Probability theory is the “logic of science” (Jaynes, 2003) and Bayesian data analysis (BDA) is the glue that brings that logic to data. BDA is a general, flexible alternative to standard statis- tical approaches (e.g., NHST) that provides the scientist with clarity and ease to address their personal scientific questions. Doing BDA in a probabilistic programming language (PPL) af- fords several additional advantages: a compositional approach to writing models, separation of model specification from al- gorithmic implementation (a la lm() in R), and continuity from articulating data analytic models to Bayesian cognitive mod- els. Furthermore, specifying one’s model and data analysis in a PPL allows you to search for “optimal experiments” for free. This tutorial will walk the participant through the basics of BDA to state-of-the-art applications, using an interactive on- line web-book and tools for integrating BDA into their existing workflow.

Language provides multiple ways of conveying the opposite:A person not happy can be unhappy, sad, or perhaps neither,just not happy. Rather than being redundant, we hypothesizethat uncertainty about the meaning of negation markers allowslisteners to derive fine-grained distinctions among these vari-ous alternatives. We formalize this hypothesis in a probabilis-tic model of gradable adjectives (e.g., happy), and use this toaddress an outstanding puzzle: how to interpret double nega-tions (e.g., not unhappy). Our model makes surprising addi-tional predictions about a putative difference between morpho-logical antonyms (unhappy) and negated positives (not happy):Listeners should judge unhappy as more sad than not happyonly when confronted with alternatives in context; when inter-preted in isolation, we predict no difference in understanding.Two behavioral experiments confirm consistent orderings ofinterpretations that interact with the presentational context inthe way predicted. These findings support the hypothesis thatlisteners represent uncertainty even about the most logical ele-ments of language.

Generic statements convey generalizations about categories,but how generic predications combine is unclear. “Elephantslive in Africa and Asia” does not mean that individual ele-phants live on both continents. In addition, such conjunc-tive generics pose interesting questions for theories of incre-mental processing because the meaning of the sentence canchange part-way through: “Elephants live in Africa” would im-ply most or all do, but “Africa and Asia” implies some live ineach. We extend a recently proposed computational model ofgeneric language understanding with an incremental process-ing mechanism that can begin to interpret an utterance beforea speaker has finished their sentence. This model makes novelpredictions about partial interpretations of conjunctive genericsentences, which we test in two behavioral experiments. Theresults support a strong view of incrementality, wherein lis-teners continuously update their beliefs based on expectationsabout where a speaker will go next with their utterance.

Habitual sentences (e.g. Bill smokes.) generalize an event overtime, but how do you know when a habitual sentence is true?We develop a computational model and use this to guide exper-iments into the truth conditions of habitual language. In Ex-pts. 1 & 2, we measure participants’ prior expectations aboutthe frequency with which an event occurs and validate thepredictions of the model for when a habitual sentence is ac-ceptable. In Expt. 3, we show that habituals are sensitive totop-down moderators of expected frequency: It is the expec-tation of future tendency that matters for habitual language.This work provides the mathematical glue between our intu-itive theories’ of others and events and the language we useto talk about them.

Pragmatic inferences are an integral part of language learn-ing and comprehension. To recover the intended meaning ofan utterance, listeners need to balance and integrate differentsources of contextual information. In a series of experiments,we studied how listeners integrate general expectations aboutspeakers with expectations specific to their interactional his-tory with a particular speaker. We used a Bayesian pragmaticsmodel to formalize the integration process. In Experiments1 and 2, we replicated previous findings showing that listenersmake inferences based on speaker-general and speaker-specificexpectations. We then used the empirical measurements fromthese experiments to generate model predictions about howthe two kinds of expectations should be integrated, which wetested in Experiment 3. Experiment 4 replicated and extendedExperiment 3 to a broader set of conditions. In both experi-ments, listeners based their inferences on both types of expec-tations. We found that model performance was also consistentwith this finding; with better fit for a model which incorporatedboth general and specific information compared to baselinesincorporating only one type. Listeners flexibly integrate dif-ferent forms of social expectations across a range of contexts,a process which can be described using Bayesian models ofpragmatic reasoning.

An essential part of cognitive science is designing experimentsthat distinguish competing models. This requires patience andingenuity—there is often a large space of possible experimentsone could run but only a small subset that might yield informa-tive results. We need not comb this space by hand: If we useformal models and explicitly declare the space of experiments,we can automate the search for good experiments, looking forthose with high expected information gain. Here, we presentan automated system for experiment design called webppl-oed.In our system, users simply declare their models and experi-ment space; in return, they receive a list of experiments rankedby their expected information gain. We demonstrate our sys-tem in two case studies, where we use it to design experimentsin studies of sequence prediction and categorization. We findstrong empirical validation that our automatically designed ex-periments were indeed optimal.

Understanding a gradable adjective (e.g., big) requires mak-ing reference to a comparison class, a set of objects or entitiesagainst which the referent is implicitly compared (e.g., big fora Great Dane), but how do listeners decide upon a compari-son class? Simple models of semantic composition stipulatethat the adjective combines with a noun, which necessarily be-comes the comparison class (e.g., “That Great Dane is big”means big for a Great Dane). We investigate an alternativehypothesis built on the idea that the utility of a noun in anadjectival utterance can be either for reference (getting the lis-tener to attend to the right object) or predication (describing aproperty of the referent). Therefore, we hypothesize that whenthe presence of a noun N can be explained away by its util-ity in reference (e.g., being in the subject position: “That N isbig”), it is less likely to set the comparison class. Across threepre-registered experiments, we find evidence that listeners usethe noun as a cue to infer comparison classes consistent with atrade-off between reference and predication. This work high-lights the complexity of the relation between the form of anutterance and its meaning.

Generic language (e.g., “Birds fly”) conveys generalizationsabout categories and is essential for learning beyond our directexperience. The meaning of generic language is notoriouslyhard to specify, however (e.g., penguins don’t fly). Tessler andGoodman (2019b) proposed a model for generics that is math-ematically equivalent to Bayesian belief-updating based on asingle pedagogical example, suggesting a deep connection be-tween learning from experience and learning from language.Relatedly, Csibra and Shamsudheen (2015) argue that genericsare inherently pedagogical, understood by infants as referringto a member of a kind. In two experiments with adults, wequantify the exchange-rate between generics and observationsby relating their belief-updating capacity, varying both thenumber of observations and whether they are presented ped-agogically or incidentally. We find generics convey strongergeneralizations than single pedagogical observations (Expt. 1),even when the property is explicitly demarcated (Expt. 2). Wesuggest revisions to the vague quantifier model of generics thatwould allow it to accommodate this intriguing exchange-rate.

Human knowledge accumulates over generations, amplifyingour individual learning abilities. What is the mechanism ofthis accumulation? Here, we explore how language allows ac-curate transmission of conceptual knowledge. We introduce anovel experimental paradigm that allows direct comparison oflearning from examples and learning from language. In ourexperiment, a teacher first learns a Boolean concept from ex-amples; they then communicate this concept to a student in afree conversation; finally, we test both teacher and student onthe same transfer items. We find that learning from languageis both sufficient and efficient: Students achieve accuracy veryclose to their teachers, while studying for less time. We thenexplore the language used by teachers and find heavy relianceon generics and quantifiers. Taken together, these results sug-gest that cultural accumulation of conceptual knowledge arisesfrom the ability of language to directly convey generalizations.

World knowledge enters into pragmatic utterance interpretation in complex ways, and may be defeasible in light of speakers’ utterances. Yet there is to date a surprising lack of systematic investigation into the role of world knowledge in pragmatic inference. In this paper, we show that a state-of-the-art model of pragmatic interpretation greatly overestimates the influence of world knowledge on the interpretation of utterances like Some of the marbles sank. We extend the model to capture the idea that the listener is uncertain about the background knowledge the speaker is bringing to the conversation. This extension greatly improves model predictions of listeners’ interpretation and also makes good qualitative predictions about listeners’ judgments of how ‘normal’ the world is in light of a speaker’s statement. Theoretical and methodological implications are discussed.