Explaining Syntax acquisition: A cross-linguistic model of the acquisition of syntactic categories

Understanding how children learn their first language is important because early progress in language learning is a key predictor of later educational achievement. It is also critical for the treatment of conditions like Developmental Language Disorder, which affects 2 children on average in every primary-school classroom. However, explaining how children learn language is challenging because it means explaining how children learn a system where the way words fit together does not depend on what the words mean, but on a complex set of language-specific rules for combining categories like Noun and Verb. For instance, English speakers say "The cat chased the mouse," while Japanese speakers use a structure like "Cat the mouse the chased" to communicate the same message. The aim of this project is to develop an explanation of how infants learn to distinguish Nouns and Verbs and so identify the building blocks of sentences. This explanation must be able to do three things. First, since infants readily learn whichever language, they are exposed to, it must explain how children distinguish between Nouns and Verbs in a way that works equally well across languages. Second, since children’s language-learning abilities depend on systems in the brain, it must be consistent with what we know about human biology. Third, since we know that children can manipulate Noun and Verb categories by 18 months of age, it must explain how children distinguish between them using skills that are present in language-learning infants. In this project we will develop an explanation that meets all of these challenges. In Work Package 1, we will apply a computer model developed in the Principal Applicant’s previous work on French to several diverse languages. This model distinguishes between words that refer to objects and words that refer to actions in the infant’s very early vocabularies; tracks the way these words are used in caregivers’ speech; and then uses this information to decide whether new words are Nouns or Verbs. The results of this work package will tell us whether our explanation works across languages and hence whether it can explain how children distinguish between Nouns and Verbs in whichever language they are exposed to. In Work Package 2, we will use Electroencephalography (EEG), to test whether 10-month-old infants are sensitive to the difference between words that refer to objects and words that refer to actions – and whether different parts of the infant brain respond to object- and action-words in the same way as those of adults. This will tell us whether infants distinguish object- and action-words in the same way as our computer model and whether this method is consistent with what we know about the brain. In Work Package 3, we will use Pupillometry experiments to test whether 10-month-olds are surprised (show a dilated pupil response) to new object- and action-words when they are presented in contexts associated with old action- and object-words. This will tell us whether the skills used by our computer model to decide whether new words are Nouns or Verbs are present in language-learning infants. We will maximise the impact of our work by working with the BBC and the Royal College of Speech and Language Therapists to communicate our findings to parents and language practitioners, and by organising a workshop to discuss their therapeutic implications for children with language difficulties.