Inference and exact numerical representation in early language development

Abstract

How do children as young as 2 years of age know that numerals, like one, have exact interpretations, while quantifiers and words like a do not? Previous studies have argued that only numerals have exact lexical meanings. Children could not use scalar implicature to strengthen numeral meanings, it is argued, since they fail to do so for quantifiers (Papafragou & Musolino, 2003). Against this view, we present evidence that children's early interpretation of numerals does rely on scalar implicature, and argue that differences between numerals and quantifiers are due to differences in the availability of the respective scales of which they are members.Evidence from previous studies establishes that (1) children can make scalar inferences when interpreting numerals, (2) children initially assign weak, non-exact interpretations to numerals when first acquiring their meanings, and (3) children can strengthen quantifier interpretations when scalar alternatives are made explicitly available.

Partial Color Word Comprehension Precedes Production

“…Color words may be unique only in that children's color word errors are more easily observed in naturalistic settings because the perceptual space is more compressed and COLOR WORD COMPREHENSION intermediate exemplars are more frequent. Emerging evidence suggests that children may acquire early partial meanings for a variety of words before they acquire full adult-like meanings, including number words (Barner & Bachrach, 2010;Brooks, Audet, & Barner, 2012;Condry & Spelke, 2008;Sarnecka & Gelman, 2004), time words (Tillman & Barner, 2013), emotions words (Widen & Russell, 2003), and object labels (Ameel, Malt & Storm, 2008).…”

Quantity implicature and access to scalar alternatives in language acquisition

“…The only missing component of implicature is that, typically, weaker items are strengthened by appeal to stronger ones (something that is impossible when interpreting unknown numerals, since stronger words have not yet been acquired). Still, once children acquire a meaning for two, they should be in a position to compute an implicature for one, meaning that even 2-year-olds could compute implicatures to derive exact meanings for numerals (for evidence that this is, indeed, what children do, see Barner & Bachrach 2010). As far as we can see, there is no difference in the processing demands required by these welldocumented inferences and the inferences required by scalar implicature.…”

“…This difficulty has often been attributed to an underdeveloped processing ability or pragmatic competence (see Musolino 2006;Musolino & Lidz 2006;Papafragou 2006;Papafragou & Musolino 2003;Papafragou & Tantalou 2004;Pouscoulous, Noveck, Politzer & Bastide 2007). This paper explores an alternative hypothesis that children fail to compute scalar implicatures because they lack knowledge of the relevant scalar alternatives to words like some (see Barner & Bachrach 2010). At the heart of this conjecture is the thesis that knowledge of scales goes beyond knowledge of lexical meanings and a capacity to compare sentences on the basis of informativeness.…”

“…For example, this account, as noted by Barner & Bachrach (2010), could explain children's ability to assign exact interpretations to numerals, which belong to an explicitly memorized list of scalar alternatives -the count list. As already noted, theories that view scales as emergent properties of semantic meaning are forced to assign numerals a qualitatively different analysis from other quantifiers, since children readily access exact meanings of numerals but cannot strengthen quantifiers used in equivalent contexts.…”

“…In their study, Barner & Bachrach (2010) argued that very young childrenbeginning at 2 years of age -routinely make inferences that are similar in structure to scalar implicatures when interpreting unknown numerals (i.e., numerals for which they have not yet acquired an adult-like meaning). As noted earlier, when a child who knows the meaning of one is shown two sets -e.g., one containing one balloon, and the other containing five -they systematically point to the larger set when asked to find five balloons.…”