Babies learn to speak by listening to their caretakers. People help them by modulating the sounds of speech in fundamentally the same way. A topic of much speculation among researchers who study language acquisition is the observation that caretakers consistently address their infants in this unique tone and manner of voice, a form that has come to be known as “baby talk,” “infant-directed speech,” or motherese. Motherese is a linguistic register based on exaggeration of pronunciation and simplification of syntax. It is found in virtually every culture and has certain common characteristics: The sentences are very short, there is a lot of repetition and redundancy, there is a sing-song quality to it, and it contains many diminutive words. It is also embedded in the context of the immediate surroundings, with constant reference to things and goings-on nearby. Certainly there is a social component to this form of speech. Infants respond more positively and listen longer to infant-directed than to adult-directed speech. Furthermore, while researchers disagree over whether exposure to this type of speech is necessary for successful first language acquisition, there is general agreement that motherese contributes to the ease with which infants are able to break into their particular language of exposure.
For the language-learning infant, identifying the components that make up spoken language is a difficult task. Instances of words vary phonetically and acoustically, depending on the discursive, syntactic, and phonological contexts in which they occur. This is in addition to variations introduced by changes in, for example, talker identity and speaker affect. At the earliest stages, word recognition must be guided by features of the individual instances of words themselves. Although it remains unclear precisely which aspects of the auditory signal initiate recognition, acoustic prominence—an important characteristic of infantdirected speech—is one factor that has been considered particularly influential in jump-starting this process. Data supporting this view indicate that infants generally prefer to listen to acoustically salient speech, where “salient” can mean either effectively or emphatically so. Conveniently, the natural form of input to the language-learning child is modified in just such a way.
Infants face another difficulty when it comes to speech segmentation. In fluent speech, words are not separated by pauses, and the cues that may serve to signal word boundaries vary from language to language. Nevertheless, and despite these challenges, normally developing infants begin to succeed at recognizing words in fluent speech about midway through their first year. This has been attributed, in large part, to caretakers’ tendency to repeat content words when addressing their infants. Repetition of the full form of a word is perfectly reasonable—even expected—in speech directed to infants, and this repetition is quite distinct from the reduction to pronominal form that occurs across mentions of content words in adultdirected speech. Although repetition is often cited as one of many characteristics of speech directed to infants, it is generally viewed as subordinate to the prosodic quality of such speech. While repetition appears to be an important feature in guiding speech segmentation, it is not the aspect of motherese that is most often referred to as influential in language learning. However, given the problems the language learner faces, it may well be just as important as other aspects of this unique register.
Generally speaking, infant-directed speech involves clear and careful pronunciation, exaggerated intonation, relatively few abstract words, reference to tangible objects that a child can see and touch, and a focus on the actions the child is doing or witnessing. Not only mothers speak motherese. Anybody who communicates with young children will adopt this modified form of speech. And while it is arguably not fundamental to a child’s ability to acquire language, its apparent universality points to its importance in aiding that process.
- Ferguson, A. (1964). Baby talk in six languages. American Anthropologist, 66, 103–114.
- Fernald, , Taeschner, T., Dunn, J., Papousek, M., BoyssonBardies, B., & Fukui, I. (1989). A cross-language study of prosodic modifications in mothers’ and fathers’ speech to preverbal infants. Journal of Child Language, 16, 477–501.
- Jusczyk, P. (1997). The discovery of spoken language. Cambridge: MIT
- Morgan, J. L., & Demuth, K. D. (Eds.). (1996). Signal to syntax: Bootstrapping from speech to grammar in early language acquisition. Mahwah, NJ: