Anne Fernald
Stanford University
To understand language, children must learn to listen for meaning in sequences of speech sounds that unfold rapidly in time. Fluent understanding requires "listening ahead", continually anticipating what is coming next using linguistic and nonlinguistic information from the context in which words are spoken. Skilled adults draw on multiple sources of knowledge to process speech with remarkable speed and efficiency, but research on the early development of this critical capacity is just beginning (1-4). We review several recent studies of early word recognition and sentence comprehension using eye-tracking methods for measuring the time course of children's understanding. Efficiency in comprehension increases as the inexperienced listener learns to exploit regularities in the structure of continuous speech, enabling the child to anticipate upcoming words and relate word meanings to one another. Even with limited experience of language and the world, infants can rely on two main sources of knowledge in listening for meaning in speech. The first is their attention to the local redundancies provided by caretakers who use a special form of infant-directed (ID) speech, i.e., who repeat themselves frequently, emphasize focussed words by placing them in perceptually prominent positions, and lighten the load of monitoring speech by presenting new information in predictable formats. The second is their early sensitivity to regularities in the speech they hear that are correlated with structural aspects of the ambient language.
First we show how infants benefit from predictability in speech at a perceptual level. Certain characteristic features of ID speech enable the inexperienced listener to anticipate the onset of focussed words, including (1) the positioning of focused words in sentence-final position, and (2) the high level of prosodic and segmental redundancy across utterances. In two studies with 18-month-olds (n=120), we found that familiar words were recognized more quickly and reliably in final than in medial position, and that the redundancy of the carrier phrase influenced the efficiency of recognition. Next we show how infants benefit from predictability at a linguistic level, as they learn to make use of language-specific cues in understanding. Further research with 18- to 26-month-olds (n=150) reveals dramatic developmental gains in the ability to exploit phonetic, morphological, and semantic cues to word identity across the second year. Even for infants with lexicons of only a few hundred words, the interpretation of sentence meaning is predictive, and linking spoken language to a mental model of the world is an active, continuous process.
References
(1) Fernald, A., Pinto, J.P., Swingley, D., Weinberg, A., & McRoberts, G.W. (1998). Rapid gains in speed of verbal processing by infants in the second year. Psychological Science, 9:72-75.
(2) Fernald, A., Swingley, D., & Pinto, J.P. (2001). When half a word is enough: Infants can recognize spoken words using partial phonetic information. Child Development, 72:1003-1015.
(3) Swingley, D. & Fernald, A. (in press). Recognition of words referring to present and absent objects by 24-month-olds. Journal of Memory and Language.
(4) Swingley, D., Pinto, J.P., & Fernald, A. (1999). Continuous processing in word recognition at 24 months. Cognition, 71, 73-108.