26 June 2011

Languages on horseback

I have just finished reading David Anthony's book The Horse, the Wheel and Language (2007), a book that I mentioned briefly in a post some time ago.

First of all, I'd like to say that David Anthony is an eminent archaeologist with vast knowledge about the prehistory of the Eurasian steppes, and more concretely about the archaeology of human societies in connection with horse domestication, an area which is vital in our understanding of Eneolithic and Bronze-age societies. He has written extensively about this issue and has also developed an innovative technique to date the use of domesticated horses by analysing bit wear in their bones. I'm sure his ideas about the subject are valuable and must be taken into account in any serious research in that field. Now, what's the problem? The problem arises when Anthony tries to fit all these data into a comprehensive explanation of the genesis and transcontinental expansion of Indo-European (IE) languages. This is when his scientific writing becomes fantasy.

Anthony bases his archaeological voyage on a series of linguistic facts which he accepts as irrefutable. We have talked about these things profusely in this blog (the traditional concept of proto-language, the use of linguistic paleontology, etc.) and we have seen how these ideas can easily be challenged. Anthony, however, takes them for granted. In his view, there is a Proto-Indo-European (PIE) people to be found somewhere, with its own language and even institutions. (p. 89): "that language [PIE] is a guide to the thoughts, concerns and material culture of real people who lived in a definite region between about 4500 and 2500 BC". In this respect, he follows Gimbutas' and Mallory's ideas, which we have extensively talked about (and criticised) in the blog. The novelty is that Anthony uses horse domestication and later developments as the use of chariots, as the main factors in the expansion of Proto-Indo-Europeans and their languages.

In his book, Anthony analyses one by one all the prehistoric cultures of the Pontic-Caspian region. He presents them in the framework of his own preconceptions, at times establishing simple correlations between culture, people and language. Pastoralist societies become PIE societies endowed with a remarkable capacity of expanding and subduing other human groups. (p. 343): "Wealth, military power, and a more productive herding system probably brought prestige and power to the identities associated with Proto-IndoEuropean dialects after 3300 BCE. The guest-host institution extended the protections of oath-bound obligations to new social groups. An Indo-European-speaking patron could accept and integrate outsiders as clients without shaming them or assigning them permanently to submissive roles, as long as they conducted the sacrifices properly. Praise poetry at public feasts encouraged patrons to be generous, and validated the language of the songs as a vehicle for communicating with the gods who regulated everything. All these factors taken together suggest that the spread of Proto-Indo-European probably was more like a franchising operation than an invasion. Although the initial penetration of a new region (or "market" in the franchising metaphor) often involved an actual migration from the steppes and military confrontations, once it began to reproduce new patron-client agreements (franchises) its connection to the original steppe immigrants became genetically remote, whereas the myths, rituals, and institutions that maintained the system were reproduced down the generations."
Franchises, military power, migratory movements... It is obvious that all of Anthony's metaphors and hypotheses can be doubted or found completely wrong. He talks about (literally) hundreds of prehistoric societies, and maks all kinds of assumptions about their language, social customs or expansive moods. One theory leads to another, in a process that can only be described as accumulative conjecture, or plain fantasy. First it is horse domestication, then the use of chariots, with the addition of a myriad of complementary elements. Aided by these extraordinary tools, PIE people started their incredible story of success. First, with the detachment of Anatolyans, then, with the emergence of proto-Slavic, proto-Germanic and proto-Italic in central-eastern Europe as off-shoots of the Pontic steppe developments, finally the expansion of proto-Indo-Aryan in the BMAC area. Let's see an example (p. 367): "The many thousands of Yamnaya kurgans in eastern Hungary suggest a more continuous occupation of the landscape by a larger population of immigrants, one that could have acquired power and prestige partly just through its numerical weight. This regional group could have spawned both pre-Italic and pre-Celtic. Bell Beaker sites of the Csepel type around Budapest, west of the Yamnaya settlement region, are dated about 2800-2600 BCE. They could have been a bridge between Yamnaya on their east and Austria/Southern Germany to their west, through which Yamnaya dialects spread from Hungary into Austria and Bavaria, where they later developed into Proto-Celtic".

The author finishes his analysis at this point, sometime at the Bronze age, with all the IE proto-languages ready for action. Their incredible run of good luck lasts centuries, millennia. In the vast poker game of prehistory Indo-Europeans seem to have the winning hand at all times!

Obviously, Anthony is not the only researcher who has felt the temptation to offer a comprehensive explanation of IE origins and expansion. Like Mallory and Gimbutas, he does so from an archaeological perspecitve, and as I said before many of the things they said might be useful and coherent, at least at a minor, less ambitious level. A similar type of analysis, enriched with population genetics data, is to be found in other authors, such as Mario Alinei. Reading his volumes about the linguistic prehistory of Europe I often felt a bit lost in the never-ending tales of prehistoric societies that follow one another. His theories are possibly quite imperfect and need a lot of refining, and in some cases must probably be rejected, but there is an important difference between Alinei and the more traditional authors such as Anthony or Mallory: his approach offers a more logical framework to understand language events through time.

Note on illustrations: on the left,
The Large Blue Horses, a painting by Franz Marc.


Belenos said...

While I disagree completely about Alinei (his assumption of language stability goes against all existing evidence, and genetic data taken from the mesolithic samples we have seems to argue very strongly against him) I have to say you have a point regarding this book.

That PIE existed is, for me, pretty much incontrovertible, but an attempt to link it it to any one material culture in one place, especially to the first equestrians, is simply speculation. IE languages in Europe can begin to be tracked about 800BC when we begin to have sources for them, in Greece, India, Persia the levant and Anatolia about a thousand years earlier.

I'm amazed that intelligent people write books based on so many unproven assumptions.

Octavià Alexandre said...

IMHO, Kurgan people would be the speakers of Proto-Indo-Iranian, but no means PIE. There's also archaeological evidence they invented the spoked wheel and the chariot, which gave them military superiority over other people. This technology was also adopted by Celts and lead to the spread of IE-speaking aristocracies in the Bronze Age, but this is very far from being the biggest military conquest of all times proposed by people like Anthony, Mallory or Gimbutas.

But also Alinei's approach is flawed in assuming the historical IE branches were already differentiated in the Mesolithic and at the same time negating language replacement. Sure, PIE was already disintegrating at that time, but the historical picture can only be the result of several expansions and extinctions over time.

Both approaches also underestimate the role of non-IE languages, specially important in the Neolithic, when Anatolian farmers colonized many parts of European, specially the Mediterranean ara.

Jesús Sanchis said...

Thank you for your comments, Belenos and Octavià.

Belenos, the earliest date for IE languages in Europe, even from a traditional perspective, is some centuries older than the date that you mention. Think of Linear B Mycenean Greek, for example.

In general, I think the presence of IE languages in Europe is a lot. I'm not talking about some kind of IE Bing Bang of the kind described in some books; in my opinion, the presence of old IE elements in Europe predates some later developments which involve other forms of IE, actually facilitating the expansion or consolidation of the new forms.

Belenos said...

Sorry, I was excluding Greek from the rest of Europe, but the sentence wasn't very clear. Linear B obviously goes back to about 1600BC which strongly suggests a presence at least 500 years before that. I'd personally be very cautious about any one who said it was certain that IE was in Europe prior to 2500BC, but I don't discount the possibility.

Jesús Sanchis said...

Alinei's theories, like Renfrew's, are important challenges to traditional views about IE, and challenges are necessary in science.

Now, there's something funny about Greece. In his book, David Anthony admits that his model does not offer any possible explanation for the emergence and expansion of the Greek linguistic group. On the contrary, Renfrew's Anatolian hypothesis solves the Greek puzzle quite simply, but finds considerable trouble on the Asian side of the IE world. This shows that it is actually very difficult to find a global explanation for the whole phenomenon of IE languages.

Octavià Alexandre said...

Renfrew's Anatolian hypothesis solves the Greek puzzle quite simply,
I don't think so. Firstly, Greek is more closely related to Indo-Iranian than to other European IE languages, and secondly, there's ample attestation of pre-Greek languages in Greece, including Minoan/Linear A as well as loanwords in Greek itself. Both facts are incompatible with Renfrew's theory.

Jesús Sanchis said...

Sorry, maybe I didn't express myself correctly. When I said that sentence, I was thinking about the internal consistency of Renfrew's theory, not about its validity.

Octavià Alexandre said...

No problem.

IMHO, all these theories lack a true understanding of language contact and replacement process, for which I've coined the term multi-layer (although this term was devised for Vladimir Georgiev for one particular Anatolian language, I've generalized it). That is, multi-layer languages are the result of multiple contact (adstrates) and replacement (substrates) processes.

Somewhat simplifying, while mainstream IE theories collapse these levels into the most recent one (i.e. the superstrate), the Paleolithic Continuity Theory does the reverse and collapses them into the more ancient one.

Anonymous said...

"there's ample attestation of pre-Greek languages in Greece, including Minoan/Linear A as well as loanwords in Greek itself."

I don't subscribe to any specific homeland theory, considering they all have their problems, but this is certainly not incompatible at all with Renfrew's theory. The very term "pre-Greek" is loaded considering we don't know with certainty whether those languages that gave Greek much of its vocabulary arrived in the greek mainland earlier or later than Greek did.