Studies in Greek, Italic, and Indo-European Linguistics: Offered to Leonard R. Palmer on the Occasion of His Seventieth Birthday, June 5, 1976 3851245318, 9783851245318

496 114 11MB

English Pages 449 [228] Year 1976

Report DMCA / Copyright


Polecaj historie

Studies in Greek, Italic, and Indo-European Linguistics: Offered to Leonard R. Palmer on the Occasion of His Seventieth Birthday, June 5, 1976
 3851245318, 9783851245318

Citation preview







Band 16



Offered to

LEONARD R. PALMER On the Occasion

of his Seventieth Birthday

June 5, 1976

Edited by








Dio Drucklegung des vodiegendon Bandes Wl.ll'dedankenswerterweise gefordert durch Subventionen aus Mitteln folgendel' BehOrden und Institutionen / Grateful acknowledgement is made fol' contributiorn, received from the followingsources: Bundesmi.:nisteriumfill· Wissenschaft und Forschung, Wien Amt der Tiroler Landesregiomng Akademischer Senat cler Universitti.t Innsbruck Universitatsbuncl lo.usbl'uck Verbancl cler wissensclmftlichon Gesollschaften 6sterreichs (auf Antrag der Innsbrucker Sprachwissenschu.fLlichenGesellschaft) The .Jowett Copyright Tl'nstees, Balliol College,University of Oxford The Boden ]I'und, UniversiLy of Oxford Worcester College, University of Oxford

ISBN 3-85 124-531-8




Herausgeber: Prof. Dr. Wolfgang Maid Institut fiir Sprachwis.senschaft der Universitat Innsbruck A-6020 INNSBRUCK, Innrain 30 Druck: Ernst Becvar, A-1150 Wien

Dear Professor Palmer, This volume comes to you on your seventieth birthday together with the warmest wishes of all those who contributed to it. They would have been far mo~ numerous if the editors had not had to yield to practical considerations and had not steeled themselves-however reluctantly-to be l'Uthless. As it is-and in spite ofiW 46 articles-the volume must only be a token of gratitude rather than a trne return for the many books and articles which you have donated in the past to the scholarly world and which you continue to produce with unfailing regularity. Your scholarly career led you among ever increasing honours from Cardiff to Cambridge, to Vienna, to Manchester, to London, to Oxford, and now back again to Austria. It seemed appropriate that these 8tudie8 should appear as a joint enterprise sponsored by the two countries which have most profited by your presence and your work and which have provided the necessary backg1·ound for them. From 1971 Oxford's loss has been Innsbruck's gain, but more than forty years ago your scholarly work first started to bear fruit in Vienna, in close contact with such leading Greek specialism as P. Kretschmer and L. Radermacher and such a revolutionary linguist as N. Trnbetzkoj. From Vienna you brought back to England .:fil'Stand foremost a wife, a remarkable scholar in her own right, but also that degree of expertise and maturity which soon allowed you to write your first two books; it may not be chance that one of them discussed the principles of modem linguistics and the other was concerned with post-Ptolemaic Greek. Now in Innsbruck your two latest books are completed or are about to be completed; once more the first deals with the principles of linguistics and the second with Greek, but while one embraces both synchronic and diachronic linguistics, the other tackles the whole history of ancient Greek. There has been no return to old work and old ideas, but progress and development. In the meanwhile the flow of books and articles has been continuous. Shorter works have made impo1·tant contributions to classical and lndoEuropean philology or to the study of such recondite languages as Hieroglyphic Luwian, The Latin Language has given students in Britain and elsewhere a. chance to see that the classical languages are not monolithic units governed bjr unassailable rules, but have a history of their own which is nothing but enthralling for anyone who is led to it by the right guide. The more specialized work on Linear B into which you tluew yolll'self in tho early .fifties with the humble devotion of a neophyte but the

Dedication sharpness and the expert..ise of a leading scholar soon made you into a world authority in this field, most often admired, frequently imitated, somotimes disliked beca11se of your devastating habit of being right so oftennever ignored. Then came archaeology. Suddenly we leamt that the Oxford Professor of Comparative Philology was no longer content with the study of language. This had led him to results which could only be supported or invalidated on the basis of archaeological evidence: never did archaeology have a more eager or enthusiastic student. This time your results were so different and so revolutionary that they became known even outside the scholarly world. In Oxford even shop assistants were eage1· to meet Professor Palmer 'the archaeologist'. The scholarly battle about the chronology of Knossos and its documents still rages, nor is this the place for a bulletin from the front, but no one-on whatever side he is-will deny that much good has come from your courageous attempt to reject all preconceived ideas, to take nothing for granted, to check all sources with dogged single-mindedness accompanied by sudden flashes of inspiration. Archaeology had to be excluded from these Studies for a very simple reason; if all those whoso work had been influenced in one way or another by your own contributions had been invited to ·write something for it, this volume would have shared t,he destiny of the biblical loaves and fishes. You once ,vrote: 'I have not everywhere been able to conceal the fact that I have opinions of my own'. You will forgive us ifwe now dare to contradict you. That sentence was an understatement and can only count as a pale approximation to the truth. It is just because you havo always had opinions of your own and expressed them with the sharpness, tl1e vigour and the originality of the true scholar that we have wanted to say 'thank you' on this very day, your seventieth birthday. June 5, Hl76

Anna Morpurgo Davies Wolfgang Meid

CONTENTS Page Dedication William Bidner J.:c.L .. ll"; Long and short diphthongs: phonological analogies and phonetic anomali• F'ran9oiae B~•:a; N oms de bergors de la racine *paHarold W. BAILEY; Indo-European 81.ler-'to colo,u· with a (lark conlour' . Thomas BURROW;Sanslu-it words dontal •8- after i, u, ancl r J"ohn CHADWICK; Hyoona,an •-r,-,o; a problem Robel·L COLEMAN; Pattern■ of synoretism in Latin N. E. Cor,r,tNGE: Global rules, active del'ivation, and Lat;in mellis Ilya GERSHEVITCS:; More moat in Iranian . Anna GV,OALONE RA111A'I.': A pl·oposito dei composti ge:r:manici con gaRobe:r:to GusMANI; Zurn Alter dos jonisohen Wandels i.i > 71 • Bryan RAlNSWOllTH:Phl· in Hom or El-ic P. HAMP; •g"•iH..-'live' Gillian R. H.AitT; HWoite !••,.,i-i#-tc, Alfr11dfuu~110:i:: Weil•1·11 .1tu Dl'l.~i11n01g dtr Kno1101-Tafeln RolfHuinsc:rm: Zur Bodeutung von ~-i,yµ[,;: Y ng Stefan Iln.LER: Das LOwentor Yon ll!yk•n• und die lda11i111heT:i·1gOdie John T. Kn.LEN; Linear B a-#o-m-jaf-jo . . Jorzy KURYLOWICZ;Phonologiscbes zum inclogormaniscben a-VokalismllS . :M:icholLlSJEUNE: Quel oeltique clans AEL\EBPATOYAEKANTEM? . Giulio 0. LEI'SCEY: Iialian Olil.t11al;n1and pel'ception predicates followed by an infiniti..-e: oompet1no• and potfonnanoe Manu LEUll-IANN: Zur :rorm Ton n•ugr. pto. ypocµµ€voc:; Olivier MASSON;La plu, anoi•nn• in,ol·iploion or'-oi1e M11m: Zur Btymologie d111\Vodea !Ur ,,Mensch" im Irisol1en Anno. Ho~filO D.LYI111:The -ecrcridatives, Aeolic -88-, and the Lesbian poets Robert MUTH: Mm·tials Bpi1l mit d•m ludu1 po1liou1 Gregorr N.AfilY:The name of Achilles: ,tym.ology nnd epic llrioh N11u: Zm R,kon,kuklion d11 indo~ermani1ell•n V•rbal1::,1t11n1 Giinter NEUMANN; Zu einigen lcr11ti1ohenPononennamen Karl 0BERHUJ3E"&; Ein zentrales Ptoblem dee altm0sopota1uischen Religionsgeschichte Oswald PANAGL: Zu:r Jlltymologie von grieoh. f:t,;o,;: Vittm·0 PISANI; San1.lcrit a~l11,.-d:-, grieoh. a,,~.16,;: James W. POW..TNEY: A p1·oblem o! 1ero-grade vocalism Ernst PU.LGRAM; Venetic ·e·kupeOari·s• E. G. QmN: Irishjemenclae Ernst RrsoH; Die Stoffadjektive anf -ejo8 im liiykenisohen Helmut Rn:: Di• ux11bri1oh1nInftnHi..-• au:f -Jiund die m·inclogerrnl'l.ni1oh•Inflniliv•ndung -d-i"t


17 29 33 43


57 63 65

77 83 87 93

97 103 107 117 127

135 153 163 109 173 181 199 209 239 255 261 273 283 285 299

305 309



Robert H. Ro:ows: Vll,t·1·0and the •ao~io1of cml\lo,5i1L ,5r1,min1,tian1 Corneh1 :I. RuLTarr: Obse1·vo..tionsSLU"la !le:idon de■ vorbo■ du type -rp(~(,), > (mttlus); nabliO tudans le pl'Csent thematique skr. Uvati {(remarquer» e~ Jes formes ttpparentCesG1, et des formes nominales: vieux neutre en *-i- sigmatisC skr. l'tvib,,av. iivi.s«C\•idcmment,», cf. v. sl. cwt, javl, gr. a.tcre&,.,oµa.t,!at. audiO; substantif animC en *-m-0-(qui pent s~ E. Vilborg, A tentative Grammar of 1l!ycenaemiGreek (1960), p. 71; P., R. Ph. 29, 1!l55, n. 2, p. 22-23.


:; f.i~,', Ph. S. 1954.(1955], p. 24; suivi par (Tlmrnb)-Schorer, dei· gi-ieohisohenDiltlekte II, § 276, 16b. sa H. F1·isk, G. E. lY., s. u. 1ta.n:-ra.l\l(,},


Nous 0,vons 6tudi0 cetto racine clans B. S. L. 66, 1971, p. 139-211. GO Cette intorprCtation nous semblo praf6rable it cello qui voit duns le vorbe le d0nomina.ti[ d'tm nom d'action du type de v. h. a. wani ~attente». 01 Voir Maydiofer, Etym. Wtb. d. Aiml., 11. 1i. Pour lat. aueO, voir \Valcle-Hofmann, L. E. 11'., s. u., et les r6serves d'Ernout-Meillot. 6•


Frano;ioise BADER

avoir ete dans un rapport de derivation *-i-f-m . .. archalque avec le neutre), v. sl. umu < *au-rrw- «inteJligence, entendement» 62 • Le degr8 zero *.iu- apparait dans le parfait (moyen) ved. uve,et le present hitt. 1ere p. sg. u"!JM, de sens «voh ». La 3eme p. sg. qui correspond a ce dernier est aus-zi. Elle est batie sur m1 radical 8largi I *iieu-s-, dont le th0me ll *.iw-e.s-peut apparaitre dans des formes nominales du vocabnlaire pastoral, en avestique et lrittite: substantifs av. vtistr3m «fourrage, nourriture», hitt. we8i- ((piture», avec son d8nominatif we§ija- «paltre (le betail)»; adjectif, nom du «berger», d8riv6 en *-tr-, av. v{istar-, hitt. weltara-63 (employ0 pour la protection humaine ou divine comme i. ir. plitdr-). Le meme theme est passe de l'id6e de «p!turei> a celle de «nourriture» (cf. av. vastrr;m), dans des noms et des verbes de langues diverses. L'on a,, en effet, un radical *wes- «manger» peut-8tre en Sanskrit, mais da,ns deux formes difficiles, rinu . . vtiva!.e R. V. 8,4,8, et vi, completed before 150 B. C. procluced homophony between the earlier III. 2 DS fine,i and III. 2 AbS finid; of. fini in fig. 1. 2. The loss of distinction between noun morphemes in all their paradigmatic representations. For instance, Latin like Sanskrit 1how1 no distinction between Dative Plural and Ablative Plural. Different patterns of syncretism elsewhere in I. E., e. g. in Greek, where DATIVE is signalled by Dative, ABLATIVE by Genitive in both Singular and Plurnl, suggest that the Latin situation is not inherited from P. I. E. except perhaps as a dialoctal feature. Moreover Latin itself before the changes specified in 1. above hacl a systematic disLinction between Dafave Singular and Ablative Singular in all pamdigms. So long as at, least one paradigm retained a distinction between DS and AbS, it is appropriate to treat not only Dative Singular and Ablative Singular but also Dative Plural and Ablative Plural as separate morphemes, even though the respective allomorphs of the lu.tter pair were homophonous in each paracligm. 3. The loss of all distinctive morphological representation of a given case function. Thus whereas Sanskrit and Lithuanian, for instance, retained a clistinctive Instrumental in Sil1gular and Plural, INSTRUMEN'l'AL is signalled in both numbers by Latin Ablative, Old h-ish Dative, etc. Various causes can be assigned to the above phenomena. 1 Thi, p•p•r i, indeb\ed tluoughout lo tho ,1and•rcl ar•mmai·• ot E. Bouroio,, C. H. Gra.ndgont •nd V. VU.nti.non, but to L. R. P~lmet•'• Lal4" La,.~1ag,, which fu .. t •~imula.tod my own inloorOII in \ho hi ■tory of Latin a.nd■ ono of tho tew work, that ,urvoy the entire di•ohmny ot tho L■.tm o••o- ■:r■ tom. Hi• theroforo a p ■.rti­ oul ■.t· plea,1ue to oont!'ibuh the prHon1 .t,ndy \o a. y0Juu11 in ProfeHox P•lm1x'1 honour. 2 P ■.racligm 011.1101■.re do■ignatod by Rom~n numou.:t■, moxphomH by the label■ oonvontionally Q.Hi1n1clto thorn in lobeara.mmau of thoir rHpooliYo la.nau•go■, tho morph■ oompri,in~ them by ihe iniiil'J ■ of 1.hoH la.b1l1: N(omin~loivo), V(ooaUY1), A(oou■•ti.. o), G(enitive), D(a.l,ive), L(oo~iiY&), .Ab(la.tin), l(m1\nun1ni~l) a.nd B(ingi.,l ■.r), P(hua.l). Gonclor i■ ignor&d throughout; 1oe p. 60 n. g, Tho ,om8ntio fl.old1 of morphemN ■.re noi ,pooifl1d, •ine• no o.-ti.1mpt i1 m•cl• hol'I to e;,cplol·&in terms ot "O .. o Grammar" ih• doep- ■ kuo~tU'e implio8\io1u or ~h• ,mfMe-,trnotul'I phenomena und&r di1onHion. Where pa.rloioul ■.;1• a.roM of \ho •em■.n\io flold1 ■.ro l'Olov■.ul, to \he ~rgumon\, ih&y Re clo■ign■.t&el in oapit•l• by th• oonY•ntiona.l lab1l1: P.4.RTITION, A.BL.A.TITITY eto. " A,1eri,k1 are omiH1d who1·ovel' iho morph in i,JUN1ion i• a.o\ua.lly a.Uested, even ihongh ih1 p■.rtioul~r oombin8iion of lexem& a.ncl morph i1 no\ r&eorded. ' 'l'ho Latin doolon1ion1 &l'I \ypifi&ii th1·ou1hont a, followe: I hina, II annus, TII. 1 (oon■ onant 1\em ■) paM,·, III. i (i-111m■) fi114,i1,IV Y dill.





Fir1tly llhonologioal changes, which, in as much as they opert1.te at the level of the allophone■ of which individual morphs are oompo1ed, .re Mpeoia.lly rele.,-..,nt to first degree syncretism. An accumulation of 1opar1.te phonologioa-1 ohl\ngel'I will moreover contribute to second and thhd degree syncretrnms; cf. the ex1.mple1 cited in 1. above. The inflections of languages like Latin and English with a strong stress accent on non-final syllables are especially prone to phonological iittrition. A priori a hierarchy of vulnerability can be posited, with -VO more durable than -'VOor- V, -V more durable than -V, etc. Morphological change■ oonkibute 1peoifloally to second and third degree synoretisms. They oon1i1t exohui,,-ely of • exten1ion1 of first degree syncretisms: for instance of the syncrctism of NP a.nd AP in one paradigm to NS and AS of the same paradigm or to NP and AP in others. Even the replacement of a gi-ven morph by one of its al\omorplrn can contl'ibute indirectly to synoretisms. Instances abound in the prehistory of Latin. The inherited I GS, preserved in the formula pater jamilias, was regularly replaced by II GS -i, beginning no doubt in such contexts as *due:ni nautiis ( > boni, nautiii) ij and *filii filiJ.sque (> fil(i)i filidique). The resultant -Oi > -ao,,homophonous with the reflexes of DLS -iii and NP -ai; whence Ulnae in fig. 1. I NP -ai II NP -oi were themselves analogical replacements of 'inhel'ited -iis, -i5s, which are directly reflected e. g. in Oscan aml Umbrian. Here the source is parbly llerhaps an inherited ii-stem dual in -ai, i-eflected e. g. in Sanskrit, but more importantly the inherited pronominal -ai, -vi occurring in such plu·ases as *istoi agr5s ( > *istoi agroi). The assimilation would have had specific motivation if the inherited AP -iins, -ons had already been replaced by -iis, -5s, as they were not in Oscan and Umbrian 7 : viz. istOs agrOs (A): *istoi agr5s (N) __,,__istOs agrO• (A): *istoi agroi (N). The resultant -oi became homophonous with II LS and, with the 1ub1equent change -oi > -ei > -i, also with II GS. Hence anni in fig. 1. The inherited III. 1 NP -ls, reflected c. g. in Oscan and Umbi-ian, was replaced by III. 2 NP · ii.s,an early instance of the interaction between these two paradigms, which is further exemplifiecl in the almost total 1:eplacemenh of III. 2 AS -im by III. 1 AS -em in classical Latin and tho encroachment of III. 1 AbS -e on III. 2 AbS 5 Hi• a oun·on• fa,hion fo trciat ,ui,h phi,nomom-..irnp1yH rule-ohang•• in •lie grammal,ioal •Y••cirnand •o ,ocik for oau,e• in •hci doepor ••ruo•uro of OM• Grammar. Bui analopoal forrna~ion• and r,-fol·ma•ion, •ro a woU-aUci••odf•••me of IUJ."fao,1kuoitno1 and mu•b b• :reoo~ni■•d a■ playiu~ an impm·•ant rolo in ph•nomena like 1ynor•ii1m. 1 ', {ha• ha, di■orodi-.ed analogical cixplanl'l•10n1 i■ -ih• faoilo \fl'IY in whioh th•y h,;i.v•bHn employed •o oonn•o• ho1nophonou, !orm1 in parl'ldigm1 withoni ai-i•nding io thoir 1yniaoiio rela"tion1hip•. To in..-oke analogy when one i1 una.ble "to poini io or oonjooiUl'e plan•ibl• context. in li..-ini •peeoh in which •he rolevan"t morph• •uflloienily olo ■-ly a.. oeia-i•d "to favotu· u,imil,;i.tion i■ io pro,,-ide no rea.l all. The d•Tiani m1101•1,1, of 11111111-, G"Pof t11i111i1, ol•rly repi:·Hent 1,naloiioal epiaraphio form■ 111u1r1oR, ore,;i.iion1:from V GP and II GP re1poo•iYely,a.nd we oan flncl an ,;i.ppropri,;i.te oonL•xi in •ho formula X a,-101·1o1n Y """'""" ~ di1nm1,•h•i i1 1i.anda.rdon funel'&l'J'in1odpi.ion1, iiven ihe unoertainiy as io whioh paradigmatic olaH r1t11W bolongod "to, M re,,-e,;i.lod by 1,hovada$ion beiw••n "oon:eoi" III. 1GP111,1111- ,ndi.hon,:r-iantIII. SM111nu,.._ in $h• lfBB of ol-ioal auihor■. 1 Of. "the replaoemon• of Groek I GB ,a,B with masculine lexemes by morphs modollod on II 08 -oo in mo•i di&l•ot.. T Wh•l'• "thel'•v•l'IO analoiY op•raied: pronominl'll NP -oi was replaoocl prehistorically by -Os.


of syncretism in Latin


-i, III. 1 AP -Mon III. 2 AP -kl (of.fi11tM, fi110,,Jint.l' in fig. 1). The assimilation here (> patriw is p&rUy due to, opera.ting in 1uoh plHa•o• as 'P'll di1tinot semantic field, to whioh both propo1ibion &nd ea.le contribute. In Li;i.tin by contrast, wit,h the exception of ,,._, -""b, SUJJera.nd ,mblo,1·( + Acouf!&ii,,-e,Abla-ti,,-e) every prepo1ilion is coupled with one case only. Aocn11-ti,,-cvri~h (l(l, propl~,· otc., Abla.tive 1..-ith ex, p1"0 elio. In such a situation the case morpheme, ha,,-Jng become ,conventional', is rodunda.nt 1ema.niioa.UJ'• In the exa.mple~ so far considered the rivalry is between a given moqlheme and preposition + the same morpheme. A second type of rivalry i1 between a giYen morpheme and preposition + a different morpheme. Thus pa,Tlbu• a.nd ptr pat,·,;, as exponents o~ IN,STRU.J\IIEN'rALITY, Genitive and ex or di+ Ablatin as exponcuhs of PARTI'rION, e. g. Unus miUtt•m, Unus ex rnilitibus, D&ti,,-ei;i.ndad+ Accu8 A .. imilaiion would ha.v• be•n ea1ier aOer -1i, > -i,, but i.hi• change i1 pl'ob1-bly-boo recent (early Snd oen$ury B. 0.) !ox•a.11trM'.lOIof -~i, •o ha.Telo•• from our da.ta,





exponents of INDIRECT OBJECT, e. g. cOnsuli n1i•,iii1'i,, ad cOMul,i:,a. ,.&11Wivit In each instance the prepositional phrase, originally di1i;inoD1emani;io...,lly but ha.ving come into a situation of rivalry, encroached steadily upon the ■im.ple oa1e morpheme. JN.g.1 1ea out the classical case 1y1tiem for nouns of the first three deolen1ion1. IV and V are omii;i;ed. They belong with a 1mall and unproduotiTe lexical group, many of which, e. g. launls, ,,, from our e1.rlie1i;records doublets in II and I. R..-en the most frequent member, of IV a-nd V wore in Vulgar La~in either 'kan1forred to o)her declensions, e. g, mano (II .AbS) for tM1'1i (IV AbS) etc., or replaced by synonyms from I and II, o. g. cmtsCi for ,·i,, di,'if'Jl.1'1for diiis. Neutern III. 1 NAS i;oo are disregarded. They had few distinctive form■ - II NS dOJ1.1on1, IM, III. 2 NAS ... are, II III NAP dBnn,·ia 1 - &nd were eTeni;.ul\1lyabsorbed ini;.oi;.he ,&nimai;e' paradigms of I and II, e. g. Jolia, uinu.s for f0Ii1o,n,ui,..1-111-. a1










a•ne VS a•••m.AB a••i

patrern,AS patrisGS patriD(L)S patre Ab(L)S 1mtresNVAP




a••OD.Ab8 lftniiAbS lUnii.sAP lftn{J,rimnGP lUnis DAb(L)P

annOs AP annOrwm GP annis DAb{L)P

patru1nGP patribus DAb(L}P



finesNVAP finisAP finiwm GP finib1w DAb(L)P

As the foregoing discussion has led us to expect, this system exhibits all three degrees of sy:ncretism, a,s compamd ·with reconstructed P. LE. Although the disappearance of Instrumental is the only instance of third degree syncretism, the maxi• mum number of morphological places has been Tcducecl .from the sixteen of P. LE. (N, V, A, G, D, L, Ab, I+ S/P) to nine, of ·which eight are occupiecl in Ill. 1, seven in the other paradigms. It is uncertain whether P. I. E. had Vocative Plural. Almost all I. R. lrmguage■ show NP in this function and the distinction in Old Irish o-1tem1, with VP Jr• < NP *uirOs, NP fir < pronominal NP -oi, looks like a local innon,tion. The corresponding 1mradigm in Latin (II) is the only one in which the language preserves even a distinctive VS, and it is surely not inelevant that this is the paradigm to which the masculine forms of personal and family names are assigned. So long 0,s anne survived it is legitimate to regard both Vocative Singular and Vocative Plural as morphemes, even though all their otl1er allomorphs are undistinctive. Voc&i;.i..,-e seems to have disappeared early in imperial Vulgar Latin - in contrast to Greek, where it still survives in a-stems - and will therefore be disregarded in what follow■. These morpl11 •te the only in■ll'l.noe■ of ov•d g•nd•t in Latin noun■. Gendet' i ■ of o•1.e1ory of thtl'll'l-lermination adj•oliYo■, •· i· d'Wnu nauta, d'Wra Ji111m1u, d'lirwM-111wlg11, - in lwo-lennin•i.ion adj•otiv-■ llfa■oulin• •nd Feminine are not d:i•iinguilh•d - and in non-p•r■ on•l ptonoun■, ill•, qui etc. 9

oom•■e I\ ■y■l•m•tio

Patterns of synorotism of Latin


Locative provides an interesting instance of wholesale first degree syncretism. In no paradigm has it a distinctive morpl1 and its occurrence is restricted to a small and unproductive group or lexemes: thus militiae but in pugna, Romae but in Italia, hurni but in solo, ri'i.re1 \ rftri but in agrO, Delphis but in .EpirO. Since no other function provicles a correspondence between GDS 111,ilitiae,GS h1~mi, AbS rftre, DAbP Delphis, it is again legitimate to reckon both Locative SingulaT and Locative Plural as morphemes in spite of the absence of any distinctive morphs. Since the rival exponents, in -1-Ablative and later arl Accusative, eventually prevailed and reflexes of 'locative' morplrn survived only in invariable place names in Romance, Locative too is disregarded henceforth. Ignoring V and L then, we can set up the following pattern of distinctiveness among the various noun-morphs:







s 1 •1 P 41 There is more distinctiveness in S than in P. The most marked morphs belong to Accusative Singular, Accusative Plural and Genitive Plural, the least marked to Dative Plmal and Ablative Plural, tlrnugh taken together they are well marked against other P morphs. ' The subsequent pattern of Latin syncretism is recovered from the diachrony of Vulgar Latin, the character of which is inferred from epigraphic and manuscript documents of the imperial and Mel'Ovingian periods. 'l'he diverse LaLinity of these documents is the producL of 01Jposing pressures. On the one hancl from the tradition of educated literary usage, as defined by the corpus of classical .Latin authors and codifiecl by the late im1mrial grammarians, notably Priscian. This tradition survived in a modified form 1-ight thl'Ough to the Middle Ages and Renaissance. On the other hand Lhere was the pressure from the spoken language, particularly that of illiterates - Vnlga1· Latin proper - which is ultimately reflected in the various Homa.nee languages. '.rhe criteria for selection of what is significant amidst the mass of vaTiants attested in the clocuments 11 are established by a comparison of the terminal points, the classical system of fig. 1n, attested in the u11.goof authors of the period c. 100 B. C. - c. 120 A. D., and the proto-Romance l'IJ■ tem of fig. 3, reconstructed by comparison of the attested Romance languages. It is convenient to analyse the Vulgar Latin evidence in terms of the three types of causation specified earlier, though of comse here too }Jl10nological, morphological and syntactic causes opera.tecl concuITently and frequently interacted. Two phonological changes are of particular importance. First the loss of -m. Inscriptions, especially from rural areas, indicate tha~ this was already widespread before 200 B. C. -m was restored in classical orthography, but it is clear from classical 10 DAbLB rwr• (< LS •-1 o!P. I. HI.) •nd DAbLP D•lf'hi• (< IP •-oi• and LP •-oi,w of P. I. HI.) may h•ve provided i.he .. adini poinh for ihe (unolionl\l enoroaohmen"lo! AblaliYe •nd p:repo■ition + Ablali..,-•on Loo•li..,-o. 11 Among the num•rou1 ooll,otion1 of m.ahx-ial, lho■- of P. A. G~ni, A_,._ Iu,;rwi,·~ fol.o Looal VarigJiio"• W\ Vwlr;a,·La/1" a, rlfl•o,,d ,:,._ ,111•ooalirM. of 0Jwi.tial4 if1.lon1.1tio"1 (Chap,l Hm fge8), P. N. PoIH.•or and B. L. Poli-•:i:, Bo,,._a,.o, 7th ""d 8th o,ntwl")' Lati,._do-i•"t. (Ch•pel Hill 1gis:,J •nd L. F. Ba■, :'.I'A•Nou" d•olo1Hio" o/ J,(,i•otoiNr;ia,._ £ajfo (Po,l'i■ 1g17) 11•..,-e been 1.aodprinoip•lly in ihi■ paper. H Bkiolly ■peaking, bhe diverieno• of Vulilil.l'and liiet•aie Latin must have begun in the pre.olanioal petiod, so J.hal fi1. 1 i■ in lb• n&iun or a t,n11-i"'"''ante quem.

t,-.,.Q, .,_




versification, whfoh regularly has -V..,#0- but -Vm#V- -+-V#V(with elision usually of the first vowel), that.. even in educated speech prevocalic -'/Jt was seldom pronounced. The early Latin situation was continued Jn VulgaT Latin, ancl on inscriptions from the early centuries of the Empire onward -m is increasingly omitted or incorrectly addedl 3 . The second change is the neutralization of the earliel' phonemic opposition of vowel-length. This resulted in the Vulgar Latin convergence of a and ii, and in most areas of e and 2 at {:or i - in contrast to ~ ( < e) and f (< i) - and of O and i1, •~ 9 or 1J,- in contrast to Q (< 0) and 11-{< U)14 . Although the early appearance of thi, group of vo,vcl-changes in the Pompeian graffiti may be due to Oscan interference, it is, like the loss of -m, abundantly tittested in insc1'iptions from all aTeas in the later centuries of the Empire. The effects of these changes can be seen in a number of first dogrM 1ynor•,i•m": (a) I AS lUnitm, AbS tuna both > l'l!,na,homophonous with NS ( < zv,.,,'1). (b) III AS pi'i,trbn >patr~, homophonous with AbS (ann1i8, subsequently homophonous with .AP ann9s (< 1InniJs). (d) ll AS in•i,n > o:1"-•9,homophonous with DAbS ( < iinnO). While annus is :frequenlly Wl"itten for APl5, the reciprocal spelling annos for NS is exti·emely rare. Thi ■ suggests that the h•o morphs converged at -1l8 l'lil.lher than -9sl 6 • Moreover II AS often has -o for -1'f)l. bul very rarely -am or -u. Bo it looks as if the .reflex of classical U and 8 in the"e morphs tended to become 9 when unchecked,¥ when checked 17 • A further problem arises with the reflex of -0. Although -u is very rare, it is found almost exclusively in AbS. Gaeng pp. 193-209 records 1463 -o, 45 -u in II AbS but l 3 -, 11al.10ot'ien orniU•d on i111oi·iplion1 before SOOB. 0., and pr•-ollil.■■ieal yer1ifloation, whioh rogu.larlyhH -V1!1' V-, admil1 -V•#-0•-+ -Y,;1,0- •• ,,-oll M lho commoner -V,,,,.0-. In oonll'l'l■t i.o _,., th• J·•1tor•tion of -,, g•n•rali••d from provoo•lio po1ilion, i1 at\1docl in both olauiol'll oriho1l'aphy and Yer1ifloaiion, and H0ln■ io h•-,,o Tofieoted tho 1ituai.ion in ,ho 1pokon lan!u&g•. Ji"Ol' lh• omi11ion Ol' inoonoo1. inHdion of -s, lhough found in dooumont, from all ov•r Lh• Lalin-■pouldng world, i1 rat·•ly more thtm 1porndio. Elv•n in H•ly, wh•t·e -• was • lo■t al1.og1i.h1x,i.ho inoro•1in1 number of omi11ion1 in JdoToTingian doomn•n•• (100Polit■or l'llld Polih.•r p. 13) inoludo■ in1bnoe1 whox·omo1·pholo1io•l lll well M phonolojioe.l. o•u101 mu1t b• reokon•d ,,-ii.b, •· g. qvi for 11ui,, t)Hdio,tat■ fol' ,.,at,ta,,ne:,,01,for ",po,. u The -,,o,,-•l1y1t•rn1 of Bioilicmand Boui.h•rn ltlil.lian dialo11ii1 oan probl'lhly b• d•d,,o,l from lhi■ pati1rn, bul .B•i:dinian pro1uppo101 i < >%, e < A, 8t1.rdinian l'lnd Rumanian o < {5, u < f, .,,,hioh ol••i:ly o•nnoi. be 10 de1·iv•d. Yor lho implication, of i.h• llil.Uorfor lho oluonology of Yl1lja1:Latin ■ee '1.'.P. B. 1Q71 p. 170 n. 8. l 1 G•eng'• labl11, pp. 1QQ-200, ■how iota\1 or :!JOO -o,, 18Q-••• .,,,m1 lho lai.t•r inor•a■in8 in l•tor in1oripilion1 of •ll i:•1ion1 oxo,pt Spain. F'l"om lt•ly Polihor D-ndPolii.■ or, p. i7, rooord 11H -o,, 7S -••· 11 The alhrnativo oxplanD-tion, that i.h• 1pollini am1.1Uin AP upn■-nt■ a"""' < dnoW,, by a.n,JolilYwii.h IV AP tllld,.,ii,, i.houjh pou1blo, •eem1 1011lik•ly, jiv•n ihat th• r11ing frequ•noy of -"WI horo ooinoid11 \fith 1.h• d•olin• of IV. ,;,aam',tllOft0I al'• commonly founcl for ola11io ■l Jl'tanii1 (NAP) and 1.hooha1·aoL• IV GB -us ( < -'W1)i1 V.PY rarely •ilte■ted (Ga,eng _pp.ui,-7). · 1T No oornpari1on i1 po11ibl• with l"tfi•xH of -i, and •Mt, 1iuo• the 1a~~•r blil.clbeen almo1t -totally 1·opl•ood by -em •lr111dy in ih• 0Jg,11ioalperiod. How•v•l' it i1 reloTanl ihat prehil'.torio •·• > •••e.g. -.ar1,/o:oii,, lhal -i tondod. to a high•r artioulalion in olc.a■ i­ oal. Latin,•· g. ll11"i,qwg,ii, ,ibi wi.ih ·• (••• T. F. B. 1962, pp. 80-8:!I) and 'tb•t in It.lit1.n

-i,> -i> -~M





of syncretism in Latin


only two instances of -u in II DS, both from Rome. annu is unlikely to have been due to analogy with IV AbB manu (cf. p. 52 n. 16), and analogy with other paradigms, e. g. patrQ(1n) : pat1·'i, lvna(m) : l1ina, can only have operated as a kind of hyperconection in conservative l'egisLers where -m was retained (viz. to produce Mwi¥m : annu) since annum was generally replaced not by "1!but by 9 18 • In fact AbS -u is probably [I, spelling contamination of -o by -um employed as an Ablative at a time when both were prnnounced 9. The totals from Politzer and Politzer's figures for Italy (p. 17) are instructive here:

II AS 1474 37 753 DS 26 323 AbS 314 32 2914 The table rnveals incidentally tlie overwhelming frequency of -o, viz. -9, in all three morphs. (e) III. 1 GS pRtrrs and J\TPp5tre8 > patris {f) III. 2 NGS fin!fs and NAP fines > finis. The spellings patres for GS, fines for NS, though much rnrer than patris, finis, are relatiYely morn .frequent than -as for -us in II NS 19 • 'l'his may be due to the inrruence of e elsewhere in these paradigms, viz. AAbS paii'Q, ffn.Q; but the numbers are sufficiently small to be accounted for by occasional uncertainty in representing 'f:.In · NAP however both i and e spellings are common 20 • This is not entii-ely due to phonological causes. For •is may represent either -is or -js. The classical system shows especially in imperial documonts the analogical extension of AP -es from III. l bo Ill. 2: NP patr&: AJ? patres ->- NP fines : AP fines beside earlier finis. Other analogical patterns between these four morphs would have given NP patri!.s,AP patris or NAP finis. Reflexes of the former would have produced Vulgar Latin patrjs patrjs, of the latter finis, fjnjs, and some such developments surely lie behind the growing frequency of -is spellings in III NAP, and also of -ium in III. 1 GP in Gaul and Italy, patrium fur pat:rum, etc. 'l'he changes specified in (a)-(f) would have produced the situation set out in fig. 2. This :is of course a purely artificial construct, embodying as it does - ·with the exception of tl1e bracketed items - only the results of phonological changes. '!'ho extent of first degree syneretism is nevertheless -very striking. Apart from the NAP allomorphs patrjs, fin1:s discussed abo-ve, the number of moqJhological places has been reduced to a maximum of six, of which III. 2 fills four, the remaining paradigms five each. In the singular NS is distinctively marked only in III. The survi-val of GP as a well-marked morpheme is clearly due to the resistant properties of the dissyllabic lij AS -wm is st-tll writton in the vast majol·iLy of Ga.,ng's insol'iptions (pp. 223-224), but -o is 1-ecorcfodas a frequent variant later on in Gaul (Sas p. 507) and in Politzer and Politzer p. 17 appears as the most frequent spelling in l\foroYingian Italy. l 9 Gaeng's figures (pp. 162--3) reveal totals of !l9 -is in NS, 207 in GS; 17 -es in NS, 18 in GS. .Politzer and Politzer do not distinguish -es and -is in NS, but their tables for GS show 89!) -is, 41 -es. 20 In Ga.eng's material III NP has -es normally, •is very rarely (pp, 135-6); ID. 1 AP has 137 -es, 52 -is, Ill. 2 AP 105 -es, 58 -is (pp. 135-6, 177-8). Politzel' and Politzer's tables show Ill NP 74 •BS, 87 -is; Ill AP (Ill. 1 and Ill. 2 are not listed separately) 88 -es, 114 -is. Sas's figures (pp. 515-0) show a sharp increase in III NAP -is in the 71,hand early 8th conl,uries, thereafter a swing back to -es, which may refloot a reaction in favour of classical forms, at least in odhography.




morphs -ar9, -9r9, -i9, Although DAbP is well-marked against other morphemes, the absence of distinctive morphs for eitl1er DS or AbS - in contrast to the classico,l fig. 2


III. 1



ann"llS NS, AP

finis NGS, NAJ?

l'lfnQ GDS, NP

annoADAbS ann§ GS, NP

pat€rNS patr,;,AAbS patris GS, NAP (patrfoNAP)

annoro GP

patr9 (patrft;i) GP




ff.nib9s DAbP

l1J,nasAP l1,1,nar9 GP l1fnfsDAbP



system - justifies the conclusion that Dative and Ablative have now coalesced in a third degree syncretism over against Nominative, Accusative and Genitive. ,ve have now to consider the changes that, as it were, bridge the gap between tllis four-case system and the two-case one that is set up for proto-Romance, as exhibited in fig. 3. fig. 3



III. 1

ann'lf,SNS, OP an119OS annf NP

pat~r NS finis NS, NOP jjn~ OS patn, OS patris (patrjs) NOP (finis OP)

*0 =

III. 2

Oblique, viz. non-Nominative

Phonologically the only difference from fig. 2 is the replacement of -I,! by ?-, as assumed for unaccented position generally in proto-Romance. 'fhe disappearance of distinctive Genitive and Dative-Ablative can be assignecl to syntactic causes. The rivah·y of ad + Accusative with Dative in Classical Latin has ah:eady been remarked (above p. 49f.), and the growing encroachment of the former upon the latter is abundantly attested in Vulgar Latin 21 ; e. g. ad iwni dfotas basilecas ... relinco (Sas p. 85), episcopus dixit ad eos (p. 174). Especially noteworthy is the frequent concord of Accusative and Dative 22 , donamus uilla nostra ... ad sacrosanctam ecdesiae (p. 37), ad heredis meos propinquis reseruaui (p. 176) etc. If the functional distinctiveness of Dative was undermined by Accusative _phrases, Ablative, which, starting from Genitive was equally under pressure from de, ex the PARTITIVE functions noted on p. 4.-ll,were extended to POSSESSIV.E and other genitival uses. Thus alongside ,...z,ie•i"' de his mm sunt inuenti (Sas p. 169)


21 'l'he examples are all taken from those cited in Sas, but can be paralleled in the material discussed by Gaeng and Politzer and Politzer. n Which renders it impossiblo to determine whether e. g. in ad ornturio ... dari uolo the form oraturio should be labelled as DS or AS.

Patterns of syncretism in Latin


we find si quis de diuersis uenatio~ibus furtum fecerit (p. 81) and per aliqua temeridate aut caliditate de successoribus ,i.01tri1(p. 169). These developments clearly reinforced the syntactic role of Ablative, bu• only in its prepositional usage. Indeed non-pre -positional uses were themselves replaced increasingly by prepositional ones (cf. p. 49), e. g. in lNSTRUMENTAL function si quis ltlterum de sagitia toz"7ata pe,·c-..t"r, uoluerit (Lex Sal. 17. 2) and in COMPARA'l'IVE function as early as Ted. 6 aliquid lle vnttatione felicius. Such developments clearly rendered the case-distinction between prep. + Ablative ancl prep. + Accusative increasingly redundant. Hence cum boue.sduos (Sas p. 194-),pro furta (p. 196) and in the area where preposipa,·111 d" predictas tional Ablative had already encroached on Genitive tp,a., ci'III.IU (p. 82). This encroachment of Accusative on Ablative wu of com1e further assisted by the reduction o( I AAbS to l1},na,II AAbS to ann9 and III AAbS to 1ialr~ (ancl fin?), so that it is impossible to say whether o. g. summilates de ligno 1a1tcto (Sas p. 106) contains Ablative or Accusative with (le. Wibh the erosion of their characteristic morphemic function the durability of the morphs of Genitive and Dative-Ablative, as exhibited in fig. 3, afforded no protection. Even the remaining distinction of Nominative v. Oblique subsequently disappeared. Although this is foreshadowed in the .first degree syncretism of I NOS and III NOP in proto-Romance, it properly belongs to the independent diachrony of the Romance languages. Thus in Italy for instance the loss of -s, already well-advanced in cent..ral districts by the late 8th century (Politzer and Politze1· p. 33), led eventually to homophony between II NS, OP ann'!_f(s)and OS ann9 and between I NOS l'lfna and OP l'(tna(s). This left NP l'lfn\>,annf as the only distinctively plural morphs in these paradigms; whence the Italian plurals lune, anni 23 • In IH NOP the effect of loss of -s depends upon the relative status of patris, finis and patrfs, finis (see above p. 53). If the former were pt·edominant, then t?e convergence of i- and?· would have produced homophony between S ancl P, vrn. patr?, fin?, and a new plural formation, reflected in padri, fini (v. paclre, fine), providecl by extension from II anni 24• If, as seems more likely, patrjs andfinfs became predominant, then the loss of -s leads directly to the Italian forms 26 • Finally in the Singular what appea1·s to be a largo-scale absorption of JIL l. nouns into tl1e III. 2 paradigm in Italian - padre, padri : fine, fini - may l)e due either to the remodelling of NS pal?r to patri_s etc. under the inJluenco offfnis 01· to the analogical extension of the first degree syncretism reflected in NO.P patrfs (or patris) to the Singular, resulting in the replacement of NS pat13rby OS pat1'?, Indeed Polibzer and Politzer's tables (pp. 17-18, 21-22) reveal a far greater encroacl1ment of AS on NS in III than in II:


"" R. L. Politzer's earlier theory, Rom. Rev. 38 (1947) pp. 159-66, that tune< l,unes, a contalllination of lune and lunas is less p1•obablo. For although liines (and perhaps partly also annis for awni, which is more frequently. attested) woul.d indeed be_explico.ble thus, I NP -es is too rarely attested and too restr10ted geogi·aplucally even Ill the Ia.te 8th century to have become the dominant allomorph which this theory assumes. •~ This is of course the situation in Fl·ench; Vulgar Latin NP mentis or rnentjs, which is i·efiocteclin Old French OP mo-nz, was l'Oplaced by mentj {> OFr. mont, of. II NP mur v bctrifft, ha,t &ber A. 8ol1erer• darauf aufmerksam gomacht, cb.13 die Enclnng - wie es bci isoliertcn fremden Eigenn•men keine Seltenheit 7 ist nach den :.;ahlreichen cchtgricchischen Ortsnamen und Appella.ti-,,en auf-11?kilvumgesta!Let sein k6nnte, so daJ3 auf cliesen Zeugen kein Verla.B ist. W &ml der Kult cler Kubaba - walnscheinlich iibe1· Lydien, wo er schon zu Beginn do1 6. Jh. inschriftlich belegt ist - in Jon.ien eingefi.i.hrt wurcle, wissen wir ni.oht: man wird wohl Laroche zugeben, da.13sichere lnclizicn zugunsten einer friiheren Anwe1cnheit der G6ttin auf clem iigii,ischen Ki.istcnland fel1len, lloch spricht auch nichts for die gegenteilige Annahme 8 • Der AussagewerL clieses Lehnwortes bleibt also im ganzcn fragwiirdig. 6 Kr1ti,oho B••ll1mgnahme zu Laroches 'These von V. Pisani uncl E. Salvaneschi in P,,idoia sg (1Q7,), 100 mi• Anm. L 1 ltr•tylo, 18 (1071}), 1U. 'Vgl. Verf., A1p0Ui del pr••tito hngnidieo (N•poli 197~), 4,6. s Obwohl m"n die phrygi,oho Kybele in iUerer Zoi, nioht. ohn• wei,•n• mi• Kul3-ijl3'1) gleioh,e•ii:en d~Tf (10 im Prm,ip rioh•ig l!'. Oa1101,.,, Inni omoi·iei, lrn•no J071}, :J27f.), ha.,te LaToohom. B. dem a1-eh•i•ch•n B•l•i von Qul3aA«,;au, Locri IClpi:i.,phyriiR,ohnung ..11.. in.,ndoneb:,n mi't1Hn, di• n"oh H. Gu.a.rduooi i.T•gen uncl ,ioh mi• d•u .A.i·gumen1.•n (Klio lit, 1070, 133ff.} die .Anwe.. nh•il de■ •ni.,pl·eohend•n Kuh■ -.n der ...,.. o:l•in•i11.ti1ohen Ki.'l,.e ,pitH.0111 :i.nBoginn dH 7. Jh. na.h,l•i•n. (D .. vrfr d•n i0n ..uon Wog, dul,:ih aus den dor,n•mo Loori orreioh, ho.t, nioh• k•nn•n, bl•ib~ dio F'r•1•, ob Qul31;1;l.•~


Roberto GusMANI

Dagegen ist die chronologische Bestimmung der Entlelumng des Medernamens durch die Jonier sicher richtig, lmd Laroches Kritik an den Hypothesen mancher Sprachhistoriker, die mit einer (geschichtlioh unmOglichen) Bekanntschaft dcr Meeler im Westen vor dem 7. Jh. (ja sogar im 10. Jh.) rechnen, trifft wohl ins Schwarze 9. Den lVt.ndel et>·.,,••iii illni1•n• &uoh du wohl m1gef&hr urn die gloioho Zoii •n\. > Il,icpao:i, woboi di0Kt'u:1nmi doa 11 ( dio Boherer l'I.. a. 0. di"Analogio yon Ri11nnam1n wie Il,po,U~ Yert,ni.worhlioh rnaohon Liquid• + mOohf..1)auoh d.-..duroh 11rltliil·t,,..,rdon kOnnh, dal3 die Folg, .,ltmg11rVokal Kon1onant" infol~e d!'I•O ■•ho:!'r■ oh•n Go■-l.■ !'11 ung!'lwi:ihnlil'lhgl'lword!'lnw■..r. In di-■l'lm Zuu;nnn!'lnh•ng ■!'Ii horvorgohoben, d•6 dt.■ on in V■rbindung mi• Iltpc:rai und mi-Ld■r ent■prooh!'lnden B!'lh ■.ndluni von ir•n. a •rwiihn•e 8itp~,i~ in Wirldiohk!'li• ■.ndeu l1lmt& N1.m11 der Pora1i: (iu,n. Par,a): Ygl. •IT-i,poi:u


•u b•u1·l.eil•n i■•: du• in letd ■rom Wol·I i ■t nieh• •l• Klll'■ ung ein•■ •u■ ir•n. ll niohh au11agen, -wiec• wei•n d■l' Z,i, dor Bn•l■lmu:ng 1ohon olmehin wahucheinlioh war. Andererseits darf man aber nicht vergcssen, daB rler fragliche Wandel schon in Metade1: zweiben Hiilftc des 7. Jh. inscluiftlich belegt ist 11 und der quantitativen these notwcndigerweise vorangeht, da z. B. 'l)O > ew auc11 die aus ifo entstandenen 'l)O bctrifft. Ferner war er scl1on zu Ende, als durch Kontraktion (z. B. o::e> &:)oder infolge der Vereinfachung gewisser Konsonantengruppen (wie sekundiires -vcr- bzw. -v,; im Wortausgang) neue ix entstandcn, weil diese sich unverandert erhalten haben. Selbstversti.indlich wurclc nach der Entstehuug des neuen ix ein fremdes /a/ getreu 'l(J"'t"licr7TTJ,; DlirayavauS, ViStifapa usw.) entnehmen kann 12 . AuBerdem jst zn beachten, daB die Entwicklung a >)] eigentlich in zwei Shufen analysiert werdcn muB, die auch dul'Ch geraume Zeit getrennt sein ki:imrten: die Palatalisierung des urspriinglichen /ii/zu /&/und das Zusammenfallen des dadurch entstandenen Phonems mit dem alten/G/. DaB es sich um einaudernichtimplizierende, chronologisch differenziertc Vorgii.nge handelt, bew.ies schon die bekannte Tatsache,

mit de1n nordw11•grioohi1ohen und eli ■ehen Vl•ndel e > • vor Liquida entdanden i■t [Ou,ola] od■l' ob d•• ■l,.,■ o; dir•k• au( di• L1.u•ung de■ Proto•yp• K•baba smii.okgeht, vor■r■• nooh o!f■n. lL B, lt1,1m ma.n jed1nft1.ll1 den Namen der ICu~ii).-i, Ton K•baba nioM •l·•nn■n.) Zmn Grt1.ffl•oKwt1a•[ t1.u1Bt1.r•di• yg}. Ver!., Neul'I epiohori,ohe Boh.rifheuini111 ,;m, S1.t'di1(Oiiombl"idg■ 1'171'1),iB. • Tret'l"ond• Kri•ik 1ohon in d■t· 81ellnngnt1.hmeTon 0. S:riemer,nyi, Btudien :rim Spraohwi11en1oh•f• und Kull,mknnd■, Ged.enlllOhrifi Br•nclendein (Inn■bruok 1~0B), 14.6, de.r •llerding1 di!'I Heren1iehuni vcn H~-'or b■i d■r D••i■l'ong van ci > l"J al• ein ,.■p■eiou■ -.rgumen•, oon■•an•Iy •nd inexou■ably t111clin •hi■ oon•■xf' b■Hiolmet ( ■. w. u.). 11 Zur Au1■pl·aohe Von ap•r•. a •.- ygJ. 1;1;r.Br•ndend■in------11. ~-'°'yrhofer, Handbuoh des AHper■i ■ohen (Wie1badon 1Q04.),1B. 11 Vgl. L. H. J•ffecy, Th■ Local Soripl.• of Aroht1.ioGre■ e■ (Oxford 106J), 71 und 3~8, A. Bar,onek, Devolopm■n• of ~h, Loni-Vow•l 671,om in .J..no1•n-L Gl'Hk Dit1.l■o• (Brno 1Qetl),QQ:tr". 11 Bei den N ■.men d,r il••1-.n Ki:inii• •.Ap,i;,paiJ,VlJoUcnXI .1.14, 175, (also XVII 63), 'Oi>ucri::IJi; OoupLx),uT6no .&uµopo:'Ccn€wv XVI 591), 1rro: (',o&v/ve&'i0p.&oxpw.p&.wv XVD.l 3, 573, XIX 344, (also VIII 23.1), ~fAwv &.x&.p,a XVIII 239, 484, 0AEil·pwvwith ~µo:.p XIX 294, ,109, noatv To:xfwcr, or c:crcn XX 189, XXI 564, XXII 8, 173, 230, 0ai:xTo:.µfvwvo:.l(l)u,vXXI 146, 301, 86-rpix.o:i; tmroure•n o"lhnio■ ("lhou1h no"I plao ■ -nam•■) and 'ool11eto1•" name, in e. li■"I of workgro1tp1 1ee o. g. Lo 0154.{h. 103); 1.t J~fjo, L! G·1tli-n,i-1i-ja, 1.4...,.,o•i•ja, 1.5 j1,;.ni-ja, 1.1'!,,. ......... jo.

Linear B a-ko-ra-ja/-jo


be eatablinhed; but it is clearly an ati.raoli.Ye po11ibilily that he i1 in 1ome 1en1e lhe o"·ner of iho two individuals conoern1d on i,he 1·eoord. Not only, ho,YeYer, i■ B 798, in lho ■ hand &11 the Ooreoordn, ali■t of 'oolleotora' · one of the 'oollootou' named on the rooord i ■ a\10 a 'collector' on the Lo (SI)tl~blet1. As we 1aw earlier, it is oxt:rnmely tempLing to reabore the name o:( the 'oolleotor' at the beginning of Lo (2) 7'!,77 1\1 i•l'5·]1"~•1·, and ;,.,, . ..,,.,·i--.;o i1 al■o, a1 we haTe jnn~ "een, one of ihe 'oolleotou' on B 798. Clearly, this evidence is well in line with the Tie"r th&t the Oo reoord■ Ji1t the 'collector' flock■ from which the Lo (i) 'oolleotor' wor:kgroup1 derived lhei.J' wool. Does in fe.ct B 798 li1i ■ome or all of tho 'oollector■' or 'm,rner1' whose fl.ookn are listed on the Oo reoord1 (and where the identity of iheae as 'collector' animals i1 indioated by a-"J,o.fa.jcr./-jo)! do-ro-jo below it cannot


Phonologisches zum indogermanischen a- Vokalismus Es ist bokannt, daB das von den Junggrammatikern aufgestellte Schema des idg. Ablauts zusammen mit der sog. Laryngaltheorie die Ansichten i.iber das idg. Vokalsystem stark bceinfluBt hat. Das ursprtinglich postulicrte Vokaldreieck mit seinen ftinf kurzen und langen Vollvokalcn uncl dem Schwundsi,ufcnvokal (ol) is-L von manchen Forschern (wie Hjelmslev oder Borgstrom) auf einen Vokal recluziert warden (wie man ilm bezeiclmet, e oder ii, ist nebensii.chlich). Bei einer solchen extremen Armut an Vokalen fanden es andere Forscher ("wieMartinet, Puhvel, Admdos) &nge1eigt, die Anzahl der hy1)othetischen Konsonanten, d. h. der Laryugale, auf 1eol11,1,,irij\f usw. zu stcigern. Die SchluJ3fo1gerung, daJl es im ldg. einmal (wann, ,,-ird nioht angegeben) bloB eincn Vokal gegeben hat, basier1, auffolgendem Gedankengang: 1) Die meisten o sind o) mit dem Ablaut zusammcngeworfon. Letzterer ist eine morphonologische Erschei.Jrnng, die erst clurch cine zugnmde liegende phonologische Erscheinung erklii.rt wcrclcn muB. Um hier gelegentlich eine Parallele anzuf-iihren: der dentscbe Umlaut, der im Ahd. cine phonei,iscl1 bediugte Erscheinung war (i, .i in der. nii.chsten Silbe), ist heutzutage ein Ablaut (d. h. eine morphonologische Tatsache). So ist im Komp. ktiU-er nicht die Lautgruppe -er, sondern das Kompai-ativsuffix -er als himeichende Bedingnng des Dbcrgangs a > ii zu bctrachten. Dagegen zieht z. B. clas -er des Norn. Sg. Mask. kaUer(Wind) keinen Umlaut nach sich. Die Rekonstruktion eines vorge1ohiohtliohen 8pr&cl11u1t~nda darf nicht, wenn sie ICrfordorni1 einea gew1Men Gr&de1 der W•ln·1chein1iohkoit (und nicht blol3 der Mi:igliohkeib)er!iillon 10\l, &llsuweit i'tber die ~e•ohiohtlioh gegebenen Fakten hinausgehim. lf•n d&rf nioht i1oliorle Einzelheiten il'I- infiniturn znri.lckvetfolgen, sonclern muB 1ic unmittelb•r •n ge1ohiohtliohe Znsammenhii,ngo anlmiipfen k6nnen. Die ~ngefiihrton P:rii.mi111111en 1tellen tlagegen chronologisch nicht georclneto Ann•hmen d&r, e.u1 denen sich keine entsprechcncle Reihe von U-ufeimmderfolgenden 8pnohsu1tinden (Sprachsystemen) ableiten Die Pdimissen sollen Beweis" grilnde d&f-iir1ein, daB das Vokalsystem ldg. (oder chcr des Protoidg.) stets im





Zuwachs bcgriffen war, bewcisen jedoch 1licht, da.13i.n•der Zwisohenzeit keine Vokal~ phoneme, etwa >(si ce type d'offrande etait la manifestation normale de reponse a une faveur ilivine accordee, par opposition aux offrandes qui ap1mllent unefaveur). I) En marge de OeOe

un problElme de signification (§ 6), un probleme d'Cty:mologie · (§ 7), un problElrue d'origine historique (§ 8), cc dernier. concernant aussi Oexrxv-re:µ.. Au formulaire gaulois, 0. S. confronte, quant a la dime votive, le formulaire latin et le formulail'e grec. a) En latin, formulations a verbo banal «donner» (tiedit, dal, etc.) 1.11ooiee1t, un 1.boll facMe note l'operation (de parhago) pr~alable a l'offrande latif absolu decumetfacUi,, elle-m&me. b) Lo grec a, on le sait, &va0eC:vix~ comrue verbe d'offrande universe!, 1\.des oent1.ine1 d'exemplaires. On le trouve aussi dans les quelques inscl'iptions vo,iTel oU la 0"1]crrxup6v, tnrco,;;, etc.) est prCcide1ign1.tion de l'offrande a l'accusatif (&y&.Aµ.rx"Ta;, s8e par l'apposition Oe:x&.'TT/v ); l'objet pout 0tre sous-entendu da.111«donner (une offrande») mais ne pouri:ait raisonnablement l'0tre dt1,nsi;;(>eest clone clairement un verbe d'o1frande. S6mantiquement, pM' consequent, il pe11t !'eleve1·de *clO-aussi bien au mains que de *dhe-. C'est a, I'etude morphologique seule (§ 7) a decider de 1'6tymologie. 7. Pour 0. S., parfait redouble 3°sg.i. e. *clhe-dhii-e> *dhedhe > gaul. d&U, paral16le a i. e. *de-dO-e> *dedO> Iepont. ded11,,ecrit 'fETU 14 • Ceci implique, bien entendu, pour la fi.nalede3e3e, un -eherite conserve, et non forme en -i (cf. noi,e 11). 6. Ici se posent:

pen■ able


Voll'n. 80.



.Argumen1 non produH par 0. B.; il qu&i,gri.,&t (adv1rbo) pour ~po;TOU,

favori ■e

d'aillours plus «ex noto» (substantif en

u In1orip,ion de Prodino. Nou ■ ,el'iYiou■ cl1.n1L1J10Mro, 1071, p. 100:: 1L'interp1·6ta:l.ion de 'J'BTU oomme d~du, & m1Ure en rappor\ avoe 1kr. dad.oV ou(iodhaV, e■• du& & A. L. Pro1dooimi. s,man•iquement, lo ehoix &1•inchff',r■n• 1n•r&'"d6- e1,•d.lii- (tdMliii! ,po•uih !). II le 11ra1t o.i.■ i pour la forme ,i l'on ■uppo ■ ai• un 1,p. ••cl,clii i11u d1

•d•-d(, 1 )-,.. ou do •&71,-dh(•ll-,..a.v•o d,ir6 x;,ro 1·a.dio1.l;mo.i1 les r,r,x·onoe1 oompa.rarenvoien• & de■ form&I i\ dO@;l,. -plein de la l'&oine (•do.dci-t., •a1i,.dhi,-,-.) ••• &noo cas, o'e■\ t. padh- de •-Ot. qu• la jn ■ lifloation d1 1,p. -t. e•• le plu■ 1.116-.~- Mai ■ O. B. {note 77) oon1id6ro -u (malsr4i le■ fa1i1 11,~in■ 1) omnme une par•ieulR'i._ de l'indien.




0. 8. &carte l'appartcnance de 8e8i::a *dO-, apparemment pour Jes deux raisons ■uiv&n•e•: 1. Exclusion d'uu parfait **cle-d( -nest acquis avant les premiers textes, rien n'oblige penser qu'il leur soit n6cessairement partout anMl'ieur de beaucoup. c) Sur le terrain, Jes isoglosses -1Jf/-N 51 suivent les limites des langues; beaucoup plus 1·aremont, SlU' le domaine cl'une langue, l'isoglosso dessine, comme en celtibCress, une frontiern dialectale. d) L'observation (c) est cependant mise en d0faut, - et peut-&tre aussi, par voie de consequence, }'observation (b) - clans un cas comme celui du venete 5~; nasale finale -n sur tous Jes sites qui nous fournissent des exemples, sauf un (L,\gole); 8, L&-gole, pendant une p0riode assez longue, coexistence de -1n et de -n (dono.1n./ tfoM.•. >;etc.). Aucune distribution ge◊graphique; aucune distribution chronologique nette; concurrence de deux traditions. Noter, d'ailleurs, qne la g8n0ralisation de -n dans le reste du domaiue, au Mmoignage de textes archaiques (nietlon a Lozzo Atestino vers 500) est anMxieure de plus de deux siCclcs aux preroieres d8dicaces de Lfi.gole. Situation confuse, dont it n'y a pas d'explication 6vidonte. Pcut81..reexemple privilCgi0 d'une neutralisation in fieri (tradition -llf archa'isante, tradide -llf Lagole (beaution -N novatrice, en coexistence), sans que l'attardement coup plus tard que dans le reste de la VCn8tie) soit assignable a un factem que nous puissions clairement reconnaitre 60 • Ces conditions g8n6rales rap1)elecs, venous-en au celtique. En celtique insulairo (gaelique et brittonique), -nest soit postule soit (en sand11i) conserve. En Iepontique, -mG 1 ; en oeltibere oriental, -m, mais -n en oeltibCre occidental 6 2• En gaulois (dossier ~pwrou excepte), partout .n 63 • Dans le dossiel' ~pcx-rou,tant6t -in (4., Organ, RG; 7, MalaucCne,RG), tant6t-n (1, 3, 3 bis, N'.i:mes RD; 2, Collias, RD; 10, St. R8my, RG).




67 Par con-,,•niion, .}ff signifie ici clnalite /-m/ f-n/ oonserv0e, .N signifie neutTa\i. 1aiion &u pl'o.lUdo •'14 • .. Oarbo1 do 1·6pariliiion (-N & l'Onc••• -M '- l'ICti); inox1oeilo•t p6rim6o, dan■ Oll1ib11·ioa, p. 4.J; mi■o jo=, dan1 ORAI 1~73, p. flli. H Of. Jie11'1-w•l tl• fo lamgue, 1'it1'il1, 1974., I 158 et (sm Ja double tradition 1,\golienne) p. JOO,-,,, 00 Non d6monilrabl• (faui• do t•xto■ o•Uiqu• dn hau• Pia-,,•) -■• Ja ■:ugg•lion quo nou ■ &Yon■ fai•o d• l'ao•ion cl'tu, ad■lrai, an OM oU lH O,lio■ OatubriiH do la r6iion do LLiolo &lll'aion• parl6 un dialooi.o -1:f. 11 AOrnan.■ ,o (PID.104.;Lopantfoa, p. 74.n.) UINOK NASOM; a Vergiato (PID. 100; L""onii.n, p. se 1v.J PRULlK •• P..A.LAJd. 11 Voirn. 0:8. 6 " D6j'- non ne1li1•abl•, 1, nombr• d•■ •x•mple■ do -1'1- gau.loi■ Hl'l'lo ■en•iblom•n• aooru p1.r l& publioation proohaino {At, 0,1,.) d• l& d•:llxio d• Chamal.i6r•■ (Puy·d•-DOm•}. F'&il ■oul& difflouli.6 l'in■arip•ion (on eori•ur• latin,; DoUin 0:1, DAG 10:1) du m•nhir do Vioux-Poilior, (I, Naintr6, Vionno): BA'l.'IN BBIVATIOM I J'BONTY TABB .. I I~P-RV. Pr6oedant ,ujo• (Li) •• -,rorbo d• d6dioao• (1. 3), acou ■ a~if 'l'ISON@S objoi rati1'1-,&YMla flnalo -11 ... ti,nduo (•iiniflca•ion ino•xiaino; rapped dou\,ux av,o lo monhir, d'troo\ion problablomon• prea•ltiqu•; o•.. , vers "pf-ti- qu• doii ■an, doul, •'ori&nt&r la quA\e t•ymologiquo). Lo -11 do raH'l4 no l'ond quo plu■ 6tl'&ni" lo -,il du ■ -cond mo\, lui-mAmo ob,aur (nom propro T) l'll'i•aiiaM; ou bii,n mo• ooroplo• on "-01'1- &YOO a■■ i• rnilAion do ,andhi do-,,ar:d/· (■ugio■tion do O. B., por liUol·a,, 14-. e. 1Q711; mai1, avee lo mot ■ui-,,an•, ni li&i ■ on ,ynlaim.,.iquo 6troito, ni voi,inai• m&•tri•l); ou bi•n abr6-



Aucune distribution gCographique dBfinissable. Orgon (4, --reµ) est 1 moins de 20 km. a l'Est de St RBmy (10, --rev);MalaucBne (--reµ), a moins de 10km. au Sud Peut-&tre seulement parce que nos de Vaison (Dottin 7, DAG 57: crocnvve:µ7J"t"O\I). inscriptions gallo-grecques sont mal claMcs61, pas de distribution chronologique ressortant de l'Cpigraphie olle-mBme (et prouvant que les textcs de Calissane et l\falaucBne seraient anMrieurs aux aubres dom16es); mais quelle que soit, la date des textes eux-mCmes, il est loisible de SUPllOSerque 4 et 7 sont Jes reprCsentants d'une tradition arcba'isante. La situation en Narbonnaise, a eel, Cgard, ressemblerait alors, en principe, a la situation v6n6te en Caclore (ci-clossus, d).Mais, ici encore, on peut se demander si l'attarclement de -111n'cst pas dU a quelque facteur local 65 • 22. Dans le traitement du *-1]? final de l;exrxv-rgµ/--rev, est remarquable non seulement l'articulation de la consonne finale, mais aussi bien le timbre de la voyelle. Traitements r6guliers: en en g1.6lique", an en b1·ibtonique, et, pour le celtiquc continental 6 7 , an en gaulois. Encore 1.-t-on, avec plus ou moins de probabilite, all8gu668 quelques exemples de en en g1.uloi1, en syllabe initiale (*grennos, *brennos, termes de subst..rat en roman) ou inUrieme (nom de mois ELEJJ1BIV a Coligny). lei encore done, posslbilite de maintenir les textes ~pc.:-rouft. l'interieur d'nn gaulois lui-m8me sans unite phonetique complebe (ce qm est le cas d'autres langues aussi, comme le groc) quant aux traitements des nasales voyelles. 23. On s'est efforc0 jusqu'ici de tenir la meme position que 0. S. quant ;\ la langue des inscriptions ~pc.:-rou: via.lion (ne ret•nant quo l'initi1.le -M( ) d'un 1eoond membro, d• oompo,,; pro,mio,r membre o,n . io-, of. :7'ovlio-rigi OIL XIII 7!50&-;1",TI1.~ion o,xplio.t>lepa-r la-limib,•ion du ohamp in1orit: plMe, au plu1, pour un• l•t•re minoo, 1,pr~ M}. L• trao, do, -ltti ro,nd diffl.oilod'y voir -NI (da•it do, d6dioa•1.ire), H Lea, ineorip•ion, iallo-gr•oque1, (" l'in'L6deur d'uno p,riode:-IlI• ,./ +ru ,.) pr,oilldent dl'lll• le tomp• l•• in,el'ip-.iona gauloi .. , en a-lpha-b•t 11.•in(montun•nbJ Ott our1if), I.TIO uno, m•rie do cih•Ta-uohemo,nt do,• deux •n•emble1 qui oorroepond, •n gro1,, 1.u + Jer 1. Lo,11,o,x••g1,uloi1en a-lphabo,•16pon)iqu, (Brionl'I.,Todi) ,on• ,uppo1611.pp1.denir "11. 1eoondo,moi-.iOdu-IT•,.,"' 1e 1i'\u•nLdono ven le milio,ude r,poqu• gallo-grMquo,. Xai, il faul. '" rappolu qu, ]1. plupar• do no ■ da•a-•ion1 1ont oonj,oiural.H. " Ber1.i'\-oe lo 1t1b1•ra• liga:re (oommun au i•uloi• de N1.rbonna-iH od•n•alo oi, au 1,pon•iquo,) f Ioi enoor,, hypo•ht1e (li1uro: 11,ngue-.M) non ..,, laub de t•d•■; [11'1. 1eule in1oripJ,ion prob1,blement lii11r" qu• nou1 a-yon1,.,,. eollo d• Ziina-go, PID. 818: M:BZUNRIMUSUB; 'Mir L,pontioa, p. UI n,]. " Voir 0. 8., p. 178, qui rl'l.ppo,llequo, •i, > an a o,pend1.n• e•e 1uppo•• dan, un po,ti• nombre de 01.1,oomme ai"M tnomt, -foM•!ha,· •••• n,, (1'agi•-iI do, ••,. dovan~ 1ona-11•e [iei, m ou )'], di1.tino) d• "1' deTl.n• oon1onn•f), Bn •ou• OM, 1' -a {impliqu•n• pro-.oga,L •-lil) dan• l• th"rne1 oon1onan'Liqu11ne oon~inue pas n60011ainm•nt •-i,• (av,o Too1.li1me aberrant de ]1. na1ale -...oyelle),mais peui Eltre analo1iqu, d• thtme1, on -o.'1-(ol. L,poJIJioa, 1Q71, I " Inoortii•udo, pour lo, e•l-.ibtre et pom Jo, lopon-.iqu•. - Oeltib, aPi- (~ liro, aMbi-), &UH1' t Botorri•a (vofr n. 80) remon•o,-t-il" •'7'bhi- on t "'"bli,i'.- f Oo,Hib, (i'U. 1g, de '\oponyme) J,OoTia, k,,Oa,, (~ lire B,go(A)lia, fo(A)ga.) 1uppo1e-•-il (ol. O,ltib,rioa, HIM, p. 1!Q) un ""'!'glio- en fao• du •lo,.go- de■ autro, JangoH (11.tin; go,rmanique; g1.u\oi1, m•m•:,1narbonnai,e1 a-veo ethniquo, i6n. pl. Aoyro-ou).l),.-o:,~), ou bien lil'l_)po,llo,-t-il tm• l'l.u•r"explioa•ion f-En 1,pontiquo,, l'aoo, pl. BITB9 .,.,•. u (of. Le,poiMi~a,1~71, p, 10ll 1v.) oelui d'un •h6m• oon1on1.n-.ique ( < •.,,._, < •-~) f - Ni pom •t1 > CM- •n o•ltibillre, ni pour •i, > '"' o,nl6pon•iqu•, on n'1. d'exemple1 1Lu·1, •• Vair Q, B. p. 27Q, qui ajouile: dt i1 not ol... r hoT •h• [G1.uli1h] df,tribu-.ion of .,~/aA i1 •o be do,flno,ch.


Quel oeltique clans t-..E6.EBPAT0l'6.EKANTEM?


1. Affaire d'arohatsme que •d~hl11to- en regard de *dekameto- (§ 17) ou quc -m en regard de -n (avec flottement -•/-• d•n• le dossier ~prx-rouJui-memo: § 2. Pom le timbre e de la nasale voyelle de (l;µ. 1 (Sexc.:v-r)-ev,part1c1pat10n a une altcrnance de traitements en/an dont le ceU,ique, pris dans son ensemble, et aussi bien le gaulois, considere isoltlment, temoignaient deja. 3. Innovation, il est vi·ai, pour l'acc. sg. de Ie d6cl,, devarn;ant l'im1ovation irlandaise {§ 20a); encore est-il non dtlmontrablo mais possible, avons-nous ajout6 (§ 20b), que le rcsLe du gaulois y pitrticipe. . . . Gaulois done, notre dossier, sans qu'il y ait beu de songer, smt (a) a un d11.lecte celtique sptlcifique de Narbonnaise 69 , dont ii serait d'ailleurs mal.aise _de tracer !e• limites sur la carte, soit (b) un formulaire emprunte par les Gaulms de Narbom1a1se a un autre parler celtique non connu 70 • Et cependant .. , . a) Si char;une des singularit0S ~pr:t.'t'o!J Jleut bien trouve1', ga ou JA,,en gaulorn ou dans le reste du celtiquc, des correspondants, le cum1tl de ces singularites n'est-ilpas le debut de la dtlfiniLion d'un dialccte? Un clialecte qui, a l'inttlrieur d'un gaulois mtldiocrement unitaire, aurait, notamment, des traits communs aveo le protogaelique (nasale voyelle do resonance e; acc. sg. de theme ?onsonanti~ue Ie dtlclinaison 1). Bien cntendu, le nombre des tllements sur gum notre doss1e1·restremt nous laisse argumenter est trop foible pour permettre une conclusion. Mais la question subsiste. b) O. S. s'est trouv0 examiner s6pare1nent deux probl6mes poses par le nom de la «di:me1>dans nos tcxtos: structure dekanta- du themo, desinence ·-emf-en d'acc. sg. de Io dtlcl. - II Jui a 6chapp0 que, conjointement, ces deux donn6es (si elles r -sin .Aeolio (of. -i. abo...-o). The aeoond hypo1.he1i1 to (.he point from which ;Te atarted; if Greek: at e.n oady 1-l;1,g11 replaced ""g"n""i·•;,iilh •g1,i.111iAeolio must have a.ltered •g,,..111i iio yCncrLby aimplifying the geminaie a-nd \hia aound oha-nge muaii ht1.veta-ken place before i;he creation of ,,;ll.11-r~crm &nd the like (of. 4.-.&bon). A corolla-ry of 1.hi, vio1T is that we ahould not expect to find any inheri1.od intervoc1.lic -ss- in the .Aeolio dialects.


26 Of. o. g. lL L•jeune, P.lio11.41iqot hi1jo1·iq- cl" '""'o•~itn d OW g1·00 cmdtn., Paris Hl71, 101 and not• 8; WMkornai•l-Dobrunnor, Altin.di,oht 0ra111,iia1i,k, I, 111; ITI l150f., 1B£1f.1'ol".&.vH\cmof. H. Reioh•U, ..'l:wo,,i,ohH Kltt1H,._t,wb11oh, lbid•lb•rg 1gog, 1515. 11 'l'hi■ ohronology i■ of course necessary • □ •xpl1,in Ionio-AUio d ( • .. ~..:> >




n The .. &rling point fol' the r•••oral.ion of -ss- could b• provided by fol·m1 like Lho Romerio llx,o'i',11,wlwi·• -s- waa p1-.Hrv•d in prMon10111!1,n\&l po■i\ion. Th1,\ •.,i in po■ t­ voealio poaition wa■ alilu•cl to -lM.baa ofton been diapuhd, bu\ it now proved by •he W:yoen&oa.n evid•no• (■ee below no•• '1).

The •ccrcrtdatives, Aeolic -ss-, and the Lesbian poets


4-.2. Ji, ia impo11ible to proceed any further m1less we are somehow in & poaition to decide between the two hypoth111e1 just considered. In mxler to do so it ia now necessary to reoon,15truot the full picture of the events which led to the croai;ion in Aeolic of the da-ti-.-eforms which we are discussil1g. In fact, we are now obliged to recom,truo1. not one buii two pioturo11, doponding onoe more on which one of iihe two hypothe1e1 we uae M our ■tarting poin-1;.The ii1Topioturoa Old.l. then be te11~edfor in1.ornal con1i1tenoy etc. and it m1,y be po11ible iiha-t wiiihin thia luger framework one of them may be rojeotad a-nd i;he othor 1npported by more concln.!5iYeeviclcnoo. In ,-.,-hatfollow, -'"e 1h1.ll i1.bnlde the faota a-s we see them. However, their interpretation may be eaaier if we mention here some aeaumption-1 which are common to both pictures. l"h-1511, we assume that the -,acn d1.iii-.-e1of t;he consonantal stems (n:cl.v-rsmnetc.) are a common Aeolio creation and were not built indopendeni;ly in the individual Aeolic di1ol11oi;11 (of. a.boYe 4.-.). Secondly, we a-ssume that the -mm datives were not bnilt ana-logioa-lly on iiho datin11 of the -s- stems (of. above 2. and 3.3.). Thirdly, we assume •hat Wa.okemagel'a propor•ion (-o~: -otm = -sc:;: -scrcn) is responsible for tl1e creation of i;heaa da-iiivea (of. abo-.-e 2. and 3.3.). From Olli' third a11umpiiion it follows that the existence of -otcrt (.in MxoLcrt etc.) was a necessary condition for t.he crea-tion of iihe new -men morphs (-mXv-rwcnetc.). However the Mycenaean evidence ma-kea i• likoly i;hab -i;he inherited morph *-oisu (soon replaced by •-oi,i) wen-1;iihrough a 1tage the -s- was restored to yield the at.tested -otcn, &nd th&t thia rcatorai.ion occured in a poab-tiycenaean period 2 8 . If so, it e.110 follows from our i;hi.rcl assumption that the oree.tion of -0:001 was later than iihe restoration of -s- in -owt. We may now turn to onr two tablesg 9 • 4.2.1. Table I:

IE 1. IE2. PGr. 1.

*-esi (•""""i) •-ehi


cons. stems *-{C)Csu *-(C)Osu

-i-stems *-isu

•-(C)O,i (•pa-.t,i) •-(O)C,i (•pa,..L!i)

*-isi (•poli,t) • (*polihi)


thematic stems *-oisu *-oisu

•-oi"i (•1Coil:oi1i) •-oilii (•1tloiioi4i)

'" I ha.v• h•ld fof a long tim■ -.1ia• \h• oorr■o• inl•l"pl'■ lalion of l{yo. -o-i and -a-i -oi,;1,ncl-i;i;•~ (of.•· g. 0. J. Ruijgh, ll"•'"°'P• 11 [1gl5B),111), bu\ I t1.mno,.- per■ u&­ dod \ba• lhoao morpll■ mn■ I be reacl .. -oi(Mi &nd .a(i)(l1)i {more likely -ri(h)i); •he re .. ona haYo been oonnni• 1umm1oriKodby L•j•\me, RPh'-1 (fg&g), JHlff. (= .M'6111oiro.,dt pliilologit lll-t'Oi6nit1tnt, vol. III, Rome 11t7i, 1158ff.;of. a.110ibid. 1/'il'lff.--,,,herehoh1-1ooll•□· the •vid•no• bet.ring on l.ho ohwnolo1y of •he l"H\or1,\ion of -,- in Myo•naHn}. In Proo. Oa1"bddgt Gollogvit.111-on ~f)'otflloGtal~ Sttd4t1, Oa.mb:ddg• 10&&,U!ff. 81emorenyi ha■ 1,rgn•d •h1.l.lh• Hyo•nael'lll. dal.-looal.iYOIondod in -'1i a-nd -oi and •ha-\ the prooe•• whioh led •o \h• or•ation of •he Ja-t•r -11.101/-~o• ■nd -0,01 ia more oomplioated th1on i ■ normally ■uggHt•d- If •hi ■ were so, \ho main line■ of our argum•n• would not be neoo111,rily aff"•o••d. .. The h•acling■ P1•o•o•G1·••kl,l)_dProlo-Aeolio are obviou ■ly oonYon-.ional N1d ■hould not bo ••ken • □o a•i:-ioUyin &family-tree 1on10.I ha.v• alao aYoid•d aeo• or propo■ini &ny thoorJ• about th• origin of the Oreok dial•o•• and •hair di■tribul.ion in the aeooud Hillonnium, though my 1ugge1\ion1, if correct, may some implioa•iona for •hi1 problom•oo. WM




PAeol.1. PAeol. 2. PAeol. 3.




(*pantsi) *-(C)Csiao

•-ui *-esi


•-i,i (•poH,i) •-ilri *-isi



•-oi.\.i (• •-oYi *-oisi


Aeol. 1. of Asia 2.

(&veecrl) -eecrcn

(1hwprt.v€wcrt) Thess. 1. 2.

i *-esi -ecrm


*-isioR *-iessi -otcrt

(mfv-remn) -(C)Cecrcn




-(C)Cecnn -(C)Ce:rrcn


-(C)Cecrcn -(O)C.:crcrt

*-isioR *-iessi !

cons.stems *-(C)Csu *-(C)Osu



•-i.,i (~oH,i) •-ihi (*polihi)


•-ibi (•poliki}

•-oiki (•woi:toi.-i) •-oilli






Boe. i. 2.

? •-ui no oTid•noe

4.2.2. Table II: -s-stems

IE 1. IE 2. PGr. l.



._,,,. (•guu,,i)


PGr. 2.



PAeoL 1.



(•gumi) •-Mi


PAeol. 2. PAeol.3. PAeol.4.


•-i,w *-isu



•-(C)C"i •-(0)011 •-(O)OM