Different Paths to Quantifier and Numeral Acquisition

• In Event: Poster Session 06
  In Poster Session: PS 06 Topic - Language, Communication

Abstract

Quantifiers and numerals both describe properties involving sets of individuals (Barwise and Cooper 1981), but the two differ in that the former requires concepts of classification whereas the latter requires cardinal knowledge. The quantifier bootstrapping hypothesis (Carey 2004) postulates that numeral acquisition is developed from quantifier knowledge; as children are not equipped with the concept of exact numerosity at the beginning, quantifiers and numerals are initially undifferentiated. On an alternative account, children are assumed to assign an exact numerosity reading to the number words right from the outset (Gelman and Gallistel 1978; Grinstead et al. 1997).

How the two types of quantity expressions are learned may vary depending on differences in the way they are linguistically represented across languages. In English, for instance, both quantifiers and numerals are placed at the determiner position inside a noun phrase. Previous studies reported that English 3- and 4-year-olds showed different preferences in interpreting the two: an “exact” reading was favored on the numerals but an “at least some, possibly all” reading on the partitive quantifier “some” (Hurewitz et al. 2006; Barner, Chow, and Yang 2009). Cantonese, the language of investigation in the current study, differs from English in two major aspects. First, Cantonese lacks plural inflection but requires an obligatory presence of a classifier between the numeral and the noun, viz., the singularity-plurality distinction is not overtly marked on the form. Second, quantificational notions are predominantly expressed by verb suffixes but not determiners. How the bootstrapping mechanism could account for the acquisition of verbal quantifiers and number words, where the two occupy different structural positions, remains to be ascertained.

The present study investigates the understanding of quantifiers and numerals by Cantonese-speaking three- to five-year-olds. Using the same “Give X” comprehension task (Wynn 1992) on both types of expressions, 69 children were tested on two suffixal quantifiers (saai3 “all” and gam2 “some”) and four cardinal numbers (jat1 “one”, loeng5 “two”, saam1 “three”, and sei3 “four”) embedded in test sentences on a pseudo-randomized order. Children’s responses on the quantifier items were coded as correct if they gave the complete set of eight items for saai3 “all”, and 2-7 items for the partitive gam2 “some” (cf. Barner et al. 2009).

Our results show that saai3 “all” is the best comprehended quantity expression and is much better perceived than gam2 “some”: 86% correct on saai3 vs. 64% on gam2 by the three-year-olds and 100% correct on saai3 vs. 80% on gam2 by the four-and-five-year-olds (Figure 1), consistent with earlier findings that exhaustivity precedes partitivitiy (e.g. Hurewitz et al. 2006). Children behaved like their English peers in favoring an “exact” reading on the numerals, except for the highest number “four” by the three-year-olds (Figure 2). Among the incorrect responses, gam2 is dominated by the “exhaustive” errors (3-year-olds: 56%; 4-and-5-year-olds: 100%) whereas the numerals are associated with “at least”, “exhaustive” as well as other errors. To conclude, our findings lend further support to the view that quantifiers and numerals follow different paths of acquisition.

Author

• Margaret Lei, The Chinese University of Hong Kong
  Presenting Author