A GRAMMAR OF CAN AND UMBRIAN WITH A COLLECTION OF INSCRIPTIONS AND A GLOSSARY BY CARL DARLING BUCK, Pu.D. PROFESAOR OF SANSKRIT AND INDO-EurorzAN Comparative. PHILOLOGY IN THE Univexsity oF Cimcaco, GINN AND COMPANY ROBTON + NEW YORK + CHICAGO + LONDON ATUANTA © DALLAS + COLUMBUS - SAN FRANCISCOCopyRIGit, 904, RY CARE, DARLING BUCKPREFACE Ture following work is an attempt to furnish in a single vol- ume of moderate compass what is most essential for the study of the Oscan and Umbrian dialects. In spite of the meagreness of the material, as compared with languages like Greek and Latin, and in spite of the many questions of detail which are still unsolved, the main features of these two dialects are well understood. And such is their relation to Latin that soine acquaintance with them. is important, not to the Indo-Europeanist alone, but to the student of the Latin language, and, in a less degree, to the student of the history and antiquities of Italy. In order that 2 knowledgo of the dialects should become more general, it is not enough that we have now such excellent works as Conway's Italic Dialects, with its full presentation of the existing material, and von Vlanta’s exhaustive Grammatik der Oskisch-Umbrischen Dialekte. Tho fullness of v, Vlanta’s treatinent, the conscientious weighing of possibilities, and the liberal citation of authorities, all add to its value as a wark of reference, but the resulting bulk of 1372 pages is likely to deter one who can devote only a moderate amount of time to the subject. ‘That there is need of a briefer grammar has Jong been the author’s conviction, which has only been strengthened by inquiries and suggestions from others in this country and abroad. In order to secure the desired brevity, it has been necessary to climinate almost wholly any detailed discussion of disputed points, as well as special references for the views adopted or rejected. Any one for whom the general bibliography given below is not suflicient may be referred to v. Planta. Only in a few cases, here 1 So Skutch, in a roview of the author's Oncan-Umbrian Verb-System, Reeliner Philotogischo Wocheuschrift, November, 18%: “Der Vert. kiime einem Reslirfniss centgegen, wenn er eine vollalindige Grammatik cles O.-U. in Masertab soines Verb= he. Denn uoben dem treflichen, aber weltschichtigen Werke v. Plantas Syrtoma sel rngcht.” Int oin kurzes Handbuch zur Einfibrung eriv Preface and there, I hare added references in footnotes, mostly to diseus- sions more recent than v. Planta, Generally I have simply stated the view which seemed to me on the whole the most probable, oF else contented myself with a non liquet, It is scarcely necessary to state that in matters of dispute I have had no predilection for my own previously expressed views, but have with equal freedom rejected them in favor of others or retained them against others, according to my present judgment, That the treatment is historical and comparative, not mnorely descriptive, is a matter of course, But tho emphasis is on Italic, rather than on Indo-European, relations. In the case of worda which are peculiar to the dialects and not found in Latin, a fairly wide range of cognates is cited, as in scctions 15, t6. But ordinarily comparison within the Italic is deemed sufficiont, and forms from other Indo-European languages aro introduced only for special reasons. Tho grammar is called 4 Grammar of Osean and Umbrian, not of tho Oscan-Umbrian dialeets, for it does not pretend to treat systematically the minor dialects included under the name Oscan- Umbrian, Most of the characteristics of these dialects (so fnr na they are clear) are mentioned incidentally, mainly in the Introduo- tion, But to discuss or even mention all the questions arising in the attempt to generalize from material consisting of only a few lines, would require an amount of space not justified by the results, Unless the material from these minor dialcets is notably incronsed, our knowledge of the Oscan-Umbrian group will be almost coinci- dont with what we know of its two principal dialects. And in this approximate senso a grammar of Oscan and Umbrian is also a grammar of Oscan-Umbrian. As the book has been practically ready for the press since the beginning of the year, and the Phonology in type since February, almost nothing in the literature of 1903 has been taken account of. But in what has appeared thero is liltlo which has entirely con- vinecd me. Special mention may be made of Brugmann’s discussion of the negative prefix an- and anter ‘inter’ (LV. 15,70 ff). I have myself wished there were somo way of equating these directly withProface v tho Tatin, instead of assuming by-forms (a8 in 98 with ¢), which indeod seems out of thequestion in the case of Anafriss if =D. mbribus (seo 98, 4). But Brugmann’s assumption that “ initial before nasal consonant hal a very open pronunciation in the Oscan-Umbrian period and had perhaps become identical in this position Italie a” fails to convince me, in view of O. embratur, Entral, and especially U. isegeles ‘insectis’. Nor do I see the necessity of soparating O. ant from I, ante because of its meaning ‘as far as’ (Gee 299, 2). Yor assistance I am indebted to Professors J. C. Rolfe and Minton Warren, who kindly offered to read proof, and especially to my pupils, Mr. W. C. Gunnerson and Mr. R. B. Nelson, who have gone over the proof with great care, devoting no small anount of time to the verification of references, citations, ete., and con- tributing in every way to the acrurney of the toxt. The remarkable keonnoss and intelligence of the proof-reader in the office of the publishers has also savod tho work from many blomishes. CD.B, Drowsnen, 1, In the interval of twenty-five years since this book was written, tho amount of new material that has come to light is disappointingly mall —less than the yield of almost any singlo year in the field of the Grock dialects. It is scarcely of sufficient consequence to neces- sitate, or justify the expense of, a revised edition, Still the new material has definitely settled some disputed points, notably in regard to tho Oscan eituna, Furthermoro, owing to more recent discussions or otherwise, the author’s opinion in various matters has been modified. ‘The oceasion of a new printing offers the opportunity to make some brief corrections and additions, with references to new atorial or new discussions. Soe pages 363 ff. - CD.TABLE OF CONTENTS Binuioanarny EXrLanations « INTRODUCTION Prorurs awn Lancuaces or Travy Crassiricaion oF tif Tratic Dratrers Osan —Exrrrwat, DATA Unmuran —Exteawat Data. Generac Cnanacrenistics or Tim O Puonorocy Inpuection Srwrax Vocanurany Summanr 0. Srectat, Cnaractenustics or Oscax Srectar, Cuanactenarics of Unnntan Borrowen Wouns . PHONOLOGY Aurnanet avn Ontnocrarny : Oscan Db 6 oe 0 oo Unite : Revation of tHe Anenanets 6. 4s Notes on Ontnoararny . ‘i ae Isrory oF THE Sounps: Vows 6 1 we oO ae... ‘i a 1 : 6 4 eee ee . 4 FoR 0 . : ee ror & eo = xi xvii ayer 28 ey 25 29 30 31 33 36 38 40 4viii Lable of Contents rar Dirntmonos . : al a see a8 ce ee ee ee ee ee os aueon ee Lexotiesina or Vowsis Oe at Suonrtenine or Vowrns © 6 eee Awartexis me Osea 2 8 Contraction any nats BB Vowst-Waanesino ix Meptan Sruvanues. 6... OD Srxcore im Memiay Sruvames 6. BT Suncore iw Fina Svuvamurs 6 ee OD SAMPRASARAMA Se rn) Loss or Fina Stont Vowsns 6 6 6 eee OL Vowsi-GRapation ee eee Consowants ee OG Conmonamran (I) 2 ee eB Consonantac u(y). eee es De Ce ee Onussion oF Nasats nerone Consonants... 70 PRO O fo 9 fo oo OO Bee ee ee eee a Inrernvocatic 8. Ruoracum. 5. 7s TH ina eee pooper ls et ce Intenvocauic 8. - eee 8 ramrvore Consonants. 6 1 ees TT Firanrs 6 6 ee ee TT ee a ee Po TB tee TB Po TD Deo TD Be PUnGeatayaMpPNAGana eee ee eee 0) bo 80 Fant 2 6 6 ee ee 0 Foon at.- Be 00Lable of Contents Unnrtan f, 18, From dd. Fiand. 6. 5 ferret Oe oe en Dewtac toe. ee Denta+ Dewan. 2 Ornen Commwations of Dextata —« we Gurrumais . ‘Unonian Panatanization or k seers eet eeceer Loss or k nurwenx Consonants. Unmniaw PALATALIZATION OF g Bho ee (lo 0 6 9 oo Go G Ce eee (5 6 09 56 6 5 oo Loss or Yin kur. 5 Cuaxce oF Surv Mutes to Sowants. Cuance ov Soxanr Murxs To SuRDS CHANGES OF THR ONIGINAL SonaNT AsriRATES Dountine oF Consonants 1x Oscan Sinrsarcation oF Dounte Consonants. Chaxors ix Sextence-Comnination. Sanpnt Accent ee ee Summany ov Tue Oscan axp Umnatan Sounns (Oeuar mma ee INFLECTION Nouns: PinnTsDRCReMmON |) eee eee Secoxn Decrenson 5. ffvSrmunb ee Oscaw Gentines i ile etc. 6 imp Decurnston see More Stems, Lignin Stems. Nasat Stems, S-Srems 0. Inneautan Nous. ee Pane 113 16 9 121 124 128 130 131x Table of Contents Fourtn Dacurnston . Firrn Decirnsion « Apsrctives: Decurnston 5 Comranison . 0. Avverms . - Nomenats: Carpimars axp ORDINALA Distaiputives axb Numenan Apvenns « ‘Puowouns : Personan Proxouxs Vossesatve, Pronouns, Demonstrative Pronouns Inrennocarive, Revative, ap Innere Broxoune TProxomnan, Excutics « Revarive Apverns ap Coxsuxctions ‘Verne: ‘Tur Persowan Expixas Exanries or Consucation Finst Consucation . Srcoxp Consucarion . ‘Tuan Consueation Fourtit Consucatiox . Inrecuian Venns FORMATION OF THE MOODS ‘Tar, Present Stem: First Consucation Secowp Consuaation Tmp Coxsucation Fourtn Coxsucation Forms or tar Tree or 1. ca} Inexourar Verns. Remanxs on THe Forms conxneten witn L. habed rere wirn L. facio REMARKS ON THR FoRMs con! Tue Iureurect Inpicative Tur Futune Inpicative ‘Tu Pensxer Invicative Tue Furune Penrvct. ‘Tre Sunsunctive . Tur Present Sunsuncrive, TH ‘Tne Penreor Sunsunctive Inrerrecr Sunmmctive, PAOR 131 132 198 134 136 187 139 130 40 140 3 ua 148 151 163 164 166, 167 109 160 161 162 163 164 165 166 167 168 169 169 169 173, 173, 14 176 176Table of Contents Tue Invarative Tue Passive. Tur, Prnirnnastio Passive, 5.2 ‘Tue, Present Inrivirive, Tar Surin ‘Te Present Active Panttorrus . 6. Tue Penexcr Passrve Panricirte, : Tus Genunpive 5 ‘WORD-FORMATION Derivation of Nouns ann Anserives Nouns : Avnwotives 6 eee Ae Srconpane Vinny, Dentvation: Drxowinatives « ee Comrostrion : Nouns ap Ansectives : Venns 5 6 6 6 SYNTAX Uses oF tim Canna: ‘Tor Gritive « 6b oo -5 6 ‘THe Darien, 0 oc ‘THe Accusative 6 6 5 Tun Locative 5 9 Tre Aviative(-Inernumenta): Apuarive Uses. 4 6 0 Instrumpntan Uses. ee Locative Uses. oo Purrosrtions (AND THe Connesroxpixe Prerixes): Winn tie Accusative oxur ‘Wirn te Apuative oxet ‘Wir tHe Accusative axp Locative . Witn tie Locative ax AnLative. Witn Oren Cases Apsrcrives Apvorns ‘Tae Vern: Voice « Tense, Moon: Commanns AxD Prourarrioxs Tur Sunsuxctive or Wien. Tue Sussuxcrive w Suastawnive Cravess Crauses or Invinect Quzstionxii Table of Contents Rewarive Crane, ‘Temrorat. Cuausen . Conprriowan CLauaee Iwrinmives any Parrionrers 6. Aommert Omisnion or Worm... ee ‘Onpen oF Wore. COLLECTION OF INSCRIPTIONS ‘Oncan Tan Crreon An ‘Tur Tamura Bawtina Intcrirnions o” Pours InsetpTions ox Puntac Worns, ann Deptcations Tue Errons Inecrirrions 2s Insonterions oF Carva ‘Tur Curse or Vina. Tw, Tovitar-Depications - : Omer Carvar Inscriptions 5. Inecetrrions rnow Orurn Casrantan Towns « Inncntrrions or Sassi ano tity, FRENTAXI: Tur Deoicatony TAnuet or AcNoxe Onsen Inscnirrions or Lucaxta, Buurrium, ann Mussaxs CL mlol dl o to lu tol ble Ummruae ‘Tue Touviian Tamura: VEVIlandt. 5. we Minon Unnntax Intentions 5+ GLOSSARY AND INDEX Onan ‘i Unnaun . 5 Pnoroonarns or Oscan Inscnirtions Facsimiue oF scan Insonirtion rnom Pourett Facsiwine of tue Tanuia BANTINA 2. Puorooarn or Vb, Iouvintax Tantes Mar or Cenznau Fray . 8m . 2B . 20 art . a . ma 223, 2 BB . 26 a) - 230 - me + 23 UT cro! + 258 . tt 256 + 268 . 8 260 202 2m, 207 01 310 - 2a 32 Prare Puare I Prare IIL Puars IV Pate V| To> NSTITV BRIEF BIBLIOGRAPHY? Ze The history of the study of the Italie dialects might be exped&t to date from the discovery of the Iguvinian Tables in 1444, but Yor several centuries all the attempts to decipher these were wholly worthless. ‘The first sign of progress is found in Tanzi, Saggio di lingua Etrusca © di altre antiche d'Italia, Rome, 1789, in which the ritual character of the contents was recognized. In the first half of tho nineteonth century fall, among others, the contributions of K. 0. Muller, who in his great work on the Etruscans (Die Etrusker, 1828; 2d ed. by Deecke, 1877) definitely disposed of the error that Oscan and Umbrian were connected with Etruscan; of the Sanskritist Lassen, who gave a critical treatment of a section of the Igavinian Tables in his Reitrige zur Dentung der eugu- binischen Tafeln, Bonn, 1833; of Grotefend, celebrated for his decipherment of the Old Persian cuneiform, who treats selected pas- sages in his Rudimenta linguac Umbrieao, Hanover, 1835-1839; of Lepsius, tho futuro Egyptologist, who in his dissertation, De tabulis Eugubinis, Berlin, 1833, cleared up the remaining difficulties of the alphabet and proposed a chronological arrangement of the tables which is still followed in tho universally adopted numbering. Lepsius also brought out the first trustworthy edition of the Oscan inscriptions together with the Umbrian, the Inscriptions Umbricae et Oscae, Leipzig, 1841. ‘A work of prime importance for the study of Oscan and tho minor dialects was Momumsen’s Unteritalische Dialekte, Leipzig, 1850. A similarly fundamental work for Umbrian was Aufrecht and Kirchhoff's Dio umbrischen Sprachdenkmiler, 1849-1851, the first really critical attempt to interpret the Tguvinian Tables as a whole. Kirchhoff was also tho first to recognize the true character of the longest Oscan inscription, the Tabula Bantina, in his claborate commentary, Das Stailtrecht von Bantia, Berlin, 1853. In Huschke's Die oskischen und sabellischon Sprachdenkmiler, 1856, and Die igu- vischen Tafeln, 1859, a wealth of knowledge on the side of antiquities 1A fall bibliography is glvon by ¥. Interpretation of tho Txuvinlan Tables, ac on xtit Santa, Hy pp. xl ally Bréal,xiv Brief Bibliography is marred by a lack of critical judgment, especially in grammatical points, so that while some of the many daring conjectures have proved serviceable, his works in general mark a stop backward. Newman's Text of the Iguvine Inscriptions, London, 1864, is without much value. Grammatical questions were also discussed in numer- fous articles by Corssen, Kbel, Bugge, and others. Rruppacher's Oskische Lautlehro, 1869, and Enderis! Oskische Formentehre, 1871, wore convenient little manuals for the time, though valueless to-day. In the last quarter of the nineteenth century the most notablo advance in the interpretation of the dialect remains was made hy the works of Bréal and of Biicheler. Tesides their exhaustive com mentaries on the Iguvinian Tables, cited below, each of these scholars has discussed in one form or another most of the more important. Oscan inscriptions. Important contributions were also male by Bugge, Danielsson, Deecke, Jordan, Pauli, and others. New editions of the Oscan and Sabellian inscriptions with facsimiles were brought out by the Russian scholar Zvetaieff in 1878 and 1884 (cited below). ‘The ltalic dialects have always held an important place in the interest of Indo-European philologists, and Brugniann especially has done much to further their study, both as anthor and teacher. It is not too much to say that the works of former pupils of his, appearing from 1892 on, especially the treatises of Broniseh and the present writer, von Planta’s grammar, and Conway's edition of the texts, all cited below, have put the whole subject on a new footing. Contributions on special points, too numerous to specify here, have been made in recent years by I. D. Allen, Barthoiomae, Ce Ehlich, Fay, Horton-Smith, Pascal, Skutsch, Solmsen, and others. The following is a list of the works which aro now the most useful to the student. Indo wuropean Grammar Brvemann-Detnnicx, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. 5 vols. Strassburg, 1886-1900. Vol. I in 2d ed., 1897. Vols. I-If (Phonology and Morphology) by K. Brugmann (abbr. Brugmann, Grd.) ; vols. ITI-V (Syntax) by B. Delbriick (= Delbriick, Vergl. Syntax, I-III). ‘The Oscan and Umbrian dialects aro treated aystomatically and as fully 1s the wide scope of the work permits,Brief Bibliography xv Bruamann, Kurze vergleichende Grammatik der indogermanischen Sprachen, Parts I=L, Strassburg, 1902-1903. In this shorter work, to be completed within the limits of a single volume, Osean and Umbrian forms are mentioned only incidentally in connection with the treatment of Latin, Latin Grammar Linpsay, The Hatin Language. Oxford, 1894 Sommer, Jandinch der Jateinischen Laut- und Formenlehre. Heidelberg, 1902. Srouz, Historische Grammatik der lateinischen Sprache. Leipzig, 1894, Oscantmbrian Grammar von Puanta, Grammatik der oskisch-umbrischen Dialekte. 2 vols. Strassburg, 1892-1897 (abbr. v. Planta), ‘A sound and exhaustive treatmen Also contains the texts. brief skolch of Osean-Umbrian graminar in inluted in Conway's Italie Dialects, and of Umbrian grammar in the commentaries of réal, Ticheler, and other, quoted Uelow. Special chapters of the grammar are treated in: Bronrscnt, Die oskischen Ie und E-Vocale. Leipzig, 1892 Bucx, Der Vocalismus der oskischen Sprache. Leipzig, 1892 (abbr. Osk. Voe.). Buek, Tho Oscan-Umbrian Ver Verb-System). fundamental for all future work. ystem. Chicago, 1895 (abbr. Texts and Commentaries* Conway, The Malic Dialects. 2 vols. Cambridge, 1897, The most exhanstive collection of the material, containing the inscriptions swith full epigeaphienl data, the glosses, lists of propor names, ete, together with a brief sketch of the grammar, and m glossary ‘A concise but complete collection of the inscriptions is also included in ¥. Planta’s Grammatik, cited above, Conway, DialectorumIalicarum ExemplaSelecta. Cambridge, 1809. Selections from the dialect inscriptions, it translation and brief notes Avrencne uxn Kinennorr, Die umbrischen Sprachdenkinaler 2 vols. Berlin, 1849-1851. See ahore, p xii, Still o be consulted with profit, Hoferonces for parxvi Brief Bibliography Brat, Les Tables Eugubines. aris, 1875 (abbr. Tab. Eng.) ‘Thia and the following aro the two leading commentaries on the I Tables. Bicnrure, Umbrica, Bonn, 1883. On the whole the moat convincing interpretation of the Umbrian remains, and followed in large measure in the present work. Mounsex, Dio Unicritalischen Dialckte. Leipzig, 180 (abbr. Unterit, Dial,). See above, p. xi, Still valuable for the epigraphical data and the goo- ‘graphical and historical notes. Zverarnrr, Sylloge Inscriptionum Oscarum. St. Petersburg and Leipzig, 1878. Zveratney, Inscriptiones Italiao Media ‘There two collections aro now mainly valuable on account of tho accom- ‘panying facsimiles. Contributions on special points of grammar and interpretation aro found in the various journals, procsedings of learned societies, and tories of stndies, eapecially the following. American Journal of Philology (abbr. Am. J. of Ph.). Beitrige zur Kunde der indogermanischen Sprachen. Fa. by A. Bezzenberger (abbr. 3.1, = Bezzenbergors Beitrage). Berichte uber die Verhandlungen der kéniglichen sichsischon Gesoll- schaft der Wissenschaft zu Leipzig. Philologisch-historische Classe (abbr. Ber. d. siichs. Gesell. d. Wiss.). Classical Review. Indogermanische Forschungen. Zeitschrift fiir indogermanische Sprach- und Altertumnskunde (abbr. 1.¥.), with the Anzeiger for indogermanische Sprach- und Altertumskunde (abbr. 1.F. Anz). ‘Mémoire de la Société de Linguistique de Varis (abbr. Mém. Soc. Ling.). Pasily Xiialiache Studion, 6 vola ‘Hanover, 1883-1887. Rheinisches Museum fr Philologie (abbr. Ih. M.)- Zeitschrift far vergleichende Sprachforschung auf dem Gebiete der \dogermanischen Sprachen. Founded by A. Kulin (abbr. K.Z. = Kuhns Zeitschrift).EXPLANATIONS Black type is used to transcribe words in the native alphabets, and italics for those in the Latin alphabet. ‘The same distinction is commonly employed for separate letters or groups of letters. Tut sometimes, to save unnecessary repetition, italics are used referring to the spelling of Loth the native and Latin alphabets. Glosses cited aro always indicated as such, except the frequently cited fume ‘The meanings of words cited are usually given, though not always, especially where they can easily be inferred from the Latin cognates cited. Vico versa, Latin cognates are sometimes left to be inferred from the Latin translations. In the case of words of doubtful meaning these translations in the grammar are to be regarded as expedients, subject to amplification or correction in the glossary. In tho texts uncertain letters are distinguished by a change of type, and where obvious mistakes are corrected the original reading is given in footnotes to the text. But in the gram- mar proper such matters are usually ignored except when bearing directly on the subject of discussion, ‘The signs j and are used for consonantal i and x, English yand 1; ym) gs 4, for the syllabic nasals and liquids assumel in Indo-European forms. The colon (:) is used to point out retation. ship, in the sense of “cognate with.” Besides the abbreviations of ‘works of reference mentioned above, the following are used. o.Bulg . O.F ng Old Bulgarian. Old English, ~ OM Irish, * Old Prussian, |. = Paelignian, ‘Sab, = Sabine. Skt. = Sanakrit. 'T.A. = Tablet of Agmono (no. 45). ‘T.B. = Tabula Bantina (a0. 2.) ‘Lithuanian, U, = Umbrian, Marrucinian, ‘Vest, = Vertinian. ©. = Onean,INSTITS OSCAN AND UMBRIAN GRAMMAJ INTRODUCTION Prorturs axp Laxauaars oF Tracy 1, The Italian peninsula, in the earliest period of history, was occupied by various peoples speaking a variety of languages and dialects. ‘The Ligurtans in the northwest have usually been regarded as relics of an aboriginal, pre-Indo-European, population, but are now thought by some to be IndeEuropean.! The linguistic remains, consisting largely of geographical names, are too meagre to be decisive. ‘The Wtruscans (Latin Htriiact or Tnaci, the Intter from *Dursct; cf. Umbrian Turskum, Greek Tuponvot, Tupprvot) occu- pied Etruria, and, previons to the Celtic invasions, much of the central part of northern Italy, in the valley of the Po. They were also masters of Campania from the eighth century B.c. down to the Samnito invasion in tho last quarter of the fifth contury n.c. The Etruscan inscriptions? number over six thou- aand, but only a few hundred contain anything more than proper names, and less than a dozen of these aro of any considerable Iength, ‘The interpretation is wholly uncertain and nothing positive can be affirmed as to the alfinities of the language Rut it is reasonably clear that it is not Indo-European. The riddle will probably romain unsolved until the discovery of on of some length. CE Kretaehimer, KZ. 8% 104M jo beg enlleted I in Corpaun Busct i FtrancaranIntroduction tt The Veneti, at-the head of the Adriatic, and the Messa- pians and Lapygtans in Calabria have commonly been grouped logether as of Ilyrian origin. ‘There are several hundred shor Venetian inscriptions,' and the Messapian is also represented by some hundred and sixty short inscriptions? From those remains it appears that the two languages, though Indo-European, do not belong to the same group, and it is uncertain whethe the Venetian, or the Messapian with the modern Alba should be classed as Ilyrian? Greek colonies occupied nearly the entire southern portion of Italy, many of them dating from a period earlier than tho Deginnings of Roman history and retaining their Greek char- acter for several centuries after Christ. Celtic tribes which poured in from the north, and in tho carly part of the fourth contury n.c. sacked Rome, maintained themselves for somo time in the central plains of northern Italy. The rest of Italy was occupied hy tribes speaking dialects akin to the Latin and with it constituting (he Italle branch of the Indo-European family. Cuassirication or THR IraAnic Dratrors 2. The Italic Dialects fall into two groups, the Latin Falisean and the Osean-Umbrian, ‘The Latin-Faliscan comprises the Latin, of which thero were local variations in the different towns of Latium, and the Fall spoken in the Faliscan plain in the southeastern part of Etruria, ‘The few short inscriptions‘ are sufficient to show that Faliscan differed but slightly from Latin. The Oscan-Umbrian group is so named from its two most important members, the Oxcan and the Umbrian, but includes 40m the Hllyrian question, ef. Paull, 1oe.; Kretechmer, Binfoltung tn lo rlechischn Sprachgeschichte, 24 (T.; Hirt, Featschrif ir Kiopert, HL ff; Pedersen, Ki, 20 ‘Collextod in Deecke, Dia Falisker.oa i CAAA ee Os s : ~ = me me 3] Osean — External Data 3 also the dialects of most of the minor tribes of central Italy, ! which may he conveniently designated as Sabettinn.! The best known of theso is tho Paelignian, which shows a very close resemblance to Osean. Much the same aro the dialects of tho neighboring Marrueinians and Vestinians, of which there are some scanty remains. Volsclan, known only from an inscrip- tion of Sour lines from Velitrac, is more strongly differentiated and in several particulars resembles Umbrian more than Oscan ; but thero is no sufficient reason for grouping it otherwise than among the Sabellian dialects. The Marstans, Acquians, and Sabines aro connected historically with the other Sabellian tribes, and their dialects doubtiess belong properly to the samo group. But they were subjected to Latin influence from a very early period, and the meagre remains that we have givo no satisfactory picture of their characteristics, Oscan — Extenwas Data 3. Oscan inscriptions have been found in Samnium (inclu- sivo of the territory of the Frentani and Hirpini), Campania, northern Apulia, Lucania, and Bruttium, and in the Sicilian city of Messana from the period after its occupation by the Campanian Mamertines. ‘Theso are precisely the rogions which we know were occupied by Samnito tribes. In calling the lan- guage Osonn rather than Samnitic we are following the usage of the Latin authors, as when Livy (10, 20) relates how in one ‘*Safinai), and Samnivon (from tradition of the Rabine origin of the Samnites and the minor tries Ike 1 witness tothe tribal relations af these pooples. The Renan writer ‘a generic term inching Samitie. Subrilian, andl their lanenage, the Osean. a Sahel iit language occupy seh a preéeinent peaition Pacligni, Shellie peaking the Sarita teibea were Tan sileet. Tut tho Saranites and namo Sahelian for the cloely related minor tribee nul dialects ‘Tho so-called OM Sabelllan inseriptions,foune in varkous parte of eqnteal Tey, fro wholly wnintlligie, and eartanty aee not tn any af the Sabe } possibly represent the Innguage of some IIlyrian tees,4 Introduction B of the Samnite wars the Roman consul sont out spies who were acquainted with the Oxean language. Now the Oscans (Lat. Orci, cartier Upsci, Grk, ‘Orveaé) wero 4 Campanian tribe, and it has been held by some that Oscan was not the original lan- guage of the Samnites, but was adopted by them after their invasion of Campania, But this is altogether improbable. Wo must, rather, asstime that the Oscans were simply a dotached branch of the Samnites, speaking essentially the samo language ; and the principal reason why this language was called Oscan rather than Samnitic is Uhat it was among the Oscans that the Grecks and Romans first came in contact with it, ‘The Sam- nites enterod the field of history as a politically distinct. people from the Oscans ; but their linguage, being the samo, was called hy tho namo already established. Moreover it was among the Oseans, hy reason of their carly contact with Greck and Htrus- can civilization, that the language was first reduced lo writing, s0 that while they did not give tho Samnites a now language, they did give them its written form, and to a certain extent, probably, sort of normalized standard of specch. ‘This last supposition would help account for the fact that local variations of Osean, outside of Campania, aro far less marked than one would expect, considering the extent of the territory in which the language was spoken. t 4. The Oscan inscriptions number over two hundred, Int more than half of these contain only proper names or fragments = of words. About three quarters of them come from Campania, where Pompeii, and in recent years Capua, have furnished the greatest number. ‘The period of time covered is nearly five centuries, the earliest remains being some coin-legends from the ond of the fifth or first half of the fourth century 1.c., while the latest are some of the graffiti of Pompeii, which there is reason to bel swore scratched on the walls after the first earthynake in G3 A.D. But by far the greater part of the material falls between 300 1.0. and the Social War in 90-89 n.c, After the Social War Oscan Crirrrerec@gwaa4 Orcan— External Duta 5 ceasod to be used in official documents, but continued to exist ) as a local patois for some time, — how long we ab Pompeii it was slill spoken, to some extent at least first contury A.b. it very likely lingered on for soveral venturies in the remoter districts of Samnium. Most of the inscriptions aro written in the native Osean alphabet, which is derived, through the medinm of the I from the Greck of the Chaleidian type. Buta few from Luca including the longest Osean inscription known, the ‘T Bantina, are in the Latin alphabet, and some from Sicily and various parts of southorn Italy are in the Creek alphabet. 5. As regards contents, many wellknown classes of in- scriptions aro represented. abula Bantina, the longest inscription, itself only a fragment of tho original, contains a sories of ammnicipal regulations. ‘The next longest, the Cippus Alvllanus, is an agreement between the cities of Nola and Alolla touching corlain temple property held in common. From Agnone in Samninm comes an inventory of statues and altars in’ a sacred grove. ‘The Curso of Vibia, from Capua, together with a few shorter curse of devotiones of which there aro many examples among Greck and Latin inscriptions. ‘There are several inseriptions on public works from Pompeii and elsewhere ; also dedications, including a peculiar series of iovitac-ledications, mostly from Capua, the nature of which is not fully understood. Cortain inscriptions painted on house-fronts near some of the streetcorners in P p in tho Social War. ‘There are numerons inscribed. coins from belongs to the secm to be guides for the allied troaps occupying the city various towns, some of them older than any of the inscriptions ‘on stone; also soveral from the time af the Social War, bearing the legond Vitetié ‘Malin’, and the names of the leaders of the allies. ‘There aro a fow epitaphs, many bricks inscribed with names, and probably one of the well-known inseribed missiles also some illegible clectionecring notices, not to mention various other insignificant serawls, on walls in Pompeii. Zu BE Eem —6 Introduction Is 6. Besides the inscriptions, thero aro some secondary sources, such as the Oscan glosses, mostly in Varro and Festus, and the gengraphical and personal names frem Osean tervitory. But they contribute relatively little to our knowledge of the dialect. 7. Oscan was not a mere patois, nor was it so regarded by the earlier Roman writers. Ennis, in boasting of having threo sottls because he could speak Greck, Oscan, and Latin, gavo to Oscan a position which he had no thought of giving to the local vernacular of his home, the Messapian, For a long time, whilo Latin was still confined to Latium and its immediate borders, Osean was spoken over a vastly wider territory. It was tho language of the people which gave the Romins the hardest fight for the hegemony of Italy. In the carly centuries the Oscans of Campania, under the Etruscan rule, and close to the Greck colonies of Cumae, Naples, ete stood on fully as high a plane of civilization as the Romans of the same period. Eminent scholars like Mommsen have expressed the conviction that there once existed an Oscan literature, and certainly the conditions for the rise of a native literature were as favorable as at Rome. But nothing has come down to us, not even a reference to anything more pretentious than the puppetshows introduced in Rome from Campania under the name of fabulae Atellanae or ludi Osci. At Rome, of course, these were no longer given in Osean, but in rustic Latin. ‘Unnrtan — Exrenwat, Data 8. Aside from a few short inscriptions from various towns of Umbria, the Umbrian remains consist of the Jiuvinian ‘Talles, discovered at Gubbio, the ancient Iguvium, in the fif- teenth century. These are seven small bronze tablets (originally nine, but two were lost soon after the discovery), most of them scribed on both sides, and containing together between four and five thousand words. This makes a far more extensive docu- inent than any representing any other dialect except Latiuy General Characteristice of Oscan-Umbrian 7 9, Some of tho tables aro written in the native Umbrian” alphabet, which like tho Oscan is derived from the Greck through the Etruscan, others in the Latin alphabet. These two divisions of the material are conveniently distinguished 88 Old Umbrian and New Umbrian, but the differences are in part merely orthographic, and, at most, far less marked than those which are usually associated with the terms Old and New in such a connection, ‘The Now Umbrian tables may date from tho eatly part of the first century n.c, How much earlier the Old Umbrian tables are it is impossible to say; different parts were inscribed at different times, and oven the relative order is not fully determined. Sce the Commentary on the Iguvinian Tables. 10. The contents of the Tables consist of the acts of a cortain corporation of priests known as the Atiedian Brothers, and in their gencral character resemble tho Roman Acta Arva- ium. They contain directions for various ecoremonies, such as the Purification of tho Sacred Mount and the Lustration of the People, as woll as the more private functions of the brotherhood, with minuto prescriptions as to the taking of auspices, manner of sacrificing the victims, ete.; also statements as to the duties of certain officials, perquisites of the priests, contributions to bbe mado to the brotherhood hy certain gentes, ctc. Some of the older tables contain matter which is repeated in an expanded form in tho later tables. Genvran Cuanacrenistics or tig Oxcax-Unmnian Grour Phonolory 11, The most. striking characteristics, as regards pho- nology, are: ; ‘Chango of the labiovelars g and g¥, which appear in Latin as qu and v (gu after 1), to the labials p and 8; e.g. O. pia squis’, U. pixi, Volsc. pia, Marruc. nipis ;—O. bivus ‘vivi'; U, benuat ‘vonerit’ ; —U. umen (from *wmben) ‘unguen’.8 Introduction [ur Extensive syncope of short vowcls in non-initial syllables ; eg: 0. actud sagito’;—U. fktu Hfigito’;—O. hire ‘hortus’; — U. tkuvins *Iguvinus'; — 0. alkatus ‘ad vocati’. Assimilation of nd to m5 —U. pikaner tpiandi’ (n for 1m, 26). Retention of a before nasals and liquids, where it is lost in Latin; e.g. O. fisnam ‘fanum’, U. feanat-e, Pacl. fesn. ; — O. kerona *eona’, U. acana ;— Pack. prism ‘pri Retention of a in medial syllables, where it is weakened in Latin to ¢ or i; c.g. O. Anterstatal *Interstitae’; — U. antares ‘integris"; — U. procannrent «*procinuerint’. Representation of original dk and dh by f, not only initially as in Latin, but also medially, where Latin has 6 or di og. O. thei, U. tefe ili’; —O. mefaf tin medin’; — U. rufra ‘rubros’ Change of final d, which in Latin is shortencd, in the direction of 3; e.g. O. molto, U. mutu, muta ‘multa’. Change of Kt to ht, and of pt to fl (Umbrian, furthor, to At); eg. O. Tntavis ‘Octavius';— U. rehte trecte’; — 0. acriftas ‘scriptac’, U. serehto. Assimilation of ks to a8 #3 eg. O. destrat ‘doxtra est’, U. deatrame, Change of ns to f, though under different conditions in Osean and Umbrian; e.g. ©. dittivf ‘usus’ from *oi U. Ace. Pl. eaf ‘eas’ (also Marruc. iaf-c) from *eans (but O. vias, Os dpsannam ‘operandam’; Inflection 12, Decrenston. The types of noun-declension are auf ficiently like the Latin to fall naturally into the same grouping of Five Declensions. But the Fifth Declension is represented by only a few forms, and in the Third Declension the consonant- stems and éstems are kept distinct in a greater number of ense- forms than in Latin. ‘The Cases are the same as in Latin, except that, in the Singular, the Localive exists as a distinct form with full syntactical functions. ‘The important differences in case- formation are as follows (for examples, see the paradigms13] General Characteristics of Oscan-Umbrian 9 First Declension. The Gen. Sg. has the original ending -a4, which is preserved in Latin only in phrases like pater familits the Nom. Pl. has the original ending -, which is lost in Latin. Second Declension. ‘The Gen. Sg. has the ending -cia, from istoms ; the Dat. Sg. has the ending -0i, which ocers in Latin only in Numasioi of the Pracnestine brooch; the Nom. PI. has the original noun-ending -d3 for both nouns and pro- nouns, while the Latin has %, from -2i, the pronominal ending + the Gen. Pl. has only the original -dm (I..-1m), there being nothing to correspond to L. -drum, which is a specifically Latin development. ‘Third Declension. ‘The Gon. Sg. always has -eis, the ending of éstoms, while Latin -is is tho proper ending of consonant- stems ; tho Acc. Sg. of consonant-stems hins -om, from o-stems ; in the Nom. PL. the consonant-stems and i-stems are kept distinct, tho former having the original ending -es with syn- cope of the e, the latter -é as in Latin (O. sumuns shomines’, but tris tees"). 13, ConsuGAtion. The conjugation-types are the same n Latin, the material grouping itself under the Four Conju- gations, leaving the relies of unthematic inflection as “Trregular Verbs.” But the type represented by Latin eapis is, in origin, more closely connected with the Fourth Conjugation than with the Third, and in Oscan-Umbrian is better grouped with the Fourth. ‘The Moods arc the same. As in Latin, the Subjunetive is fusion of original Subjunctive and Optative forms, and the distribution of the forms is the same as in Latin, except in the Perfect Subjunctive (see below). ‘Tho ‘Tenses are the same, except that, perhaps accidentally, there is no example of a Pluperfect. ‘The Voices aro the same, but of the Passive there are only forms of the Third Singular and ‘Third Plural. Of the non-finite forms there are found a Present Active Panticiple, Perfect Passive Participle, Gerundive, Present Active as10 Introduction [is Infinitive, Perfect Passive Infinitive, and Supine. The Ger- und, Perfect Infinitive Active, Future Infinitives, P Infin- itive Passive, and Future Active Participle are lacking. ‘The absence of examples of some of these forms is possibly a mere accident, but it is probable that most of them are specifically Latin formations. ‘The important differences in formation are as follows: The Pres. Infin, Act ends in -om5 4g. Ox eztent, U.crom esse’. ‘The Future is an «formation, in origin a short-vowel Sub- junetive of an #Aorist; e.g. O. deivast siurabit, U. ferest ‘foret’. The Fut. Perf. is an e-formation, probably based on an old Perf. Act. Partic. in -us combined with a short-vowel Sub- junetive of the vorb *to be’; e.g. O. dicnat sdixerit’, U. benvat “tvenerit’, Among the different formations making up the Perfeot System, the f-Perfcot is characteristic of Oscan-Umbrian ; eg. dated ‘docrovit’, U, andiraafuat ‘circumtulerit’. (Oscan- erfeot, and Umbrian an LPerfect and Perfect.) ‘The Latin vi- and sPerfects are lacking. The Perf. Subj. is a real Subjunetive form with the mood- sign é not an Optative with mood-sign i as in Latin ; eg. O. tr barakattins ‘acdificaverint’, U. combifiandi ‘nuntiaverit’. In the Third Singular and ‘Third Plural there is a distine- tion between primary endings, which are -,-nt, and secondary endings, which are -d (lost in Umbrian), -ne; c.g. O. faamat shabitat’, but faidiad +facint'; — O. stahint stant’, U. furfant spurgant’, but O. deicans ‘dicant’, U. dirsane ‘dent’. Latin shows -d in some of the earliest inscriptions, but nothing cor- responding to -ns. ‘The unthematic form of the Third Plural, -ent, which in Latin is always replaced by the thematic form -ont, unt, is pre- served, and even extended to thematic formations; eg: O. eet, U. sent ‘sunt’, O. censazet ‘eensebunt’. ‘The Third Singular and Third Plural of the Passive have an ending -ter, unknown in Latin, while the Latin -tur appears1) General Characteristicn of OscanUmbrian 1 only in Umbrian secondary tenses ; e.g. 0. vincter ‘convincitur’, karanter ‘vescuntur’, U. herter ‘oportet!; U. emantur ‘aceipiantur’. ‘The Third Singular Passive has also a peculiar act of forms in which the ending is neither -ter nor -tur, but simply -r 5 eg. U. ferar ‘feratur’, O. sakrafir (Perf. Subj.) «sacrato’. ‘The Imperative Passive has an ending -ma(d), O. -mé-r, which is of similar origin to the early Latin -mind ; e.g. O. cen samur ‘consetor’, U. peranihimn ‘precator’. Syntax 14, Tho Syntax shows a remarkably close resemblance to the Latin, ‘There are no uses of the moods and tenses which cannot be paralleled in the Latin, the agreement being closest, in some respects, with early Latin prose. ‘The Passive forms nclude both genuine Passives and Deponents, as in Latin, but the frequent impersonal use is characteristic of Oscan-Umbrian, In the use of the cases there are many interesting constructions, of which the following are the most noteworthy. The Locative, ~ Doing preserved as a distinct ense-form, is used where the Latin requires in with the Ablative, e.g. O. else terefin eo territorio’. ‘The Partitive Genitive has a wider scope than in Latin, eg. U. iuenga peracrio tursituto siuvencas ex opimis fuganto’. A Genitive of Time is seen in O. zicolom XXX nesimum ‘in diebus XXX proximis’. ‘The Genitive is uscd more frecly than in Latin to denote the matter involved; e.g. O. eizazune egmazuns sin these matters’, U. puai ocrer pihaner ‘a8 in the case of the purification ‘of the mount’. ‘The prepositions corresponding to Latin inter and tran are used with both Acousative and J. those corresponding to ob and post aro used with the Ablative. Vocabulary * 15, Of words which are characteristic of Osean-Umbrian as compared with Latin, the following are the most important examples : 1 Special attention In given here tou arin lace tenn are ok, Tike the other charactetatlen, tho eubject of fuller treatment in the gramme proper12 Introduction [1s 1. her-+velle’. O. herest ‘volet’, herlam ‘arbitrium, vim’, Herenta- tele ‘Voneris’ (acl. Herentas); U. nerisvult', herieat volet', elc., herter ‘oportel’, herie ‘vel’, pis-her ‘quilibet’. Cf. L. horior, hortor, Grk. xaipa, Skt. huérydmi *be gratified, delight in’, Goth, sgairns ‘eager’, Eng. yearn. ‘This root completely nto yel- (L. wold) in the meaning *wish’, Uhe latter appearing o inaspecialized meaning ; e.g. U. veltu'deligito’, hella sinboto! 2. toutd- ‘civitas, urbs, populus’. O. rw¢ro Maweprvo ‘civitas Mamertina’, toutad pracsentid ‘populo pracsente’, tonticom ‘publicum’, ete.; U. totam Jiowinam ‘civitatem Iguvinam’, tuderor totcor ‘fines urbiei’, etc.; Marrue. toutai Maroweai ‘civitati Marrucinae’; Volse. toticu spublico’. Cf. Lith, tauta ‘people’, O.Pruss. tauto country’, O.Ir. tuath ‘people’, Goth. piuda ‘people’, O.Eng. péod speople, nation’, ete. 3. ais-‘sacer, divinus’. O. aisusie ‘sacrificiis’; U. esona ‘sacra’, crificium’; Marruc. aisos‘dis'(?); Mars, esos +di Volse. craristrom sacrificium'; aizot~ Geot tnd ‘Tuppyvav (Hesychius), aesar Btrusca lingua deus (Suetonius). Per- haps related to Germ. Bhre (Goth. *aiza), and lo Goth. aixtan srevere’, L.. aestimd, from aiz~ 4. komno- ‘comitium’. ©. comono tcomitia’; U. super kumne ‘super comitio’, kumnahile‘in conventu’. From kom ‘cum’ tsullix -no- (ef. L. pro-nus, trdns-trum). 5. hontro- tinferus’. ©. hufajtruis ‘inferis'; U. hondra tinfra’, Superl. hondomusinfimo’. From hom. related to L. humus, Grk. xapat, xOdv, ete. Vor meaning of. L. humilis, Grk. xOaparss, Lith. Zmas ‘low’, demgn down’, from zémé ‘earth’, 6. medes-tius’, U. meta, mere tins’, merato siustum’, mersuva siusta’; 0. meddies ‘meddix’, officin) title (ef. Festus *meddix apud Oscos nomen magistratus est”; Livy 26, 6, 13 “medix tuticus [O. meda see above, 2} qui summus magis- tratus apud Campanos est”; cpd. like 1. tive from *iia-dik-), medicim‘magistracy’,meddikiai tin the meddixship’; medicatinom siudicationem’, medicatud ‘iudicato’; Pacl., Volse., medix(Nom. PL); Mars, media, Cf, 1, modus, modes-tus, Grk. péBopar, ele. esono | |15) General Characteristics of Oscan-Umbrian 13 T. nor- ‘vir, princeps’, title of rank. O. nerwm (Gen. PI.) nite (Nom. Sg.);. U. nerf (Ace. PL), nerve (Dat. VL). For related Sabine forms ef. Suetonius Tib. 1 “inter cognomina autem et Neronis adsumpsit, quo significatur lingua Sabina fortis strenuus”; Aul. Gellius 13, 23 “id autem, sive Nerio sive Nericnes est; Sabinum verbum est, eoquo significatur virtus ct fortitude”; Lydus de Mens. 4, 42 “nepien yap % avdpia. éori nat vépavas Tobs dvBpeious of EaBivor wadoiow". Cf. Grk, davip, Skt. nar“man’, O.ln, nert ‘strength. 8. nessimo-‘proximus'. O. nessimas (Nom. Pl.), nesimum (Gen. Pl), nesimois (Abi. Pl); U. nesimei ‘proxime’ (alv.). Cf O.lr. nessam ‘nearest’, cto. Cf. also O. nistrus ‘propinquos’. 9. pert trans’. Q, pert viam trans viam’, am-pert not more than, dumlaxat?; U. pert spiniam ‘trans columnam’(?). An extension of per, Umbrian also uses traf = L. trans. 10. postin ‘according to. O. pistin slagim ‘according to the territory's U. pusti kastruvuf ‘per capita’@) ete. An exten- sion of *posti (early Latin poste), A. pir U. pir ignis', pure-to ‘ab igne’; O. purasiat tin igniaria'. Cf. Grk. wip, mupds, O.1LG. fuir, four, Eng. fire, otc. 12. scyo-‘totus’. 0, sivom ‘ommino’; U. sewom ‘totum’, sev-akne ‘sollenme’. Cf. L. aé-lue, Goth. aé-la(?). cfro-Iumntoffering’. ©. saahtim tefirim ‘sacred burnt- offering’; U. tefra ‘enrnes cromandas’, tefrw-to ‘ox rogo’. Probably from *teparo-, related to 1.. tepor, Skt. tapas, ete. 1d. treb-thabitarc’. U. trebeit «versatur’, tremnu ‘tabernacul 0. teffbim ‘domum’,, tribaralekiuf ‘aedificium’, triberakavim ‘aedificare’, ete. Cf. O.Ir, treb tdwelling-place’, Lith. troblt “building’, Goth. patrp ‘field’, Germ. Dorf, ete. 15. wero-, g, is present In Oscan, but not in Umbrian, is sometimes explained by the supposition that the Oscan alphabet. was borrowod earlier than the Umbrian, Mut at all periods Btruscan possessed both characters, > and ¥, used as by-forms for the suri. Umbrian took only %, possibly ecause thi was proferred in the local typo from which it waa derived. Oscan took over both charactera and differentiated them again. ‘That in this process the original value of tha signa in Grosk was restored, instead of the opposite (sce follow- ht be accidental, but is very likely due to the influence of Campanian ‘Tho apparent transponition of the signs for d and r ie accounted for ax follows. ‘Tho Etruscans had no sound d, but used Qa a hy-form of 4 =, in fark preferred it, as less likely to be confused with ‘1 = p; and with this valuo lopled by tho Oscana and Umbrians, But We old signs for + wero also taken vor and employed for the sound «, — A by the Oscans, 4 by the Umbrians, arly Umbrian we of 4 as d is seen in some of the minor inscriptions. ith tho chango of intervoealie d the letter was retained for the new sound, Wcribo F, and thenceforth the unchanged d was expressed by a rounded form ico versa in Latin of tho sign 8, f, in disputed. Possibly i of G, nsed frst. in combination with C, and then alone, as first FB, then F. ‘Tho relation of the alphabets may be seen from the following! : Chalcidian Grock Latin Primitive Etruscan, Campano-Ftrusran Osan Etruscan Umbrian (of Etruria) Notes on Orthography 28, Résumé of methods of indicating vowol-length. The Tength of a vowel may he indicated : 1) by doubling of the vowel sign,—in Osean of the native alphabet, rarely in Umbrian of the Latin alphabet. See 22, 26. 2) hy vowel +h, — in Umbrian of both alphabets. See 26, 26. 3) hy vowel ++ vowel, —in Umbrian of the Latin alphabet. Sco 26. 8, Part Ty whieh also contains a eomparati t fora af tho letters, 1 From Conway's Hate Din talilo of tho nlphahots with the: vaEB BSS] | 26 Phonology [28 But oftenest there is no designation of the length, and in such cases it is not customary to supply marks of quantity, as is done in the case of Latin, where metrical usage furnishes a criterion lacking in the dialects. Vor oxample, we write O. aasas, cituas, U. totar, though in this caso there is no doubt, ‘of the vowel-length in the last syllable (Gen. Sg. ending ~as). In Oscan the designation of length is, with a few excep- tions, confined to root-syllables. 29, Use of ei, ci, in Umbrian. While in Oscan the digraph ti, ci, uniformly designates the diphthong ei, its uses in Umbrian, where the original ef had become a monoplithong, are various. Sometimes it designates a secondary diphthong, the ¢ of which comes from a guttural, e.g. teitu, dette ‘dicito’ (143). But it is frequently used in the Latin alphabet, and rarely in the native, much as in Latin inscriptions of the first century B.C. as one of the various spellings of 2 monophthong. It is notably frequent in the first thirty-odd lines of Table Via. Oftenest it stands for original i, e.g. sereihtor (L.. scriptus); sometimes for the close ¢ resulting from oi in final syllables (67, 2), e.g. Dat.-Abl. VL uereir, or from original é, e.g. nesimei ‘proxitne’ (adverb in -é), heriiei (Perf. Subj. with mood-sign ¢). ‘There are also a few reasonably certain instances of its use for a short i, namely Dat.- Abl. Pl. aueis (*-ifa), Acc. Sg. Fisei (-im), 3d Sg. Pass. hertet beside herti, herter (tir from -ter; see 39, 2). Puzzling is the use of ci in neip, neip ‘nec’ (with neifhabas ; see 84), in eikvasese, eikvasatis, of uncertain meaning, and in eiacurent ‘arcessicrint’, For eitipes see 264, 2. 4. For eikvasatia and eievasene conncetion with Ts, aoqio in plausibl for eiscurent the comparizon with 0.11.G. eiscon (Germ. heiachen), Lith. jéazkatt ‘eeek’, ele, pointing to a Present *aiskd (Skt. techdnt from sis.rk0 vith reduced grado of root) in tho most probable of all miggestions offered, Yet according to the nsuab orthography wo shuld expect e for the open & coming from ai (63). 1. is conceivable however tat wo hare here isolate murvivale of archaistic apolling, representing not the earliest period whon ef wae still pro- nounced aa a diphthong, but a second period, in which the apetling et wan retained for the sound resulting from ei and extended to the raine sound resnlLing andPe, ee, ee eee ae 30} Notes on Orthography 27 from ai (both ei and ai resulted in an open 2 ; reo 68, 65). Cr. early Latin deicd and inceits. ‘Tho ordinary use of ei for f, close 4, otc., ar described above, cannot he tho result of auy such orthographical development within the Umbrian, ainco it does not appenr where the sound was originally ei. Tt must rather be regarded as borrowed from contemporary Latin spelling. 8. For neip, neip wo might also assuine archaistic spotling (cf. 0. neip), but its almost uniform appearance in this particular word (aeip, neip 0 times, ‘onco nep) would romain to be accounted for. A suggested derivation from *nt (from ne, 0. ni, or *nei, O. nei) + particle -i4 p would explain the spelling, as reprosenting a gonuine diphthong, but for various reasons seems isnprobable. 30, While Oscan orthography, barring the inconsistency in the designation of vowel-length and a few other, mostly localy.- <° variations, is remarkably uniform, Umbrian orthography is 4" ! diverse as possible. Various spellings of the same sound afd: used, sometimes wholly promiscuously, sometimes with a mark preference for one spelling in cortain portions of the tables or, . in certain classes of forms. Among the commonest variations jy55 | are the following : ieee 1. Variation between ¢ and &. In the great majority of instances this cecurs whero tho sound lies between ¢ and i, or, moro correctly, between tho extromes of an opon ¢ and a closo #; that ia, it ie elther the open # from original hott 4 (45), oF the cloge 4 from original # (42) or from of in final eyllablon (67, 2). ‘The apelling ¢ is rolatively moro froquent in the native than in the Tatin alphabet. ‘The use of ¢ for clawed ¥ from original 2, of, vico versa, of ¢ ; for tho open e from original short ¢, oF for open 2 from original ai ore ‘The variation between e ant eorresponds then in general to the Oscan use of f 2, Variatjen between ei and eor i, Seo 29. 3. Variation hotween o and x (only in the Latin alphabet, of course, 1 Uhe ease of original 3 (64), some in the Tatin alphabet always 0) for the ronnded a (asin English call), coming from final -& ($4). . Variation in the designation of vowel-lengih, 0.g. 6, eh, or ehe (iu native alphabet only eh), oF, oftenest, simply e, ¢, without indication of length. Seo 25, 26, 28. 6. Variation between p and b in the native alphabet, e.g. habina, hapia It is doubtless owing to the double value of t and k, which answer for both urds anid sonanta, that p is also used not infrequently for b. 7. Variation between single and double consonants. Double consonanta are not indicated in the native alphabet, and only occasionally in the Latin.28 Phonology [30 8. Prosence or absence of A. ‘Tho weak pronunciation of in Umbrian is responsible for considerable inconsistency in spelling, just as is the ease in Latin, See 149, a. The use of has a sign of hiatus is common to both Osean and Umbrian, .g. O. atahint ‘stant’, U. ahemnes ‘alt 9. Presence oF absence of n before » consonant (108, 1) 10, Presence or absence of r before # (115, 116). 11, Presence or absence of most final eonsowants (164, 9). 31. An important difference between the orthography of the native alphabets and that of the Latin alphabet, in both Osean and Umbrian, is the following. The glide sound which naturally intervenes between i or w and a following vowel is regularly expressed in the native alphabets, but nearly always omitted in the Latin alphabet, as in the spelling of Latin, So U. triia, but trio (L. tria); U. tuves, but duir (L. duo) ; O. eitiuvam, but eituam. @. i, & In Umbrian, of words occurring in both spellings tho examples are: triia (9), trio (2); heriiel (1), Aerici, heric (4); Atiietiur etc. (17), Atioraur etc. (5) ; Klaverniie (2), Claverniur (1); Vehiies (2), Vehier (4); in all, 46 oveur- Fences with no exception to the distribution of tho two spellings as stati, In Oscan too the spelling ii is cmployed consistently, as in the oblique caus of names in -iia, contrasted with i in the oblique caxes of names in -is; ¢-g- Dekkieis Rahiieis Gen. Sg. of Dekis Rahiis (174). Since ii is so evidently tho normal spelling in the case of vowel 4, Unere is the strongest presumption that, whero the spelling in the native alphabots is simply 4, this must represent something different, namely the consonantal i, And this is often corroborated by other evidence, such as doubling of consonants, in Osean, occasional omission of tho i in Umbrian, ete. (100, 3). Yet some exceptions imust be almitted. In 0. Dekkviarim and U. tokvias 4 cannot possibly represent a consonantal i; 0. Iiviass is not to be separated from fiviia; in O. vid, U. via, vea, consonantal ¢ is of course impossible, and that the vowel is othor than original i (ef. I.. via) is improbable ; consonautal i is also itnpossible in U. arvia, and improbable even after v preceded by a vowel, as in aviekla ctc, It is perhaps for the very reason that there would be no ambiguity, that { is s0 often used in place of il after v. ‘A differont case is that of tho Oscan i coming from original ¢ before a vowel (33, 1). Hero too in tho earliest inscriptions the spelling is ii, but after the introduction of the character { this alone is used ; e.g fink, later Sik ¢ 4. uy, w. In Umbrian the contrasting examples are: tuves ete. (5), duir (2); kastruvuf (4), castruo (11); prinuvatur (6), prinuatur (B); vatuva (8), uatwo (6); in all 47 occurrences with no exception to the distribution of he two spellings as stated. But wo find saluxom, saluua, onco each beside 24 examples32] Vowels 29 of satuom ete., and tuua ‘tua’ once beside 18 examples of tua, fuer (once also toucr). ‘Tho omission of v in purtueta is doubtless accidental, and aruvia besiddo usual arvia is probably an angraver's error. In Osean, v is useil instead of uy hosido sakruvit, in eltiy. for eftiuy(ad), and probably in minive (no. 315). So possibly in U. iveka ‘invencas’, though hero tho omission of u seems much stranger, and many assume an actual phonetic cbango of juy- to ake HISTORY OF THE SOUNDS! VOWELS a 32. 1. a@ in initial syllables remains unchanged, as in Latin, So O. actud : Lu. ago ;—U. ayer: Le ager; —O. allo: 1. aliua ; —O. patir, U. patre: 1. pater ; —O. fakiied, U. facia: L. facts ; — O. castrous, U. castruo: L. castrum ;—O. ant: 1. ante, 2. Final a ia also unchanged, as in the Umbrian Vocatives Tursa, Iouia, te. See 168, 6. 8. Likewise in medial syllables, where in Latin a has been weakened Lo é or ¢, it is regularly preserved. So O. Anterstataf: 1, "Interatita (cf. Praestitia); —O. tribarakaviim ‘acdificare’: L. (co)-erced ; — U. antakres: 1. integer ;—U. procanurent : L. (0¢)-cinui ; — U. atkani ‘cantum’: L. *accinium ; — U. tuplak ‘furcam’ @): L. duplex (cl. Gri. dérdak). See 86. ie arrangement of the material and the choice of head considerations of convenience, Since we are dealing primarily with the relat the sounds of the dialects to one another, rather than with their rclations to tho ronnds of the other Indo-Furopean langttages, tho material is arra ence to what belongs together fram tho Italic point of view. ‘TI ais considered the history of Italie a, regardless of ite various IF. sourees (a, 2, etc.) from LE. 9 las the saine history as original en, and need wot bo treated reparately similarly with or, of, from 7, 1, ow from en, ete. Only in tho treatment of Vowel- Gradation is thero any necessity of reverting to the LE, vowelaystem. But the headings do not always represent the Italic sounds. It is often simpler to take the E. nonnuls as tho starting-point, as, for example, in the case of the sonant aspirates, dh, dA, ote., for which the precise stage of development reached in the Italie perio is not in all cases certain, Or, again, it may be desirable to discuss in one place tho history of a sound or group of sounds, which is partly of Indo-European, partly of Halle, and partly of atill Inter origin, as, for example, in tho ease of ms. In Keneral, tho author has not hesitated to anctifico consistonicy to convenience.30 Phonology [2 4. But a weakening in the direction of u, where a labial consonant precedes or follows, is seen in a few words. Sec 33. 4, except when final, remains unchanged, as in Latin. So O. fratrim, U. fratrum: 1., frdter ;—O. Maateeis, U. AMatrer : L, mater ; — O. assas, U. asam: 1. dra ;— Abl. Sg. of First Decl., O. toutad, U. tota: L. -@; — suflix -~dno-, O. Abellanis, U. 7'reblani L, Romanus, 34, Final a, which in Latin is shortened, preserves its quantity, but is changed in quality to a rounded sound like tho a of English call. In Osean it went so far in the direction of 6 that it is never denoted by the letter a, but always by 3, 0, 0, or, rarely, by a. In Umbrian the sound is written both « and u in the native alphabet, but always o in the Latin. Ex- amples are the forms of the Nom. Sg. of @-stems, which ended in 4, a8 shown by Greek, Sanskrit, cle., and of the Nom.-Acc. PL. Neuter, in which tho 4, belonging properly to ostems, was extended in the Italic period to other stems. Oscan. vid ‘via’, fiend ‘fanum’ (Ace. fiisnam), fiw, fi-k, ioe ‘ea’, molto ‘multa’, allo ‘alia’, touto ‘civitas’ ; — comono ‘comitin’, teremennid ‘termina’, peliru-pert, petiro-pert ‘quater’ (192, 2). UMbrtan. muta, mutu ‘multi’, panta ‘quanta’, etantu ‘tanta’; — veekla, veskdu ‘vasculn’, vatuva, vatuvu, uatuo ‘exla'(?), prosescto sprosecta’, atru, adro atra’, See also 238, 236, 2, 297, 300, 9. 35. In Umbrian this rounding of the @ takes place also before final -te (from -to-s or -ti-s by vowel-syncope). So pihaz, pihoa‘piatus’, kunikaz, conegos ‘conixus’ (in form as if L. *eont- ‘gitus), Casilos ‘Casilas’ (Dat. Casilate), -vakaz, -uacos ‘vacatio, intermissio’ from *yakat(i)-t. @ A similar variation in spelling, which ean hardly bo reparated from ‘the phenomenon just dencribed, in neen in Prestate, Prestote, and Teseaakes, Tesenocir, The former word, although L. Pracaitia suggests * prae-ita-, may be from a by-form *praca (ef. L. prae-sfatus Lesido prae-atitus), anil forf a] Vowels 31 the lather word a mufix -Ako- ie in itself more probable than -ako-, the existence of which is doubtful. But the oxplanation is dificult, since elaewhern there is tio indication of a change of @ except under the conditions described above, It is porsiblo that in tho lator Umbrian even tho @ of medial ayllables changed slightly in tho dircetion of 4, but not enough to affect the usnal spelling. Yet It is atrango that tho i so consistently employeit in these two words, and never found as a variant in the great majority of words containing a. But to regard the o as standing for short a only increases the difculty. Such a weakening of @ whero thore is no contiguous labial consonant (88) ia unsupported and ly. A somewhat different, but equally difficult, caso is subotu if this ta the ‘samo word a eubaktu ‘doponito’ with secondary a (121, 76). e 36. 1. egonerally remains unchanged. So O.edum: L. edi — 0. ezum, est, estud, U. erom: L., ease, eto.;— O. destrst *dex- tra est’, U. dertrame: L dexter; —O. mefat: I. mediue;— U. ferest, forte: I. ford; —O. aragetud: L.. aryentum, 2. ¢ also romains before 2+ consonant, or final % where in Latin it becomes first o, then u. SoU. pelmner : [.. pulmentum ; — (also U. veltu ‘deligito’, eh-vetkta ‘sententiam’, but in these wel- is from yole-: L.. volt, vult);—O. famel: I, fant; — U. eumel: L. simul (early inser. semol). 3. ¢ also remains generally in medial syllables, where in Tatin before a single consonant itis weakened to i. So U. tagez, tasctur: I. tacitus;—U. maletu: L. molitus;—O, Genetal: LL. genitus. 4. But before a lnbial in medial syllables a weakening occurs, resulting, just as in Latin, sometimes in w, sometimes iwi, Seo a 37. A change of ¢ to 0 is seen in *pompe ‘quinque’ (0. pum- periaa, U. pumpetias (*quincuriae’, O. pomtis «quinquiens’) from *hYenkYe (160), where it sce two Bs. 4, The combination aye which Becomen s0 in Tatin (toror from #ayesin, ete.) romaina unchanged in O, everrusel (86), but Umbrian shows the same cchango as Latin in aonitu: L. sond, from *ayen- (Skt. svan-).82 Phonology [se fore 38. Oscan. 1. Before another vowel, e becomes an open ¢ and is invariably denoted by an character (( in the native alphabet, carlicr iis see 31, «). Compare Hal. mio from L. meter, ceria from I. ereat, clo. So iiu-k, fi-k, éo-¢ ‘ea’, tone ‘eum’, fus-c, fussu ‘idem’: 1, ea ete. (cf. also Marruc. tafe teas’); —fatium Mari’: La fateors pitiad ‘posit’, as if L, *poteats — turumiiad “lorqueatur, as if L, “lorment; — Tianud “Peano, Loc. Sg. Tiianei ;— Tiiatium ‘Teatinorum’, 2. Before r the ¢ had a closer prontneiation than usual, as is shown hy amiricatud +*immercato’, with which may be com- pared rustic Latin Mircurios, stireus, ele.; further by Tirentlum *Terentiorum’ and Virriis ‘Verrius’. But the change was so slight as to be commonly ignored in the spelling (cf. pert, pereke, pumperias, ete.) 3. Tintirii, if, as probable, from * Zinktrio- and related to L. tingus, tinctur (Girk. réyya), is evidence of the same change as occurs in Latin before » + guttur 4. In aistrus ‘propingnos' beside messi ably only a misspelling 5. For fat beside ent a, see 217, 2 9 ‘proximne’ ele., the 1 I¥ prob- 39. Ummitan. 1. Before another vowel ¢ had a rela- tively close pronunciation, as shown by farsio, fasiu ‘farrea’, turriandu ‘lerrcantur’, and by iepi, iepru, in ease these are from the stem co-. But the change did not go so far in the direction of as in Oscan, and the spelling is regularly ¢, e.g. cam, eaf, 9, cle. . 2. From ostensendi for Fostensender (ending -ter, 298, 1) and herti (4 times), hertei (once), beside herter, herte, we may assume that ¢ before final r had a close pronunciation verging on é 3. In eringatro, krikatru ‘cinetum’ beside krenkatrum, from *krengh- (O.Eng. hiring, O.Bulg. krayii tcirele’), we have a change of e in the direction of i, as in Latin before m+ guttural (tingua, lingua, ete.) Sev 98, tr a Ae ee ee ee ee eei al Vowels 4. In tse his! bose enir, — ito, inte ila’ Veto et, exe — tee, beh Siuon, —twunt*iten’y tho # Hs perhaps dio to a partial cmamivation of the lene etm nl LL ne the following. B. "the winglo oceurrence of tavin against 21 examplon of tase te, and of viata against 18 examples of vestige, vestigia, ele, shove that in the following forms, which occur Int co cacy wo ay havo, neckientlly, the abornal rather dhan the nornal spalling: tigi a dorcl;— isgsles Hnsectis', with {for tne; vagetumel ae *in viliatum si?) with {for postpositive e(n). hut it i fnnmot bo wholly accidontal that in all these eases (el also fair et, above) the rowel is followed by a sibilant. Apparently tho Fwalty of the mibiant hn Ia some effect om tho preceding ¢— Tit a alight tat in ont won ts nover in tho apelling. 1a U thy vit’, witht, th fs Hate (Ie it, also 0. Viet ‘alia’, hough probably from original e (“yearling ef. I. ete ‘ealf), "Where ad how the change cause about In unknown, a é 40. ¢ had a closer pronunciation in Latin than the short ¢, a we know from its development in the Romance Inmguages A and from statements of the grammarians. I was the French | of éé rather Uhan the @ of mire. It probally had this rela- tively close promneiation in the Talic period, and in Oscan and Umbrian progressed still further in the direction of %. | 41, In Osean it has gone so far that we may speak of a | change to %, since it is invariably denoted by an i-character. 1 This # was « relalively open %, indicated in the native alphabet by (or ff, being Unis distinguished from original 7, which was close! Examples: ligud loge’, Mgatits ‘legatis': 1. ler, legit ; tus jsni, fiisnam, flenam ‘fawn’: L.. firatus, fériae (99, 1)3— h Wikitud, livinud Miceto’: L. lieeta; —hipid shabuerit’, from *heped, belonging to the same Perfeottype as L. edpi, Wf and with the Subjunotive-sign 7;— fus{d ‘forct’, herring ‘eaperen™ with the same moodsign & as the Latin Imperfect Subjanetive, but with- out the shortening seen in La. et, -ent (78). ode Thin ans similar atatemente an to the distinction in we letween Fail i rote fo the normal Oman orthography, Ie must bo remensberod that the {i Tneking i in oldest ineeriptions, and also that after its inteaduction it was aw earlesely cunployeot in some inscriptions, mostly those of Capua, that their evidences in this rojgard ie Lo bo ignored, f Vvaananvantenentantet34 Phonology (a 4. An 2 which is the result of contraction in tho Halle period has the same dovelopment as original 2. Thus tris: L. trés, from *lrejes ;— hirtia ‘tin horto’ from *hortei-en. See 82, 1. . Dut an 2 resulting from rome later process of Vowel-lougthening retains tho quality of the short ¢, and is not written {; e.g. keenzatut, censlur,eeatint, ete. (13, 77). 42. In Umbrian the spelling i occurs frequently, especially in the Latin alphabet, but e is far more common. The Imperatives of the Second Conjugation always havo é in the Latin as against e in the native alphabet, but this dis- tinction does not hold for other words. ‘Thus habitu, babetu: L, habéta; — tursitu, tusetu: 1, terréta ; — filiu, feliut ‘laotentes? from the fé- of L. fé-mina;— plener, plenasier: L.. plénut j— rehte: L. réeté, earlier *réctéd. In a few cases the spelling «4 occurs ; eg. heriiel ‘volucrit? with the Subjunctive-sign ¢;—nevimei ‘pro- xime’, adverb ¢ rehte; sei-( podruhpei) ‘scorsuin's L. scum, Evidently ¢ in Umbrian had a very close ‘pronunciation, but had not gone as far in the direction of 7 as in Oscan. i : 43. i remains an open i. This open quality is shown by its designation { in the Oscan alphabet! and for Umbrian by the frequent spelling ¢ (80, 1). Final i, unless dropped, remains é in Osean, but becomes ¢ in Umbrian, as in Latin. Thus U. ute, ote ‘aut’: O. auti; —U. ealre, eacre, etc. (Nom.-Ace. Sg. N. of i-stems). 44. Oscan. Examples: dadikatted ‘dedicavit’: L, dédicd; — meddfes, meddis ‘meddix’, Gen. Sg. medikets, Nom. P'1. medias, peBSak (for e see 24; compare also exverp = {nlm), a compound of dik, like L. diden, iidicis (18, 6);— Ueltud, licitud ‘liceto’s L, iced; —uineter ‘convincitur’: L. vined ; — tiurt: L. turrim; — pla, pin, pid, pilpit: L. quis, quid, cte.;—sullix -iko, eg. tiv- tike ‘publicus’, toutico «publi @ When the consonantal ¢ intorvening between tho vowel ¢ and another expressed in the writing, as in nearly always the case in the native vowel ¥ Bee footnote, p. ih.SSSR OPP F 4 a7] Vowels 35 alphabet (31), tho vowel i ia then written f, not $; e.g, fakilad ‘facia’, need ‘eapint’, Heleiiele Helv 6." An é arising fom consonantal é by samprastrana (81, 1) acoms to have 1 differed in quality from original i, judging from the spelling of péstris *poste- riuw* with-in, nol fa, from fr, in consequence of which tho anaptyctie vowel ik alno 4, not 1. Ch. also the proper names like Vibie*Vibius’ eta, (172-174). For {in Mals cte., s00 176, 8. . Isolated examples of ¢ for é aro: meavum ‘minuore' on the carelessly } written Curso of Vibia ;—esldum, eslde[m), for the unual fldum ‘idem, on two inscriptions of Sarmnium, possibly due to a loca! contamination with the stem of caaut “pw. 45. Umnnran. The spelling is either ¢ or ¢, oftoner the former. As is the case also with other sounds which are rep- resented by both spellings, the ¢ is more frequent in the na alphabet than in the Latin. See a0, 1. or the rare ef, s Examples: tigel ‘dedicatio’, tikamne ‘dedicatione’: 1. died ti6 ele.3— dersioust ‘dixerit’ from *de-dio-uat : O. ddicuat (44, 95); —uirseto ‘visum’ from *uideto-: L, vided ; —eteplatu, stiplatu, ; atiplo: L. stipulor; — sestu, seste: L. sisté;— teta, diraa, dersa ‘det? (from Redupl. Pres. *didd): O. did it pirei, perei, ctc.: 1. quid, O. pid;— sullix -iko-, e.g. Puprike “Publico’ @), fratreke, fratrez **fratricus’. 4a. Tho Accusative Singular of items nearly always appears as -e(m), -e(n), e.g. uvem, ure, ocrem, acre ete. (178, 4), indicating that before final m an é waa more than ordinarily open. Contrast tho -(m)y-(m), i, of io-tema, in which th & comes froin consonantal j hy sxmprasaraia (01, 1). Cf. 44, 8. 1 46. Thad a closer pronunciation in Latin than the short i, as is proved by the Romance development; e.g. Ital. chi, scritto from L. qui, scriptus, contrasted with che, lece from L. quid, zet. Thosame qualitative difference existed in Oscan-Umbrian, as is shown by the fact that original @ is indicated in Oscan by 4, not §, and that in Umbrian the spelling ¢, 60 common for short 4, is rare. 47, Oscan. Tho spelling is i, not {, but where doubling is employed as a mark of length we find if, not ii like aa, ete.36 Phonology tr ‘This spelling if may possibly it something like ie, but more probably it is purely orthography, it being avoided on account of its other v imple: liimité{m] dimitum’: L. limes; — imaden sab imo’: Ja Tats; —soriftas scriptac’: L. seriptus; — Ab. S; slaagid contrasted with Ace. Sg. slagim ; — sullix -Ino- nais “ivinis’, Zantins ‘Bantinus’, Mapeprevo *Mamert 48. Umnniax. ‘The spelling in the native alphabet is i, ih, rarely e} in the Latin it is i shi, ed (very frequent in the first thirly-odd lines of Via), rarely e. Examples: persnimu, persnihmu, persnimi, peranihion ‘pr calor’, Imperative of the Fourth Conjugation ;— screhto, sereih- pins ;— pica, peica, pein (10 times in VI 1-17) Pics ;— pehatu, pihatu ‘pinto’, pihancr, pehaner, peihaner ‘piandi’s L. pid from *pid, O. Pithidi +I*io'; — sullix Tuo og. Tkuvins Iguvinus’, Jouénam, Joueine (ci oneo only in over 100 occurrences). cate a nuance of pronuneiation utter of oy L. pie ° 49. 0 vemains for the most part unchanged, and in alphabet as 0, in the native Osean alphabet as a. the native Umbrian, and also in the earliest type of the the V did duty for both 9 and n, ATl forms from these sources must therefore be ignored in dis of o and x, Examples: O. dp, op, U. ostendu tostendito’ from ops: 1. 0b, obs;—O. pist, post, U. post: 1. post; —O. pid, port, U. puke, porte: 1. quod; —O. ipsannam ‘faciendam’, U. sate ‘facito': L.operor;—O. hire: L.hortua;—U. poplom: L. populus. The o is also preserved hefore Z-+ consonant and before n+ guilinral, where in Latin, except ly inscriptions, it appears as wu Thus O. molto ‘multa’, miltasikad tmullatica’, U. motar: 1, mutta, early molta, moltiticéd; — 0. éitiumam ‘ultimam inguishing the sounds L.. nltins from *oltimo-;— 0. ione ‘cum’, with which compare L. hme, early hone,a] Vowels 37 uforo 50. In Oscan, before final m the 0 heeame 1, or at least was changed so far in the direction of waa to be commonly writlen w. ‘Thus the Present Infinitive (ending -om, 241), with the possible exception of tribarakavim (i not certain) on the Cippus Abellanus, shows -m, -um; eg. acum, deiewm, czum, censawm, detkum, fatiam, ‘The enclitie particle -om (201, 6) always appears as-um; cag. pieisamcuinspiam’, pidum‘quidguam’ (C.A.), fsidum ‘dem’. The Ace, Sing. of o-stems, however, though sometimes showing -um, as in dolim, trutum, Nivellum, etc., usually appears as -om (more frequent than wm on the ‘Tabula Bantina) or -dim (always on the Cippus Abellanus). It is alto- gother probable that this spelling of the Acc. forms is a sort of pedantic orthography,due to the 9 of other ease-forms (-0f,ii-,ete.), while the spelling of the other forms, which were not subject to such influence, represents more faithfully the actual pronunciation, Nore. In Umibrian not only docs o remain wichanged before final m, ‘but even w becomes o (67). 51. In Umbrian before + + consonant, or even before r alone, we find so many eximples of the spelling x, although 0 also occurs, as to make it evident that the vowel was consider- ably modified in this position. ‘Thus eurnaco, curnase (5 times): L.. corniz:; — prefix pur- in purditom sporrectum’ ele, (10 times, never por): 1. porss—tursite ‘terrelo’ ete. (4 times) from *tor- 808 (91); —courtnst ‘reverterit? for *counrtuat beside covortua: carly 1. vormua, advortit, ote. (91); —Juro forum’: 1. forum (it is unnecessary, though possibile, to assume that furo contains the reduced grade dhur- like Crk. Opa, as compared with divor~ in L., forum); — trsiandu Tugentur? with de for tur from tar; — nr ‘illo’ beside orer sillius'@). rt, Possible examples of x fora before ra from d (131) are daepursus, pet pursue (Init e094); alsa atripurant*tripordate the explanation of which deperile co the view taken of 1. friputinn ete, besita early fripa (late weakening of to 1 oF contamination with a derivative of a rou pnud- related to pai 4. An isolated instance of w eleowhore is swift beside sonilu: Te son,