Modeling the Effect of Cross-Language Ambiguity on Human Syntax Acquisition

A computational framework is presented which is used to model the process by which human language learners acquire the syntactic component of their native language. The focus is feasibility is acquisition possible within a reasonable amount of time and/or with a reasonable amount of work? The approach abstracts away from specific linguistic descriptions in order to make a 'broad-stroke' prediction of an acquisition model's behavior by formalizing factors that contribute to cross-linguistic ambiguity. Discussion centers around an application to Fodor's Structural Trigger's Learner (STL) (1998) 1 and concludes with the proposal that successful computational modeling requires a parallel psycholinguistic investigation of the distribution of ambiguity across the domain of human languages. 1 Principles and P a r a m e t e r s Chomsky (1981) (and elsewhere) has proposed that all natural languages share the same innate universal principles (Universal Grammar UG) and differ only with respect to the settings of a finite number of parameters. The syntactic component of a grammar in the principles and parameters (henceforth P&P) framework, is simply a collection of parameter values one value per parameter. (Standardly, two values are available per parameter.) The set of human 1The STL is an acquisition model in the principles and parameters paradigm. The results presented here are not intended to forward an argument for or against the model, or for tha t matter , for or against the principles and parameters paradigm. Rather, the results axe presented to point out (the possibly not-soearthshattering observation) tha t the acquisition mechanism c a n be extremely sensitive to ambiguity. grammars is the set of all possible combinations of parameter values (and lexicon). The P&P framework was motivated to a large degree by psycholinguistic data demonstrating the extreme efficiency of human language acquisition. Children acquire the grammar of their native language at an early age generally accepted to be in the neighborhood of five years old. In the P&P framework, even if the linguistic theory delineates over a billion possible grammars, a learner need only determine the correct 30 values that correspond to the grammar that generates the sentences of the target language. 2 given that a successful syntactic theory must provide for an efficient acquisition mechanism, and since, prima facie, parameter values seem transparently learnable, it is not surprising that parameters have been incorporated into current generative syntactic theories. However, the exact process of parameter setting has been studied only recently (e.g. Clark (1992), Gibson and Wexler (1994), Yang (1999), Briscoe (2000), among others) and although it has proved linguistically fruitful to construct parametric analyses, it turns out to be surprisingly difficult to construct a workable model of parameter-value acquisition. 1.1 P a r a m e t r i c A m b i g u i t y A sentence is parametrically ambiguous if it is licensed by two or more distinct combinations of parameter values. Ambiguity is a natural enemy of efficient language acquisition. The problem is that, due to ambiguity, there does not exist a one-to-one correspondence between the linear 'word-order' surface strings of the input sample and the correct parameter values that generate the target language. Clearly, if ev230 binary parameters entails approximately a billion grammars (230 = 1,073,741,824.)