Learning words from speakers with false beliefs

Abstract

During communication, hearers try to infer the speaker's intentions to be able to understand what the speaker means. Nevertheless, whether (and how early) preschoolers track their interlocutors' mental states is still a matter of debate. Furthermore, there is disagreement about how children's ability to consult a speaker's belief in communicative contexts relates to their ability to track someone's belief in non-communicative contexts. Here, we study young children's ability to successfully acquire a word from a speaker with a false belief; we also assess the same children's success on a traditional false belief attribution task. We show that the ability to consult the epistemic state of a speaker during word learning develops between the ages of three and five. We also show that false belief understanding in word-learning contexts proceeds similarly to standard belief-attribution contexts when the tasks are equated. Our data offer evidence for the development of mind-reading abilities during language acquisition.

The Role of Speaker Knowledge in Children's Pragmatic Inferences

“…The same findings are also significant in the context of broader debates about children's early communication skills by showing that young children are sensitive to social pragmatic reasoning in calculating speaker meaning. In this sense, they are consistent with evidence demonstrating, among other things, that preschoolers consult a speaker's knowledge when learning a novel word (Papafragou, Fairchild, Cohen, & Friedberg, 2016;Sabbagh & Baldwin, 2001) and take into account their interlocutors' perspective during online reference assignment (Nadig & Sedivy, 2002).…”

Developmental and cognitive aspects of children’s disbelief comprehension through intonation and facial gesture

“…Verbs like think, know and remember refer to physically unobservable states, and relationships between their acquisition and ToM development have been reported (Moore, Pure, & Furrow, 1990; Olson, 1988). Papafragou, Fairchild, Cohen, and Friedburg (2017) found that the tracking of speakers’ mental states is used when acquiring a new word from a person, and that this ability is developing between the ages of 3 and 5. For syntax it has been argued by de Villiers (2007) that the syntax of sentence complements under certain verbs is what facilitates reasoning about the knowledge states of others, claiming that language helps the development of ToM reasoning.…”

Actions do not speak louder than words in an interactive false belief task

“…These studies found that, compared to their performance in standard FB tasks, even 3-year-olds (and older children) performed proficiently in tracking an agent's FB to assign reference in situations that involve learning of new object labels. However, a recent study by Papafragou et al [32] suggests that this effect may be an artefact of miss-matched control conditions: once both task types were closely matched, no performance-enhancing effects of word learning contexts over standard FB tasks were found.…”

“…3 that did not refer to the child's memory of previous object locations (thus did not provide a semantic shortcut). Note that we did not include the novel label for the referred object in the request, because Southgate et al's study showed that it is not necessary for infants' success (see also [32]). We coded for the box that participants approached or pointed at first, which was either the referred box (correct in TB, incorrect in FB) or the non-referred box (incorrect in TB, correct in FB).…”