Foundations of Language: Brain, Meaning, Grammar, Evolution

(ff) #1

regular rule suddenly“popping into place,”more careful micro-observation often shows a gradual emergence of
constructions, appearingin use withoneverbata time(Tomasello2000a, withwhom I agreethatchildrendo nothave
full adult grammars merely constrained by performance; see also Da ̧browska 2000). In addition, memorization of
higher-frequency regulars is to be expected, as nothing in principle prohibits it.


I hardly clai mto have provided a thorough account of language acquisition here. Rather, I have tried to show that the
parallel architecture leads to an attractive formal view of the lexicon and of rules of language. In turn, this view
reframes the acquisition problem, bringing traditional questions of linguistically oriented acquisition research closer to
theneuroscientific approach tolearning. Thegap is stillnotyetbridged, but perhaps itis nowat least possible tosee to
the other side.


6.10 Universal Grammar as a set of attractors


We are now in a position to make an interesting conjecture about the nature of Universal Grammar and the way it
determines particular grammars.


I have alluded a few times to a position taken by both Principles and Parameters Theory (P&P, Chomsky 1981) and
Optimality Theory (OT, Prince and Smolensky 1993): Universal Grammar explicitly provides all the grammatical
possibilities for thelanguages oftheworld.InP&P, acquisitionofa grammar involvessettingafinitenumberofinnate
parameters; in OT, itinvolves ranking afiniteset of innateconstraints. Thus learning a grammar is conceivedof rather
like customizing a software package: everything is there, and the learner has only to set the options to suit the
environment.


Such an approach seems improbable on grounds both external and internal to linguistic theory. The external grounds
are that Universal Grammar is required to be extremely elaborate and veryfinely tuned. This lies behind the critiques
ofUniversalGrammar offered byElman etal. (1996), Tomasello(1995), and Deacon(1997) discussed inChapter4. It
is hard to imagine all this structure emerging in the brain prior to experience, much less being coded genetically. As
mentioned, Chomsky's response (e.g. 1995) has basically been that language simply is unlike any other biological
capacity.


Buttheinternalgroundsagainstsuchanapproachareevenmorecompelling. As Newmeyer(1998b),Culicover(1999),
and Ackerman and Webelhuth (1999) point out, few of the actual parameters determining differences among
languages have been successfullyworkedout, despitenearly twentyyears of intensiveresearchin theP&P tradition.At
the same time, this research has uncovered many phenomena across languages whose differentiation from each other
would require parameters of such niggling specificity that they are hardly plausible as universal possibilities. Culicover
and Ackerman and Webelhuth


190 ARCHITECTURAL FOUNDATIONS

Free download pdf