Erard on Australian Languages.

My favorite reporter on linguistic issues, Michael Erard, has a fine Science piece about recent studies of Australian languages and the controversies they help address; after surveying some of the problems (the members of the hypothetical Pama-Nyungan family have lots of similarities but few cognates), he writes:

Now, a new generation of researchers is attacking the problem, and a small but growing group is taking its cue from evolutionary biology, which relies on genetic clues to decipher relationships between organisms. They are using computers to sort giant databases of cognates and generate millions of possible family trees based on assumptions about, say, how quickly languages split. The method, called computational Bayesian phylogenetics, forces researchers to explicitly quantify the uncertainty in the models, says linguist Claire Bowern of Yale University, a pioneer of the approach and co-author of the new study. “That’s useful in Pama-Nyungan,” she explains, “because you don’t have good data, and you have to rely on single authors who may not be that familiar with the languages.” Based on a set of parameters, researchers can winnow millions of trees into groups of the most plausible ones.

The first such computational efforts, done by biologists borrowing linguistic data, drew harsh responses from many linguists. “Most look exclusively at words, seen as something like the equivalent of the gene as a unit of analysis in genetics,” says Lyle Campbell, a historical linguist at the University of Hawaii, Manoa. But linguists traditionally determined historical relationships through sounds and grammar, which are more stable parts of language.

Bowern counters that the “instability” of words can actually be a boon, serving as a tracer for how languages change over time. In 2012, she and Quentin Atkinson, a biologist at the University of Auckland in New Zealand, constructed a family tree for the elusive Pama-Nyungan, using a massive database of 600,000 words to compensate for the low number of cognates. They analyzed 36,000 words from 195 Pama-Nyungan languages and compared the loss and gain of cognate words in 189 meanings through time.

This initial work found that Pama-Nyungan has a deep family tree with four major divisions tied to the southeastern, northern, central, and western regions of the continent. For the study published in Nature, Bowern drew from an expanded database of 800,000 words, which contains 80% of all Australian language data ever published, and looked at cognates from 28 languages across 200 meanings. Then she compared her tree with genomic data from Willerslev’s new survey. […]

To the researchers’ amazement, the genetic pattern mirrored the linguistic one. “It’s incredible that those two trees match. None of us expected that,” says paleoanthropologist Michael Westaway of Griffith University, Nathan, in Australia, a co-author on the Willerslev paper. “But it’s confusing: The [genetic splits] date to 30,000 years ago or more but the linguistic divisions are only maybe 6000 years old.”

He addresses counterarguments (R.M.W. Dixon “says these languages are so unique that new theories of linguistic change must be invented to explain them”; others “argue that the computational models, built for genes that can only be inherited, deal poorly with languages that spread by diffusion”) and finishes by saying that Aboriginal stories describe the birth of languages “much the way Bowern thinks it happened”:

In 2004, Evans recorded an Iwaidja speaker, Brian Yambikbik, explaining how his language might be related to the one spoken on distant islands. “We used to speak the same language as them, but then the sea came up and we drifted apart, and now our languages are different.”

Comments

  1. Interestingly, much importance is given to the supposed inability of these computational tree-reconstruction methods (inspired by genetics) to evaluate borrowings – at the same time when these same papers are heavily involved in identification of borrowings / admixtures in the genetic material. (The youtube link gives a short summary of one of the most fascinating of yesterday’s reports, demonstrating that the earliest wave of the human migration out of Africa may not have disappeared without a trace. It’s traces are found in the “borrowed” genetic material in New Guinea). One might hope that computational linguistic analysis will develop abilities to detect and analyze admixtures, too?
    https://www.youtube.com/watch?v=z9LRdoMeXC4

  2. Great article, thanks!

    A bit baffling to see “pama-nahyoongan” offered as a pronunciation of Pama-Nyungan, though. I guess it’s sort of equivalent to pronouncing Kyoto “kee-yoto”, but I feel like anyone reading this article could probably have handled “pama-nyoong-an”—and I would definitely recommend adding that second hyphen, to show that the last bit is more [ŋən] than [ŋɡən].

    (Unless there’s a standard Anglicized pronunciation in academic circles that really is like “pama-nahyoongan” and I’m just out of the loop, I guess.)

  3. That “h” is also possibly confusing, as to me “hyoo” suggests the sequence of sounds found at the start of the word “human.”

  4. Atkinson must have thick skin to still be publishing on linguistics after the philologists explained what was wrong with his paper on PIE (“Bouckaert et al.”)

  5. Oh, he’s a serial offender. I doubt if he reads any of the rebuttals, except long enough to offer a counter-rebuttal.

  6. A bit baffling to see “pama-nahyoongan” offered as a pronunciation of Pama-Nyungan

    I too was bothered by that.

  7. Bowern’ and Atkinson’s paper on Pama-Nyungan classification suffers from the usual flaw of phylogenetic linguistics, in that there’s no way offered to confirm the results other than run the black box again. However you can’t say that Atkinson did the work apart from traditional linguists.

  8. J.W. Brewer says:

    Indeed, the right-hand column of this page includes Prof. Bowern’s own irregularly updated blog in its list of Linguablogs with the coveted Hat Seal of Approval.

  9. David Marjanović says:

    other than run the black box again

    …What makes you think it’s a black box? Everything is documented. Les molécularistes don’t use these programs blindly, they understand what they’re doing.

    I’ll download the paper on Monday if I won’t forget again.

  10. What I meant was, data goes in and trees come out, but the algorithms are not usually configured to “show their work”, that is, to present the intermediate stages in a way that could be evaluated by people. In the case of linguistic phylogenetics, I would want to see the inferred innovations for each node, as well as the characters which are inconsistent with the tree.

  11. Stephen C. Carlson says:

    I would want to see the inferred innovations for each node, as well as the characters which are inconsistent with the tree.
    Based on what I understand of the algorithms, this should be possible. In fact, I’ve seen publications where they did this (though not in linguistic phylogenetics, IIRC).

  12. David Marjanović says:

    Based on what I understand of the algorithms, this should be possible.

    Yep, it’s easy to do. I might do it myself on Monday.

  13. Perhaps it might not be helpful, but I wonder if actual genetics could help in this case.

    The genetic part of these recent studies showed that there has been very little genetic mixing among the various Australian groups for tens of thousands of years. (In fact, southwestern Australians are almost as genetically distant to northeastern Australians as the Spanish are to native Brazilians.)

    This just doesn’t make a lot of sense if you believe that their languages are all connected by less than a few thousand years. How could that happen on a continent wide scale, while the languages kept the same phylogeny as the genetics the entire time?

    In Eurasia, it is easy to see how revolutionary changes in lifestyle and technology might drive language change, yet even there it seems to be that genetics and technology are very closely linked.

    I think it is crazy to apply that logic to Australia.

    All the evidence, so far, points to Australia being a unique case of a single language being introduced, and then evolving (mostly) with the people that were speaking.

  14. Jim (another one) says:

    “All the evidence, so far, points to Australia being a unique case of a single language being introduced, and then evolving (mostly) with the people that were speaking.”

    Then the genetic relationships that comprise Pama-Nyungan shouldn’t be discernible, given the time depth of colonization event. But they are. So either Pama-Nyungan is really a Sprachbund rather than a valid family, or at some point lots of disparate peoples shifted language for no apparent reason. These same groups have mostly shifted to English; what similar social dislocation as a causal factor can anyone even hypothesize around whatever time the proto-language broke up?

  15. I’ve seen arguments that Pama-Nyungan was brought by Indian migrants a few thousand years ago along with the dingo and perhaps some other innovations. I don’t know enough about the (human) genetics, though, to tell whether the South Asian influx idea holds water.

  16. The Dravidian-Australian idea used to be in vogue long ago, because, I think, black skin and retroflex consonants. Nothing to it, though.

  17. J.W. Brewer says:

    This piece has links to various others giving the ups and downs of the influx-of-South-Asian-DNA-to-Australia-a-few-millenia-ago argument, and claims the idea is currently the minority view but the issue is not yet definitively settled. Don’t know how fair a presentation it is, but maybe by following the various links someone sufficiently interested could get enough context to discount any excessive spin. http://theconversation.com/an-ancient-australian-connection-to-india-55935

  18. Christopher S says:

    Investigating the homepage of J.W. Brewers link, I found an unfortunate headline: “autonomous cars provide an experiment in policy-driven pubic health intervention”.

  19. Jim (another one) says:

    “I’ve seen arguments that Pama-Nyungan was brought by Indian migrants a few thousand years ago along with the dingo and perhaps some other innovations”

    Even if it were quite solid, there would still be the question of diffusion. Why would the language of that group come to cover most of the continent? This kind of language shift is rare, if it ever happens, in forager settings. We know how English came to cover the whole continent. That’s about the level of disruption it takes.

  20. “I’ve seen arguments that Pama-Nyungan was brought by Indian migrants a few thousand years ago along with the dingo and perhaps some other innovations”

    These new papers (with much much more data) found zero genetic evidence for this theory. So, if it happened, they essentially left no trace of mixture with the native Australians.

    It is also hard to understand how if “at some point lots of disparate peoples shifted language for no apparent reason”, the language relationships would mirror the genetic relationships, when the genetic relationships clearly arose from very ancient divergences.

  21. David Marjanović says:

    Yep, it’s easy to do. I might do it myself on Monday.

    …or Tuesday, or Wednesday.

    because, I think, black skin and retroflex consonants

    And a distinction between alveolar and dental, or apical and laminal alveolar, consonants, not found anywhere else to the best of my knowledge. Retroflexes are much more widespread.

  22. -All the evidence, so far, points to Australia being a unique case of a single language being introduced, and then evolving (mostly) with the people that were speaking.

    note that this appears to contradict archaeological evidence which says that humans reached Australia 45 thousand years ago.

    Certainly proto-Pama-Nyungan is nowhere that old.

    However, the riddle is nicely solved if we assume that the Aboriginals simply lacked language (or complex language anyway) before Pama-Nyungan came from New Guinea.

    Hence, extraordinary spread of Pama-Nyungan without a trace of existence of previous languages.

  23. Capra Internetensis says:

    Regarding forager spreads, most of Canada is covered by three language families: Athapaskan (Na-Dene), Algonquian (Algic), and Inuit (Eskimo-Aleut). If you throw in Alaska, Greenland, and relevant bits of the continental US that’s close to twice the area of Australia.

    It’s fun to write native Australian names with Tamil romanization, they look elegant.

    There was a theory that the spread of Pama-Nyungan was associated with the invention of a way to detoxify cycad seeds, which led to increased ability to hold feasts I think and the spread of a prestigious ceremonial system, or something like that iirc. There was also one connecting it to the spread of backed blade technology, only it turns out the backed blades are kind of superfluous and everyone just stopped using them after a while.

  24. Australian archaeologists have one weak point – apparently there is a taboo on comparing Pleistocene Australian stone technology with contemporary stone technologies in Europe.

    European archaeologists are not bound by this taboo and they say that Australian Aboriginal stone technology corresponds to Lower Paleolithic in Western Europe (ie, what was used by Homo Erectus and Homo Habilis hominids).

    Upper Paleolithic level of stone technology appeared in Australia sometime in mid-Holocene (around 5000 BP – so called Australian Small Tool Tradition).

    So judging by stone technology alone, an objective observer, not bound by considerations of political correctness, should indeed conclude that it is very likely that Australians before Holocene probably had level of intelligence comparable with Homo Erectus (lack of language and all that).

    Of course, the story is not that simple. There are other signs that at least some Australians and New Guineans (recall that New Guinea and Australia formed a single continent back then called Sahul) were somewhat more advanced showing other signs of behavioural modernity like rock art or use of shell ornaments.

    So a reasonable suggestion would be that before Holocene, Australia was mostly populated by creatures with Homo Erectus level of intelligence (and certainly lacking real language). Some of them were anatomically modern humans without language skills while some were hominids related to Homo Erectus.

    However, at least since the Last Glacial Maximum (circa 20000 BP), some parts of Sahul continent were inhabited by real humans who had real language and human level intelligence.

    For about ten thousand years, they shared the continent with their less intelligent cousins, but then something, perhaps, climatic changes or technological advance, tipped the balance in their favour.

  25. Using stone tool technology as a proxy for intelligence is unconvincing. Early skeletal remains are anatomically modern. As you say, rock art indicates people as intelligent as modern humans. Either we twist ourselves around to say that the tools of these attested people disappeared, yet the tools of their unattested primitive cousins remain. Or, the stone tools people were using then were adequate for their purposes.

  26. Capra Internetensis says:

    It isn’t only Australia that was using “Lower Palaeolithic” stone technology (i.e. large simple choppers and ad hoc flakes) down to the Holocene. It was also pretty much all of Southeast Asia, southern China, and parts of India. What’s more, ethnographic Australian Aborigines were *still* using mainly “Lower Palaeolithic” technology. (They did have ground stone axes though, which are ‘Neolithic’ or at least ‘Mesolithic’ in the traditional scheme. These go back to 35 000 years ago.)

    There is a very simple explanation for this: blade technology just isn’t that big a deal, at least not in that kind of environment.

    There are also a lot of Australian families which *aren’t* Pama-Nyungan, of course. And there are no Homo erectus fossils from Australia.

    So yeah.

  27. -rock art indicates people as intelligent as modern humans.

    earliest Australian rock art reliably dates to terminal Pleistocene; mostly concentrated in the north (Arnhem land in Northern Territory).

    -It was also pretty much all of Southeast Asia, southern China, and parts of India.

    and these are exactly the regions where native populations have significant Denisovan ancestry.

    The most recently known prehistoric Hominin population that did not look like modern humans was found here. https://en.wikipedia.org/wiki/Red_Deer_Cave_people are dated to between 14,500 and 11,500 BP, they had flat face, broad nose, jutting jaw with no chin, large molars, prominent brows, thick skull bones, and moderate-size brain. Thought to be a result of mating between Denisovans and modern humans.

  28. -And there are no Homo erectus fossils from Australia.

    there are finds of robust hominids which don’t quite look anatomically modern (Wilandra Lakes hominds, Kow Swamp hominid, etc).

    they could be a result of mating between South East Asian hominids (Homo Erectus from Java for example or Denisovans) and anatomically modern humans.

  29. There is no evidence that the first Australians were anything other than the direct ancestors of the natives that now live there.

    They have over 90% ‘modern human’ and the rest of their ancestry is Neanderthal and Denisovan, and maybe 1-2% from an early wave out of Africa. Across the continent, they have around 50,000 years worth of genetic drift.

    They have no Homo erectus ancestry, and there is no evidence that Homo erectus ever set foot on the continent.

    All of the ‘evidence’ from early skeletal forms is reminiscent of the Americas, where most older bones were originally classified as non-Native American. More recent genetic studies have shown them to be 100% Native American. In fact, many Native Australians today have prominent supraorbital ridges, sloped foreheads, and robust bones, yet speak and act normally.

    I find it quite ridiculous that someone would suggest that they didn’t speak at all when they arrived.

    It is actually more likely that even Homo erectus had some type of language skills, and certainly Neanderthals and Denisovans. Do you really believe they could pass along that amount of technology and skill for that long without any language at all?

    Modern humans even interbred with these archaic people on multiple occasions, and then the hybrid children clearly joined into the human population, because they had children of their own that survived, and now they have billions of ancestors.

    There are African populations who split off from the rest of humanity well before the main Out-Of-Africa migration (up to 200,000 years ago), and they have no problem using very complex languages, and no problem learning and using multiple unrelated languages.

  30. I’m confused. Is anyone proposing that all five of the following statements are true?

    – Australia was colonized by speech-capable humans 45.000 years ago.
    – Genetics allows the pre-European population of Australia to be organized in a tree by descent.
    – Proto-Pama-Nyungan was a unitary language about 5.000 years ago.
    – Comparative works allows the current Pama-Nyungan languages of Australia to be organized in a tree by descent.
    – The linguistic and genetic trees largely correspond, including the (approximate) position of the root.

    Of course the genetic tree could be rooted in a dispersion/conquest event corresponding to the introduction of Proto-Pama-Nyungan to Australia. In that case the timing would make sense, but then it has no relevance to the question of whether there was language in Australia before that.

    If the genetic tree is rooted at the colonization event, the timing makes no sense.

  31. prominent supraorbital ridges, sloped foreheads, and robust bones

    Indeed, I have had technical conversations with at least two people who looked like this (but were not Native Australian). I can’t swear to their bones (they looked stocky), but they passed the Turing test with flying colors. A few shared primitive traits does not a species boundary make.

  32. Lars,

    I would classify your statements into these groups:

    Absolutely true:
    – Australia was colonized by speech-capable humans 45.000 years ago.
    – Genetics allows the pre-European population of Australia to be organized in a tree by descent.

    Likely true:
    – Proto-Pama-Nyungan was a unitary language about 5.000 years ago.

    Unlikely to be true:
    – Comparative works allows the current Pama-Nyungan languages of Australia to be organized in a tree by descent.

    Very unlikely to be true:
    – The linguistic and genetic trees largely correspond, including the (approximate) position of the root.

  33. J.W. Brewer says:

    What do we make of the facts that: a) the non-Pama-Nyungan languages are more diverse (i.e. not thus far assigned to a single family or even macro-family by most scholars; b) the non-Pama-Nyungan languages are all (although it may be hazardous to assume geographical range as of first European contact = geographical range further back in time?) concentrated in one modestly-sized piece of the continent, which thus has greater linguistic diversity than the much larger balance of the continent; and c) coincidentally or otherwise that place in which the diverse range of non-Pama-Nyungan languages are concentrated is more or less the part of the continent where one would expect post-Ice-Age external contact to have been most likely, i.e. the north coast that has been reachable from Indonesia by sea once comparatively-modest out-of-sight-of-land sailing technology and skills had developed.

  34. I must say, SFReader, I’ve never seen anyone else so confidently pinpoint the beginning of language (and “sentience”, as you call it) at the onset of behavioral modernity – itself a fraught concept.

  35. It reminds me of Julian Jaynes and his briefly popular The Origin of Consciousness in the Breakdown of the Bicameral Mind.

  36. Jim (another one) says:

    David,
    “And a distinction between alveolar and dental, or apical and laminal alveolar, consonants, not found anywhere else to the best of my knowledge.”

    Some varieties of Irish distinguish alveolar from dental consonants, where others have an alveolar/palatal distinction.

    Capra,
    “Regarding forager spreads, most of Canada is covered by three language families: Athapaskan (Na-Dene), Algonquian (Algic), and Inuit (Eskimo-Aleut). If you throw in Alaska, Greenland, and relevant bits of the continental US that’s close to twice the area of Australia.”

    There is a difference between spreading over uninhabited territory that you happen to be able to exploit and make habitable and spreading into existing populations. An analog is the spread of Lakota and various Algonkian speakers onto the Great Plains when the horse made that are truly habitable. they shouldered the very scant Athabaskan population aside but never dislodged the Pawnee or Crow (or Mandan, or Hidatsa etc.) who were much more numerous and established.

  37. J. W. Brewer: The diversity of N. Australia has a great deal to do with its biological productivity, which supports a lot of small territories, rather than just an old age of diversification.

  38. Capra Internetensis says:

    Jim (another one),

    None of those territories were uninhabited at the time of the spread. Well, we don’t know exactly when Algonquian or Athapaskan spread, but almost no one thinks they’re anywhere near *that* old. We do know with near-certainty when the Inuit spread (very recently), and the Dorset had already successfully inhabited the High Arctic for thousands of years before them.

    As for the example of the horse, well, some technology giving an advantage in subsistence and/or warfare has certainly been proposed for the spread of Pama-Nyungan (backed blades, cycad detoxification). Part of the spread was into the western desert, and that area probably was largely depopulated prior to that due to a climate downturn.

    Really though there is very little agreement on the cause of most prehistoric language spreads, it’s not like Pama-Nyungan is unusual in this respect.

  39. J.W. Brewer says:

    Although we have no idea what sort of language the Dorset people spoke and I think the claim is that they died off w/o issue rather than being assimilated (linguistically or otherwise) into the newly-arriving Inuit society. The puzzle here seems to be the tension created by the belief that the current-day genes spread out many millenia before the current-day language did, meaning massive language shift among existing populations over most of the continent.

  40. marie-lucie says:

    David: “And a distinction between alveolar and dental, or apical and laminal alveolar, consonants, not found anywhere else to the best of my knowledge.”

    Several “Penutian“ languages of California such as the Miwokan and Yokutsan families, as well as the (most likely related) Takelma language of Southern Oregon, have this distinction. My interpretation is that at least one of the series derives from an earlier cluster such as *tK (K being a velar of some sort) atttested in some languages of the group:

  41. m.-l., what would be an example of some of these tK-t correspondences?

  42. @Jim (Another one), Capra Internetensis: In the case of the spread of Cree I had mentioned some relevant factors on this thread (see my July 28 comment):

    http://languagehat.com/muskogean-and-lambs-quarters/#comments

    It is perhaps relevant to mention, in this connection, that one Algonquian scholar claimed that a number of Cree words without an Algonquian etymology could be assumed to derive from the pre-Cree, pre-Algonquian language(s) which Cree replaced in the Canadian subarctic.

    I know even less about the spread of Pama-Nyungan than I do about Cree, but one datum which is relevant to the present discussion is that one scholar has claimed to have found some widespread vocabulary items in various Pama-Nyungan languages which can be reconstructed back to Proto-Pama-Nyungan, and which are argued to be Austronesian loans into Proto-Pama-Nyungan. If true, this would confirm that the Proto-Pama-Nyungan spread must be comparatively recent (i.e. it must postdate the expansion of Austronesian into what is today Indonesia and Coastal Papua New Guinea) and that Proto-Pama-Nyungan must have been spoken in Northern Australia. Or perhaps even further North, outside Australia.

    Marie-Lucie: I second Y’s request: examples of comparative Penutian are always of interest to me, and I suspect to others here as well.

  43. marie-lucie says:

    Y; Etienne: Thank you. Give me more time and I will find some for you.

  44. Trond Engen says:

    The dingo is one such possible innovation, of roughly the same age as Proto-PN. That would indeed mean that the expansion originated in a contact event. A contact event could also mean that the Pama-Nyungan expansion was helped by some imported contagious disease which the Proto-Pama-Nyungans of Cape York and the assorted Northerners of Arnhem Land had been exposed to before. Or so I thought. The new genetic evidence seems to refute the hypothesis with brutal force.

  45. Trond Engen says:

    Or the Proto-Pama-Nyungans of the Northern Territory, if that’s were they expanded from.

  46. David Marjanović says:

    …or Tuesday, or Wednesday.

    I have now read all three papers (among others). But they’re Nature papers, so they’re extended abstracts, while the actual papers make up most of the “supplementary information”, and I haven’t read that yet. The language tree is entirely buried there; the “paper” doesn’t say more than we’ve already heard.

    I have, however, read much of a long paper that argues for using the settlement of Australia as the last possible date for the origin of full-blown language, because one does not simply cross 8 to 10 stretches of open ocean without being able to talk about how to build a sea-going boat.

    However, the riddle is nicely solved if we assume that the Aboriginals simply lacked language (or complex language anyway) before Pama-Nyungan came from New Guinea.

    There are non-Pama-Nyungan language families (and isolates) in northern Australia; and some of them have been shown to be related to Pama-Nyungan.

    without a trace of existence of previous languages

    The Tasmanian languages weren’t Pama-Nyungan. And there’s evidence from Victoria that the local languages allowed word-initial [r], which was found on Tasmania but not in the known Pama-Nyungan languages.

    There was a theory that the spread of Pama-Nyungan was associated with the invention of a way to detoxify cycad seeds, which led to increased ability to hold feasts I think and the spread of a prestigious ceremonial system, or something like that iirc.

    Ooh, that makes a lot of sense! Similar things have been proposed to explain the spread of Indo-European in Neolithic Europe.

    This could even explain how the backed blades spread if they were as useless as you say. 🙂

    However, at least since the Last Glacial Maximum (circa 20000 BP), some parts of Sahul continent were inhabited by real humans who had real language and human level intelligence.

    Some rock art is older than that, IIRC.

    It is actually more likely that even Homo erectus had some type of language skills, and certainly Neanderthals and Denisovans.

    Yep. Claims that Neandertal anatomy didn’t allow spoken language have turned out to be erroneous.

    It reminds me of Julian Jaynes and his briefly popular The Origin of Consciousness in the Breakdown of the Bicameral Mind.

    Yeah.

    Some varieties of Irish distinguish alveolar from dental consonants

    Oh! Yes! I learned this on Wikipedia and then forgot. 🙁

    m-l, I didn’t know about the languages you’ve mentioned, though. I third Y’s request. 🙂

  47. The “alveolar” stops Marie-Lucie mentioned are, in some languages, more like postalveolar or retroflex stops, and they are often described in the literature as retroflex. Perhaps that is why this contrast is not so well-represented in lists of phonological curiosities.

  48. David, the Nature papers are genetic, are they not? I thought you were going to run Atkinson and Bowern’s linguistic data from their paper a few years back.

  49. I maintain that anatomically modern humans, Neanderthals, Denisovans, Homo Erectus and possibly even earlier hominids were CAPABLE of speech if somebody TAUGHT them.

    In other words, they had hardware for speech (brains and anatomy), but it doesn’t mean they necessarily had software for it.

    And the software of language is essentially a cultural tradition.

    Someone has to teach children the language (ie, speak the language around them and speak with them, so they would pick it up) and these children would teach it to their own children when they grow up and so on and on for tens of thousands of years.

    If there is no such tradition (of if it is interrupted and there are no groups nearby from which it could be picked up again), then there is no language (and sentience as we understand it).

    I hope this clarifies the misunderstanding.

    Note that I don’t believe all anatomically modern humans had language or that other hominids couldn’t have had it. They could and probably had, it’s not an inherent biological trait, but just a cultural tradition, one which could be learned and taught.

    Now, do we have hard evidence for existence of language in Lower Paleolithic?

    Of course not, we don’t have and couldn’t have such evidence, because before invention of writing, language couldn’t and didn’t leave any archaeological traces.

    But somehow, tradition of making same stone tools in the same rigid pattern, without a slightest change for millions of years doesn’t strike me as something which is transmitted by speech.

    Surely it’s a sign of instinctive behaviour, not intelligent human activity.

  50. sentience

    Sapience. Sentience has a meaning, and it’s not what you’re using it to mean.

    This also seems to presume an awfully binary conception of language and not-language. You don’t envision there being much of a transitional phase between the two? I mean, it’s hard for me to imagine a social group of anatomically modern humans, who have both the mental ability to speak (unlike chimps) and the physical ability to speak (again unlike chimps), not inevitably having some form of vocal communication which, if not quite at the level of modern language, would at least be fascinatingly similar to it.

  51. @ SFReader: But the thing is that there are signs of cultural traditions used by Neanderthals, e. g. funerary rites. Also, there has been spread of technology already in the Lower Paleolithic, there are 380 k year old javelins found in Europe and archeological evidence for building shelters. One thing about the Stone Age that bears repeating is that it actually was mostly a wood and natural fiber age, and that the stone tools are probably the most uninteresting product of the humans of that time, but unfortunately the one product that was conserved best. The reason why innovation was so slow here is most probably not that people were incapable of coming up with new ideas, but that innovation didn’t bring sufficient advantage. And culture can be an obstacle to the dissemination of innovations as well as supporting it. So, in summary, I would neither take slow innovation in stone tools as a sign of absence of language, nor does the record show no innovation or signs of culture before the relatively late dates you propose.

  52. Additionally, I don’t think that “instinctive” is a correct term to use for the creation of tools by our paleolithic ancestors. Non-instinctive group-specific practices, like use of specific tools or avoidance versus embracing the crossing of water courses has been observed in ape species, so it is reasonable to assume such non-instinctive behavior also for the early hominids. I may be undermining my previous argument here, as this shows that a certain degree of learning and transmission of culture and techniques is possible by imitation and primitive communication (showing approval / disapproval, emitting signals of warning, etc.). But while a certain degree of culture (in the sense of group-specific practices transmitted by learning) is clearly possible without language, we actually don’t know enough about these things to confidently state that a certain degree of culture or innovation is proof for the presence or absence of language.

  53. Since sign languages have been created quite a few times, pretty much ex nihilo, by people who had had no language before, I think spoken language would have been invented the same way, soon enough, by your non-speaking hominids. In other words, the situation of people who can speak but don’t wouldn’t hold long.

  54. Stephen C. Carlson says:

    I think spoken language would have been invented the same way
    I agree and this makes me skeptical of the assumption of the monophyly of language inherent in “Proto-World.”

  55. David Marjanović says:

    What Hans said.

    this makes me skeptical of the assumption of the monophyly of language inherent in “Proto-World.”

    …But without a Proto-World, you have to postulate that people who could speak but didn’t existed for long enough to spread, divide into reasonably isolated groups, and then invent language separately.

    I think language developed quite slowly, and Proto-World could have been a rather odd language by modern standards that didn’t necessarily have all of… whatever few linguistic universals there are today. But downright separate developments? That’s hard to imagine.

  56. David Marjanović says:

    David, the Nature papers are genetic, are they not? I thought you were going to run Atkinson and Bowern’s linguistic data from their paper a few years back.

    On one of the papers, the one specifically about Australia, Bowern is a coauthor, and chapter 15 or so of the supplementary information presents a linguistic tree. I haven’t looked into it yet; maybe it’s straight from that paper.

    Surely it’s a sign of instinctive behaviour, not intelligent human activity.

    Instinctive? Innate, not learned? Like the way we’re born knowing how to drink?
    How would that have evolved in the first place?
    How did this instinct disappear so thoroughly?
    How did it ever get so elaborate? Even bird migration is only partially an instinct. At certain times of the year, migratory birds will try to fly in a certain direction; the exact direction, and also whether they’ll do this at all, are heritable and can mutate. But birds that aren’t doing this the first time can correct their course when they notice that their starting point is different this time.

    I’m not saying there are no instinctual behaviors in humans. Certain kinds of smiles (most of them fake, but that’s beside the point) automatically make me aggressive. (…Not afraid, interestingly, but directly angry.) But sitting down for a good long while and making a stone tool is a different category.

    Even the use of unmodified stones by one chimpanzee population to crack nuts is cultural.

  57. Stephen C. Carlson says:

    I’m not doubting that an original human community spoke a language (which one I suppose could call Proto-World), or even that some languages today are descendents of it. I’m just skeptical that all languages that humans speak today are descendants of it. Languages can be and have been invented, for example, Klingon. Twins have been known to invent their own language (cryptophasia), and all it takes is a couple of these to take root over the many millennia to have independent language families.

  58. “How did this instinct disappear so thoroughly?”

    This isn’t a language observation, but I have a few children, and it is nearly impossible to stop them from making perfect Acheulean Handaxes from stones in the garden.

    I never learned to do it myself (as I only had legos to play with as a child) but I guess it is just a pure instinct passed down genetically from the million years of non-cultural transmission of this technology.

  59. – by your non-speaking hominids

    You guessed right.

    Indeed, I used to be a non-speaking hominid.

    For about two years.

  60. Capra Internetensis says:

    Thanks, Etienne, seems like a plausible idea (whether or not accurate in any given case).

    SFReader, how would you go about falsifying your theory?

  61. David Marjanović says:

    Still no time to burrow through the supp. inf.. Maybe only on the extended weekend.

    Languages can be and have been invented, for example, Klingon. Twins have been known to invent their own language (cryptophasia), and all it takes is a couple of these to take root over the many millennia to have independent language families.

    Sure, but how likely is it that a whole community would take up such a language?

  62. Cryptophasia produces a language that is parasitic on the parents’ language.

  63. David: there exist creoles today, some spoken by millions of native speakers, which arose through the spread and nativization of a pidgin language. I speculated a little, here at Casa Hat, about how a purely invented language could have spread and nativized in a similar fashion:

    http://languagehat.com/the-bookshelf-in-the-land-of-invented-languages/

  64. how likely is it that a whole community would take up such a language?

    There have been suggestions that some of the many language families in California arose when a few very young children were lost or orphaned at very young ages — since the living conditions are such that a couple of five-year-olds could plausibly survive to adulthood and found a new tribe, only retaining a few content words, if that, from their parental language.

    Whether that counts as a new glossogenesis is a good question — the idea of language as a useful thing to have would at least not be new to such children. And another question is whether a few children can found a speech community robust enough to survive the inevitable meeting with other languages.

    Also I don’t know if it’s still seen as plausible that early modern humans lived in dispersed nuclear families — but if we accept that, it seems likely to me that over a few hundred thousand years it will happen that children are brought up by parents who both have lost or never acquired language — because of congenital deafness for instance.

    The same caveats apply there, of course. Siblings don’t usually form families themselves, and successfully retaining and transmitting their new language in nuclear families where the other parent speaks a pre-existing language might not be very likely. But not impossible.

    Rambling ends.

  65. There have been suggestions that some of the many language families in California arose when a few very young children were lost or orphaned at very young ages — do you have a source for that? It seems very odd idea, with nothing to recommend it over the usual processes of language diversification.

  66. -SFReader, how would you go about falsifying your theory?

    Short of time travel?

    I suppose we could try to reconstruct and date proto-Australo-Papuan language. If it’s around 40-45 thousand years old, then my theory is obviously wrong and first settlers of Sahul continent already possessed language. (assuming there was one single wave of colonization)

    If it’s considerably younger, say, 20-25 thousand years, then the theory that the original settlers had no language will be strengthened (of course, short of time travel it can’t be proven either)

  67. @Y, nothing academic. It was the sort of thing that was quoted in popular books on languages 30 years ago, but I think I remember seeing it as a factoid within the last few years as well.

    And one point against it is that (as far as I know) the recorded events where sign languages arose spontaneously were at schools where there were tens if not hundreds of children. I can’t imagine a pre-technological scenario that would bring so many language-less children together if the older generation had language.

  68. Stephen C. Carlson says:

    It was the sort of thing that was quoted in popular books on languages 30 years ago
    I vaguely recall that too (and something similar for Papua New Guinea). I don’t know if it is still or ever valid.

  69. marie-lucie says:

    “There have been suggestions that some of the many language families in California arose when a few very young children were lost or orphaned at very young ages”

    I find this very doubtful. It is true that those languages are very diverse, but few of them (or the families they belong to) are limited to California. Besides; very young children supposedly lost or orphaned are unlikely to survive on their own for long: they might be adopted by some animals (in which case they lose language or were too young to have acquired it), or more likely, found and adopted by other humans whose language they learn, like any other normal children.

    The hypothesis seems to have come from people overwhelmed by the diversity of the languages and not competent enough as linguists to study them seriously. As for a reference, I wonder if Golla 1911 mentions the hypothesis.

  70. David Marjanović says:

    David: there exist creoles today, some spoken by millions of native speakers, which arose through the spread and nativization of a pidgin language.

    Of course: if a language is useful for wider communication, or acquires enough prestige, it will spread, no matter what its origins are.

    But how, in preliterate times, does one go about inventing a language and making it useful or prestigious enough for this? The closest thing, as you mentioned, is Damin, which isn’t close. I’m not saying it’s impossible, but I can’t imagine that it’s at all probable.

  71. Here is a genetic explanation of the Australian languages, but without a definite reason for the genetic and linguistic expansion:

    We find that Papuan and Aboriginal Australian ancestors diversified 25–40 thousand years ago (kya), suggesting pre-Holocene population structure in the ancient continent of Sahul (Australia, New Guinea and Tasmania).

    However, all of the studied Aboriginal Australians descend from a single founding population that differentiated <32 kya. We infer a population expansion in northeast Australia during the Holocene epoch (the past 10,000 years) associated with LIMITED GENE FLOW from this region to the rest of Australia, consistent with the spread of the Pama–Nyungan languages.

  72. Trond Engen says:

    Interesting. Where is it from?

  73. A genomic history of Aboriginal Australia

    Malaspinas et al.

Speak Your Mind

*