English-learning preschoolers can correctly parse and interpret negative sentences to guide their interpretations of novel noun and verb meanings

ABSTRACT

We studied English-learning children’s ability to learn the meanings of novel words from sentences containing truth-functional negation (Exp1) and to use the semantics of negation to inform word meaning (Exp2). In Exp1, 22-month-olds (n = 21) heard dialogues introducing a novel verb in either negative-transitive (“Mary didn’t blick the baby”) or negative-intransitive (“Mary didn’t blick”) sentences. When then asked to “Find blicking!” while viewing two-participant versus one-participant actions, children who heard negative-transitive sentences looked longer at the two-participant event than children who heard negative-intransitives. Thus, the mere presence of negation does not disrupt sentence processing and word learning in young children. Experiment 2 tested whether 2-to-4-year-olds (n = 20) use the semantics of negation to restrict the meaning of novel nouns when categorizing objects varying along a perceptual continuum (from 10 to 90% exemplars). Children initially heard “These are blickets” paired with certain exemplars (e.g., yellowish creatures, exemplars 10 and 30%). They then saw additional exemplars (e.g., pinkish creatures, 70 and 90%) while hearing either “These are not blickets” (negative condition) or “These are also blickets” (affirmative condition). At test, when seeing two novel exemplars from the continuum (e.g., creatures 20 and 80%) and asked to find “a blicket,” children in the negative condition selected the exemplar from the bottom of the continuum (i.e., the 20%) more often than children in the affirmative condition. Thus, English-learning children as young as 22-months of age correctly parse negative sentences and 2-to-4-year-olds can use negation to understand the boundaries of a word’s meaning.

Acknowledgements

We thank all the bright “actors” who participated in the videos illustrating actions in Experiment 1 and specially Marianna Olave and Abigail Marcus for performing so well in the dialogue videos. Special thanks to Abigail Laver for acting so beautifully in the teaching videos of Experiment 2. We also would like to thank Yiyi (Yelina) Chen and Bosede Ajiboye who assisted us in the recording of these videos and with recruitment and testing of participants, and Abigail Marcus who also assisted us in programing Experiment 1. We are also extremely grateful to Cindy Fisher for sharing and allowing us to use the videos of actions used in the test trials of Experiment 1. We also thank Sandy Waxman and Alexander LaTourrette for sharing the stimuli of Havy and Waxman (2016) for Experiment 2. We also thank Isabelle Dautriche for her thoughtful comments and suggestions for Experiment 2.

Declaration of Conflicting Interests

The authors declared that there were no conflicts of interest with respect to the authorship or the publication of this article.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹ We did not include an additional comparison between performance in the negative intransitive condition versus an affirmative version of these intransitive sentences, because this comparison wouldn’t be very informative. According to previous studies using the same verb-learning paradigm (e.g., Arunachalam & Dennis 2018; Arunachalam et al. 2016; Arunachalam & Waxman 2010; Dautriche et al. 2014; de Carvalho, Dautriche, et al. 2021; Messenger et al. 2015; Yuan & Fisher 2009; Yuan et al. 2012), when children interpret the novel verb as intransitive and therefore infer that the verb’s meaning involves one participant role, they tend to show no preference for any of the events at test because the intransitive verb could refer either to the one-participant event or to a subcomponent of the two-participant event (Fernandes et al. 2006; Fisher 2002; Yuan & Fisher 2009). As a result, the negative intransitive and the affirmative intransitive conditions could give rise to exactly the same pattern of results (i.e., performance around chance) but not for the same reasons. In this case, it would be unclear if children performed around chance (i.e., showing no preference for the two-participant event) because they were confused by negation or if they correctly understood the sentence but were unable to decide whether the verb referred specifically to the one-participant event or to a subcomponent of the two-participant event. We therefore preferred to focus our comparison between affirmative and negative sentences specifically on the transitive sentences because these are the conditions in which the effect of learning (i.e., preference for the two participants event) should be clearly observed.

² Note that there is evidence in the literature suggesting that children in the age range we tested (i.e., 1-to-4-year-olds) might be capable to understand “also” in our sentences, beucase they spontaneously produce “also” in their own speech and they were shown to understand it in different tasks and in different languages (see, e.g., Berger & Höhle 2012; Bergsma 2006; Höhle et al. 2009; Kurokami et al. 2021).

³ The static images were animated (i.e., moving up and down) to better hold children’s attention and interest.

⁴ Note that since adults were 100% correct in the Negative Condition and we did not have any pointing towards the alternative picture in the negative condition, in order to run our mixed model in the adult data, we added a bit of noise to this data by randomly shifting the condition of 3% of our data, which is a standard procedure in these cases.

Additional information

Funding

This project has received financial support from the CNRS through the MITI interdisciplinary programs award to Alex de Carvalho. It was also supported by a Fyssen Foundation Research Grant award to Alex de Carvalho, (grant Eotp RH03J20AFO04_ EDENFCOGN°H03R8240) and by grants from the University of Pennsylvania Research Foundation (URF) award to Alex de Carvalho and John Trueswell and from the National Institute of Child Health and Human Development (grant R01HD37507) award to John Trueswell.