The Paleolithic Indo-Europeans


Until very recently, it was universally believed that the major language families of Europe had separated from one another in the course of a general migration from east to west.  Both the old chariot-warrior hypothesis and the more recent Neolithic hypothesis of Indo-European were based on this idea of a branching pattern of diffusion from a homeland somewhere in or beyond eastern Europe.  There could obviously be no similarity between this scenario and the archaological story of a post-Ice Age repopulation of Europe flowing along several routes from south to north.

However, information recently gained through DNA analysis not only weighs heavily against the old east-to-west scenario for the arrival of the Indo-European languages in their present locations but also offers a simple and persuasive alternative.  This radical shift in perception hinges entirely on the history of the Celts.

There are presently three Indo-European languages families in western Europe -- Italic, Germanic, and Celtic.  Most of the region is dominated by  Italic languages in the south (including Italian, French, Spanish, and Portuguese) and by Germanic languages in the north (including English, Dutch, German, and the Scandinavian languages.)  The Celtic languages are confined to a marginal position out on the Atlantic fringe, -- in Ireland, Scotland, Wales, Cornwall, and Brittany -- and even there they are threatened with extinction.

However, two thousand years ago -- back when Julius Caesar wrote that "Gallia est omnis divisa in partes tres" -- things were very different.  A thriving array of Celtic tribes and kingdoms then dominated not only Gaul but much of Europe.  There were Celtic-speakers in Ireland and in England, in France and central Spain, in northern Italy and southern Germany, in Austria and the Czech Republic, in Romania, and even on the far side of the Black Sea in Anatolia.  At this time, Latin -- the ancestor of all the modern Italic languages -- was still native only to Italy, while the Germanic languages were restricted to northern Germany and Scandinavia.

Although Caesar could not have been aware of the fact, this situation of Celtic pre-eminence was relatively recent in his day, probably extending back no more than a thousand years or so, and it was destined to last for only a few centuries more.  Much of it was the product of elite conquest, which may explain why it proved so vulnerable when it came under pressure from the Roman Empire to the south and the German barbarians to the north.

However, the relevant issue here is not the dynamics of Celtic dominance but its source -- exactly where had all those Celtic-speakers originally come from?

A century ago, when the Indo-Europeans were regarded as Bronze Age chariot-warriors from the steppes, it seemed obvious that the Celts must have originated near the eastern end of their historic range and spread westward, probably reaching England well after the era of Stonehenge and the other megalithic monuments.  The Neolithic hypothesis modified this scenario slightly by starting the expansion a few centuries earlier and bringing the Celts to England late in the megalithic period, but it maintained the idea of a relatively recent eastern origin.  This had the undesirable result of leaving the Celts kicking their heels aimlessly in some unidentified corner of central Europe for thousands of years, but there seemed to be no better alternative.

Recently, however, the notion of an east-to-west expansion for the Celts has broken down completely in the light of genetic analysis.  The primary factor arguing against it is that there turns out to be an extremely sharp dividing line between the Y-chromosome DNA of the Celtic-speakers of Ireland, Wales, and Scotland and that of the English, French, and Germans -- a dividing line that has remained constant ever since northern Europe was resettled at the end of the Ice Age.  In fact, the DNA of the insular Celts is so specialized that there seems to be no possibility of any later arrivals, not even to the limited degree that might be found in the case of elite conquest.

These findings have recently inspired archaologists like Professor Barry Cunliffe of Oxford to offer the revisionist view that the spread of the Celtic languages proceeded not from east to west, but from west to east.  He believes that this expansion was originally propelled by the enormous prestige of the navigators and astronomer-priests who carried the megalithic culture of Ireland to England, France and southern Germany a couple of thousand years before the Celts' final brief period of military dominance.  If this theory is correct, then the Celtic homeland must have been in precisely the areas around the Irish Sea where Celtic languages are still spoken today.

Cunliffe's theory is a radical one, not least because it calls into doubt so many assumptions about the nature of cultural superiority.  The chariot-warrior hypothesis of Indo-European was born of the era of European imperialism in the late 19th century, which exalted war-making abilities as a prime example of Darwinian survival of the fittest.  The Neolithic hypothesis had its own ideas about cultural superiority, which reflected the late 20th century era of American hegemony and market-based globalization.  In contrast, Cunliffe's revisionist image of cultural change as something carried abroad by mystics, geeks, and wandering poets may offer a foretaste of the new values of the 21st century.

One additional piece of evidence in support of Cunliffe's ideas comes in the form of what is known as the "Old European hydronymy," a seemingly exotic phrase which simply refers to the system of river-names in the region.  River-names are important in the study of prehistory because they tend to be retained even when the local language changes.  (Think about how how many rivers in America still bear Indian names like Mississippi or Housatonic.)

Many of the ancient river names of Germany, France, and England -- such as "Rhine," "Rhone," and "Thames" -- appear to arise from a common system of naming, and although these names are Celtic in form, they do not appear to be Celtic in origin.  There has been a great deal of lively argument as to what this "Old European" language might have been, but it is widely accepted that it was probably Indo-European and most likely an early form of Germanic.

However, even if the Celts did spread eastward from Ireland and western Britain, those areas could only have been only a secondary staging point.  The original Celtic homeland has to have been located in the west of the Iberian peninsula, the source of those seafarers who settled Ireland at the tail end of the Ice Age.  Irish DNA gives powerful testimony to this -- it is almost identical to that of the Basques, who would have been their immediate neighbors before the northward migration.

There is linguistic evidence for this migration as well, in the form of an obscure but apparently Indo-European language called Lusitanian, which the Romans encountered when they colonized western Iberia.  The handful of surviving inscriptions in this language suggest that it had distant Celtic affinities, and yet it was not at all similar to the Celtiberian languages of central Spain, which had arrived from France as part of the recent Celtic expansion.  It seems quite possible that Lusitanian was a survivor of the proto-Celtic spoken in the late Ice Age.

Late Ice Age migrationsWith the relocation of the Celtic homeland to the Atlantic fringe, the linguistic map of western Europe falls neatly into place.  Proto-Germanic was the language of those late Magdalenian reindeer hunters who migrated from southern France to England, Germany, and Denmark  around 16,000-13,000 BP.  Proto-Celtic was the language of the Iberian seafarers who set out for the north about 10,000 BP.  And proto-Italic was the language of the people who lived in central Italy both during and after the late Ice Age and didn't go anywhere at all.

The separation of these three language families may have become final as a result of their various migrations, but it must have begun some thousands of years earlier, when they were isolated from one another during the Last Glacial Maximum.  Italic would have been cut off first by the glaciation of the Alps -- presumably before 20,000 BC, since the Solutrean culture never reached Italy.  Celtic and Germanic, which are thought to be more closely related to one another, would have been separated both by the Pyrenees glaciers and by the Basques of northeastern Spain, but the split is likely to have been both later and more partial.


Just as is the case with archaeology and genetics, Indo-European linguistics displays a fundamental division between eastern and western Europe.  This division is sometimes spoken of as the centum/satem split, because one of its more obvious markers is that the Celtic, Germanic, and Italic languages all have (or formerly had) a hard "k" sound in certain words where the eastern languages have an "s."  ("Centum" is the Latin word for "hundred," while "satem" is the same word in Avestan, the ancient Iranian language of the Zoroastrian holy texts.)  If the western languages had already started splitting up by by 20,000 BP, then the separation between east and west must have begun even earlier, at least on the level of pronunciation.  There may have already been distinct eastern and western accents emerging during the long millenia of increasing cold between 25,000 and 20,000 BP.

The details of linguistic history are more difficult to reconstruct in the east than in the west, in part because so many long-distance migrations and invasions have flowed back and forth over the open glasslands and steppes.  Nonetheless, there are some useful conjectures that can be offered.

Late Ice Age migrationsThe situation of the Balto-Slavic languages is the most clear-cut.  The Baltic family today (orange) consists solely of Latvian and Lithuanian, although other members of the family were spoken in northeastern Poland as recently as a few centuries ago.  These languages are extremely conservative, retaining many archaic elements, and give every sign of having been firmly established in their present location for many thousands of years.  There is thus good reason to associate them with the people who migrated north from the Dniepr River to the Baltic coast about 13,000 BP.

The history of the Slavic languages is more complex, but it is generally believed that their original homeland was somewhere in the vicinity of the Pripet Marshes, which lie just west of the Dniepr along one of its major tributaries (pink).  This would suggest that the speakers of proto-Slavic were simply the people who stayed behind when their proto-Baltic-speaking cousins moved north.

Proto-Balto-Slavic, in turn, may have been part of a larger eastern European linguistic community before the Glacial Maximum.  There is a string of very similar river-names -- all of them derived from the Indo-European "danu," meaning "river -- which seems to define the original territory of this community.  As the growing cold isolated one river-valley from another, the eastern dialect of Indo-European would have broken up into three daughter languages:  Proto-Balto-Slavic on the Dniepr, proto-Indo-Iranian on the Don, and what we might call proto-Balkan along the Danube.

This third family is now nearly extinct, but at the time of ancient Greece there were a number of languages in the northern Balkans of which only fragments have been preserved.  To the northeast, Thracian and Dacian were spoken on either side of the Danube River where it flows into the Black Sea, an area corresponding to the present-day countries of Bulgaria and Romania, .  To the northwest, Illyrian was spoken between the Danube and the Adriatic, in roughly the territory of former Yugoslavia.  (Modern Albanian is thought to be a descendent of Illyrian, though it has been so much altered by later contact with Greek and the Slavic languages that its original affinities are hard to decipher.)  All these languages seem to be part of the satem group and display similarities to Balto-Slavic, making the idea of a common ancestor before the Glacial Maximum quite plausible.

The history of proto-Indo-Iranian is in many ways as hazy as that of proto-Balkan, but for very different reasons.  Far from being extinct, the contemporary Indo-Iranian languages form the most successful branch of Indo-European.  They are spoken by hundreds of millions of people and have produced a disproportionate share of the worlds' literary and philosophical masterpieces.  And yet our knowledge of their circumstances prior to 1500 BC is sharply limited, and it has never been possible to connect them decisively with any archaeologically documented culture.

The Iranian languages were relatively late arrivals in the Middle East, reaching Iran from Central Asia and northern Afghanistan about 1000 BC.  In contrast, the Indic group, consisting of Sanskrit and its descendents, seems to have been in northwestern India much longer.  This is an area of great cultural continuity, going back to at least the start of the Neolithic, and there are no traces of any pre-Indic languages even in the form of river-names.  It is thus altogether unlikely that the speakers of Sanskrit could have entered India as recently as 1500 BC, as claimed by the chariot-warrior hypothesis, or even around 3000 BC, as suggested in some versions of the Neolithic hypothesis.  The only question is just when they did arrive.

DNA studies make it clear that there was a substantial population influx from Central Asia into India at some point, and the most likely moment for that to have happened was near the end of the Ice Age.  During and just after the Last Glacial Maximum, northwestern India was in the grip of desert and semi-desert conditions.  Then, about 15,000-13,000 BP, the monsoons returned to their more normal pattern, bringing with them a renewal of moisture and abundance.  At this time, the area could easily have been repopulated simultaneously from the west and from the east, producing a mixed population and a distinctive Indic culture.

Indo-Iranian migrationWith this in mind, it becomes possible to trace out a tentative scenario for the migration of the speakers of proto-Indo-Iranian.  Their first movement would have been a short hop eastward, from the River Don to the northwestern shore of the Caspian Sea.  Once there, they would have found themselves squeezed between the speakers of Caucasian languages in the mountains to their south and the Uralic-speaking ancestors of the Finns and Estonians on the middle reaches of the Volga to their north.  The only direction in which they could have continued onward was around the upper end of the Caspian and along its eastern shore.

It is likely that this section of the journey occurred during the Glacial Maximum.  At that time, the major rivers of Siberia were prevented by the ice cap from flowing north, so instead they flowed south into the Caspian and Aral Seas.  The Caspian was greatly extended, mostly at its northern end, the Aral was roughly twice its post-Ice Age size, and one or more rivers flowed between them.  Even though the area east of the Caspian was generally deep desert at this time, with almost no rainfall, the local environments along the shores and rivers would have formed lush and welcoming oases.

The next stretch of the itinerary would have brought the Indo-Iranians southeast along the Amu Darya, which flows into the Aral Sea from the hills of northern Afghanistan.  This is precisely the area that was at the heart of the proto-Iranian homeland in early historical times.  It is also a territory that played an important role in the development of agriculture.  The people who lived in the Afghan hills between 15,000 and 10,000 BP were the first in the world to domesticate goats and sheep.  Domestic plants also appeared at a very early point in the entire area between northern Afghanistan and the Caspian Sea.

It seems highly probable that those early shepherds and farmers were speakers of proto-Indo-Iranian.  They may even have begun their experiments in domestication while they were still confined to the oasis spots around the Caspian and Aral.  In those places, plant and animal life would have been densely concentrated and an intensive use of resources would have come naturally.

In the final leg of the journey, the speakers of proto-Indic would have left the Amu Darya to press on through the foothills of the Hindu Kush and into the valley of the great Indus River -- probably no later than 13,000 BP.  When agriculture reached India from the northwest about 10,000 BP, it would not have represented the arrival of a new people but merely the result of cultural transmission among close relatives.

Previous | Next

Return to Worlds of Wonder
Return to Trogholm

Background courtesy of Eos Development