Finding agreement through a free exchange of views is often difficult. Collective deliberation can be slow, difficult to scale, and unequally attentive different voices. In this study, we trained an artificial intelligence (AI) mediate human deliberation. Using participants' personal opinions critiques, the AI mediator iteratively generates refines statements that express common ground among group on social or political issues. Participants (N = 5734) preferred AI-generated those written by...

Access to data is a critical feature of an efficient, progressive, and ultimately self-correcting scientific ecosystem. But the extent which in-principle benefits sharing are realized in practice unclear. Crucially, it largely unknown whether published findings can be reproduced by repeating reported analyses upon shared (“analytic reproducibility”). To investigate, we conducted observational evaluation mandatory open policy introduced at journal Cognition. Interrupted time-series indicated...

Pragmatic inferences are an integral part of language learn- ing and comprehension. To recover the intended meaning utterance, listeners need to balance integrate different sources contextual information. In a series experiments, we studied how general expectations about speakers with specific their interactional his- tory particular speaker. We used Bayesian pragmatics model formalize integration process. Experiments 1 2, replicated previous findings showing that make based on...

Abstract Before formal education begins, children typically acquire a vocabulary of thousands words. This learning process requires the use many different information sources in their social environment, including current state knowledge and context which they hear words used. How is this integrated? We specify developmental model according to consider an age-specific way integrate them via Bayesian inference. accurately predicted 2–5-year-old children’s word across range experimental...

Consider the process of collective decision-making, in which a group individuals interactively select preferred outcome from among universe alternatives. In this context, "representation" is activity making an individual's preferences present via participation by proxy agent -- i.e. their "representative". To end, learned models human behavior have potential to fill role, with practical implications for multi-agent scenario studies and mechanism design. work, we investigate possibility...

Human communication has been described as a contextual social inference process. Research into great ape inspired by this view to look for the evolutionary roots of social, cognitive and interactional processes involved in human communication. This approach highly productive, yet it is partly compromised widespread focus on how apes use understand individual signals. paper introduces computational model that formalizes multi-faceted process integrates (a) information contained signals make...

Before formal education begins, children typically acquire a vocabulary of thousands words. This learning process requires the use many different information sources in their social environment, including current state knowledge and context which they hear words used. How is this integrated? We specify developmental model according to consider an age-specific way integrate them via Bayesian inference. accurately predicted 2-to-5 year-old children’s word across range experimental conditions...

Language is a remarkably efficient tool for transmitting information. Yet human speakers make statements that are inefficient, imprecise, or even contrary to their own beliefs, all in the service of being polite. What rational machinery underlies polite language use? Here, we show speech emerges from competition three communicative goals: convey information, be kind, and present oneself good light. We formalize this goal tradeoff using probabilistic model utterance production, which predicts...

Pragmatic abilities are fundamental to successful language use and learning. Individual differences studies contribute understanding the psychological processes involved in pragmatic reasoning. Small sample sizes, insufficient measurement tools, a lack of theoretical precision have hindered progress, however. Three addressed these challenges three- five-year-old German-speaking children (N = 228, 121 female). Studies 1 2 assessed psychometric properties six pragmatics tasks. Study 3...

Language is learned in complex social settings where listeners must reconstruct speakers’ intended meanings from context. To navigate this challenge, children can use pragmatic reasoning to learn the meaning of unfamiliar words. A critical challenge for that it requires integrating multiple information sources, which have typically been studied separately. Here we study integration process. We isolate two sources and -- using a probabilistic model conversational – formalize how they should...

Logic tells us that two negatives make a positive, but in language, things are not so black and white: A person "not unhappy" may be entirely happy. We hypothesize innovative uses of double like stem from listeners entertaining flexible meanings for negation markers "not" "un-", which context can then help disambiguate. formalize this hypothesis computational model language understanding, predicts means something different "happy" also makes the additional prediction single negations...

Generic statements convey generalizations about categories, but how generic predications combine is unclear. “Elephants live in Africa and Asia” does not mean that individual elephants on both continents. In addition, such conjunctive generics pose interesting questions for theories of incremental processing because the meaning sentence can change part-way through: Africa” would imply most or all do, “Africa implies some each. We extend a recently proposed computational model language...

Generic language (e.g., "Birds have hollow bones") conveys generalizations about categories and is a fundamental ubiquitous mechanism of learning the world. How exactly this occurs, however, unclear: Generics exhibit so much flexibility in how they are interpreted that quantitative theory generics update beliefs has not only been developed or tested, but often explicitly dismissed. Here, we explore hypothesis introduced by Tessler Goodman (2019) via an uncertain threshold like vague...

Generic language (e.g., “Birds fly”) conveys generalizations about categories and is essential for learning beyond our direct experience. The meaning of generic notoriously hard to specify, however penguins don’t fly). Tessler Goodman (2019) proposed a model generics that mathematically equivalent Bayesian belief-updating based on single pedagogical example, suggesting deep connection be- tween from experience language. Relatedly, Csibra Shamsudheen (2015) argue are inherently pedagogical,...

Abstract Pragmatics is foundational to language use and learning. Computational cognitive models have been successfully used predict pragmatic phenomena in adults children – on an aggregate level. It unclear if they can be behavior individual We address this question (N = 60, 3- 5-year-olds), taking advantage of recent work cue integration. In Part 1, we data from four independent tasks estimate child-specific sensitivity parameters three information sources: semantic knowledge, expectations...

Understanding a gradable adjective (e.g., big) requires making reference to comparison class, set of objects or entities against which the referent is implicitly compared big for Great Dane), but how do listeners decide upon class? Simple models semantic composition stipulate that combines with noun, necessarily be- comes class “That Dane big” means Dane). We investigate an alternative hypothesis built on idea utility noun in adjectival utterance can be either (getting listener attend right...

Abstract The meanings of natural language utterances depend heavily on context. Yet, what counts as context is often only implicit in conversation. utterance it's warm outside signals that the temperature relatively high, but could be high relative to a number different comparison classes : other days year, weeks, seasons, etc. Theories sensitivity agree class crucial variable for understanding meaning, little known about how listener decides upon class. Using case study gradable adjectives...

Human communication has been described as a contextual social inference process. Research into great ape inspired by this view to look for the evolutionary roots of social, cognitive, and interactional processes involved in human communication. This approach highly productive, yet it is often compromised too-narrow focus on how apes use understand individual signals. In paper, we introduce computational model that formalizes multi-faceted process relies information contained signal,...

Replication is the cornerstone of science – but when and why? Not all studies need replication, especially resources are limited. We propose that a decision-making framework based on Bayesian philosophy provides basis for choosing which to replicate.

Pragmatics is foundational to language use and learning. Computational cognitive models have been successfully used predict pragmatic phenomena in adults children -- on an aggregate level. It unclear if they can be behavior individual We address this question (N = 60, 3- 5-year-olds), taking advantage of recent work cue integration. In Part 1, we data from four independent tasks estimate child-specific sensitivity parameters three information sources: semantic knowledge, expectations about...

(Find newer version of preprint here: https://psyarxiv.com/67ne8) Language is a remarkably efficient tool for transmitting information. Yet human speakers make statements that are inefficient, imprecise, or even contrary to their own beliefs, all in the service being polite. What rational machinery underlies polite language use? Here, we show speech emerges from competition three communicative goals: convey information, be kind, and present oneself good light. We formalize this goal tradeoff...

The meaning of an utterance can change depending on the context. Yet, what counts as context is often only implicit in everyday conversation. “it’s warm outside” signals that temperature outside relatively high, but could be high relative to a number different comparison classes: other days year, weeks, seasons, etc. Theories context-sensitive language use agree class crucial feature understanding, little known about how listener decides upon class. We extend Bayesian model pragmatic...