New Papers by Polly Jacobson

Polly Jacobson has posted some new papers on her website. Let me highlight two of them:

  • “I can’t seem to figure this out”, to appear in B. Birner and G. Ward (eds.), Drawing the Boundaries of Meaning: Studies in Neo-Gricean Pragmatics and Semantics in Honor of Laurence R. Horn, J. Benjamins.

This paper is a study of the “can’t seem to” construction (as in “I (just) can’t seem to figure this out”). The challenge posed by this construction is that the semantic scope of “can”, “not”, and “seem” appear to be the opposite of their surface scope: the above can be paraphrased as “I seem to be unable to figure this out” (thus surface “can-not-seem” corresponds to semantic “seem-not-can”). We can note immediately that the fact that “not” has semantic scope over “can” in this construction is a general (well-known) fact about “can’t”, but the puzzle remains: why is it that in this construction, “can” and “seem” appear to reverse scope? The seminal work on this is Langendoen (1970), who argued for a syntactic scope reversal rule, allowing semantic “seem unable” to surface as “unable (can’t) seem”. Another solution treats “can seem” as a single idiom, and a Negative Polarity item.

This paper argues against both the scope reversal analysis and the diom analysis and proposes, instead, that the construction does have a purely compositional interpretation. A scope reversal solution is suspicious, for the requisite mechanisms would be restricted to happening only under not and would be restricted in some ad hoc way to apply to just these lexical items. Additiional evidence against scope reversal (and also the idiom analysis) comes from the interaction of this construction with VP Ellipsis. In the end, then, I suggest that scope reversal is an illusion: the appearance of scope reversal is an artefact of the existence of available reasonable paraphrases. I claim that “can” here does not actually mean “able”, but is rather the “can” which quantifies over occasions (the fact that it takes on a suggestion of “ability” in certain cases follows from more general facts). Moreover, “seem” here does not have the same meaning as the homophonous verb that occurs in the construction “seem to X”, but is rather just an assertional hedge. Once one carefully analyzes the meanings of these two words - meanings which occur independently in other contexts - the scope reversal illusion disappears, and the construction yields to a compositional analysis.

Since at least as early as Sag (1976), Antecdent Contained Deletion (ACD) has provided a classic argument for the existence of LF - i.e., for the claim that quantified DPs in object position are pulled out of their surface position at the level of representation at which interpretation takes place. As such, it appears to provide strong evidence against a strong direct compositional view of grammar. This paper argues first that the textbok wisdom is incorrect - in fact it is entirely based on the fallacious assumption that all functions must be “saturated” in order for semantic composition to proceed. Once we rethink this assumption (an assumption which is neither motivated nor is really consistent with other things which are standardly assumed), then we open the door for an analysis of ACD as “Transitive Verb Phrase” ellipsis. This idea was first put forth in Cormach (1984); the paper here explores a number of consequences of this idea, and shows how it can handle a variety of more complex facts surrounding ACD. I also give a full analysis of the VP and Transitive verb phrase ellipsis within a categorial grammar/direct compositional architecture. Finally, I argue that ACD is not only compatible with the direct compositional (and variable-free) architecture, but that there is an interesting interaction with Pied-Piping which is straightforward under the account here but in the standard view.

Post a Comment | Trackbacks Closed