Development of Reality-Status Semantics in Early Childhood

Abstract

The distinction between reality and various forms of non-reality, i.e. apparent, imaginary, counterfactual or fictional objects or events, is central to many domains of cognition (Flavell, 1988; Clark, 1996; Harris, 2000; Gelman, 2003). Reality is an abstract and broadly applicable concept; for instance, the word real is used to express different types of information in statements like Unicorns are real, This is a real diamond and The patient's life is in real danger. Despite its semantic complexity, this domain is invoked by children's language early in development. By three years of age children use such expressions as real/really, pretend/make-believe, looks/sounds/is like to distinguish reality from non-reality (Woolley & Wellman, 1990; Woolley, 2002), though their construal of these terms' meaning may not be adult-like (Lillard 2001; Hansen & Markman, 2005). Despite long-standing interest in children's developing conception of reality vs. non-reality, their use and comprehension of the relevant vocabulary have not been directly addressed. The present study examines spontaneous use of common reality-status terms by young children and their parents in everyday conversation. Our main hypothesis, drawing on the cognitive-semantics analysis of polisemy (Lakoff, 1987) and on the conceptualchange account of cognitive development (Carey, 1988), is that children's conceptualization of reality and non-reality undergoes progressive elaboration and differentiation, which is reflected in their use of reality-status vocabulary. We present a semantic analysis of reality-status terms and posit predicate modification as these terms' common function. For instance, by saying This is like an orange or This is a real orange speakers convey information about the strength and scope of the current application of the predicate is an orange, analogous to the information conveyed by hedge terms (e.g. John is somewhat/definitely/extremely tall; see Lakoff, 1972). From this analysis we derive empirical predictions about the developmental trajectory of young children's use of reality-status vocabulary, including the order in which different words and different word senses enter children's lexicon. We predict that children should initially use reality-status expressions as simple intensifiers/detensifiers and subsequently express more specific information about the nature of modification, e.g. fiction vs. reality, apparent vs. actual kind membership, reality vs. representation of reality, etc. Our predictions are then tested against a longitudinal corpus of spontaneous parent-child conversations, obtained from 64 families. Despite considerable variation in how early and how often children produce reality-status terms, these data reveal a common pattern in line with our predictions. We find furthermore, that while the frequency of parental use of reality-status terms predicts the overall timing of children's acquisition, it can not account for the order in which different senses of these terms are expressed in children's speech. We argue that these findings support our semantic analysis of reality status. Taken as a case study of children's acquisition of abstract vocabulary, our findings suggest that it involves stepwise structuring and augmentation of conceptual frameworks by children using input from adults as evidence. At the same time, our study exemplifies the potential utility of child language data as empirical evidence on semantic analysis of polisemy.