Desktop version

Home arrow History arrow A Minimalist View on the Syntax–Semantics Relationship: Turning the Mind into a Snowflake

Source

Foreword

According to one kind of analysis of de re modal talk, there may possibly be a world in which the White Rabbit is my counterpart. In that world, I would find it perfectly easy to follow King’s order: ‘Begin at the beginning' said the King to the White Rabbit, ‘and go on till you come to the end: then stop' Alas, that world is either spatiotemporally entirely disconnected from the actual world (if you are a modal realist), or it is merely a story (if you find ersatzism more appealing). In either case, there being a White Rabbit counterpart of me offers no help with finding an appropriate beginning for a minimalist inquiry into intricate relationships between narrow syntax and interpretive components of the C-I system in the actual world, the task being only harder if semantics is to syntax as the hole of the doughnut is to the whole of the doughnut—as one may, reversing the dictum of Sige-Yuki Kuroda (see Kuroda (1979: vii)), conceive of their relationship in accordance with current minimalist views in mind. For the relationship has become more intricate on the one hand and tight on the other than it has ever been. At the outset of a story of interactions between logic and linguistics since the beginning of the 20th century, Lenci and Sandu (2009) observe:

On one hand, the rationalist turn in linguistics has actually allowed for an unprecedented convergence of linguistics with important areas of mathematical logic. On the other hand, the generative paradigm has also set constraints on the study of natural language and formulated hypotheses on its architecture that have often dramatically conflicted with the logicomathematical approach. Thus, the history of the relationship between logic and theoretical linguistics in the past decades is rather a deeply dialectic one. It is a history of profound and synergic efforts toward the common aim of understanding the nature and universal principles of human language and its formal structure, but it is also an history of harsh conflicts and divergences on the nature of universal grammar itself.

At the core of this confrontation lies the issue of the relationship between grammatical form and logical form, that is to say, the possibility itself of carving out the natural language syntactic and semantic space as a logical space. (Lenci and Sandu 2009: 776)

The survey of ‘profound and synergic efforts’ and ‘harsh conflicts and divergences’ which Lenci and Sandu (2009) provide ends, basically, with the end of the GB period. The minimalist enterprise, while continuously stressing the importance of the relationship of narrow syntax with non-syntactic components which receive the output generated in the former and the priority of the C-I compo- nent—either in terms of ‘output conditions’ which narrow syntax should satisfy for its output to be ‘legible’ at the interfaces, or by bringing syntax and the C-I component closer by taking a syntactic process, viz. labeling, to be required for interpretive purposes—makes a rapprochement between formal semantics and syntactic theory even more difficult. The theoretical framework of minimalist syntax presents a desert landscape, with a constant effort to eliminate concepts and mechanisms not grounded in general computational principles as obstacles to explaining properties of the human language faculty within limits set by problems of language acquisition and language evolution—far away from the abundance of primitive concepts and a cornucopia of mechanisms offered by formal semantics. The divergence may invite pessimism about current possibilities of both enterprises coming together:

If you look at, for example, formal semantics where an awful lot of work is done, there’s no principles-and-parameters. It’s assumed to be uniform and invariant. Maybe it isn’t, but it’s assumed to be invariant. So it’s just like a straight, one-to-one, instantaneous mapping from syntactic structures to some kind of interpretation. And then comes the question: What are the details of the mapping? If you look at it, it’s kind of like phonology; it violates every imaginable computational principle. It’s just not a question for people; it’s not the kind of question they care about. They want to get it descriptively correct, get the scope relations correct, that sort of thing. I don’t want to exaggerate. There are some deep and very suggestive ideas. But I think it is fair to say that the main problems, and they are huge and challenging, tend to be descriptive at this stage of understanding. (Chomsky 2004b: 186-187)

While minimalist syntax went well ‘beyond explanatory adequacy’ the semantic theory remains so far mostly at the level of descriptive adequacy, the difference in the richness of their theoretical repertory reflecting a difference not of the matter, but of the degree of theoretical development. Moving towards the goal of ‘going beyond descriptive adequacy’ in the theory of the C-I component as it is directly interfacing with narrow syntax has a point of reference: the theory of narrow syntax as currently envisaged. Although it still does not provide a distinguished starting point, there are then far fewer potential points to begin the journey from narrow syntax to the C-I component than there are on the anchor ring. Given the role that the minimalist theory assigns to the emergence of narrow syntax and its sole structure building operation—‘the operation Merge—an operation that takes human concepts as computational atoms and yields structured expressions that, systematically interpreted by the conceptual system, provide a rich language of thought’ (Berwick and Chomsky 2016: 87)—it is in effect a search for sources and underlying mechanisms of the uniqueness of human thought. Insofar as an explanatorily fruitful harmonization between syntactic theory and semantics may be achieved, and insofar as the hypothesis that syntactic operations might be computationally perfect, or close to it, hence the result of physical laws independent of humans’ (Berwick and Chomsky 2016: 87) proves correct, the part of the C-I component responsible for handling objects delivered from narrow syntax will turn to be ‘more like a snowflake than like giraffe’s neck’ (see recently Moro (2016a,b) for comments on the appropriateness of the analogy). It would be a little suprising for a ‘language of thought’ and the locus of humaniqueness. ‘After all, there is no reason in a snowflake,’ as Moro (2016b: 37) remarks.

 
Source
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >