Melissa Bowerman

STRUCTURAL RELATIONSHIPS IN CHILDREN'S UTTERANCES: SYNTACTIC OR SEMANTIC?1

Publisher Summary nThis chapter reviews the structural relationship between syntactic and semantic in childrens utterances. According to the view of language acquisition, the linguistic knowledge that lies behind childrens initial attempts at word combining may not and need not include information about the basic grammatical relations or the constituent structure they entail. There is, in any event, no compelling evidence as yet that it does. The characteristics of cross-linguistic data suggest the alternative view that children launch their syntactic careers by learning simple order rules for combining words, which, in their understanding, perform semantic functions such as agent, action, and object acted upon, or perhaps other even less abstract semantic functions. Through additional linguistic experience, children may begin to recognize similarities in the way different semantic concepts are formally dealt with and to gradually reorganize their knowledge according to the more abstract grammatical relationships, which are functional in the particular language they are learning.

THE CHILD'S EXPRESSION OF MEANING: EXPANDING RELATIONSHIPS AMONG LEXICON, SYNTAX, AND MORPHOLOGY*

Studies of first-language acquisition typically have shown strong respect for the major components into which linguistic analysis divides language: lexicon, syntax, morphology, and phonology. Thus, researchers explore the acquisition of word meaning (for example), or the characteristics of children’s early word combinations, or the acquisition of inflectional morphemes, but only rarely compare the elements of the child‘s developing linguistic system across the major formal categories. The picture of language acquisition built up in this way is fragmented. We may know a great deal about the development of particular subsystems, but we do not yet have a clear understanding of how the different parts fit together, how they interact and are affected by each other in the course of development. Interrelationships among the components of the child’s developing grammar can be approached in various ways. The most studied problem to date is whether children’s initial rules for combining and inflecting words are bound to particular words or groups of semantically similar words rather than extended across all words of the relevant part of speech (e.g., References 1 and 2). Limited attention also has been paid to the influence of the infant’s phonological system on the “selection” of first words to be l e a ~ n e d . ~ The present paper asks still a third question: Given that the child has a certain type of meaning he wants to communicate, what are his lexical, syntactic, and morphological options for encoding that meaning, and how do these options change and affect each other over time? This question is elaborated in the first section below. Two issues raised there are considered in more detail in the next two sections. Finally, some possible implications of these issues for second-language acquisition are discussed in the last section.

Cross linguistic similarities at two stages of syntactic development

Do all children, regardless of the language they are acquiring, pass through a similar sequence of developmental stages, each of which is marked by the production of particular kinds of constructions? Samples of early speech from American, Finnish, Samoan, and Luo children are compared in an initial investigation of this question. Extensive similarities are found across languages in the kinds of construction patterns produced at each of two periods of syntactic development, called here “Early Stage I” (mean length of utterance 1.30–1.50 morphemes) and “Late Stage I” (mean length of utterance 1.60–2.00 morphemes). These similarities suggest that there are many commonalities in the developmental order in which children in different linguistic communities learn how to combine words to express various relational concepts.

PHONOLOGY AND PHONETICS: GENERAL DISCUSSION

M. STUDDERT-KENNEDY (Queens College, Flushing, N.Y.) : Dr. Cole, if we recognize words so quickly, why do they last so long? R. A. COLE (Carnegie-Mellon University, Pittsburgh, Pa) : According to Marslen-Wilson and Welsh [see Reference 11 in Dr. Cole’s paper], listeners can usually recognize a word in a neutral context after processing the first three or four segments, and recognition is even faster if the word is predictable from prior context. Given this “fact” of perception, why should language contain so many polysyllabic words longer than three or four segments? Note that this question assumes that the duration of words in a language somehow depends upon the time that it takes listeners to recognize words. I do not know of any evidence for or against this interesting hypothesis, but let us assume that there is a causal relationship between the time course of word recognition and the duration of words in a language. The question then becomes, What is the optimal form of this relationship? Your question implies that words should only be as long as listeners need to recognize them. A language of this form, however, would have little redundancy; there would be no “unnecessary” syllables. I suspect that a language of this form would be very difficult to understandlike reading Gertrude Stein poetry, which has little redundancy. Long words, which can be recognized before they are complete, give the listener extra processing time when listening to speech. This time can be used for comprehension, to recover from noisy input or lapses in attention, or to frame a reply. S. BELASCO (University o f South Carolina, Columbia, S.C.) : Dr. Cole, principally what you said about the acoustic cues of word boundaries is true. However, I believe you overstated the case because it is generally known that acoustic cues are sometimes used to differentiate word boundaries. UNIDENTIFIED SPEAKER: Dr. Ferguson, the examples you gave were from speech production. In L1 [first-language] acquisition, children sometimes are able to recognize differences they don’t make in production. Do you find the same parallelism in adult L2 [second-language] acquisition? C. A. FERGUSON (Stanford University, Stanford, Calif.) : This is an important question, the relation between perception and production. Perhaps 85 or 90% of our research has had to do with production, not because we weren’t interested in the very kind of question that you are raising, but because we find it very difficult to contrive experimental