In this talk, we propose an account of this split based on independently observable properties of agreement in YM together with the Movement Theory of Control (Hornstein 1999, Hornstein and Polinsky 2010 inter alia). First, we develop a clausal syntax for a variety of YM clauses in which absolutive arguments, including intransitive subjunctive subjects, remain low in the clause. Second, we show that this independently motivated syntax together with a particular approach to control predicts the ungrammaticality of intransitive subjunctive control complements. Finally, we argue that the attested bare forms are in fact nominalizations and therefore have a quite different syntax than the transitives.
While this expectation is consistently borne out for most evidentials, we show that reportative evidentials – i.e. those which indicate that the speaker’s source is what some second or third party has told them – consistently do allow for exactly this. Whereas previous authors have proposed semantic accounts for such data, we argue that these exceptional cases are due to pragmatic perspective-shift. Such shifts are only readily possible in the case of reportatives since they introduce another perspectival agent, whereas other evidentials (even including intuitively ‘weaker’ ones like conjecturals) do not. Beyond explaining the cross-linguistic behavior of reportatives, I argue the proposal also makes correct predictions for languages like Bulgarian where a single evidential form has both reportative and inferential uses.
Here, I’ll present the first Bayesian model which can simultaneously learn word meanings and perform pragmatic inference. In addition to capturing standard phenomena in both of these literatures, it gives insight into how the literal meaning of words like “some” can be acquired from observations of pragmatically strengthened uses, and provides a theory of how novel, task-appropriate linguistic conventions arise and persist within a single dialogue, such as occurs in the well-known phenomenon of lexical alignment. Over longer time scales such effects should accumulate to produce language change; however, unlike traditional iterated learning models, our simulated agents do not converge on a sample from their prior, but instead show an emergent bias towards belief in more useful lexicons. Our model also makes the interesting prediction that different classes of implicature should be differentially likely to conventionalize over time. Finally, I’ll argue that the mathematical “trick” needed to convince word learning and pragmatics to work together in the same model is in fact capturing a real truth about the psychological mechanisms needed to support human culture, and, more speculatively, suggest that it may point the way towards a general mechanism for reconciling qualitative, externalist theories of social interaction with quantitative, internalist models of low-level perception and action, while preserving the key claims of both approaches.