A Brief Guide to [A Brief Guide to Language Evolution for Linguists – by Derek Bickerton]

I’ve been rereading the Derek Bickerton paper, A Brief Guide to Language Evolution for Linguists quite a lot recently. Its a really good attempt to synthesize the diverse debates about language evolution into something that an average linguist can quickly refer to. Unfortunately I think that Bickerton is not very impartial in his representation of the issues and I feel duty bound to pick my way through it and give a little commentary. It is a very useful paper and he makes some excellent points, however there are some parts where I think he is promoting his position whilst claiming to be an impartial guide.

In the paper Bickerton attempts to streamline the debate into some sort of hierarchy of importance. The two issues he holds as central to the language evolution programme are the evolution of symbolic units, and the subsequent evolution of syntax. He argues that these are only two central issues that we should be concerned with, and these evolved in order because of the commonsense notion that syntactic systems couldn’t evolve until there were some sort of symbolic units to be organised.
Secondary to the ‘big two’ are a set of questions that Bickerton considers of lesser relevance but still important enough to be investigated.

  1. What was the initial selective pressure for the first steps toward language?

  2. Was the subsequent development of language gradual or abrupt?

  3. How did the development of phonology relate to (1) and (2)?

He apologizes for the relegation of evolutionary phonology to the second tier but argues that without first explaining the evolution of symbolic meaning and syntax, phonology can’t be easily investigated. He is spot on about the importance of the analysis of selective pressures and the speed of language development. These are both essential for helping to contextualise the evolution of any aspect of language.

Bickerton’s bugbears
After this confident and accurate start, Bickerton reveals the secondary agenda of this essay – to cast doubt on the areas of investigation that he feels are unhelpful or irrelevant. Whilst some of his criticism is justified, the evidence simply doesn’t exist to provide a comprehensive unifying evolutionary pathway yet. Bickerton does acknowledge how complex and conflicted the field is, but gives us a reductive approach, dismissing lots of research programmes out of hand. Given how much of the field he dismisses, I think it’s in any linguist’s interests to go through these dismissals and try to put it in a broader context.

The question of whether language began with speech or with signs is irrelevant. (p512)
Bickerton argues that his own definition of protolanguage is sufficiently broad to cover both gestural and verbal communication, and for him, the important fact is the presence of symbolism, not the modality. I would argue that this is too restrictive a view as it presumes that the mental representation of language is distinct from the modality. Whilst the linguistic consensus on modern sign language is favourable to the idea of parity between gestural and vocal structure and complexity, there is evidence of some underlying neural differences in the use of sign language. Bickerton can’t dismiss the importance of gestural/linguistic divide until we know the selective pressures under which language evolved. The motor-cortical theory of language evolution certainly makes a strong case for motor systems being exapted in language evolution. The language/gesture divide is important for understanding this hypothesis. It also matters as to whether linguistic pragmatics is reliant on auditory or visual systems. These are firmly divided systems with potentially different advantages and evolutionary pressures.


The question of whether language evolved out of prior means of communication is redundant (p512-513)
The idea that language bears no relation to existing communication systems is highly problematic. I feel Bickerton’s intended point is that it bears no resemblance to existing call systems, but this is an entirely different thing to bearing no relation. The point of biological approaches is that everything that has evolved has a relation to everything else. Its just a matter of degrees, distance and kind. Holding language up as being unrelated to any other communicative system is dangerously close to nativist assertions that it just is, as if divinely manifested. Evolutionary biology’s comparative approach is valid insofar as it places language in the context of genes and selection. Understanding why language is so different to other animal communication systems is a valid question to help understand why language exists in the first place. I also think Bickerton is also a little unfair in implying that the majority of this research programme believes that language evolved from monkey calls. The Hauser and Fitch research programme has considerably widened and refined since then. I would agree that this question perhaps has a lower priority than other questions, but we can try to look at language as a unique adaptation, motivated by unique and isolated pressures, or as a unique adaptation in the context of comparisons with a wider evolutionary system. Neither of these positions are mutually exclusive or even contradictory.


The location of language evolution is irrelevant (p513)
I’m largely in agreement with Bickerton here. We can’t trace the location of language evolution, nor is there any evidence it might have contributed to selective pressures for its emergence. The interaction between geography and language however, is very interesting indeed, and might well reveal something about the genetic basis for language.


Simulations are interesting but tell us very little about real language evolution (p522)
Okay, if I’m criticising Bickerton for bias I’d better confess my own prejudices. I’m studying at the LEC which is an institution specifically devoted to the computational modeling of language evolution. However, I’m not a modeler myself, so perhaps not best placed to defend modeling from a detailed mechanistic point of view. However, I think most population geneticists working would take great umbrage at the basic notion that evolutionary modeling on computers is redundant. Yes, modeling is not reality, but it a means of exploring the mechanics that can reasonably be seen to exist in reality. I think simulation should be approached with caution, rather than derision, and none of the computer modelers make the claim that their evidence is definitive. They are just trying to find evidence of how different parts of the system can be demonstrated to work under mathematical conditions.

FOXP2 is not the ‘Language Gene’ and genetics is unlikely to provide any quick fixes.
Bickerton is right about FOXP2 and the evidence is piling up that it is an interesting genetic sideshow, rather than a fully functional gene for language. A recent paper has provided yet further evidence that FOXP2 manipulations are more to do with sensory-motor functions than language. The interaction between genes and language is still a long way out of reach, and the idea that any part of language, even grammatical universals, might be traced directly to genetics seems more distant than ever. However, this is not an excuse for us to ignore this important genetic dimension of language evolution. Nothing in biology makes sense except in the light of genetic selection and nothing, not even language, is an exception. Admittedly language does intuitively feel like it has a different, more ethereal component than say, an elephant’s trunk,. But just because the link between genetic selection and language is unclear shouldn’t dissuade us from looking at it. However, I would agree with Bickerton that the relationship is likely to be a lot more complex and subtle than the nativists first assumed

Mirror Neurons are not the missing neurological link that make language possible
Mirror neurons are fascinating and potentially very important. However, as Jim Hurford has powerfully argued, they fall a long way short of making language possible. In particular they fail to account for how we crossed the symbolic threshold and went on to develop syntax. I suspect they will end up being considered as mere cogs in the mechanisms of early language evolution, but at the moment the evidence isn’t firm either way.


Bickerton concludes the paper with a very apt and necessary warning: Take care” he chides, “it’s a minefield out there, strewn with explosive charges of little-known fact ready to blow up the fanciest new theory.” Wise words indeed, the only problem with this is that anyone with even a cursory knowledge of the field can see that the gatekeeping role that Bickerton has assumed implicitly promotes his own agenda. A linguist reading this paper would be wise to apply the cynicism he is encouraging to Bickerton as well.

Overall the paper reflects a classic linguist’s desire to atomise the problem and focus on understanding each little aspect in isolation. Whilst I agree totally with the logic behind this, I would suggest that there is also a need for broader exploratory and comparative approach as well. Bickerton notes that we linguists are well prepared for the problems of language evolution because we already work in a tradition that is used to unending debates and unscientific extrapolation. But whereas Bickerton sees this as an experience we should use to sharpen our cynicism, I would argue that sometimes it can be a necessary evil. We do need high standards of proof and rigorous adherence to scientific method, but dismissing all programmes before they have a chance to refine and test their ideas is counterproductive. I think we need plenty of what Howard Bloom called ‘creative bickering’, even if it does throw up the odd idea in need of a good firm slap. The defining thing about the field of language evolution is its need to think creatively about how to generate evidence. Linguists, as much as anyone, have to be prepared to walk down some unfamiliar alleyways. However this should not be considered an excuse for not keeping your wits about you.



3 Responses to “A Brief Guide to [A Brief Guide to Language Evolution for Linguists – by Derek Bickerton]”

  1. 1 jcm December 16, 2007 at 6:23 am

    Most of what you say I agree with. But ;-), I’d like to add some:

    DB skips over the “evolution of symbolic units” like it meant something. For the term ‘evolution’ to have any meaning, except metaphorically, requires a change within a biological entity that produces a second, different, entity with traits more favorable to reproduction: This second, evolved entity would be a symbolic unit? NOT! The second entities would be living *creatures with a newfound ability to *use symbolic units. Our human “symbolic competence” requires brains that can produce and manipulate neurological patterns as symbols. [Once such units exist, we also need to translate these neuro-electro-chemical representations into different formats, audio or visual or motor. To broadcast these symbols requires only a dozen or so recombinant physical forms,in some sensory mode, and virtually any animal can produce those.]

    To avoid being Zen-like, I think science needs to avoid Bickertons language-as-virus and focus on finding an observable difference between actual brains: those with the ability to use symbols and syntax, and those without. Since only humans seem able to use syntax, those brains won’t be found in other species. However there are many humans, with brains capable of both symbols and syntax, that lack language. These would be deaf persons raised without any signed language and without any social group whose dynamics enable them to create their own as was done in Nicaragua or Israel.

    DB consistently ignores these populations, probably because his theories cannot explain them. Both the Nicaraguan/Israeli kids, who did create their own languages, and the alingual persons, who did not, lacked parental input. They lacked any previous generation of language that could provide an external source of information. Also, their innate bioprogram/LAD/whatever did not provide a sufficient internal source of information for language creation either. Nevertheless, the Nicaraguan/Israeli kids did create language after interacting socially with their peers. These human populations invented language iff they experienced certain social interactions.

    Bickerton’s theories don’t explain this, but then neither does anyone else. It seems to me that by the uniformitarian principle, if language “evolved” this way in these cases, it would’ve evolved the same way in other cases at other times.

  2. 2 jcm December 16, 2007 at 6:33 am

    how do you leave a comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: