Congratulations to Ang Li, who successfully defended her dissertation on Oct. 28!
Find the details of her dissertation below:
Title: Comparing Alternatives
Committee: Simon Charlow, Troy Messick, Yimei Xiang, Rajesh Bhatt (external member, UMass)
I present a new analysis of the meaning of comparatives. In the past few decades, theorists who approach gradation in natural language using degrees have typically analyzed comparatives as imposing an ordering relation between two degrees. I argue that a comparative is fundamentally about comparing two structurally derived alternatives, as opposed to two degrees, on a given measurement dimension.
The motivation of this new analysis comes from the context dependency of comparatives without an overt standard phrase/clause. While the traditional approach predicts this dependency is no more than degree anaphora, I show evidence that these incomplete comparatives are actually sensitive to a larger linguistic context that contains the intended antecedent degree. I argue that the empirical pattern automatically follows if there is an additional requirement, namely that the antecedent degree is the measurement of the alternative we make comparison to.
This alternative-based meaning turns out to be the key unlocking a unified account of incomplete comparatives that brings together a class of phenomena, which have led to wildly different comparative meanings in existing proposals. In addition to the infelicity condition in discourse anaphoric comparatives, the alternative-based meaning can also provide a compositional account of the recurrent ambiguities between comparison, additivity, and continuation across languages, as well as the so-called internal reading.
For explicit comparative constructions, the alternative-based new analysis challenges a long-held assumption in this literature, i.e. that an explicit comparative is like a transitive construction, in that the comparative marker takes the overt standard clause (e.g. the than-clause) as its internal argument. In my theory, this transitive relation is replaced by anaphoric binding between the comparative marker and semantic objects introduced in the standard clause. I show that by treating explicit comparatives as intrasentential anaphora, the theory not only can still handle all kinds of comparative constructions, it even offers new and better solutions to some long-standing problems in comparative semantics.
Throughout this dissertation, I aim to show that the new analysis provides a better empirical coverage for each of the empirical phenomena under discussion. Above all that is an argument from uniformity: it is through thinking of comparative as comparing alternatives that we can start to understand how these seemingly unconnected phenomena all stem from a common core, and how the differences emerge from the ways we find the right alternative to compare.