Observations of early morning conversations between twin boys (2;9 at onset of research) indicate that young children are able to sustain a coherent dialogue over a number of turns. Contrary to the views of Piaget (1926), the interlocutors generally attend to one another's utterances. Coherence is achieved to a large extent by attending to the form of one another's utterances. Extended exchanges are maintained by speakers focusing on a sequence of sounds (sound play) or a constituent within an antecedent turn and reproducing it (with or without modification) in the next turn.

Studies of comprehension of the quantifiers all and more are reported. The subjects were children between the ages of three and five. There were two main conditions. In one of these the objects to which the quantifiers related were enclosed in containers which either were or were not filled by the objects. In the other no containers were present. These conditions yielded substantially different response patterns. The relation of the findings to those typically obtained from Piagetian conservation tasks is discussed; and the implications for theories of semantic development are considered.

Two experiments tested the semantic feature hypothesis with children younger than those previously studied. Two predictions were considered: (1) that order of acquisition of dimensional adjective pairs corresponds with their order of semantic generality; (2) that at an early stage of development, the marked members of dimensional pairs are interpreted as having the meaning of the unmarked members. In general, both experiments supported prediction one but failed to support prediction two. In Experiment II, a possible confounding variable (non-semantic size preference) was examined. Two general conclusions were reached: (1) that it is difficult to rule out semantically irrelevant factors (size preference) from experiments investigating the learning of dimensional adjectives; and (2) that refinements of the semantic feature theory may be necessary to accommodate individual differences (which may be age dependent) in strategies for learning dimensional adjectives.

Perception and production data were collected from 28 children, ages 1; 8 to 3; 11, to test four specific hypotheses on the acquisition of initial fricatives and glides in English, based on the assumptions that perception precedes production and unmarked precedes marked. Perception data were collected by the Shvachkin-Garnica technique for ‘phonemic perception’. Pairs of objects are given nonsense names, CVCs differing only in the initial sound; the child is asked to perform certain actions with the named objects and if he is correct on 7 out of 10 trials he is assumed to have demonstrated phonemic perception of the opposition in question. Production data were compared to the perception data. Results only partially confirm the hypotheses and indicate that (a) children as late as age 3; 0 do not have complete phonemic perception, (b) phonemic perception develops gradually, generally in advance of production, and (c) the order of acquisition shows trends toward uniformity but is not universal.

A frequent claim has been made that the acquisition of intonation precedes other language learning. Little is known about developing comprehension of language in general and of intonation in particular. This article reports the result of an experiment testing the comprehension of certain intonation patterns between the ages of seven and ten. It is found that development is still going on during these years and is not wholly complete by the end of this period. Some suggestions are made about the order of acquisition.

This paper presents a proposed strategy of phonology based on acquisition data from two children, one French and one English. Both children showed a process of FRONTING, where the sounds in the child's words are ordered from those produced at the front of the mouth to those produced in the back. While this first appears to be an uncommon strategy, an examination of its implied universals suggests that Fronting at some more basic level may represent a general phonological process. Specifically, Fronting suggests that a marked or unmarked relationship exists between the initial and final consonants of a CVC syllable dependent upon their adherence to Fronting. It is proposed that back consonants are less marked in syllable final position than initial position and vice versa for front consonants. Thus Jakobson's theory, based on markedness in initial consonants, is just part of a more general phenomenon of markedness in syllables. Data are given that suggest that such is the case.

An attempt is made to explain how the child comes to relate his acquisition of the form of language with the categories that he is establishing in the organization of his non-linguistic experience, by hypothesizing a basis for language in pre-linguistic cognitive development. The child's task is seen as being to match the organization of language with the cognitive organization that he has already imposed upon his experience. This is made possible, it is argued, because the organization of meaning within the language system is closely related to the universal categories of early experience. Evidence from recordings of a small sample of children in the early stages of language acquisition is advanced in support of this hypothesis, and the results of the linguistic analysis are discussed in the light of findings from the Genevan school of developmental psychology.

Data on children's errors in marking pronominal case are surveyed and the pattern of errors described: children substitute objective forms for nominative ones but do not commit substitution errors in the opposite direction. It is proposed that two general cognitive principles underlying language acquisition might explain the uniformities in the data: avoid exceptions, and pay attention to the ends of words. It is further suggested that the last of these should be extended in scope, to cover short-term processing of larger constituents.

The following describes a study of the acquisition of ‘pragmatic’ structures by Italian children. PRAGMATICS refers to the study of the use of language in context, by real speakers and hearers in real situations. It therefore fails to meet the definition of ‘competence’ outlined by Chomsky, emphasizing the ideal speaker abstracted from particular situations, and perhaps for this reason has been neglected in psycholinguistic studies. However, there has been a renewed interest in pragmatics in recent semantic theory, and a number of proposals have been made regarding a formal representation for pragmatic structures. Three major positions can be distinguished, all of which in varying degrees separate pragmatics from the propositional content itself, as a set of procedures and assumptions for the appropriate use of propositions. Interpretive semanticists, such as Jackendoff (1972), place pragmatics entirely outside the syntactic component, in a separate and heterogeneous semantic component. Fillmore (1968) and Weinreich (1963) offer proposals for a separate ‘modality component’ which operates on a more articulated semantic structure. For example, in the sentence Could John have hit the ball?, the proposition ‘John hit the ball’ is spelled out in terms of the predicate ‘hit’, the two arguments, and the semantic relations holding among them. The modality component would contain simply a set of unanalysed symbols like QUESTION, CONDITIONAL and PAST PERFECT that are applied to the proposition. A third approach to pragmatics is offered in natural logic models like those of Parisi & Antinucci (1973) or Lakoff (forthcoming). In these models, as contrasted with the above modality component, all the meaning underlying a sentence is broken into minimal elements. Thus the above sentence is described not only with the nuclear proposition ‘John hit the ball’, but with a performative proposition describing the speaker's interrogative intention (i.e. ‘I ask you…’) and with various presuppositions1 describing the conditions that are necessary for the sentence to be appropriate. For example, instead of an unanalysed symbol for conditional, there is an ancillary proposition describing the fact that the central proposition is not necessarily true at time X. This latter approach was of more heuristic value for a developmental study of pragmatics because a representation is available for all the meaning underlying a given sentence. For example, the articulated presupposition ‘Proposition X is not necessarily true at time X’ is more easily translatable into psychological terms than an unanalysed symbol like ‘Conditional’.

Two general classes of particle may be recognized in Japanese: those with syntactic function, and those which express emotions and desires. The former can be divided into three types: CASE (Kakujoshi), CONJUNCTION (Setsuzokujoshi), and ADVERBIAL (Fukujoshi).