Geoffrey Pullum and the Argument from Poverty of Stimulus

The room was packed to the rafters; a basement lecture theatre, as sweaty as the armpit of Satan and woefully inadequate for the nature of the event it was hosting. It was stuffed with what seemed like half the academic community of Edinburgh, all eager to see the new Chair of General Linguistics, the renowned Geoffrey Pullum, deliver a stinging rebuke to linguistic nativism. Pullum, a tall, pleasant man with an implausibly high waistband and the look of a high school physics teacher, was in a playfully combative mood. The talk was remarkably balanced, whilst serving quite powerfully to rally the troops. Pullum restated his and Barbara Scholz’s position on linguistic nativism and systematically dismantled the attacks against them from the Chomskian camp.

For those not from a linguistics background, the argument from poverty of stimulus is the idea that some of children’s linguistic abilities must be innate because children are not exposed to sufficient stimuli during ontogeny to account for the remarkable effectiveness with which they learn language.

Pullum’s main argument was that this position was ill defined and accepted as truth despite never having properly been submitted to empirical assessment. He went through the four attempted empirical assessments that had been identified in the literature and systematically raised serious doubts about their methodology. He then went on to argue that if linguists were to actually look carefully for examples of these rarities, rather than looking for exceptions to learnable grammar , they will probably find plenty of examples.

Pullum’s biggest obstacle was tackling what he called Chomsky sentences, a special kind of two way grammatical construction that Pullum outlines here. Chomsky says that these sentences don’t regularly exist and therefore constitute the strongest argument yet for linguistic nativism, yet Pullum has clearly found many such examples in corpora of everyday speech. Pullum was keen to stress that he doesn’t claim to have disproved linguistic nativism, just significantly muddied the waters. What was perhaps absent from the talk was some discussion of what would constitute a sufficient stimulus to rule out the argument from poverty of stimulus for a given grammatical rarity. I wanted to ask him how he would go about quantifying this difficult problem but frankly I was too wimpy to put my hand up.

The main thing I took away from it was a very important philosophical point about the division between nativism and linguistic nativism. Pullum described himself as a nativist, a believer in some innate facility for language, but reminded us all that this was different to being a linguistic nativist. Clearly we are imbued with some sort of biological capacity to adapt to linguistic environments but this by no means forces you to assume that this innate capacity must extend to specific grammatical universals. It is entirely possible to be nativist without being Chomskian. I will try to enforce this distinction with clarity in future entries on this blog.

Overall the talk was a minor triumph and Pullum was convincing, academically broadminded and intellectually robust. I have the feeling that if I had stood on a chair at the end of the lecture and screamed out “To the language lab! Let’s check the corpora for Chomsky sentences!” everyone would have followed in a Simpsons-esque mob. He certainly had me convinced that the future exploration of language evolution will have to focus on a broader definition of nativism, one that looks properly at the interaction between social learning and and any innate facility we have to deal with the linguistic environment into which children are born.

Advertisements

2 Responses to “Geoffrey Pullum and the Argument from Poverty of Stimulus”


  1. 1 jordi January 24, 2008 at 9:38 am

    I recommend you a book from Jeff Ellman called Rethinking inateness

    Basically. Pullum is right, the babies and children receive an enormous amount of data. Some kind of selective processing toward faces, language, etc. May be prebuilt.
    But, complex systems and sure the brain is one of the most complex, can extrapolate further than some given examples. It can extract new grammar and more complex from previously seen grammar with no explicit rule.

  2. 2 languageevolution January 27, 2008 at 10:34 pm

    I agree completely. I’ve just discovered a lot of the connectionist work on language and what is so stiking about the models is the extent to which similar effects can be produced by such amazingly simple networks. I hope to post a few entries once I’ve read up a lot more about them.
    I also think this distinction between innateness and linguistic innateness is going to become increasingly important for linguists in years to come. There seems no tenable position from which you can argue that humans aren’t naturally built for language, but this is not the same as possessing innate linguistic knowledge. Language is also a complex adaptive system, and language is built for our cognitive equipment by the processes of natural selection.


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: