Research

(Note: Some publications are relevant to multiple categories, and are thus listed multiple times.)

Semantics of attitudes

In linguistics and philosophy, the term attitude is often used ambiguously to refer both to cognitive states like desire, belief, and regret, and to words and constructions that in some sense "talk about" such states, such as want, believe, and regret. In my dissertation, as well as related published work in Linguistics and Philosophy and the proceedings of Semantics and Linguistic Theory (SALT), I have argued that in the natural language ontology—the model of the world used to interpret linguistic meaning—states of desire, belief, etc. have non-trivial part-whole structures, with these part-whole relations having significant semantic repercussions (Pasternak 2018a,b, 2019a). These arguments have been based on two types of evidence: constructions measuring the intensity of attitude states (e.g., There is a lot of desire on Maria's part for a change in leadership), and cases in which a group of people are said to believe something that no one person in that group believes (e.g., Sam's six clients think she built six houses for them). In addition, Kai von Fintel and I have a manuscript in revision for Linguistics and Philosophy exploring the semantics of attitude conditionals like If Laura becomes a zombie, she wants you to shoot her (von Fintel & Pasternak in revision).

Relevant publications and manuscripts:


Grammar and processing of quantification and measurement

Languages have a variety of ways of quantifying over and measuring objects, including quantificational noun phrases like every student and measurement structures like three feet of rope. In addition to theoretical work on the grammar of quantification (Pasternak 2020) and measurement (Pasternak 2016, 2018a, 2019a,b, 2022; Pasternak & Sauerland to appear), I am also engaged in computational research on the processing of ambiguous quantificational structures (Pasternak & Graf 2021).

Relevant publications and manuscripts:


Pragmatics of co-linguistic content

People gesture constantly when they speak, and often such gestures are used to communicate truth-conditional information. For example, if I say The girl will use the stairs and point upward while saying use the stairs, it typically conveys that the girl will take the stairs upward. Lyn Tieu, Philippe Schlenker, Emmanuel Chemla and I have experimentally tested some peculiar pragmatic properties of gestural inferences and have found evidence that they behave like another type of inference found in spoken language, namely presuppositions (Tieu et al. 2017, 2018). Out of a desire to see how specific these inference patterns are to gestures, I have suggested in a short manuscript (Pasternak 2019c) that co-speech sound effects—meaningful sound effects aligned with recorded speech—could serve as a useful test case, as these lack many of the traits that make gestures special. With this in mind, Lyn Tieu and I have experimentally tested inferences from co-speech sound effects, as well as co-text emoji, and found evidence that these give rise to the same inference pattern as gestures, suggesting that the inference pattern of co-speech gestures does not arise because of some special trait of gestures, but rather because of their more general status as co-linguistic content (Pasternak & Tieu 2022). Finally, in a 2021 manuscript I have argued that the meanings of gestures, sound effects, emoji, etc. are integrated with the meanings of spoken content not at the level of the grammar, but rather at the level of the human parser, which I refer to as the parsing hypothesis (Pasternak 2021). I also provide an explicitly defined formal parser illustrating what an analysis in accordance with the parsing hypothesis might look like.

Relevant publications and manuscripts: