Chomsky and language learning.  



Noam Chomsky Critics The evidence Wild children

Evidence from mother-tongue


Powerpoint Presentaton LAD The Universal Grammar Theory Neurological evidence The blind and the deaf Conclusion



Noam Chomsky

Noam Chomsky is perhaps the best known and the most influential linguist of the second half of the Twentieth Century. He has made a number of strong claims about language : in particular, he suggests that language is an innate faculty - that is to say that we are born with a set of rules about language in our heads which he refers to as the 'Universal Grammar'. The universal grammar is the basis upon which all human languages build. If a Martian linguist were to visit Earth, he would deduce from the evidence that there was only one language, with a number of local variants. Chomsky gives a number of reasons why this should be so. Among the most important of these reasons is the ease with which children acquire their mother tongue. He claims that it would be little short of a miracle if children learnt their language in the same way that they learn mathematics or how to ride a bicycle. This, he says, is because :

  1. Children are exposed to very little correctly formed language. When people speak, they constantly interrupt themselves, change their minds, make slips of the tongue and so on. Yet children manage to learn their language all the same.
  2. Children do not simply copy the language that they hear around them. They deduce rules from it, which they can then use to produce sentences that they have never heard before. They do not learn a repertoire of phrases and sayings, as the behaviourists believe, but a grammar that generates an infinity of new sentences.

1. 2. Children are born, then, with the Universal Grammar wired into their brains. This grammar offers a certain limited number of possibilities - for example, over the word order of a typical sentence.

Some languages have a basic SVO structure


The teacher


a lecture




75% of the world's languages use either this (English, French, Vietnamese) or SOV (Japanese, Tibetan, Korean) - others prefer VSO (10 - 15% - Welsh) or VOS (Malagasy)

 * Some languages, such as Latin, appear to have free word order, but even here, SOV is very common. OSV is very rare - but you will find an example in the speech of Yoda, in Star Wars.


When the child begins to listen to his parents, he will unconsciously recognise which kind of a language he is dealing with - and he will set his grammar to the correct one - this is known as 'setting the parameters'.

It is as if the child were offered at birth a certain number of hypotheses, which he or she then matches with what is happening around him. He knows intuitively that there are some words that behave like verbs, and others like nouns, and that there is a limited set of possibilities as to their ordering within the phrase. This is not information that he is taught directly by the adults that surround him, but information that is given. It is as if the traveller were provided at the beginning of his journey with a compass and an astrolabe.

This set of language learning tools, provided at birth, is referred to by Chomsky as the Language Acquisition Device . (Notice that he uses the term "acquisition" rather than learning).


How did you learn to speak your native language? Notice, this shouldn't be such a puzzling question. We often ask questions such as, do you remember when did you learned to tie your shoes, ride a bike, and eat with a fork. Sometimes we can remember because a parent helped us learn how to do these things. Now, since we always speak the language of our parents, they must have helped us learn to speak our first language. But do you remember when your mother taught you the past tense? When your father laid down the rules for passive sentences? We don't remember these important moments of our childhood because they never occurred.

Our parents didn't teach us how to walk and they didn't teach us how to talk. Yet we learned from them. How can this be? Certainly there must have been a subtle, perhaps intuitive teaching process that neither our parents nor we were aware of. We begin by imitating what we hear our parents say as best we can, repeating random phrases. Our parents in subtle ways punish us for the childish speech errors we make (by not responding, correcting the error, etc.) and reward correct phrases (by responding positively). As our speech improves, our parents respond more positively and less negatively. No?

First, let's examine the assumption that children begin speaking by trying to repeat what they have heard their parents say. Have you ever heard a child say things like this:

1a Daddy go
1b He hitted me
1c No eat cake

Who did they hear utter such phrases? Daddy go is an attempt to express 'Daddy is going'. But if the child were merely trying to repeat this common phrase, choosing random two-word combinations, he or she would also occasionally say Daddy is or simply is going? Yet these two phrases do not occur as normal speech errors of children while Daddy go is a common one.

Second, research shows that while mothers often respond to the semantic content of what their children say ('No, that's not a doggie, it's a cow'), they very rarely respond to the grammatical status of their children's phrases. Indeed, when parents do respond to speech errors, they most often respond positively. Here are a few advanced errors from the history of my family. What do you think our response was—correction or laughter (which I take to be a positive response)?



Mama, mama, there's a tree-knocker in the back yard!


It's raining, where is the underbrella?


Give me the beach-lookers! (binoculars)


In fact, parents themselves make grammatical errors when they speak. Despite the fact that children don't know when their parents are speaking grammatically and when they are making errors, all children grow up knowing (if not always speaking) the language perfectly.

So how do we learn to speak? Take a look at example No. 1b above for a clue. Although hitted is not a word children hear adults utter, it is wrong for an interesting reason: the verb, in a sense, has the 'right' ending on it for the past tense. In other words, the only way a child learning language could make such an error is that he or she is learning a rule that derives past tense verbs from verb stems. What the child hasn't mastered at this stage is the exceptions to the rule. Notice also that the words in the erroneous phrases are all in the correct order. No child would say go Daddy for 'Daddy is going' cookie mommy for 'Mommy's cookie'. By the time a child begins putting two words together, he or she has already mastered the basic rules of syntax and applies them correctly even in their erroneous speech. It takes the child a little longer to master the rules of morphology.

The evidence then indicates that children do, in fact, absorb a massive number of sentences and phrases but rather than parrot them back, they abstract rules from them and create their own grammar which they then apply to create new utterances they have never heard before. Over the years from 2-7, when language is mastered, children constantly adjust their grammar until it matches that of the adult speaker population.

This critical period between the ages of 2-7 suggests that (first) language learning, like walking, is an innate capacity of human beings triggered by a level of development more than feedback from the environment. That is, so long as a child hears a language-any language-when they reach this critical period they will learn it perfectly. If this is true, any child not hearing language during this period not only should not learn to speak but also should not be able to learn to speak. The ethical implications of research on this question are obvious. However, there have been a few tragic non-scientific bits of evidence that supports the innateness + critical period hypothesis.


If I wanted to start the course off with a silly pun, I could say 'Learning a language is a child's game'. But perhaps it is more accurate to say 'Creating a language is a child's game'. Let us look at an example of how a language may be created :


But a pidgin can become a language - Creole. How does this happen?

According to Derek Bickerton, who has reconstructed the process of creolisation in Hawaii, it takes one generation.

When children begin to use a pidgin, they automatically enrich the vocabulary and the syntax - it becomes a full language. The community of young children in Hawaii took the pidgin used by their parents - workers from China, Japan, Korea, Portugal, the Philippines and Puerto Rico - and created a language.

According to the followers of the American linguist, Noam Chomsky, this can stand as an emblem for what the process of acquiring a language consists in - at least for a mother tongue. The child does not learn the language, but creates it anew.

Does this have anything to tell us about learning a foreign language?


Chomsky's critics

Those linguists who do not agree with Chomsky point to several problems, of which I shall mention just four.

  1. Chomsky differentiates between competence and performance. Performance is what people actually say, which is often ungrammatical, whereas competence is what they instinctively know about the syntax of their language - and this is more or less equated with the Universal Grammar. Chomsky concentrates upon this aspect of language - he thus ignores the things that people actually say. The problem here is that he relies upon people's intuitions as to what is right or wrong - but it is not at all clear that people will all make the same judgements, or that their judgements actually reflect the way people really do use the language.
  2. Chomsky distinguishes between the 'core' or central grammar of a language, which is essentially founded on the UG, and peripheral grammar. Thus, in English, the fact that 'We were' is considered correct, and 'We was ' incorrect is a historical accident, rather than an integral part of the core grammar - as late as the 18th Century, recognised writers, such as Dean Swift, could write 'We was ...' without feeling that they had committed a terrible error. Similarly, the outlawing of the double negation in English is peripheral, due to social and historical circumstances rather than anything specific to the language itself. To Chomsky, the real object of linguistic science is the core grammar. But how do we determine what belongs to the core, and what belongs to the periphery? To some observers, all grammar is conventional, and there is no particular reason to make the Chomskian distinction.
  3. Chomsky also appears to reduce language to its grammar. He seems to regard meaning as secondary - a sentence such as 'Colourless green ideas sleep furiously' may be considered as part of the English language, for it is grammatically correct, and therefore worthy of study by Transformational Grammarians. A sentence such as 'My mother, he no like bananas', on the other hand, is of no interest to the Chomskian linguist. Nor would he be particularly interested in most of the utterances heard in the course of a normal lecture.
  4. Because he disregards meaning, and the social situation in which language is normally produced, he disregards in particular the situation in which the child learns his first language.

That some kinds of migratory birds navigate thousands of miles toward their destination by calibrating the positions of stars against time of day and year, poses no serious problem for many scientists, who can easily attribute this amazing success to the birds' instinctive behaviour is apparent, after all, that these animals cannot learn such complicated astronomical facts through a trial and error fashion; they neither have enough time nor necessary cognitive capacity.

The same scientists, however, including some professional linguists, are quite reluctant to attribute any form of instinct to human infant, who arrives at complex linguistic knowledge within a remarkably short period of time. The infant's is no less a complicated task than that of the bird's as the linguists themselves have spent decades (or even centuries) to discover the intricacies of the very same system and with no final theory. Infants, on the other hand, not only arrives at an almost complete knowledge of grammar in their brinds (brain+mind) but also accomplish this task within less than a decade.

Although a human infant and a migratory bird are essentially alike in terms of the complexity of the task to be accomplished and their inability to handle the task with their current cognitive capacity, only the latter is believed to rely on its instincts.

There are, of course, some differences between an animal and a human baby; it would be unwise to equate the cognitive capacities of the two. And it is also impossible to underscore the importance of environmental factors in child language acquisition. After all, thousands of hours of exposure is required in order for a child to acquire his mother tongue, whereas animals like sonar-using bats or web-building spiders seem to be ready to use their instinctive knowledge with minimum, if any, learning experience. It is equally unwise, however, to suggest that a cognitively immature child can accomplish a task which has yet to be accomplished by professional linguists.

A child may well not have grasped the property of conservation of volume nor be able to perform but the most rudimentary arithmetic calculations, yet will have the knowledge linguists formulate as the binding principles, none of which has been explicitly taught.

The amazing success of children in picking up their mother tongue is no recent discovery. Slobin (1979) quotes Rene Descartes commenting on human beings' distinctive ability to formulate a linguistic system:

Even those who were born deaf and dumb, lacking the organs which others make use of in speaking, and at least as badly off as the animals in this respect, usually invent for themselves some signs by which they make themselves understood by those who are with them enough to learn their language

In the literature of child language acquisition there are cases in which infants, deprived of linguistic input, invent a rudimentary grammar not attributable only to the external factors. Children are also known to build a natural language when exposed to unsystematic pidgin data. The resulting creole is almost as systematic and sophisticated as any natural human language and more interestingly contain rules that are not attributable to the languages forming the pidgin, out of which the creole is driven.

The Universal Grammar Theory

Among theories of language acquisition, Universal Grammar (UG) has recently gained wider acceptance and popularity. Though noted among L2 acquisition theories, the defenders of UG are not originally motivated to account for L2 acquisition, nor for first language (L1) acquisition. However, UG is more of an L1 acquisition theory rather than L2. It attempts to clarify the relatively quick acquisition of L1s on the basis of 'minimum exposure' to external input. The 'logical problem' of language acquisition, according to UG proponents, is that language learning would be impossible without 'universal language-specific knowledge' (Cook, 1991:153; Bloor & Bloor: 244). The main reason behind this argument is the input data:

"…Language input is the evidence out of which the learner constructs knowledge of language – what goes into the brain. Such evidence can be either positive or negative. … The positive evidence of the position of words in a few sentences the learner hears is sufficient to show him the rules of a language." (Cook, 1991: 154)

The views supports the idea that the external input per se may not account for language acquisition. Similarly, the Chomskyan view holds that the input is poor and deficient in two ways. First, the input is claimed to be 'degenerate' because it is damaged by performance features such as slips, hesitations or false starts. Accordingly, it is suggested that the input is not an adequate base for language learning. Second, the input is devoid of grammar corrections. This means that the input does not normally contain 'negative evidence', the knowledge from which the learner could exercise what is 'not' possible in a given language.

As for L2 acquisition, however, the above question is not usually asked largely because of the frequent failure of L2 learners, who happen to be generally cognitively mature adults, in attaining native-like proficiency. But why can't adults who have already acquired an L1, acquire an L2 thoroughly? Don't they have any help from UG? Or if they do, then how much of UG is accessible in SLA? These and similar questions have divided researchers into three basic camps with respect to their approach to the problem:

Direct access -L2 acquisition is just like L1 acquisition. Language acquisition device (LAD) is involved.

No access - L2 learners use their general learning capacity.

Indirect access - Only that part of UG which has been used in L1 acquisition is used in L2 acquisition.

Proponents of UG, for example, believe that both children and adults utilize similar universal principles when acquiring a language; and LAD is still involved in the acquisition process. This view can be better understood in the following quote.

Advocates of UG approach working on second-language learning... argue that there is no reason to assume that language faculty atrophies with age. Most second-language researchers who adopt the UG perspective assume that the principles and parameters of UG are still accessible to the adult learner.

To support the view above, the acquisition of the third person “-s” can be given as an example. According to research  both child L1 and adult L2 learners (e.g. Turkish learners of English) acquire the third person “-s” morpheme at a later stage of their overall acquisition process and have a great difficulty in acquiring it when compared to other morphemes such as the plural morpheme “-s” or the progressive morpheme “-ing”. This shows that such learners are somewhat affected by UG-based knowledge. However, in the case of foreign/second language teaching it is very well known that the third person “-s” is taught at the very beginning of a second language learning program and presented in a great majority of textbooks as the first grammatical item.

Accordingly, Fodor’s views have some parallels with the UG Theory. Jerry Fodor studied the relationship between language and mind and his view that language is a modular process has important implications for a theory of language acquisition. The term modular is used to indicate that the brain is seen, unlike older views such as behaviouristic view of learning and language learning, to be organized with many modules of cells for a particular ability (for instance, the visual module). These modules, according to Fodor (1983:47), operate in isolation from other modules that they are not directly connected. The language module, if we are to follow Fodor’s ideas, is one of such modules. This modular separateness has been termed as “informational encapsulation” by Fodor. To put it simply, each module is open to specific type of data. In other words, modules are domain specific. This is another way of saying that conscious knowledge cannot penetrate your visual module or language module or any other subconscious module.

Basically, Fodor’s arguments are somewhat similar to that of Chomsky or the proponents of UG Theory in that the external input per se may not account for language acquisition and that language acquisition is genetically predetermined. Add to this, such a modular approach to language acquisition is totally different from the views of Piaget and Vygotsky who have laid the primary emphasis on the role of social or environmental factors in language development.

In the case of foreign/second language teaching, the common view is that inductive learning (teaching a language through hidden grammar or) leads to acquisition. However, dwelling on Fodor’s views as discussed above, it is obvious that inductive learning is confused with acquisition and that by learning something via discovery learning, students just improve their problem-solving skills, but not acquire a language.

As for the problems with Universal Grammar, it can be said that UG’s particular aim is to account for how language works. Yet UG proponents had to deal with acquisition to account for the language itself. “Acquisition part” is thus of secondary importance. A second drawback is that Chomsky studied only the core grammar of the English language (syntax) and investigated a number of linguistic universals seems to be the major problem. And he neglected the peripheral grammar, that is, language specific rules (i.e., rules of specific languages which cannot be generalized). Thirdly, the primary function of language is communication, but it is discarded. The final and the most significant problem is a methodological one. Due to the fact that Chomsky is concerned only with describing and explaining 'competence', there can be little likelihood of SLA researchers carrying out empirical research.

In summary, UG has generated valuable predictions about the course of interlanguage and the influence of the first language. Also, it has provided invaluable information regarding L2 teaching as to how L2 teachers (or educational linguists) should present vocabulary items and how they should view grammar. As Cook (1991:158) puts it, UG shows us that language teaching should deal with how vocabulary should be taught, not as tokens with isolated meanings but as items that play a part in the sentence saying what structures and words they may go with in the sentence. The evidence in support of UG, on the other hand, is not conclusive. If the language module that determines the success in L1 acquisition is proved to be accessible in L2 acquisition, L2 teaching methodologists and methods should study and account for how to trigger this language module and redesign their methodologies. The UG theory should, therefore, be studied in detail so as to endow us with a more educational and pedagogical basis for mother tongue and foreign language teaching.

  Chomsky - the Evidence

1) Acquisition under extreme conditions

a) Neurological evidence

Language functions do appear to be localised in the brain, much as we would expect were Chomsky to be correct in his surmise that language is innate. However, language functions appear to be distributed throughout the brain, and in normal use, the whole brain is brought into play. It is also important to recognise that although neurobiologists now know a lot about the brain, there is also a lot that is not known. The brain is an extremely complex organism.

1)           Normal development of L1 in young children

We saw that Chomsky is certainly mistaken in believing that children hear only partial and ungrammatical sentences. Studies of the ways in which parents, and particularly mothers, interact with their babies and infants show that they use a special kind of language, and take great care to speak in full correct sentences to their children. Nevertheless, the rapidity with which children do learn their mother tongue does suggest that there may be some underlying mechanism that fits them for this task.

It is necessary to note that children in some cultures are not spoken to by their parents directly, and yet they learn their mother tongue all the same. Pinker suggests that the neurotic behaviour of Western middle-class mothers is a parallel to that observed in some African societies, where mothers are very anxious to teach their children to sit up.

2) Language learning under extreme conditions

2.1. Wild Children

From time to time, there appear in our midst beings who challenge our conception of what it means to be human. These beings are often referred to as wild children or wolf children. They are often tragic figures, offering glimpses of what might have been, of fully human intelligence that somehow does not enable them to live a social life. This is particularly true if they are already through puberty when they are found. They suggest to us that there may be a 'critical age', an age beyond which any child who has somehow missed out on learning a language will never completely master one.

The most striking recent case, however, is rather more ambiguous in its results:

In 1970, two women, one of them suffering from cataracts, and partially blind, stumbled into the social services bureau of Temple City, in California, bringing with them a child. At first, the staff thought that the child was about 6 or seven years old, and that she was autistic - she weighed four stone, and stood 4' 6" high. She did not appear to talk.

On further investigation, she turned out to be 13 years old. She could understand some words - about 20, including the colours, red, blue, green and brown, the word 'Mother' and some other names, the verbs 'walk' and 'go' and a few other nouns, such as 'door' or 'bunny'. She could say only two things - 'Stopit', and 'Nomore'.

Why was she in this condition? When she had been about 20 months old, her father, who was suffering from a severe depression, sparked off by the accidental and brutal death of his mother, decided that she was severely retarded, and that she needed protection from the world. This protection he provided by shutting her up in a small bedroom, and leaving her there for the next eleven years.

Genie was attached to a potty by a special harness for most of the day, and then, at night, she would be fastened into a sleeping bag, unable to move her arms, and put into a cot. There was very little sound in the house, for the father forced the rest of the family to speak in whispers. If Genie herself attempted to make any noise, her father would beat her with a stick. On those occasions upon which he felt the need to communicate with his daughter, her father would bark or growl like a dog.

Genie had very little visual or physical stimulation. Hung up in the room were a couple of plastic raincoats, and she was sometimes allowed to play with them. Other small toys - plastic containers, or the TV journal - were sometimes given her. Her feeding was swift and silent, and she had eaten nothing but baby foods and cereals - she did not know how to chew.

Genie was immediately surrounded by a team of scientists. These people were particularly interested in her progress in language. Would she ever learn to speak?

According to the neuropsychologist, Eric Lenneberg, in his book Biological Foundations of Language, 1967, the capacity to learn a language is indeed innate, and, like many such inborn mechanisms, it is circumscribed in time. If a child does not learn a language before the onset of puberty, the child will never master language at all. This is known as the critical period hypothesis. If Lenneberg was right, then Genie, at over 12 years old, would never be able to speak properly. If, on the other hand, she did learn to produce grammatically correct sentences, then Lenneberg was wrong.

At first, a number of the people working with her were convinced that she was going to demonstrate the falsity of the critical period hypothesis. One year after her escape, her language resembled that of a normal 18-20 month old child.

It is at this point that the language of the normal child begins to take off - there is a sudden qualitative change, and the infant learns not only more and more vocabulary, but also more and more complex grammar. But with Genie, this did not happen.

So had Genie's case proven that Chomsky and Lenneberg were right? No, she had not. Lenneberg himself observed that Genie's personal history was so disastrous, that it would not be at all clear why she had been unable to make more progress. It could be that she had been so emotionally damaged by her father's treatment that all learning processes would be interfered with.

Others suggested that perhaps her father had been right in judging that she was mentally abnormal. Brain scans had shown some unusual features - in particular that Genie's brain was dominated by her right hemisphere. Language is mainly situated in the left hemisphere. Was it her brain that was interfering with her language, or was it the lack of linguistic stimulation, and resulting under utilisation of the left hemisphere that had resulted in right brain dominance?

Genie's lack of progress with language is, as so often with the evidence that I have quoted, capable of interpretation either in a Chomskian framework, or in line with Bruner's ideas. Her experience does suggest that, over a certain age, any child who has not learnt a language will have great difficulty in acquiring one. Lenneberg's hypothesis is not proven, but it is strongly supported. Is there further evidence?

2.2. The blind and the deaf

Blind children, particularly when born to sighted mothers, do not receive the same degree of stimulation, and that they therefore fall behind in their linguistic development. In most cases, they catch up pretty quickly - thus comforting the Chomskian line ; parents may hasten the speed of progress a little, or they may hold it back a little, but in the long run, all children brought up in normal circumstances achieve fluency.

What about deaf children? Here there is some evidence that being unable to hear can have long-term effects upon language acquisition. This is true not simply of spoken language, but also of sign language.

ASL (American Sign Language) is a fully articulated language. It has its own grammar, which is not the same as that of English - nor the same as that of French sign language. Often, it is learnt late in life, and when this is the case, the learner 'speaks' it with a foreign accent - and makes the same kind of grammatical errors that a foreigner makes. If the deaf person learns the language as a child, however, they learn it fluently, and can use all the resources that it offers.

A particularly interesting case is that of 'Chelsea'; her behaviour gave her parents cause for concern. They took her to see a series of doctors, who diagnosed her as being retarded. Her family refused to believe this - she was brought up in a very sheltered and loving environment, but never learnt how to speak. Then, at the age of 31, she was taken to see a neurologist, who recognised that she was, in fact, deaf. She was given hearing aids, which brought her auditive capacity up to about normal levels. After therapy, she now scores on IQ tests at levels for a normal ten year old, she works at a vet’s, reads, writes and communicates. But when she speaks, she produces strings of words, with no apparent underlying syntactic structure. Her utterances may be comprehensible in context, but they look nothing like normal sentences.

Other evidence from deaf people is also interesting. Recently, linguists have been showing more and more interest in the language of the hard-of-hearing - Sign language. We now know that Sign Language is a full language - it has a full lexical range, it has a complex syntax, and a complex system of signs, whose relationship to referents is as arbitrary as is that of other languages - even when they seem most iconic. There is not simply one sign language - people who use British Sign Language cannot understand people who use ASL - neither language is directly related to English.

People who learn to sign in adolescence or adulthood are very similar to people who learn a foreign language - they have an accent, and they never master the more arcane syntactic rules. Children who learn do master the language - and, according to Steven Pinker, they master it even when they learn from parents who do not speak it properly. Once again, this is suggestive - children are specially programmed to learn a language, and they lose this skill at puberty - once again, both Chomsky's and Lenneberg's positions appear to be vindicated.

Evidence from neurology is also suggestive - many children who have suffered damage to the left hemisphere are able to acquire a language by transferring language to the right hemisphere. Adults are not able to perform the same feat as easily. Once again, it would seem that Lenneberg may be right - there is a critical period for first language learning.

This obviously interests us as teachers of a second language. Many observers have noted that a second language appears to be more difficult to learn after puberty. Later on, we shall see that this observation has not gone unchallenged, and that for certain kinds of linguistic knowledge, adults and adolescents apparently learn more quickly than children - but it may be that the way that they learn is totally different - whereas children may still call upon the LAD to learn a second language, adults and teenagers have to use other strategies, and in particular, they have to lean heavily upon their first language.


1. For a readable and touching account of Genie's history, see Russ Rymer, 'Genie : A Scientific Tragedy', Penguin, 1994. There are numerous accounts of 'wild' or 'feral' children : I have found Douglas Keith Candland, 'Feral Children & Clever Animals ; Reflections on Human Nature' particularly useful.

2. In this section, I have mainly relied on Pinker. David Crystal, 'The Cambridge Encyclopaedia of Language' is also very informative, as it is on other aspects of language disability, such as aphasia.


3)  Evidence from mother-tongue acquisition

Now let us look at how children actually do learn language.

They may begin to learn in the womb. We know that they react to their mothers' voices from birth - they have been listening to her over the last three months of pregnancy. However, the first noticeable active vocal activity begins at about 8 weeks - the baby begins to coo - at first producing individual sounds, but later stringing them together in a rhythmical pattern. Then, at around 20 weeks, the baby diversifies the sounds she is producing, and gradually starts babbling. Babbling involves a selection process.

- in the first stage, the child appears to produce the whole gamut of sounds used by human beings in the production of speech - it is the tower of Babel indeed.
Bit by bit, however, the range of sounds used narrows down, and the child concentrates more and more upon the sounds used by the mother tongue. She is listening to you. So what is being said to her?

We remember that Chomsky claims that children only hear very partial and ungrammatical input. It is now known that this claim is almost certainly false - adults in our culture, when speaking to children, take great care to phrase their utterances correctly. This is probably not because they are thinking primarily about offering the correct syntactic model, but because they are aiming for clarity of expression. It has been noticed that mothers and other caretakers, when speaking to children, adopt a certain number of specific verbal strategies. The style of speech that they use is sometimes referred to as 'Motherese', although non-sexist linguists prefer to call it 'caretaker talk'. What are the characteristics of this kind of language?

So the language that children hear is by no means necessarily partial and ungrammatical. It has been suggested that these characteristics offer the child such clear samples of language, that there is no need to posit a Chomskian black box, or UG. However, supporters of the UG approach point out that -

·        - large number of Wh- forms.

In Samoa, for example, adults very rarely speak directly to their children, and among some black communities in the US, it is considered a waste of time to speak to children who are too young to give sensible replies - why talk to them, they don't know anything yet? And yet, these children also learn language.

Not only do they fail to copy the utterances their mothers give them, they also produce utterances that they have never heard, and use structures that they have never heard.

Child : Mamma isn't boy, he a girl.

Mother : That's right.

Child : And Walt Disney comes on Tuesday.

Mother : No he does not.

Indeed - and this is of direct interest to language teachers - correction of grammatical form appears to be a waste of time.

The mistakes that the child makes do not appear to be simply random errors. Linguists argue that they are not, in fact , ungrammatical, but that they are based upon the child's own grammar. Interestingly enough, all children tend to make the same kinds of mistakes at roughly the same period in their linguistic development. For example, English-speaking children working on negation go through a predictable sequence :

  1. First the negative words 'No' and 'Not' appear as single word sentences.
  2. These combine with other words to form two-word sentences - 'No car', 'Not gone' etc.
  3. During third year - negative words used within constructions

You no do that, Mummy

You not got it

while negative auxiliaries also appear. - Won't, can't

As we shall see, there are interesting similarities between this sequence, and the sequence of acquisition of the negation in English by second-language learners.

Followers of Chomsky claim that the regularity of these errors, and the fact that they are not based upon what the child hears, demonstrate that they are derived from the Universal Grammar. The child works through from the simplest possibilities offered by the UG to the more complex, until his own grammar is the same as the grammar of the mother-tongue. The claim is almost that the child does not make mistakes, but simply has a different grammar to the grammar of the adult.


The evidence from neuroscience and from first-language learning is suggestive. We find a number of observations that do fit in with what we would expect if Chomsky were right. However, the evidence needs to be treated with caution.

We have also seen that Chomsky is certainly incorrect in his claim that children do not hear well-formed language. On the other hand, children do seem to understand almost instinctively that language is a rule-bound system, and are capable of discovering the rules underlying their mother tongue with remarkable rapidity. But it needs to be borne in mind that the fact that children seek out the rules underlying language does not mean that they necessarily have a specific approach to language itself. It may simply be a product of the peculiar nature of human intelligence, which makes us look out for and be sensitive to the underlying rules which govern phenomena in the world - this is one of the main characteristics of all human cultural activities, and not just of language-learning.