On the syntax of disjunction scope

Conclusion

The account of disjunction proposed above is an interesting one, I believe, not simply because a principled account has been given of a certain collection of data, but also because the relation between the syntactic and semantic analyses is an intuitively satisfying one. Under the account argued for here the syntactic properties of elements such as either and whether are almost entirely predictable given three components of information: (i) the meaning of disjunction as explicated by Rooth and Partee (1982); (ii) a number of general principles and conditions, and (iii) certain very simple lexical facts such as the fact that either is [−WH], while whether is [+WH]. In view of the first component we know that disjunction takes scope through binding of a free variable. The second component presumably gives us that scope is syntactically represented, that scope assignment involves movement to an Ā position, that the trace of this movement is subject to ECP, that the domain of this movement is bounded by Subjacency, &c. Finally, given either and whether as scope indicators, the third component entails that the former adjoins to S, and so marks scope within the minimal sentence containing its associated disjunction, while the latter moves to COMP, and so may mark scope in broader domains.

From the standpoint of language acquisition these represent promising results, particularly when considered with reference to recent thinking in the philosophy of mind. Suppose we assume, following Fodor (1983), that certain meanings or concepts — perhaps a large number of them — are given to the language learner as a part of universal biological endowment, and that the task of acquiring these meanings is essentially one of ‘identification’, i.e., of matching concepts with morphemes from the spoken environment. The meanings of basic logical connectives such as and, or, if are plausible candidates for membership in this universally available class. Then adopting the semantics of disjunction discussed above, the language learner will know that or takes scope, and that he or she may tacitly expect scope markers. The task then reduces essentially to identifying such markers in the stock of morphemes encountered, of picking out either and whether as the relevant items, or of postulating their null counterparts in languages where phonetically realized markers are lacking. Once this identification is made, the syntactic properties of either and whether then follow immediately, as we have seen. Correlatively, given our results concerning conjunction, the language learner will also know that and does not take scope, and hence that no such markers are to be expected or hypothesized. Thus he or she would tacitly know that both is not and could not be a scope indicator, and hence must be assigned some other syntactic status, say, quantifier phrase. If this general picture is correct, therefore, we appear to move toward a genuinely explanatory account of disjunction.