The Effect on Language Development of Speech Addressed to Children
Developmental psycholinguistics studies infants and childrens ability to learn and process language, usually with experimental or at least quantitative methods. It examines how speech emerges over time and how children go about constructing the complex structures of their mother tongue. In order to acquire a first language, the young child needs to be exposed to and interact with persons whose language use is more advanced than his own. The speech of adults and of older children and that heard on radio and television are all potential sources of linguistic data for the child learning to speak. Studies of adults speech to children indicate that adults make a number of modifications in their speech when interacting with a young child and that some of these modifications may facilitate the childs language acquisition. Modified repetitions of the childs utterances, along with certain syntactic, semantic, and cognitive simplifications, appear to be potentially effective language-teaching devices. However, the special lexicon, phonological simplification, and higher pitch that also characterize this "baby-talk" style are much less likely to play important roles in the childs language learning. However, in the past several years, there has been a growing body of literature indicating that adults as well as older children introduce a number of modifications into the speech they address to young children (A-C speech). The results of these studies show that, in contrast to the irregularities of adult-adult speech (A-A speech), the young child usually hears well-formed utterances that are short, internally consistent, redundant, and related to the immediate situation. These studies have helped lead to a disenchantment with the view that language acquisition is largely an innate process, and have served to focus new research on the effects of language input on the childs language learning.  During the last few decades, the presence of a number of distinctive aspects of A-C speech has been documented by more than a score of investigators examining a wide variety of language communities. The stability and the breadth of this evidence suggest that "baby talk" (a popular term for adult speech to children between the ages of 1 and 4 years) may play an important role in the childs linguistic development. Cross-cultural researchers have reported baby talk styles in European, Asian, American, and African languages, both in nonliterate and in literate communities. In the United States, systematic modifications in speech to children have been found in communities as diverse as suburban Boston and inner-city Oakland . Indeed, Slobin (1969) suggests that there may be a standard way" of simplifying English. Adults, even while talking to each other, sometimes modify their speech in the presence of children. The modifications that mothers introduce have been emphasized most, but fathers, too, use a special style of speech with children , as do women who are not mothers . Even children perhaps as young as 4 years of age adopt a baby talk style to address younger children . Systematic modifications do not appear to be limited only to spoken language-deaf parents reportedly adjust their signing in particular ways to deaf children.  The data from 31 studies converge to yield a fairly consistent description of A-C speech. These data indicate that adults frequently address children in short, high-pitched sentences spoken clearly and slowly. This A-C speech often includes its own uniqu e lexicon of about 20-60 items , as well as numerous words which are modifications of those that normally occur in adult speech. This special vocabulary usually includes kinship terms, animal names, nicknames, words referring to body parts and functions, terms for basic qualities (such as good, bad, and dirty), and names of games and playthings. Also characteristic of the baby talk lexicon are diminutives and terms of endearment. When unmodified A-A words do occur in the speech to children, they art sometimes used in different ways. For example, adults name certain objects differently for children than for other adults. The terms that children hear are often the less specific ones. In comparison to the speech that adults address to each other, A-C speech is simplified phonologically (reduplication, lengthened vowels, dis- tinctive consonant-vowel clusters, etc.) and syntactically (fewer subordinate and coordinate clauses, fewer embeddings and conjoinings, more "content" words, etc.). Baby talk reduces cognitive complexity, too. There are at least three types of evidence for this conclusion. First, mothers usually speak to children about recently completed actions or about immediately present objects and pictures. This emphasis on the here-and-now introduces a verbal-contextual redundancy that may allow the child to gasp the meaning of a sentence without necessarily understanding its structure. Second, younger children are more likely than older children to be asked questions with known answers-probably a cognitively less demanding task than answering other types of questions. Finally, adults seem to communicate fewer concepts per unit of time to children than to other adults.  A-C speech often includes repetitions and expansions of the childs utterances. The exact proportion of repetitions and expansions in different samples of baby talk, however, is not consistent. Thirty percent of the childrens utterances in one American study prompted parental expansions, but only 5% of the utterances of Luo and Samoan children were followed by any sort of repetition or expansion. However, in the latter study, the parents appeared to be actively controlling the verbal interaction, rather than reacting to and elaborating on the childs sentences. Thus it is not surprising to find that all but 3-11% of the Luo and Samoan childrens utterances were elicited by adult utterances.  The evidence on adults use of different sentence types is also inconclusive. In four different A-C speech samples, there is a preponderance of interrogative forms. Two of these also include a high frequency of imperatives-a rare occurrence in A-A speech. In contrast to these findings, Brown and Hanlon (1970) report many more simple active affirmative declaratives than negatives and interrogatives combined. The difficulties in obtaining clear indices of sentence types are probably magnified by the demand effects of the task structure (the adults in the Sachs study, for example, were told to elicit the childs attention, and may have asked more questions as a result) and by the parents concern with possible evaluation of their child (Blount notes that the Luo girls parents were preoccupied with showing the experimenter how well the child could carry out instructions; hence they produced a large number of imperatives).  The results of a number of recent studies reveal many parallels between the speech addressed to children and the language the children actually use. Childrens speech often preserves the stress, the rhythm, and the intonation of the adults most recent utterance. This finding is reported both in studies that elicited imitations from the children and in those that recorded spontaneous speech. There are similarities in content, too. Mothers speak primarily about their children and the events their immediate activities. Snow found that more than two-thirds of the semantic relations in childrens utterances involved those prevalent in mothers A-C speech. One final similarity between A-C speech and the childs speech is that the mean length of utterance of 2-year-olds is positively correlated with the mean length of their mothers A-C utterances. The occurrence of parallels between A-C speech and the childs own speech does not necessarily imply that children will adopt any linguistic form that adults model. At different points in his development, the child appears to attend to different perceptual categories and to use different strategies to process linguistic information. The appropriateness of the childs perceptual and cognitive approach may determine the accuracy of his comprehension of a linguistic structure. Syntactic forms which are misinter- preted will be difficult for the child to reproduce. Hence the usefulness to a child of a certain sample of language input must be assessed with respect to the childs current level of cognitive and perceptual skills. If linguistic input becomes too complex, the child may simply not respond at it.  The possible effect of an absence, or a relative dearth, of adult-child language interaction can be seen in the retarded linguistic development of a young hearing child of deaf parents. The childs parents did not use the sign language of deaf in his presence, as they wanted him to learn spoken English only. Thus, during the first 4 years, the child was exposed to language only while watching television, playing with young neighborhood children, and participating in a program of group activities. It is unlikely that the language used by either the television programs, the young playmates, or the group supervisors incorporated many aspects of A-C speech or encouraged the childs own productivity. As a result, the paucity of A-C speech modifications appears to have had a marked effect on the boys language growth-even after being placed in a special intervention program, he could form only rudimentary sentences.  Among the many specific aspects of A-C speech that might help the child learn the relationships between the adult grammatical construction and the structure of his own utterances are the adults use of "occasional" questions and incomplete sentences. The occasional question is a request that the child repeat an unintelligible word or phrase or provide a missing answer. Two subsets of this form of question are the "say constituent again" interaction: Child: I want milk. Mother: You want what? Child: Milk. and the "constituent prompt": Mother: What do you want? Child: (no answer) Mother: You want what? Children more frequently give appropriate responses to an occasional form of a question than to regular question forms, and the use of the occasional form by parents is positively correlated with grammatical development (Brown et al., 1969; Moerk, 1972). Apparently the recasting of the question and the prompting of the child to reply aid the childs understanding of constituent structure and help the child infer the antecedents of words like what and it. It is likely that a similar effect is produced by a related language-teaching device, the incomplete sentence, which probably focuses the childs attention on the missing sentence constituent in his mothers utterance. For example: Child: I want milk. Mother: You want...? Child: Milk. The slower pace of A-C speech may also prove helpful to the child. McNeill (1970) cites an unpublished study by Psotka (1969) which tested childrens comprehension at rates of 0.5 word/sec, 1 word/sec (the rate at which the children in the study spoke), and 3 words/sec (the experimenters rate of speaking). Childrens understanding was best at their own production rate and poorest at the slowest rate. This study suggests that the helpfulness of the A-C rate modification cannot be determined from a comparison of A-A and, A-C rates alone; the childs own rate of speaking must also be considered. This implication is true of other features of A-C speech as well-an A-C sentence which is shorter than the average A-A sentence, for example, may still be longer than the average utterance produced by the child. Further research should evaluate each baby talk feature in terms of A-A speech and the childs speech.  The mutual modifications of responses that characterize the adult-child interaction may provide the richest type of data for the child learning the language. However, as seen in the studies of adult expansions of the childs utterances, the effects of these patterns of interaction on the childs language growth remain unclear. Such interactions, in fact, do not appear indispensable to acquiring language comprehension. The child Lenneberg (1962) described who was physiclly incapable of producing verbal feedback showed no impairment of understanding. [However, this child could have provided other types of responses (no action, appropriate action, inappropriate action) which were functionally similar to verbal feedback.] More generally, as yet there is no strong evidence that any particular kinds of A-C speech or adult-child interaction are necessary for language acquisition. Instead, the data are primarily correlational, showing that both childrens speech and the speech of others to children change with the childs increasing age.  Exposure to A-A speech in addition to A-C speech may, because this A-A speech adds complexity, enhance rather than hinder the childs progress. For example, A-A repetitions may serve to introduce a wider range of word categories to the young child. These repetitions usually involve articles, possessive pronouns, personal pronouns, and conjunctions (Maclay and Osgood, 1959)-just those parts of speech that emerge late in the childs linguistic development. It is possible that a child exposed to normal A-A speech would learn functors more rapidly than a child hearing only A-C speech. It is also true that the false starts in A-A speech often involve the inappropriate repetition of content words; this, however, may pose no problem to the child, who, when imitating, may simply ignore semantically irrelevant words. The hypothesis that pitch acts as a cue to the child to attend to certain utterances may also be interpreted to imply that the child also learns to disregard other low-pitched, non-baby talk utterances which contain valuable linguistic data. It has been posited that A-C patterns of intonation, pauses, and stress make grammatical units more salient; actually, the relation may be the reverse-perhaps it is proper sentence fragmentation that the child learns first, and which determines where he hears pause, stress, and intonation. With respect to sentence length, it may be more beneficial for the child to be exposed to speech samples of varying MLUs (means of length utterance) than to be addressed with consistently short utterances. As the childs MUL grows, he imitates increasingly long A-C utterances, even if the adults MLU remains stable or increases less rapidly than the childs (Lord, 1975). Finally, the child might acquire a more differentiated vocabulary if he were to hear objects named for him in the same way as they are named for other adults.  The elements of baby talk seem to fall into two categories:(1) those that play no obvious facilitative role in the childs language learning-the squeaky voice, the phonological distortions, the diminutives, the unique lexicon of "bow-wow," "beddie-bye," etc.; and (2)those that may promote linguistic development-shorter MLUs , slower rates, recast sentences, occasional questions, and perhaps a few other components. Perhaps a linguistic sample constructed with elements from the second category might be most effective if it is (a) designed in such a way as to invite the child to be an attentive and active participant in verbal exchanges, and (b) moderately novel, in comparison to the childs linguistic output. These are not independent factors; input which is moderately novel is most likely to hold the childs attention. Expansions are hypothesized to be especially effective because they are emitted at a time when the child is particularly likely to be attentive. One possible reason why Cazdens group did not excel is that they were overexposed to expansions and consequently ceased to attend to them. The role of recastings in facilitating language growth has been documented experimentally. Like other types of expansions, recastings might derive their potency from the interest and attention they elicit from the child, who is listening to a reworked version of his own words. Shipley et al. (1969) postulate that the low frequency of appropriate response to commands beginning with nonsense words may be a result of the childs failure to attend to a sentence that begins with meaningless items. Even the 4-year-olds seem alert to the importance of the variable of attention-they used significantly more attentional devices with 2-year-olds than with adults.  Because the childs comprehension usually outdistances his production, perhaps his mothers speech should not match the childs but instead stay a step or so ahead. Adams mothers sentences, for example, are, on a very general level, the type that Adam himself will produce the following year. Equal or better understanding of slightly more complex adult speech by young children is seen in the study by Shipley(1969). Those children who were themselves producing holophrastic utterances responded just as well to two-word commands as to one-word commands, while the children who were producing telegraphic utterances responded best to mature adult forms (involving a verb, one or two function words, and a noun). As the authors note, "just those utterances that they [the children] themselves did not use were more effective as commands". Brown and Hanlon (1970) "suspect that the only force toward grammaticality operating on the child is the occasional mismatch between his theory of the language and the data he receives". The maximally sensitive adult may respond differentially to children of different ages and different MLUs, and may even adjust his speech to the details of the childs progress: for example, some adults may provide different linguistic input to a child who has mastered the possessive inflection than to one who has not. Such an A-C interaction, if it occurs, is a far cry from the initial characterization of baby talk along static, global dimensions.  In addition, it should be noted that adult verbal interaction with children is, in most cases, a phenomenon of staggering proportions. One recent estimate of the total amount of adult speech heard by a child between the ages of 2 and 3 or 3-1/2 years is about 1 million words. In addition, the childs corresponding productions were estimated at 1/2 million words. Clearly, the language-learning child is exposed to an enormous amount of speech, which may facilitate his induction of the rules of his language, and he is also actively producing a large quantity of his own utterances. In summary, although the studies reviewed above are only the first steps toward an adequate model of the role of adult speech modifications in childrens language acquisition, there is every reason to believe that as more detailed investigations are completed the specific effects of such modifications will be much better understood. Regardless of the exact form of the final model, the above studies make clear that some of the special characteristics of the language spoken to a child may play an important role in the childs language development. Indeed, if one were to design a language-teaching device for a young child, one would probably try to incorporate those aspects of the baby talk style that appear to encourage the child to be an active and attentive language learner, and that challenge him to deal with utterances that are slightly more complex than his own.
REFERENCES: 1. Berko Gleason, J. Talking to children: Some notes on feedback, Boston, 1974. 2. Brown, R. A First Language, Harvard University Press, Cambridge, Mass, 1973. 3. Cazden, C. Environmental assistance to the childs acquisition of grammar. Harvard University, Cambridge, Mass, 1965 4. Thomas Scovel. Psycholinguistics. Oxford University Press, 2002. 5. Slobin, D. I. Questions of language development in cross-cultural perspective. Working Paper No. 14, Language-Behavior Research Laboratory, Berkeley, Calif. 1969.