Chomsky（1959），in an extensive discussion of Skinner's（1957）Verbal Behavior，delivers the first serious attack on these behaviorist ideas about language learning.Chomsky argues that human behavior is considerably more complex than animal behavior.Moreover，certainly language behavior is so specific to humans that it could never be explained through animal behavior.According to Chomsky，a description of language behavior cannot be just a description of external stimuli and concomitant responses，but it primarily has to be a description of the innate ability of human beings to learn a language.Chomsky discusses Skinner's theoretical concepts point by point and questions their relevance by demonstrating that the conclusions Skinner draws from laboratory experiments with animals cannot lead to conclusions about human behavior，let alone language behavior.Chomsky furthermore calls Skinner'sspeculations premature in the sense that little can and should be said about the language learning process before we have gained a better understanding of the linguistic system that is learned.

Chomsky's（1959）Skinner-review heralds a revolution in ideas about language learning；a revolution that took place in the 1960s.Until that time，most attention had been paid to external linguistic factors.After 1960，the contribution of the main factor in the learning process，the child itself，began to play a more and more dominant role.This revolution was strongly influenced by the rapid rise of a new development in linguistics that can also be traced back to Chomsky：transformational-generative grammar（T-G G）was a source of inspiration for all sorts of experiments in language learning research.

In T-G G it is assumed that the ability to learn languages is innate.The so-called Language Acquisition Device or LAD enables the child to make hypotheses about the structure of language in general，and about the structure of the language it is learning in particular.This is not a conscious process.The hypotheses the child subconsciously sets up are tested in its use of language，and continuously matched with the new linguistic input that the child obtains by listening to what is said in its immediate environment.This causes the child's hypotheses on the structure of language to be changed and adapted regularly：the child develops its rule system through a process of systematic changes towards the adult rule system.

This view of the language learning process，therefore，stresses the mental activities of the language learner himself，and strongly questions the relevance of such external factors as imitation，frequency of stimulus，and reinforcement.

For the behaviorists，learning can be the same for every individual because it is socially conditioned.In teaching we can ensure that everybody learns equally well by making sure that the conditions of learning are the same for each.An alternative view，which is called‘mentalist’，contradicts the behaviorists at almost every point.Everybody learns a language，not because they are subjected to a similar conditioning process，but because they possess an inborn capacity that allows them to acquire a language as a normal maturational process.This capacity is by definition universal.In a sense then，the mentalists too argue that language learning is the same for everybody，but the similarity ends there，because for the mentalists what the learnersshare is a capacity the existence of which the behaviorists would deny.

For the mentalists，language is far too complex a form of behavior to be accounted for in terms of features external to the individual.In a review of Skinner's account of verbal behavior，the linguist Noam Chomsky demonstrates that this brand of behaviorism at least is quite incapable of explaining our ability to learn and use our mother-tongue.He attacks with particular vehemence the notion that language responses are under the control of external stimuli—that，as he puts it，the individual is merely the focus of behavior and not its cause.He suggests that Skinner himself cannot maintain this view and that there is evidence in his book to contradict it.For him the most important thing of all is that human beings use language whereas other animals do not.It is no use applying principles of learning that have been derived from research with animals，as he says the behaviorists do，to explain a form of behavior that animalsare not capable of.Since all normal human beings learn the language successfully，they must possess some internal capacity for language that other animals do not have.Since this capacity cannot have been acquired socially it must be innate.It is their willingness to admit the possible existence of unobservable internal mechanisms that leads these linguists to be considered mentalistic.Such mentalism，however，need not be regarded as an escape from rigorous scientific procedures.They would argue that the nature of language is such that it is impossible to explain it without postulating an innate mechanism of a fairly well-defined kind.

To the innate mechanism that they put forward the name‘language acquisition device’has been given.It is said to operate in the following way.A child，starting from birth，is exposed to language that acts as a trigger for the learning device.The device has the capacity to formulate hypotheses about the structure of the language to which it is exposed.The child is，of course，quite unconscious of this process.The hypotheses are tried out in the child's own language production and are regularly checked against the further data that his exposure to the language provides.As he finds that his hypothesis cannot account for all the data，he modifies the hypothesis and checks it again.The first hypotheses are very simple indeed.Most children pass through a stage of twoword utterances for example，in which they appear to operate on the hypothesis that there are two classes of words，one limited and the other more or less unrestricted in number，which occur in a fixed sequence.As the child gets older，the hypotheses become more and more complex and，applying them to his own use of language，he brings hisspeech closer and closer to the adult model to which he is for the most part exposed.What the child is doing is constructing an internal grammar of the language.This grammar passes through successive modification until it becomes the complete grammar of the adult language.At this point it should be identical with the descriptive grammar that the linguist attempts to write.

The arguments in favor of this view are twofold.First，the nature of language structure is such that the child must have some such device.Any other attempts to explain language learning are at best incomplete because they cannot account for the learning of all structural relations.Secondly，there is some evidence from the observation of the language or young children that seems to support the mentalists'account of language learning.If their theory is correct，many of the rules that the child formulates will be incorrect or incomplete.If these are then applied to the child's own language production，the result should be error in the child's speech.This is just what happens.Here，for example are forms that have been observed in the speech of children：

I breaked（or even‘broked’）my lorry I better go to bed now，bettern't I？

The significant thing is not that these are mistakes.Anyone can make mistakes in speech.They are mistakes that could not be due to faulty hearing or imitation，since they quite unlike any utterance that the child will have heard from an adult.Something other than imitation of adult speech is going on.The child is operating two overgeneralized rules，the first saying that the past tense of break is formed by the regular process of inflection，the second that better is a modal auxiliary verb like can or must and can be repeated in question-tags like those verbs.The making of error is now seen as an inevitable part of the language learning process.It is possible that it is not only inevitable，but also necessary since it provides the only means that the child has of finding out the limits to the domain of the rules that he is formulating.What is the role of the social factors to which the behaviorists attach so much importance？For Chomsky they have virtually no role at all.The nature of the language acquisition device and its mode of operation are inviolable.The belief is that with research，a clear developmental sequence will emerge，implying that given the nature of the device and the language to which it is exposed，learning will follow a predictable path.The stages through which children pass will be very largely the same.They do not，however，necessarily learn at the same rate and it is here that the possible influence of schedules of reinforcement can be admitted.But from the linguist's point of view，rate of learning is far less significant than the nature of the learning mechanism itself.

Observation of children learning language also suggests that there are occasions on which pieces of language are learned simply through being heard.A word may be heard，perhaps once，perhaps more than once，but not produced by the child at the time it is heard.There is therefore no active responding and consequently no reinforcement and no repetition.In spite of this，the child may suddenly produce the word quite correctly in a totally new context.However useful the behaviorist notions of reinforcement and repetition may be，then，they do not relate to conditions that are essential for learning to take place.

There is another way in which the behaviorists and the mentalists may differ，though it is possibly only a difference in the language that they prefer to use.Whereas behaviorists will talk of the child using‘analogy’in the construction of sentences，the mentalists prefer to think in terms of the production and application of‘rules’.