Speaker: Michael Diercks (Pomona College)
Title: Building Bridges: Developmental Minimalist Syntax
Date: Thursday 9 December
Venue: Online (get the Zoom-link through the Mailing List)
Time: 16.15 – 17.30 hrs
By many measures, the generative grammar project in its current iteration (the Minimalist Program) has been highly successful as a framework motivating and organizing syntax research on the world’s languages. Nonetheless, contemporary generative grammar remains relatively intellectually isolated, with a relatively minor impact on the cognitive sciences more generally. While there are likely many reasons for this, among the major ones is that there has not been much success linking particular components of the Minimalist syntactic framework to particular cognitive properties outside of the grammaticality judgments it is based on. In particular: why does a Minimalist derivation of a sentence build structures bottom-up when language is neither produced nor perceived in that way? Is a Minimalist derivation nothing more than an extended metaphor?
This talk sketches a framework that has the potential to be a bridge out of our isolation. Building on a long history of similar ideas, Developmental Minimalist Syntax (DMS) proposes that the reason that bottom-up derivations model adult syntax so well is that adult grammars (in part) encode the developmental pathways by a child acquires grammatical knowledge. Specifically, DMS claims that so-called ”Universal Grammar” (UG) is in fact a precise description of cognitive biases about the configuration that we mentally represent grammatical knowledge in when we acquire it. That is to say, Merge does not only build adult grammatical structures; the reason it does so is because it describes the format that occurs when children successfully arrive at a grammatical generalization in acquisition. DMS claims that the Minimalist derivation of a sentence recapitulates the child’s developmental pathways precisely because adult grammatical knowledge retains those earlier stages of knowledge. DMS therefore predicts fairly tight correspondences between timelines of acquisition and sequences of derivational steps in a Minimalist analysis.
Ideas along these lines have persisted in the field for a long time (e.g. Radford 1990, Rizzi 1994). Part of the the novel contribution of DMS is that it suggests that acquisition does not correspond specifically to structural height in adult grammar, but to sequences of the Minimalist derivation. Because structures are canonically built bottom-up, this generally does correspond to structural height. But it also opens the door to instances where acquisition is out of sync with structurally height: what this predicts is that it should correspond to an ”out-of-sync” Minimalist derivation. We suggest that such patterns do in fact occur in adult grammars: they are the counter-cyclic phenomena that the field has long considered a thorn in its side. Look-ahead effects are where something happens in a derivation before it ”should” (according to a strict bottom-up cycle): we suggest this corresponds to ”early” acquisition of a piece of grammar by children. Likewise, Late Merger inserts material into the structure after it ”should:” we suggest this corresponds to a child acquiring a piece of grammar after its surrounding structures.
Acquisition data suggest that this is plausible, though our own novel empirical work is in a fledgling state. The reason we continue to pursue the idea, though, is that it has the potential to connect the large, abstract structures (and complex derivational sequences) of Minimalist analyses with a large data set (child language) that is distinct from the grammaticality judgment data that Minimalist analyses are built on. If this idea were to hold up, it would begin to construct a bridge both for syntacticians to make direct predictions about acquisition, but also for acquisitionists to make direct claims about adult syntax. A lot of work is necessary before we get to that point, though. This talk sketches a framework based on what we know so far.