8 May 2007

The Kurgan Hypothesis is... hypothetical

I hate to state the obvious but when something is called a "hypothesis", it naturally means that it's hypothetical, not a fact written in stone. For some reason, it confuses quite a few on the net.

I say this because after contenting myself at dnghu.org by pointing out the errors or exaggerations on their site concerning Proto-Indo-European, one of its purported members volleyed back a claim that the Kurgan hypothesis is consensus among IEists. That's news to me, but I've been in this game long enough to have a good idea of what is consensus and what is fantasy. So many times, I've encountered people who after reading one book, deify the author of said book into an infallible Saviour that no one should dare challenge, not even other knowledgeable authors. Was there ever a time when the tenet "Question what you read!" was taught in schools?

How Indo-European languages expanded over time according to the Kurgan Hypothesis

What is the Kurgan Hypothesis? This is from the Encyclopedia Britannica (2003) :

Kurgan culture, seminomadic pastoralist culture that spread from the Russian steppes to Danubian Europe about 3500 BC. By about 2300 BC the Kurgans arrived in the Aegean and Adriatic Regions. The Kurgans buried their dead in deep shafts within artificial burial mounds or barrows. The word kurgan means "barrow" or "artificial mound" in Turkic and Russian.
Then in Encyclopedia Britannica, volume 22, page 987, it clearly states:

Some scholars believe that the Indo-Europeans were the bearers of the Kurgan (Barrow) culture of the Black Sea and the Caucasus and west of the Urals. The Kurgan culture, however, was only one of a number of related steppe cultures extending across the entire Black Sea-Caspian region, an area that was transformed about 4000 BC by the advent of the horsedrawn wheeled vehicles and related innovations. It is probably best therefore, to follow J.T. Mallory (In Search of Indo-Europeans [1989]) in locating the speakers of Proto-Indo-Euoprean among the populations of this region, but not to attempt a more precise indentification until further evidence is available. [I add boldface here to emphasize key points.]

We should pay close attention to Encyclopedia Britannica here because the encyclopedia has a continued dialogue with university academics and is carefully edited by professionals. If it doesn't have a full grasp of what's going in the academic world, what encyclopedia does? Wikipedia? I don't think so.

Even if we could claim that it were consensus in academe however, the theory is full of too many gaping holes to be taken as seriously as many linguistic hobbyists choose to take it. Unfortunately, the holes are just obscure enough that they can be easily glossed over by the unknowing masses. Take for example, the simple revelation that culture and language are not the same thing and that they don't even spread across the world in the same way. Don't believe me? Well, we can easily imagine for example a case where a single person adopts a new culture but retains one's old language. It happens. It happens a lot. And so if it happens, it undermines the narrow-minded assumption that the Indo-Europeans must be *strictly* identified with the Kurgan culture. Common sense will tell us that archaeology, being a study of material culture, can tell us absolutely nothing about linguistics, the study of an abstract and thus non-material means of communication. If I observe in the archaeological record that a people in Eastern Europe have adopted a new style of pottery at a certain period, I honestly don't know whether they merely adopted a new style of pottery or also adopted a new language. Archaeology here is utterly useless to the sensible linguist.

It also doesn't help the credibility of Marija Gimbutas, the one who pushed forth this decades-old theory, when she shows her strange bias towards making Indo-Europeans a completely "male-dominated", "woman-reviling", "war-loving" people to force an imaginary, black-and-white opposition between Indo-Europeans and "Old Europe" which we are to believe by her accounts is the exact opposite. Example:

Moreover, in contrast to the mythologies of the cattle-herding Indo-European tribes that, wave upon wave, from the fifth millenium BC overran the territories of Old Europe and whose male-dominated pantheons reflected the social ideals, laws, and political aims of the ethnic units to which they appertained, the iconography of the Great Goddess arose in reflection and veneration of the laws of Nature. (words of Joseph Campbell in Foreword included in The Language of the Goddess by Marija Gimbutas, 1989)

Yet, fundamentally with these very Wiccan-stained assertions, she proves to us that she doesn't understand one important fact about the Indo-European-speaking population that undermines her whole claim: The Indo-Europeans were never a unified people. They couldn't have been because they were supposed to be pastoralist nomads and such people, as any qualified ethnologist understands, would never have understood the abstract notions of a single government or of a "unified people" as we do today. At most, we can be sure that an average Indo-European could very well have been aware of which peoples around one's area spoke languages similar to one's own and which did not but this would be the reasonable limit of an Indo-European speaker's sense of cultural unity with peoples it may have only occasionally interacted with many kilometers away.

So if we now grasp that a "single, unified Indo-European people" is an anachronistic fantasy, then we must also grasp that Indo-European speakers had many dialects and many cultures spread over a large area from the very beginning, similar to the situation of the Inuit today. However, if we accept that, we must inevitably reject the Kurgan Hypothesis altogether, for it is then nothing more than an overly simplistic vision tied to a black-and-white belief that prehistorical archaeological cultures have a one-to-one relationship with protolanguages and that a single language must necessarily have a single culture, which is clearly a false premise if we choose to contemplate on the inherent complexity of humanity for a moment.

Where does that leave the Indo-European homeland debate? In Eastern Europe of course. Ultimately, it's linguistics that must decide the core homeland of the linguistic theory and there is ample linguistics to show that Eastern Europe is the best choice. On a whole, Indo-European must be in a position to be accessible to the northern shores of the Caspian Sea (nb. the words borrowed between Indo-Iranian and Finno-Ugric languages c. 2500 BCE) and also to be capable of receiving Near-Eastern loans at an early date (nb. the issue of "six" and "seven" and other cultural loanwords of clear Semitic origin). Ironically, the Encyclopedia Britannica even concedes, when discussing the Kurgan culture, that by 3500 BC there was a secondary homeland that "was established in the Danube river basin" which is "what many linguists consider to be the Proto-Indo-European homeland" (Enc.Brit., vol 18, p. 763).

So while many online like to be one-minute experts and call the Kurgan Hypothesis "widely accepted", it is in fact the most received only by those who know the least about Proto-Indo-European or what absurd contradictions the Kurgan Hypothesis implies.


  1. Glen, words borrowed between Indo-Iranian & Finno-Ugric c.2500 BC don't say much on PIE, only on I-I. The Semitic borrowings (eg, "6", "7") are more telling IMO & rather point to an (earlier) Anatolic homeland, in accordance to Renfrew's ideas (eg, Gray & Atkinson 2003 Nature 426:435), no?

  2. Ah, Nature magazine. The current issue online has the headline "How to survive in a black hole". How scholarly. I'm sure we'll need to know this when we risk falling into one during a physics exam :P Perhaps they should write a corollary entitled "How to survive in a k-hole" which would better serve these failing students.

    Somehow I'm not surprised that the same magazine that actually suggested that Wikipedia is "just as good as" Encyclopedia Britannica to the shock of academics, also published a minority, non-mainstream view like the "Anatolian Hypothesis" to the shock of academics.

    The keyword here is shock value. Shock value helps sell magazines. True scholars don't like shock value. Please read a variety of sources and ones with a higher standard of scholarship than just Nature magazine.

    You act as though there are no logical criticisms towards your Anatolic homeland and that my views should somehow be at its mercy. Please read a number of the issues you need to address in the links below to be better informed:

    Alan Little's weblog
    Language Log [upenn.edu]

  3. Hi Glen!
    I am not even an amateur linguist, but just curious:
    What is the status of the IE / PIE (I don't even know the difference!) theory?
    How sane is it to assume a single geographical origin for a proposed PIE-talking people?
    As you yourself said, this people could not have a unified culture. Why then assume that they had a unified language?
    Why should I belive that there ever has been such a thing as PIE???

    Regards Gunnar Øyro
    email: gunnar[at]operatic.no

  4. Hey Gunnar!

    What is the status of the IE / PIE (I don't even know the difference!) theory?"

    First of all, "IE" means "Indo-European" and "PIE" means "Proto-Indo-European". Those two terms are often used interchangeably, although "PIE" is specifically the language and nothing else, whereas "IE" might refer also to the people(s) speaking the language(s) or their culture(s).

    If by "status" you mean university credibility, the status is quite high. Many universities have courses devoted to PIE and other protolanguages.

    How sane is it to assume a single geographical origin for a proposed PIE-talking people?

    Not very, hehe ;) Actually, to be fair, some who assume this may simply be inexperienced with languages and honestly don't yet see how searching for protolanguage "homelands" is not as cut and dry as it seems. As I say, we talk about "general environs" of a language but not about clear borders, so it depends on what is meant by "IE homeland".

    As you yourself said, this people could not have a unified culture. Why then assume that they had a unified language?

    Exactly my point! Still confused ;) I was thinking a few days ago about explaining this better through "wave theory". My point is that while there never was a single PIE language, reconstruction is still important because it gives us a "rounded-out" version of what took place. Imagine if one said "What's the point to finding the value of 'pi' at all? Afterall it's infinite and unattainable anyway." The point is that approximation is better than nothing. So Proto-IE is still much better than nothing at all.

    Why should I believe that there ever has been such a thing as PIE???

    Because the evidence from languages throughout Europe and India show clear and regular similarities to each other that can't have arisen by chance. The reason why I may seem self-contradicting to you is because you're still thinking about language evolution as a large tree, with languages "splitting" and forming "branches". That's how we normally see the IE family represented in diagrams. However, in the real world, languages don't split or branch. The "wave" model works better. Languages and dialects are more like waves in a sea. The waves represent "unique features" of a language (anything such as grammatical quirks, pronouncing "p" as "pf", a new word for "4", etc). There is no clear border between one wave and the next and a wave can spread to another dialect area or even a completely different language (areal diffusion). This is a more accurate picture of what PIE was. It was a puddle of closely-related dialects in a larger sea. It was from that soup that modern IE languages formed.

  5. A few nits with your article attempting to debunk the Kurgan hypothesis.
    -- Encyclopedias are notoriously bad sources of scientific information and do not necessarily represent the consensus or even the cutting edge in a particular field for each entry. Ironic that you mentioned Gimbutus' Wicca influence further down. (Although I thought she influenced the Wiccans and not the other way around.) It's ironic because one of Britannica's most infamous failings, at least among historians, was in for many years perpetuating as the "censensus" the view that the victims of medieval European witch hunts were followers of remnant pagan relgions -- a la the claims of Wiccans -- when in fact the consensus among historians on the subject was that this idea was entirely without merit. Having said that, I think EB's article you quoted summed up this issue quite accurately. But it didn't make quite the same case that you did. Mallory's conclusions, which they held up as the best word on the subject, are not a rejection of the Kurgan idea as you have implied. Instead, he overs a softer form of the hypothesis and deemphasizes the primary emphasis on solely the Kurgan culture. For him, it is just a component of the PIE picture.

    Your statement that culture and archeology are irrelevant go much too far. Of course, culture can be transmitted without language and language can be transmitted without culture. But they're also transmitted together. More to the point, if we were postulating PIE out of thin air without any linguistic evidence, then your point would be an excellent cautionary. However, that is not the case. The linguistic evidence for there having been a PIE-speaking group and therefore a probably PIE-homeland, seems insurmountable. That's why we're even talking about this. Further, it's equally clear that this LANGUAGE did spread widely. So evidence for cultural and/or ethnographic spread along the same lines would be useful in discerning the linguistic history of this known spread. It is fair to say that we can't latch on to every sharing of culture and say that's the clincher. But neither would it be wise to totally devalue the archaeological evidence. All things being equal, you would think a linguistic spread of such great consequence would have left other traces. Nothing wrong with looking for them.

    Finally, doesn't your entire anti-Kurgan piece beg the point? The division in this field right now is between those advocating a steppe-PIE homeland (the Kurgan view, but also that of Mallory and others) and those advocating an Anatolian homeland (Renfrew). Whatever the shortcomings of an absolutist position on the Kurgan culture, which is the version of it you've chosen to critique, the steppe homeland still holds up much better than the Anatolian homeland in the face of all the evidence. Renfrew's side seems to be very astute at getting the attention of the popular media, however. I hope that doesn't work to their advantage in establishing their unwarranted view as the new consensus.

  6. Thanks for the comment, fourtoes. Congratulations, you've entirely *misread* my article <:(

    I purposely emphasized the word "strictly" in one of my sentences: "And so if it happens, it undermines the narrow-minded assumption that the Indo-Europeans must be *strictly* identified with the Kurgan culture."

    Quite obviously then, to say that PIE speakers could have been behind multiple archaeological cultures at once, and consequently that archaeological cultures could represent multiple languages at once *does* indeed undermine the "Kurgan Hypothesis" as it was presented by Gimbutas herself. Read a published critique of the Kurgan Hypothesis in Matthew Spriggs, Archaeology and Language II (1998) on p.267. How exactly are you defining the "Kurgan hypothesis" in your mind?

    fourtoes: "Encyclopedias are notoriously bad sources of scientific information [...]"

    To the contrary, I would argue that well-crafted encyclopedias like the EncBrit are notoriously misread by poor readers who don't properly use them for what they really are: valuable sources of generalized information that aid in locating more primary sources of information.

    I view my blog as a similar service which notifies its readers of available online information that may be pursued further in one's free time.

    Your statement that culture and archeology are irrelevant go much too far.

    No, not irrelevant! I explained that it was "useless to linguistics" and if you had read carefully was meant for the topic of naively searching for an Urheimat. Don't take statements out of context.

    Do you not understand the implications of the fact that PIE is a linguistic construct, a theory? Are you not aware of the illusion of reification that hampers logical reasoning?

    I presume that the very second you read of my satisfaction of the exceptional work put into the Encyclopedia Britannica (in marked contrast to most other encyclopedias), you abandoned all further reading in order to set out on a crusade to dish me your cold comeuppance in my commentbox for my heinous expression of appreciation for a (*gasp*) tertiary source of information as an educational tool.

  7. Oops missed one more thing that I need to respond to:

    "Finally, doesn't your entire anti-Kurgan piece beg the point? The division in this field right now is between those advocating a steppe-PIE homeland (the Kurgan view, but also that of Mallory and others) and those advocating an Anatolian homeland (Renfrew)."

    I reject the Anatolian Hypothesis even more so than the Kurgan Hypothesis. What is misunderstood here is what my actual position on the territory of the Indo-Europeans really is, which gives me incentive to write another blog precisely about what a 21st-century view of the "Indo-European Homeland" and protolanguages in general should look like. For that, perhaps I should make a colourful map... Stay tuned.

  8. Ok, after seeing your responses to my comments, I'm probably less in disagreement with you than I thought. (However, I would blame lack of clarity in you original argument. But let's not berate that point.)

    While I did not originally decide to write a response because of the encyclopedia issue as you have implied, I do have a concern with your high regard for works like EB. Encyclopedias tend to oversimplify and have been guilty of ignoring ongoing academic debate in order to present a concise clear summarization to their generally popular audience. (And as that is their primary goal, that's understandable. They just shouldn't be understood as anything other than a summarization for popular consumption.) But I do agree that EB is the best of the lot. My example of what EB did in its entry on witchcraft is notorious among medieval historians and is cited as an important factor in skewing the historiography on that subject. Admittedly, however, I'm not aware of any other such gross example from them. My cautionary is that EB or any similar work is a poor choice to hold up as evidence of the state of academia on a subject. My guess is that you would hardly have mentioned them if they had not happened to incline toward the position with which you agree. (But then, serious scientific thinkers should always be wary of the very idea of consensus anyway.) As it happens, part of my misunderstanding was with regard to whether you were attacking the entire idea of a steppe-PIE homeland or merely the strictly Kurgan idea. I have to agree with you that the strict Kurgan idea is no longer viable in the face of additional evidence. Am I correct, then, in understanding that you would tend to favor a position similar to Mallory's on the likely PIE homeland?

    And as I said, my other caution is that you seemed to me to be downplaying the value of archeology in what is a linguistic construct. As clarified, your position is not as extreme as I had thought. But some food for thought: I tend to come at this subject from the entirely opposite perspective -- Surely this linguistic construct has great value in helping us interpret the archeology, ethnography and genetic evidence, and thereby arrive at something that is neither linguistics nor archeology nor ethnography: a vague but useful narrative of the history of these "prehistoric" peoples.
    Would you agree with that statement?

  9. I don't think Mallory or anyone else really represents my personal perspective on PIE. To me, the idea that a) a singular "Proto-Indo-European" language *never* existed and b) the idea that Proto-Indo-European is well proven by fact are not contradictory to each other.

    I regard PIE as an "averaging out" of actual dialects. PIE will be the seed of protolanguage dialectology in the centuries to come.

    Languages should be thought of as a "collection of waves" like those on the surface of a lake. Each wave represents an isogloss or quirk in language. These isoglosses are not bound by language boundaries and can move into neighbouring languages (i.e. areal influence). These waves feel nothing for differences in genetics or culture. They are inanimate and travel wherever human whim takes them.

    So when I think about this nonsense about "IE Homeland", what *I* see is a package of isogloss waves which we may call "PIE" eminating outwards from a Pontic core area from about 4500-4000 BCE into previously non-IE and para-IE territories. Some waves travel farther than others; some waves are from outside PIE or only partial (e.g. the Satem wave). Para-IE languages would exist in outlying regions where only *some* of the isoglosses attributed to PIE would have reached.

    The spicy masala of innumerable non-IE, para-IE and IE dialect waves spreading in various directions on this turbulent linguistic ocean led inevitably to the creation of regional dialects in PIE as it continued to evolve while spreading into the steppelands, Anatolia and throughout Europe.

    My idea isn't represented by the Kurgan Hypothesis, Out-Of-India Hypothesis, Anatolian Hypothesis, Oui-Ja-Board Hypothesis, Farmer-Joe Hypothesis or any other simplistic us-versus-them scenario that narrow-minded theorists dream up. It's naturally way more complicated than all of these hollywood theories combined.

    Again, I stress to everyone to reject the illusion of reification. Language is not a singular entity with definable boundaries and ergo PIE is not either. Ergo it cannot have a homeland, merely a vague "core area" to point to. The reconstruction of PIE is being misunderstood by those who don't grasp the nebulous meaning behind the word "language" itself.

  10. I should also be more clear by stating that I think the PIE core was in the *North-West* Pontic area. I also reject the notion that PIE or even Pre-IE came from Anatolia because I'm convinced of the potential of its linguistic relationship with Uralic (to the north) and Altaic (far to the east).