Variable Input: What Sarah Reveals About Nonagreeing Don't and Theories of Root Infinitives

Abstract

Two recent proposals link the use of nonagreeing don't to the Root Infinitive (RI) Stage. Guasti & Rizzi (2002) argue for a misset parameter involving how agreement is spelled out. Schütze (2010) proposes that Infl is underspecified in child language and that do surfaces to support the contracted clitic/affix n't. Both proposals obtain partial support from the Sarah corpus (Brown 1973), yet neither proposal drew on Sarah's parents’ use of nonstandard, nonagreeing don't as a possible explanation for Sarah's production. In this article I argue that much of the nonagreeing don't produced by Sarah is not part of the RI Stage. Once Sarah's data are removed from the analysis, the remaining data support Schütze's proposal.

ACKNOWLEDGMENTS

I would like to thank two anonymous reviewers for their helpful comments on this work. All errors are my own. Previous versions of this article were presented at The Penn State Center for Language Science Forum and BUCLD 36 2012.

Notes

¹Schütze notes that Sarah's parents used nonagreeing don't in their own speech and asked how this might bear on the interpretation of the results. He concludes in a later footnote that Sarah's unique input did not appear to have made her an outlier among the group of children in terms of her production of nonagreeing don't, and he suggests that nonagreeing don't may have a different analysis in her grammar from what it has in her mother's grammar. These conclusions lead him to include the Sarah corpus in his analysis. Also, as noted by Schütze, Guasti & Rizzi (2002) make no mention of the use of nonagreeing don't in the input to Sarah.

²Guasti & Rizzi (2002) report that Ross's last occurrence of nonagreeing don't occurred at 6;03 in file 50. I believe this is a misprint. Ross was 4;03, not 6;03, in file 50, and he does not produce any nonagreeing don't in file 50. Instead, his last occurrence was at 6;11.19 (File 83b2).

³These numbers differ from Schütze (2010) because he carried out a different analysis of the data. While I included only Sarah's mother and father in the analysis and also included all 3sg pronouns and all 3sg full NPs that occurred with don't, Schütze reports that he carried out automated searches on utterances by all people other than the target child, suggesting that he may have also included the utterances of the university researchers present during the recordings, and his automated searches looked for sequences of the form “{he/she/it} {don't/doesn't}” which excludes structures with inversion (e.g., doesn't it get snarls when it is long?; Sarah's Mother File128), utterances where some element intervenes between the pronoun and the auxiliary (e.g., she usually doesn't give up; Sarah's Mother File135), the pronoun that (e.g., that doesn't come out; Sarah's Mother File61), and full NPs (e.g., Poor Inky doesn't have any teeth; Sarah's Mother File59).

⁴For details on the CLAN program and transcriptions see MacWhinney (2000).

⁵It is important to note that when counting tokens of sentences like those in (7), Schütze analyzed only a subset of the data because he followed the counting procedures used in G&R. It is possible there are more tokens than reported. In order to determine this, I carried automated CLAN KWAL searches for all tokens of do and then went through by hand and found those tokens containing a 3sg subject. The results showed a slight increase in the number of utterances like (7) with nonagreeing do for Sarah (four additional tokens). However, there was no increase for Nina. This means that Sarah produced six tokens of nonagreeing do like that illustrated in (7).

⁶It should be noted that although it looks like Sarah continues to omit the auxiliary does after 3;09, 18 of Sarah's 22 tokens of omissions between 4;01–4;06 years of age involve repetitions of the question, “What does this/that spell?,” as illustrated in (i).

(i) a. what dis [: this] spell ?

b. what dat [: that] spell ?

c. what this spell then ?

d. what all that spell ?

If these repeated sentences are treated as only 1 token, then Sarah would have only 6 tokens of does omission during Stages III and IV instead of 27 tokens.

⁷Guasti & Rizzi (2002) argue that the cases of nonnominative subjects with nonagreeing don't could be interpreted as involving truncated structures above T (and therefore including do), but under Agr (and therefore excluding nominative marking on the subject).

⁸A reviewer asked how Sarah's parents’ variable usage of nonagreeing don't could be accounted for and suggested the Spell-out Account as one possible explanation for adult behavior (and also child behavior). It is beyond the scope of this article to provide a proposal about the variable usage of nonagreeing don't in the adult speech. Instead, the goal of this article was to examine how variable input affects acquisition in children. Nevertheless, as far as I can tell, the Spell-out Account is unable to explain the social factors constraining variable usage in adult speech or explain Sarah's usage of nonagreeing do in interrogatives. Furthermore, adult speakers who use nonagreeing don't in declaratives also use them in negative interrogatives (e.g., Don't it matter?). This is true of Sarah's parents as well. Nonagreeing don't in interrogatives would be inconsistent with the Spell-out Account.

⁹Sarah's Stage II seems to represent true optionality. I looked for patterns to determine whether verb type, subject NP type (full NP, he, she, or it), or mother's own speech (i.e., whether the mother had recently used nonagreeing don't) predicted the use of nonagreeing don't in Sarah's speech, and none of these factors appeared to have an impact.