The Indo-European question reconsidered Ellegård, Alvar Fornvännen 1990(85), s. 128-132 http://kulturarvsdata.se/raa/fornvannen/html/1990_128 Ingår i: samla.raa.se

128

Debatt Referenser Animcrman. A. J. & Cavalli-Sförza, L. L. 1973. A population model for lhe diffusion ol early farming in Europé, Lhe explanation oj culture change: Models m prehistory. Ed. 1>\ (... Renfrew. London. — 1979. The wave ol .ulv.mee model tor ilu- spread of agriculture in Europé, Lranspirniations: Malhematical approaches la culture change. Ed. hy C. Renfrew k: K. L. Cooke. New York. — 1984. The Neolithie transition mid the gemiiis oj population in Europé. Princeton. Baldi. P. 1988. Rec. av Renfrew 1987, Current Anthropology 29. Ehrel, C. 1988. Language change and lhe material corrdates of language and ethnic shift, Antiquity 62. Gimbutas, M. 1988. Accounting for a great change, Times Literary Supplement june 24-^0. v. Heland, M. 1989. Makt, ordmakl ocli maktord, Maktens /ormer. red. Y. Hirdman, Stockholm. Malloiy, J. 1988. Rec. av Renfrew 1987, Antiquity 62. — 1989. ht Searrh oj the Indo-Europeans. London. Renfrew. t:. 1978. Dags aii omvärdera folkvandringarna, Forskning och Framsteg 1978: 8. — 1988a. Archaeolo^ and Language. Ilu- Puzzle of Indo-European Ori^ns. Author's Precis, Current Anthropology 29. — 1988A. Replv, Current Anthropology 29. Zvdebil, M. 1989. On lhe transition lo farming in Europé, or what was spreading with the Neoliihic: a reply to Ammerinan (1989), Antiquity 63. Pontus Hellström Meddhavsmuseet Box 5405, S-114 84 Stockholm

från arkeologins horisont förefaller vara cn elfenbenstorn ocb att få dem att hjälpa till an lösa ett för arkeologin grundläggande problem. Det går inte för sig att språkvetenskapliga teorier som leder fram till kulturhistoriska konklusioner får leva ett liv fritt från den arkeologiska verkligheten. Teoremct om ett indoeuropeiskt urhem på de ryska stäpperna, från vilket krigiska indoeuropéer strömmat ut ocb med vapen i hand betvingat Europa från 3000-talet f. Kr. ocb framåt har enligl Renfrew ingen reell arkeologisk verklighetsanknytning. Detta är dock ett påstående som inte alla arkeologer år villiga alt skriva under. Som den senaste exponenten för denna syn, i modifierad förin, lanserar nu Malloiy (1989) området norr om Svarta Havet o d i bort mot Kaspiska Havd som det indoeuropeiska urhemmet i en stor och välargumcntcrad genomgång av problemkomplexet. Detta visar att del indoeuropeiska urhemmet i sin traditionella form ännu inte är färdigt att avföras från kartan. Avslutningsvis kan j a g bara hoppas att Colin Renfrews provokation leder till snabbare slutliga resultat än vad Expansionsmodellen förutsäger. För övrigt ställer jag mig frågan om inte trots allt en så komplex framtida lösning på problemet måst c länkas att den kommer att bygga såväl på vissa resultat av den linguistiska paleontologin som på Expaiisioiismoddlen och på teorier om språkkonvergens. Tills vidare får vi vara tacksamma för de nya aspekter på problemet som Renfrews bok och den av honom initierade debatten har givit.

The Indo-European question reconsidered
Indo-European origins used to be a central topic of discussion among linguists, and to a lesser extent, archaeologists, during the century preceding the Second World War. But when Nazi Germany built its ideology of the Master Race on the myth of the blond and blue-eyed Aryans, the subject löst its appeal in the scholarly world. Publications on the subFornvännen 85 (1990) jecl bave been remarkably sparse during the lasl few decades. However, it seems the quarantinc is ovcr. In 1987, Colin Renfrew of Cambridge published his Archaeology and Language: The Puzzle of Indo-European Origins. Now, only two years after, J. P. Malloiy, of the University of Belfast, and assistant archaeological editor of

Debatt the (linguistic) journal of Indo-European Studies, presents us with a new full-length treatment of the subject. The two books are very different. Renfrew launches an overall hypothesis, namely, tbat the Indo-Europeanization of Europé came about as a result of the spread of farming, roughly between 7000 and 4000 BCE (a convenient abbreviation, increasingly populär among theologians, for Before lhe Common Era). Renfrew realizes, of course, that his overall hypothesis oversimplilies matters. Reality is always more complex tban the theories we invent to describe and explain it. But his main aim is obviously to establish his hypothesis as a serious competitor to the "established" one, propounded ovcr the last three decades above all by Marija Gimbutas, Lithuanian-born archaeologist at Berkeley. In her view the Indo-Europeanization was effected by several waves of horse-riding invaders from the Pontic steppes between 4000 and 2000 BCE. Their archaeological counterpart is the Kurgan (burial mound) culture of Southern Russia. Renfrew stresses the weaknesses of that theory, but does not attempt to cover all the relevant evidence. This holds bolh on the archaeological and on the linguistic side. Renfrew presents a very effective brief, but it must be admitted that it is a one-sided one. Mallory's book contains considerably more factual information, especially on the linguistic side. This means tbat Malloiy puts his reader in a much better position to draw his own conclusions. While Renfrew is an advocate who assumes that his readers bave already a general acquaintance with tbc field, Malloiy is more of a teacher concemed with displaying the facts of lhe case, induding their complexity. Undoubtedly Malloiy offers more meat for the beginner, while Renfrew provides a salutaiy cballenge for the more advanced student. Since Malloiy wrote two years after Renfrew, he naturally bad to take a stånd on the latter's hypothesis. But he does not allow his discussion of it to play a prominent part in bis book. By and large, Malloiy comes down on the side of lhe established theoiy, placing tbc
9-9081)42

129

origin of the Indo-Europeans in the North Pontic—Caspian region about 4500—2500 BCE. He differs, however, from Gimbutas by extending the original territory westwards to include also the Donets—Dnepr area. Malloiy arrives at this condusion above all on the basis of "linguistic paleontology", the shared Indo-European vocabulary. Agricultural terms like corn and to plough, names of domestie animals like sheep, cow, and horse, names of technical inveiitions like wheel and wagon, terms of social organization like Latin rex and Sanskrit raj, all point to a stage of development which roughly indicates the period 4500-2500 for the area concemed, southwestern Eurasia. Further, Malloiy points out that the considcrable number of apparently early Indo-European loanwords in Finno-Ugrian languages, and the meagre harvest of traces of linguistic exchange between Indo-European and the Semiiic languages, favour a location to tbc north ratber than to the south of the Black Sea and tbc Caspian. As for the Anatolian languages, there are many hints that they are intruders rather than aboriginal in Asia Minor. On this basis, Malloiy finds a strong case for placing the original Indo-Europeans where he does. What I would like to question is the base itself. Unlike Renfrew, I do not think we should reject the evidential value of linguistic paleontology. Without it, we have hardly any foundation at all. The Indo-Europeans are necessarily and by definition people speaking some Indo-European language. It is through their language tbat we identify them. On the other band, I wisb to query Mallory's assumption that we cannot go further back tban the year 4500 BCE in our reconstruction of tbc Indo-European speecb community. Ii is true that we have no evidence of the domesticated horse before c. 4000 BCE, and that, in the relevant area, we do not encounter the wheel until c. 3500 BCE. But in the first place we cannot c o n d u d e that the words must have come into being before the IndoEuropean languages had diverged from each other. To simplifv a little: centurn speakers may have recognized that a satem term like, for instance, *esvos, could bc loan-translated into
Fnrrwnnnen ,S'i (1990)

130

Debatt the end of the paleolithic were at least in the main developed out of the languages brought by the various populations of Homo sapiens sapiens which first spread into these parts of the world some 30,000-60,000 years ago. This does not by any ineans imply that they then all spöke one and the same language. There is no more reason to assume tbat all humanity ever spöke tbc same language, tban that they arose from a single pair. O n the contrary, the evidence that we have from present-clay hunter-gatherers, in Australia and in Soulb America, is that the groups speaking a certain language are normally quite small, from a few hundred to a few thousand. Historical evidence suggests that a few thousand years of separation is enough for two originally similar languages to split into mutually unintelligible ones (though, to be sure, linguists will recognize their relatedness). With a time perspective of several tens of tbousands of years, and speech communities of a few thousand each, it is probable tbat many hundreds of mutually unintelligible languages were spöken in paleolithic Europé. They were indeed distantly related to each other, like the American Indian languages of to-day. But again, like those languages, they would be so different from each other tbat even professional linguists would not bc able lo assign them to a common stock. However, wben the European languages begin to come into view, in the second and first millennium BCE, they are in fact demon-

their own *ekwos. Moreover, the wheel terms might be natural extensions of older words for "roll": such semantic shifts occur all the time, and are among tbc well-known pitfalls of linguistic paleontology. In the second place, we must remember that our discussion of the original Indo-European language rests on a rather narrow foundation. Even if we demand no more tban tbat the word in question should be found in at least three language families, with at least one each from Europé and Asia, the number of such words (or rather, roots) is no more tban a few hundred, out of tbc tens of ibousands that we know are in current use in any natural language. Though linguistic paleontology is an indispensable tool, we must never forget how narrow and dim is the window that it opens u p on tbc linguistic reality of the past. We must assume that the original home of the Indo-European speakers covered a considerahly smaller area tban the one in which we find the daughter languages wben they first come into view, some of them as early a.s c. 2000 BCE. Let us look a little more dosely into this matter, which will take us into tbc field of anthropology and human origins. Thanks to advances both in paleontology and in molccular biology we can now say with some certainty that all modern humans belong to the same subspecies, Homo sapiens sapiens. The origin of tbat sub-species should almost certainly be sought in Africa, perhaps 100,000 years ago. From there, human populations spread into Eurasia, and eveiitually into Australia and America, which were reached perhaps 30,000-40,000 years ago. Work done especially by L. L. Cavalli-Sförza et al. (1988) reveals a striking amount of correspondence between biological measures of relatedness för tbc world's aboriginal populations, and the major language families of tbc world.

strably related to each other. A considcrable

linguistic unification must bave taken place. O u r problem is, when? Malloiy, as we have seen, opts for the period 4500-2500 BCE, refusing to speeulate about what may have happened before tbat period. I think tbis is a mistake, since we then have to explain virtually tlie whole of the Indo-European expansion into Europé as due to migrations during that We have indeed many indications from hisperiod, for which wc bave very little bard torical times that a population has switched evidence. I think Renfrew is right in insisting from one language to another. But such tbat massive replacements of whole populashifts, at least across language families, have tions seem very unlikely in the largely agriculapparently not been very common in tbc past. tural Europé of that time. Elitc invasions apAc < ordingly there is reason to believe that tbc pear a little more plausible, hul only towards languages found in soutliwestern Eurasia at tbc end of tbc period. Hence I find Renfrew's
Fomvåmun 85 (1990)

Debatt

131

hypothesis appealing, since it connects the linguistic unification with an archaeologically idcntifiablc technical and economic innovation, which is also demonstrably able to explain the demographic change without largescaie invasions. The many small, old European languages, differentiated över a period of tens of tbousands of years, could thus be replaced by a new set, whose differentiation had only recently started. Linguistically, too, the longer time perspective provided by Renfrew's theory seems to mc highly desirable. Even at their first appearance some 4,000 years ago the Indo-European language families are remarkably differential cd. Mycenean Greek is veiy clearly Greek, far different from its near contemporary Hiitite, and also from its neigbbour, Latin, which first comes into view in the middle of the first millennium BCE. Also, to j u d g e from the Indo-European loanwords in the Mitanni kingdom of North Mesopotamia, c. 1500

BCE, tbc distinctive character of Indo-Iranian
was already apparent at that lime. In view of what has happened to those languages in the subsequent 4,000 years, it seems to me very unlikely ihat tbc earlier differentiation could bave taken much less time to develop. The mere iwo millennia tbat tlie Gimbutas model allows is simply not enough. Now Malloiy does not rcallv follow Gimbutas. Rather, he accepls that the Indo-Europeans, around 4000 BCE, occupied a wider area than the original Kurgan culture. He also allows tbat they may have been linguistically differentiated. But he refuses to speculate on ihal differentiation, since linguistic paleontology does not c a n y him further than the domesticated horse and the wheel. No doubt our reconstruction beyond that period is hypothetical. But wc are not completely in the dark. We know sometbing of tbc råte of change in languages, and about the factors inducing a population to adopt a new language. Some of ibosc factors—differential population growth, migration, élite dominance—may leave archaeologica] traces. It is here. 1 lliink, that Rcnfraw's hypothesis comes into its own. l.ci mc finish this review by presenting an

outline sketch which may be looked upon as a synthesis of Renfrew's and Mallory's views. At the end of the Paleolithic, the southwcstern parts of Eurasia, together with Northern Africa, were inhabited by groups of huntergatherers who had lived in those regions for tens of tbousands of years. There were probably bundreds, if not tbousands, of mutually unintelligible languages there, each spöken by a few hundred or a few thousand people only. O n e part of this väst region, Eastern Anatolia—Northern Mesopotamia—Iran, saw the origin of agriculture. O n e of the peoples who adopted the new economy were speakers of a language which was lo develop into the IndoEuropean group. However, that was only one mu of dozens or more languages in the area, among which, many thousand years låter, we can identify not only Semitic ones, but also apparently quite u n r d a t e d ones such as Sumerian, Hattic and H u m a n . If we assume that lhe original Indo-Europeans lived in the Western parts of Anatolia, and that their adoption of agriculture led to a marked population increase, then their natural search for virgin land—not already occupied by other agriculturalists—would lead them towards Europé, and the expansion would not be likely to stop until it reached the western and northern ends of that continent. Needless to say, the spread would not be uniform, for various reasons. O n e of them has to do witb soil and climatc. Another bas lo do with demograpby. In some places, the indigenous populations may have adopted or developed agriculture themselves, before the general European wave of advance originating in Anatolia reached their region. The remnants of non-Iiido-European languages in Spain (Basque, and perhaps Iberian), Italy (Etruscan), and Greece ("Pelasgian") may be accounted for in this männer. The main wave of advance, accordingly, may have been towards the northwest and west, following the Danube, and towards the north and east, into tbc Pontic steppe. In tbc fifth millennium BCE the steppe Indo-Europeans domesticated the horse. Considerably låter they, or some of tbem, developed a nomadic economy, and also a more
Fornvännen S, (1990)

132

Debatt the Finno-Ugric languages in the north, while the scarcity of the Semitic-Indo-European correspondences is natural, if the first, and not very numerous, Indo-European speakers were separated from the Semitic speakers by such already agricultural communities as the Hatti and the Hurrians. The Hittites may, as Malloiy insists, have been newcomers in the region where we first come across them. But they need not have come from afar. And the fairly extensive similarity of Armenian and Greek may be due to a common origin north of the Black Sea in the third millennium BCE. The above is of course no more than a possible scenario. Necessarily so, since we are trying to trace a development embracing some 10,000 years, of which only the last 4,000 contain at least some bard linguistic evidence. Still, information continues to accumulate in such neighbouring fields as archaeology, anthropology and biology. The present-day state of the art is very well covered in Mallory's book. If read together with Renfrew's, it provides an excellent starting-point for whoever wishes to take u p again the fascinating problem of Indo-European origins. References Cavalli-Sförza, L. L., et al. 1988. Reconstruction ol human evolution: hringing together genetic, archaeological and linguistic data. Proceedings of the National Academy of Scienees of the USA, 85. Mallory, J. P. 1989. In Search of the Indo-Europeans. Thames 8c Hudson, London. Renfrew, C. 1987. Archaeology and Ixinguage: The Puzzle of Indo-European Origins. Jonathan Cape, London, 1987. Renfrew, C. el al. 1988. Discussion of Archaeology of Language, in Current Anthropology, 29. Alvar Eliegård Rattgatan 2 S-42176 V. Frölunda

stratified social system exploiting tbc military possibilities of the riding horse. This is when the Kurgan culture begins to expand. That expansion was naturally directed towards the already rich agricultural regions, such as India, Anatolia and Greece, where there were riches, and a food surplus, to exploit. Låter, historically documented nomad invasions have taken a similar course—Huns, Turks, and Mongols. Against this background, it is quite possible to regard Hittite and Luwian as the direct descendants of the "original" Indo-Europeans. And we seem to have no reason to regard the Thracians, Illyrians, Slavs and Balts, Germans, and Celts as other than descendants of those Indo-Europeans who came with the first wave of advance. The same may well be true of the Tocharians, who would therefore have arrived in Turkestan before the nomadic Kurgan culture developed. But with the development of horse riding, of new military techniques and of new forms of organization, during the third millciiniuin BCE and after, conditions arose for language shifts due to élite dominance. The spread of Indo-Iranian languages south into Iran and India in the third and second millennium BCE is probably a case in point. So is also, at the other end of the world, the spread of Celtic in the first millennium BCE into Britain, where it presumably replaced earlier Indo-European languages. Very possibly the spread of Greek, and of Armenian, into the areas where they are now spöken, may be similarly late shifts due to invading élites. This kind of scenario, it seems to me, makes a great deal of sense both archaeologically and linguistially. The long stay of the Indo-Europeans in the Pontic steppe, and their secondary expansion from that area, accounts for the fairly substantial number of words shared with

Fornvännen Ä5 (1990)