Somewhere between birth and maturity the highly structured linguistic system characteristic of the adult mind evolves. While young children commonly group words idiosyncratically or according to a thematic principle, adults prove more homogeneous in obeying conceptual categories. A gradual transition between two age extremes and their modes of organizing language can be mapped in detail. The several experiments presented in this volume are designed to tap the growth of appreciation of relations among 20 specific words, though inferences drawn from the data may be applicable to much of the lexicon. Where empirical evidence to date is scant in the area of semantic evolution, Jeremy Anglin establishes the hypothesis that development proceeds from the concrete to the abstract. As tools to survey the architecture of cognitive capacities, he adopts and adapts traditional verbal learning and psycholinguistic techniques and applies them to subjects ranging in age from 7 to 26 years. The experiments and their findings are described clearly enough to be readily understood by the general reader.
A blend of preconceptions and perplexities gave rise to Dr. Anglin's experiments, which included sorting, free-recall, free-association, and concept-formation tasks. Four biases concerning the nature of words governed selection of the tasks, the set of words, and the methods of analysis. First, the word contains meaning, which can be identified in part with the features or criterial properties associated with it. Next, words cohere in a system in which many features can be organized hierarchically or in theoretical nests. Third, the meaning of a word is often derived from the contexts in which it occurs, and similarity of meaning among words relates to privileges of occurrence within the same context. Finally, the word, as a social tool whose function is to communicate, is useless unless it means the same thing to different members of a linguistic community. All the verbal tasks chosen could be administered to children with ease, and words selected were presumed to be within the vocabularies of the youngest subjects tested.
Major findings of most tasks supported the concrete-abstract progression. A persistent puzzle for proponents of this theory, however, is the ubiquitous phenomenon that children of about 4 years of age can spontaneously speak the language, giving parts of speech their proper grammatical treatment, despite the fact that young children in the various experiments consistently ignore the form class distinctions that are so important to adults. Employment of such principles and cognizance of them may reflect very different abilities. Semantic development appears to be an extremely prolonged process which may never be complete.
Patterns of interrelations revealed among the host of experimental methods suggest that many results may reflect different aspects of the same underlying cognitive capacities. Dr. Anglin's novel application of multidimensional scaling procedures to meaningful psychological data provides a graphic illustration of the various stages of the growth of the subjective lexicon.
MIT Research Monograph No. 63