The spread of Indo-European languages in the third millennium bc

Importantly, if somewhat circuitously, an argument about the spread of Indo- European languages through the steppe proceeds smoothly enough from the premise that the Maikop culture, with its cartwrights and wainwrights, was an important part of a wider PIE language community. It is now widely recognized, thanks in large part to arguments put forward by Andrew Sherratt and David Anthony, that wheeled vehicles enabled the exploitation of the grassland steppe during the late fourth and the third millennium bc.⁷¹ During this period the Yamna (pit-grave) archaeological culture spread over the grassland steppe from at least the Dnieper river in the west to the Ural in the east, and Indo-Europeanists are quite sure that Indo-European languages were spoken in most if not all of the Yamna culture.⁷² Before the Yamna period food-producers had lived along the rivers of the steppe, where fish were plentiful, winter foraging was available for animals, and agriculture was not terribly difficult. Away from the rivers, the steppe had been roamed by hunters and foragers (many hunter-gatherers of the Dnieper- Donets culture seem to have lived on the open steppe) but had not been occupied by food-producers. Out in the steppe the sod was too tough for farmers in the fourth millennium bc to plow, although it did offer ample vegetation for grazing.

Wool, in contrast to milk and meat, was durable and marketable, and when woolly sheep began to replace hairy sheep (wild sheep have a hairy coat, and so did the early domesticates) the steppe may have become more attractive. As noted at the beginning of this chapter, all the branches of Indo-European inherited from PIE a word for wool and we must assume that wool was of some importance in the PIE culture. The earliest woolen textiles that archaeologists have found anywhere date to the fourth millennium BC. There is indirect evidence that in western Iran and a few other places certain varieties of sheep may have been bred for both meat and wool already in the sixth millennium BC, but Bill Darden concluded that “significant indications of wool technology do not occur anywhere until the fourth millennium.” Eventually wool became an important item of trade, and in Mesopotamia “workers’ wages were paid in wool, and exports of woolen cloths paid for the imports of raw materials.”⁷³ How important wool sheep were in the Yamna culture is debated. Some, but not all, osteological evidence suggests that the Yamna pastoralists depended more on their cattle than on their sheep.⁷⁴

However that may be, late in the fourth millennium BC it evidently became profitable to take flocks of sheep as well as herds of cattle deep into the steppe. Exactly how this pastoralism worked in the Bronze Age is not entirely clear. Until 1000 bc the steppe was not fully nomadic, because the pastoralists coexisted with agricultural villagers living along the river valleys (by ca. 700 bc all the villages were gone and the steppe was fully nomadic). The pastoralists of the late fourth and most of the third millennium bc are identified with the Yamna archaeological culture, which is well known from its burials in kurgans (each of which usually contained multiple burials) and most of the kurgans lie in the deep steppe. The pastoralists must have kept their flocks and herds not too far from the kurgans that they erected, but very few settlements have been found. This is in contrast to the subsequent Srubna (Timber Grave) period, when a good part of the steppe population lived in small settlements. A typical community of Yamna pastoralists must have had a temporary settlement, or nothing more than a camp, and moved from one spring, water hole or well to another when the immediately accessible steppe began to be overgrazed. Although evidence for agriculture is slight, a few crops were evidently planted and harvested.⁷⁵

Sherratt and Anthony argued convincingly that what allowed people to live on the open steppe, and to provide fresh pastures for their flocks and cattle over an immense area, was the wagon. In the first millennium bc the “Skythians” had no houses or huts but lived in their covered wagons,⁷⁶ and what evidence we have (and also what evidence we do not have) indicates that most pastoralists did the same already in the late fourth and the third millennium bc. The pastoralist’s wagon would have been covered with a canopy or tilt: a fabric—probably of felt—stretched over a wooden frame.⁷⁷ Every wagon was ponderous, with heavy planking and with tri-partite disk wheels, often a meter or more in diameter.⁷⁸ For the ancient pastoralists about whom we have information wagons were their “home”: in a wagon they were conceived, in a wagon they were born, and their hope in life was to become wealthy enough to own several wagons, so that at their death they could be sent to the Underworld with a wagon.

Some of the vehicles that the nomadic pastoralists used may have been twowheeled carts. Wagons carried more than carts, but in the fourth and third millennia BC wagons did not yet have pivoting front axles and were therefore difficult to turn. A two-wheeled cart, the wheels revolving on a fixed axle, could turn with no difficulty at all. The animals that drew both carts and wagons were of course oxen, usually castrated bulls. Teams of oxen in paired draft were sometimes sacrificed and placed in the burial chamber of a kurgan along with the funeral wagon.⁷⁹

The nomadic pastoralism that wheeled vehicles made possible must have spread the language of the vehicles’ occupants through much of the Pontic-Caspian steppe and well beyond. A great virtue of Anthony’s The Horse the Wheel and Language is its detailed argument for a connection between wheeled vehicles and what seems to have been a remarkable expansion of Indo-European languages in the third millennium bc. By the end of that millennium almost everyone in the Yamna archaeological culture, which extended over half a million square miles in the grassland steppe and even into the forest steppe, may have been speaking a language descended from PIE. During the third millennium bc the Dnieper-Donets culture of hunter-gatherers disappeared. We must suppose that the hunter-gatherers of this and other Mesolithic cultures were converted either to mobile pastoralism or to settlement on the river villages, and in either case they probably learned the pastoralists’ language and forgot their own.⁸⁰ Over the vast territory from the southern Urals to the Dnieper the evolution from PIE into several daughter languages will have begun.

The expansion of PIE eastward from the Volga river in the third millennium bc evidently led to the formation of various languages, including what by 1500 bc can be recognized as Indo-Iranian. Michael Witzel has proposed that the language ancestral to Indo-Iranian was a lingua franca or a koine along the steppe as far to the east as the upper Yenisei river (the Yenisei languages, now almost extinct, included some loan-words from a very early stage of Indo-Iranian).⁸¹ Witzel also suggested that the language ancestral to Indo-Iranian was carried south from the steppe into the semi-desert and into the BMAC culture, where its speakers borrowed words that would eventually appear in both Vedic Sanskrit and Avestan. Still further to the east, the expansion of PIE apparently brought to the Tarim basin the seeds of what eventually would evolve into the Tocharian language, attested by Buddhist texts from the first millennium CE. Whether the earliest of the famous mummies from the Tarim basin were related to the continuation of such an eastward expansion is not certain, but James Mallory and Victor Mair have cautiously concluded that they were.⁸² DNA analyses tend to support that conclusion.

As Proto-Indo-Iranian began to form east of the Volga in the third millennium bc, at the same time Proto-Baltic apparently began to take shape west of the Volga and eventually to have been brought into northeastern Europe.⁸³ In modern times the Baltic subgroup of Indo-European, with some very archaic features not far removed from PIE, is represented only by the Latvian and Lithuanian languages and was for that reason labeled “Baltic” by nineteenth-century philologists. Once upon a time, however, “Baltic” languages were apparently spoken over a much larger part of Eurasia. Because the oldest documents in any Baltic language are the catechisms written in Lithuanian and Old Prussian (now extinct) during the Lutheran Reformation, evidence for the extent of Baltic in ancient times comes mostly from river names. Mallory noted the presence of Baltic river names somewhat to the west and far to the east and the south of Latvia and Lithuania, and he reasonably assumed that various Baltic languages had been spoken there until they were submerged by the expansion of Germanic and Slavic languages in the first millennium CE. According to Mallory, “the minimal view would see the Balts, during the first millennium BC, occupying the area from west of the Vistula’s mouth east to Moscow and the upper Volga (which itself carries a Baltic name), and south almost as far as Kiev.”⁸⁴ Because “Baltic” names are attached to rivers more than 600 miles from the Baltic, Wolfgang Schmid argued that the PIE homeland was near the Baltic, and that migrations from that homeland spread the Indo-European languages through much of Europe and Asia.⁸⁵

Recent DNA analyses indicate that demic diffusion tended to go in the other direction: in the third millennium bc many people from the east moved westward to the Baltic and beyond.⁸⁶ Population geneticists have found that although people living in northern Europe in the fourth millennium BC had little or no DNA indicating what the geneticists call a “Yamna” (or “Yamnaya”) ancestry,⁸⁷ throughout the third millennium bc northern Europe experienced a considerable influx of people from the “Yamna” cultures of the east. More specifically, the geneticists concluded that the population of the Corded Ware culture of northern Europe (and to a lesser extent of the Bell Beaker culture) was in large part descended from ancestors who came from the Corded Ware culture further east.

The Corded Ware culture, which lasted through almost all of the third millennium bc, is an archaeological assemblage that stretched from the upper Volga to the Rhine. It is known mostly from graves, and although considerable evidence for settlements has been found the settlements were apparently small and of relatively short duration. Domesticated animals were important in the Corded Ware culture, and many of the people may have been mobile pastoralists. It is not at all unlikely that over the centuries a drift from the Yamna culture headed northward and westward and helped to form the Corded Ware culture. Northeastern Europe— from the Baltic to the Rhine—was in several respects more attractive than the lands that lay to the east in the same latitude. Agriculture and animal husbandry were far better established in the west: the Corded Ware culture was essentially the first culture east of the Baltic to depend on domesticated animals rather than on hunting,⁸⁸ while the Neolithic had come to northern Europe 2000 years earlier. Winters in northern Europe were not so cold as those in Russia. Also important was that proximity to the Baltic and the Kattegat extension of the North Sea meant access to amber, which was highly prized throughout the Corded Ware culture and also in the Yamna culture.

Evidently it was along the shores of the Baltic and the Kattegat seas that an influx from the southeast especially changed the population’s DNA. One of the geneticists’ studies shows that in modern Europe the highest percentage of Yamna ancestry is found in Norway, Lithuania and Estonia (with Iceland and Scotland close behind).⁸⁹ Even along the Baltic, however, the influx must have been gradual since here, according to a recent archaeological report on the region’s Corded Ware culture, “no mass migrations have been observed.”⁹⁰ What language or languages the immigrants were speaking on their arrival in northern Europe cannot of course be determined, but it is very likely that those who came from the southeast did speak an Indo-European language. The geneticists’ findings therefore add some support, however shaky, to the possibility that in the third millennium BC the considerable area in which Proto-Baltic was spoken expanded into northeastern Europe.

Another area into which an Indo-European language seems to have expanded in the third millennium BC was southern Caucasia. Archaeological evidence is clear that after 2500 bc nomadic pastoralism became prevalent here. Eventually most of the Kura-Araxes settlements were abandoned, and those that remained through the Trialeti culture were relatively small.⁹¹ If PIE did not evolve in the Kura-Araxes culture, and we have good reason to think that it did not, it may have been with the arrival of nomadism in the third quarter of the third millennium BC that an Indo-European language first came to south Caucasia. Where the pastoralists may have come from, or whether most of them were simply Kura- Araxes villagers who had become pastoralists, is unclear. Those pastoralists who did not have local roots may have come from north of the Caucasus, but it is also possible that they came from the southeast. Some specialists have argued, that is, that pastoralists from northwestern Iran began filtering into southern Caucasia in the third millennium bc.⁹² We know only that in the middle of the second millennium BC Indo-Iranian was spoken in or near to southern Caucasia.

Most of this is very uncertain and unclear, but frames my suspicion—shared by many who know more about the topic than I do—that by the end of the third millennium BC several Indo-European languages were spoken over a very wide area, including much of the grassland steppe and the forest steppe far to the east of the Volga, the Corded Ware culture westward from the upper Volga into northeastern Europe, and probably also southern Caucasia. Informing the speculative is the certain and the near-certain. PIE was descended from Proto- Indo-Hittite, the indigenous language of western and southern Anatolia. Early in the seventh millennium bc people on Anatolia’s western coast were fully engaged in the production of food, harvesting crops and depending on domestic animals for their meat. A growth of population followed the production of food and encouraged the colonizing of arable lands outside of western Anatolia. Among the first lands to be colonized were central Crete, Thessaly and the southern Balkans, from all of which the colonists remained in communication with the Proto- Indo-Hittite heartland. PIE must have evolved because Indo-Hittite colonists sailed—still on log boats or some other primitive craft—to a land farther away. The most likely locations for such a distant colonizing are lands along the northeastern shores of the Black Sea and the Sea of Azov. And the most likely location for the evolution of PIE is northern Caucasia and the steppe along the Don and the middle Volga, as settlers speaking their Indo-Hittite language came into close and continuing contact with the native Mesolithic population.

After the maturation of PIE, late in the fourth millennium BC, it soon spread (thanks in large part to wheeled vehicles) through and well beyond the grassland steppe. As it spread, PIE must have evolved into several regional dialects and then— surely by the end of the third millennium bc—into several distinct languages. We have no reason to think, however, that at that time Indo-European languages were yet being spoken in most of Europe (in Greece and Italy, we can be quite certain, they were not), or that several important Indo-European subgroups—Keltic and Germanic, as well as Greek and Italic—had yet begun to form.