| Abstract | Language lies at the heart of human cognitive and social development. Infants, who are by definition “without language”, become speaker-hearers of particular languages within their first few years through the experience with the speech of their caregivers and other significant people in their environment. The foundation for the emergence of language proper is the infant’s discovery of sound-meaning correspondences in the utterances produced by those significant people. Social and physical contexts provide support for the semantic meaning of an utterance, although determining the specific referent of an unknown word from nonlinguistic context alone may be no simple task (see Quine, 1960). The present discussion, however, focuses on the other side of the sound-meaning relation - the sound pattern itself. It is still far from clear how the infant comes to recognize in the stream of connected speech the sequence of consonants and vowels that may underlie the diverse pronunciations of a given word in different sentences, by different speakers, and under different speaking conditions (e.g. in rapid casual speech versus slow, exaggerated infant-directed speech). Presumably, these accomplishments are built on the infants’ prior abilities to discriminate and classify the audible properties that correspond to various levels of organization in speech, for example, consonants and vowels (phonetic segments), rhythmic stress patterns, prosodic phrases, and so forth. |