The role of constructional competition in children's infinitival-to omission errors

Abstract

Most 2-3-yr-old English-speaking children sometimes omit infinitival-to in obligatory contexts (e.g. *I want hold it vs. I want to hold it). In a series of studies, we explore whether omissions can be accounted for by competition between constructions based on their relative frequencies in the language children hear and on their presence or absence in the immediate discourse context.  Children hear some verbs both with (e.g., I want to go/get/have…’ - a ‘WANT-to’ construction) and without (e.g., I want it/a drink/that… - a ‘WANT-X’ construction) to-infinitive complements. Competition between these constructions could result in omission errors when the wrong construction is selected for production based on either relative input frequencies or recent use.

In one study, we examined relative error rates across children based on the frequencies of the WANT-to/going-to and WANT-X/going-X constructions in the language they hear, using corpus data from 13 children aged between 2-3 yrs.  We found that children who heard a higher proportion of WANT/going-X utterances produced higher proportions of infinitival-to omission errors with these verbs. Moreover, children’s error rates were higher with the verb that was relatively less frequently produced with infinitival-to in the input (Kirjavainen et al, 2009).

In a second study, we further investigated whether the WANT-to and WANT-X constructions affect rates of provision of infinitival-to by looking at their effects in the immediate discourse in naturalistic conversation.  We found that relative to discourse contexts in which the target verb 'WANT' had not previously appeared, the presence of the WANT-X construction was associated with a higher proportion of to-omission errors whereas a the presence of the WANT-to construction was associated with a higher proportion of grammatical WANT-to utterances. (Kirjavainen & Theakston, 2011)

Finally, we are currently investigating the constructional competition account in an experimental context using a priming paradigm to determine the effects of different constructions on subsequent provision of infinitival-to.

These findings suggest that competition between constructions may provide an explanation for children’s infinitival-to omission errors and provide support for a usage-based model of language where the distributional properties of the input play a significant role in acquisition.

Duration of the project

2009 - 2012

Funding bodies

British Academy Grant SG-54173 (A. Theakston)

Max Planck postdoctoral stipend to Dr. Minna Kirjavainen