5 December 2009

The spread of agriculture and languages

Being a vegetarian, I have mixed feelings about the concept of agriculture. On the one hand, I feel fine with this way of life: it has helped me have a more balanced diet and a more natural approach to life in general. On the other hand, however, vegetarianism is ultimately a product of the agricultural revolution, a relatively recent phenomenon in the history of mankind. Our species was shaped in the context of hunter-gatherer societies where technological innovations took a very long time to be developed or perfected. The long-lasting balance between humans and their environment came to an end with the advent of agriculture and pastoralism: human populations became more sedentary and the surplus in production led to a complete change in the way humans related to one another. The rest of the story is well known: exploitation of resources, overcrowding, famine, war, slavery and many other elements (including some positive effects) that have shaped the world ever since (you can read this article by Jared Diamond for a complete view on this topic). The Neolithic revolution, starting in the Fertile Crescent at about 9,500 BC, is the real turning point for humanity, even for the fate of Homo Sapiens. It would be impossible to go back to pre-Neolithic times and advocate for a hunter-gatherer existence, like some people have proposed, e.g. John Zerzan and the followers of anarcho-primitivism. We are just too imbued with agriculture, we can only try to do things better for once.

It's obvious that the emergence of agriculture must have made a considerable impact on the languages of prehistoric people. First, as a source of new vocabulary and linguistic structures for the new social reality; second, as a new scenario for the spread of languages and the appearance of stratified variants. Unfortunately, it is impossible for us to know the languages that the first farmers spoke, as there is no direct evidence of them. However, the analysis of indirect sources (linguistic, archaeological, genetic, etc.) has allowed linguists to come up with some interesting results. Associating the spread of a given group of languages with the spread of agriculture is a strong argument, and it has been used in the various areas where agriculture was independently developed: the Near East, the Indus Valley, China, Mesoamerica, etc. As far as I know, there isn't a generally agreed pattern of how agriculture and language may have expanded in parallel. Authors such as Colin Renfrew, in the context of the processual approach, have made interesting proposals about the combined process of agriculture-language spread, but it seems there's still much to be done in this field of study.

What about the Indo-Europeans? The expansion of IE languages has traditionally been seen as occurring some millennia after the expansion of agriculture. In this blog I have variously criticized this traditional view, which is based on a series of assumption which are quite dubious. I have presented some alternative views, including Renfrew's Anatolian Hypothesis (see here), which links the spread of IE with the spread of agriculture from the Anatolian Peninsula. Needless to say, Renfrew's theory has been strongly criticised by mainstream Indo-Europeanists, who stick to their intra-linguistic approach. I find the Anatolian Hypothesis quite interesting, and I'm sure that the spread of agriculture and pastoralism must have made a crucial impact on the languages of the IE area, as anywhere else in the world, but there are some problems associated with this theory that are not easy to solve. It is true that Renfrew's hypothesis offers a plausible scenario for the spread of IE languages in Europe, but for the events in the other areas (Iran, India, Central Asia) a different, possibly more complex explanation is required. Was agriculture a local development in the Indus Valley or was it imported from another place? The former option looks more likely. And one more, and essential, question: what kind of language did the first Indus Valley farmers speak?

We tend to think that new languages, e.g. the ones brought by the first farmers or by any other migratory or expansionist group, is basically different from the language of the original population, and that the new situation triggers a process of language substitution whereby the old language simply disappears. But it doesn't have to be this way. Maybe the languages associated with the Neolithic expansion were not so different from the ones spoken by European Mesolithic populations. And something similar could be said about the Indus Valley. What we would have here is a common IE background and a double process of language expansion (or re-expansion) associated with agriculture. This is just a hypothesis, of course, but I think it makes at least some sense.

Note: the map (see above) has been taken from this page.


Joan-Carles Martí i Casanova said...

Great article. Just thought I would let you know I've read it. One of my favourite blogs.

JK said...


Have you seen this recent paper:
Separating the post-Glacial coancestry of European
and Asian Y chromosomes within haplogroup R1a by Peter A Underhill et al.?

This paper discusses this issue a bit, introduces a new M458 marker, and states that "it would exclude any signi?cant patrilineal gene ?ow from East Europe to Asia, at least since the mid-Holocene period."


Jesús Sanchis said...

Thanks, Joan-Carles.

JK, welcome to this blog and thanks for the information about this recent paper by Underhill et al. I'll try to read it soon.

In my post I have mentioned some possible ways of linking agriculture and language expansion, with Renfrew's theory as an interesting reference point. I don't think I have mentioned any possible gene flow from Europe into Asia. Anatolia is not even in Europe! In any case, I'm generally against any explanations based on massive migrations in prehistory, and India is not an exception.

Tackling the prehistory, genetic ancestry and linguistic diachrony of the Indo-Aryan region is a complex task. Please notice that in my post there are more questions than answers about this particular topic. As I mentioned in my post, the eastern half of the IE area poses a considerable problem for Renfrew's theory, and he was well waware of this. In fact, he has proposed a series of alternative explanations (vid. Chapter 8 of Renfrew, Colin ( 1987). "Archaeology and Language. The Puzzle of the Indo-Europeans").

One day I'll write a post about Indo-Aryan languages. I'm familiar with Michael Witzel's writings and with the debate around invasionist and non-invasionist theories, sometimes referred to as the 'Indo-Aryan controversy'. But I still have to read more about it.

JK said...


I have not read Renfrew's book yet, but saw the following criticism in Edwin Bryant's "The Quest for the Origins of Vedic Culture: The Indo-Aryan Migration Debate:" Maybe he has addressed it in his book.

When a new group invades an area, they don't change place names or river names. Anatolian place names did not make any sense in Indo-European. Also there is no word in PIE for wheat and barley - two major early agricultural produce.

Jesús Sanchis said...

Even if we accept the continuity of IE languages from a pre-Neolithic date, i.e. in the Paleolithic or Mesolithic, a view that I'm fond of, it would be absurd to overlook the phenomenon of agriculture expansion in connection with language change, at least in some areas.

As I said, I think Renfrew's Anatolian Hypothesis does not offer a good explanation for the whole issue of IE languages, but it might be useful in some respects. At least it is a step forward if compared to the traditional IE chronology, which sets the process of IE expansion at a much later date.

Generaly speaking, the task of trying to understand the languages of prehistory is really complicated. A lot of factors must be taken into account before we can test a hypothesis. And it is also necessary to challenge traditional assumptions, even if they look apparently solid. I haven't read Bryant's book but I might do one day (there are hundreds of books I'd like to read!). In the last part of your comment you provide some ideas which, apparently, derive from Bryant's view, or Bryant's arguments, but I'm not sure I can understand them well, they seem a bit confusing. You talk about invasions and Anatolian place names, but I get a bit lost. I don't really know what you mean. And I also didn't quite understand your first comment, or its relevance to the topic.

Octavià Alexandre said...

There's no doubt IE languages weren't native to India. I recommend you Michael Witzel's article: Substrate Languages in Old Indo-Aryan http://www.ejvs.laurasianacademy.com/ejvs0501/0501ART.PDF

IE continuity theory's postulate that IE languages are native EVERYWHERE they're historically attested is false. The original area where IE was originalkly spoken is of course more reduced.

Andrew Oh-Willeke said...

The crops used in the Indus River Valley Civilization were domesticated in the Fertile Crescent at a time preceding their appearance in the IRV. So, some influence of the Near East on the IRV civilization is necessarily present.

The problem with Renfrew's theory is that we have documented pre-IE languages in Anatolia (e.g. Hattic), Crete (Minoan), Italy (Etruscan), Iberia (Basque and probably Aquitannian), and Mesopotamia (Sumerian) within a few hundred years prior to the appearance of Hittite or later.

The continuity of population that Renfrew proposes between the early and late Neolithic populations also isn't supported by the ancient DNA evidence (e.g. the virtual disappearance of mtDNA haplotype N1a).

The shared roots in Indo-European back to proto-Indo-European are also a poor fit for Anatolian conditions.

One could argue for an LBK language family derived substrate to Central and Eastern European IE languages, and a Mediterranean Neolithic substrate to Celtic-Italic and perhaps even Germanic IE languages. But, Renfrew's timing for IE itself simply must be too old.

I wouldn't be as dismissive as Octavia of the possibility of IE languages originating in the Harappan culture. The lack of a legendary origin elsewhere, the lack of evidence of violent conflict, the Saravasti River system's collapse as a motivator for IE expansion, the appearance of Cemetary H cremation close in time to the appearance of cremation in IE cultures in the Balkans and Greece, the clear source of Indo-Iranian to the East all the way through the Mitanni, and the likely origin of R1a in South Asia all point to that hypothesis. Indeed, even if the IE language has origins in Central Asia (leading to the Tocharians and early Hittites), the IE religion, and in particularly the cremation that became characteristic of IE peoples until Christianity subdued the trend, could very well have it roots in the Indo-Aryan experience. Bronze and Iron artifacts are as old in NW India as they are in Anatolia.

But, a Central Asia origin for it is probably a stronger argument, particularly in light of the evidence of Tocharian origins, which pretty much force an earlier date for IE people and suggest a religion that was different from the one in place by the time that IE people took the world by storm.