Monday, April 7, 2008

"The Pied Piper of Cambridge"?

In my last post I wrote a bit about Edmund Blair Bolles’ comment that as made evident by many contributions to evolang 08, some aspects of Chomskian linguistics are in trouble, especially if we approach questions of grammatical nativism from an evolutionary perspective. Some even go so far as to claim that Chomsky’s approach to “Language and Problems of Knowledge” (Chomsky 1987) was flawed from the start and led the field of linguistics in a fundamentally wrong direction, ultimately leading to alienation between linguistics and other fields of biologically based inquiries into human cognition.

Philip Lieberman, ‘enfant terrible’ of language evolution studies,[1] for example, calls Chomsky “The Pied Piper of Cambridge” (Lieberman 2005; Cambridge refers to Cambridge, Massachussetts, where Chomsky is professor emeritus of linguistics at the Massachusetts Institute of Technology.), because he led linguistics away from the methodological path indicated by Charles Darwin and Theodosius Dobzhansky’s (1973) famous credo that

“Nothing in biology makes sense except in the light of evolution.”

But what exactly is it that Chomsky claims nowadays and what is this “generative enterprise” anyway? I have already written a bit about Chomsky’s view on language evolution (see here, here, and here; there’s also a recent post on Babel’s Dawn discussing Chomsky’s theories of language and its evolution, which is much more detailed then mine will be).

In one of his latest papers, “On Minds and Language” (Chomsky 2007), which appeared in the inaugural issue of Biolinguistics, Chomsky basically reiterates the saltational, exapatational, single mutation claims he has been making for a long time, (for a discussion, also check out Babel’s Dawn) but first let’s look at briefly what else he talks about in his paper. According to Chomsky, the primary focus of the ‘generative enterprise’ is

To determine the structures that underlie semantic and phonetic interpretation of expressions and the principles that enter into growth and development of attainable languages” (Chomsky 2007: 10)

Chomsky general claim how to account for such the general propensities is that there is

“a rich, innate, species-specific component of the mind dedicated to language (the faculty of language [FL]).” (Boeckx & Piatelli-Palmarini 2005: 448)
and that languages can be described formally by elaborate “tree”-diagrams and abstract computational rules such as

Sentence = NP (noun phrase) + VP (verb phrase) (Chomsky 1957), a subrule being VP = Verb + NP which would map out as wikipedias’ :

As these were the humble beginnings of the generative enterprise, and by now the theory has gone through a series of transformations with such stunning titles such as Transformational generative grammar (TG; c. 1957), Standard Theory (c. 1965); Extended Standard Theory (EST; c.1970), Government/Binding Theory (GB, c. 1970) and (finally ?) the Minimalist Program (MP, post 1990) (V.J. Cook & M. Newson 2007: 4; just to put some more irrelevant information in here: V.J. stands for Vivian James and he’s a guy.).

In this final instalment of Chomskian generative grammar there are computational mechanisms with such lovely titles as s-select, c-command, assignment of theta-roles, UTAH, feature checking and many, many more. The general idea is something on the lines of having a ‘basic’ sentence structure and then deriving all other sorts of sentences via movement and merge rules.

Given that for complex sentences, a lot of intricate changes have to be made, the formal analysis of a sentence like “It was in the gallery that he met the panther” by now looks like this:

(taken from on of my lecturers on minimalist syntax, Mike Schiffmann)

This apparent complexity has sparked many criticisms, especially on the grounds of the psychological feasibility of such a theory. Some people object whether such trees are adequate description of what really happens in the brain (such as neurolinguist Friedemann Pulvermüller )

But more importantly, there are by now serious doubts if the complexity of language construction and comprehension Minimalist Program is in accordance with psycholinguistic research on how we parse and understand real-life utterances. As Fernanda Ferreira puts it

“One of the fundamental problems is that the model derives a tree starting from all the lexical items and working up to the top-most node [of the syntactic tree], which obviously is difficult to reconcile with left-to-right incremental parsing (Ferreira 2005: )

MP thus seems:

“highly unappealing from the point of view of human sentence processing” (Ferreira 2005: 370)

She also argues that in MP,

“Data from other areas such as neurolinguistics, computational linguistics, and psycholinguistics were not taken into account at all, nor were any insights from the rest of the cognitive sciences.”

These apparent problems with the Minimalist Program and generative approaches in general have sparked a lot of competing theories of grammatical description, which break with a lot of the foundational ideas of generative grammar, but follow the Chomskyan approaches in having lovely, opaque acronyms as their titles such as: CG (Construction Grammar), HPSG (Head-Driven Phrase Structure Grammar), OT (Optimality Theory), L-TAG (Lexical Tree-Adjoining Grammar), LFG (Lexical Funtional Grammar), APG(dunno), etc.

Another area in which Chomsky’s position has been criticized is his assertion that grammatical competence is innate, which of course is directly related to his views on the evolution of language. I’ll write a bit about that in my next post.



References:

Boeckx, Cedric, and Massimo Piattelli-Palmerini (2005). Language as a natural object – linguistics as a natural science. The Linguistic Review 22: 447–466.

Chomsky, Noam 1957. Syntactic Structures. Berlin; New York: De Gruyter.

Chomsky, Noam 1987 . Language and Problems of Knowledge. The Managua Lectures. Cambridge, Mass. / London, England: MIT Press (Current Studies in Linguistics Series 16)

Chomsky, Noam (2007): Of Minds and Language. In: Biolinguistics 1, 9-27.

Cook, V.J.m abd Mark Newson. (2007) Chomsky’s Universal Grammar: An Introduction. 3rd ed. Blackwell

Dobzhansky, Theodosius (1973). Nothing in biology makes sense except in the light of evolution. American Biology Teacher 35: 125–129.

Ferreira, Fernanda (2005): Psycholinguistics, formal grammars, and cognitive science. The Linguistic Review 22: 365–380.

Lieberman, Phillip (2005): The pied piper of Cambridge. The Linguistic Review 22: 289–302.


[1] Although, given further consideration, the title would probably be much more apt for Derek Bickerton

2 comments:

Anonymous said...

Your main objection seems to be that the technical terms are unknown to you and the esthetics of the diagrams don't appeal to you. That's not an argument.

Michael Pleyer said...

Please note that "I" didn't intend to state any objection to the generativist approach. I was merely pointing out that there are people who are making such arguments. Also, I didn't mean to say that the aesthetics of tree diagrams don't appeal to em (I could make the same argument about family trees or the evolutionary "tree" and the argument would be just as vacuous). I was merely stating that there are doubts on whether they are adequate in terms of characterising the actual cognitive processes involved in parsing and producing utterances, without taking anyones side (I hope).
I do know about (to be fair,I don't claim to have fully or even partly grasped them) techncial terms such as the Uniformity of Theta-Roles Assignment, Spellout, Phi-features, etc. I wasn't implying that there is anything wrong with them because I'm unfamiliar with them (that would indeed be prepostorous)but that there are people who turned away from generativist approaches because these concepts are hard to grasp intuitively (which is much more the case in ,say, Cogntiive Linguistics, which on the other hand doesn't have formalisms as consitent and elaborate as generativism), and at first glance seem rather obscure and abstract.
I have my problems with some versions of MP, because I feel (in the approaches I know of, that is to say, e.g. Adger 2003) that the Conceptual-Intentional System, as well as the Sensori-Motor/Articulatory-Phonological System are underspecifiied, that they are inconsistent in making transparent their choices for the formalisms of category and verbal features, in first rejecting projection hierarchies because they aren't minimal and then reintroducing them, its problems in defining the phenomenon of adjunction, and inconsistencies with the use of little v.
But this doesn't lead me to say that these problems couldn't be worked out and that generativist approaches are flawed in general because, honestly, I don't really know enough about it to make such an argument.

I hope to have clarified these points
kind regards,

Michael

References:

Adger, David. 2003. Core Syntax: A Minimalist Approach. Oxford: OUP.