First, third, and fifth graders judged a speaker's belief about entailment and propositional truth in sentences containing five subordinating conjunction-verb combinations: if, because, although, unless combined with indicative verbs, and if with a subjunctive verb. Children were more accurate when a speaker believed the truth of his utterance (because and although) than when he voiced disbelief (if + subjunctive) or doubt (unless and if + indicative). They were also more accurate in judging entailment when the speaker thought the events in the two clauses were positively related (if and because) than when the speaker expressed a negative relation (unless and although). The pattern of errors in judgement conformed to a pragmatic analysis of the conjunctions. First graders based their judgements on semantic cues and did not differentiate propositional truth and entailment concepts. Third and fifth graders understood the difference between entailment and propositional truth and used syntactic cues to differentiate positive from non-positive statements. However, they thought that a subordinating conjunction only determines belief about its own clause, and they did not understand the appropriate cues expressing uncertainty vs. disbelief.

Research has indicated that both the lexical content and the discourse structure of the experimental instructions in traditional conservation tasks enhance task difficulty. Investigations of the comprehension of the relational terms more, less, same and different have revealed significant discourse and task demand effects upon comprehension strategies and cast doubt upon strong versions of hypotheses derived from linguistic marking theory. Five experiments were carried out on children aged between 3 and 6 years. Results were consistent with the following claims: (a) young children do not necessarily understand less to mean more, but do confuse the meanings of same and different; (b) function-based strategies of comprehension are superseded in development by attribute-based strategies for relating discourses to represented referents.

Sixty 3- to 6-year-old children were tested in order to assess certain aspects of their ability to answer questions alluding to hypothetical events. Each child was asked four hypothetical questions from each of four topic areas: (1) past-parent, (2) past-fantasy, (3) future-parent and (4) future-fantasy. Although correct responses increased with age for each of the four categories, children more often responded correctly to future hypothetical questions than to past hypothetical questions and to hypothetical questions involving fantasy characters than to hypothetical questions involving the children's parents. Moreover, the interaction of the past-future and parent-fantasy factors yielded an overall pattern of difficulty for the four topic areas. These findings are discussed in terms of the factors which influence the development of hypothetical reference skills.

The present study concerned the acquisition of double-object, bi-transitive sentences which can be expressed in either BASIC (S-V-DO-to-IO) or TRANSFORM (S-V-IO-DO) constructions. Three-, four-, and five-year-old subjects were tested on comprehension and production of these two alternative forms, investigating the distribution of Basic and Transform sentences in acquisition, and also the effect of prototypicality of the described event on performance. The results showed that the Basic structure was earlier in acquisition, with the Transform appearing later and, at first, only in descriptions of the more prototypical events. Overall, the more prototypical an event, the more competent the children were in comprehension and production and the better able they were to handle the more difficult Transform sentences. Decreased prototypicality resulted in lower performance and greater dependence on the Basic as opposed to the Transform bi-transitives. The findings are discussed within the framework of a general theory of language performance being developed by the first author.

Evidence produced by Chomsky (1969) has suggested that 5-year-old children comprehend the verb tell more easily than the verb ask. This paper analyses the interpersonal context of appropriate use of these verbs and reports the results of two comprehension experiments with 5-year-olds. The first experiment, in attempting to correct an apparent methodological weakness in Chomsky's study, reverses her results and elicits a significantly higher frequency of correct responses to the verb ask than the verb tell: the second experiment demonstrates that this age group can consistently identify the speech acts to which the words ask and tell refer. The results are interpreted as support for the conclusion that 5-year-olds can comprehend the meaning of ask and tell, but that their linguistic comprehension is easily dominated by contextual and interpersonal expectations.

Forty-eight normally developing children, twelve each at ages three, four five, and six years, were asked to define and describe ten concrete nouns expressively with no attendant visual stimuli. Responses were coded according to the semantic relations expressed. Nineteen different semantic relationships were isolated in these definitions, and significant differences were found between the age levels in the various configurations of semantic relational categories represented in the first five relations elicited from each child by each noun. Compatible with recent hypotheses concerning the development of internal semantic memory structures, these differences also proved to be highly predictive of a subject's chronological age. The results suggest that semantic relational analysis may be a productive method of documenting the characteristics of lexical semantic systems of very young children.

The ability to differentiate the directive senses of ask and tell was investigated using a production task in an experimental setting with children between the ages of 2; 6 and 6; 6. Results showed that the children at all ages distinguished the verbs along dimensions sensitive to adult ratings of politeness. There was a trend towards increased politeness with age under ask instructions, but not with tell instructions. Analyses of structural features of the directives produced showed that the interrogative form was more likely to be used with ask instructions, as was the word please, while tell instructions elicited the imperative, with less frequent use of please. The findings suggest that children before the age of 7; 0 are able to differentiate the illocutionary forces of utterances which have the same communicative intention.

This study investigated the development of noun definition by grade school children and adults (N = 100). Subjects were asked to define 15 common nouns in one of two conditions: with objects present or without objects present. Definitions were classified on the basis of perceptual attributes (CONCRETE), everyday usage (FUNCTIONAL) or a combination of these (COMBINATORIAL). Results were analysed with a 5 (Grade) × 2 (Condition × 3 (Category) Analysis of Variance which indicated no effects for condition, but a Grade × Category interaction. With increasing age, there was a shift from FUNCTIONAL to COMBINATORIAL definitions; CONCRETE definitions were infrequent for all grades. In addition, age changes were found in the forms of the definitions subjects produced. In contrast to the children, adults formulated definitions in a more complex Aristotelian form. It was concluded that noun definition is a metalinguistic ability which follows a course of development similar to that for the acquisition of word meaning.

Cambridge University Press apologises for an error in its advertisement in Volume 8, No. 2. The title of Alison Elliot's book is Child Language.