The role of child-directed speech in language acquisition: a case study

Abstract

This study examines the nature of child-directed speech (CDS) from the perspective of functions [M.A.K. Halliday, Learning how to mean: Explorations in the development of language, Elsevier North-Holland, Inc., New York, 1977] and social interactionist theory. It is argued that previous explanations of CDS, often called motherese or caregiver speech, have either minimalized or neglected the functionalist–interactionist dimension of input in language acquisition. Far from being merely a novel way of describing the language caregivers use with infants, CDS is presented as a crucial catalyst in the complex process of L1 acquisition.

At the heart of CDS is negotiation between caregiver(s) and infant. The infant need not always respond with complete or near-complete linguistic units or constituents such as an adult might during a given negotiation, yet the context of the negotiation remains crucial to the infant. As physical maturation increases and the infant begins to produce more adult-like utterances, the negotiation between interlocutors becomes more balanced, syntactically and phonologically, but not necessarily semantically/functionally.

This paper presents the results of a case study which specifically examines the utterances or input which family members direct at a Japanese infant during the early part of his language development. The data generated by the subject and his parents provide an interesting glimpse into one of the ways in which infants absorb language. The results of the data analysis show that while the parents of the subject were seen to use roughly equal amounts of language with the child, the distribution of language functions used by the mother was importantly different from that used by the father; therefore, it is suggested that this difference in CDS aids the language development of the infant by providing more interactive negotiation, which is argued to be the crucial factor in language development.

Introduction

On the face of it, the following short piece of actual dialogue would seem to contain an ordinary exchange between a mother (M) and father (F) and their young child (H). That it is in Japanese or that its contents relate to mundane events in the life of the child are not the especially noteworthy parts, except, perhaps, to suggest the universality of such parent–child exchanges. However, whether this language exchange contains components crucial to the child’s language acquisition process is extremely important.

Superficially, this kind of language interaction, which is replicated daily by parents and their children around the world, seems unremarkable, even meaningless. However, it is anything but meaningless for it contains utterances which urge this child into action, question him, and confirm things for him, all of which are rather complex concepts. Parents use language to help reveal the world to their children. However, infants are not born with adult-like language competency with which they can comprehend the meaning of the language their parents direct at them. So, where, one may logically ask, do they get that ability? Though it is likely that children are born with some sort of innate capacity for language acquisition (the nature of which is only theorized at present) which might play some role in the language learning process, we know that normal infants successfully manage to acquire the language(s) of their environment. Exactly how infants become such skilled manipulators of a communication tool as complex and nuanced as language, and to do it within a relatively short time frame, still remains a fascinating riddle without a completely satisfying solution.

Chomsky (1988, p. 3), near the outset of his Managua Lectures, states that

[a] person who speaks a language has developed a certain system of knowledge, represented somehow in the mind and, ultimately, in the brain in some physical configuration. In pursuing an inquiry into these topics, then, we face a series of questions, among them:

• What is the system of knowledge? What is in the mind/brain of the speaker of English or Spanish or Japanese?

• How does this system of knowledge arise in the mind/brain?

• How is this knowledge put to use in speech (or secondary systems such as writing)?

• What are the physical mechanisms that serve as the material basis for this system of knowledge and for the use of this knowledge?

Chomsky’s second question, which deals with the thorny issue of acquisition, necessitates direct, empirical study of child or first language (L1) learning in order to help explain the ‘system of knowledge’ mentioned in his first question. Yet, to approach any sort of answer to the question of how language is acquired, an examination of how and for what purposes that language comes to be used by the infant (Chomsky’s third question) should also be conducted.

To begin to answer this multi-faceted L1 acquisition question and how the acquisition relates to L1 usage, one must ask, ‘Assuming that a language acquisition system of some sort exists within the mind of an infant and is operating normally, how does it work?’ In the simplest of terms, the function of human language is to encode and decode signals that are sent between individuals. These signals can range from an infant’s simple expression of pain or pleasure to a teacher’s explanation of a sonnet. Typical production and comprehension of a language (the potentially infinite set of linguistic signals and functions of that language) assume an intelligence which creates, directs, and understands specific linguistic output, and which can accurately decipher and appropriately respond to linguistic input.

Chomsky’s questions, especially the acquisition question, have long been considered either from a biological perspective or, once an infant is old enough to begin producing ‘language’ which somewhat resembles typical adult forms, from a developmental perspective, both of which are influenced by the environment. It is almost a truism that the theory of acquisition one adopts and the method of studying L1 acquisition one chooses determine to a great extent the types of questions one asks and the way one examines the data gathered from such questions, and so, theories of language acquisition abound. Piper (1998, pp. 141–164) provides a succinct overview of the principles, strengths, and shortcomings of behaviorist, nativist, cognitive, and social interactionist theories of language acquisition. Even though each of these general theories has strengths and weaknesses, the one that I believe provides the greatest insight into the process of language acquisition and, therefore, the one on which I base the present study, is social interaction theory.

More will be said about this theory later, but for now a short quote from Piper (1998, p. 161) sums up my motivation for choosing social interaction theory as a basis for examining L1 acquisition.

The question that is of primary interest in language acquisition theory is how children acquire the ability to express their intentions or meanings in language. [Social] interactionists believe that they do so through a process of negotiation with their mothers or principal caregivers [emphasis added]. This negotiation occurs partly as a result of mothers treating children’s speech, even if it is babbling, as meaningful and intentional [emphasis added].

Investigations of L1 acquisition which focus on biological aspects and which are unquestionably of great importance, neglect, or at least minimize, what I believe to be the crucial aspect of that acquisition: namely, the interaction (or in Piper’s terms, the ‘process of negotiation’) of the child and mother (or primary caregiver), arguably the most important person in the infant’s environment. It is this interaction, and its concomitant facilitation of the development of functional language use, which demands the closest scrutiny and, therefore, is the aspect of L1 acquisition on which I will focus.

The research on child language acquisition has examined from a variety of perspectives the language that infants produce. If, however, we assume that infants do not learn language in a vacuum, an examination of the infant’s linguistic environment seems logical and appropriate. In a different but related area of study, second language acquisition (SLA) research, a great deal of work has focused on the input to which learners are exposed (see Gass, 1997, Gass and Selinker, 2001; for comprehensive and insightful discussions of issues regarding input). The implication of this vein of SLA research is that the input to which second language learners are exposed is a very important component in determining the output that they produce in the target language. That this should be even truer for first language acquisition is almost too obvious to mention. Nevertheless, this specific area of L1 research, the linguistic input directed at infants by their parents, has not been examined from the perspective of what language functions it may contain and how those functions may affect L1 acquisition.

In this study I attempt to answer, at least in part, the question of ‘how the system of knowledge arises in the mind/brain.’ Specifically, I examine L1 acquisition through a study of the interaction between a young child and his family in order to discover how it might be that infants come to understand the relationship between the phonetic and semantic net which is thrown over them in the first few years of life and the world which that net represents. First, I will examine what research into L1 acquisition says about how language acquisition is believed to take place. Next, I will present what has been said about one of the most crucial factors in that acquisition: child-directed speech (CDS—variously referred to in the literature as ‘infant-directed speech, parentese, caretaker speech, nursery talk, nursery language, and caregiver speech’ (Cattell, 2000, p. 104)) and its contribution to the language learning process. Then, after having presented the research plan for this study, I give the results of this study with their analysis. Finally, I will explain why previous analyses of the L1 acquisition by infants are incomplete and will propose a more appropriate perspective from which to view the relationship of CDS and L1 development.

For this study I examine the interaction between an infant and his environment in the context of social interactionist theory and using the taxonomy of language functions proposed and defined by Halliday (1977). Halliday’s work posited that the language children use contains functions which show what children do with language. These functions, which Halliday believed to be present in the child’s output system, do not, of course, appear fully formed and functioning at birth. The functions (and of course the language which is used to convey them) must have developmental roots within either the children themselves or the environment, or perhaps both. Although Halliday examined a child’s linguistic output system, it is reasonable to ask what influence the environment, in other words, the input, has on the development of that output system. If the environment plays any role at all in the linguistic development of an infant, presumably the parents or primary caregivers are responsible for a substantial part of the input needed for the infant’s linguistic development to commence and then flourish.

But, recalling Chomsky’s question of how the system of knowledge arises in the mind/brain, we may ask ‘What is the nature of the input, the “linguistic net,” which is thrown over the infant? Does it contain some or all of the types of functions seen in the output that Halliday has proposed? If so, which functions appear and in what concentration? Can we discover any sort of relationship between the input and the output?’ If this functional language ‘net’ does indeed exist, it should be detectable in the language spoken to infants by their family members or those with whom they have sustained contact. It is perhaps reasonable, then, to suppose that CDS may be more complex and play a more crucial role in acquisition than has been thought.

By examining the kind of language parents or caregivers use with infants according to the functions Halliday saw in the output of his subject, it may be possible to discover a better and more appropriately quantifiable way of describing CDS than has been available before. It is logical to believe that normal adult speakers have the ability to use their language for any and all functions that exist in a given language with other adults; however, one might assume that there is some observable and quantifiable variation among parents or caregivers in their use of functional language when they interact with their young children. But because there are no set rules about what sort of language people must use with their children, an empirical examination of just what they actually do is necessary. In an attempt to do this, I believed that recording the language and interactions in which an infant and his parents engaged would provide the raw data from which a clearer picture of CDS might emerge. Since much of the literature in the field of L1 acquisition examines the relationship between infant and mother, I hypothesized that it would be in that dyad where functional language use and L1 development would be most easily observable.

Since mother–child interactions would likely be the most profitable for investigation (Bloom, 1993, de Boysson-Bardies and Vihman, 1991, Snow, 1995), I also surmised that this relationship would produce results that were quantitatively very different from that of the other relationships the child would have with other family members. Based on my understanding of Hallidayan functions (which will be explained below), I presumed that the interactional function would be the most important and prominent and that the regulatory function would also be prominent in the mother–child interactions but to a lesser degree. Putting all of these beliefs, assumptions, and guesses together, I formulated and proceeded to test the following related hypotheses:

• The CDS used by the mother with her child in this study will be quantifiably different in content and quantity from the language that other family members use with the subject.

• Though all functions of language may be present in the linguistic input to which the child is exposed, a specific subset of those functions, the interactional and regulatory functions, will dominate the interactions between mother and child and will be less prominent in the language uttered by other family members.