| Linguists have analyzed agreement from different perspectives. Some have argued that it is a syntactic phenomenon, others that it is a semantic phenomenon, and still others that it is a discourse phenomenon. While each perspective is correct to some degree, no single perspective is sufficient. Agreement cannot be adequately described in any single realm. Rather, it is determined by many factors: syntactic, semantic, discourse, referential, and pragmatic.; In most cases, the factors tend to align in producing the same agreement forms. For this reason, one analytic method—say syntax, semantics, or discourse—can account for a vast amount of data. To see the role of each factor, we need to look at cases where there are mismatches. For this reason, linguists have been paying more and more attention to agreement with coordinative subjects.; Although linguists have progressed in understanding coordinative agreement, they have mostly concentrated on conjunction, assuming their analyses apply equally to disjunction. We can obtain a better picture by looking at disjunction as well, since disjunction commonly leads to modular misalignments. This dissertation marks the first extensive study of agreement with disjunctive subjects.; In order to analyze the grammatical structures of disjunction, I use an Autolexical framework, which allows for independent semantic and syntactic representations of grammatical structure. Such independent structures make it possible to isolate the factors involved, a crucial step in analyzing agreement data. I supplement this framework with a discourse level that is mapped off of semantic structure and with a set of weighted agreement constraints.; Among the constraints I posit are syntactic, semantic, discourse, and referential constraints. Individual speakers differ in how they assign relative weight to the constraints, and this difference leads to variation, which is widespread with agreement (especially disjunctive agreement). For example, some speakers place more weight on semantic factors than others do. Indeed, some speakers even make use of contextual cues in determining agreement. This last fact may force us to reevaluate pragmatics' role in grammar. Traditionally pragmatics is viewed as outside grammar; however, I argue that it permeates grammar. |
