Was Proto-Human an SOV language?

Oct 17, 2011 by

Most scholars agree that all of today’s languages descend from a common ancestral language, Proto-Human, which was spoken by behaviourally modern humans (BMHs) some time between 200,000 and 50,000 years ago. (Some theories equate BMHs with anatomically modern humans, which developed around 200,000 years ago; others believe that BMHs arose around the time of the “cultural explosion” 50,000 years ago. We might discuss this issue at more length in another posting.) An alternative possibility is that Proto-Human was spoken not by the first BMHs but later, during a period of population bottleneck, when other languages became extinct and only Proto-Human survived and left modern day descendants. One way or another, Proto-Human is in effect the most recent common ancestor (MRCA) of all modern human languages. If we adopt this Proto-Human hypothesis, the next question to arise is what Proto-Human was like as a language, what properties it had and how it compared to modern living languages.

Recent research by Merritt Ruhlen and Murray Gell-Mann, published in the Proceedings of the National Academy of Sciences and discussed in an article in LiveScience.com, attempts to address those very questions. As the title of the LiveScience article — “The Original Human Language Like Yoda Sounded” — suggests, Proto-Human had a very different word order from the one found in English. In particular, Gell-Mann & Ruhlen claim that Proto-Human had the order Subject-Object-Verb (SOV). While not familiar from languages like English, or Spanish, French, Italian, or Russian, SOV is nonetheless the most commonly found word order in today’s living languages: according to Gell-Mann & Ruhlen, 1008 out of 2011 languages studied (or just over 50%) exhibit the SOV order. According to Matthiew Dryer’s work in the World Atlas of Linguistic Structures Online (as well as Mallinson & Blake 1981), the predominance of the SOV order is less pronounced (41% according to Dryer and Mallinson & Blake), but it is nonetheless the most frequently found order among the six possible combinations. Examples of languages with the SOV word order include Japanese, Korean, Turkish, Georgian and Basque, among others. For example, ‘The woman has seen the man’ in Basque is rendered as Emakumeak gizona ikusi du, literally ‘the woman the man seen has’.

The other five possible orders of S (subject), V (verb) and O (object) are found less frequently: SVO is found in 35% (Dryer) to 38% (Gell-Mann & Ruhlen) of languages; VSO — in 7% (Dryer) to 8% (Gell-Mann & Ruhlen) of languages; VOS — in 2% of languages (according to both studies); while the rarer OVS and OSV are found in the mere 0.8% and 0.3% (Dryer)-0.6% (Gell-Mann&Ruhlen) of languages, respectively. Examples of VSO languages include Irish, Tagalog and Zapotec; VOS languages include Malagasy and Tsotsil; the best known OVS languages are Hixkaryana and Mangarrayi; and the best known OSV language is Nadëb.

While predominance of the SOV order among today’s living languages suggests that Proto-Human was SOV too (given the overall tendency for conservatism), the question of the word order in Proto-Human is not decided solely “by the show of hands”. Gell-Mann & Ruhlen’s argument is a bit more complicated: they argue that there exists a pattern as to how the word order may change in time as a result of what they call “natural drift”. For example, without any external influences SOV languages may change into SVO languages (with the change to OVS or OSV being possible but much rarer). SVO languages, in turn, may change into VSO and more rarely into VOS. Verb-initial orders may develop from each other or may develop back into SVO. Crucially for Gell-Mann & Ruhlen, in the absence of external influences from SOV languages, no word order can mutate into SOV. This is schematized in the diagram below (thicker lines indicate the more common patterns of change):

This schema is based on Gell-Mann and Ruhlen’s examination of known word order changes. Let’s look more closely at just one example — the Indo-European family. Different word orders are found in various branches: SOV is found in Anatolian, Tocharian, Indic, Iranian, Italic (think Latin!), and (early) Germanic; SVO in Greek, Armenian, Albanian, and Baltic; and VSO in Celtic and perhaps (early) Slavic. Since Anatolian, Tocharian, Indic and Iranian are the first branches to split off the rest of the Indo-European tree, Gell-Mann & Ruhlen decide that Proto-Indo-European was SOV as well. The change from the earlier SOV to the present-day SVO is illustrated by Latin-to-Romance and Old-to-Middle-English.

Gell-Mann & Ruhlen examine 11 other language groupings (either widely accepted languages families like Afro-Asiatic or Nilo-Saharan, or more controversial macro-families like Nostratic and Dene-Caucasian) and conclude that in those families too the change into SOV is found only as a result of external influences.

Based on this hypothesis that the change into SOV is not possible in the process of “natural drift”, they ask the following question:

“If new SOV languages arose only from contact with older SOV languages, then where did the prior SOV languages come from; and if they too are assumed to be due to contact with SOV languages, then how did the very first SOV language come about?”

Their answer:

“the very first SOV language was in fact the language from which all or most attested languages derive and that most modern languages with SOV word order merely preserve this initial state, except for cases where SOV has been borrowed from neighbors.”

While Gell-Mann & Ruhlen rely purely on arguments from historical developments and typological facts about the range of possibilities in the modern languages, their theory finds indirect support in the earlier work by Tom Givon, a linguist at the University of Oregon, who argued that SOV had to have been the first word order, based on how children learn language; according to him, the SOV word ordering seems to come most naturally to humans.

It should be noted though that very different views have been expressed on the subject of the “most natural word order”. For example, Richard Kayne of NYU has argued in his work on the antisymmetry hypothesis that SVO is the most basic, underlying word order. Conversely, Etienne Bonnot de Condillac, an 18th century French grammarian, considered OVS as

“the ‘most natural order’: first the noun indicating the object one was talking about, then the verb indicating the operation one intended to carry out on that object: for example, fruit want; the subject of the verb came at the end of the whole series: for example, fruit want Peter.” [cited in Lepschy 1998:196]

Overall, Gell-Mann & Ruhlen’s work on the order of Proto-Human is very interesting, but it is not without problems. In the next posting, I will consider some critical objections to their work.

Subscribe For Updates

We would love to have you back on Languages Of The World in the future. If you would like to receive updates of our newest posts, feel free to do so using any of your favorite methods below:

  • John Cowan

    By the same drift argument, clicks should have been present in Proto-Human, and yet they are extremely rare in the world's languages.

  • iamflower2

    I have heard that Proto-Oracle bone script was SOV language. But, Chinese is SVO language. Why is that?

    • Bla

      It’s often claimed by linguists that proto-Chinese and proto-Sino-Tibetan was SOV and later became SVO

  • Asya Pereltsvaig

    @John Cowan: I am not sure why you say that "clicks should have been present in Proto-Human" — is it because we don't know of any languages acquiring them?

  • Asya Pereltsvaig

    @iamflower2: not sure about Proto-Oracle… And yes Chinese are SVO languages. I am not sure I understand your question "why"? What do you expect it to be? More than half the world's languages are not SOV and most of those are SVO. Why a question specifically about Chinese?

  • iamflower2

    My hypothesis is that Mongolian(Japanese, Korean, Turkish, etc) was the origin of Chinese culture. Because, Mongolian is SOV Language.

  • Pakistani

    I don't believe in "Proto-Human."
    I actually am planning on writing an essay on my theories on the origin of human language and evolution. I believe a common human brain is why all the languages have similar features.

  • Asya Pereltsvaig

    @iamflower2: There is a great deal of cultural influences between the Chinese, Japanese, Korean etc.; however, there is no linguistic evidence for a common descent. Do you have any?

  • Asya Pereltsvaig

    @Pakistani: so your theory is that language arose multiple times in the process of evolution?

  • John Cowan

    Just so: we know of no languages that have acquired clicks other than by borrowing (the Bantu languages of South Africa), so by R&GM's logic they must have been present in Proto-Human — and yet that seems unlikely given how rare they are in the world's languages.

    It's true that Mandarin has mostly SVO word order, but many of the typically associated characteristics are SOV-ish (modifiers must precede their heads, for example), suggesting a possible transition toward SOV in progress (and a refutation of R&GM). In addition, the notion of subject is not well-defined in Mandarin. There's an excellent book chapter by Li & Thompson on Mandarin typology which discusses the issue.

    (The other Sinitic languages are firmly SVO.)

  • Asya Pereltsvaig

    @John Cowan: Thank you for your comment! I like your argument about clicks very much. It does show that the logic of G-M&R is somewhat suspect.

    As for Mandarin, G-M&R claim that Proto-Sino-Tibetan was SOV, which would mean that Mandarin might be developing away from SOV, not into SOV. Could it be that any of the phenomena you've had in mind is associated with transition from rather than to SOV?

    And while Sinitic languages tend to be SVO, aren't Tibeto-Burman languages mostly SOV?

  • Asya Pereltsvaig

    Not that I necessarily believe that G-M&R are correct (see my next posting)!

  • Randy Hudson

    I don't see the point about clicks calling the G-M&R into question. The scarcity of clicks vs. the commonness of SOV could merely mean that a language is much more likely to lose clicks than lose SOV order. By analogy with radioactive decay we could say that SOV has a much longer half-life.

  • Asya Pereltsvaig

    @Randy Hudson: Thank you for your comment! You are correct in saying that it is possible that Proto-Human had both SOV and clicks, even though many languages retain SOV, but not many retain clicks. However, for this theory to have traction we'd need to figure out WHY languages are more likely to retain SOV than clicks. There doesn't seem to be any particular usefulness to SOV, as far as I can tell.

  • Pakistani

    My theory is language arose in different parts of the world when humans learned to make use of vocal chords and other body parts to produce sound.

  • Asya Pereltsvaig

    @Pakistani: Any evidence to support your theory?

    And by the way, your theory implies that human language arose later than the Proto-Human theory implies. Any evidence for that?

  • John Cowan

    Most of the Sino-Tibetan languages are indeed SOV. But the Sinitic languages are the descendants of Middle Chinese, which was firmly SVO. Therefore, Mandarin's SOV features must be of secondary origin. However, Mandarin is not a full counterexample to R&GM, because it does still have VO order. For that matter, English has a few SOV features like adjectives before nouns, but no one would call it incipiently SOV because of that alone.

  • Asya Pereltsvaig

    @John Cowan: Thank you for the comment on Mandarin. Of course, English used to be SOV, so the adjective-noun order might be a conservative retention from those old days. Russian might be a problem too, as it is an SVO and adjective-noun language too, but does not descend from SOV (that we know of). Of course, this is precisely the kind of evidence that G-M&R use to "prove" that various reconstructed proto-languages *were* SOV! But let me point out that noun-adjective order is far more common for both SOV and SVO languages (http://wals.info/feature/combined/81A/87A), so from the order of the noun and the adjective we can conclude nothing about some "latent" SOV or SVO nature of a language! This is one of the problems with G-M&R's analysis (using putative correlations to reconstruct the order in a proto-language), which I was going to discuss but haven't gotten around to yet.

  • John Cowan

    I misstated the Chinese story here. The universal is that if languages have relative clauses before nouns, then they have SOV order. Chinese is the one and only known exception to this universal, hence the idea that although it is still SVO on the surface, it is shifting towards an underlying SOV ordering.

    Unfortunately, I was also wrong to say that the other Sinitic languages beside Mandarin lack relative clauses before nouns. I have skimmed a paper on Cantonese relative clauses that sums up as "the situation is not well understood"; I hope to read it more closely soon.

  • Asya Pereltsvaig

    @John Cowan: Thank you again for a great comment and for sharing this important info. Don't you just love it when an article says "the situation is not well understood"? Is it ever? 😉

  • Pingback: Are We “At the Origin of Language Structure”?—Another Blooper-Riddled Article from Science Daily - Languages Of The World()

  • Pingback: Reconstructing the Lifestyles of Three Pre-Historic Amazonian Tribes - Languages Of The World()