Robert . ing theories (such as Paul Wexler's Sorbian hypothesis)?

Such radical hypotheses sometimes lead ( interesting new questions and answers, soeties they lead generation of linguists down false path. we detennine when we [ former case, when the latter? Such are the questions that retlections [ origins of Yiddish lead us (. 1 do ! think that [ are all that different from questions that linguists who work Nostratic are required to pose and try to answer.

BEFORE INDO-EUROPEAN AND URALIC JanosMakkay

establishment of [ period of existence of the Indo-European proto-language rests ( considerabIe extent ' s conception the formation and succeeding development of proto-Ianguages. Basically we have two choices: . Separation (subgrouping) along [ rules of the well-known faily ( ode/ when the 'daughter' languages of former Nostratic language protofamily (Altaic, Uralic, Dravidian, artvelian, Afro-Asiatic and Indo-European) as constituents of [ protofamily show great number of m archaisms and partly m innovations unknwn elsewhere, and ! [ sarne time they keep the shared retentions of core vocabulary (more rarely morphology, phonology or perhaps syntax) [! was m ( [whole protofamily or to other branchlbranches of its groups. . other suggested way of [ evolution of parent language (for instance the Uralic, Indo-European or an other protolanguage), i.e. the recently preferred covergence (heory (model of language league or language alliance) must ruled out in this case. According ( this theory several protolanguages did ! about from m ancestral superfamily but developed integrating small triballanguage units complicated (and hitheno totally unknown) convergence processes. basic idea of the Nostratic protofamily ought ( [ widely accepted principle that related language families diverge with the passage of time isolating from other. apply the convergence model ( the memberslprotodialects of the Nostratic protofamily would push time limits of [ formation of the Nostratic protofamily back ( times beyond unanalysabIe depths in the Middle and Lower Palaeolithic (see [ s Fig. 5). the other hand, 1 consider [ 'Sprachbund or language league-theory' an artificial and forced creation and as such untenabIe. Th theories of language league and the idea of Nostratic protofamily are incompatibIe with other. As result of these shon considerations [impottant points emerge: . Th relatedness of these six protolanguages i.e. [ question whether the genetic relationship of these languages an eminently proved the relevant linguistic criteria. This apparently is ! duty and topic. . Another imponant point is ( detennine and describe [ period underlying [ formation, existence and split of the Nostratic protofamily, i.e. the chronologies of these related events which, [ other hand, are dispersed over very wide area and also very long way of development in time. As result, their closer archaeological study would strongly need cooperation of several expens of [ Palaeolithic and Mesolithic which is apparently ! the case ! the moment. Professional archaeologists unfortunately rarely dwell such and similar questions. An impottant of this point is ( calculate the duration of the formation (etc.) periods of [ protofamily, since its succeeding phases an fall different developing phases as for example the Upper Palaeolithic and Mesolithic of given area (selected as possibIe homeland of the

142

143

Jdl10s

.'
ro

/ Indo-Eropean & Ura/ic

Nostl"atic processes linguistic grounds57). As rnatter [ [act, several lts [ continuity observed between [ Final Palaeolithic and [ MesoJithic [ ( Levant and bordering pan5 Syria and Turkey ( details , below), and in also between their Mesolithie and Early Neolitl1ic eultures. l think tl13t suggested population continuity { Final Palaeolithie and { Lower Mesolithie give" [ { Near East i"s ( 05t! opportunity [ ( Nostratie hypothesis.

around15,000 (between 15 and 10 thousand) is reaIistic (Fig. 1), . live

thousand years remain accurate [or intemal and diverging processes of the Nostratic protofamily. exclusion principle [the homeland should ! set in an where
there is evidence [ prior existence of different tongues; MaIIory

Adarns 1997, 295]

Fig. 1: Locatiol1 oft/Je Nostratic hoe/and " suggested " /d. After Mallory and Adams 1997, . 292. The s/dd il1dicates the gCllera/ized distribution oft/ze Nostratic /anguage faily. The darker s/dd illdica~s tlzeNostratic Izoelad . 15,000 accordig { Allall /d. Tlle terius nt que af tlle ist of assumed Nostratic pl1ase given sil: [ independent Indo-European, i.e. the separation [ the lndo-European speeeh 11uit. s here the ti length 01' the independent existence 01' lE. This ssud length here ratl,er depends edueated guesses, balaneed judgrnents and idspeeulatio"s than observabIe ts. 1 eonsider the time depth around 5000 as punetuation point when the final di[titi processes [ the IE speeeh dialeet tiuuhad staned. This suggestion gives enough t; [ struetural proeesses and intemal dvlts [ the ' [il toward separations [ its earl;est daugl1ter languages ;.. dialectal groups (e5peeially Anatolian and Indo-Iranian). and also [ tran5itional develop,nents Nostrat;e bases toward its "' daughter languages [ 5000 . ' hd's suggested loeation [ his Nostratie hld (into Southem Turkey, Nonhem Syria and Irak). and it5 ehronology 57 vcry
importanl point is that the sclected should guantt: t.:mm linguistic

does ! play role in this case. . FinaIIy comes the material background [ the whole proeess, i.e. some arehaeologicaI facts or theories whieh related to lingustic processes some means.58 Bomhard's mentioned dating around . 15,000 determines the archaeologieal character [ this period: according to tbe recent stand [ Near Eastem prel1istoric archaeology (whieh shouId k into consideration) this datum sds weII to the earliest emergence and succeeding floruit of tbe Mesolithic way [ li[e in the Near East and Anatolia, and an ltd witb the end of the Upper PaIaeolithic in titris of Europe lying south [ tbe Ice . short eontribution wiII coneentrate reconstructable events [ these periods, mostly [ tbe MesoIithic at the end and after the 1ee Age. I[ views tbe Nostratie theory t, there is simple way to eharacterize the relationship between the protolanguage groups and the superfamily, and it is the extension [ the [amily tree model backward in time, into time-depths weII be[ore the period [ existence [ tbe daughter proto-Ianguages. basic principle [ this approach is tbat if the cognate stocks of language [amily (Indo-European, Uralic, etc.) more or less similar or ouIy somewhat related (and presumably geographicaIIy proximate) to another, so aIso someone argue tbat simiIar relationships exist between di[ferent language [amiIies59 bordering other. 1[ we take sueh backward step, it does ! change tbe general ruies of linguistic reconstruction but causes them to go back stage d gives (or give) them specilie (or different) temporaI, spatiaI and culturaI perspective. temporal category wiII , [ course, the ehronology [ tbe whole process depending two [actors: . First, the dating [ the existenee tbe stiII undifferentiated, original speeeh ! - Nostratic - munit of tbe daughter languages i.e. be[ore Nostratic began to diverge into diaIectal groups, i.e. the Nostratic daughter protolanguages. 1tan argued that this systematic dating approach wiII mueh [aciIitated i[ tbe ! speech 11unitis (both of the assumed Nostratic super[amily and its descendants, or only an [ them) an identilied with archaeologicaIly attested1attestable [aets and suggestions. From this in! of view tbe ! speeeh groups [ the UraIic protolanguage delinite advantage: spatially, it k as resolved that tbeir speekers had aIways Iived the northem periphery of (Indo-European), two (EIamite andJor artIi) or more (?) dughtr-Igugs (west or northwest [
different, mostly undelined PaIaeosibirian and other Asiatic

Altaie?

groups).

Teporally great ! of the UraIic protohabitat had covered the !-.

S8 he g large distribution [ Upper PaIaeolithic archaeologicaI entities takeo into consideration, the application of the method of contiguous and large cultural territories would usefuI here. Por tltis method ' Makkay 1992,200-201. S9 J. . MaI10ry in MaIlory-dams 1997,291-292.

deve10pment and territoriaJ closeness of protodialccts of { superfamily .. continuity of their material culture) during long period of time.

144

145

Jaos Makkay he first settlers migrated ( tbese temtories after ( final retreat tbe !- in continuous flow after 12-8,000 , and if these time limits are granted, tbey mn sure teris post qe for tbe aival Proto-Uralians ( tbeir prehistoric habitat. It surely means (! around 8,000 ! 1atest, speakers of Proto-Uralic had already diverged from ( Nostratic protofami1y. Or, what I would consider absurd, groups speaking Late Nostratic tongue had who migrated in [ wake of tbe retreating ice toward tbe Nonhem ! . Considering tbe time around 8,000 or between 12 and 8 tbousand as tbe phase of emergence of Proto-Uralic language family and ( slow ! progressive moving of its speakers to tbe Nonh, dating of around 15,000 for ( existence of ( still non-diverging Nostratic superfami1y seems ( realistic. . Second, some significance should attached ( tbe lengtb of tbe - hitherto mostly unknwn - linguistic processes and also tbe speed of tbe man progress of that assumed period which led ( ( emergence of tbe six language groups. Fortunately, two eminent scholars of tbe field gave us estimates which I think are based ( available evidence. Vitaly Shevoroshkin calculated in 1989 tbat Nostratic was language spoken some 14,000 years ago, i.e. 12 millennia . (Shevoroshkin 1989, 7). As mentioned before, . Bomhard suggested Nostratic homeland lying between ( Levant and Turkish-Iranian Kurdistan, and dated it ( around 15,000 . relatively early time taken into consideration, tbe difference between tbese two calculations is ! very important, and it simply an ( result of tbe continuous (and continuing) backslip (or downslip) ( time-depth in ( last two decades (for more details see Makkay 1992, . 199).' backslip is partly tbe result of tbe use of scientific dating methods in prehistoric archaeology (as for example tbe application of higher calibrated and recalibrated radiocarbon data for tbe same development phase), and at tbe same time because of tbe discovery of increasingly earlier and earlier cultural horizons the Neolithic. tbird time factor an called tbe stretching tbe single periods. It was Gordon Childe ." discovered this when likening [ prehistoric chronology "to flexible bellows which could expanded or contracted ! will: end was fixed ! 1500 .., tbe otber earlier was free to , giving longer or shoner chronology very according ( the wish of tbe archaeologist" (Makkay 1989, 177, witb furtber reference). Conceming recent results of radiocarbon-based chronology in Ear1y Neolitbic Near and Middle East, instead flexible bellows 1 would use ( analogy of an extendable rubber band: end is fixed somewhere in tbe third quarter tbe second mill. . dating of cu1tures, phases or types before this date simply depends tbeir relative position in [ relevant sequence, and also tbe stretching of tbe band, i.e. tbe stretching tbe relevant part of ( band. clustering and scatter of absolute dates remain in tbis case in good agreement witb tbe broad outlines of tbe traditional re1ative chrono10gy, i.e. tbe general sequential pattem is a1ready clear. tbe otber hand, bowever, tbe deeper tbe position Neolithic Cesolithic, !.) artefacts/phases in tbe

/ /do-Eropea & Uralic stratigraphic sequence, ( higher their absolute - 14 - chronology (and tbe otber way round) (Makkay 1996,221). furtber aspect of chrono10gical irnportance is when an earlier or ch earlier archaeo10gical entity (group, culture, etc.) is selected ( stand for tbe material representation of tbe antecedents of given protolanguage or superfamily. For more details see below. best example of these (especially tbe second) factors (discovery increasingly earlier and earlier horizons) is tbe dating of tbe famous // c/tre of Northem Mesopotamia, Southeastem Turkey and Northem Syria, which p1ays special role in ( search for, and identification of, tbe Indo-European homeland (as for exarnple in theories of Diakonoff and Gamkrelidze-Ivanov). Before tbe excavation of its principal site, ll rp in 1933 (Mallowan and Rose 1933), ( earliest knwn pottery phase of tbe post-Mesolithic developments of the Near ! (and tbe whole world) disregarding some earlier beginnings in Natufian times tbe Levant - was tbe al'Ubaid culture (Burkitt and Childe 1932). After 1933, however, tbe extreely fine painted pottery of ( Halaf culture was considered tbe type fossil of tbe earliest sedentary phase tbe Near East, and was dated around 5000 usingl1re-radiocarbon high chronology, 1 Now tp dates ( tbe more later if using traditional dating systems. period some time between 6000 and 5000 , and according ( ( presently established sequence it follows series of Pre-Pottery Neolithic, Proto-Neolithic, Late Aceramic Neolithic and Hassuna phases from tbeir beginning around 10,500 (Bienkowski and Milard 2000, 30, and ( chrono10gical . vii; Charvat 2002,42-71).62 ( earliest Neolithic sites (excavated recently) is Hallan t;emi Tepesi in tb Eastem Anatolia, dating ( ( end of ( 11 millenniu (Fig. 4).63 site's inhabitants were dependent primarily hunting-gathering, ! were a1ready experimenting with animal doestication. also possessed rich and 1 culture witb affinities ( botb earlier (Zarzian, i.e. Mesolithic), conteporary, and later sites along (Taurus-Zagros flanks (Rosenberg 1999,25).

61 Childe (939. _ Makkay 1989, 177-181. - correct, Childe dated rp after hypothetical "Neolithic" and before the al'Ubaid sequence because presence of copper. Copper, however, is also present in Early NeoJithic deposits as for exarnple il in kish distan: Ozdo!lan 1999, 58: copper beads was widely used during the Second and hird Stages of the PPN and PPNC. In any case, as . Mallowan writes. Arpacblayah's excavations opened new and enthralling chapter and will for ever stand as milestone the long road ofpreblstory: Mallowan 2001, 100. 62 As for the phases of the Pre-Neolithic and Early NeoJithic sequence see Ozdo!lan 1999, !, 41-59. 63 Hallarn <;emi Tepesi, irnportant and very early centre of Early Neolitblc developments, is located in the foothills of the us Mountains in Eastern k, and it is ! 500 kilometres nonhwest of Shanidar. Its ult was derived fro the regional EpipalaeoJithic Zarzian tradition and as sucb can testify to the succeeding continuity of Late Palaeolithic
EpipalaeoJithic traditions (Solecki ! al. 2004, 118).

60 J. . Mallory in Mallory - Adams 997, Fig. . 292.

146

147

Jdnos

Makkay

Before /ndo-European & Uralic perhaps to thc north-east, in the later part of [ Pleistocene, over the time range from 15,000 to 10,000 .." (Renfrew 1989, 137): "What, then, if the Nostratic hypothesis were right? What if all these languages were indeed related some farnilial affinity? What are the implications, ifwe should thinking of an early, proto-Nostratic language, spoken perhaps around 15,000 in some area within Europe and western Asia, from which all these later languages would in some sense descended?" (Renftew 1990,7.) suggestion of Nostratic homeiand lying between Levantine and Kurdistan territories is beyond the realms of Pte-, Ptoto- and Early Neolithic archaeology of the area around and shortly before 10,000 . 1 fuHy agree with the view that "! date around 10,000 these proto-languages already have existed as distinct ianguages or dialects." (Renfrew 1991, 14.) totally dillerent picture emerges, however, if we advance (backwards) beyond the 15" i., to Upper Palaeolithic periods. II someone would incline to accept Bornbard's proposal for an original mnn Nostratic homeland in the Levantine-Kurdistan area, he/she should matiall subscribe to the theory of Renfrew. It identified East Anatolia as part, although ! necessarily all, 01 the early "homeland" of people speaking very early fonn of IndoEuropean, around 7000 (Renfrew 1987, 174). reception, however, of his theory .. ! n lavourable n, especially its suggestions concerning Anatolian origins of the Indo-European stock. 1 fuHy agree with the wording of . Szemerenyi (without going into details of horing but astonishing archeological trivia1ities): "1 must confess that the whole of [his] statement, obviously 01 the highest importance for Renfrew's theories, seems to me utterly without rational basis".67 Recently well-infonned expert Levantine matters, the late J. Cauvin remarked the sensational synthesis of Renfrew 1987 as foHows: In his view the neolithisation of Europe an understood as genuine colonisation that started in the seventh millennium from source in Anatolia and involved the arrival in Europe of new population elements. considers these as the carriers of the Indo-European proto-languages [Ianguage], moving forward in accordance with "wave of advance" model. More recently, .. extended his theory beyond this family 01 languages referring to the so-called 'Nostratic hypothesis' proposed Russian linguists. ! level, cornparative methodology .. allowed the definition of farnilies of languages (for exarnple, the 'Indo-European' farnily) in tenns of affinities of vocabulary, morphology of words and phonology, all the languages 01 farnily being derived from hypothetical COll1fOOn ancestor. Now it is further proposed that certain of the larnilies themselves present affinities that allow them to grouped in their turn into 'macro-farnilies' that equa1ly point back to common source. herefore 'Nostratic' proto-Ianguage was the origin olthe Indo-European, Harnito-Semitic, Elarnite-Dravidian, Ura1ic and Altaic farnilies, that is of the great majority of the languages spoken in Europe, Asia and North fri. Renftew nks his conclusion with 'four-Iobed' theory of the f1fst Near Eastern Neolithic
67 Szemerenyi 1989, 158-164.

Zarzian industry testifies ( the continuity between Late Palaeolithic, Epipalaeolithic and Ptotoneolithic inbabitants of the Eastern fringes of. the F~rtlle Crescent. In Shanidar cave, the Proto'Neolithic horizon Iles ahove the Eplpalaeollthic Zarzian horizon, but [ basis of radiocarbon-14 dates, th~re. as break of some two w thousand years between the two occupations. Zarzian lithic dustr, however, n viewed as " direct development out of the underlying late (Palaeollthic) Baradostrian industry ! Warwasi" Zarzian date as early as 22,000 ~P, and have lasted until 12,000 . After time gap of some two thousand ye~s ' ProtoNeolithic in the Zagros area was time of much cultural change and penod that an viewed as transitional between the Epipalaeolithic and the later, Aceranuc and full~ developed Neolithic" dated into the time period from circa 10,900 (~10,500 (Solecki ! al. 2004, 114-116). ak this territory of Zarzian cultural tradibons conslstent W\th the initial Nostratic distribution of . Bornbard would need to expan~ his area. . western area of Bornbard's candidate for the loca1izabon of the Nostra~c [ somewhat earlier homeland of relatively restricted distribution is the territory 6~ Mesolithic and Proto-Neolithic cultures in the Levant and Syna. Both seem to .v~ry promising as the place of very early lndo-E~opean. and Pal~eosennbc (AfroaSlabc) contacts i.e. [ scene of their final dlSlntegratlOn: the1f separatlOnfrom thelr mnn superfamily. exclusion principle, however, argues that the 1 hom~land should n! set in an area where there is evidence of prior n-1occupabon. This also relates to Hattic and Hurrian Anatolia, almost the entire Near st (Semitic, Sumerian) and Iran (Elarnite).65 This principle locates the earliest possible 1 homeland q~far north of the dark shaded area of Bornbard (Fig. 1), ! the present mom! wlthout clear understanding of the great spatial gap of the suggested protohabitat of the ancestors of Early Indo-Europeans in Nostratic phase and their calculated -, mostly European distribution in the Early Neolithic. As it is weH known Renfrew s chOlce for an 1 protohabitat (from range of prehistoric cultural groups) in the close~ area of Catal ii(Iying an expanded territory of Bomard's Nostratlc) was eqUlvocally refu~ linguists, prehistorians and cultural histriatsusing g~eatnumber of arguments. only possibility to reconcile such appare:nt\y contradictory suggesbons won1d pushing back the dating of the dispersal of Proto-lndoeuropaeans from Nostrabc protohabitat lying in territories of the Fertile Crescent (see below).. . . Renfrew's model was n of the first that placed the Nostratlc homeland to the Near East and dated it to the inibetween 15,000 ( 10,000 . According to m "the historical background to the Nostratic group would in the Near East and

64 recen!summarysee Cauvin2000, part1. 65 J. . Mallory in Mallory and Adams 1997, 295.

already considered E1amites . 3000 : Po!ts

- he inbabitants 01 Susa
1999,43.

and Fars nU~t

This argumcn! is, 01 coursc, valid

1 from tbe poin! 01 vicw 01 tbc cxclusion princip1c il tbc E1amite was ! part of bypotbctical Elamita-Dravidian protolanguagc (i.c. Nostratic) group. I general, Szemerenyt 1989, passim. 66 Scc Transactions 01 tbc Pbi1010gical Socict:y 87:2, 1989, 158-171, and also Cue! Antbropo10gy 29:3,1988,441-463.

148

149

Jtinos Makkay [68] which were, according to , the Levant, Anatolia, the Zagros and peoples, perhaps Turkmenia. Based an original comrnunity of 'Nostratic' language, of these four lobes given rise to of the four directions of simultaneous agropastoral expansion and linguistic differentiation: Anatolia gave rise to the westward direction (the Balkans and Europe) with Proto-!ndoEuropean, the Levant to the South (Arabia and Africa) with Proto-Semitic, the Zagros towards the east with the Elamo-Dravidian languages, and finally Turkmenia towards central Asia with the dialects of the Uralic and Altaic families [Fig. 2]. It goes without saying that this attractive theory very wel1 relate to reality, but, ! the present stage of archaeological and lingustic research, it is Renfrew's opinion that it is ! possible to consider it as anything more than speculative hypothesis, only the European direction seeming for the present to sufficiently supported (Cauvin 2000, 139). We an add: important parts of his hypothesis (as for example that concerning the origins of the Proto-Greek language from the East, frorn Proto-!ndo-European developingaroundCatal Htiyiik)must and has - considered false (Makkay 23,9). No atter how attractive Renfrew's theory of Turkmenian seat of Proto-Uralic is, it is in the strongest opposition to variety of suggested Uralic proto-homelands lying wide belt between the Altai Mountains and the Baltic (with the exception of unfounded claims for an Uralic hoeland in Southern Central Asia in the MesolithicNeolithic (Makkay 2003, 240). the other hand, speculative hypotheses based informed guesses are the only available ethod for approaching to linguistic atters of Upper Palaeolithic times. It is interesting to note from this respect that the Hungarian linguist JlSHarmatta was the ft who asserted that the earliest post-glacial habitats of !ndo-European people [around . 15-10,000 ] were in the Near East in the neighbourhood ofProto-Smiti peoples (Harmatta 1989, 162 and 17).69

Before Indo-European & Uralic

Fig. 2: Schematic representation offaring origins and language dispersalfrom nuclear in Anatolia and aod. When transition (primary faring occurs in nuclear with some linguistic diversity (shown within the broken circle), the consequence ofthe ensuing agricultural dispersal is likely ( linguistic replacement in adjoining areas. he lobes represent the areas ! the subsequent language families derived from the corresponding proto-languages. Originally published Sherratt and Sherratt 1988, Fig. I, with comments ! Renfrew 1990, Fig. 4 page 12.
I

68 Renfrew 1991,6-7. theory was first suggested . Shelatt and S. SheIatt in 1988: Transactions of the Philological Society 87:2, 1989, Fig. 7 . 135, seen here as Fig. 2. According to the ! interpretation of Renfrew, "the curreot distributions of languages comprising the constituent language families of the hypothetical Nostratic macrofamily (including the Indo-European languages) could plausibly explained the wider application of the farming/language dispersal model". Renfrew 2000, 9. 69 For further comments see Szemer6nyi 1989, 162 and Renfrew 1989, 173.

There is undeniable fact which an! escape our attention: the Nostratic homeland suggested Bornhard is identical with both of the mentioned Early Neolithic focuses, and ! the same time it is confined - or if wishes differently, relatively confined area as compared to the supposed distribution of our daughter languages in the final Neolithic: two-thirds of Eurasia. Here again reference shouJd ade to suppleentary note of . Szemerenyi: experts in possession of the relevant evidence should re-examine the question whether the area now selected for the hoeland an in fact proved to inhabited ! the tirne required, and that in sufficient numbers to appear as likely candidate for being the cradle of an enterprising go-ahead race.70 Another important factor is that the relatively confined homeland of Bornhard the Levantine-Cappadocian-Kurdistan belt - is exactly the area where archaeological investigations into the NeoJithic accelerated to an astonishingly swift within the last twenty-thirty years. century-old tendency continues: such relatively small geographic areas used to claimed as homelands where spectacular archaeological researches had taken place.

I I '1

70 Transactions

ofthe Philological

Society

87:2,1989,164.

150

151

Janos Makkay

Hefore /ndo-n & li Age graves at the end ofthe 19. ty.71 Without doubt, the mentioned Central Asiatic theory of { Finno-Ugrians was founded S. . Tolstov's thoughts, which leaned excavations in horezm in CentraI Asia, especiaIly { so-aIld Kelteminar lithic technology and contemporary pottery, during World War II and afterwards (Laszl6 1961, 103-110).

/ ~ -.\

,/

/"--'\ '''1

"

.''

~-'L

''!
/

Fig. 3: Hypothetical application ! the odel seen in Fig. 2 to the Neolithic Near East. It is postlated tOOt rond 10,000 ,Proto-Afro-Asiatic (1), Proto-E/ao-Dravidian a (2), nd Proto-/ndo-Eropean (3) langages were spoken in the Near East within the OOtchedareas, [iWsibly 4000 . A/'terRenfrew 1990, Fig. 5, page 13. During the first decades of research in { last third of the 19. century, this was an entirely mmo view both in investigations about Indo-European and Uralic protohabitats. New and intensive discoveries and excavations took place commonly in relatively large region of some country, and researchers estabIished chronology and cataIogue of especiaIly funerary remains but aIso recovered rnaterial from settlement sites. An enormous amount of previously unknown archaeological rnateriaI surfaced, which spurred ( create new theories. After some years, based such new discoveries and knowledge, [1S! rchaeologists, and then historical linguists founding their theories a those of the archaeologists, enthusiasticalIy present certain area, which {en becoes the new Indo-European, UraIic, or Nostratic homeland. This happened the first time in the middle of { 19'" century in { case of the Swiss lake dwelIings, MegaIithic graves and later Bronze Age cultures of Southem Skandinavia, and the Late NeolithiclEarly Bronze Age Corded ware culture of CentraI rman. h of these cuItures and (mostly confined) regions were presented in tum as the ancestraI cuIture and hoeland of { Indo-European peoples. For the study of FinnoUgric prehistory, this kind of idea of geographicaIly confined hoeland area - the ama-river region - has faciIitated { studies of { Ananjino Bronze and Iron

Fig. 4: Thepresent-day distribtion ! the langage grops which were classified ll/ich-Svitych within the Nostratic acrofaily. /: Altaic, 2: Afro-Asiatic, 3: /ndon, 4: Soth-Cacasian, 5: Ura/ic, 6: Dravidian. After Renfrew 1990, Fig. 2 . 7, In the last third of { last century { assumed Indo-European protohabitats were foIIowed two other regions, Kurgan region of the South Russian steppe has quite welI (but ! completely enough) known Russian and Ukrainian research (more than 50 thousand excavated graves!), homeland of the proto-IndoEuropeans the favorite research topic of . Gimbutas ( basis of thousands of these so-lId Kurgan graves (the exclusive sequence of Copper Age Pit-graves, Bronze Age Catacomb and Timber graves, and pre-Scythian and Scythian tombs) - and
her inadequate knowledge of Soviet and Russian data.

l' I i
71 Fodor, 1973,47-55. Makkay 2003, 242. note 22. In fact, most of the Ananjinn especia11y nf metal - wbicb an considered graves contain arcbaeo1ogica1 materia1(Scytbian) groups. remains of Old Iranian

152

153

Janos Makkay other, recently preferred is Anatolia. J. ll. significant discoveries in Hacila and <;:atalH(jyUk, . well as results of mostly former Halaf culture studies (together with spectacu1aexcavations of the last three decades in the Levant) given the exclusive basis for Renfrew's - and Gamkrelidze72and Ivanov's - previously mentioned location of the Indo-European (and also Nostratic) hornelands. Lying between the suggested homeland areas of the protolanguages proposed in the last two centuries, there were far larger territories than the homelands. Most1y stray finds recorded . resu1tof nearly totallack of extensive excavations from these extensive territories between the small homelands. Shall we igi that ! the end of the 19thcentury an enormous area, 3,000 kilometer wide, was archaeologically entirely unknown or nearly unknown, and tbls area separated the Ananjino culture of the region (the Uralic and!or Finno-Ugric homeland in man theories) from the Corded ware groups of Silesia and Czech region (wblch an had considered . the oldest region where the Indo-European protolanguage was spoken)? , it was ! questioned how language contacts cou1d possible between two such sll culture areas - in tbls case between the Indo-European and the UraliclFi-Ugrian confined homelands - located so far apart. ! the sae tirne, there was lot of discussion about the lingnstic interrelations (or primevallinguistic affinity, within the framework of kind of Nostratic) between the Indo-European and Uralic protolanguages (the IndoUralic hypothesis) - but the question itself is clearly to further discussion. h.ts to intensive archaeological field researches of the last decades, the interrnediary blank areas the archaeological s disappeared: museurns of ! only Europe but also Turkey and the Near East are literally overpacked with excavated finds. Numerous other argurnents an also brought forward against the idea of smaJJ (confined) Indo-European, Uralic - and also Nostratic - hornelands. Gyula L.szI6, an ll! Hungarian archaeologist, who wrote pioneering book the archaeology of the earliest Uralians (Usz16 1961), has expressed tbls the most accurately. wrote that if we assue that the "proto-people" had really lived in say - three centers (speaking three different tongues i.e. dialects of - say - Nostratic superfaily), who had in this case lived in other regions from wblch there finds providing proof of continuous settlernent? If linguists' assuptions (! small homelands) were right, then we should presuppose the ex.istence of widely separated and intemally homogeneous cultures with dense populations, and that there were extensive inhabited regions between these cultures ! only with different trial cu1ture but also with languages belonging to totally alien protolanguages. W then correct in asking the linguists how Uralians and Indo-Europeans, starting from their small hornelands lying definable part of the Nostratic protohabitat - cou1dsuddenlyspreadover such an enorrnous and assimilated more sizeable populations, who spoke different protolanguages, and who lived in geograpblcally larger areas than those from which the

Before Indo-European & Uralic Indo-Europeans and Finno-Ugrians rne (Laszl6 1961,71; Uszl6 1987, 37-38). istorical exaples indicate, . we know from the Indo-European dialects that spread to regions where other languages were spoken (for exarnple, Greek and the Anatolian IE-languages), that these kinds of assimilation processes slow and variety of different substrates is large.

72 iusl enough. dig the formu1ation of their Anatolian homeland theory (Gamkre1idze and Ivanov 1995, the origina1 Ru"ian edition 1984) Gamkre1idze held to the old opinion that Indo-European group' pene1rated into to ftom their Northern homeland _cro" the C_ucasus, ! later than the end of the third mil1ennium (Gamkre1idze 1970).

Fig. 5: Maps representing the alleged coincidence ofthe anthropological (), archaeological () and linguistic () odels ofM. lin; (1996, Vol. 1,Figs. 14.3. and 5-6.) relating the Lower Palaeolithic (5. -) and ( correlating tripartite division ! large language groups (5. ): agglutinative languages (AG), synthetic (inflectional) langages (FU), and analytic (isolating or root) languages (IS). 155

154

Jdnos kl

/ [ndo-European & Uralic settled territories inbetween . for exarnple the weH-knwn Cucuteni-Tripolye culture living east and northeast of the Carpathian Range, An assumption of contemporary tribal language groups living in geographically small regions does ! help , 1 this case, we againsl man questions of principles. Why would tribal group speaking the Proto-Uralic language eerge in s area . resull of Sprachbund-like integration of previously isolated language groups representing different ethnicities when in its neighborhood there lived other groups that were also more or less independent from other, and integrated themselves into another language group, in this case the Indo-European - or AItaic? language family? Nor an know why these dispersed tribal language groups would start suddenly to group themselves into language family especially in the arna river region and the South Russian steppe (or Central Germany) ! ! elsewhere, 1 addition, all of this would have occurred during the Mesolithic, which we know was ! period of cultural integration ! rather period of isolation and divergence into s technocomplexes and ethnic groups, According to different ! reliable countings there would have 1,54, 90 or even 700 isolated tribal societies living in the . million square kiloeter area of the Central an forest zone, or in the later an Finno-Ugric territory of similar extent, and the sae number of language units of different genetic origins when integration processes resulting in the forrnation of Sprachbunds would have started at the dawn of the Mesolithic. he start of the assumed integration processes ought to dated to the time of the Mesolithic, because supporters and believers of the language league theory equa1ly postulate the existence of parent speech communities, the disintegration of which began around the time of the tum between the Mesolithic and Neolithic periods.
1l0rtll
Pl'Oto-Ls

1f this kind of rapid assimilation happened when several dialects of Nostratic superfamily spread from their relatively smaH hoelands (. for exaple lobes from their initial distribution territory), then where are the substrate features - expected to foundespecially in the peripheralregions - providing proof of this kind of expansion? As it is weH knwn, Central uroan regions of 1ndo-European dialects did ! show substratal remains (only their northem peripheries where most part of words indicating substratal influence belong to Uralic i.e. Lapponic protostratums73), while the existence of traces of assirnilated local antecedents territories of the FinnoUgric languages is theoretically impossible. From this latter point of view, of the moSI important and frequently neglected archaeological facts is the Janislawice (Poland), the faous Mesolithic skeleton found in Tardenoisien grave, attesting 10 the presence of ! Lapponic anthropological type the North an Plain, lying far to the south of the suggested prehistoric protohabitat 01'Lapponic tribes (Makkay 200 1, 327-32874).

Small - cltraUy independent, loosely bonded, related? triballangages nd the convergence theory. ..

- aeas / <ll

sum , archaeology has developed great deal since the fst formulations 01' theories ! the eergence of protolanguages and 'Iocalizations 01'protohabilats, and the site number and complexity of archeological finds has increased in all regions of Eurasia. We an mention IWO factors for this. is that there are longer white spots in an archaeological sense within and Anatolia, ! in the most peripheral regions, the most northem areas, and high mountain territories. regions between the arbitrarily selected centres (selected homelands) are longer empty, ! they are fuH of sites: ual grounds, setllements and hoards. 1f someone still maintains that the homeland of the Uralic (or only of the Finno-Ugric) languages should the Bjelaya region or between the Pechora and rivers, must assume that there were other proto-Ianguage groups in the zone between the Uralic and 1ndo-European homelands (for exaple, between the arna and Oder rivers). Similarly, if somebody would adhere to the orthodox belief that Central Germany or part of the Kurgan area in Southem Russia an considered the original homeland of the Indo-Europeans, he/she must speculate ! the nature and linguistic identity of populations living in densely

easl
Proto-Finno-Ugric Indo-Europeans Indo-Europe.ns
Caucasians

ilMle
sotJI

Indo-Europeans lberians-Basques

Sicans

Indo- ur1S

I ProloHattians

3 PrOIO2 Urarraeans PrOloElamites

73 kk

2001, 327-328. Neolithic

the other hand, there is absolutely popu1ation in Westem other later periods. analysis the archaeological

archaeological Siberia

evidence

prove the presence Mesolitblc, Janislawice

rto-Lap '

in the Palaeolithic, problems around the

Sketch (spread (Weslern Eurasian protolanguages ! Iheend of (Ice Age. Fig. 6: SkRtch/ the spread / the Western Eurasian protolaguages at the nd / the [ . represented in the ode/ /. Gdbori nd J. Harmatta. After ldy 2000, Fig. 1 . 73.

74 For more detailed

and preblstoric

man see 2002 ' published only in Hungarian, . 119-129. J. armatta was

the first who made reflection ! the role the Janislawice man in questions of the Uralic (Proto-Lappooic) ethnogenesis. unfortunately viewed it negatively. See bls short comment in the Archaeologiai Erteslt 94:2,1967,215.

156

157

Janos Makkay Of course, the existence of such vast territories hetween proposed homelands might offer some support for the Sprachbund theory, if supposed that numher of tribal groups speaking independent ( only partly related) tribal tongues of unknwn language types, lineages and stages lived in these territories. However, there is general consideration that should accounted for when we to consider some specific aspects. Supporters of the Sprachbund theory usually assign the period of !arge-scale integration of tribal even so-called group languages into larger protolanguage language family units to Pre-Neolithic i.e. Mesolithic times. the other hand, the Eurasian Mesolithic is associated with gradual adaptation to loca! resources and conditions in response to post-glacial environmental changes. Regional specialization and rationalization observed in the sphere of food-gathering and the various communities could follow various trajectories in space and time.75 This adaptation finally !ed to the diversification of the Mesolithic aterial and spiritual culture, and as result the Mesolithic assemblages show great variation region to region. This means that paralle! with the assumed language integration there would cultural disintegration and dijferentiation. Moreover, correlating with the assumedlanguageintegration processes during times cultural disintegration - the Sprachbund theory continues there developed true language families, arnong others the IE and UIFU ! languages. ! the supporters of the Sprachbund theory never indi8ate the causes and reasons for the apparently unwarranted and sudden change in the course of developments that then took place: after the postulated emergence of ! languages way of ethnocultural and language integration 'why did the process abruptly change direction without an ! s, and the final differentiation of ! languages begin? Nor is it possible to deny that these diverging processes (i.e. the final separation of IE UIFU languages) actually took place, since the differentiation of the speech communities of the parent languages into separate daughter tongues s continued ever since the Neolithic, and it an clearly docurnented as far back as the first occurrence of written IE (ittite) lingnistic sources from the early second millennium . As result of these and other considerations, suggested Mesolithic processes and their language outcomes an! reconstructed within the frarnework of relatively confmed and late Nostratic homeland in the Mesolithic. assumed of . Bomhard, however, gains credibility if we advance backwards into phases of the Upper Palaeolithic. ! such an early time (during the supposedly first arrival of relatively sall groups of sapiens from Northem Africa) confmed is particularly advantageous to further (biological, aterial and linguistic) developent as an original centre starting . Prospects and perspectives of Turkish prehistory in this ! of the country wiH surely contribute to the success of discovery of solid proofs.76

Before /ndo-European & Uralie Accordingly, it seems to that much better s could rad out for this decisive change l the much earlier time of the first an of sapiens fossilis ! the beginning of the Upper Palaeolithic weH hefore 15,000 . It an he assned that the spread of Modem an from Mrica an he equated with the initial dispersal of the genetic precursor of Nostratic languages. In this model mo sapiens would displace the Hominid populations of Asia and of Europe ( ereetus and mo Neandertha/ensis) around 35-40,000 years ago even much .77 According to ! researches the whole developent in the direction of the presently existing uman species, sapies sapiens, seems to have taken place in Mrica, perhaps as earlyas 100,000 years ago. thal continent our ancestors first crossed to South-West Asia and than continued 10 Europe (35,000 ago?). their way they tnight have integrated and completely assitnilated the communities of classical Neanderthals of the Near s! (! 2002, 3). Such assitnilating processes would have led 10 the sapientization of Neanderthals, and this an give us the possibility to tak further suggestions about the origins of Uralic peoples (Makkay 2000, 78-79). Integration and assirnilation an he applied with two reservations. he frrst is that fuH integration and assitnilation extended over territories of non-Nostratic languages which were occupied during the foHowing development and distribution of Nostratic groups including their very late and large scale spread during modem colonizations, mostly l the expense of small tribal langauges of Asia and Africa. State languages supported established religions (hin, Japan, India, the Moslim world) have mostly remained unaffected. Secondly, integration and assitnilation processes include ousting of groups of indigenous trihes their native lands lying periferies of expanding Nostratic languages as for exarnple the Lapponic in Scandinavia and soe Palaeoasiatic languages of Northem Asia. his process continues in these first , of the third iHennium. Even this assumption and also suggested Nostratic homeland in Southem and Eastem Turkey - annl solve complicated questions of the Uralic and Indo-European interrelations, whether they were genetically related (Nostratic) languages of independent origin (including now also the convergence theory). ! this mo! nothing an said these questions in responsible nn. view briefly descrihed says thal the time depth of I formation of the Indo-European groups must he pushed back ! least to the final phases of the Upper (Late) Palaeolithic, but in the s of Proto-Uralic to even times, to the iddle Palaeolithic age of sapies presapies, archaic sapiens or mo sapiens neaderthalensis, its equivalent: the sapietized Neaderthal he of emergence of these (Indo-European and Uralic) peoples and their languages is question to he decided: did both to Europe common Near Eastem centre fro

Nostratic homeland

together

with the dispersion

of the first Upper

Palaeolithic

cultures,

the Aurignacien and Gravettian, the ancestors of the Proto-Uralic stem were the local pre-Aurignacien (iddle Palaeolithic) trihes living the southem border of the lce

75 rv! 2002,6-12. 76 See especia11y " in Ozdogan ! 1. 2003. Recently see a1so Aurenche ! 1. 2004, esp. ' aout the transition from the Epipa1aeolitblc to the Neolitblc.

77 Szemereoyi

1989. 159.

158

159

Jdnos Makkay surviving during the Gravettian period as an independent tundra Ul.78Such Near Eastem origin of the Indo-European ran during the beginning of the Upper Paleolithic suggested Harrnatta and ikl6s aabori had obviously preceded the Renfrew model that was considered novelty at the time, a1most decade.79

/ 1ndo-European & Uralic where ( environment is [was?] favourabIe (contrary ( ( curious theory ofthe Mrican Eve). of the consequences was that the dernographic rate increased and led ( further movernents (Fig. 7). these elements contributed to the evolution of thinking and of language. time depth of these developrnents strongly need modification of the 'short' theory, i.e. the Kurgan theory which locates the early Indo-European distribution territory within the westem half of the Eurasian steppe. took the view that the Kurgan theory an! long withstand criticisD1, and also the 'Iong' theory presented Renfrew based
the diffusion of agriculture

is ! considered rnore convincing , 'despite of its

...

evident har' (Otte 1997,76). Otte has put forward his rnodel based his 'Eurasian Steppe PaIaeolithic Revolution' which an caIled the 'superlong' theory. sum it up, it says that foHowing to the transition ( anatornicaIly modem an the Eurasian steppes diffusion occurred between 40 and 35,000 years ago, towards Northem India (arrow 1), the Levant (Aurignacian around 32,000; arrow 2) and the Balkans (around 40,000; arrow 3), then along the ditanan route. "This radical break in the archaeological evolution is the only (! an explain profound ethnic modification eqnivalent to the appearance of the Indo-European peoples." Later processes during the European Upper PaIaeolithic and especiaHy the Magdalenian cultures probabIy led ( the fnati of the Indo-European language and cultures (Otte 1997, 80). Such synthesis of archaeological and lingustic processes is strong contradiction of what the Nostratic hypothesis and its archaeological interpretation says about suggested Nostratic homeland in the area between the Levant and Westem Iran, dated ( (hitherto undtnid) periods of the Upper PaIaeolithic. further progress to achieving better understanding clearly depends reconciling opposedIy forwarded beliefs and opinions the origins of modem man, its archaeological cuItures and time depths. Fig. 7: the Nuc/ear Zon o/the Eurasian steppes (east o/the Don river) at the " beginning / the , /. After Otte 1997, Fig. 5.3. . 79.

strongly opposing view was advanced . Otte in his short contribution questions of changes ! ( end of the Lower and again the iddle Paleolithic in Eurasia (Otte 1997, 74-76). According ( , ! the end of the Lower Palaeolithic muItiple technological innovations appeared evoking recognizabIe distinct traditions maintained autonornousthni groups. During the iddle Palaeolithic Mousterian tirnes (between 100,000 and 50,000) these (hnil innovations multiply and ( density of human sites increases. llis phase is crucial in Europe, because it directly precedes the appearance of anatornically rnodern man and what is called the 'Upper Palaeolithic' way of life. Evolutionary steps happened slowly and the outside probabIy in the steppes of Eurasia
78 For details see 2000 paper, 72. 79 Makkay 2000, 73-74, with further reference to 1976. 1977 and 1981 papers of the late Mikl6s G1Ibori.

160

161

Jcos Makkay

Before ldo-European & Ura/ic

References

Alinei,Mario.1996,2000.Originide//e /igud'Europa,vols.1-. ologna,. B

_ Aurenche, . . Le Miere - . Sanlaville (eds.). 2004. From /he River /0 /he Sea: /he Pa/aeo/i/hic and /he Neolithic the Euphra/es and ' /he Nor/hern Levant. R S1263. Oxford. Bienkowski, . - . Milard. 2000. DictionoryoJ/he Ancient Neor East. London. Cauvin, J. 2000. Their/h oJ/he Gods and the Origins ! Agricu/ture. Cambridge. Charvat, . 2002. Mesopo/amiabeJore his/ory. London- New York. Burkitt, Miles - . Gordon Childe. 1932. Chron%gica/ ( oJprehis/ory. Supp/emen/ /0 An/iquity,June 1932, columns2 and 3. Childe, . Gordon. 1939. The Orient and . AJA 44, 1939, 10-26, and the chronological tabIe. Gamkrelidze, Thomas . 1970. "Anatolian languages" and the fJd-uan

migrationinto Asia Minor. S/udies in Genera/ and Orien/a/ Linguis/ics,

presented /0 Shir6 Hattori. Eds. RomanJacobson and Shiego Kawamoto. Tokyo, 138-143. Garnkrelidze, .. and .. [. 1995. lndo-European and lndo-Europeans, [. Translated from Russian J. Nichols. Berlin - New York, 1995. Harmatta, Janos. 1989. Das Pelasgische und die alten Balkansprachen. Linguistique Ba/canique 9:1,1964,41-47. Laszl6, Gyula. 1961. Ostorree/iink/egkor6bbi szakaszai. (In Hungarian: The earliest periods of the prehistory of the Finno-Ugrians. Archaeological remains of the Finno-Ugric prehistory in the Soviet Union).Budapest, 1961. ed., Budapest. . 1987. Os/or/ene/iink (1Hungarian: prehistory).3" kk, Janos. 1989. ". . Childe chronological correlations between the Orient and ". In: Neo/i/hic! Sou/heas/ern Europe ond its Neor Eas/er connec/ions. VarioArchae%gica Hungorico2, Budapest, 177-181. 1992. " Neolithic model of 1ndo-Europeanprehistory". Jouro/ ! lndo-European _' S/udies 20: 3-4, 193-238. . 1996. "-14 Chronology: Eastem Europe." / Archoe%gico 67. Supplementa 1, 1996: Absolute Chronology.Archaeological 2500-500 , ed. lavs Randsborg. K0benhavn, 1996,219-225. . 2000. "The ear1iest prehistory of lndo-European and UliclFi-Ugi speaking peoples in Mikl6s Gabori's model". [: / ~echerche de /" Prehis/orique, sous /0 recherche de 280// Mester and Arp6d Ringer. E.R.A.U.L. 95, Liege, 69-82. 2001. "The earliest Proto-[ndo-European- Proto-Uralic contacts: Upper Palaeolithic model". [: / Con/oc/s Uro/icond lnd()-Europeon: Linguis/ic and Archoe%gico/ Considerotions, popers presen/ed / in/ero/iono/syposium he/d / Tviirminne,8-10 Jauary, 1999. Eds. Chr. Carpelan, . Parpola and . Koskikallio. Memo;res de la Soeiete FinnoOugrienne 242, Hensinki, 2001, 319-343.

...

2002. "Egy magyar mt velemenye az urali finnugorsag szarmazasar61, 11.resz: kk vege". (In Hungarian: Hungarianamateur's opinion the origin of the Uralie Finno-Ugrians, part II: The end of the Stone Age). Budapest, 2002. _.2003. " arehaeologist speeulates the origin ofthe Finno-Ugrians". The MankindQuarter/y 43:3, 2003, 235-271. 2003. Origins ! the Proto-Greeks and Proto-Ana/o/iansfro _' perspective. Budapest. Mallory, James . - D.Q. Adams (eds.). 1997. Encyc/opedia ! /ndo-Europea Cu/ture. London and Chieago: Dearbom. Mallowan, M.E.L. 2001. Ma//owan's Meoirs. London. Mallowan, M.E.L. and J.c. Rose. 1933. Excavatios ! //Arpachiyah, 1933. lRAQ 2, 1935, 1-178. Otte, Mareel. 1997. "The diffusion of modem languages in prehistoric Eurasia". [: Archae%gy and /aguage, /. 1. Theoretical and ethod%gica/ orientations. Eds. Roger Bleneh and Matthew Spriggs. London - New ork: Routledge, . 74-81. 6zdogan, . ] 999 = . 6zdogan: <;::U. (eds.) . 6zdogan and N. B~gelen: in Neolithic in Turkey, the eradle of eivilization. Istanbul, 1999, Text, 3563. 6zdogan, . and . Hauptmann- N. Ba~gelen. 2003. d Kente. From vilIages to eities. / villages in the Near East. Studies preseted to Uji/kEsin, vols. (-11.Istanbul,2003. Potts, .. 1999. Thearchae%gy ! /. Cambridge. Renfrew, Colin. 1987. Archae%gy and Language. The Puzz/e of/ndo-Europea Origins. London. _.1989. "Models ofehange in language and arehaeology". Trasactios ofthe Phil%gica/ Society 87:2, 1989, 103-155 and 172-178. 1991. "Before Babel: speeulations the origins of linguistie diversity". _' Cabridge Archaeologica/ JournaI1:1-3, 1990,3-23. _ 1997. "World linguistic diversity and farming dispersals". (: Archae%gy and /anguagu, /. 1. Theoretica/ and ethod%gica/ orientatios. Eds. R. Bleneh and . Spriggs. London- New York: RoutIedge, . 82-89. _ 1999. "Nostratie . linguistic maerofamily". (: Nostratic: exaiig lingustic acrofai/y. Eds. . Renfrew and . Nettle. Carnbridge,. 317. _ 2000. "! the edge of knowability: towards prehistory of languages". Cabridge Archae%gica/Joura/1 : 1, 2000, 7-34. Rosenberg, . 1999. " <;::emi".( Neolithic in Turkey, the crad/e / civilizatio. Eds. . 6zdogan andN. Ba~gelen.Istanbu1,Text, 25-33. Sherratt, . - S. Sherratt, 1988. "he archaeology of Indo-European: an altemative view". Atiquity62,1988,584-595. 163

_'

162

}(os Makkay

V Shevoroshk , V ., 1989 . " Symposium the deep reconstructionof langages and c1tres".In Reconstructing languages and cultures. Abstracts d Materials from the First /nternational /nterdisciplinary Sympos/um Language and Prehistory , 8-12 Noveber, 1988. Ed. .. Shevoroshkin.Bochm: km,. 6-8. . . . ~~~..k R S - R ., L Solecki - . . Age1arakis. 2004. The Proto-Neol/th/c 1 . .
'

VOWEL HARONY AND OTHER FORMS OF VOCALIC ASSIMILATION IN MONGOLIC Peter . Michalove
1. lt was in the 19'" try that scho1ars, primarily native speakers of lndo 1angages, began to systematically investigate the Uralic and Altaic langages and discovered the apparently nsal[eate of vowel harmony. hey conclded, the basis of the Uralic and Altaic langages, that vowel harmonywas [rm of (+back) agreement, both within the stem and between the stem and sffixes. ( recent work, however, has shown that vowel harmony in the Tungsic langages goes back to TR system. See Ard 1980.) Earlier researchers also concluded, the basis of these langages, that vowel harmonyis an exclsively progressive (rightwardfrom initial trigger) process. Nowadays of corse, we know that vowel harmony is ! sch an exotic feature;it occurs in nmerousAfrican, native American, and other languages. And despitethe difficlty of defining vowel harmonycross-lingistically (see Anderson 1980),(. is dobt that it is ( of vocalic assimilation, common phenomenon. We also know that vowel harmony is not exclsively progressive. There examples of vowel harmony spreading bidirectionally from non-initial triggersyllable, sch as the Bant langage !:'!), spoken in . There a!so cases of dominantJrecessive vowel harmony, sch as Chkchi and Koryak, spokenin the Rssian northeast. In these langages dominantset of vowels is opposed to recessive set: the phonological word contains vowels from the recessive set ! ifall ofthe vowels in the word recessive. Ifthe word contains dominant vowel in position, then the other vowels harmonize with it shiftingthe relevant featres to their dominant conterparts. hs, dominant vowel trigger vowel harmony progressively (rightwards) and regressively (leftwards). Tb.is will deal with the qestionof directionality in Mongolian vowe! hanand other forms of vocalic assimilation. While disptes that vowel hanoperates in exclsively progressive direction in Mongolian, it is striking thatalmost of the other cases of vowe! assimilation in Mongolian regressive. As the title of this indicates, we see vowel harmony simply as particlar f11 vocalic assimilation. The qestion before us, then, is ( find consistent of explanatiQnof vowel assimilation that covers both the progressive vowel harmony, as well as the other, overwhelmingly regressive cases of vowel assimilation in Mongolian. further note abot terminology is in order here. Classical Mongolian refelSto the literary langage of the 17'" and 18'" centries, written in the Uyghr 165

. Szemer~nYl,. 1989. "Conceming Professor Renfrew's . Vlews . the Indo. rohomeland". Transactionsofthe Phllologlcal Soclety 87.'2 , 156165.

Ceetery in Shanidar cave. Texas UvSlty Press, 2?04.

...

164