Not All Children Agree: Acquisition of Agreement When the Input Is Variable

Abstract

In this paper we investigate the effect of variable input on the acquisition of grammar. More specifically, we examine the acquisition of the third person singular marker –s on the auxiliary do in comprehension and production in two groups of children who are exposed to similar varieties of English but that differ with respect to adult production of agreement on the auxiliary. In the first variety, the input to children for agreement on the auxiliary is consistent as the marker is always produced when there is a third person singular subject. In the second variety, however, there is a variable input for agreement marking on the auxiliary do: within and across speakers, agreement marking is sometimes present (e.g., He doesn't like venison) and sometimes absent (e.g., He don't like venison) with third person singular subjects. In other words, while both groups of children are exposed to an input where the third person singular marker is always present on the main verb (when it agrees with a third person singular subject), the input to the two groups of children differs in the context of agreement on the auxiliary do. The results of the present study show that while the two groups of children differ from each other in their own production of agreement marking on the auxiliary, a finding we attribute to the input type they are exposed to, they do not differ in their (in)ability to associate the marker to a third person singular subject in comprehension.

ACKNOWLEDGMENTS

Many thanks to Juliana Litts and the students at the Michigan State and Penn State Language Acquisition Labs for recruitment, scheduling, and testing. I would also like to express my gratitude to the teachers at Oakdale Christian Preschool (GRCS), Michigan Head Start, and the MSU Child Development Center for allowing us to work in their schools. Very special thanks to Cristina Schmitt and Alan Munn for their helpful comments throughout the elaboration of this work and also to the journal editors, Cynthia Fisher and Susan Goldin-Meadow, and three anonymous reviewers for their very helpful and insightful comments on the manuscript. This project was funded by the Calvin College Center for Social Research and by funding from Penn State University.

Notes

¹ Input like that shown in (15) may have potential implications for studies reporting that young typically developing children and SLI children omit do in their own speech and that children also rate omissions of do as grammatical (see Paradis, Rice, Crago, & Marquis, 2008; Rice, Hoffman, & Wexler, 2009).

² One potential problem with the spontaneous speech data presented here is that it only included adult-directed speech. It is possible that parents use more or less nonstandard, local variants when talking to their children than when talking to other adults. This raises the possibility of an input that looks very different from what is presented here. While we cannot exclude this possibility, the differences between adult-directed and child-directed speech with respect to the use of labovian type variation is not completely understood. It has been shown that parents sometimes use more local, nonstandard variants when talking to their own children than when talking to other adults. However, the opposite pattern has also been reported (Smith et al., 2007). (We thank an anonymous referee for encouraging us to examine this issue.)

³ This term was used by Schütze (2010) to describe Guasti and Rizzi's (2004) proposal.

⁴ A reviewer suggested that the experimental prompt which contains agreement marking on the main verb (e.g., Ask Zach if his dad writes with glitter glue) may have primed children to produce questions with agreement marking on the main verb and not on the auxiliary (i.e., utterances like (19b)).