Tom Bever of University of Arizona will speak at the UCSD Linguistics Department Colloquium on April 20, 2009, at 2:00 pm in AP&M 4301.
I will discuss three examples of universals of attested languages, all of which have been treated as part of Universal Grammar, the essence of language structure. Two of the universals may have a functional cause related to language acquisition: the EPP and the mutual opacity of representational levels. A third may derive from long-evolved neurological foundation of visual species recognition: the upward movement of constituents.
The EPP isn't. EPP is a stipulated configurational filter on distant derivations, and accordingly violates the general principle that all syntactic constraints are "local" (be it in minimalist or connectionist frameworks). I show that EPP phenomena fall out of a canonical form constraint on attested languages - every learnable language has a canonical surface form and canonical mappings of that onto thematic relations, to facilitate learning with a traditional hypothesis-and-test learning model.
The functional roles of opacity and recursive operations between linguistic levels - making grammar learning a kind of problem-solving "fun". The argument builds on the above model of learning, as problem solving. Since problem solving elicits a general human motivation, language has to have the form of a problem that stimulates that motivation for the individual language learning child. This becomes a critical part of the intrinsic motivation in individuals to learn language.
....The pied piper of vision.... It has been a long standing fact and puzzle why linguistic movement has always been "upward", from more embedded parts of a linguistic structure to less embedded parts. I propose an explanation based on "coaptation" of representational principles of vision, relating to species specific recognition. In particular, I show experimentally that perception of real "upward" movement (from a more to less embedded portion of a scene) is easier than of "downward" movement. I then argue that this upward hierarchical organization of vision is related to how species recognize the motion of conspecifics. I conclude with a brief discussion of "linguistic universals of the third kind", so called "laws of form". Recent explorations of the ubiquity of the Fibonacci sequence in growth patterns at various linguistic levels, have raised the possibility that language is constrained by that mathematical "natural law". I suggest that such numerological phenomena are a result of conflicting representational constraints at each level, not a cause.