Northeastern University
I discuss "semantic integration", the idea that elements of a message vary in the degree to which they are linked at the message level in language production. I show how this difference in conceptual representation has consequences for grammatical encoding, the process which translates an abstract message into an ordered and inflected sequence of words. In particular, I argue that greater integration leads to increased overlap in planning between message elements, so that, for example, tightly integrated lexical items are more likely to be active simultaneously in the lexicon. Evidence for this view comes from examining error rates during the production of subject-verb agreement, and from examining exchange error rates and speech onset times during the production of simple noun phrase - prepositional phrase sequences.
Reception to follow in 1413 Marie Mount Hall.