Color word learning is a gradual inductive process

Color word learning is a gradual inductive process Katie Wagner kgwagner@ucsd.edu Dept. of Psychology, UCSD Karen Dobkins kdobkins@ucsd.edu Dept. of Psychology, UCSD Abstract According to some, the moment of abstraction resembles a conceptual epiphany. For example, according to Franklin (2006), “Children seem to struggle with their first color word yet learn most of the other basic terms fairly rapidly over the next several months…. This seems to suggest that there is some kind of ‘switch’ for children’s ability to learn and map color words correctly¬” (p. 324). On this view, once children have mapped their first color word, mapping of other color words to adult-like meanings is relatively simple and fast (see Franklin, 2006; Soja, 1994). The idea that the mapping process ought to be rapid comes from two main lines of research. First, in an often- cited study of color words in 110 languages, Kay and colleagues reported evidence for cross-linguistic universals in linguistic color categories (Kay et al.,, 2009) and argued that the number of color categories cross-linguistically is relatively small and constrained. Second, there is mounting evidence suggesting that pre-linguistic infants possess perceptual color categories very similar to those found in adults (e.g., Bornstein, Kessen & Weiskopf, 1976; Bornstein, 1976; Franklin et al., 2008; Franklin et al., 2005). In each case, the purported existence of constraints on language and perception have led researchers to conclude that color word learning is a simple mapping problem, whereby largely innate perceptual categories are associated with labels provided in language input. Examples of this view are common in the literature, with important consequences for how color word learning is studied. For example, according to Pitchford and Mullen (2004), “Developmental studies have shown young children's perceptual colour space is organized in a similar manner to that of the adult… Thus, when children engage in the learning of colour terms, they already possess colour percepts on which colour concepts can be mapped.” (p.53) The implication of such arguments is that, because color words can be mapped to pre-existing perceptual categories, the lag between production and adult-like understanding must not be due to the problem of determining boundaries. Instead, the delay must be due to the prior problem of identifying color as a domain of linguistic meaning. There are good reasons, however, to believe that the acquisition of color words is not a simple mapping problem. Despite being restricted by universals of human perception, languages vary both in the number of color words they have (2 to 12) and how these words encode color (Kay et al, 2009). For example, some languages that have four basic color terms mark a boundary between red and yellow (e.g., Culina, spoken in Peru; Waorini, spoken in Ecuador) whereas others do not (e.g., Chacobo, spoken in Bolivia; Mura-Piraha, spoken in Brazil; Kay et al., Most current accounts of color word acquisition propose that the delay between children’s first production of color words and adult-like understanding is due to problems abstracting color as a domain of meaning. Here we present evidence against this hypothesis, and show that, from the time children produce color words in a labeling task they use them to represent color. In Experiment 1, an analysis of early color word production errors finds that, before acquiring adult-like understanding, children make systematic hypotheses about color word meanings, which are best characterized as overextensions of adult meanings. Experiment 2 analyzes comprehension errors and finds that these overextensions result from overly broad categories, rather than a communicative strategy. These results indicate that the delay between production and adult-like understanding of color words is largely attributable to the problem of determining language-specific color boundaries. Keywords: Concepts and Categories; Acquisition; Cognitive Development David Barner barner@ucsd.edu Dept. of Psychology, UCSD Language Introduction Color words like red, green, and blue pose a difficult problem to children learning language. According to early reports from the turn of the 20th century, children did not acquire the meanings of color words until as late as 8 years of age. Recent reports suggest that children now acquire color words earlier around 3 or 4 years of age (possibly due to early education, see Shatz et al., 1996), but nevertheless struggle to learn them (e.g., Backscheider & Shatz, 1993; Sandhofer & Smith, 1999). The primary evidence of children’s difficulty is that, similar to the domains of number (Wynn, 1990) and time (Shatz et al.,2010), children produce color words well before they use them with adult-like meanings. Also, it’s often argued color word use is initially “haphazard and inconsistent” (p.70, Pitchford & Mullen, 2003). By most current accounts, this delay between production and adult-like understanding is caused by a difficulty abstracting color as a dimension of linguistic meaning (e.g., O’Hanlon & Roberson, 2006; Kowalksi & Zimilies, 2006; Sandhofer & Smith, 1999). Here we present evidence that children’s initial use of color words is in fact systematic rather than haphazard, and that children have abstracted color by the time they begin using color words. We argue that the main source of children’s delay is the problem of inferring category boundaries for color words. Current accounts of color word acquisition typically assume that once children have conceptualized color as a domain relevant to word meaning, the mapping of color words to their target color categories proceeds quickly.