Search

There is considerable debate about the origin, mechanism, and extent of humans' capacity for recursive mindreading: the ability to represent beliefs about beliefs about beliefs (and so on). Here we quantify the extent to which language exposure could support this ability, using a Large Language Model (LLM) as an operationalization of distributional language knowledge. We replicate and extend O'Grady, et al. (2015)'s finding that humans can mindread up to 7 levels of embedding using both their original method and a stricter measure. In Experiment 2, we find that GPT-3, an LLM, performs comparably to humans up to 4 levels of embedding, but falters on higher levels, despite being near ceiling on 7th-order non-mental control questions. The results suggest that distributional information (and the transformer architecture in particular) can be used to track complex recursive concepts (including mental states), but that human mentalizing likely draws on resources beyond distributional likelihood.

Context changes expectations about upcoming words—following a story involving an anthropomorphic peanut, comprehenders expect the sentence the peanut was in love more than the peanut was salted, as indexed by N400 amplitude (Nieuwland & van Berkum, 2006). This updating of expectations has been explained using Situation Models—mental representations of a described event. However, recent work showing that N400 amplitude is predictable from distributional information alone raises the question whether situation models are necessary for these contextual effects. We model the results of Nieuwland and van Berkum (2006) using six computational language models and three sets of word vectors, none of which have explicit situation models or semantic grounding. We find that a subset of these can fully model the effect found by Nieuwland and van Berkum (2006). Thus, at least some processing effects normally explained through situation models may not in fact require explicit situation models

Can we know a word by the company it keeps? Aspects of meaning that concern physical interactions might be particularly difficult to learn from language alone. Glenberg & Robertson (2000) found that although human comprehenders were sensitive to the distinction between afforded and nonafforded actions, distributional semantic models were not. We tested whether technological advances have made distributional models more sensitive to affordances by replicating their experiment with modern Neural Language Models (NLMs). We found that only one NLM (GPT-3) was sensitive to the affordedness of actions. Moreover, GPT-3 accounted for only one third of the effect of affordedness on human sensibility judgments. These results imply that people use processes that go beyond distributional statistics to understand linguistic expressions, and that NLP systems may need to be augmented with such capabilities.

Previous work has found competing evidence for how contextual diversity influences early word learning. In support of contextual diversity facilitating learning, the corpus-derived diversity metric from Hills, Maouene, Riordan, and Smith (2010) was found to correlate with earlier ages of acquisition in children. We extend this work to five languages, accounting for a nonlinear relationship between the raw contextual diversity metric and word log-frequencies, and we account for additional covariates such as word length, concreteness, and lexical class. In contrast with the original result, we find that contextually diverse words are acquired later by children across languages. Our findings support the hypothesis that contextual diversity introduces an excess of possible meanings for contextually diverse words, adding noise to the word learning process. This hindering effect overshadows any benefits of syntactic or semantic bootstrapping during early word learning, when children are still in the early stages of vocabulary and conceptual development.

Words in human languages cluster together in phonological neighborhoods more closely than would be expected by chance. But why? One explanation is that large neighborhoods are directly selected for, possibly because they scaffold word learning and production. But it's also possible that they emerge as a byproduct of other constraints or selection pressures operating over real lexica. We advance one such selection pressure as a candidate explanation. A pressure to avoid overloading unique wordforms with homophones may lead to clusters of words that are not identical but similar. Using simulated baselines, we test the viability of this alternative account. We find that a pressure against loading too many meanings on unique wordforms––paired with the phonotactics of a target language––produces lexica with neighborhoods that are at least as large on average as those in real lexica. This does not rule out the possibility of a pro-neighborhood pressure, but it does demonstrate the viability of a parsimonious alternative account based on a pressure against homonymy for which there is independent evidence.

When do people use knowledge about the world in order to comprehend language? We asked whether pronoun resolution decisions are influenced by knowledge about physical plausibility. Results showed that referents which are more physically plausible in described events were more likely to be selected as antecedents of ambiguous pronouns, implying that resolution decisions were driven by physical inference. An alternative explanation is that these decisions were driven instead by distributional word knowledge. We tested this by including predictions of a statistical language model (BERT) and found that physical plausibility explained variance on top of the statistical language model predictions. This indicates that at least part of people's pronoun resolution judgments comes from knowledge about the world and not the word. This result constrains psycholinguistic models of comprehension—world knowledge must influence propositional interpretation—and raises the broader question of how non-linguistic physical inference processes are incorporated during comprehension.

Despite being designed for performance rather than cognitive plausibility, transformer language models have been found to be better at predicting metrics used to assess human language comprehension than language models with other architectures, such as recurrent neural networks. Based on how well they predict the N400, a neural signal associated with processing difficulty, we propose and provide evidence for one possible explanation—their predictions are affected by the preceding context in a way analogous to the effect of semantic facilitation in humans.

Having cancer is often described metaphorically as a battle(“my fight against cancer”) or as a journey (“my path throughcancer treatment”). Previous experimental work has demon-strated that these metaphors can influence people’s reason-ing and emotional inferences about experiences with cancer(Hendricks, Demj ́en, Semino, & Boroditsky, 2018; Hauser &Schwarz, 2019). However, it is currently unknown how theuse of these metaphorical frames translates into behavioralchanges, such as the likelihood and magnitude of charitablegiving. Using hand-labeled data from more than 5,000 Go-FundMe cancer-related campaigns, we ask how a campaign’suse of metaphor predicts several measures of donation behav-ior beyond what other control variables predict (e.g. shares onFacebook). We find that the presence of either metaphor fam-ily (battle or journey) has a positive effect on campaign successand donation behavior.

Do individual sounds carry meaning? The relationship between sound and meaning in human languages is typicallyassumed to be arbitrary, though recent research provides evidence for the existence of both iconicity and systematicitybetween word forms and their meaning. However, this research has not asked whether individual sounds in a languagecovary in systematic ways with aspects of meaning. In two analyses, we find evidence for more systematicity betweenthe initial phones of words and those words concreteness ratings than one would expect in a truly arbitrary lexicon. Thissuggests that initial phones may act as cues to aspects of word meaning, and raises questions about whether languagelearners detect and exploit these cues.

Ambiguity pervades language. One prevalent kind ofambiguity is indirect requests. For example, “My office isreally hot” could be intended not only as a complaint aboutthe temperature, but as a request to turn on the AC. How docomprehenders determine whether a speaker is making arequest? We ask whether the prosody of an utterance providesinformation about a speaker’s intentions. In a behavioralexperiment, we find that human listeners can identify whichof two utterances a speaker intended as a request, suggestingthat speakers can produce discriminable cues. We then showthat the acoustic features associated with an utterance allow aclassifier to detect the original intent of an utterance (74%accuracy). Finally, we ask which of these features predictlistener accuracy on the behavioral experiment.

Human communication has a remarkable capacity todescribe events that occurred elsewhere and at othertimes. In particular, when describing complex narratives,speakers must communicate temporal structure using amixture of words (e.g., “after”), gestures (e.g., pointingrightward for a later event), and discourse structure (e.g.,mentioning earlier events first). How do listenersintegrate these sources of temporal information to makesense of complex narratives? In two experiments, wesystematically manipulated gesture, speech, and order-of-mention to investigate their respective impacts oncomprehension of temporal structure. Gesture had asignificant effect on interpretations of temporal order.This influence of gesture, however, was weaker than theinfluence of both speech and order-of-mention. Indeed,in some cases, order-of-mention trumped explicitdescriptions in speech; for instance, if ‘earlier’ eventswere mentioned second, they were sometimes thought tohave occurred second. Listeners integrate multiplesources of information to interpret what happened when.

Languages exhibit striking semantic diversity, but differentlanguages often share core metaphors. Conceptual MetaphorTheory (Lakoff & Johnson, 1980) claims that universalhuman experiences give rise to conceptual representationsthat are then expressed in language. But languages changeslowly, making it difficult to observe implicitconceptualization affecting linguistic convention in realtime. Here, we describe a shared conceptualizationpreviously absent from speech that has now becomeconventionalized in linguistic metaphors. In two studies, wedocument how members of the US military talk about timeusing conventionalized lateral metaphors (e.g., ‘push themeeting right’ to mean ‘move the meeting later’). We showthat military members, unlike civilians, consider suchsentences to be acceptable—sometimes even moreacceptable than more standard phrases. Moreover, militarypersonnel seem unaware that these lateral metaphors arespecific to their linguistic sub-community. Our findingssuggest that implicit mental representations can becomeconventionalized metaphors in language.

We test the robustness of American college students’ mentaltimeline to dual tasks that have interfered with spatial andverbal reasoning in prior work. We focus on the left-right axisfor representing sequences of events. We test Americancollege students, who read from left to right. We test forautomatic space-time mappings using two established space-time association tasks. We find that their tendency toassociate earlier events with the left side of space and laterevents with the right remains under conditions of visuospatialand verbal interference. We find this both when participantsmade time judgments about linguistic and non-linguisticstimuli. We discuss the relationship between these results andthose obtained for mental timelines that result from learningnew metaphors in language (Hendricks & Boroditsky, 2015),and the effects of the same interference tasks on number tasks(mental number-line and counting; van Dijck et al., 2009;Frank et al., 2012).

Mathematical expertise builds on a foundation of space,especially the ability to map exact numbers to linear space.This “mental number-line” is known to vary cross-culturally,but there is debate about the mechanisms responsible for itscultural elaboration. We investigated the role of co-speechgesture, a ubiquitous cultural activity, in stabilizing andentrenching the mental number-line within a community.Imitating culture-specific gestures systematically shapedgesturers’ mental number-line. Moreover, gestures were usedspontaneously to infer speakers’ spatial understanding ofnumber, and merely observing these gestures was sufficient toshape the observer’s own mental number-line. These findingsestablish co-speech gesture as one mechanism for propagatingand perpetuating the number-line