Once upon a time and a very good time it was there was a moocow coming down along the road and this moocow that was down along the road met a nicens little boy named baby tuckoo.

Fittingly enough, Joyce began Portrait of the Artist with a bedtime story in baby talk. But the process by which we learn to speak (“we” refers to all human beings, speaking every language on earth) is considerably more complex than baby talk. Among the mistaken notions about baby talk are that it represents some “natural,” instinctive way children begin to speak and that it is an adult imitation of childish speech. The truth is that baby talk is simply a variant of the adult language, taught to children by adults who then, after a few years, encourage the children to stop using it.

The really fascinating thing about baby talk is the universality of its linguistic form and content, as demonstrated by a comparison of the way it is spoken in six quite different languages: American English, Spanish, Syrian Arabic, Marathi of India, Gilyak of Siberia, and Comanche Indian of North America. The actual baby-talk vocabularies in the six languages are of course different; nevertheless, the words reveal surprising similarities in linguistic characteristics. All six languages simplify clusters of consonants (as English speakers do when they substitute tummy for stomach); they reduplicate syllables (choo-choo); they alter words in consistent ways to form diminutives (such as the y in doggy); they eliminate pronouns (daddy wants instead of I want); and most of the languages drop unstressed syllables (as when good-bye becomes bye or bye-bye). The existence of such similarities in widely different languages suggests that adults with no knowledge of one another’s tongues have arrived at much the same linguistic formulas.

It is now also becoming apparent largely because of the research by the authors of the three books under review—that all children pass through stages in learning to speak, regardless of the language of the speech community they are born into, and that these stages have nothing to do with traditional ideas about “baby talk.” Psycholinguists have collected much evidence that English-speaking children of all ethnic backgrounds and socio-economic classes go through strikingly similar stages of development in learning to speak; furthermore, these stages appear to be very much the same for children whose native tongues are more than thirty languages from widely disparate language families.

No wonder that the recent growth of interest in the way children acquire their native language has been phenomenal, beginning a little over fifteen years ago and rapidly increasing in the past few years. As Roger Brown writes:

All over the world the first sentences of small children are being as painstakingly taped, transcribed, and analyzed as if they were the latest sayings of great sages. Which is a surprising fate for the likes of “That doggie,” “No more milk,” and “Hit ball.”

The intellectual capacities of infants and young children are of course very limited. Yet, beginning at about one and a half years of age, in a span of usually less than four years every human child, whether stupid or intelligent, masters most of the exceedingly complex structures of his native language—and he does so before he even enters what used to be called “grammar” school. Parents can encourage and drill the very young child in speech, but their influence is negligible in making the child do a substantially better or worse job in producing the grammar of his language. A child does not speak much sooner or later because of parental support or neglect; and when he finally begins to speak he does so in a set way that he shares with children who speak every other language. Middle-class parents often think they have an impact on improving their children’s speech, but that is true only of corrections in pronunciation and of such irregularities as goed, and the elimination of taboo words. Numerous observations have revealed that parents almost never correct errors in syntax such as Why doggy no eat? (and, in fact, do not even seem aware of most of them).

The fascinating and complex way in which we learn to speak is being understood increasingly well, largely because of the pioneering work of Roger Brown at Harvard and Dan Slobin and Susan Ervin-Tripp at Berkeley, as well as the numerous students they have influenced. The initial impetus for a new approach to language acquisition came some fifteen years ago from ideas put forward by Noam Chomsky of MIT and later elaborated by him and his colleagues in discussions of transformational grammar. This is not the time to go into the pros and cons of Chomsky’s positions regarding many facets of language (John Searle has already examined the controversy over Chomsky’s work in NYR, June 29, 1972), but rather to emphasize those specific points which, regardless of whether they eventually prove true or not, have been a liberating influence on psycholinguistics. With a brevity that I hope does not do. violence to Chomsky—or to his detractors—let me summarize those of his views that have particular reference to the child’s acquisition of language.


Chomsky has argued that an infant is endowed at birth with an innate capacity to acquire a language, the one which he hears spoken around him, just as he possesses at birth an innate capacity to learn to walk. This human blueprint for language makes every child extraordinarily creative in producing and understanding sentences he has never heard before, and this is particularly obvious when the child’s early constructions are examined. No child could conceivably remember all of the sentences he hears in his native language (and, in fact, the number of such sentences is theoretically infinite). Instead, he acquires grammatical rules for the creation of sentences, thereby allowing him to produce an infinite variety of utterances in his native language, almost none of which he has heard previously.

A note of caution: “rules” should not be misunderstood. Chomsky and the psycholinguists do not say that speakers, whether adults or children, can state the explicit (and extraordinarily complex) rules of their language’s grammar. In fact, no one can—if for no other reason than the inability of linguists to discover as yet the complete and adequate rules for any language. The important thing is that from the time of our earliest sentences we speak as if we understood those rules. Show a very young child a picture of a fanciful creature and tell him it is called a flamp; then show him another picture with two such creatures and he will unhesitatingly call them two flamps, a clear indication that he possesses a rule for producing English plurals, even though he could not possibly have heard the made-up word flamp previously. In the same way, adults follow considerably more complex rules, even though no adult, unless he be a linguist, could possibly articulate those rules.

While Chomsky’s transformational grammar provided the initial impetus for the work of Brown and Slobin, and to a lesser extent Ervin-Tripp, the transformational linguists nevertheless have attacked the methodology of psycholinguistic research. Their contention has to do essentially with the question of competence versus performance, a distinction seen more clearly in the analogy of arithmetic. The fact that someone has grasped and internalized the rules for multiplication (his competence) does not mean that he will always produce correct answers to problems (his performance). Such a distinction becomes even clearer in the case of children, whose performance is likely to result in an unusually high number of multiplication errors.

Chomsky feels that, in the same way, children’s speech is composed largely of stumbling, awkward, and ungrammatical statements—and that such performance is almost useless to the linguist attempting to arrive at an insight into the child’s competence. Chomsky’s position, which he has expounded with considerable logic, has much to recommend it—except that Brown, Slobin, Ervin-Tripp, and others have produced extraordinary insights by studying performance. Nor have the transformational linguists offered an alternate methodology that has been as productive.

Psycholinguists were also encouraged to do research on how we learn to speak by a discovery that on the face of it is rather commonsensical to the nonspecialist. The discovery was, simply, that children, even very young ones, intend to convey meanings by what they say (earlier studies had ignored semantic considerations in favor of syntactic abstractions). By the time most infants reach the end of the first year, parents usually identify in their babbling the sudden appearance of intelligible, single-word utterances. These are technically known as “holophrastic speech” in which a single word stands for an entire complicated sentence. An infant uttering the word ball may mean (depending upon the particular speech situation) “I see the ball,” “Give me the ball,” “I lost the ball,” or any number of other things.

Shortly thereafter, at about the middle of the second year, the child puts two words together to form primitive sentences, such as more milk, give toy, or daddy gone. Recent studies show that even at this early stage the child is using language in the basically human sense of intending to communicate meanings that go beyond mere naming or the description of events. Every parent knows that the child is, in fact, asserting the existence of particular relationships in his environment. He understands, for example, the semantic relations between an agent (such as a thing that hits) and an object (the thing that is hit by the agent), relations that in English are often expressed simply by word order. That can be shown by his different behavior when he is told to “Make your truck hit the car” and then later “Make your car hit the truck.” Both commands involve the same words and action, but the change in word order defines which toy is the agent (the hitter) and which the object (the thing hit).


A truly amazing fact that emerges from cross-linguistic studies of the early speech of children is the similarity in its progress everywhere. The English-speaking child’s expression of the function of demand, as in more milk, has a similar function and linguistic form when used by German, Russian, Finnish, Samoan, East African, and other children at equivalent stages of acquiring their native languages. In fact, a mere eighteen relations account for practically every utterance by very young children speaking English and all other languages that have been studied for comparison purposes. These relations include the nominative (that ball), the possessive (daddy chair), locative (go home), attributive (big ball), and so on.

Even at this early stage, children speaking any one of a wide variety of unrelated languages tend to talk about the same things—such as the location of objects, description of a situation, or an expression of demand—and almost never to talk about themselves or about social relationships. Nearly all their words are nouns and verbs, never articles or prepositions; they never use case endings; they use an early form of negation which emphasizes nonexistence (no hat) rather than rejection or denial. Apparently, the similarities in incipient speech in the many languages studied reflect the kind of cognitive growth that children everywhere have at about this age. Children who are beginning to speak merely express what they already know; first they think, then they talk.*

The child who is beginning to speak very soon produces sentences of more than two words. The transition from two-word to longer utterances can actually be detected whenever a child starts a short sentence, then pauses and adds words to create a longer sentence, as in Want toy…Cathy want toy. And hesitations further indicate the child’s understanding of grammatical relationships. A child who says Stand up…cat stand up…cat stand up (hesitation) table has unwittingly offered a view into the grammatical system he has begun to acquire.

His repetitions and hesitations reveal that he has unconsciously analyzed this sentence into its hierarchical constituents. He knows the difference between the subject (cat) and the predicate (stand up table). His hesitation before table indicates that he has further divided the predicate into a verb (stand up) and a locative (table). The child may stumble, hesitate, say foolish or cute things, but he is nevertheless always operating on his own grammatical principles. Even though his speech in these early years deviates from adult speech, it does so in a systematic and regular way. The conclusion is inevitable that such utterances are creative constructions that result from the child’s preliminary analysis of his native language.

A further insight into the cognitive processes of the language-building child concerns irregular verbs. Parents know that children incorrectly inflect the past tense of irregular verbs as if the verbs were regular—in that way producing forms like comed or breaked—and that they continue to do so well into elementary school. This phenomenon, known as “regularization,” has also been noted among children speaking many other languages. The glib explanation used to be that children simply extend the regular forms for the past tense, like walked or helped, to irregular verbs like come or break. The actual sequence of events, though, is more complicated than that.

At first children use the irregular verbs correctly; they say came and broke. But then, as their knowledge of grammar expands and they begin to learn the past tenses of the many regular verbs like walk and help, they replace the correct forms with the incorrect ones. Apparently they do so because the preponderance of regular forms allows them to discern a rule for past tenses; they then over-regularize the rule by unconsciously applying it as widely as possible, even to the extent of eliminating from their vocabulary the correct forms (came, broke) which they once knew. As Slobin comments, “One cannot help but be impressed with the child’s great propensity to generalize, to analogize, to look for regularities—in short, to seek and create order in his language.”

I have done no more than suggest that developmental psycholinguistics is a field in which important research is being undertaken, with implications for language, child development, mind, personality, our knowledge of the world around us, and a lot of other things. For example, the maturing child constantly improves his grammar, moving ever closer to the adult form. It would appear sensible that parental approval or disapproval would inevitably encourage the child to move toward better formed sentences—but evidence does not support that sensible conclusion. What, then, impels the child to put aside, one after the other and in sequence, childish levels of speech until finally the adult form is reached? The answer to a question such as this one would go to the roots of the process of mind. Even more, an answer would go far toward defining what is uniquely human. At one time or another human beings were defined as tool-making animals, social animals, or playful animals, but all of these definitions have been undermined as animal behaviorists discover equivalent behavior in the beasts. The uniqueness of human beings now must be defined in view of behavior depending on language—and this behavior begins with the first words of the helpless, crawling infant.

All three books under review are extremely valuable works on this subject, although for quite different reasons. In only 148 pages, Slobin has brilliantly presented a clear summary of the present state of psycholinguistics, including, but not limiting himself to, language acquisition by children; his book can be read by anyone untutored in this very complex subject. It is good to have Ervin-Tripp’s major essays collected in one place, if for no other reason than that they show the interesting questions that might be raised in psycholinguistic research, such as the way in which bilingual children manage to acquire two languages. Brown’s volume is written in the grand manner of major pioneering works in science, but it is of such complexity that it demands almost Talmudic study. Study and patience, though, are rewarded; it is an exhilarating book about a new field of human knowledge.

This Issue

February 21, 1974