6 Overcoming errors

There have been a number of ideas put forward to explain how children eventually overcome over-generalisation errors. These include the following:

• Entrenchment: The more often a child hears a verb in a particular syntactic context (e.g. I suggested the idea to him) the less likely they are to use it in a new context they haven’t heard it in (e.g. ∗I suggested him the idea; (Ambridge, Pine and Rowland, 2012). Note the convention is to mark ungrammatical sentences with an asterisk∗.
• Pre-emption (Goldberg, 2007): If a child repeatedly hears a verb in a construction (e.g. I filled the cup with water) that serves the same communicative function as a possible unheard generalisation (e.g. ∗ I filled water into the cup), then the child infers that the generalisation is not available (Ambridge, Pine and Rowland, 2012).
• Construction semantics (Pinker, 1989): Constructions are associated with particular meanings (e.g. the transitive causative with direct external causation). As children refine this knowledge, they will cease to insert verbs that do not bear these meaning elements into the construction (e.g. ∗ The joke giggled him).
• Cognitive mechanisms (Ibbotson, 2020): Evidence from a number of cross-linguistic studies shows that a maturation in cognitive abilities independent of language, such as inhibitory control and working memory, can explain the pattern of individual differences in over-generalisations errors.
• Conventionality: There is a huge amount of evidence from social psychology that people conform, obey and act the way they do because they are quite simply not willing to behave differently from others in their group. Eventually overgeneralisations get subsumed into examples that essentially just mandate: we say it this way. In this context children pay attention to more expert speakers who offer feedback in conversational turns right after errors, with the result that children make repairs (Clark, 2020; Chouinard and Clark, 2003; Saxton, 2000; Strapp and Frederico, 2000).
• Intention-reading (Tomasello, 2003): Children form constructional categories based on trying to comprehend the overall communicative intention behind the utterance; as if every utterance demands an answer to the question: ‘what is the relevance of directing my attention in this way given the situation?’. If the intentions of other people aren’t equally possible in all situations then neither are all possible linguistic generalisations or over-generalisations. Using this type of intention-reading can help children home in on the way adults use language.