Children’s early negative auxiliaries are true auxiliaries

ABSTRACT

This study investigates young children’s acquisition of functional categories through their use of negative words and negative auxiliaries in particular. Drawing from CHILDES, we analyze twelve months of spontaneous speech by 14 children (youngest age 1;9, oldest age 3;1) and their mothers, in order to assess whether children’s earliest negative productions are morphological combinations and reflect possession of abstract syntactic categories or are instead input-driven formulae. In five analyses we show that (i) two-year-olds use a wide and overlapping range of negative and positive auxiliaries; (ii) the range of the negative auxiliaries children produce is strongly correlated with the range of the positive auxiliaries they produce; (iii) children’s most common negative auxiliary, don’t, is used grammatically with respect to the syntactic category being negated and with respect to overt markings of tense; (iv) children’s subject agreement errors with don’t are mirrored by subject agreement errors with do, have, and haven’t; and (v) children omit auxiliaries with not at rates that cannot be attributed to properties of their input. Our findings support the hypothesis that children’s earliest negations are syntactically adult-like and reflect the possession of abstract syntactic categories. By age 2, English-learning children productively combine auxiliary, negation, and tense categories and syntactically differentiate different negative morphemes.

Acknowledgments

We would like to thank Martin Chodorow, the members of the Hunter College Language Acquisition Research Center, and audiences at BUCLD46, LSA 94, and RALFe 2018 for their helpful suggestions and feedback at various stages of this project.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10489223.2024.2356668

Notes

¹ This proposal assumes that only syntactic heads are available for morphological composition.

² All appendices are provided in the supplementary material, available at https://doi.org/10.1080/10489223.2024.2356668.

³ These items were tagged as a negator in CHILDES, either erroneously by the corpus creators or by the part-of-speech tagger that generates morphological information for CHILDES corpora. We chose to ignore them rather than attempt to infer what the transcriber intended.

⁴ We call these “genuine” negative auxiliaries because Misanalysis proposals have typically considered only don’t and can’t to be monomorphemic due to their prevalence in children’s speech.

⁵ Liz, the child with the lowest percentage of verbs following can’t, produced only 10 utterances with can’t. The 2 non-adult-like utterances were “can’t bricks” and “can’t Mummy.” While the former might be a category selection error, the latter seems like a vocative use of “Mummy.”

⁶ In particular, across both studies children gave a range of responses to the prompt in question involving a number of different negators. When looking at all responses given to the prompt, children produced don’t with an additional tensed item only 4% of the time in both Thornton & Tesan (2013) and Thornton & Rombough (2015).