To appear in K. Hirsh-Pasek & R.M. Golinkoff (Eds.), Action meets words: How children learn verbs. New York: Oxford University Press.

Discovering Verbs through Multiple-Cue Integration

Introduction

Before children can ride a bicycle or tie their shoes, they have learned a great deal about how words are combined to form complex sentences. This achievement is especially impressive because children acquire most of this syntactic knowledge with little or no direct instruction. Nevertheless, mastering natural language syntax may be among the most difficult learning tasks that children face. In adulthood, syntactic knowledge can be characterized by constraints governing the relationship between grammatical categories of words (such as noun and verb) in a sentence. However, acquiring this knowledge presents the child with a "chicken-and-egg" problem: The syntactic constraints presuppose the existence of grammatical categories because syntactic knowledge is generally couched in terms of categories of words and not in terms of individual words. On the other hand, grammatical categories have little value in and by themselves, rather they are only useful insofar as they support syntactic constraints. A similar "bootstrapping" problem faces a student learning an academic subject such as physics: understanding momentum or force presupposes some understanding of the physical laws in which they figure, yet these laws presuppose the very concepts they interrelate. But the bootstrapping problem solved by young children seems vastly more challenging, both because the constraints governing natural language are so intricate, and because young children do not have the intellectual capacity or explicit instruction available to the academic student. So how does the child solve the bootstrapping problem in language acquisition? In this chapter, we pursue a possible solution in the form of multiple-cue integration.

By one year, infants will have learned a great deal about the sound structure of their native language (for reviews see Jusczyk, 1997, 1999; Kuhl, 1999; Pallier, Christophe & Mehler, 1997; Werker & Tees, 1999). Thus, when they face the problem of bootstrapping syntax at the beginning of their second year, they are already well acquainted with the phonological and prosodic regularities of their native language. The multiple-cue integration hypothesis suggests that this perceptual attunement provides an essential scaffolding for later learning by biasing children toward aspects of the input that are particularly informative for acquiring syntactic information (e.g., Christiansen & Dale, 2001; Gleitman & Wanner, 1982; and contributions in Morgan & Demuth, 1996; Weissenborn & Hohle, 2001). Specifically, the integration of multiple probabilistic cues derived from the co-occurrence of words (distributional), their sound properties (phonological), and their intonational (prosodic) as well as situational (semantic) context by perceptually attuned general-purpose learning mechanisms may hold the key to how children solve the bootstrapping problem. In this way multiple cues can provide reliable evidence about linguistic structure that is unavailable from any single source of information.

A further initial strategy that the child may bring to bear on the problem of bootstrapping syntax is to focus on the discovery of nouns and verbs, perhaps the most salient groups of content words in that they refer to objects and actions in the environment. Even in this much reduced version, the child's learning task remains formidable given that they are still in the process of making sense of the nonlinguistic world as well. And it is in this context that verbs may be particularly difficult to pin down. Minimally, early verb learning requires that the child masters three different complex learning tasks. Firstly, the child needs to be able to segment fluent speech to locate possible verb forms using distributional, acoustic and other types of language-internal cues (see the other chapters in Part I). Secondly, the child has to be able to find the appropriate parts of actions to be named among unfolding event sequences involving many types of language-external cues (see chapters in Part II). Finally, the child has to learn to integrate language-internal and language-external cues in the service of acquiring the form and meaning of verbs (and other words; see chapters in Part III).

In this chapter, we discuss how the child may accomplish the difficult task of verb learning, focusing on the integration of multiple language-internal cues to verb forms. We first review previous work on multiple-cue integration. We then report results from novel analyses of corpora of English child-directed speech, pointing to different roles for distributional and phonological cues in the learning of nouns and verbs. Finally, we relate the differential roles of cues to differences in semantic support for nouns and verbs in language-external information, and discuss possible implications of our results for the understanding of word learning more generally.


Click here to download a PDF version.