita 7 == xatahstnovta. The divider is here set incor- rectly in the interior of the word, which is an error of Addition. Pos- sibly the fact that ov is a common word ending and ta is a common form of the article, contributed to its insertion here. 4. vets 10 = petx. The p lacks the fourth stroke, thereby becom- ing in reality y ; a Change by Subtraction. 5. amovttoy II == an’ ’O(xo)vttwy. An excellent example of Haplo- graphy, in which the group zomo is reduced to a single zo. 6. texva. 12 == téyve. But in view of teyva. 38, the x for x is to be regarded not as an error of writing, but as a testimony to the slight difference between the sounds indicated by the two letters, the only difference being that the aspirate y has a puff of breath following the k- sound, which is lacking in the non-aspirate x. 7. hosottg 14, 35-6, 38 == bdotrg. The regular doubling of ¢ in this word, in contrast to the frequent simplification ofa doubled consonant, seems to indicate a particularly sharp pronunciation of the o, rather than an error of writing. If however it be a mere error, it is one of Dittography. 8. toubumoxvayrdiorg 21 == toi(¢) Humoxvays3torg. This is an Omission, without apparent motive or assignable cause. 9. vaumaxtig 22 == Navmduri(4¢ tt)c. A typical example of Haplo- graphy with Skipping, in which og has been lost along with one set of the repeated letters between which it stands. 10. howtwverxanratesevttpotes 35 == hotzivés xx ’martég Evtepot (wrt). Here coves has become e¢ by the influence of the preceding <<, found twice’ in the clause as word-ends. This error may be termed Tele-Dit- tographic Change. 1. homovtioy 39 == Horovtiwy, which has no right to the A, as is 20 LANGUAGE MONOGRAPH NO. 2, 1926 seen from the correct writings in 13-14, 14, 33, but has acquired it by the influence of the forms of Humoxvauidtes 1, 5, 6 bis, etc. This is an example of Addition, by the influence of a word associated in meaning. 12. varraxtiov 40 == Narraxttwy ; this is, in view of the writing vaur- found eighteen times, an orthographic variant to represent the same pronunciation, and testifies alike to the diphthongal value of av and to the semivocalic value of f. 13. Wagit/Ew 45-46 == vaorkw, with dittography of the & induced by the division between the two lines. This is not a bungling attempt to divide the lines of writing at the division between the syllables, intended to show that part of § went with each syllable ; for, of the 35 lines of this inscription which end in the interior of a word, no less than 21 fail to divide at the end of a syllable. IG IX 1. 333 THe TREATY BETWEEN THE OEANTHEANS AND THE CHALEIANS. The treaty between the Oeantheans and the Chaleians is likewise in the Locrian dialect, and is on a bronze tablet inscribed on both sides. It is Buck’s No. 56, on pages 218-19 of his Greek Dialects, and is given in facsimile by Roberts, Introduction to Greek Epigraphy, p. 239. This inscription contains however two documents, the treaty which gives it itsname, and certain provisions on legal proceedings. Curious- ly also, two engravers are involved in the work ; the first did a trifle over seven lines of the obverse, the second did the remaining line and a fraction of the obverse and the nine lines of the reverse. They used slightly differing alphabets, as appears in the Table, and the first used the triple interpunct, the second a double interpunct (once a triple interpunct). This does not however affect the dialect. The alphabet lacks 4, w, and 9, but has ¢ and 4. H is lacking in the nominative singular masculine of the article (7, 11 bis, 14), but is written elsewhere when it is etymologically justified ; cf. F. Bechtel, 1. c. Doubled consonants of whatever origin are simplified : etag 1 == &(z) cag for 2x t- 3 atticudcr 2-3 == at ci(¢) ovA@ ; Oaracag 3 == Oaddooac. Elision and crasis are found. The interpunct separates phrases, less often single words, but is often lacking where it is expected. There are a few words in which an error may be suspected : 14. hayev 1, 3 == oye. The h is wrongly prefixed, despite the KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 21 absence of in the compounded omayoy 11 == (with crasis) 6 éxayuv ; cf. also b3pxov 18, hopxopotas 16, 17 = hopxwpydtac, but ze/vroextay 16-17, and F. Bechtel, Gr. Dial. 2. 5-6. The incorrect h of hzyev seems to be a downright error, though one may suspect the influence of some word related in meaning, such as apnalw, aioe, etdov. 15. he/prodrovopAetopotucvAacat 5-6 == hutddrov doAcitw A Ott cuAdcoat. The peculiarity is in fott, which does not have an etymological right to the digamma ; but Bechtel, J. c., has solved the difficulty by regard- ing f as an « minus the lowest horizontal ; an error of Change by Subtraction. 16. ootvecto 9 == Owifjotw, Attic -xc0w. The first character lacks the dot or cross in the center that would make it into a 6; a Change by Subtraction. 17. Fasotos 14 == fastic, cf. f/acrdy 14-15. The doubling is evi- dently similar to that in hogottg in the preceding inscription, but in view of the spelling in 14-15, with a single, it may be regarded as an error. In fact, there are only three other places in this inscription where a letter is repeated, tao cvvGeA%: I5 without interpuncts, and two in- stances in which a final « is immediately followed by an initial a, without interpuncts : yeewataattiov/Aor 2-3 = yovpata at t(¢) ovAd ; tovdecvAovtaavatocuAcy 3 == dv 38 avddvta avatw(¢) cvAyy. In these few passages of the two Locrian documents, we find accord- ingly the following kinds of error : I. Errors of Omission : Omission, unmotivated : 8. Omission, helped by Phonetic Dissimilation : 1. Haplography : 5. Haplography with Skipping : 9. II. Errors of Addition : Addition by the influence of familiar words : 3. Addition by the influence of a word associated in meaning : II, 14. Dittography : 2, 7, 17. Dittography induced by the division between lines : 13. Ill. Errors‘of Change : Change by Subtraction : 4, 15, 16. Change by Tele-Dittography : 10. IV. Examples listed, but perhaps not errors : 1, 2,6, 7, 12, 14, 17. CHAPTER IV OSCAN : THE TABULA BANTINA The Tabula Bantina is a small bronze tablet inscribed on both sides; the inscription on the one side is composed in Latin, and that on the other side is in the Oscan language, but in the Latin alphabet. The two inscriptions are of different subject matter. The Oscan inscription, which is to come under our attention, is Buck’s No. 2, in his Grammar of Oscan and Umbrian, pp. 230-9, and is presented in facsimile in Buck’s Plate II, as well as in Zvetaieft’s Inscriptiones Oscae, Plate XIX. The text is given, with detailed criti- cal annotations, by von Planta, Grammatik der oskisch-umbrischen Dia- lekte, 2. 494-7, and by R. S. Conway, The Italic Dialecis, 1. 22-9. The words, many of which are abbreviated, are separated from each other by a single interpunct, the presence or absence or which is often difficult to determine. This interpunct is certainly lacking in the following groups : pieisum brateis 6, comono ni 8, hafieist meddis 8, exum nep 10, pru meddixud 13 (cf. pru.medicatud 24), comono ni 14, pon pos- mom 16, con preiuatud 16 (cf. com.preiuatud 15), ni hipid 17, paet eixeis 22, pae ancensto 22. Whether suae pis ‘si quis’ should be writ- ten with the separating dot, as in 4, 13, 17, 23, 25, or without it, as in II, 12, 17, 20, 26, 29, may fairly be disputed. The interpunct is wrongly inserted in the interior of the following words : com.parascus- ter 4, pocapi.t 8, ex.elg 11 (=exeic), medicat.inom 16, anget. uzet 20, ex. aiscen 25. | There are the following errors of writing remaining for discussion : 1. pocapi.t 8 is the same word as plocapid 31 (Avellino fragment) and pukkapid ofthe Cippus Abellanus 52. It owes the final ¢ apparent- ly to the influence of the final of post, which immediately follows ; the t of pocapit is therefore an instance of Tele-Dittographic Change. Buck, § 127. 1a, calls it merely an error, and von Planta, 1.578, suggests some other possibilities. One may note that the final ¢ of the preposi- tion dat, from earlier *dad, seems to represent a generalization of the form before voiceless sounds, as it occurs four times (6, 8, 9, 10 ; cf. Buck, § 190. 3 a). 22 KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 23 2. post. post 8 is a normal example of Dittography. 3. hafieist 8 seems to be the reading of the tablet, for hapiest. The p has been changed into f by the addition of the medial horizontal stroke, a Change by Addition. The character or characters between ¢ and # are more puzzling ; Conway, 1. 25, thinks that there is an 7 which has been rather unskillfully changed into is, the 7 being merely the vertical hasta of the r, remaining after the correction. Von Planta, 2.495, reads is ; Buck, p. 231, reads r. But a'careful examination of the facsimiles shows that the vertical hasta is crowded in, much closer to the preceding and the following letters than is normal on the tablet ; which indicates that it a later insertion, though made by the original engraver. It so happens that in the next line, less than three inches away, there is an omitted 7: eizasc for eizaisc ; 1 suggest there- fore that on scanning the tablet for errors, the engraver discovered the omission in eizasc, and inserted the i — but in the wrong word. The same letter, s, follows the omitted 7 of eiza(z)sc and the inserted i of hafieist. This error, ifthe explanation be correct, may be termed Addition by Metathesis. 4. pous 9 stands for pus, with ou for u by the influence of the ou of touto, which immediately follows ; an example of Tele-Dittography. 5. deiuatuns 9 is certainly a perfect passive participle in the nomi- native plural masculine ; if it be a correct form, it must have been transformed into an n-stem (cf. Buck, § 244. 2a). But it is more pro- bable that it owes its 7 to Tele-Dittography from deicans, the second word thereafter. 6. stom 9 stands for siom, the 7 being transformed into ¢ by the addition of a cross-stroke. 7. eizasc 9 stands for eizaisc, with omission of the 7. This seems to be the 7 which was later inserted in hafieist 8, in an effort to correct the omission ; cf. on No. 3 above. 8. tadait 10 stands for tadaid, with a change of the letter. Perhaps the Latin, with the ending ¢ in the third singular, has affected the writing of the word. 9. fepacid 10 stands for fefacid, with f becoming p by the omission of the middle cross-stroke ; a Change by Subtraction. 10. docud 11 stands for dolud, with change of one letter. Possibly the c of deicum in the preceding line, almost directly above, has had an influence. ‘1. ex.elg 11 stands for exeic, with an incorrect interpunct, and two letters changed by the addition of single strokes. 24 LANGUAGE MONOGRAPH NO. 2, 1926 12. con 16 for com is probably not an error of writing, but a use of n to denote a less heavily pronounced nasal ; cf. von Planta, 1.307. 13. comonom 17 displays a final m which is lacking in comono 5, 7, 8 bis, 11, 14, and therefore, in view of the regularity with which final m is written in this document, seems to be an error. I suggest that it is a Dittography of the following mi, a group of four strokes easily changed into m ; cf. the present phrase comonom. nihipid with comononi. hipid 8, 14. : 14. sansae 19 stands for bansae, with the initial letter changed by the influence of the initial of the second syllable ; a Tele-Dittographic Change. 15. tautam 19 is for toutam. Either the a in the first syllable is by the influence of the vowel of the second syllable, or of the vowels of the preceding word sansae, a Tele-Dittographic Change; or the engrav- er has replaced the unfamiliar Oscan diphthong ou by the familiar Latin au, because he was using the Latin alphabet, and ou is so abnor- mal in Latin as to be virtually non-existent. 16. ausc 20 stands for iusc. The and the first stroke of the w meet at the top, and they are united midway byashort horizontal line ; this produces an apparent ligature au. A Change by Addition. 17. cebnust 20 has been suspected of being an error for *bebnust, a reduplicated form, the c being by the influence of censtomen, the second word preceding. If we accept this interpretation, there is a Tele-Ditto- graphic Change; but a reasonably certain interpretation of cebnust as it stands has been given. Cf. von Planta 2. 329, with bibliography. 18. licud 24 is the apparent engraving for ligud, with absence of the stroke which differentiated ¢ from c, a Change by Subtraction ; but the reading of the character is not absolutely certain. 19. acum 24 stands for agum, unless the root ag- has been changed to ac- by a leveling process ; cf. Buck § 159a. This seems more pro- bable than that there is here an error like that in the preceding example. 20. phim 25 is an error for pim, either from the Latin confusion in the employment of p and ph (Buck, p. 144 ftn.), or by the influence of the following prubipid (von Planta 1.62). In the former case it is an error of Addition by Latin influence, in the other it is an example of Tele-Dittography. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 25 The Tabula Bantina contains therefore errors, certain or probable, of the following kinds : I. Errors of Omission : Omission, unmotivated : 7. II. Errors of Addition : Addition by Metathesis : 3. Addition by Latin iufluence : 20. Dittography : 2. Dittography, with Ligature : 13, Tele-Dittography : 4, 5, 20. Ill. Errors of Change : Change by Addition : 6, 11. Change by Ligature and Addition : 16. Change by Tele-Dittography : 1, 10, 14, 1s, 17. Change by Subtraction : 9, 18, 19. Change by Latin influence: 8, rs. IV. Examples listed, but certainly or probably not errors : JS, 12,17, 19. CHAPTER V UMBRIAN: THE BRONZE TABLES OF IGUVIUM The Iguvine Tables are seven bronze plates, which are inscribed on both sides with a document in the Umbrian language, giving the ritual ofthe Atiedian Brothers ; except the third and the fourth plates, which are inscribed on one side only. They were found in 1444 at Gubbio, the ancient Iguvium, where they are still kept in the town-hall. Two other similar tables were found at thesame time, but have been lost. The most important literature onthem is as follows : S. Th. Aufrecht and A. Kirchhoff, Die umbrischen Sprachdenkmaler ; 3 vols,, Berlin, 1849-51. M. Bréal, Les Tables Eugubines ; Paris, 1875. A separate album con- tains excellent facsimiles of the inscriptions. F. Biicheler, Umbrica, Bonn, 1883. R. von Planta, Grammatik ie oskisch-umbrischen Dialekte; 2 vols., Strassburg, 1892-97. The text of the Iguvine Tables, interpretation, critical and explanatory commentary, and glossary, are to be found 2.557-85, 667-71, 724-64. R. S. Conway, The Italic Dialects ; 2 vols., Cambridge, 1897. This contains the most careful critical study of the text, 1.399-433, witha glossary, 2.596-672. C. D. Buck, Grammar of Oscan and Umbrian, Boston, 1904. The Igu- vine Tables are given on pages 260-310, with interpretation, critical and explanatory notes, and a glossary on pages 327-52. R. G. Kent, Studies in the Iguvine Tables, in Classical Philology, 15.853-69. This is listed not for its importance, but because it will be convenient to refer to it rather frequently, as it also discusses possible errors in the text. | These works will for brevity be referred to by the names of their authors; excepting the last, which will be cited as Iguv. Stud. Epigraphically, the Iguvine Tables fall naturally into two divisions, according to the alphabets in which they are engraved. Tables I, IJ, Ill, IV, the obverse of V, and the first seven lines of the reverse of V, 26 KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 27 A 4 a b q 7 die Jv Fok © |b ae Wik Yo! MA m von aL Gy 2M s XV Vou 8 f ag QO 9 FIGURE 3 THe NativE Umsrian ALPHABET are in the native Umbrian alphabet. The remainder of V, and all of VI and VII, are in the Latin alphabet. Tue TABLES IN THE UmBriAN ALPHABET. The native Umbrian alphabet is given in Figure 3, with the custo- mary transcriptions of the letters. The five-stroke m is found in Tables I-IV, but is replaced by two-stroke min Table V. The form of t in which the cross-stroke does not cross the hasta is used in Tables III- IV. San (M) is twice used tors (salu II a 18, serituIla 24). Theta (6) is twice used for ¢ (furfa) Ib 1, purtuvidu IV 20). Otherwise the writing is exceptionally regular. The words are separated from each other by a double interpunct, which is identical with an English colon. This is not placed at the beginnings of lines, but is normally placed at the ends, if, asis usually the case, the line ends with the end of a word. For a properly faithful representation of the text, words and passages cited for epigraphic pur- poses will be provided with the interpuncts before and after them precisely as on the original Tables; an interpunct which is lacking in the Tables will be replaced by a plus sign to show the word division. Citations which start without interpuncts or + sign belong therefore at the beginning of the line, and those which close without interpuncts or + sign belong at the end of a line where the interpuncts have been omitted. There is an evident attempt to avoid the division of words at the end of lines, for in the 256 lines written in the Umbrian alphabet, all but 20 end with the close of a word. Of the remaining 236 lines, 58 lack the interpuncts at the close, and 1 line, IV 31, has but asingle 28 LANGUAGE MONOGRAPH NO. 2, 1926 dot at the end. Those lacking the interpuncts may be divided into two groups, those where there is no space between the end of the line and the edge of the Table, and those where there is ample space for the interpuncts : The Line ends with the end of a word, but without interpuncts : Space is lacking Space is ample Lip eeE tik 2, 275128 10 Ib TT sa 02h. by 11h 07) Shsc as 27,44 IIb 3, 4, 28 ven Gy 1o,.2Laaorae 22, NA, 2745325. 335535 Wie 31, Gy. 10, 145 Le, 20s 8 20, Zit, ds, Oy Jos ates 24) wooly) 13, 15, 23,20, 29, 30/9200 8 Vila ¥ 251 20;329 Vb 6 Of the remaining 20 lines, where a word is divided by the end of the line, the division is always caused by lack of space; except in II b 7, where the interpuncts are wrongly placed at the end of the line, despite the fact that the word has not yet been completed. There are four other divisions attended by errors of some sort (I a 31 ;I b 20; Ila 33; Il bir); these will be dealt with later. Of the other 15 in- stances, 6 are divided between thesyllables of the words, 9 are divided in the interior ofthe syllables. The instances are as follows: Line Divisions not Coinciding with Word Ends Division inside Syllables Division between Syllables Division with Errors I a 26 : pesnim/u: la 28: ukri/per +- Ia 29 : feit/u + Ia 30 : staf/li: I a31 I bro : pernaia/f : Ib4 -+6e/fi: I b 20 Il ag : kuma/Itu -+- II a 23 :sutent/u : Il a 24: ahtrepura/tu : II a33 II b 10 : seva/kne: Il b 7 II b20 + pesni/mu: II b 27 + apel/us : Il bir Va27_ : atiier/iu: Vais :kumnah/kle: V bs : pepurkure/nt : The syllabic divisions in staf/li, atiier/iu, perhaps in sevak/ne, kumnah/kle, cannot be regarded as absolutely certain. As for the interpuncts in the interior of the lines, reference may be made to von Planta’s enumeration (2.667-68) of their omission, and to Conway’s critical annotations to the text. The first two Tables show about 13 percent of such omission, but Tables III and IV almost KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 29 never omit the interpuncts, and V never omits them. There are some instances, however, of wrongly ‘inserted or misplaced inter- puncts, and the occasional use of single interpuncts instead of double, as follows : Errors in Interpunctuation, Apart from Simple Omission. Ib 8 : anter: vakaze(= antervakaz se), cf. VI b47 .ander. uacose. I b 36 : ant. akre: (= antakre), withasingle dot wrongly inserted. Ib 45 kvestre: tie : (= kvestretie) Ilaro + perak : nem: (= perakném) Ila13 :ekvi:ne (= ekvine) Ila 16 anter:menzaru (= antermenzaru) I1a33 :tuve: rekapirus : (—tuvere kapirus) II b r seme:nies : (—= semenies) Il b 4 :etrep:eieriate: (= etre peieriate ) Ilb 7-8 :pera:/kne : (= perakne); the interpuncts are wrongly set at the end of the line, where the word does not end. Il b 8 :upetue : veietu: (= upetu eveietu) II bio : pera: kne : (— perakne) IIb 11-12: purtu:/etu : (= purtuvetu) ; the interpuncts are wrongly set at the end of a line, where the word does not end, and a letter is omitted. Il b 12 + fer:tu: (= fertu), cf. :fertu: earlier in the same line. Il b 14 :svi:se:ve: (= sviseve), cf. sviseve: at the beginning of the same line. II b17 + purtu : vetu : (= purtuvetu) II b 18 :atre:puratu (— atrepuratu) Il b 20 pesni:mu : (= pesnimu) ll b 25 :trisiuper-+ (= triiuper) IV 1 : es. kamitu: (= eskemitu), with a single dot wrongly inserted. IV 7: et. vempesuntres (— et vempesuntres), with asingle dot between the words instead of the double interpunct. V a 1 esuk. frater: (= esuk frater) with a single dot betwen the words. Va 3t:t. kastrucie :—TT Kastrucie), with a single dot after the abbreviation of the father’s praenomen. Vais: k. t. kluvier: (= K T Kluvier) with a single dot after each praenomen, In the matter of word division, enclisis alsoaffects the orthography. The words which are in question are the Umbrian equivalents of Latin quis, est, sis, sit. In the Tables written in the Umbrian alphabet, pis ‘quis’ is regularly enclitic to sve ‘si’ (I b 18, IV 26), and there- fore not divided from it by theinterpuncts ; est ‘est’ is always preceded by the interpuncts (I b 18 bis, II a 15); si ‘sit’ is set off by the inter- puncts or starts the line (V a6, 24, 27; Vb 3, 7) except in two somewhat doubtful instances (I b 8 :anter:vakaze-+ — anterva- 30 LANGUAGE MONOGRAPH NO. 2, 1926 kaz se;I b 8 +vacetumise+ = vacetumi se); the equivalent ot Latin sis does not occur. Besides these, there is a considerable equip- ment of enclitic prepositions, which need not be enumerated. Where the final of one word and the initial of the next agree, there is always an opportunity for fusion in speech, if the words are closely united in the phrase, or for haplography in writing. The engraver of these Tables took little advantage of either factor, for we find about 85 places where the final and the initial are identical in the Tables written in the native alphabet, out of about 1380 word-ends in the interior of lines. There are even five phrases in which the last two let- ters of one word agree with the first two letters of the succeeding word and yet there is nocrasis or haplography:I b 7 : inuk+ukar: Ib 20 :apretu:tures: II b 12: ife: fertu: IV 19 :super: erecle: V a 4: atiierier: ere: The only passages where the identity of finals and initials may have produced crasis or haplographyare in I b 8, 18, 19, 40, Il a 9, II b 22, which will be discussed later. There are also in the Umbrian alphabet certain variations in the orthography, which are not errors but indications of weakness in sounds or of similarity in sounds. Thus 1) e ei i vary with one another, within certain limits imposed by the etymological value of the sounds in question ; cf, Buck, p. 108. 2) h may have etymological value, or may be the sign of length of vowels ; it is therefore sometimes set where it has no warrant. 3) The sound } may be represented by the letter p as well as by theletter b, since in the lack ofthe letters g and d the letters k and t represent the sounds k g and ¢ d respectively. 4) f may become r before consonants, especially before s and v (Buck, p. 83), and may even disappear before s. 5) r before s,and n before mutes and spirants, may be omitted in writing, as they were weak sounds, 6) The final long -a was represented in some parts of the texts by a, in others by u (Iguv. Stud. 356-58), since it was an intermediate sound, and the Umbrian alphabet had no o. 7) Final-n normally, but not always, was written m, the weakness of final nasals and the frequency of final -m assisting the process (von Planta, 1.572-73 ; Buck, pp. 70-71). 8) Most final consonants were weak and often optionally omitted in writing : -m; -n ; -F (lguv. Stud. 360-61); -f; -t (with limitations, Jguv. Stud. 361-64); -k ; -s (except from -fs -ks). The facsimiles show blurs which indicate that the original engraver has correctéd his first version. These corrections are most frequent in Table Ia, and are found in smaller numbers in I b, II a, II b. In the later Tables they are very rare. In most instances they cannot serve as KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 31 a basis for, our studies ; but where they are clear enough for that pur- poses, they will be utilized. This revision of the engraved text is shown also by insertions, above the line, of omitted letters and words. In the consideration of special words and passages, everything will be given with epigraphic fidelity, as has been said ; letters which may optionally be omitted, if they have been omitted and it.is desired to indicate them, will be represented by letters above the line. The fol- lowing thenare the words and passages writted in the native Umbrian alphabet, in which error may with probability or with certainty be found : 1. la1o: are*arv*es stands for arepes arves, a phrase which, with slight orthographic variations, stands I a 6, 13, 19, 23, 27,1b 4, 7,26, 30, 33, 44, Il a 7. The two asterisks indicate characters which have been obliterated on the original Table. I suggest that the engraver first wrote ares--arvs for are(pe)s arv(e)s, the first omission being an Haplography with Skipping, the second a Pseudo-Haplography, for the letters v ande in the Umbrian alphabet differ in that the v lacks the middle horizontal or slanting stroke. The engraver then corrected the words to the best of his ability by deleting the twos’s, and adding es atthe end, which makes the last word correct, but he could not insert the wanting -pes inthe first word, since spacewas lacking. Why he did not place the letters below cannot be conjectured; for the next line is a blank separating two paragraphs. Conway however thinks that p stood originally in the first blur, and not s. 2. [ari : krapuvi stands atthe end of the line, and the space was insufficient ; the engraver has therefore placed the last two letters in the vacant space just above the end of the line, so that vi stands above pu. This can hardly be called an error, since there in no reason to be- lieve that the two letters were not engraved there before the next line was begun. 3. I ar2 :arvia-t is probably a correction for arviu (so Conway ; but von Planta and Buck read arviu), for the blur shows a correction of some kind, and the section la 1 tol b9 uses a normally, perhaps exclusively, and not w, to represent original final long a (lguv. Stud. 357 3 von Planta 1.79). 4. La 13 arpes:arves : has been placed at the end of the blank line between two paragraphs, and seems to be added when the engra- ver reviewed his work ; for if done at the first engraving, we should expect it to stand in a normal position at the beginning of the line. Further, arpes stands for arepes, the omission of the vowel not being oR aR 32 LANGUAGE MONOGRAPH NO. 2, 1926 true syncope, for in the 12 other instances of the word the vowel e or i is always written after r. 5. la 15 fise-+ was first written fisu, and then changed by the engraver. Perhaps the final of the immediately preceding word :fetu : was the cause of the error; it would then be Tele-Dittographic Change. On the Table, of course, this preceding word is at the end of the pre- ceding line, and thereforeunlikely to affect the engraver; but this sepa- ration into lines is not necessarily that of the draft from which the engraver was working. 6. 1a 16: arviu-+ is read by von Planta, Conway, and Buck, but the u is blurred, and may really stand for a u corrected to a; cf. No. 3, Lahr. 7. 1a 17: fiiuvi : is wrongly written for fisuvi, with s changed to iin imitation of the preceding letter ; Dittographic Change. There is a suspicious blur in the v which suggests that it was first written e, and that the middle stroke has been deleted ; e for v is an instance of Change by Addition. 8. la 18 kapir: stands for kapif; for kapi occurs twice and capif three times. The r is written here because of association with other forms of the paradigm, such a kapire in the dative and ablative singu- lar and kapirus in the ablative plural, but is still to be regarded as an error of Change. g. I a 23 arviu : isthesame case as No. 6, 1a 16. 10. la 30: erel : stands for erek, another error of Change ; the 1 and the k have the vertical hasta in common. 11. La 30/31 :staf/li:siuvesmik: is a disputed passage. Von Planta suggests that it is wrongly written for staflare esmik (2.558), while Conway preters staflii uv° esmik, and would like to read staflarim; both follow essentially Bucheler (74-75), who compared staflare in the paral- lel account in VI b 37. Buck follows Conway’s text in reading staflii uve esmik. I have proposed (Jguv. Stud. 354-56) that the correct reading is stafli : iuv(ie:) esmik, with Pseudo-Haplology caused by the almost complete identity of v and e in the Umbrian alphabet ; my warrant is in the parallelism of Ia 30-32 with I a 27-30, where the iuvie of 31 corresponds in function to the : tefri : iuvi: in 28. 12. | a 31/32 : ikuvinp/a: is for ikuvina. The p at the end of 21 is only a partly written a for which the engraver concluded that there was insufficient space ; he left it at the end of the line and cut a at the beginning of 32. This is therefore a Partial Dittography, by division between lines. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 33 13. La 33 +purtitius: standsfor purtiius, which occurs four times. Thisis an instance of Tele-Dittography. 14. la34 : kumats: is for kumates, which is foundII a 42, [Va 29, and (without the final s) I b 37,38, Ila 10. This is a mere Omission; if there had been any phonetic significance in the loss of the e, the ts would have been written z in the Umbrian alphabet. 15. I b 3 :kutep: stands for kutef. This may be a Dittographic Change, as the next word, pesnimu, begins with p ; or the explana- tion may be asin No. 16. 16. I b 4 vitlup: turup: is for vitluf turuf. Unless this isan outright change for which we cannot find the motivating cause, we must assume that the p is a variant form of another form of f, either like early Latin f with slanting strokes, or like the Faliscan f, which consisted of -a vertical hasta from the top of which, on both sides, a short stroke extended diagonally downwards; cf. Bréal 227, 313; von Planta 1.465 n.; Buck § 25 a. With this theory, p for f is an example of Change by Subtraction. 17. 1b 8 :anter: vakaze--is for antervakaz se ‘intervacatio sit’. As zZ represents ts, Zhere stands for ts + s, an instance of Haplography. It is hard to exclude here a phonetic element, since the verb may easily have been enclitic, which will excuse the lack of interpuncts; but such graphic simplification between words is extremely rare in the Tables, although there are many opportunities for it, and we may therefore regard “this as a phenomenon of the orthography rather than as one of the spoken language. 18. I b 10 :anzvriatu: is for anzeriatu; the failure to engrave the middle oblique stroke of e leaves the character a v. This is a Change by Subtraction. 19. Ib 15 :menes: stands where we expect benes. There is an m just above, which may have caused Tele-Dittographic Change. Or the form menes may be by Decomposition from compounds in which, as in *kum-benes, the bstood after mand was assimilated to it ; cf. von Planta 1.433-34, Buck § 125.2<. 20. 1b 18 : purtatulu: stands for purtatu ulu. This might be an instance of crasis, for the two words are closely connected in the phrase; but in view of :feitu:uru: in the same line, and of .portatu.ulo. at VI b 55, where there is an equal opportunity for crasis but the words are written in full, I regard the present example as one of Haplography, with Skipping of the interpuncts. 21. I br9 :staheren: termnesku: for staherent termnesku. While 34 LANGUAGE MONOGRAPH NO. 2, 1926 the initial of the second word may serve as an excuse for the omission of the final t of the preceding, it is not a normal example of sandhi, for the interpuncts remain, preventing the coalescing of the two words. It is therefore an example of Tele-Haplography, in which one charac- ter is lost because of identity with another, butthe intervening symbol or symbols remain; here, the interpuncts which separated the two t’s, are kept. 22. 1b 19 : armanu: tor armanu, cf. arsmahamo VIb 56. Ther for risan error of Change, in the size of the loop attached to the ver- tical hasta; it can hardly be calleda phonetic change (von Planta 1.294- 95 ; Buck § 132 a). Then form is an error of Change by Subtraction, since nin the Umbrian alphabet consists of the first three of the five strokes which compose the m. 23. I b 20/21 :amprefu/us: stands ror amprefus, with Dittography of the u induced by the division of the word at the end of the line. The engraver obviously tried to put the whole word into the space at the end of line 20 by crowding the letters, but could not quite do it. 24. Ib 25 :ferime: for ferine, as is shown by the recurrence of the phrase in which it is found, 12 times in other parts of the Tables. This is an errorof Change by Addition, the reverse of the n for min armanu I b 19. 25. Ib 25 :feiu: for fetu, wherein the t lacks the cross-stroke ; a Change by Subtraction. But as fetu has a variant spelling feitu in the Umbrian alphabet (45 and 20 times respectively), feiu may be for feitu with graphic omission of the t; this is however less probable. 26. Ib 27 rupinié:e: has the preposition repeated after the locative, which looks like a dittography, but is not so to be regarded, for we find identical or closely similar uses in IIb 12 : tafle : e+ ‘in tabula’, II b 27 :testre:e-+-uze: ‘in dextro umero’, II b 28 :testre: e+-uze+. 27. I b 4o pustertiu: equals pust tertiu, but the lose of t here seems to be phonetic when the preposition is proclitic (von Planta 1.576). The examples are postertio. VII a 46, pusveres: Ia 7, 14, 24; but the separated post is found in post. uerir. VI a 58, VI b 3, 22, and in .post. uerir. VII a 38. Asimilar variation between enclisis and inde- pendence is seen in the preposition pre, with : preveres : ‘ante portam’ Ia 2; preveres: larr, 20 ; pre.uereir. V1 a22; pre. uerir. VI a 59, VI b 4, Vil a7; preuerir. Vib 1,19; + pre. uerir. VIb 235 .pre uerir (with space between the words, but no interpunct) VI b 20. 28. I b 4o :tuseiu : for tusetu, where the cross-stroke of the t is lacking ; cf. feu I b 25. This is a Change by Subtraction. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 35 29. I b 43 :afviu: for arviu ; the loop of the r is too small, which changes the character to r. This isa Change. 30. Ib 45 :usaie: stands for usace, cf. +usace+ II a 44. The engraver has failed to attach to the vertical hasta the loop which makes the ¢; a Change by Subtraction. 31. I] a2: esum stands for esunu. The slight crowding at the end of the line has forced the u over against then, and a slight alteration in the slant of the strokes has resulted in a perfect m. This might be termed a Change produced by Crowding. 32. Ila4 -+eretu: stands for heretu ; in this common word the initial } is always written except in this instance. The error is one of Omission of a Letter representing a Weak Sound, cf. von Planta 1.445. 33. Ila 4 :fefure: has often been considered a third person plural with omitted-nt, but if it is, it is the only example of such omission of -nt, without some accessory factor (Iguv. Stud. 363). It is preferable to take it as a third personal singular ending in secondary -d, which is — lost in Umbrian. There is therefore no error in this writing. 34. Ila 9 :purtiiusuru: is for purtiius suru; Haplography with Skipping of the interpuncts. 35. II aro :iuvip: seems to be an abbreviation of the dative of Iupater, but that is found only as iuvepatre (five times ; also : iuve : patre : IIb 7), and thereis no certain dative of a consonant stem ending in -1 (von Planta 1.148). It is probably therefore an error for iuvie, dative of the stem iowio-, found many times. The p is fairly close in form to an e which lacks all but the topmost cross stroke ; a Change by Subtraction. 36. Ia 28 :prusektu: stands for prusekatu, which is found earlier in the same line, as well as in III 33, 35, IV 2. This isan Omission, without assignable cause. 37. Ia 30 :eenpersuntra: is for venpersuntra with e for v, by the addition of the middle cross stroke ;a Change by Addition. 38. Il a 33/34 :a/anfehtaf: is for anfehtaf. The engraver started the word on line 33, then saw that the space would be insufficient, and leaving unoccupied the space of about four letters, began again on the next line. Thus results a Dittography from the line division, as at Ib 20/21. 39. Il a 35 :petruniapert: has an extra t at the end ofthe postpo- sition, cf. petruniaper II a 21. Here the t may be induced by the petr- at the beginning of the word, with the same letters in a different order, or by the similar pert ‘trans’ which occurs in the next line, and may 36 LANGUAGE MONOGRAPH NO. 2, 1926 have caught theengraver’s eye in the original copy. In either instance, this is an example of Tele-Dittography. 40. Ia 42 :persmhniu: for persnihmu. Apparently the engraver omitted thesecond oblique stroke of the m, which produced persnihniu; he then tried to insert the stroke, but misplaced it in the wrong -ni-. This might be Change by Metathesis of a Stroke in Correction. 4t. Il b 9 :facefete: is for fagefele, with 1 changed to t by mispla- cement of the oblique stroke, and the extension of it across the verti- cal hasta. 42. II b 10 :eaputu: was originally written, but the engraver has tried to obliterate'the middle stroke of the e and correct it to v, giving vaputu ; Change by Addition, corrected in the original. 43. Il br1/12 :purtu:/etu: stands for purtuvetu, which is found a number of times in variant spellings. Besides the wrong division of the word by the interpuncts, there is an omission of the v, by Pseudo- Haplography, since its neighbor e is so closely similar in form. 44. II b 19 :ranu: stands for ranu. This Change was caused by the engraver’s failure to bring the loop down far enough to the bottom of the hasta; he seems to have tried to correct it later, leaving a blur. 45. Ib 22 :eruhu:ticlu: stands for eruhunt tiglu. This is a Tele- Haplography, to be explained precisely as in :staheren:termnesku: I b 19 (No. 21), with a further graphic omission of n. 46. I b 26 feiu: stands for fetu, asin Ib 25 ;a Changeby Subtraction. 47. Ul 16 :ferime: is for ferine, as at I b 25. a Change by Addi- tion. Buck, p. 335, accepts this writing as an error, but von Planta, 2.736, and Conway, 2.621, take ferime hereas a different word, an accu- sative with postposition -e(n). It is true that the phrase here, :kazi: ferime:antentu:, is different from that in which ferine elsewhere occurs, which is always vatuva ferine fetu, with or without permis- sible variants in the spellings. In its other occurrences, antentu and its variants atentu and andendu have a direct object expressed or implied, and a form which is probably a dative, less probably a locative (so von Planta in his glossary), or else an adverb ; only in VII. a 25 do we find a prepositional phrase, .superne.adro.trahuorfi.andendu. It is more likely then that the formin III 16 conforms with the idiom found in the other passages in its neighborhood (II a 19-20, II b 28, 1114-15, 16-17, 17, 21-22, IV 21, 27), than that it is an accusative with the postposition, only partly paralleled VII a25. The writing ferime is therefore in this passage also to be regarded as an error ; for ferine is the normal spel- ling both for the dative and for the ablative. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 37 48. III 31 vatra: is an error for vatuva, for the word occurs in a formula, vatuva ferine feitu, found (with permissible orthographic variants) Ia 4, 13, 22,1 b 3, 6-7, 25, Via 57, VIb1, 19, 43-44, 45, VII a 4, as well as here. The change of two letters to one of an en- tirely different shape is one of Change ; it may be styled Change with Reduction of the Number of Letters. 49. Ill 31 :aruvia: stands for arvia, for this word and its conge- ners occur 35 times inthe native alphabet, always with -rv- except in this one passage. The only reasonable explanation is that it isan attempt at phonetic analysis of the sounds, with a slow pronunciation of the word; an influence of the Latin alphabet seems hardly likely. It is therefore an instance of Addition, with an attempt at phonetic accu- racy. 50. IV x :aveitu is for arveitu (occurring five times ; arveitul b 6; arsueitu eleven times; arueitu VI b 23), with Omission induced by crowding at the end of the line, where the space was insufficient for the word. Von Planta’s suggestion, 1.415,seems unlikely. 51. IV 5 :erererunt: is for ererunt, with Dittography. 52. IV 6: ereclamar is for ereclumar, where the erroneous -a- seems tobe caused by the -a- of the next syllable ; a Tele-Dittographic Change. The true value of the vowelis seen in the accusative ereclu, which is found twice, and in erecluma, found three times, the same with the postposition -a(r). 53. IV 7 :iseceles: is for isecetes; a ‘ee with misplacement of the cross-stroke. 54. IV 9 :sevakne stands for sevaknes, at the end of the line, where there is no room for the additional letter; Omission because of Crowding. 55. IV 12 :puprikes: is for puprike, with a Tele-Dittography, induced by the apparent parallelism of the phrases puemune: puprikes: et : vesune : puemunes although in reality this puprikes is an adjective modifiyng puemune, just as puprikes atthe beginning of the next line modifies puemunes at the end of 12. 56. IV 14 :erus: There is some mistake here; a verb tertu or tetu ‘dato’ has been omitted, or erus should be eru, a genitive plural of the pronoun (cf. Jguv. Stud. 368). 57. IV 14 : purtupite is for purtuvite ; the lower cross-stroke of the v is lacking. A Change by Subtraction. 58. IV 17 :vesveca: is for vesteca, by the influence of the initial ve-; a Tele-Dittographic Change. 38 LANGUAGE MONOGRAPH NO. 2, 1926 59. IV 18 :inuntek is for inumek ; the second diagonal of the m is set on the wrong side of the hasta. An Outright Change. 60. IV 23 inumk: is for inumek, an instance of Pseudo-Haplo- graphy, since the e and the k alike begin with a vertical hasta. 61. IV 25 :persihmu: for persnihmu, shows Omission of the n, hardly-a Pseudo-Haplography, for in this passage the first and third strokes of then are not vertical, like the single hasta ofthe i, but are somewhat slanting. 62. IV 26 :puprces: for pupricges ; the word occurs in eight other passages, always with -rig- or-rik-, never with omission of the vowel. This seems to bea Pseudo-Haplography, for the letters on each side of the i have as main components a vertical hasta. 63. IV 28 tertu: is the only occurrence of this spelling ; we find tertu twice, dirstu five times. It seems likely then that tertu is an error, with too large a loop on the left of the vertical ; cf. von Planta 1.408, Buck § 1324. 64. IV 33 neithabas i is for nei arhabas or even for nei arhabias. Is this spoken crasis ? Hardly, for -ei +-a- would give -eya- -ea-, anda contraction to a long vowel, not to a diphthong. Is this an example of syncope in the word group, because the verb is enclitic on the preced- ing negative? This is without parallel in Italic, at least. But this Table contains, for its length, an unusual number of errors, and I believe that we have here errors of Omission, possibly of pseudo-haplographic nature because of the numerous vertical strokes. Cf. Buck § 84, §218; von Planta 2.241-42, with note. 65. V a22 :furenr: stands forfurent. The engraver had insufficient roomat the end ofthe line, and turned his letters downward; as he made the oblique line of the t, histool slipped, and cut around to the top of the hasta. This is therefore a Change by Accidental Slipping of the Engraver’sTool. THE TABLES IN THE LATIN ALPHABET. When we turn to those parts of the Tables which are engraved in the Latin alphabet, we meet with some new kinds of error. For these texts are evidently transliterated from an older version in the native Umbrian alphabet ; andas the twoalphabets do not correspond letter for letter, the process of transcription may result in mistakes. Some letters, it is true, correspond precisely in the twoalphabets ; but there are the following variations : 1) The r of the Umbrian alphabet becomes rs in the Latin alphabet, which is often written merely s before a consonant. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 39 2) The ¢ ofthe Umbrian is represented by Latin § (s with a grave accent above and slightly to the left) ; the diacritical mark is often omitted. 3) The Umbrian consonantal v becomes the Latin « in writing, as the Latin alphabet does not distinguish the vowel uw and the consonantal v. 4) While the Umbrian alphabet indicates both the consonantal v and the conso- nantal i before vowels, as glides after vowel « and vowel i respectively, the Latin alphabet represents neither glide in writing (except v in wvi before another vowel). 5) The Umbrian u represented several sounds, as the native alphabet lacked the letter o. Therefore an Umbrian u might become either oor w in the Latin alphabet. 6) The final open 6, from earlier -d, which is written a or u in the Umbrian alphabet, is always o in the Latin alphabet. 7) While long vowels were indicated in the Umbrian alphabet only by an added h (type, ah), they were indicated in the Latin alphabet by two other methods as well : the doubled vowel (type aa), used sporadically in Latin from about 130 B.C., and a mixture of the preceding types, the doubled vowel separated by an / (type, aha). _ 8) In the Umbrian alphabet, there was no g or d; k and t represented there fore two sounds each, and despite the presence of a bthe p was by analogy sometimes used for ) as well. The Umbrian k, t, p became accordingly in the Latin alphabet respectively c (q before w) or g, tor d, por b. 9) The Umbrian z, sounded ¢s, became Latin s ; the same Latin letter of course represented the Umbrian s. 10) The long consonants are occasionally represented in the Latin alphabet by the gemination of the letter, a practice which in the writing of Latin texts began soon after 200 B.C. ; this gemination is unknown in the Umbrian alphabet. 11) Other graphic variations, listed in the account of the Umbrian alphabet and not inconsistent with the transcription, are still found as in the earlier texts (Nos. 1, 2, 5, 7,8 as listed on p. 30). In the Tables written in the Latin alphabet, no word is divided at the end of a line. The word-division is indicated by a single dot, placed at half the height of the letter ; it is never used at the beginning or at the end of a line. Though its presence or absence in the interior of a line is sometimes difficult to determine, it seems to be lacking in about four percent of the possible positions — oftener in Table VI than elsewhere. There are the following wrong uses of the interpunct : Errors in Interpunctuation, Apart from Simple Omission : VI a 40 pequo + c.astruo. (= pequo castruo) Mi ad ye uk: er ver C=='erer) Vi a 48 .peracri+-p.ihaclu. (= peracri pihaclu) VI a 54 .no.mneper. (== nomneper) VI b 5. . t.estisiat+et. (= uestisia et) VI b 8 era.rnomne. (= erar nomne) VI b II .wow,se+-auie. (= uouse auie) VI b 13. fist.er. (= fisier) VI b 29 iowie+-orer. (= toute orer) ; the interpunct was later inserted by the engraver, above the o of orer. 40 LANGUAGE MONOGRAPH NO. 2, 1926 VI b 37. staflar .e. (= staflare) VI b 47. .ander .uacose. (= anderuacos se) VI b 49 .destra.mescapla, (= destrame scapla) VI b 51 -.ponisia.ter. (= ponisiater) VI b 59 .tarsinat.er. (= tarsinater) Vil a7 .t.ua. (= tua); the dot was first wrongly placed, and was then correctly inserted also, before the /. VIL a 43. «.dirsu.st. (= dirsust) VII a 49. preplo. hotatu. (= preplohotatu) In three places, free space is found between letters, in the middle of the word ; in the first two, according to Conway, a letter has been expunged : VI b3 filiu. VIb4 .destre. VI b 20 .pre uerir Once there is free space between two interpuncts : VII air .serfia. .serfer. In a considerable number of places there is free space between the interpunct and the next letter, usually but not always where a new sentence begins, or there is an equally important division. It seems hardly necessary or desirable to list these passages. Enclisis causes the omission of the interpunct before the Umbrian equivalents of Latin quis in siguis (once ; only instance), of est (20 times enclitic ; 17 times separated), sis (once enclitic ; 5 times separated), sit (4 times enclitic). There isample opportunity for crasis or for haplography, since over four percent of the final letters in the interior of lines are identical with the initials which follow ; there is even a triple identity in VI b 59 ¢otar.tarsinat.er. and in VII a 12 .totar.tarsinater. But there is little advantage taken of this ; apart from instances where the initial of (the Umbrian equivalent of Latin) sis and sit combines graphically with a final s of the preceding word, there are only two possible instances, VI b 54 .ehesu. (= ehe esu), VIL a 46 postertio. (== post tertio). Besides this, we find that the original engraver went over his work with considerable care after he had completed i it. Some of his corrections have been mentioned already ; we shall find, as we go on, a deletion in V b rt, insertions in V b 12, 15, VI b 15, 30, 61, VIL a 14, and changes in VI a 35, VIb 38, VII a 37. We may now pass to the consideration of the certain or probable errors in the Tables written in the Latin alphabet. The items will be numbered serially, from the point reached in the errors made in the Tables written in the native alphabet. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 4I 66. Vb it .fratxer. stands for frater. According to yon Planta 2.59 n., an expunged r is still faintly visible. The other plural cases of the word seem to have influenced the engraver, for they had r immediately after the ¢. One may term this an Addition by Para- digmatic Analogy, a defect linguistic rather than orthographic. But as this instance is an individual misspelling, we may keep the category in our list. As the engraver did not place thee in the space left by the deletion of the r, he cannot have noted his error until the word was completed ; this shows that the error is not one of mere omission of the e : fratr for frater. 67. V b 12 .pretra The second r was originally omitted, and after- ward inserted above the line, between the ¢ and the a. 68. V b 15 martier. The first r was originally omitted, and later inserted above, between the a and the t. 69. VI a 3 .auuei. stands for auei, with Dittography of the uw. The word and its derivatives occur 14 times in the Latin alphabet (7 times also in the Umbrian script), but no other instance gives any authority for this doubling of the u (auu- would equal auv-). 70. VI a 6 .courtust. stands for couortust, cf. kuvurtus I b 11, co- uortus VIL a 39, couortuso VI b 64. This omission of the o may be by Tele-Haplography. Or it may be due to the confusion in the ortho- graphy where -ov- -uv--u- are involved, as in Latin; cf. Trans. Amer. Philol. Assn. 43. 41-42. 71. VI a 7 .andersesusp. is for andersesust, witht changed to p by a slight alteration in the cross stroke. 72. Vl a 14 .hoier. may be for .holer., cf. :hule IV 17, though von Planta 1.413 thinks itunlikely. Ifan error exists here, the / has become i by failure to engrave the horizontal line; a Change by Subtraction. 73. Vl a 16 +-anclar. stands for anglar, since spellings with g are found five times. Here the engraver has wrongly transliterated his Umbrian k, which had the value of both ¢ and g. 74. VI a 17 .meersta isthe sole example in which the vowel before an r plusconsonant is written double. Even when ther is not written, there is no tendency to double the vowel. This word occurs 14 times with one e, and in other derivatives 6 times in the Latin script and 4 times in the Umbrian. For these reasons, despite Buck § 76. 1, I am inclined to regard the doubling of e in this word as a Dittography. 75. Vl a 18 .ancla. stands for angla ; cf. No. 73. 76. VI a 22 .iuue. should be ioue ; itis a too faithful transcript of 42 LANGUAGE MONOGRAPH NO. 2, 1926 Umbrian iuve, violating the Latin practice of not writing -wu-, and failing to give the proper representation to the 0, which of necessity was written u in the Umbrian script. 77. Vl a 27 .crabouie. should be grabouie, the g being found in 35 other instances ; another wrong transliteration of Umbrian k, cf. Nos. 73 and 75. 78. Via 31 .erlr. (or .erir. with a slight line at the bottom of the i, according to Conway) stands for erer ; e becomes / when the top and the middle cross-strokes are wanting. Change by Subtraction. 79. VI a 35 .pihaclo. was written first, then changed to pihaclu at once. That the correction was made at once, is evident from the posi- tion of the interpunct, which is on the curve of the 0; for o is a wider Jetter than u, in the Latin alphabet, which after the correction to allowed the dot to be superposed on the line of the o. This error in the original writting was one of Wrong Transliteration, since Umbrian u represented also thesound of 0, and the engraver has at first chosen the wrong equivalent. 80. VI a 37 .crabouie. for grabouieis a Wrong Transliteration, iden- tical with No. 77. 81. VI a 41 .saluuom. is for saluom. The Umbrian alphabet wrote the glide after the antevocalic u, but the Latin alphabet did not. In this example the engraver has made an error of Addition by Literatim Transliteration. The word occurs with one u, 24 times. 82. VI a 42 .saluua. is for salua ; the same error as in No. 81. 83. VI a 42 .tuua. is for tua, which with the single wu is found 13 times; the same species of error as Nos. 81 and 82. 84. VI a 54 .tribrisine. should be .tribrsisine, on the evidence ot tribrigu Va 9; unless the s of the rs (= r) is lost by dissimilation against the following s. Another possibility is that the engraver rebelled against so monstrous a combination as -brs- and omitted the s by the influence of Latin consonant groups. Or even it may be an instance of Tele-Haplography, only one of the two s’s being written, though they are separated by a letter. 85- VI a 54 .peracnio. is for peracrio, a Change with out assignable reason, except that both words existed and were used freely. Here the wrong word has been engraved. 86. Vl a 58 .trebo. was corrected by Aufrecht-Kirchhoff to trebe, because of trebe I a8; but trebo is kept by Biicheler as dative of a u- stem. Later editors have followed Biicheler in recognizing change of stem-class. If there be an error here, it is one of Change without assignable reason. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 43 87. VI b 3 ++/iso. may be for fise, but is to be judged like the pre- ceding. | 88. VI b 3 .ocrifer. for ocriper, has f for p, by a slight change in the strokes. | 89. VI b § .confgos. for conegos, has f for e, a Change by Subtraction of the lowest horizontal bar. 90. VI b 15 +-nomne.erar. has per inserted above, to make nommeper erar ; an Omission, corrected on the Table itself. gt. VI b 15 -erite. is for frite, with e for f, a Change by Addition of the lowest horizontal stroke. 92. VI b 25 .perstico. is for persico, a simple Addition, for which no cause can be assigned. 93. VI b 30 .touer. occurs twice in this line, perhaps by wrong transliteration of Umbrian tuver, assisted by the Latin repugnance to a repeated vowel sign. The spelling twer occurs five times ; cf. also No. 83. 94. VI b 30 .uasetomesf-+- is for uasetom est, with t changed to f by addition of the medial stroke, and shortening of the top one. 95. VI b 30 .frosetomes. was corrected to frosetom est by the inser- tion of ¢ above the line. The error made in the first engraving was one of Omission, without assignable cause. 96. VI b 32 .pfquo. is for pequo, with f for e, a Change by Sub- traction of the lowest horizontal stroke. 97. VI b 35 .tioute. is for iouie ; for the stem is found 17 times with io- and 8 or 9 times with iu-. It is of course a graphic assimi- lation to the stem iiouino- ; cf. Conway 1.405 n. The alteration is in this word still to be accounted an error, since it did not become rea- sonably customary. Addition by Graphic Assimilation to another word. 98. VI b 36 .atropusatu stands for atripursatu, with o for i in the second syllable ; Change, without reasonable explanation (despite Buck § 86. 7). The proper spelling is proved by ahtrepuratu II a 24-25, 25, 31, 38, atrepuratu II b 18, atripursatu VI b 16, ahatripursatu VII a 23, 36. 99. VI b 38 .nestisiar+ was first engraved, and then the error was corrected by deleting the first vertical of the n, so as to leave v ; the word is vestisiar. The original error was one of Change by Addition. 100. VI b 45 .itouinar. for iiouina, has an added r by the influence of neighboring words, for the four words preceding all end inr. This is an instance of Tele-Dittography. 44 — LANGUAGE MONOGRAPH NO. 2, 1926 ror. VI b 45 .watue. is for watuo, an example of Change, possibly tele-dittographic in nature, since the group -ue- occurs almost directly above in the preceding line, .werisco. VI b 44. 102. VI b 46 .ficlmrsueitu. stands for ficlam arsueitu. The m repre- sents therefore AA, without the cross-bars and the interpunct, so that the four strokes make an M; a Change by Subtraction and Ligature. The final m is so rarely written in Umbrian of the Latin alphabet, that the alternative explanation is unlikely : that AMA was reduced to M by the omission of both A’s, on acount of the resemblance of each to one half of the M; which would have been Pseudo-Haplography. 103. VI b 47 .ander.uacose. seems, like :anter:vakaze-+ I b 8 (No. 17), to stand for anderuacos se ‘intervacatio sit’. For the same reasons asin the other occurrence, this may be regarded as an instance of Haplography. 104. VI b 48 .sururo. stands for sururor, by Dittography for suror ; cf. Iguv. Stud. 361. 105. VI b 50 esonomf. is for esonome, with the same error as in No. 96; a Change by Subtraction. 106. VI b 50 .ffrar. is for ferar, with the same error as in No. 105. 107. VI b 50 .entelust. is for endelust, by slavish transliteration of the Umbrian alphabet. I cannot agree with von Planta t.553, that the t remains after 2 by influence of other compounds such as ostendu, for we have the forms endendu andendu actually written. 108. VI b 53 hebetafe. is for ebetrafe, which occurs VI a 12; there are two errors, one of Addition of a Letter representing a Weak Sound, the other of Omission. 109. VI b 54 -ehesu. stands for ebe esu, which also is written in the same line. It is unlikely that this is true spoken Crasis, as has been said above ; it is probably Haplography. 110. VI b 55 fsme. is for esme, a Change by Subtraction; cf. Nos. 96, 105, 106. 111. VI b 60 .preuilatu is the same word which is written preuislatu at VII a 49 ; there is therefore an error of Omission, as there is no evidence that s (= ¢) was lost before /. 112. VI b 61 .prestota.serfer. The word serfia, between these two words was lost by Haplography (with Skipping), but was inserted after- ward by the engraver, above the line, when he compared his work with his original copy. 113. VIl a 3 .seree. for serfe, anerror of Change by Addition of the lowest cross-bar of the e. The engraver seems to have tried afterward to obliterate the extra stroke. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 4S 114. Vila 14 .erom.nomne.erar.nerus. The words nomne.erar were omitted before nerus, by Haplography, and were inserted later by the engraver, above the line. 115. Vil a 22 .serfiar+ is for serfia, an example of Tele-Ditto- graphy, since most of the neighboring words end in r. 116. VII a 37 .meca. was engraved for mefa, and then corrected on the Table; an error of Change by Addition. 117. VIL a 43 .abrons. is for abronf, a Change of a single letter. 118. VII a 46 postertio. Cf. No. 27. 119. VII b 1 + fratrexs. Cf. fratreks V a 23, V b 1, in the Umbrian alphabet. The xs may be due to a mixture of alphabets, for the Umbrian used ks and the Latin used x. Yet as xs was a not too infrequent writing in Latin inscriptions, it is perhaps better to regard the xs in fratrexs as an over-zealous attempt to show the double nature of the sound, even asin those Latin inscriptions. In these 119 errors, certain or possible, we have found the following varieties ; examples of uncertain interpretation are in italics : I. Errors of Omission : Omission, unmotivated : 4, 14, 36, 56, 61; 67, 68, 90, 95, 108, 111. Haplography : 114. Haplography, with Skipping : 1 ; 112. Haplography, with Skipping of the interpunct only : 17, 20, 34; 103, 109. Tele-Haplography : 21, 45 ; 70, 84. Pseudo-Haplography : 1, 43, 60, 62, 643 102. Pseudo-Haplography, with Skipping : 11. Omission caused by crowding at the end of the line: 50, 54. Omission of a letter representing a weak sound: 32. Omission by the influence of Latin orthography : 70, 84. - II. Errors of Addition : Addition, unmotivated : 56 ; 92. Dittography : 51 ; 69, 74, 104. Dittography induced by division between lines : 23, 38. Partial Dittography, by division between lines : 12. Tele-Dittography : 13, 39, 55 ; 100, II5. Addition by the influence of familiar words : 97. Addition by association with other forms of the paradigm : 66. Addition for phonetic accuracy : 49. Addition of a letter representing a weak sound : 108. Addition by wrong transliteration into the Latin alphabet : 81, 82, 83. Addition by the influence of Latin orthography : 119. III. Errors of Change : Change, unmotivated : 10, 16, 22, 29, 41, 44, 48, 53, 59, 63; 71, 86, 87, 88, 94, 98, 117. 46 LANGUAGE MONOGRAPH NO. 2, 1926 Change by Addition : 24, 37, 42, 473 91, 99, 113, 116. Dittographic Change : 7, Ij. Tele-Dittographic Change : 5, 19, 52, 58; 101. Change by Subtraction : rs, 16, 18, 22, 25, 28, 30, 35, 40, 46, 573 72, 78, 89, 96, 105, 106, II0. Change by Subtraction and Ligature : 102. Change by Crowding and Ligature : 31. Change by slipping of the engraver’s tool : 65. Change by metathesis of a stroke in correction: 40. Change by the influence of familiar words : 85. Change by association with other forms of the paradigm : 8. Change by preserving older orthographic practices : 3, 6, 9. Change by wrong transliteration into the Latin alphabet : 73, 75, 76, 77, 79, 80, 93, 107. IV. Examples listed, but certainly or possibly not errors : 2,19, 26, 27, 33, 72, 74, 84, 86, 87, 118. APPENDIX As is noted by von Planta 1.445, with a full list of the words involved, the } is weak in Umbrian when it is initial in the word or stands at the beginning of the second element of a compound ; so that it may fail to be written where it belongs, or may be written where it has no etymological justification. It has been necessary to divide such instances into those of linguistic nature, and those of orthographic nature. Accordingly, I have not listed the following instances as errors of orthography : 1) Where as the second element of a compound the / stood imme- diately after a consonant ; for the forms without h may be phonetic, and those with ) may be analogical : a) anostatu etc. 4 times, anhostatu etc. 3 times. b) in the enclitic -hont, which is always -ont after a conso- nant : eur-ont erir-ont if-ont surur-ont is-unt etc., 30 times in all. c) in eitipes V a 2, 14, if from *eitom-hipens. 2) Where there is a strong possibility that the word has regularly gained or lost an / in all its forms, by the analogy of another word : a) in habina and its torms, 6 occurrences, all with h. b) in holtu VI b 60, VIL a 49; with which may belong hule IV 17, and perhaps hoier V1 a 14 (cf. No. 72). The only actual instances of wrong orthography are eretu Ila 4 (No. 32) without the initial b, and hebetafe VI b 53 (No. 108) with a prefixed h. CHAPTER VI LATIN : THE LEX JULIA MUNICIPALIS The Lex Julia Municipalis is a document of the year 45 B.C., on a bronze tablet which has been broken into two pieces of almost equal size. One piece was found in 1732, near Heraclea in South Italy, and has been since 1754 in the Naples Museum; the other, found 1735 at the same place, is also in the Naples Museum, after having been for some years in private possession in England. The other side of the ‘Tablet bears a Greek inscription, CIG 3.5774-75. The Latin inscription which we are to examine is published in the Corpus Inscriptionum Latinarum I‘, No. 206, pp. 119-125 (ed. Th. Momm- sen, Berlin, 1863) =I. 27, No. 593, pp. 482-487 (ed. E. Lommatsch, Berlin, 1918); it is given in lithographic fac-simile by F. Ritschl, Priscae Latinitatis Monumenta Epigraphica, plates xxxiu and xxxtv (Ber- lin, 1862 ; a supplement to the CIL). The letters are in a good style, virtually identical with our capital letters, but without the ornamental serifs; they are slightly irregular in shape, as is common when letters are engraved on bronze. The words are separated by a single interpunct, which is rarely if ever omitted. The engraver was both careless and ignorant, as was pointed out by H.A. Seidel, Observationum Epigraphicarum Capita Duo, diss. Vratislav. 1880, who examines in detail certain points. He changed some letters, and added others, he omitted parts of letters, entire letters, and even words and phrases; he had a very imperfect knowledge of Latin, and engraved a plural verb under the influence of near-by plural nouns, while theactual subject was in the singular but at a distance, or (more often) changed the case of nouns, apparently by the influence of their neighbors, in some instances, but more often without any observable reason. The text, it is true, is highly abbreviated, many words being reduced to their initials ; and some series of such initials can be expanded only by comparison with other documents, where the formulae recur in full. Now whereas these abbreviations were present in the copy from which the engraver worked, we cannot identify some errors as those of the original copy and others as those of the engraver ; for we +7 SRR AAS 48 LANGUAGE MONOGRAPH NO. 2, 1926 do not know whether the engraver was expected to copy his model slavishly, or to expand some of the abbreviations, nor do we know whether the model was written in a different script, of cursive or partly cursive character, which the engraver was to change into the formal lettering, or the model had the same script as that of the in- scription which he cut. If we could determine these points, it would be very illuminating. Particularly it would help us to know whether the model wrote II and I' for E and F, for then the repeated confusion of E and F and I would be most easily explained. In the absence of any direct evidence fora different view, we must assume that the engraver attempted to copy his model text without alteration : without change of the forms of the letters, without expan- sion of the abbreviations. From the linguistic standpoint, we should note that EI is very fre- quent for long I, and in fact almost regular in endings; that the doub- led consonant is sometimes written single; that M is used freely for the reduced nasal before T and D. We may now proceed to the exam- ination of the errors and other similar peculiarities. 1. EAFDEM 2 = cadem ; there is a faulty insertion of F, written without the lowest cross-stroke. 2. OMNIA IN IISDEM 5 has a wrongly added IN. 3. EI 5 =et;a Change by Subtraction, the lack of the horizontal of T changing it into I. 4. DIIBVS 5 =diebus ; the second I lacks the horizontals which would make it into an E. Another Change by Subtraction. 5. EST 6 =esset. The engraver has omitted two letters and changed the form into one more familiar. 6. ATQVE 7 = ad quem. The variation of final d and ¢ is common, especially when sound assimilation would justify it, as here, and the product isa familiar word. The final m is a weak sound which was often neglected in writing. 7. ET QVOD EVM EVM 11 = quod eum. There isa wrong insertion of et and Dittography of eum. 8. VTET 11 = ute. The horizontal stroke has changed I into T; a Change by Addition. 9. PR VRB EVMQVE 11 = pr(aetorem)que urb(anum) eumque. The QVE has been omitted after PR. 10. QVE 13 = quei, nom. sg. masc. ; the final 7 is omitted. 11. ETA 15 —ita; a Change by Addition, since the I has been wrongly provided with the horizontal strokes. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 49 12. curato has been omitted in 15. 13. EORVM 15 = forum ;a Change by Addition. 14. DABVNT 17 = dabit, in queiquomque frumentum populo dabunt damdamve curabit, nei quoi eorum...; but the subject, queiquomque, is in form ambiguous with the nominative plural, which may account for the error. 15. QVE 17 = quoi, dat. sg.; possibly a confusion with the com- mon enclitic -gue ‘and’. 16. AD COS 17 = a co(m)s(ule), to judge by the formula in 147; we may regard it as confusion of two common prepositions by an engraver who knew the language very imperfectly. 17. QVEI 18 = quoi, another confusion of forms. 18. QVAE/VIAE ERVNT 20-21 = quae via erit; for the copy ran -QVAEVIAERIT, and when the engraver had cut quae viae he felt obliged to change rit to erunt. 19. AO 23 = aqua; the O stands for Q, a Change by Subtraction, and the other two letters are omitted. 20. QVEI NVNC SVNT 24 is evidently a later insertion in the copy and lacks sense ; it must be deleted. 21. MAC 24 = mag(istratum), an error of Change by Subtraction, and not a use of cin the value of g, for MAC occurs but three times in the inscription (here and Nos. 22, 85 = lines 25, 105) and MAG is found fourteen times (lines 25, 69,73, 80, 84, 90, 95, 98, 100, 133, 140, 143, 144, 1533; in 80 the word is written in full). 22. MAC 25 = mag(istratu); see No. 21. 23. IN VRBEM ROMA 26 = in urbem Romam ; graphic omission of the weak sound, the final m. 24. L 26ora character much like it, stands for M ‘ thousand’ ; the engraver has not understood the symbol, which has the form of an 8 on its side. 25. VIAM PER 29 = via inter; the four strokes of IN are run together to form M and T is then easily changed to P, whereupon the words are redivided, according to their apparent meaning. 26. SVVM 32 = suum, is an early example of dissyllabic -u-u-, and not an error, although this inscription usually shows -w-o- in such groups, cf. suom 34. 27. EI 35 = et, as in No. 3. 28. OSQVE 38 = eosque ; an error of Omission ; but there is a gap on the Table before the 0, which indicates that the E may have stood there but have become illegible. 50 LANGUAGE MONOGRAPH NO. 2, 1926 29. TABVLA 39 = tabulas, since there is no room at the end of tae line, tor the s which should stand there. 30. QVII 40 = quei; with I for E as in No. 4. 31. PROXVMIIS 41 = proxumeis ; the same error as in No. 30. 32. OVOI 42 = quoi ; where lack of the tag has turned Q into 0. 33. DAPE 43 = dare; again a change by subtraction. 34. RIM 44 = rem; asin Nos. 4, 30, 31. 35. IVDICIVMQUE 45 = iudiciumve; a letter has been inserted, changing one common enclitic to another. Before this word, iudicem has been omitted, on account of the close ressemblance ; Haplography with Skipping. 36. OPORTEBIT 45 is miswritten for the present subjunctive, since it is governed by dato utei. The future of this word occurs many times in this inscription, and either the drafter or the engraver has become confused in some way. 37. TVENDAM 45 was first engraved wittt initial L or D, and was then corrected. 38. gquei redemerit 47 seems to have been omitted after wte7. 39. M 50 ‘ thousand ’ has been omitted, probably because of its unfamiliar appearance ; cf. No. 24 40. PL VESCS C 52 = pl(ebeive sc(iteis) s(enatus Jue c(onsulteis). VE has been omitted after S = senatus. 41. EVM 52 should be eius,cf. the formula in CILI. 27. 589. I. 30 == CILI'. 204); that it is the genitive plural is less likely. 42. PROCVRATOR ERIT 55 stands for procuratio erit, with an omit- ted letter and a Tele-Dittography. 43. HABETABETVR 56 = habitabitur, with I twice transformed into E by the addition of the horizontals: 44. ADVHEI 58 = advehei; an Omission of a single letter. 45. DEMOLIENDA LOCA ERVNT 59 = demoliendae locatae erunt. The omission of the final of the first word can hardly be motivated, unless it is to make it agree apparently with the LOCA which was actually to follow. The omission of -tae in locatae can be explained as a passing over from the first A to the second A, and haplography of the E which ends the word and the E which begins the next word ; the product is moreover a familiar word. 46. REX 62 = regem; because the preceding virgines and the fol- lowing flamines, though really accusative, are identical in form with the nominative, and affect the intervening word. 47. TRIVMPHAVIT 63 = triumphabit ; an Error of Change. The KENT, TEXTUAL CRITICISM OF INSCRIPTIONS Sue inscription is too early for us to accept a change in the pronunciation of b toward a spirant sound. 48. caussa 64 seems to have been omitted after LVDORVM. 49. QVEI ROMAE AVT VRBEI ROMAE PVBLICE FEIENT 64 stands on the Tablet, but PP M == Pp(ropius) p(assus) m(ille) has certainly been omitted after the second Romae ; cf. the formula in line 77. This is a Haplography with Skipping, since the letter P is the first of the omitted letters and is the first letter to follow those which are omitted. Curiously, in CIL I. 2? (but not in CIL I*), aut urbe Romae is omitted from the text; a fine example of Haplography in modern times. 50. QVOVE 65 = quo, by Tele-Dittography ; since there are many ve’s in the inscription. 51. H X 67 stands for ad h(oram) X, with Omission of ad; or, less probably, for h(oris) X, in which case there is no error in the text. 52. BVBLICAE 68 = publicae; a Tele-Dittographic Change of P to B, helped by the similarity in form of the two letters. 53. FORVMVE 69 = ecorumve ; E becomes F by omission of the bottom horizontal ; the reverse of No. 13. A Change by Subtraction. 54. INMOLITOMVE 70 = inmolitumve. The 0 for V in the declen- sional ending does not denote a survival of the earlier vowel, since this change took place by the beginning of the second century B.C., but is either the product of a Greek influence (since the unchanged vowel persisted inthe corresponding Greek form, and this inscription comes Greek-speaking territory), or is without assignable cause. 55. E,at the end of 72, stands for est; the last two letters are omit- ted, since it is at the end of the line, and there is no room. 56. DIXET DIXERIT 74 = dixit dixerit; the second I of the prior word has been written E, either by the influence of the next word (Tele-Dittographic Change), or simply by the erroneous addition of the horizontals. 57- QVOQVE 75 = quoigque, with omission of a letter ; the product being a familiar word. 58. quoi 80 has been omitted before QVISQVE, by Haplography with Skipping. 59. SVERAGIO 84 = sufragio; the E for F is a Change by Addi- tion. 60. MVNICIPIA 84 = municipei; the Change produces a familiar word. 61. FORO 85 = fori ;a Tele-Dittographic Change, as séveral neigh- boring syllables contain 0. 52 LANGUAGE MONOGRAPH NO. 2, 1926 62. QVE 85 — quem, as in line 7. Omission of a letter representing a weak sound, though the inscription normally writes the letter. 63. COLONIAE 75 = colonia ; an error of Addition. 64. PRAEFECTVRAT 85 = praefectura ; an error of Addition. 65. EORO 85 = foro; a Change by Addition. Cf. Nos. 13, 53. 66. The preposition in is omitted before senatum 86. 67. COAPTATO 86 = coptato ; apparently an instance of Tele-Dit- tography; or else the engraver was etymologizing, and doing so wrongly. 68. MAIORE 91 (so Ritschl’s plate ; but CZ reads MAIORA) stands for maiorem; an error of Omission. 69. ANNEIS 92 = annueis; a Pseudo-Haplographic Omission, for V resembles the last two strokes of the N which precedes it in the word. 70. After praecedere oporteat 92, there is an omission of dum taxat quod ei legibus pl(ebei)ve sc(iteis) procedere oportebit ; an excellent example of Pseudo-Hapiography with Skipping. The formula 1s restored from lines 102-103. 71. VOCATO 93 = vocatio; an Omission, perhaps Pseudo-Haplo- graphic, since I is identical with the hasta of T, which precedes it. But there is a vacant space on the bronze between T and 0. 72: FOIDERE 93 = foedere; a Change by Subtraction, asin lines , 5, 40, 41, 44, etc. This is not merely the older oi, for we have FOE- DERE in line 103, and of was no longer used at the date of this in- scription. 73. QUE 94 = quei; an error of Omission. 74. COLONIAE 98 = colonia (abl.); an error of Addition. 75. QUICT 98 = Quinct(iles); an error of Omission, since N is not elsewhere in this inscription omitted in such a group. 76. If VIR IIR VIR 98 = Ilvir(eis) IlIIvir(eis); Change by Tele- Dittography. For the correct text, cf. line 136. 77. ANNEIS NATVS 99 = anneis XXX natus, cf. line 89 ; an error of Omission. 78. II VIR II VIR 99 = Ilvir(um) IlIIvir(um) ; Haplography. 79. QVA EI 102 = quae ei ; Haplography. 80. PROCEDERT 103 = procedere; the change of E to T is aided by the partial identity of the strokes composing the two letters. 81. LIBITINANVE 104 = libitinamve. A ligatured MV in the copy may have been separated into two letters by the engraver, with the result NV; orasingle stroke of the four making up the M may have been omitted. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 53 82. OVID 105 = quid, an error of Change by Subtraction. Note that the line division comes not between this quid and the following faciet, as is given in CIL 1? and by E. Schneider, Exempla Selecta, p. 86, but four words earlier, between Jibitinanve and faciet. 83. After the second faciet 105, there seems to be an omission of in municipio colonia praefectura ; cf. lines 94-95. 84. WIF VIR 105 = J/lIvir(um); a curious error of Change by Addition. 85. MAC 105 = mag(istratus), as in No. 21. 86. After conscriptorum 106, the enclitic -ve is omitted. 87. COLONIA PRAEFECTURA 108 = coloniae praefecturae ; the final letter is omitted in each word, to make them agree in ending with the preceding municipia. 88. Before conciliabulo 109, there is an omission of foro. 89. Before senatu 109, there is an omission of in. 90. QVOE 109 = quoi; a Change by Addition of the horizontal strokes. g1. SENTENTEMTIAM 110 = sentemtiam; Dittography, with repe- tition of NTE induced by the recurrence of E. 92. ISPE 110 = ipse; an error of Change by Metathesis of the letters, perhaps aided by the familiar pronoun is. 93. FIDVCIO 111 = fiduciae; an error of Change, Tele-Dittogra- phicin nature, since the preceding word is iudicio and the following words are pro socio, all of them ending in 0. 94. FECET FECERIT 112 = fecit fecerit; an error of Change, as in No. 56, where the identical phenomenon occurs. 95. After im iwre 113, there is an omission of bonam copiam abiuravit, or the like. 96. After quei 113, the enclitic -ve is omitted. 97- QVID PRAEFVIT 116 == quei i(ure) d(eicundo) praefuit ; an Omission of two letters, the second one by Haplography. 98. In line 117, there is an omission of bona possessa proscriptave sunt erunt before possessa proscriptave sunt erunt ; virtually by Haplogra- phy, although that does not explain the loss of the first word, bona, which differs from the words which cause the contusion. 99. RESTITVS 118 = restitutus; an excellent example of Haplo- graphy. 100. INGNOMINIAE 120 =onis; cf. No. 104. . destantis 17 = S. Il. 3: for destandis ; Haplography with also a Tele-Dittographic Change. Errors of Addition : 7. summus 7 =S.1.10 : for swmus ; Dittography. 8. deiprachensa 9 ==S.1.14 : tor deprehensa. Aw Sw ty 60 LANGUAGE MONOGRAPH NO. 2, 1926 9. quaedam 9 =S.1.15 : for guadam. 10. prorumprimus 10 = S.1.16 : for prorumpimus ; Tele-Dittogra- phy. Ir. exsercitos 12 = S.1.20: for exercitos; an attempt to indicate the double sound of x, and probably not to be counted as an error of orthography. 12. adfaciam 14=S.1.25 : for adfatim ; the inserted letter changes the rare word into a familiar one. On ¢ for t, see No. 25. 13. difficiale 15 = S.1.27: for difficile, perhaps by the influence of speciali in the same line ; Tele-Dittography. 14. remediis 15 == S.1.26 : for remedii. . Errors of Change : 15. Sarm 2: for Carpic, repeating the abbreviation which is used a few words before, with the same words preceding and following ; Tele- Dittographic Change. 16. dignatas 4 = S.1.2 : for dignitas ; Tele-Dittographic Change. 17. cordicionemque 6 = S.1.7 : for condicionemque ; by likeness of the letters. 18. raene 8 =S.1.13 : for paene, by likeness of the letters. 19. aefas 9 = S.1.15 ; with a for xn, by likeness of the letters. Or possibly the initial 2 was omitted, and ae was written for e. 20. diu rerum 10 = S.1.16: for diu verum ; Tele-Dittographic Change. 21. obtunisi 11 == S.1.18 : for obtumsi, with splitting of the m into two letters. The presence of -ims- and not -s- in the original is assur- ed by -ms- in the Stratonicean copy and in the Plataean copy. 22. excorris 11 == §.1.18 ; for extorris ; by likeness of the letters, though not in the forms used in this particular copy. 23. senpen perdere 12 ==S.1.21 : for semper pendere; with metathe- sis of the letters, so that two n’s fall in the one word and two 7’s in the other. 24. ros 13 = S.].22: for imbri]bus ; by likeness of the letters. 25. adfaciam 14 = S.1.25 : for adfatim ; cf. No. 22, and No. 12. 26. patio 14 = S.1.26 : for ratio ; by likeness of the letters. 27. semn omni 16 == S.II.1: for sed in omni. 28. demque 17 == S.I1.2: for denique ; crowding has changed ni into m, the converse of No. 21. 29. destantis 17 == S.II.3: for destandis ; see No. 6. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 61 THe STRATONICEAN Copy. The errors are classified as in the Egyptian copy. Errors of Omission : . licetranquillo l.1, for licet tranquillo ; Haplography. . qua 1.4, for quae. . Increnta 1.5, for incrementa ; Haplography with Skipping. . debachandi 1.6, for debacchandi ; Haplography. . in dieiusmodi 1.6, for in dies eiusmodi ; Haplography with Skip- . adque ticendi 1.7, for atque reticendi ; Haplography with Skipping. . ut 1.13, for aut. . remedi tuenta 1.13, for remedia inventa. 38. iura natura 1.13, for iura naturae ; Omission by Graphic As- similation. . nef 1.15, for nefas. . uergent 1.15, for vergentes . . atrocissima 1.16, for atrocissimae. . planeiusmodi 1.20, for plane eiusmodi ; Haplography. . Offici exercitos 1.20, for officia exercitos . . afluntiam 1.23, for affluentiam. . divitis 1.24, for divitits ; Haplography. . popul 1.25, for populos. . potuisent 1.25, for potuissent ; Haplography. . or 1.27, for orbe. . iutior 1.28, for iustior. . posit Il.2, for possit ; ree . se I1.8, for esse. . quadamensitate 11.8, for quadam immensitate ; Haplography with Skipping. 53+ . praefinitauaritia 11.12, for praefinita avaritia ; Haplography. . discursum 1.15, for discursuum ; Haplography . . subigetur IL.19, for subiugetur. . idem 11.20, for eidem (dative singular). . auaritia I1.20, for avaritiae. . ab iusmodi I1.21, for ab eiusmodi. . habenspecies 1.21, for habens species ; Haplography. . stui 1.22. for statui ; Haplography with Skipping. Or possibly igitur a pretia \l.9, tor igitur ea pretia. for sibi, with v for 6, and 7 converted into ¢ by a crossbar. 62 LANGUAGE MONOGRAPH NO. 2, 1926 62. ese II.22, for esse ; Haplography. 63. obsequis 11.23, for obsequiis ; Haplography. 64. religione atur 11.23, for religione teneatur; Haplography with skipping. The reading teneatur is found in line 99 of the Delphian copy, and supplants the custodiatur of the CIL and other.editors. Errors of Addition : 65. festinant 1.5, for festinat ; Tele-Dittography, by the influence of the preceding words momentis, incre>nta, augmenta, each con- taining the group -nt-. 66. fidest clamat 1.13, for fides clamat ; Tele-Dittography. 67. cobiubemus 1.13, for cobibemus ; perhaps Dittography, since w (as consonant) and b represented the same sound, when intervocalic. 68. ferenddae 1.14, for ferendae ; Dittography. The adn. crit. in CIL 3.1129 gives two d’s, but the typeset transcript on p. 804 gives only one. 69. |tuss 1.17, for interventus ; Dittography. 70. cconuersatione 1.19, for conversatione ; Dittography. 71. poterant Il.9, for poterat. 72. subbliti 11.10, for subditi ; see also No. 90. 73- breuiis I.10, for brevis; Dittography. 74. uttique Il.11, for utique ; Dittography. 7S eSSEO ALLO TS OL esses 76. iinmunis Il.21, for inmunis ; Dittography. Errors of Change : 77. at 1.8, for ac. 78. relicio 1.9, for religio ; by likeness of the letters. 79. puplico nefas 1.15, for publicum nefas; p for b by Tele-Ditto- graphic Change. 80. quereliarum 1.16, for querellarum ; by likeness of the letters. 81. praeceptiutem 1.18, for praeceptricem ; by likeness of iu to ri. 82. quit 1.22, for quid ; but this variation may be phonetic, as in at and ad, etc. 83. prouenise 1.23, for provenire. 84. laceratricet 1.25, for laceratrices . 85. ted 1.26, for sed ; perhaps by Dittographic Change, as the pre- ceding word is persuadet . 86. Jatuere II.1, for exto|rquere; by likeness of at to rq. 87. diet Il.5, for dies. 88. sensuant Il.5, for censeant ; two changes, that of the c to s being presumably Tele-Dittographic. The Delphian copy, lines 68-9, has KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 63 the curious variation nesciant, which is possible only if a preceding ne has been lost by Haplography. 89. conprdhensa 1.5, for comprehensa. 90. subbliti 11.10, for subditi ; by likeness of / to d. Cf. also No. 72. 91. ubscruantia IL. 10, for observantia; by likeness ofc to e. 92. eaciem II.11, for eadem ; since the strokes of d, if separated, make ci. 93. beatitudone IL. 12, for beatitudine. 94. extstamaverit Il. 22, for existimaverit ; Tele-Dittographic Change. 95. erismodi I1.24, for eiusmodi ; by likeness of ri to iu; cf. No. 81. THe PLarakan Copy. _ As in the consideration of the Egyptian copy, the equivalent place in the Stratonicean copy will be given. Errors of Omission : 96. tranquillorbistatu 1 == S$.1.1 : for tranquillo orbis statu; two instances of Haplography . 97. paen 7 = S.1.5 : for paene. 98. glicentis 12 = 8.1.8 : for gliscentis. 991 16 26=-'5.1. 17's for tot. 100. uertatibl 30 = S.1.20: for uuertatibus = uber- ; Haplogra- phy. 101. fluentiam 35 = S.1.23 ; for afluentiam or affluentiam. 102. officis 36 = S.1.24: for officiis ; Haplography. 103. quarum necessitas 39 == S.1.27; has the final m and the initial m run together into a ligature like a five-stroke m ; virtually a partial Haplography . 104. nominaestimonis 46 == S.I1.2: for nomina aestimonis. The prior omission is haplographic ; the second, that of -ati-, is unmo- tivated, but was in the model copy, since the same omission recurs in the Egyptian copy (cf. No. 5), and in the Delphian copy, line 63, which has apparently eptimonhs (the h being of course the Greek +, and not a consonant). The words fall in a lacuna of the Stratonicean (ext. 105. sec{torius 49 == S.II.4: for sectorivus = -ibus ; Haplography. 106. permoti cum 51 = S.11.6: for permoti ut cun.. 64 LANGUAGE MONOGRAPH NO. 2, 1926 Errors of Addition : 107. ardataauaritia 6 =S.1.4: for ardet avaritia : Dittography. See also No. 118. 108. baccanidiligentiam 9 =S.1.6 ; for [de|/bacchandi licentiam. ‘Tele- Dittography ; see alsoNo. 121. 109. nustraetcunf| 17 = S.1.12 : for nostrae confleratur|; the t is pseudo-dittographic for c, since the two letters are very similar in the script of this inscription, and the false writing produces an apparent nostra et conf-. 110. superilloumedillae 25 (sic! in facsimile) == S.1.17 : for super- fluo medellae. f and i are confused, there is dittography of /, and there is metathesis of wo. See also No. 131. III. énstilutorum 36 == S.1.24 : for institorum ; by Pseudo-Ditto- graphy. 112. consenctentur 37 = S.1.25 : for consectentur ; the inserted let- ter denotes a weak sound. 113. idirigi 44 = S.1.30: for dirigi ; Tele-Dittography. Errors of Change : 114. gentrum 4 =S.1.3: for gentium, Tele-Dittographic, by as- similation to the endings of barbararum ipsarum, which immediately follow. 115. sundatal 5 = S.1.4: for fundatam ; by likeness of the letters. 116. dibitum 5 = S.1.4 : for debitis. 117. iusticiue 5 = S.1.4 : for iustitiae ; by likeness of the letters. 118. ardataauaritia 6 = S.1.4 ; for ardet avaritia ; Tele-Ditto- graphic Change of the e to a. See also No. 107. 119. qut6=S.1.4: for quae. 120. eontinentiae 8 = S.1.5 : for continentiae ; by likeness of the letters. 121. / baccanidiligentiam 9 == S.1.6 : for [de|/bacchandi licentiam ; perhaps thec is changed to g by the influence of the word diligentiam. See also No. 108. 122. tad 9 == S.1.6: for laclerantur]; by likeness of the letters. 123. detestaniam 10 =S.1.7: for detestandam ; by likeness of the letters. 124. ahbbere 12 ==S.1.8: for habere ; Metathesis of the letters. 125. religic 13 == S.1.9: for religio ; the result of an uncompleted stroke. 126. existimatup 13 = S.1.9 : for existimatur ; again, the result of an unfinished stroke. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 65 127. promisionis 17 = S.1.12 : for provisionis ; the result of added strokes to the w. 128. / cepe 19 =S.1.12: for [ha] /c spe. 129. /are direptionis 21 = S.1.14 : for fere] /ndae direptionis ; for nd is easily corrupted into ar, and e¢ stands for ue. 130. pualicum 22 =S.1.15 : for publicum ; by likeness of the letters. 131. superilloumedillae 25 = S.1.17: see No. 110. 132. odtumsi 27 = S.1.18: for obtumsi ; by likeness of the letters. 133. cenpestalisque 32 = S.1.21 : for tempestatesque ; by likeness of c and ¢. 134. capiare 32 = §.1.21: for captare ; by likeness of the letters. 135. indribus 33 = S.I 22: for imbribus ; by likeness of b and d. 136. pualicae 35 = S.1.23 : for publicae ; as in No. 130. 137. adaritiae 38 = S.1.26 : for abaritiae = avar- ; for v and J intervocalic were sounded alike, and d and 5b were almost alike in shape. 138. scatii 38 = S.1.26 : for statui ; with two changes, the for- mer by the likeness of the letters. 139. difficile 40 = S.1.27 is somehow miswritten, so that the final letters are quite illegible. 140. colo 40 = S 1.27: for toto; by likeness of ¢ and ?. 141. protoelari 41 = S.1.28: for revelari. 142. intepl| 41 = S.1.28 : for intell[egatur]. 143. mentiorn (sic!) 42 == S.1.29 : for mentium; the o is more or less phonetic for u, but the rn is the product of careless writing and splitting the m into two letters. 144. acnosdebe 43 = S.1.29 : for agnoscere ; by likeness of the forms of all three pairs of letters which are changed. 145. interdam 47 ==S.II.2: for interdum. 146. detesdandis 48 = S.11.3 : for detestandis ; Tele-Dittographic Change. 147. emcritus 49 =S.II.4 : for emeritos ; by likeness of c and e. 148. /1 depraedatores 50 = S.11.4: the lost word before depraeda- tores is quo, so that some error lurks in the hasta forming the i. 149. censuamus 54 =S.II.8: for censuimus. The errors in these three copies of the preamble to the Edict of Diocletian have already been clasified in the three major groupings, but it is worth while to index them in the more exact sub-divisions. No attempt has here been made to distinguish Changes by Addition 66 LANGUAGE MONOGRAPH NO, 2, 1926 and Changes by Subtraction, since the letters are rather irregular in shape and it would be hazardous to make the distinctions ; but a new category, that of Change by the Likeness of the Letters, takes their place. Note that Nos. 1-29 are in the Egyptian copy ; Nos. 30-95 are in the Stratonicean copy ; Nos. 96-149 are in the Plataean copy. I. Errors of Omission : Omission, unmotivated : 1, 2, 3, 4, 5; 19 3 31, 37, 39, 40, 41, 43, 44, 46, 48, 49, 51, 53, 59, 57, 58, 59 397, 98, 99, 10I, 104, 106. ; Omission of a letter, to produce a familiar word : 36. Haplography : 30, 33, 42, 45, 47, 50, 54, 55, 60, 62, 63 ; 96, 100, 102, 103, 104, 105. Haplography with Skipping : 6, 32, 34, 35, 52, 61, 64. Tele-Haplography : 38. Il. Errors of Addition : . Addition, unmotivated : 8, 143; 71, 75. Dittography : 7; 67, 68, 69, 70, 72, 73, 74, 76 ; 107, 110. Tele-Dittography ; 10, 13 ; 65, 66; 108, 113. Pseudo-Dittography : 109, 111. Addition to produce a familiar word : 9, 12, 14. Addition of a letter representing a weak sound : 75, 112. UI. Errors of Change : Change, unmotivated : 27 ; 82, 83, 84, 87, 88, 89, 93 ; 116, 119, 128, 138, 139, 141, 142, 145, 148, 149. Change by likeness of the letters : 17, 18, 19, 22, 24, 25, 26; 61, 78, 80, $1, 86, 90, 91, 95 3 115, 117, 120, 121, 122, 123,125) 1126) 127, Rag} nae, 131, 132, 333, 134, 135, 136, 137, 138, 140, 144, 147. Dittographic Change : 85. Tele-Dittographic Change : 15, 16, 20, 29 ; 79, 88,943; 114, 118, 146. Change by Ligature: 27, 28. Change by division of a character : 21 ; 923 143. Change by Metathesis : 23 ; 124, 131. Change to a familiar word : 77. IV. Examples listed, but probably not errors : Pigjio2% GRAPT ER Vill CONCLUSION It is time now to sum up the results of this investigation, which has been carried through seven inscriptions (one of them in three copies), in five languages, and has listed the presence of about 450 errors. Such a summary is the more necessary, because the study has inevitably advanced item by item, without any generalizations except those implied by the tabulations at the end of each chapter. The inscriptions dealt with have varied from the very carefully engraved Inscription of Darius to the slovenly Lex Julia Municipalis and the incompetently cut Edict cf Diocletian. As it happens, the first and the last of the inscriptions examined are cut in stone, and all the others are on bronze tables. It might have been desirable to include some of the official documents of Athens, cut on marble in letters which have a regularity of form matched only by the mechanical pro- cesses of modern times. But the errors in them are quite infrequent , and a scrutiny of the critical apparatus in W. Dittenberger, Sylloge Inscriptionum Graecarum, 1, for a very considerable number of inscrip- tions, reveals no phenomena not found in the inscriptions already studied in detail. Apart from the ordinary omission or insertion of single letters, I have noted in them the following errors of change of single letters: Correct. Letter Actually Inscribed A changed to A, A r= Y> b, P » | A= » A E » Cc, 1,1 H aa cf) » Il, | Oo » O | » ob M » IN N » Pe F » r Y » T ¥. » Y 68 LANGUAGE MONOGRAPH NO. 2, 1926 The slightest examination shows that these changes are almost all effected by the omission or addition or transposition of a single stroke, or of part ofa stroke, and are therefore precisely similar to the changes which have been found elsewhere. It is true that the Latin inscriptions studied have been documents written by persons with little knowledge of Latin, presumably Greeks or persons with a knowledge of Greek. A hasty review of the longer official decrees in the first volume of the CIZ will show varying degrees of care exercised in the work of engraving. Thus: The Epistula Consulum ad Teuranos de Bacchanalibus, commonly known as the Senatusconsultum de Bacchanalibus (CIL 1'.196 = 17.2.581), engraved in 186 B.C., found 1640 at Tiriolo in Calabria, and now in the Imperial Museum in Vienna, is written with great care, and has but six or seven errors. The Lex Bantiae (CIL 1'.197 = 1*.2.582), written on the reverse of the bronze tablet whose obverse bears the Oscan Tabula Bantina, was engraved between 133 and 118 B.C., and was found in 1791 near the ancient Bantia; it is now inthe National Museum in Naples. It has few or Lo errors. The Lex Repetundarum (CIL 11,198 = 12.2.583), of 123 or 122 B.C., is ona bronze tablet now in many pieces, probably found together and soon brought to Rome; one piece was edited in 1521, but there is noother clue to the date of discov- ery. The most important pieces are in the National Museum at Naples, others are in Vienna, still others have been lost. The inscription contains a great many errors, including a long repetition near the end (lines 72-79 = lines 79-85). The Sententia Minuciorum (CJL 11.199 = 1°.2.584), of the year 117 B.C., was found about 1506 near Genoa, and is now in Genoa, in the Church of San Lorenzo. It contains very few errors. The Lex Agraria (CIL 1.200 = 1?,2,586), of the year 111 B.C., is inscribed on the reverse of the tablet which bears the Lex Repetundarum on the obverse; it contains a great number of errors. The Lex Cornelia de XX Quaestoribus (C/L 11,202 = 1*.2.587), of 81 B.C., was found at Rome in 1528, and is now in the National Museum at Naples ; it is almost free from errors. The Senatusconsultum de Aesclepiade Polystrato Menisco (CIL 1',203 = 1?,2. 588), of the year 78 B.C., was found in 1570 at Rome, and is now in the National Museum at Naples. The text is inscribed first in Latin, then in Greek, only a little of the Latin version being extant ; it is correctly written. The Lex Antonia de Termessibus (CJL 1'.204 = 17.2.589), of 71 B.C., is on the reverse of the Lex Cornelia; it is correctly written. The Lex Municipii Tarentini (CIL 1*.2.590), engraved between 89 and 62 B.C., was found in 1894 at Tarentum, and is now in the Public Museum of Naples. It con- tains a few errors only. The Lex de Gallia Cisalpina (CIL 1,205 = 17.2.592) of 49 B.C., was found in 1760 near Veleia, and is now in the Museum of Parma. It has but a few errors. The Lex Coloniae Genetivae Juliae sive Ursonensis (CIL 1?.2.594), of 44 B.C., is on four tables, found 1870-71 near Osuna, the ancient Urso, in Spain, and is now in the Museum in Madrid, It contains many errors. KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 69 This hasty review of certain more important Latin documents of good date, all on bronze tablets, shows the variety in the quality of the engraver’s work, for some inscriptions have few or no errors, while others are full of them. In the examination of the errors found in those inscriptions which have been under detailed examination, an attempt has been made to define, so far as may be, the nature of the error and its probable cause. It is impossible to arrive at objective certainty in such matters ; another investigator would doubtless classify and motivate a consi- derable number of the errors in a different way. But it is probable that the kinds of error and the causes of the errors would, for the total number of examples, differ but slightly from those given here — except for a somewhat differing terminology. It is likely also that an examination of additional inscriptions would add but few additional kinds oferror and but few additional causes of error ; and these could readily be identified by the application of the procedure used in this investigation. For these reasons, I may here present the various kinds of errors in a tabulation which combines the items of the indexes at the ends of the foregoing chapters. The capital letters at the ends of the lines are used to indicate the inscriptions in which the special kind of error is exemplified, as follows : P = Old Persian : the inscription of Darius at Behistan. G = Greek : the Locrian treaties. O = Oscan : the Tabula Bantina. U = Umbrian : the Iguvine Tables. L = Latin : the Lex Julia Municipalis. E — Latin : the Edict of Diocletian. Errors of Omission. 1. Unmotivated omission acronone of moreletters of a: word. .225, sedis ein ale PGOULE MRSA TNC VO a Oo cs minus ats id enna liege 4a 5 LE Peon awe OF IMOte, SUCCESSIVE “WOTCS a. ai iie Ages an aistine «#4 c= 94 bi 2. Haplography emaPOrYial NADIOOTADINY coo 00. os sieccart are nia ning Aatebiatans Sto @ PGULE b. with skipping of the interpuncts only,......-..... 00. eee eens U c. with skipping of letters or words..............00eeeeees PGULE Sreereer ADIOUTADO Yoo Ct Atal, cds em al phlei tat hs bud Ao td eas 28 PULE e. pseudo-haplography................ eerie Be sp ties Re PUL f, pseudo-haplography with skipping......-............0ee000- UL 3. Omission caused by lack of space and crowding at the end of the line ee ei is seek eR athe a hee A ile aah ae a sod iw em oe UL 7O Si 6. LANGUAGE MONOGRAPH NO. 2, 1926 Omission by the influence of the orthography of another language fa- miliar to the engraver 0.0 be pies jew Ve 5 e oils oe elem aie ee 6 Shee iki ele & Oke eels Omission of semi-linguistic nature a, tO produce 2 familiar WOrda ss vie cine th ae LE b. to produce apparent syntactical agreement.................008 LD, eliof aletter:representing a weaktotnd i! 20), GV aen eee eee UL d.\ for apparent phonetic accuracyysiere. Ho. Wise. GO ee P é..;helped by phonetic dissimilation. ..\.)5 cil sis a sae lee bt G Errors of Addition. 1; Wnmotivated Addition. | mest ..2 yf satel ai ole ve ana ee ene ULE 2. Dittography armoritial Gittomraphy eit epic bee cate kare 2 (ah Rey nee PGOULE b. dittography caused by division between lines................- GU ci dittography-and/ ligatures spa, Hovieiie le. LRG SIR, GS. OE oie O d.,,tele-dittogTApn yb. \iunedaps tee hatte A aes a be tiee ere eae POULE ¢., pseudo-dittography 5 0s Fontes :. vin Wistersls,a sich4g ete tae lee ee E f. partial dittography, caused by division between lines............ U 3. Metathesis of a stroke inserted in correction............... 00000 O 4. Additions due to change of alphabet a., by wrong transliteration fr j.:5!.2a hil vp opie eal ble sles he Mey AI U b. by the influence of the orthography of the other language familiar to the COSTAVES J .c's vee a Ley Abilis RPh ee aE eee RO a hy Meer OU ¥ Additions of semi-linguistic nature a. to produce: a familiar, word..;. UP Moe ee oe EL Sg GULE b. to produce apparent syntactical agreement..............+.00- LE c. by assimilation to other forms of the paradigm................ U d. of a letter representing a weak sound.................+.05- UE e. of a letter representing a weak sound, helped by the influence of a word associated in, meaning’. 2. 55; soe vic e's Bele alin ly aon Ve ee G f, for apparent phonetic AGCUTACY. «vas. ss wives smut 6 cane + Re OR U g. of a phrase inserted by a wrong idea of the meaning of the passage L h. by popular etymology cece ere eee eee ee te te eee reer ereeeeeeesne Errors of Change. 1; ;Unmotivated (Gnange rs Cro... ac nt oe oe ait eae ae een ULE 2. Change by likeness of the shapes of the letters.................. E 3. Change by addinon of a ‘stroke (sas... «spine cs eee Be POUL 42S. dittographic Change ee a ce. Leek ae ty alr oene eae UE bi ‘telé-dittoprapbic Changer en re ean atke aes GOULE SLUR. DY (SUDCCACHIONSi0t ste enttns ck tae aad nenae ee PGOUL by ‘by subtraction and lgarure ss. so. <<, aps esa jnsiqud sie cess ed U cr by semi-haplography 0225 < hfe. | wa ach 5b ORE ea P 6; ta. by dipature oe tr Tee rie ane ie ot au hen OLE by ‘by crowding ‘and Lipature.', | ues se «cis ne * puna eee U FL ae DY WYONP AIVISION INTO WOLdS «ae cis» «+ vk» og he acne E bi by division‘of a’ thatucter inte twos... a. cam eee ae en L S Change'by metathesr or the tetters.ic 7. 224 «ake eae Lape LE KENT, TEXTUAL CRITICISM OF INSCRIPTIONS aus g. Change of an unfamiliar or unintelligible symbol. ................ it 10. Change by the slipping of the engraver’s tool................... U 11. Change by metathesis of a stroke in correction.................. U 12, Changes due to change of alphabet en ye WrORe VEADSHCEATIOLY ons cS rien, Pou ee OR Bed U b. by the influence of the orthegraphy of the other language familiar to she pagravey. ity iad . wey e el way Gets dou Busy a) OL 13. Changes of semi-linguistic nature a. by preserving older orthographic practices................... O Datonproduce atamillar WOrdis | ones Tel sreteur rete te, ULE c. to produce apparent syntactical agreement.................... ie d. by assimilation to other forms of the paradigm............... U It would be unsatisfactory to close without a classification of the causes of error, with illustrative examples drawn from the previous text. In so doing, the abbreviations used in the preceding table will be used, with the serial number assigned in the text ; the versions of the Edict of Diocletian will be distinguished as Ee, Es, Ep for the copies found in Egypt, at Stratonicea, at Plataea respectively. I. Identity or Similarity of Letters or Groups of Letters is a very fruitful cause of error, leading to partial or complete loss ; this is nor- mally to be termed Haplography, or some variety of Haplography. a. Ordinary Haplography, where a repeated letter is simplified, as is Es 30 licetranquillo == licet tranquillo ; where a repeated series of letters is simplified, as in L 99 restitus = restitutus ; where a repeated series of words is simplified, as in U 114 nomne erar = nomne erar nomne erar (the second nomne erar was inserted above the line when the engraver examined his work for errors). b. The Haplography may involve the loss of interpuncts between two words, as in U 34 :purtiiusuru: — : purtiius : suru: c. The Haplography may involve the loss of one or more letters or words, standing between those which are identical. Thus we find quoi quisque reduced to quisque L 58; incrementa reduced to increnta Es 32 ; conciliove habeto neive quis quem sei adversus ea comitieis conciliove reduced to conciliove L t1o. d. Tele-Haplography is a term which may be applied where one of two neighboring but separated units is lost, but the unlike element between is retained: thus :staherent:termnesku: has become :sta- heren:termnesku: U 21; iura naturae has become iura natura Es 38. e. Pseudo-Haplography occurs where a character is lost because of partial identity with its neighbor; thus annueis becomes anneis L 69, since V is virtually identical with two strokes of the N which precedes it. 72 LANGUAGE MONOGRAPH NO. 2, 1926 f. Pseudo-Haplography with skipping is sometimes to be found, as in U ri :iuvesmik: for :iuvie: esmik: since the v and the e are almost, but not precisely identical in the Umbrian alphabet. Or it brings about a longer omission, where the condition of absolute iden- tity of the last letters written before the omission, with the last letters omitted (or of the first omitted letters, with the first written after the omission), is not met, though much of the omitted passage is a repe- tition of words before or after it; as in L 70, L 98. g. Semi-Haplography is the loss of part of a character, because that part is identical with the adjacent character or part of a character ; Che Pee II. Duplication of Letters or Groups of Letters is another fruitful cause of error in inscriptions. There are the following varieties : a. Normal Dittography, as in sententemtiam L 91 for sentemtiam ; and in eum eum L 7 for eum. b. Tele-Dittography, where a letter or letters at a greater or smaller distance are repeated ; as in iiouinar U 100, for tiouina, the r being repeat- ed from the final r of the preceding words. c. Pseudo-Dittography, where the dittography is not exact; as in nustraetcunf- Ep 109, for nostrae conf-, since t and c in the alphabet there used are very similar. Also, in institutorum Ep 111, for instito- rum, where the inserted -tu- is a dittography of -ti- or of -to-. In both these instances, other common words are the result of the incorrect writings. d. Partial Dittography, where half a letter is made and then left unfinished, the complete letter being placed afterwards. We have found one example : ikuvinp/a: U 12, for ikuvina; the engraver decided that the space at the end of the line was insufficient for the a, and placed it in the next line, after having engraved part of the letter, which happens to be identical with p. e. Dittographic Change is found where a letter is altered to be iden- tical with a neighboring letter; as in ted Es 85, for sed, the preceding word ending ina ¢. f. Tele-Dittographic Change is found where a letter is altered to be identical with a letter not contiguous; as in II VIR IIR VIR L 76, where the IIR is changed from IIII in imitation of either the preced- ing or the following word, which contains 7 after 7. III. Lack of Space of the end of a line is the cause of errors of seve- ral kinds : KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 73 a. Omission of letters, as in :sevakne U 54, for :sevaknes, and ine L55, for est. Cf. also : purtu:/etu: U 43, for purtuvetu. b. Change of letters through the fact that they are united by liga- ture, as in :esum U 31, for esunu. c. Dittography of the final letter of the line, since it is sometimes written as the first letter also of the second line; asin baie /E G 13 == d4gr&y ; and in amprefu/us: U 23, for amprefus. d. Partial Dittography of the final letter ; cf. II d, above. IV. Crowding of characters, so as to produce ligatures : a. At the-end of a line, see III b, above. b. In the middle of a line, as in viam per L 25, for via inter, with in run together into m, and other alterations attending this; cf. also O 13, where dittography is combined with ligature. V. Division of a single letter into two letters, as in obtunisi Ee 21, for obtumsi. VI. Wrong Division of Words : a. The indication of word separation may be wrongly omitted ; see the introductory remarks to each inscription, especially those on the Iguvine Tables. b. The indication of word separation may be wrongly inserted in the middle of the word ; see the introductory remarks to the inscrip- tions, and also G 3. c. The indication of word division may be misplaced, as in viam per L 25, for via inter, cf. IVb, above. VII. Metathesis : a. Two letters may be transposed, as in abbere Ep 124, for habere ; senpen perdere Ee 23, for senper pendere. b. A stroke inserted to correct an error in the engraving, may be inserted in the wrong place in the word, where there is a similar group of characters, as in :persmhniu: U 4o, for persnihmu. It may even be inserted in another word, as in hafieist O 3, for hafiest ; the added letter should have gone into eizasc O 7, for eizaisc. VIII. The slipping of the engraver’s tool may make a mark which unintentionally converts one letter into another, as in :furenr: U 65, for furent. TX. An unfamiliar or unintelligible character may be omitted, as in L 39, or it may be changed, as in L 24. 74 LANGUAGE MONOGRAPH NO. 2, 1926 X. The engraver may know another language better than that which he is engraving, with certain results : a. That he uses his own grammatical endings, as in O 8, L 54. b. That he avoids unfamiliar vowel combinations, as inO 15, U7o. c. That he avoids unfamiliar consonant combinations, as in U 84. d. That he uses his own combinations of consonants, unknown in the other languages : U 119. XI. The change of the alphabet in which a document is written, may produce errors of transliteration : a. A letter of the old alphabet may have two values which are dis- tinguished in the new alphabet by the use of different letters ; and he who transliterates may choose the wrong equivalent : so anclar U 73, for anglar, and entelust U 107, for endelust. Also, iune U 76, for ioue, and pihaclo U 79, for pibaclu, into which the engraver did actually cor- rect it when he reexamined his work. b. The engraver may follow the practices of the new alphabet in such a way as to produce errors in the representation of the language ; see X b, c, d. c. The engraver may transliterate letter by letter, slavishly, and keep results that are at variance with the orthographic practices of the new alphabet : so in the writing of the glide consonantal -w- after vowel -u- in saluuom saluua tuua U 81-83. XII. Miscellaneous Errors of Carelessness : a. A letter ora word may be omitted, without determinable reason ; cf. advhei L 44, for advehei; the omission of curato L 12; the omission of a long phrase, L 83. b. A letter or a word may be inserted, without determinable reason ; Ci sor avr eter c. A phrase or clause may be inserted, by wrong understanding of the passage ; cf. L 20. d. Asingle letter or a group of letters may be changed, because the shapes of the letters are closely similar, and the change is easily pro- duced, usually with the omission or the addition or the change of a single stroke or part of a stroke; there are many examples of all these scattered through the examples studied in the preceding chapters. Or there may be no resemblance between the letters interchanged, so that only an undefined carelessness of the engraver is to blame. XIII. Errors of Semi-Linguistic Nature. a. The error is one which produces a familiar word, whether by KENT, TEXTUAL CRITICISM OF INSCRIPTIONS 75 omission of one or more letters, as when quoique becomes quogue L 57 ; or by addition of one or more letters, as when adfatim becomes adfa- ciam Ee 12; or by change of one or more letters, as when quoi becomes que L 16 or quei L 17. b. The error is one which produces an apparent syntactical agree- ment with near-by words, whereas the proper agreement is with a word at a distance. Thus we find colonia praefectura L 87, agreeing with the preceding municipia in ending, although both words should end in -ae; eas res L 121 has eas as though res were accusative plural, though it is really nominative singular, and the engraver should have cut ea res ; agunto L 119 stands for agito, since numerous plural words stand between it and its singular subject. Errors of this sort therefore result in Omission, Addition, Change. c. The error may be that of omission or addition of a letter repre- senting a weak sound. Such are the failure to write initial h in eretu U 32, and its wrong prefixing in hebetafe U 108, for ebetrafe ; the loss of final m in in urbem Roma L 23, the failure to write in cesendi L 124 = censendi, the wrong insertion of m in consenctentur Ep 112 = consectentur. d. The error may be caused by association with another word, as in itouie U 97 for iouie, and perhaps in coaptato L 67 = coptato. So also hayev G 14 = eye (q. v.; cf. alsoG 11). It will be noticed that these errors verge respectively on Dittography, Popular Etymology, Addi- tion of a Letter representing a weak sound. '_e. The error may be produced by the general scheme of the para- digm. Thus fratrer U 66 was first engraved, after the other forms of the plural (gen. fratrom, dat. fratrus) and was, later corrected by the engraver by the obliteration of the prior 7. Similarly, kapir U 8 stands for kapif (= capif, occurring three times), after other forms of the paradigm, such as kapire and kapirus. f. The error may be merely one of following a different orthographic practice ; thus arviu seems to have been written first in U 3, U 6, U 9, and to have been changed to the older arvia. g. The error may be one of violating an orthographic convention, to secure greater fidelity to the actual spoken sounds, as when’ the abso- lutely final short a is not written with the indication of length, P 6; so also in aruvia U 49, which seems to show a slow pronunciation of the usual arvia. h. The error may be one of omission assisted by a phonetic dissimi- lation, as in hoxotsvoyv G 1 == hinws Gévov, where -s ks- is graphically reduced to ks-; though this is an uncertain example. 76 LANGUAGE MONOGRAPH NO. 2, 1926 It should be noted that these last varieties of error, under the heading Errors of Semi-Linguistic Nature, may,or may not belong in this study. Are they errors of writing, or errors in the reduction to writing, or indications of the quality of the sounds which the letters represent more or less faithfully ? Different scholars would decide differently. Certainly XII c-h embraces kinds of error which must be included in the treatment of phonology and morphology; and yet they are categories which cannot here be ignored. For the loss of initial 4 in “hanser, under the influence of anas, and the development of a regu- lar spelling (and pronunciation) anser, isa definitely linguistic matter ; but the loss of an initial ) in Umbrian (h)eretu in one occurrence, as against 36 instances of writing the / in the forms of the word, is a distinct error, since itis against the orthographic norm. This leads us to the problem of defining Epigraphic Error, and at the same time gives the answer. I should define an Epigraphic Error as a writing which is at variance with the normal orthography of the inscription, and is not common enough in practice to be regarded as a permissible variation. In most instances, the engraver has examined his work for errors, either as he went along, or after he finished his work. The Umbrian Tables are notably full of corrections, where a letter has been deleted (U 1, 66) or changed (U 3, 5, 6, 9, 42, 44, 79, 99, 113, 116), or insertions of omitted characters or words has taken place, usually above the line (U 2, 67, 68, 90, 95, 112), but once below the the line (U 4), and once, when space permitted, on the same level (U 1); and in one word the correcting stroke has been inserted in the wrong posi- tion (U 40; cf. a similar phenomenon in O 3). Other corections by the engraver are attested by the blurs in the facsimiles of Tables La and II a and b; and still others have been mentioned in the study of the interpuncts which accompanies the discussion of the Tables. The editor of an inscription, or the scholar who desires to utilize the evidence of an inscription for linguistic purposes, must accordingly identify the epigraphic errors of these various kinds, before he can pro- ceed further with his task. If in this study the various kindsof erfor which occur, and the causes which lead to them, have been put on a somewhat more definite basis than heretofore, with a hint at what the engraver himself might do in the way of correction, the writer is amply satisfied. PROTAT BROTHERS, PRINTERS, MACON (FRANCE. — MCMKXVI)