The present paper reports an analysis of correct wh-question production and subject–auxiliary inversion errors in one child's early wh-question data (age 2; 3.4 to 4; 10.23). It is argued that two current movement rule accounts (DeVilliers, 1991; Valian, Lasser & Mandelbaum, 1992) cannot explain the patterning of early wh-questions. However, the data can be explained in terms of the child's knowledge of particular lexically-specific wh-word+auxiliary combinations, and the pattern of inversion and uninversion predicted from the relative frequencies of these combinations in the mother's speech. The results support the claim that correctly inverted wh-questions can be produced without access to a subject–auxiliary inversion rule and are consistent with the constructivist claim that a distributional learning mechanism that learns and reproduces lexically-specific formulae heard in the input can explain much of the early multi-word speech data. The implications of these results for movement rule-based and constructivist theories of grammatical development are discussed.

A mother's ability to identify consistently what she perceives to be communicatively salient behaviours is considered to be fundamental to the process of responding sensitively to prelinguistic infants. The present study investigated the ability of 35 mothers to identify consistently what they perceived to be communicative acts by infants at three ages (0;6, 0;9, and 1;0). Each mother coded the same videotape of her infant's behaviour on two occasions, three months apart, and observed measures of agreement (consistency) between coded records were obtained. A randomization procedure was used to provide distributions of chance levels of agreement between coded records with which observed measures were compared. Comparisons of the mothers' coded records indicated that they were able to identify infants' communicative acts consistently at each age.

Marcus, Pinker, Ullman, Hollander, Rosen & Xu (1992) claim that when the irregular past form of a verb is known, it is immediately known to be the correct form, such that overregularizations only occur as speech errors, not as a genuine grammatical alternative; as a result, they argue, overregularization rates are, when carefully inspected, very low. In the present paper: (1) it is shown that even if overregularizations are a genuine grammatical alternative, overall rates in samples would still be low for most children; (2) careful analysis shows evidence for substantial overregularization periods in three longitudinal subjects ages 2; 5–5; 2 (Abe), 2; 3–5; 2 (Adam) and 2; 3–5; 0 (Sarah); (3) Abe's much higher rates follow from general developments in his past tense acquisition, in ways not consonant with Marcus et al.'s formulations.

This study set out to explore the contexts in which preschool children use private speech, or self-talk, in the naturalistic setting of the preschool classroom, and age-related changes in the contexts in which preschoolers talk to themselves. A total of 2752 naturalistic observations of fourteen three-year-old and fourteen four-year-old children were conducted using a time-sampling procedure in two preschool classrooms over the course of one semester. Results from logistic regression analyses revealed that both age groups were (a) more likely to use private speech during the self-selected activity classroom context as opposed to both large group and outside free play classroom contexts, and (b) most likely to talk to themselves when alone, next likely in the presence of peers, and least likely when in the presence of a teacher. Although the probability of private speech among three-year-old children did not vary as a function of the child's immediate activity, four-year-old children's private speech was more likely to occur during sustained and focused goal-directed activity as opposed to rapidly-changing and non goal-directed activity. The findings suggest that private speech appears systematically in young children and that, in several ways, four-year-old children use private speech more selectively than three-year-olds.

This methodological study in 33 two-year-olds shows that child speech (total utterances, word roots, MLU) occurs at about the same level in different settings (the familiar home vs. the unfamiliar laboratory), but that children speak more and in more differentiated ways with different people (mother vs. stranger). Child speech also shows significant short- term stability. Girls use more different word roots and speak in longer utterances than do boys. In spontaneous child speech, cross-context generalizations appear warranted, but they also depend on conversational partner and gender of child.

Communicative Development Inventories (CDIs) were collected from 669 British children aged between 1;0 and 2;1. Comprehension and production scores in each age group are calculated. This provides norming data for the British infant population. The influence of socio-economic group on vocabulary scores is considered and shown not to have a significant effect. The data from British infants is compared to data from American infants (Fenson, Dale, Reznick, Bates, Thal & Pethick, 1994). It is found that British infants have lower scores on both comprehension and production than American infants of the same age.

In my paper Patterns of Pronoun Case Error (Rispoli, 1998b), I documented two aspects of pronoun case errors arising out of principles of lexical retrieval that are operative during the acquisition of the word specific pronoun case paradigms of English. The two aspects were; (a) the ‘double-cell effect,’ a supercharging of the her for she error rate; and (b) the ‘antagonism’ between the stereotypic objective for nominative replacement (e.g. him for he) and the rarer nominative overextensions (e.g. he for him/his). In my response, I shall address Schütze's criticisms in both of these areas. The response begins by addressing issues surrounding the double-cell effect. It then moves to issues surrounding antagonism.

This paper explores two Mandarin-speaking children's (3;2 and 3;3) ability to refer to the past in mother–child conversation. The approach encompasses morphosyntactic, semantic and discourse-pragmatic perspectives. The results show that the children tend to refer to immediate past spontaneously, but rely heavily on elicitation when referring to earlier past. It is suggested that maternal scaffolding functions as a discourse support for children to participate in conversation involving earlier past. When establishing past reference with overt temporal markers, the children resort mainly to aspect markers. In addition, they also rely on semantic and discourse-pragmatic resources for temporal inferencing, such as inherent semantic aspect, shared background knowledge and situational context.

Review essay on: B. MACWHINNEY (ed.), The emergence of language. Mahwah, NJ: Erlbaum, 1999.

An old joke that has been circulating for the past decade or so goes as follows: a biologist, a physicist, and a cognitive scientist were sitting around discussing the great achievements of their fields. The biologist waxed eloquent about the insights of Darwin's theory of evolution; the physicist expounded on the implications of Einstein's theory of general relativity. Then the cognitive scientist spoke up: ‘Our great discovery is the thermos. You put a cold drink in, the drink stays cold. You fill it with hot soup, the soup stays hot. This is amazing, for how does the thermos know?’

Obviously, this cognitive scientist has asked the wrong question about how the thermos maintains temperatures. In The emergence of language (henceforth, EL), an edited collection of chapters authored by an interdisciplinary group of computer scientists, linguists, and cognitive and developmental psychologists, it is suggested that perhaps language acquisition researchers have been making the same mistake as the errant cognitive scientist. To be sure, language is an extremely complex phenomenon, yet it is also elegant. Recognition of these characteristics in all aspects of language (thanks in large part to Chomskyan approaches to linguistics) highlights a well-known apparent paradox: language is hopelessly complex but children acquire it with ease. Solutions to this paradox have typically inspired researchers to posit rules or other kinds of blueprints – knowledge (innate or acquired) that children have to guide their language acquisition. But are these rules truly necessary? Perhaps children do not KNOW how to acquire language any more than a thermos knows how to maintain temperature.

Sabbagh & Gelman (S&G) present an insightful criticism of the emergentist approach to language acquisition. The analysis takes as its starting point an expressed frustration with the fact that emergentism is not packaged as a single theory or formalism. As a result, S&G decide to focus their critical attention on a particularly strong version of emergentism in which, ‘only domain-general tools are required to account for language development.’ This strong formulation of the emergentist position matches up well with the disembodied connectionism of the 1980s (Rumelhart & McClelland, 1986). However, it misrepresents the richer expressions of emergentism being developed by the authors of this volume. In particular, this ‘strong’ version fails to properly appreciate the degree to which emergentists view cognition as grounded on the body, the brain, and the social situation.

Consider a simple example from phonological development. There is a universal tendency to avoid sequences of nasal consonants followed by voiceless obstruents, as might arise in forms like ‘manpower.’ This constraint is grounded on the facts of speech production (Huffman, 1993) and figures prominently in recent elaborations of Optimality Theory (Kager, 1999). Languages use at least five phonological processes to deal with this problem. These processes include nasal substitution, post-nasal voicing, denasalization, nasal deletion, and vowel epenthesis. Initially, children may apply a variety of these processes (Bernhardt & Stemberger, 1998). Which processes are preserved and which are dropped out will depend on the shape of the target language, whether it be Indonesian, Quechua, Toba Batak, English, or Kelantan Malay. In the terms used by S&G, each of these phonological processes is a small emergentist ‘buzzsaw’ cutting patterns that are shaped not by some innate cognitive ‘blueprint,’ but by the emergent facts of articulatory phonology.

The review article by Sabbagh & Gelman (S & G) on The emergence of language (EL) mentions several criticisms of strong emergentism, the view that language emerges through an interaction between domain-general learning mechanisms and the environment, without crediting the organism with innate knowledge of domain-specific rules, a view that successful connectionist modelling is taken to support. One criticism of this view and the support for it that connectionist modelling putatively provides has been made frequently, and is noted by S & G: it is arguable that connectionist simulations work only because the input to the network in effect contains a representation of the knowledge that the net seeks to acquire. I think it is worth adding to this another criticism that to my mind is a fundamental one, but which has not featured so strongly in critiques of connectionism. A primary goal of modern linguistics has been to account not merely for what patterns we do see in human languages, but for those that we do not. The concept of Universal Grammar is precisely a set of limitations on what constitutes a possible human language. The kind of example used in teaching Linguistics 101 is the fact that patterns of grammaticality are structurally, not linearly, determined: in English we form a yes–no question by inverting the subject NP and auxiliary verb, not by inverting the first and second words of the equivalent declarative sentence, or the first and fifth words, or any number of conceivable non-structural operations. Could a connectionist mechanism learn such non-structural operations? Perhaps I have asked the wrong people, but when I have queried researchers doing connectionist modelling, the answer appears to be ‘yes’. If that's the case, then connectionist mechanisms as currently developed do not constitute an explanatory model of human language abilities: they are too powerful.

In The emergence of language, it is argued that much of what generative linguistics has characterized as ‘rules’ in fact can be derived from domain- general cognitive mechanisms. One example of this perspective is Bates & Goodman's contribution ‘On the emergence of grammar in the lexicon’, which Sabbagh & Gelman say offers ‘…a series of compelling arguments detailing how development and early acquisition shape the subsequent acquisition of new information.’ The thrust of the argument presented by Bates & Goodman is that there is no need to posit the existence of an independent grammar domain, because grammar can be reduced to the lexicon, which in Bates & Goodman's account, can be acquired using general purpose mechanisms. One of several major arguments in favour of this position presented by the authors is that grammar and vocabulary grow at the same rates in child speech. Thus, the authors argue that the development of the lexicon and grammar correlate because there is no grammar outside of the lexicon.

Since Noam Chomsky's famous 1959 review of B. F. Skinner's Verbal Behavior, a linguistically based paradigm for research in first language acquisition has been strong, even in some senses, dominant. Even Piaget, whose main concern was the nature of cognitive development in general, did not deny the essential claim of a linguistically based paradigm for the study of language acquisition: ‘I (Piaget) also agree with him (Chomsky) on the fact that this rational origin of language presupposes the existence of a fixed nucleus necessary to the elaboration of all languages’ (Piaget in Piatelli-Palmarini, 1980: 57). This linguistically based paradigm has led to a developed theory of what it is that the child must acquire when s/he acquires language, and to precise scientific hypotheses regarding the nature of this knowledge. These hypotheses can be, and are being, subjected to empirical test, thus advancing the scientific foundations of the field. In this paradigm, the postulation of a ‘Language Faculty’ in the mind of the human species, and in the child, has allowed the formulation of specific components of linguistic knowledge which are now being tested in language acquisition as well as in grammars of languages of the world.

Sabbagh & Gelman have written an unusually thorough and careful commentary, and furthermore, I agree with virtually everything they say. Most of what I want to do in this commentary (given the space) is to try to situate these issues in a broader perspective.

To start with, I'll assume that my sympathy for statistical analytic processes as an important part of children's language acquisition is well-established (Maratsos & Chalkley, 1980; Maratsos, 1998), even if I think there are qualifications and limits of some sorts that are necessary (Maratsos, 1998). Simultaneously, I strongly agree with the implications of many of Sabbagh & Gelman's remarks: these mechanisms may be variously biased or specially set for various activities, and probably are not the only thing going on, either.