The Analogy between First- and Second- Language Learning

Vivian Cook

IRAL, 1969, VII/3

Later comparisons in The Relationship between First and Second Language Learning Revisited 2010 and First and second language learning 1979

In recent years a number of the techniques that have come into language teaching have relied implicitly on their being a close analogy between the way that a child acquires his native language and the way that a student learns a foreign language. This paper does not try to test the strength of this analogy. Instead, it takes one side of the analogy, first-language acquisition, and argues that the implications for second-language teaching are rather different than generally supposed1.

One of the main areas of interest to transformational generative grammarians has been the process of language acquisition. "The problem of internal justification - of explanatory adequacy - is essentially the problem of constructing a theory of language acquisition, an account of the specific innate abilities that make this achievement possible."2 This interest has provided linguists and psychologists with new insights into the way that a child acquires and develops his competence in his native language, particularly into the development of grammatical systems3. The chief difference from earlier studies is that the competence of the child used above all to be compared with that of an adult: his competence was analysed in terms of what it was eventually to become —the competence of an adult native speaker of the language. In other words, the child's speech was treated as a simplified telegraphic version of adults'.

Recent work treats the problem rather differently by regarding the child's competence at a given age as a self-contained internally consistent system not dependent on the full adult system. It seems absurd to describe, in effect, the child as possessing all the rules of adult competence together with a set of deletion and reduction rules to account for his ungrammatical sentences. The aim now is to describe the different stages through which the child progresses towards adult competence, with each stage having a grammatical system of its own that does not need to be explained by reference to the adult system. The child is believed to make a series of hypotheses about the structure of the language which he tests and abandons or preserves. Each successive hypothesis is an interim grammar accounting more successfully for the data he is exposed to. The last hypothesis is the final adult grammar of competence in the language.

An illustration of this is provided by the work of Klima and Bellugi on negation in English.4 They describe three stages in the child's development. At Stage 1, the rule for negation merely states that a Sentence Nucleus can be preceded or followed by 'no' or 'not'. The child produces sentences such as "No singing song." "Not a teddy bear." "Wear mitten no." In terms of the child's grammar, all these are grammatical; they bear little relation, however, to grammatical sentences produced by an adult. To account for Stage 2, this simple rule has to be modified to permit 'don't' and 'can't' (which are only found in the negative) and negative elements "within the sentence but not connected to an auxiliary verb". The child produces sentences such as "I don't sit on Cromer coffee." "No pinch me." "That no fish school." At Stage 3, the rules must allow for the appearance of auxiliaries in other than negative forms and for 'some' occurring in both positive and negative sentences. The child produces sentences such as "I don't want cover on it." "I didn't see something." "That not turning." It is apparent that he has still to acquire some rule to account for the relationship of negation and 'any'.

There appears to be very little connection between evidence of this kind and the conventional theory of second language learning. One can distinguish three areas of divergence: development, error, and grading.


The form of development theory implicit in second-language teaching has been similar to the older first language acquisition theory outlined above. The competence of the second-language learner has been compared with the full native competence at which he is aiming. The products of his successive grammars are evaluated, not by their consistency with his interim grammars, but by their conformity with the rules of full native competence. Teachers usually demand that the sentences of second-language learners should be grammatical from the very beginning, a demand not imposed on children acquiring their first language. It is possible that an approach which heals the stages through which the learner progresses as self-contained will be as fruitful in second-language learning as in first-language acquisition.

In addition, the second-language learner is not expected to make interim hypotheses about the language he is learning; instead, he is assumed to learn the rules of native competence one by one. He is required to build up his grammar as one would a house, brick by brick: "Add each new element or pattern to previous ones."5 The new element is not supposed to make him have cause to modify those rules he has already learned. The child, on the other hand, constructs and destroys or modifies a series of bigger and better tents. The two processes involved appear quite distinct; we learn our first language by a series of evolving hypotheses; we are assumed to learn a second language by building it up rule by rule. It is clear that there is no analogy between the conventional theory of second-language development and what we know today about development of the first language. Only further research will show whether this conventional theory is wrong and whether the two processes do, in fact, develop in a similar manner.6


Tested by adult competence, the child's sentences will contain errors. Similarly, tested by native competence, the second-language learner's sentences will also contain errors. While, in the theory of first-language acquisition outlined above, 'errors' are an integral part of the process and show what the child's interim grammar does not yet include, in second-language teaching it is usually thought that errors are extremely harmful.7 Usually the teaching situation is carefully controlled so that the possibility of error is minimised. This limitation prevents the learner from making errors by trying to use parts of the language he has not yet been taught. According to Newmark and Reibel, the explanation for interference is that the learner fills in the gaps in his experience of the second language with forms from his first language.8 If the second-language learner is not to make constant errors, the environment in which he is using the language must be tightly controlled so that the number of grammatical rules and lexical items necessary is severely limited.

In first-language acquisition, an error shows that adult competence has not yet been reached and the grammar is still an interim hypothesis; in second-language learning, an error is taken to show that an item has been wrongly learned.9 If the second-language learner is to proceed by a series of makeshift hypotheses, he too must be allowed great freedom to err (in terms of native competence) so that he can test his hypotheses and abandon those that are unsuccessful. Take, for instance, a recent book The Teaching of English to Immigrant Children.10 The authors suggest that immigrant children should be discouraged from using 'pidgin'. Their examples of pidgin are "Black pencil no", "Me cut paper no", and "Ghulam no give glue." It is immediately apparent that the immigrant children are following the stages outlined by Klima and Bellugi through which native children go; the first two sentences conform to the negation rule for Stage 1, the third to Stage 2. If the analogy holds, then far from deploring these errors, we should be commending the children's progress towards native competence along the same road followed by the English child.


The child hears a virtually unrestricted input of gram­matical and ungrammatical sentences; he produces more and more comprehensive systems of rules to account for them. It is thought that, apart from certain inherent limitations in the child's environment, the language he hears is not graded systematically. (This disregards, of course, the effects, if any, of baby-talk.) In second-language learning, the input has usually been highly restricted and systematically ordered. It has also consisted solely of grammatical sentences. The way to native competence has been assumed to be through minimal steps in carefully restricted situations.

Let us now turn to some more specific results with implications for second-language teaching. In her article "Imitation and Structural Change in Children's Language", S. Ervin deals among other things with past-tense formation in English.11 She found that the forms occurred in the following order in her material:

i. Irregular past forms.

ii. Irregular extensions of regular past forms, e.g., "buyed", "corned", "doed". 

iii. Regular past forms.

As she writes, "The odd, and to me astonishing thing is that these extensions occurred in some cases before the child had produced any other regular past tense forms according to our sample". McNeill uses her data to show that practice is by no means essential to first-language acquisition.12 The children had learned the irregular past forms correctly and had practiced them a number of times yet, when they started to produce regular past forms, the irregular forms rapidly underwent analogous change in spite of the practice they had received. "Apparently patterns weigh more heavily with children than frequency of repetition does."

Until recently, the position in second-language teaching was almost the reverse; practice was thought to be the most important element in learning a second language. "The student must be engaged in practice most of the learning time."13 Indeed supporters of the language laboratory often argue in favour of overlearning, i.e., further practice beyond the point when an item is learned. According to McNeill, practice is not relevant to acquiring the native language. If it is also irrelevant to second-language learning, then the second-language learner must be given more opportunity to perceive patterns at the expense of time devoted to practice.

Other implications are that, once again, 'error' is integral to first-language acquisition; that the child comes to no harm by using an incorrect form a large number of times; that the grading of language courses appears to be entirely different from the progression revealed in sentences produced by a child: in short, that there are few similarities between the way in which a child acquires the past tense in English and the way in which a foreign learner is usually taught it.

Another article of importance is by Bellugi and Brown, "Three processes in the child's acquisition of syntax."14 The processes they describe are: (i) imitation and reduction, (ii) imitation with expansion, (iii) induction of the latent structure. The third has already been discussed above. Imitation and reduction is the process at work in the following examples;

Model utterance.                           Child's imitation.
Wait a minute.                                 Wait a minute

will be unhappy.                    Fraser unhappy
not the same dog as Pepper.        Dog Pepper.

What the child omits in his imitations is usually the grammatical items; what he preserves is the content words. Indeed his ability to repeat grammatical items is little in advance of his ability to produce them.15 Brown and Bellugi have two explanations for this 'telegraphic' language. One it that "Perhaps the children are able to make a com­munication analysis of adult speech and so adapt it in an optimal way to their limitation of span". The other that it is due to 'differential stress'; "the heavier stresses fall, for the most part, on the words that the child retains." In English the stresses would tend to fall on content words and so the child retains content words.

It is true that most second-language teaching today makes extensive use of repetition of one type or another. Yet this is always, as far as I am aware, complete repetition; a learner who repeats only part of a sentence is corrected and encouraged to repeat it in its entirety. He is not expected by his repetition to show evidence that he has abstracted the grammatical rule or content words he is being taught. That is to say, he is required to repeat "The cat sat on the mat" neither as "cat sat mat" nor as "the on the".

Second-language teaching also seems to attach much less importance to content words. While they are usually carefully selected and graded, the chief aim has been to give the learner grammatical rules he can use rather than lexical items he can use in them. The primary grading has then been grammatical. Yet in the child's imitations the grammatical items appear the least important part.

Imitation with expansion is the process involved in the following examples;

Child                     Mother
Baby highchair.      Baby is in the highchair.
Mommy eggnog     Mommy had her eggnog
Sat wall                  He sat on the wall
Throw Daddy         Throw it to Daddy

This may be considered almost the opposite to the preceding process; instead of the child repeating a sentence in a reduced form, the mother expands a sentence produced by the child. She supplies the grammatical items absent from the child's sentence (in terms of adult competence) while preserving the content words in the same order. Unlike the child reducing his mother's sentences, the mother has to choose one out of a number of alternatives when she is expanding his sentences. "Baby highchair" might be expanded into "Yes, that's baby's highchair, not yours" or "Where is baby's highchair?"; the mother's actual expansion "Baby is in the highchair" is due to the nonlinguistic situation at the time it was spoken. The mother's expansions encode "additional meanings at a moment when he is most likely to be attending to the cues that can teach that meaning".

As the demand in second-language teaching is usually for the learner to produce grammatically flawless sentences, no use has been made of this process. The learner is not allowed to produce what appear to be reduced sentences en route to full native competence; it is not then possible to expand his sentence as a teaching technique. Instead his sentences are corrected rather than expanded. The mother appears to treat her child's sentence as sacrosanct in order and in content words; the foreign-language learner will have both frequently corrected. His sentence is treated as 'wrong' rather than as deserving sympathetic situational interpretation.

This process lends little support for the value usually given to correction in foreign-language teaching. Nor does it provide a justification for the concept of reinforcement as used in structure drills; the child is rarely rewarded by automatically hearing one correct response, as the second language learner is. Instead his reinforcement consists of hearing an expansion of his sentence into a situationally appropriate form that is grammatical in terms of adult competence. As the evidence suggests that about thirty per cent of the child's sentences are expanded in this way, the process of imitation and reduction seems likely to prove one of the most important ways in which children learn to speak and one of the most neglected ways in which people are taught a second language.

One further process should be mentioned - that of verbal play. This is described by Ruth Weir in her account of her son's bedtime monologues Language in the Crib.16  The following is a typical example; "Stop. Have to stop. Stop. Stop it. Stop the ball. Stop it. Stop the ball please. Take it. Stop it. You take it. Please take it."

Superficially, this activity is very similar to second-language teaching by means of structure drills.17 Where it differs from a structure drill is in the lack of external 'inputs'. The child supplies the item to substitute out of his own head rather than having it supplied to him by an outside source. He appears to be playing with language divorced from situation; his 'drills' seem to have the function of exploring his interim grammar to see what sentences it will produce. In second-language teaching, drills are commonly used for intensive practice rather than for testing of hypotheses.

Other aspects of first-language acquisition are not yet sufficiently well defined to make any analogy with second-language learning possible. In particular, one might mention the controversial issue of what the child brings to language acquisition; is he tabula rasa or does he contribute a set of innate ideas that enable him to construct a grammar of competence out of the language input he receives? While transformational generative grammarians agree at present about the existence of innate ideas, they differ as to their nature. Some think they are chiefly "the basic grammatical relations and the general idea of a transformation"18; others that they are the semantic features underlying grammatical categories.19 Whatever they may be, their very existence poses two problems for second-language teaching: first, that there are far more similarities at a deep level between human languages than hitherto supposed, thus shedding a new light on the question of interference; secondly, that the capacity for language acquisition and the innate ideas possessed by the child may not be so readily available to the second-language learner, suggesting that teaching may be more effective if it makes use of faculties other than that of language in the learner.

There seems to be little similarity, then, between the process of first-language acquisition as it is understood today and the process of second-language learning as implicit in present-day teaching. A method for teaching foreign languages that could justifiably claim to be based on first-language acquisition would have to meet at least the following requirements:

1.    That it would allow the learner to progress by forming a series of increasingly complete hypotheses about the language.

2.    That, consequently, it would permit, and indeed encourage, the learner to produce sentences that are ungrammatical in terms of full native competence, in order to test these hypotheses.

3.    That it would emphasize the perception of patterns rather than the intensity of practice.

4.    That its teaching techniques would include partial repetition of sentences, verbal play, and situationally appropriate expansions of the learner's sentences.

No method can at present claim to fulfil these requirements. It remains to be seen whether they can in principle be fulfilled, whether, in fact, the analogy between first- and second-language learning is sound.


1. For further discussion of this analogy see: L. Newmark and D. A. Reibel, "Necessity and Sufficiency in Language Learning," IRAL, VI/2 (1968). Wilga M. Rivers: The Psychologist and the Foreign-language Teacher (Chicago and London: The University of Chicago Press, 1964), pp. 99-104.

2.  Noam Chomsky: Aspects of the Theory of Syntax (Cambridge, Massachusetts: The M.I.T. Press, 1965), p. 27.

3. For a summary of this approach see:
David McNeill: "Developmental Psycholinguistics," in:
The Genesis of Language: A Psycholinguistic Approach, ed. Frank Smith and George A. Miller (Cambridge, Massachusetts, and London, England: The M.I.T. Pr (1966).

4. E. S. Klima and Ursula Bellugi: "Syntactic regularities in the speech of children," in: Psycholinguistics Papers: The Proceedings of the 1966 Edinburgh Conference, ed. J. Lyons and R. J. Wales (The Edinburgh University Press, 1966).

 5. Robert Lado: Language Teaching: A Scientific Approach (New York 1964), p. 53.

 6, For some indications of what such experiments might consists of see A. S. Reber: "Implicit Learning of Artificial Grammars," Journal of Verbal Learning and Verbal Behavior, VI, 6 (1967).
Sol Saporta, Arthur L. Blumenthal, Peter Laekowski, and Donald G. Reiff "Grammatical Models and Language Learning," in:
Directions In Psycholinguistics, ed. Sheldon Rosenberg (New York and London, 1965)

7. For a fuller discussion on similar lines see: S. P. Corder, "The Significance of Learner's Errors," IRAL, V/1 (1967).

8. Newmark and Reibel: pp. 158-159.

9. Cf. Rivers p. 102. "... in the audio-lingual method, the student . . . must

10 John Stoddart & Francis Stoddart (1968), The Teaching of English to Immigrant Children, London

11 S.M. Ervin, 'Imitation and structural change in children's language, in New Directions in the Study of Language, ed. E.H. Lenneberg, MIT Press

12 McNeill Developmental Psycholinguistics, p.71

13 Lado, p.55

14. R. Brown and U, Bellugi: "Three Processes in the Child's Acquisition of Syntax,'' in New Directions in the Study of Language, ed. Eric H. Lenneberg. Massachusetts: The M.I.T. Press, I (MM).

15. Cf Ervin: p. 172.

16. R.H. Weir (1962) Language in the Crib, Mouton

17 V. Cook (1968), 'Some types of oral structure drill', Language Learning, XVIII, 3/4

18. D.I. Slobin (1966), 'Comments on "Developmental Psycholinguistics",' in The Genesis of Language, ed. F. Smith and G. Miller, MIT