Perspectives in discourse: The case of the motion verb come*

It has long been assumed that using the deictic motion verb come to describe a motion event involves taking the perspective of an individual located at the destination of the motion event (Fillmore 1966). I argue that this perspective-taking is built into the meaning of come, drawing on novel data concerning what I call the “anchoring implication” of come. This is the implication that a particular individual, the “anchor”, is located at the destination. I develop three empirical generalizations about the anchoring implication: acceptable anchors are limited to individuals with salient, relevant perspectives, the anchor argument satisfies classic diagnostics for anaphorically interpreted implicit arguments, and the anchoring implication is necessarily a de se commitment of the anchor. To account for these observations, I supplement the dynamic semantics of AnderBois et al. (2015) with representations of the perspectives of particular individuals, defined in terms of Stalnaker’s (2008) theory of belief and Roberts’s (2014) theory of doxastic perspectives. The anchoring implication of come is then modeled as an update to the anchor’s perspective. The resulting analysis is empirically superior to previous accounts and provides a lexical semantic justification for the intuition that using come involves perspective-taking. The paper concludes with a sketch of how the analysis can be generalized to account for another arguably perspectival phenomenon: non-speaker-oriented appositives.