How do children’s linguistic and socio-cognitive skills interact with language input in learning modal and mental state terms?

The aim of this project was to investigate whether and how learning the lexical items and grammatical structures that we typically use to talk about mental states and perspectives(e.g. What do you THINK? HE KNOWS it’s in the red box) help children to understand mental states such as beliefs and desires. In particular, we were interested to see whether different linguistic tools that we use to represent mental states can provide the same amount of scaffolding for children’s understanding of beliefs, especially when these beliefs are false (i.e. false belief). For example, both mental verbs (e.g. think, know) and modal verbs (e.g. might, must) can indicate speakers’ degree of certainty (e.g. I THINK/KNOW it’s in the red box. It MIGHT/MUST be in the red box). In addition, some languages also use evidential markers: In Turkish we can also say it is-DI in the red box, where the -di indicates that the speaker is certain about the location of the object. So, we also took a cross-linguistic perspective to compare English- and Turkish-speaking children’s understanding and use of mental-state language and their understanding of false belief. 

Our first study was a longitudinal study, which allowed us to take a closer look at the developmental relationships between children’s understanding of different forms of mental-state language and their understanding of false belief. Around their third birthday, 48 children took part in the first two sessions, where we looked at their understanding and production of complement clauses (e.g. John saw that his mum ate sweets), their understanding of mental and modal verbs (e.g. He KNOWS that the sticker is in the red box. The sticker MUST be in the red box), and their understanding of false belief. In addition, we also measured their general linguistic skills (vocabulary and grammar), their executive function (e.g. working memory, inhibition), and other Theory of Mind skills (e.g. to understand others’ desires). Unlike previous studies, we also made sure that our false-belief tests did not contain any of the mental-state language used in the other tasks. Six months later, when children were around 3;5 years old, we looked at their understanding of false belief and mental-state language again. Our main results suggest that children’s false-belief understanding at the age of 3;5 can be predicted by their understanding and production of complement clauses (e.g. John saw that his mum ate sweets) and their understanding of mental verbs, but only when used with third-person subjects (e.g. HE knows that the sticker is in the red box). 

Our second study was a training study where we investigated whether an increased exposure to mental-state language can boost children’s false-belief understanding within a short period of time. Children who took part in this study were between the ages of 2;9 and 3;10. In a four-week eight-session program, 70 English-speaking children had activities with mental-state contrasts (such as mistakes, disagreement and surprise) mediated linguistically in three conditions: 1) Simple clauses (The book is on the table), 2) First-person complements and mental verbs (I THINK that the book is on the table) and 3) Third-person complements and mental verbs (HE THINKS that the book is on the table). The only group to improve significantly in their false-belief reasoning was the group trained with first-person complements and mental verbs. Analysing predictors of false-belief improvement, we found a significant advantage of training with first-person complements and mental verbs over training with simple clauses as well as a positive effect of inhibitory control. These results suggest that having mental states spelled out with first-person complements and mental verbs helps children learn from situations with mental-state contrasts, and the effect of inhibitory control aligns with previous conclusions that the ability to suppress one’s own perspective is central for developments in mental-state reasoning. 

In our third study, we compared English- and Turkish-speaking children’s production and understanding of different linguistic tools that can be used to mark relative certainty (e.g. mental verbs and evidential markers) and how this relates to children’s understanding of false belief and other social-cognitive skills such as source monitoring (i.e. understanding the difference between direct and indirect evidence). Overall, we have tested 100 children between the ages of 3;5 and 5;0. Preliminary analyses suggest that both in Turkish and in English children’s use of evidential markers and mental verbs, their source monitoring, and their understanding of false belief are interrelated. 

In a PhD project related to this work package, we have also been looking more closely at children’s use and interpretation of modal verbs and how this is driven by usage patterns in their input. For example, in an utterance such as He MUST be nice to the dog, the modal verb must can have an epistemic or a deontic function. If interpreted epistemically, the utterance could be paraphrased as I THINK he’s nice to the dog. If interpreted as a deontic modal, the utterance could be paraphrased as He NEEDS TO BE nice to the dog. Our results suggest that children tend to acquire the deontic function before the epistemic function. In addition, children’s use and interpretation of specific modal verbs is driven by form-function mappings in their input.

Outputs

Boeg Thomsen, Theakston, A., Kandemirci, B., & Brandt, S. (2019). Do complement clauses really support false-belief reasoning? A longitudinal study with 2- to 3-year-olds. Poster presented at the Boston University Conference on Language Development, Boston, MA.

Bell, K., Brandt, S., Lieven, E. & Theakston, A. (2019). Does caregiver input influence children’s acquisition of modality? Poster presented at the Child Language Symposium, Sheffield, UK.

Theakston, A., Bell, K., Lemen, H., Brandt, S., & Lieven, E. (2019). Form-meaning relations in acquisition: the case of polysemous constructions. Paper presented at the Societas Linguistica Europaea 52nd Annual Meeting, Leipzig, Germany.

 

Project Team: Silke Brandt (Lead), Ditte Boeg Thomsen, Birsu Kandemirci, Elena Lieven and Anna Theakston. 

Start date: March 2017

Duration: 2.5 years

(Work Package 8)