hand, the "recency" effect in human serial learning is an acquired tendency to pay particular attention to material in final position, a tendency acquired because sentences are so constructed as to place in final position words carrying a lot of information. Perhaps it is differential stress that selects what the child will reproduce, and sentences are nicely adapted to this predilection in that the heavier stresses fall on the less predictable forms. Or perhaps it is some combination of these ideas.
This paper began with an argument that the correct English sentences produced by a child are not good evidence that he possesses construction rules since we can never be sure that a correct sentence is not directly copied from a model. It seemed to us that systematic errors and manipulation of invented words were better evidence as a child is not likely to have had exact models for these. In the second section of the paper we discussed techniques for inducing construction rules or a generative grammar from the child's natural speech. Since this speech, for Eve at least, is not good English, it can be argued that she had no models for it, and so it is legitimate to infer rules from such data. In the third section, however, we have seen that child speech can be rather well characterized as a systematic reduction of adult speech, and so, after all, there were models for Eve's sentences. She could have learned most of them by selective imitation if not by imitation per se.
Eventually children must do more than imitate and memorize if only because there is not enough time for them to learn as particular verbal responses all the sentences they will be able as adults to produce and evaluate grammatically. (For a detailed statement of this argument see Bruner, 1957, p. 156; or Miller, Galanter, and Pribram, 1960, pp. 146- 147.) In addition to the logical argument that children must learn construction rules in view of their terminal linguistic achievement, there is much empirical evidence that children older than Eve do, in fact, learn construction rules. Some of this evidence is available to every parent in a child's systematic errors (sheeps, I bringed, etc.), and some of it has been collected in a controlled fashion by Berko ( 1958) and Brown and Berko ( 1960).
While children must and do eventually induce construction rules, it is not necessary that they do so from the very earliest age at which words are combined. Eve, after all, is not yet prepared to produce an infinite set of sentences, nor, so far as we know, is she able to distinguish all grammatical sequences from ungrammatical sequences. It is possible that for the earliest linguistic accomplishments one sort of learning theory will serve--a theory developed largely from the study of animal be-