Evolutionary nonsense
It’s a matter of word play, I used to think, when people talk about the evolution of language. Evolution can simply be the synonym of change, development, improvement, optimization, adaptation, etc. More often than not, the evolution of language is talked about in these senses, which have only tangential relationship with Darwinian evolution.
I am reading Bernard Bichakjian’s 2002 book "Language in a Darwinian Perspective" — only starting. But I am already struck by how one can move freely between the two senses of evolution to make an — or any — argument. Half way through the book, it’s clear to me that he wanted to say language changes is directional rather than random. That’s fine, and it’s evolution in the first sense. But Darwinism?
What’s the unit of selection? Bichakjian’s answer is linguistic features, but that doesn’t make sense at all. A frequent argument used in the book is that if something is easier to learn, it’s closer to the biological programming and therefore adaptive (need to find a quote here). Doesn’t sound right, does it? All babies in the world start cooing with /ahh/ or /ohhh/. So the most evolutionary language is Hawaiian?
I refer here to the vast literature on pidgins, language changes, OT, language acquisition, on and on to make the simple point that language is a system, not a collection of features. Whatever the force is, the pressure is on the language as a system. (This is reminiscent of the debate — e.g., somewhere on http://www.edge.org/ — on whether the unit of evolution is the invidual gene, but the analogy doesn’t go all the way.)
Where’s the evolutionary pressure? Presumably speakers who "possess" the adaptive linguistic features will outlive ones who do not? Or may be a speaker community? Whatever it is, it ignores the basic fact that languages change (for the most part) because of linguistic contact. Who dies in a linguistic contact?
The acquisition argument is the most utterly shocking one I’ve seen so far (and I am still reading). Here’s a quote (also in a lead article in 1999 by Bichakjian):
Terry Deacon states unambiguously that ‘languages do not change, they EVOLVE’, and adds further on that ‘the world’s … languages have all evolved to become learnable at the earliest age possible’ (1997:109 and 137; with original emphasis). This is very encouraging.
What? So initially all languages were only learned in, say, adolescense, and were only later mended to be learned by age 3? And what is the reproductive advantage for a 3-year-old to learn a FEATURE of a language instead of a year or two later? Absolute nonsense.
June 6th, 2007 at 7:50 am e
Very good points. If anything, it’s likely that language was initially only learnable by the youngest of the young.
June 6th, 2007 at 9:12 am e
Can’t agree more. There are some really exciting work on game-theoretical modeling of language evolution (I try to keep track of them on my citeulike site). Recursive grammar, or more generally, consistency in syntax, seems to result from the acquisition bottleneck. So in a sense the learnability argument has to be turned up side down — the starting point is not if a language is learnable, but language is by definition learnable, not only by individuals but by a community of speakers.
It’s perfectly ok to talk about the evolution of language, but the mechanism may or may not be Darwinistic. In fact, chances are, the analogue will only work in spirit, but not in details.
I was reading this morning the part of the book on why the Proto-Indo-European language switched from head-last to head-first. As soon as I see citations on Michael Posner’s early work on hemisphere specialization, I knew it’s time to close the book. You can imagine where it goes — dead wrong.
August 3rd, 2007 at 3:50 am e
Evolutionary sense
A reply to Gary Feng’s “Evolutionary Nonsense”
By Bernard H. Bichakjian
BHB@Post.Harvard.edu
http://www.bichakjian.com/bernard/
Aug. 3, 2007
After a short introduction on the uses of the word evolution, Gary Feng (henceforth GF) writes that he is reading my latest book Language in a Darwinian Perspective (henceforth LDP) and concludes: “Absolute nonsense.” Since Stephen Gould and Ernst Mayr had re-acted positively to the ms of an earlier book of mine and to a nutshell presentation of LDP, respectively, I am not terribly shaken by GF’s judgment.
I shall discuss his objections point by point. GF writes: “Half way through the book, it’s clear to me that he wanted to say language change is directional rather than random. That’s fine, … But Darwinism?” He questions the Darwinian part because he apparently is focusing only on half of what I have said. The important changes that have taken place in the history of languages are indeed unidirectional, but the second point that I have made is that when a unidirectional takes place the output has selective advantages that the input didn’t have. It’s this second point that justifies my studying language in a (not the) Darwinian perspective.
FG writes further: “A frequent argument used in the book is that if something is easier to learn, it’s closer to the biological programming … (need to find a quote here).” He can stop looking for the quote, because he won’t find it. What I have said is that if a linguistic feature, and by that I mean a speech sound, a marking device, a grammatical distinction or a syntactic strategy, is acquired earlier and is at least equally functional has selective advantages over a late acquired homologue. I have never said anything about its being “closer to the biological programming.” In fact, I have specifically argued against the al-leged existence of a biological program. If we compare the obstruents (true consonants) of the Indo-European protolanguage and those of its daughter languages today, we see that practically all of them have been replaced by earlier-acquired alternatives. That does not mean languages have been brought closer to an alleged biological program, it simply means the model was made ever more efficient, just like everything we do in life: the second, third, fourth time around we try to do it with greater efficiency.
GF asks what the unit of selection is and, by way of corollary, where the evolutionary pressure is applied. These are not easy questions to answer, and he rightly refers to the situation in biology, where Ernst Mayr and like-minded evolutionists will argue that the species is the unit of selection, while Richard Dawkins et sui will defend the view that the gene is the unit. Likewise, one may ask whether in language the unit of selection is the system or the feature as I define it here (see supra). If the item that changes is correlated with another one, the system also changes. That’s especially the case in phonology where the system reorganizes itself as a phoneme changes. But it certainly is the linguistic item that is replaced that bears the selection pressure.
GF also wonders, I presume seriously, whether “speakers who ‘possess’ the adaptive lin-guistic features will outlive [the] ones who do not.” The answer depends on what is meant by ‘possess.’ The last generation of spats wearers died with the use of the ankle gaiters, perhaps someday the last generation of necktie wearers with die with their silk accessory, just as a generation of “as I said” sayers are living their twilight years. That a generation dies with a given feature is a common fact of life, but speakers do not “pos-sess” a linguistic feature the same way the Irish elk possessed antlers.
GF contends that the selection pressure approach “ignores the basic fact that languages change (for the most part) because of linguistic contact.” This is inaccurate on both counts. First, languages change considerably without contact. French has not been in sig-nificant contact with another language since the Carolingian period, and yet Modern French and ninth century French are quite different languages. Old and Modern English are also quite apart from each other and the phonetic and grammatical changes that have taken place were not caused by language contact. Second, when there is contact and bor-rowing takes place, the imported items always have social or linguistic advantages.
Finally, GF quotes Terry Deacon, who had written that “languages have all evolved to become learnable at the earliest age possible” and asks facetiously: “So initially all lan-guages were only learned in, say, adolescence, and were only later mended to be learned by age 3?” Deacon did not mean that at all. What he meant – and I am the one who has revealed those facts – is that the mastery of ancestral languages required many more years than that of modern languages. There is nothing odd about that, and there are plenty of empirical data supporting the statement that a variety of ancestral features is acquired around the age of ten. There are also plenty of advantages cognitive, social, and neuro-logical to early-acquired features.
In a subsequent post, GF questions my tentative explanation of why ancestral languages were head-last. My relying in part on Posner and Raichle’s findings, though I do so in part only, is in his view a grave error. Perhaps. I let him debate that with them, but if GF has a better explanation of why ancestral languages were head-last, I am certainly willing to listen.
In closing, perhaps I could submit to GF that all of us in the field of science are trying to find the truth, and our chances of finding it are far greater when we carefully examine and discuss one another’s hypotheses than when we reach rash judgments and wield depreca-tory comments.
August 3rd, 2007 at 9:49 am e
In the spirit of blogs, I will say what I think, and admit where I was wrong. And I will try to keep it brief.
======
0. Readers and Scholars
I resonate strongly with the last paragraph of Bernard’s comment. But somehow I felt immune from its intent. I am not an expert in language evolution, and will never be. As a reader, I rely on scholars in the field to feed me theories and facts, and I can take them or leave them.
Maybe this is too much consumerism, but I felt that every reader is entitled of his or her opinion and, in the era of blogs, to publish whatever they have to say, without the prerequisite of Ling101 (though Bernard criticized Intro Ling on his newly established blog for not introducing evolutionary theories).
I never expect the hear from any author mentioned on my blog, so it’s a honor and pleasure to have Bernard here.
1. Re: Absolute nonsense.
Bernard was offended by the phrase, and the hurt feeling is misplaced. I was talking about the Deacon scenario, not the book as a whole.
2. The missing quote
Bernard is right. I may never be able to recover a verbatim quote from the book, so I won’t even try.
But if obstruents in the proto-Indo-European were replaced with “easier” consonants, what is the basis for saying one sound is easier than another? The way our articulatory system (and the brain, and the bones, …) is shaped, right? You see what I meant.
3. Unit of selection
No argument here. Bernard’s comment entertained both “language” and “feature” as units, and does not come down to a verdict. My reading of the book is “feature”, and I disagree.
4. Possessing a linguistic feature
Thanks Bernard for explicating my point. I double-quoted “possess” precisely because I don’t believe linguistic features are inherent to an individual, as genes are. The fact that you can learn a new language (or a new linguistic feature) without having to be re-born suggests the evolution (read: change) of language may not follow the classic Darwinian model.
Think this way — what if we can freely acquire any gene we want, try it, and return it if you don’t like it (sounds like a good business plan, unfortunately)? Are we still playing the same Darwinian game?
5. Language change.
I was clearly being careless in my original post. Language contact is a primary force in language change, but not the only way. Bernard is right.
6. The last point:
This is what prompted my “nonsense” comment — that language is evolved to be acquired earlier. I am not familiar with the facts here — per Bernard’s advice to stick to the facts I should be mute on this one.
But you know I won’t. Being a psychologist I am more or less familiar with the Age of Acquisition (AoA) effects of words and morphological/syntactic structures. So I will go on.
My original question was two fold: (a) why emphasize on FEATUREs rather than other units, and (b) what’s the REPRODUCTIVE advantage of learning a feature earlier. The reasons still escape me.
========
So in sum, my original comments on Bernard’s book could have been better crafted — but that wouldn’t make it a blog entry — and I accept responsibilities for errors in it. The substantive questions I had — unit of evolution, advantage of early acquisition, selection pressure — are still open, though.
I am not surprised that Bernard wasn’t “terribly shaken”, as these are basic questions that scholars in this field (should) have been struggled for years. Bernard may have Stephen Gould and Ernst Mayr on his side, but I am still not satisfied with the answers he gave.
There goes my short reply.
August 4th, 2007 at 4:03 pm e
(Bernard emailed me the follow response and asked me to post — gary)
======
First, my thanks to Greg for posting my comment. Blogging is admittedly a genre of its own, often an impulsive one; yet, behind the abrupt rhetoric there may hide a sensitive and gracious person. I am glad Greg is such a person.
I will try to give two short reactions to Greg’s remarks.
First, about early acquired features. I have argued – and this is what gives my account of language evolution a Darwinian perspective – that in unidirectional changes the output has selective advantages over the input. One of the advantages can be in the form of early acquisition. If a linguistic feature – speech sound, marking device, grammatical distinction, or syntactic structure – is acquired earlier by children learning their native language, its neuronal interface is likely to be stronger, making it less vulnerable to fatigue, psychological stress and degenerative erosion. Since the neuromuscular algorithm supporting them is likely to be simpler and smaller, earlier acquired features are also more economical: their storage requires a smaller neurological investment, and their operation necessitates a smaller expenditure of energy. Moreover, the sooner children are proficient in language, the sooner they can initiate and expend their learning efforts and their socializing activities. In turn, the earlier this knowledge and these skills are acquired the stronger their neuronal support is. So, there is a wealth of cascading advantages to early acquired features.
There are also functional advantages, but I won’t discuss them here.
The second point that I will make is about the current Zeitgeist. When the nineteenth century naturalists found the courage to challenge the Creationist scenario and suggest instead an alternative based on empirical data and rational thinking, they found themselves fighting against an all-powerful establishment and the crushing weight of a tradition nearly three millennia old. Later, in the early part of the twentieth century, geneticists took a few years before dropping their objections to Darwinian thinking.
The situation in linguistics today displays a striking parallel. The establishment has changed and the Church has given way to a rising phalanx of self-proclaimed custodians of social justice – in linguistics and elsewhere – who consider it their sacred duty to see to it that evolutionary thinking be rigorously confined to the study of bones and other body parts. In addition, there is the hostility of those in mainstream linguistics who are distraught to see that important questions are asked and answers sought outside their theoretical framework. But that’s in the nature of science: sooner or later, orthodoxies are challenged and amended. That’s because not only languages, but also theories evolve!
August 4th, 2007 at 6:10 pm e
I thank Bernard for replying and agree whole heartedly with the two points he made above.
I am fully aware of cognitive benefits of early acquisition, and being an atheist I do not have to be convinced of evolution. The argument here is whether language change is Darwinian in a strict sense. Barnard made his points here and in his book. With due respect, I reserve my judgment.
Barnard, being a leading scholar in the field, probably feels much more responsibility as he fights the establishment. For me, though, Darwinian evolution is just a theory that at the moment is better than alternatives. I am would not hesitate to replace it with another if the latter better fits evidence.