Spring 2026
Title: What's Important in Discourse?
Abstract: Discourse is not a shopping list of sentences: some utterances, participants and sections in text and conversation are more important or salient than others, though measuring this can be challenging. In this talk I explore differences in the importance of content using a newly developed methodology leveraging multiple summarization, in which information captured in more summaries is considered more salient than the less ‘summary-worthy’ information that does not make the cut. Multiple analyses of the linguistic means that signal salience at the discourse level show considerable variation across text types, revealing that how we express pertinent versus supporting information varies broadly between fiction, academic writing, spontaneous conversation or YouTube videos. To investigate these effects, I propose an adversarial genre analysis using models trained to fit one genre and tested on data with perturbed inputs, which shows for example that properties flagging a character as important in a biography could actually correspond to a tangential one in a Reddit forum discussion, and vice versa. I will also present some recent results on the sensitivity of both humans and LLMs to the memorability of salient information, and how human and model-generated summaries compare and diverge.
Title: New horizons in evaluating pragmatic competence in language models
Abstract: For the first time in history, artificial models are using language like and with humans, sparking interest in whether LMs have learned the pragmatics of natural language. A dominant approach to evaluating pragmatics involves benchmarking how LMs capture standard phenomena such as implicature and figurative language. In this talk, I will explore approaches beyond this paradigm, using cognitive theories to interpret LM behaviors and highlight overlooked aspects of pragmatic competence in LMs. First, I will investigate “micro pragmatics” in LMs, showing that basic aspects of language (like when to use “the” and “that”) pose a pragmatic challenge even for internet-scale, Transformer-based LMs. Second, I will compare how humans and LMs make pragmatic inferences that are not linguistically mandated (“elicitures”) and arise through world modeling, in contrast to Gricean implicatures. Finally, on a methodological note, I will show how cognitive models can be used to interpret value tradeoffs in LMs’ utterance choices. These case studies highlight the relationship between pragmatics and world-building mechanisms, and suggest a blurry distinction between formal and functional linguistic competence.
Title: Questions that are merely posed, not asked
(Joint work with Sabine Iatridou)
Abstract: A number of languages have ways of marking a question as one that is merely posed and that the addressee is not obligated to answer. Examples include Greek "araye", Turkish "acaba", Japanese "naa", German verb final questions with "wohl", and many others. These merely posed questions raise important issues for the semantics and pragmatics of questions and for the general picture of how sentences are being used in context to perform speech acts. In this talk, we explore the phenomenon and its theoretical consequences.
Title: Pragmatic Indecision (or "When meaning is like voting")
Abstract: Vague predicates are obligatorily tolerant. For example, the positive form gradable adjective ‘long’ cannot be used to draw a sharp distinction, even when the facts of the context of utterance and the semantic properties of the sentence in which the adjective occurs otherwise conspire to make such uses possible. In this talk, I will do three things. First, I will argue that no existing account of vagueness — neither epistemic nor expresssivist nor contextual nor semantic nor cognitive nor distributional — provides a satisfactory explanation for why tolerance persists in such cases. Second, I will show that state-of-the-art attempts to eliminate the challenge presented by these kinds of examples by appeal to a decompositional semantics for positive form gradable predicates fail, because they do not generalize to other kinds of examples that manifest exactly the same behavior but cannot plausibly be analyzed in the same way. And finally, I will outline an alternative, pragmatic account of tolerance, in which application of a vague predicate is an expression of a preference or “vote” for a particular resolution of linguistic indeterminacy, and tolerance emerges from a model of the context as a socially optimal aggregation of such preferences.