Child-directed speech. Modifications in linguistic input to children and their possible functions.

Table of Contents

I. Introduction

II. Some Remarks on Terminology

III. Grammatical Aspects of Child-Directed Speech
III.1. Phonetic and Prosodic Aspects of Child-Directed Speech
III.1.1. Phonetics of Child-Directed Speech
III.1.1.1. Principles of Phonological Modification
III.1.1.2. Functions of the Phonetic Features
III.1.2. Prosodic Features in Child-Directed Speech
III.1.2.1. Principles of Prosody in Child-Directed Speech
III.1.2.2. Functions of Prosodic Features
III.2. Lexical Components of Child-Directed Speech
III.2.1. Principles of Word Use in Child-Directed Speech
III.2.2. Functions of Lexical Components in Child-Directed Speech
III.3. Syntactic Aspects of Child-Directed Speech
III.3.1. Simple and Complex Syntactic Strucures in Child-Directed Speech
III.3.2. Redundancy Features

IV. Non-Verbal and Pragmatic Aspects in Interaction with Language-Learning Children
IV.1. Context in Early Interaction
IV.2. Conversational Aspects in Early Interaction
IV.3. Non-Verbal Signs in Early Interaction: Gestures, Smiles and Gaze-Coupling
IV.4. Concluding Remarks on Non-Verbal and Pragmatic Aspects

V. Conclusion

VI. Bibliography

I. Introduction

Anyone who has seen and heard a child speak his/her first words and who has noticed how much language determines the progress of a child in becoming independent of his/her parents and in mastering interaction with others must have wondered why it is that children can fulfil such an obviously complex task in such a surprisingly little time with seemingly little effort. It appears almost impossible that the child can achieve so much all by himself/herself. Thus, quite frequently, people consider the linguistic input that is available to the child as a crucial factor for the child’s language acquisition: “It is no wonder that he has learned to speak so fast, with so many people around speaking to him and showing him how to speak!” is an utterance I heard quite often when my son was at the age when he started producing his first words. But did I really help him to learn language? Of course, I noticed my own speech modifications as well as that of other people when conversing with him. But does this special register, which nearly everyone (at least in our western culture) seems to adapt when adressing children, really make a difference? Do we even teach our children to speak? I will not attempt to answer these questions in the present paper, as much more research is needed in this area. However, I want to give an overview on the different features of the language used when adressing children. It will be shown on which principles of modification these features are based and which functions they might serve for the child. Moreover, it will be argued that some features of the speech to children may possibly facilitate language acquisition, but that the language used to them is not necessarily simplified language.

II. Some Remarks on Terminology

Before examining the different features of the language that is used when talking to a child, one has to consider a much more basic, yet in my view elementary problem: What exactly are we talking about when we speak of Motherese as compared to other terms used to label this register? What differences – if any – are there between Motherese, Baby Talk, Child-Directed Speech or Caregiver Talk. Do all these terms imply the same concept? When having a look at the literature, one might even get the impression that the labelling of the phenomenon seems to be merely a question of taste. It may seem pedantic to insist on a clear and reasonably defined terminology; but, due to the amount of labels which all seem to designate the same phenomenon - though they actually do not, as will be shown -, any analysis of this register which does not consider the various terms and their use in psycholinguistic literature and which does not give a clear definition of its subject matter is bound to fail in that its results will remain vague.

When linguists became interested in the nature of linguistic input to young children in the 1960s, the common term to describe this special kind of input was Baby Talk. This term is nowadays often felt to carry “adverse connotations of triviality” and, thus, is considered inadequate to signify an input which is said to serve a “very positive purpose” (Cattell, 2000: 104). Nevertheless, the reason for choosing this term may become clear when one considers the circumstances of its emergence: it was not designed to describe a closely-analysed register of one language, but rather a newly discovered sociolinguistic phenomenon, which seemed to occur in several languages:

“Anthropological linguists and fieldworkers were publishing papers in the 1950s and 1960s about the phenomenon of ‘Baby Talk’ as a special register of the languages they had studied. They documented Baby Talk primarily as an interesting sociolinguistic phenomenon, but often reported as well the local beliefs that it aided children in learning to talk.”(Snow, 1994: 3)

Thus, the term Baby Talk was not coined by psycholinguists who were interested in research on linguistic input to children learning a first language, but by sociolinguists or anthropological linguists gathering information about the different registers in a language. What they found out was that people change their way of speaking when adressing a baby or a child, and consequently they called this register Baby Talk.

Of course, one may argue that the term is ambiguous in that it might refer to adults adressing babies, but also to the special kind of ungrammatical and phonologically reduced speech of little toddlers. (cf. Crystal, 1992: 37) The Oxford Companion to the English Language even gives this meaning of the term as first choice. Baby Talk is described as the “kinds of speech used by small children. (…) In the utterances of young children there is little grammar, vocabulary is idiosyncratic, and pronunciation immature (…)”. (MacArthur, 1992: 101) One example of the two-fold use of the term Baby Talk to refer to both of these at least conceptually different registers at a time - without explicitly calling attention to the ambiguity of the term - can be found in Crain & Lillo-Martin (1999). While they use the term to describe how “(…) toddlers or young children (…) speak in only simple sentences, and use incorrect grammar (…)” (p. 5), only some pages later do they state that adults “(…) talk differently to children than to other adults, using what is sometimes called ‘Motherese,’ [sic] or ‘baby-talk’.” (p. 14) In fact, Clark & Clark (1977) use the term with yet another meaning. They refer to only one specific feature of adult talk to young children as Baby Talk, namely to the “(…) ‘baby talk’ words that are considered appropriate in talking to very young children. For example, adult speakers of English often replace the words for animals by words for their sounds – miaow, woof-woof – or by a diminutive form of the adult word – kitty, doggie. (…)” (p. 322) Interestingly, however, they consider these words “(…) to be modeled on the sounds and combinations of sounds that young children tend to produce when trying their first words (…)”, thus establishing a certain semantic interdependence between the two notions of the term Baby Talk. Similarly, the Oxford Companion to the English Language claims that “[a]dults speaking to small children adopt simplified grammar, special vocabulary, and exaggerated intonations (…)”. (Macarthur, 1992: 101) Adult Baby Talk is modeled on babies’ first verbalisations, but is most probably also taken up as a model by the child. In general, it seems that, unless a clear definition of the ambiguous term is provided within the respective paper or work, as in Ingram (1989: pp. 131), Baby Talk is not an adequate term to describe the special register used when adressing young children or infants. (cf. Crystal, 1987: 235)

Although we have seen that there are still authors who have decided to stick to the term Baby Talk (see also: Finegan, 1994: pp. 449; Cruttenden, 1994: pp. 135; Holzman, 1997: pp. 144), its ambiguity and the fact that it, as mentioned above, connotates triviality probably led to a preference for “(…) various alternative terms (…), such as child-directed speech (CDS for short), infant-directed speech, parentese, caretaker speech, nursery language and caregiver register.” (Cattell, 2000: 104) Due to the fact that most of the analysed linguistic input to language-learning children was provided by mothers, another term emerged, which can still be found in many papers published in this field today:

“Many recent investigations do reveal that mothers’ speech differs from speech among adults (…). Similar effects are apparently found for any speaker addressing a very young child (Shatz & Gelman, 1973; Sachs & Devin, 1976). This stylistic variant of every day speech is systematic enough to deserve its own name; we have called it ‘Motherese’.” (Newport et al., 1977: 112)

While “’Motherese’ is a term that has historically referred to the prosodic exaggerations that are typically found in mothers’ speech to their infants and young children (…) [, to the fact that] when mothers talk to their infants, the melodic, rhythmic qualities of maternal voices increase (…)” (Cooper et al., 1997: 477) (cf. Fernald et al., 1989; Fernald & Simon, 1984; Papousek & Hwang, 1991), it is nowadays mainly used to refer to the complete register of mothers when adressing infants. According to Crystal (1992: 258) it includes “(…) such features as short sentences, repetitive discourse, simplified vocabulary, and expressive intonation (…)”.

Since research on fathers’ language input to children showed that there were some similarities to the maternal speech style, some linguists argued that Motherese could not be the appropriate term. (cf. Ingram, 1989: 131) Instead, more general terms, like caretaker speech, caregiver register or even parentese were invented. (cf. Crystal, 1992: 258; Crystal, 1987: 235; Cattell, 2000: 104) Again these terms were questioned when more research showed that not only adults adopt a certain speech style when adressing young children, but also siblings.

“Within Western middle-class culture, members of the family other than the mother are beginning to be investigated. Early research seemed to show that fathers and siblings from Western middle-class families adjust their speech to language-learning infants in many of the same ways as mothers in terms of higher vocal pitch, restricted vocabulary, and shortened sentence length – implying that what the child learns from them is similar as well.” (Barton & Tomasello, 1994: 111)

However, although there were many similarities concerning the input provided by mothers, fathers and children, some important differences could also be found between the language directed to children by these groups. Barton & Tomasello continue that “[f]athers and siblings do not adapt their conversational interactions for young children (…) nearly as much or in as many ways as do mothers from the same families.” Thus, the question of terming the linguistic input to young children may simply be a question of putting emphasis either on the similarities or on the differences that exist between the various kinds of input. For many publications, however, over-extension of the term Motherese, that is, the fact that fathers’ or siblings’ input is subsumed under the same term, does not even occur, as they are exclusively concerned with maternal speech styles and do not consider other forms of linguistic input. In these cases, of course, there can be no objection to the use of the term Motherese. The problem, in my view, lies in the comparability of findings concerning maternal speech styles with research on other forms of input.

Before I come to the core subject of this paper, I want to demonstrate yet another problem concerning terminology. Research on language input to young children has undeniably proven that, whatever name we give to this kind of register, it is not a fixed and stable one which shows the same features at any given stage of the child’s language acquisition process. As Brown puts it in his introduction to Snow & Ferguson (1977: 13), it rather seems that the “(…) mothers’ speech is fine-tuned to the child’s [sic] psycholinguistic development (…)”. Mothers and probably other persons interacting with young children, too, adjust their linguistic input to the psycholinguistic capacities of their children to ensure successful communication. “What adults are chiefly trying to do, when they use BT [i.e. Baby Talk] with children, is to communicate, to understand and to be understood, to keep two minds focussed on the same topic.” Whatever reasons there may be for the fine-tuning of input to children is not important for the point to be made: linguistic input to newborns is different from the language used to 2- or 3-year-olds in several respects. Consequently, we need to raise the question whether a single term, such as Motherese or Baby Talk, suffices to describe these various stages in linguistic input to children. The following table has been adapted from Grimm (1998: 747) and describes maternal adjustments as they are made during three developmental stages of the child. Unfortunately, Grimm uses the terms Baby Talk and Motherese to refer to different stages in this development. This, again, might be a source of confusion and is not consistent with other literature on maternal speech styles.The problematic nature of both these terms has already been discussed. Interestingly, Grimm refers to the early stage, in which prosodic features are of major significance, as Baby Talk, which is contradictory to the statement put forward above that Motherese has historically referred to these features. (cf. Cooper et al., 1997: 477) It might have been better to simply use the terms early, mid-stage and late child-directed speech (or similar terms) to avoid the misleading terms Motherese and Baby Talk and to indicate the differences between these stages without being too specific about their nature.

Table 1: Maternal speech styles (adapted from Grimm, 1998: 747)

illustration not visible in this excerpt

Again, I have to point out that it might seem somewhat pedantic to reject the term Motherese for studies including fathers’ and children’s input, but, unlike Cattell (2000: 105), I am not “(…) prepared to sacrifice purist correctness in favour of a more user-friendly term(…)” [i.e. Motherese ], but rather want to stress the fact that it is not only mothers’ input this paper is dealing with, but also that of fathers, other adults and other children.

In this paper, I am going to use the more general, self-explaining – and to my view therefore also more user-friendly - term child-directed speech and, whenever necessary and possible, I will try to be explicit about the different sources of the respective input. Child-directed speech, in this paper, is understood as any kind of verbal input that is adressed to the child. It should become evident from the following chapters that, most of the time, any person interacting with a child modifies his/her language. I will also shortly comment on non-verbal aspects, such as gestures, accompanying the linguistic input, but, again for reasons of clarity, I do not want to subsume these aspects under the term child-directed speech, but rather see it as complementary to child-directed speech and, thus, I will subsume it under the broader concept of interaction with the child.

III. Grammatical Aspects of Child-Directed Speech

While the previous chapter was concerned with the problem of giving a name to a phenomenon which “(…) has reached a level of general consciousness beyond any other register (…)” (Brown, 1977: 7) studied by linguists, the features of this special register as well as their principles and functions have not yet been pointed out. Thus, the present chapter will try to give an overview on the different features of child-directed speech. Moreover, these features will be analysed according to the general principles of language modification on which they are based and according to the functions they might possibly fulfil in the child’s language development. In order to give a full account of the features of child-directed speech, it is necessary to consider the different aspects that form the grammatical system of a language. It is this “(…) entire system of rules and principles that account for our linguistic behaviour” (Crain & Lillo-Martin, 1999: 74) which the child has to learn in order to be capable of complete comprehension and correct production of language. It remains to be seen whether any aspects of the grammar of child-directed speech actually include principles of simplification and to what extent these principles might play a role in a possible facilitation of the child’s language acquisition. Some pragmatic aspects, which are not strictly related to the grammatical system, but are rather concerned with the contextual implications of language, and non-verbal aspects of the communicative situation will also be analysed according to these questions.

III.1. Phonetic and Prosodic Aspects of Child-Directed Speech

The foundation of a (spoken) language lies, of course, in the fact that it is made up of different sounds, which are combined to form meaningful words and utterances. We call this the phonological component of grammar. (Crain & Lillo-Martin, 1999: 74) It is these phonological rules a child has to learn in order to be able to use language appropriately.

“An important landmark is reached in a child’s linguistic development when he begins for the first time to use sounds contrastively. Once he has two words in his vocabulary, say [mama] and [dada], the process of acquiring the phonological system has begun.” (Cruttenden, 1979: 16)

But even before the child is able to produce these sounds, he/she will be capable of discriminating these sounds perceptually:

“In general, however, perception of a specific phonemic contrast precedes its production. Children first achieve a perceptual contrast and then try to reproduce a similar contrast in their production.” (ibid.)

Given the fact that language directed to children is different from that directed to adults and that phonology is at the base of language-learning, it seems obvious that the differences in register also concern the phonological level. Moreover, it has been found out that, in particular, the melodic and rhythmic qualities of speech directed to children, the prosodic features, have proved to be different from that directed to adults. (Cooper et al., 1997: 477; Sachs, 1977: pp. 51; Garnica, 1977: pp. 63) This analysis will therefore begin at the ground-level of language and take into account the special phonological and prosodic aspects of child-directed speech.

III.1.1. Phonetics of Child-Directed Speech

In the dicussion of the term “Baby Talk”, it has already been put forward that the language used to adress children is in some ways similar to that used by children themselves. While Clark & Clark (1977) only referred to the lexical items in child-directed speech as being “(…) modeled on the sounds and combinations of sounds that young children tend to produce when trying their first words (…)”, the phonetics of child-directed speech is also similar to that of language by children:

“Many of the processes involved in children’s phonological treatments of adult phonological systems (like regular substitutions and consonant harmony) are similarly represented in BTPh (…)” [i.e. Baby Talk Phonetics^[1] ] (Cruttenden, 1994: 136)

[...]

^[1] n.b.: Cruttenden is one of the authors to stick to the term Baby Talk