India

Essentially the same issues as for the Americas feature in the uses that scholars have made of lexical evidence for early agricultural history in India. India, like Mesoamerica, had multiple centres of early food production, and each major centre can be associated with a different language family.

southern India, the Ganges basin, and the region around the middle and lower Indus River. Similarly to work in the Americas, linguistic attention thus far has focused on the evidence of old reconstructed names for particular crops.

The southern Indian core areas extended from Gujarat southwards through central parts of southern India. Early cultivated crops in the northern sector of this zone, in Gujarat, included little millet (Panicum sumatrense) and the urd bean (Vigna mungo). Further south the early farmers also concen­trated on grains and pulses: their early crops would have included browntop millet (Bracharia ramose), bristly foxtail millet (Setaria verticillata), mung bean (Vigna radiata), and horsegram (Macrotyloma uniflorum).^[117] The languages associated regionally with the southern Indian Neolithic belong to the Central and Southern branches of the Dravidian family.

The names for several of these crops trace back to the common ancestor language of Central and Southern Dravidian, although the insufficient evi­dence currently available for the North Dravidian branch does not allow the reconstruction as yet of any of these crop terms back still earlier in time to the proto-Dravidian language. As was true for the Americas, the high retention of old crop terms probably suggests that the period of the common ancestor of Central and Southern Dravidian may have been the era in which the crops became fully domesticated and not just cultivated.^[118] In the archaeological record, the processes of domestication were well underway by 2000 BCE, and the beginnings of cultivation surely go back much earlier.

A second major region of early cultivation lay in the middle and lower Ganges basin. Before the spread of the Indo-Aryan languages after 1500 bce, the farming peoples of these areas spoke languages of the Munda branch of the Austroasiatic family.^[119] The borrowing of early Munda crop terms in the later Indo-Aryan languages of this region confirms the former much wider presence of Munda speakers in this region.^[120] A number of early crops originated there, including cucumber, luffa, and pigeon pea. Most important, it now appears that an independent domestication of Indian rice, separate from the domestication of rice in the Yangtze valley, provided the early staple of this agricultural tradition.^[121]

Northwestern parts of the Indian subcontinent constitute the third major region of early cultivation and the earliest region in which domestication took place. The major crops arrived already domesticated, via diffusion across the Iranian plateau from the Middle Eastern origin areas of agriculture, but linguistic studies suggest that the populations that took up the new crops may have been indigenous to the Indus regions rather than intrusive.^[122] To the incoming domesticated food resources, these peoples added two new contributions of their own: the raising of indigenous zebu cattle, which they may already have practised before the arrival of the new crops, and the cultivation of indigenous cotton.

Again, in the Indian subcontinent the reconstruction of crop names takes our historical knowledge back to the periods of early domestication, but not to the beginnings of cultivation. In the Ganges basin and in southern peninsular India, the pre-domestication eras of cultivation probably go back another 2,000 or more years beyond where the linguistic evidence yet reaches, and the study of the historical implications of old verbs and nouns specifically connoting cultivation, which might open up new knowledge about those times, remains a task still to be pursued.

Implications of linguistic reconstruction for agriculture in Europe

Finally, the linguistic evidence also affords a quite sharp lesson for the arguments about the agricultural revolution in Europe. First and foremost, the linguistic evidence flatly rules out the possibility that Indo-European- speaking peoples might have introduced agriculture to Europe. The evidence from the Indo-European language family presents two fully fatal objections to this now quarter-century-old proposal.^[123]

The first fatal objection has to do with the linguistic reconstruction of technological knowledge and practice. The proto-Indo-European language and society passed through two stages. The early proto-Indo-European stage, or Indo-Hittite stage, came to an end when the ancestral Anatolian-speakers moved away from the rest of the proto-Indo-European communities. The late proto­Indo-European period came to an end several centuries later, when the remain­der of the Indo-European communities began to spread out and diverge from each other. The evidence is overwhelming that the late proto-Indo-European communities possessed and used the wheel: their lexicon contained a suite of at least five and possibly six terms relating to wheels.

The possession of wheel technology as a key element of culture shows that the late proto-Indo-European period flat-out cannot date to earlier than the fourth millennium bce - well after agriculture had already spread far to the west across Europe. Since the Indo-Hittite period preceded the late Indo-European period by no more than several centuries, the lack of words for the wheel at the Indo-Hittite stage, in contrast to their abundance in late proto-Indo-European, indicates a dating of the Indo-Hittite stage to the close of the fifth millennium, towards the end of the pre-wheel era, and a dating of the late proto-Indo-European period to the early or middle fourth millennium.

The second fatal objection applies the criterion of cultural salience to the proto-Indo-European lexicon of food production. Not one single root word of unambiguous reference to even one of the ancient crops of Middle Eastern agriculture can be reconstructed with certainty to the proto-Indo-European language. Instead, the proto-Indo-European communities (PIE) possessed an abundant domestic animal vocabulary, including even breeding terms, the most diagnostic markers of all for livestock-raising. Along with a collective term for domestic animals in general (PIE *peku-), these ancient root words include terms for sterile domestic animal (PIE *ster-); cow (PIE *guou-); bull (PIE *tauro-); he-goat (PIE *ghaido-); she-goat (PIE *dik-/*dig-); kid (PIE *kago-); sheep (PIE *owi-); (young?) male sheep or goat (PIE *bugo-: seen in English buck); pig (PIE *su-: Latin sus; English sow, swine; and PIE *porko-: Latin porcus; English farrow ‘young pig'); and horse (PIE *ekuo-).

The implications are unequivocal. Cultivated crops lacked cultural salience for the proto-Indo-Europeans. Instead, a variety of domestic animals held first place in their economies and daily lives. The early PIE communities were pastoralists with little or no cultivation, and the late PIE communities of a few centuries later were not only pastoralists, but pastoralists who made and used wheeled vehicles and raised horses along with cows, sheep, goats, and pigs. The proto-Indo-Hittite stage may date to around 4000 bce or just before, but the wheel-using late proto-Indo-European era cannot be dated earlier than sometime in the fourth millennium bce. In all the world in the fourth millennium bce, only on the Pontic steppes did there exist a wholly pastoral economy with highly valued - as their burial customs show - wheeled vehicles. The linguistic testimony puts it beyond reasonable doubt that the far-flung expansions of Indo-European languages began no earlier than the fourth millennium and that they emanated outwards from the lands along the north of the Black Sea.^[125]

So who were the peoples who brought food production to western Europe in the seventh and sixth millennia bce? The Basque language is the most likely candidate for a surviving descendant of the languages spoken by those early farmers. The alternative hypothesis - that Basque is a survival of a late Palaeolithic hunter-gatherer language - is highly improbable, and here is why: nearly always when farmers spread into regions occupied previously by people who are solely gatherers and hunters, the languages of those foragers in the end die out. The languages die out because the incoming farmers form larger local population groups and because the intrusive farm­ing populations grow more rapidly in numbers as they expand into more and more of the hunter-gatherers' lands. When a language survives the spread of farming peoples, as Basque has, it is normally because the speakers of that surviving language were already themselves farmers. They belonged to communities that were already large enough and numerous enough to compete culturally and demographically with the incoming groups.