The Comparison of Language Development in Native Children and Foreign Adults

Vivian Cook
SLA Home
Later relative clause experiment


IRAL 1973, 15, 1-20
NB Footnotes have become endnotes.
More up-to-date papers: 1977, 2008. To my knowledge, this is the first SLA paper to use the elicited imitation method borrowed from L1 research.


It has recently been suggested that second language teaching can benefit from a closer study of current ideas about first language acquisition. This suggestion is at the moment highly speculative and rests on an assumption that the two processes of first language acquisition and second language learning are identical, or at least very similar. Two separate issues are involved in this assumption; evidence in favour of one may prove nothing about the other. One issue is whether children can learn a second language in the same way they acquired their first. At the moment research such as Daniel Dato's suggests that this may very well be true. The other issue is whether adults can learn a second language in the same way as native children acquire their first. Little concrete evidence has yet been produced for or against this proposition. Teachers have, however, often asserted quite flatly that it is untrue, supported by anecdotes of their own teaching experience. The present article tries to provide more information about this second issue by adapting some procedures from developmental psycholinguistics to the comparison of native children and foreign adults. It must be considered more in the light of a pilot experiment than as a full-fledged scientific investigation. It must also not be forgotten that even if it is established that foreign adults can learn a second language in the same way as native children learn their first this is not the same as saying that they should.

The chief problem in this comparison are the differences between children and adults that are not a question of language. The adult is mature in many other respects than language. We can then anticipate finding differences between children and adults that reflect the adult's superior stage of general mental development rather than different processes of language learning; features of child language not encountered in foreign adults may be characteristic of limited mental development rather than limited linguistic competence. To a great extent this involves the relationship between language acquisition and cognitive development. At the same time that the child is acquiring language, he is also developing other cognitive processes; the extent to which language interacts with these is as yet unknown. It might be that language is central and that cognitive development depends upon it; it might be that language is subordinate to cognitive development or that it is entirely independent. Or each of these relationships may be true of some part of language. The studies that have been done within a Piagetan framework suggest that, while children at different stages of conceptual development have particular preferences for syntactic structures, these are symptoms rather than causes.

A further problem is the factor of language teaching. In first language acquisition, the question of teaching hardly arises: despite wide ranges of environment and of types of interaction with adults, most children acquire adult competence; 'teaching' cannot be of crucial importance. However, when looking at foreign adults, teaching almost invariably plays some role; few adults have not been taught the foreign language by one method or another; it is hard to find adults who have not been tampered with, so to speak; probably more information about untaught foreign adults will become available through the studies being conducted of immigrants. But for the most part second language learning is inevitably seen through a distorting lens; it may prove impossible to separate the effects of the learning process in the adult's mind those of the teaching method or technique. What consistent differences there may be between native children and foreign adults may be the accidental by-products of teaching rather than the inevitable consequences of two distinct processes.

1. A Comparison of the Imitation and Comprehension of Relative Clauses

This compares the performance of native children and foreign adults on the same tasks. The children numbered 24; they ranged in age from 2:11 to 4:9, the average age being 3 :6; they were all attending a nursery school or play­group in the London Borough of Ealing. From now on they will be referred to as the C group and their imitations will be prefixed by (C). The foreign adults numbered 20 and were all attending elementary classes in English for Foreigners at Ealing Technical College; their mother tongues were mixed: Spanish (3), Serbo-Croat (4), Persian (4), Italian (2), Indonesian (2), Arabic (2), Polish (1). They had all been studying English in England for less than a year though they had had a varying amount of English teaching in their own countries. These will now be referred to as the F group and their imitations marked as (F).

The children and adults were tested individually. They were shown a picture and were read a sentence which the picture illustrated. They had to repeat the sentence and, with some of the sentences, to answer a comprehension question. So far as the children were concerned this activity was thought to approximate to an adult reading aloud to a child; so far as the adults were concerned it was similar to the techniques involved in an audio-visual course, with which they were all familiar. The test started with two warm up sentences to ensure that the instructions ("Say after me" or "Repeat what I say") were understood. The remaining 13 test sentences were then read in four different random orders in rotation and after four of the sentences a comprehension question was asked. Everything that was said was recorded on a tape-recorder and later transcribed.

 This assumes that the learner's imitations will depart from the model sentence in systematic ways and that these departures will reveal information about the learner's competence. D. Slobin and C. Welsh, for example, infer parts of the child's competence from his imitation of "The man who I saw yesterday got wet" as "I saw the man and he got wet"; R. Troike shows some of the differences between standard English and Negro English by means of the Negro child's repetition of "She cleans her teeth with her brush" as "Her clean her teeth with her brush". The technique is open to the objection that, by divorcing language from meaningful use, it sets a lower limit to the learner's capacity rather than showing him at his best. For this reason a minimal context was provided for the test sentences by means of the pictures.

Table 1.
Test sentences and questions for Comparison 1.

The sun is shining.
The horse is pulling the cart.
The girl that is washing the dog is little.
Q. Is it a little girl or a little dog?

The food that is cooking is nice.
The hammer that is breaking the cup is big.
Q. Is it a big hammer or a big cup?

The boy that is helping the man is wearing a shirt.
The ball that, the girl is bouncing is small.
Q. Is it a small ball or a small girl?

The lady the boy is drawing is funny.
Q. Is it a funny lady or a funny boy?
9. The dog eating the bone is black.

This is the man that drives the bus.
11. This is the door that opens.
12. This is the knife that cuts the meat.
13. The girl is watching the boy that is climbing the tree.

This is the horse that the boy rides.
This is the paper the man reads.

Table 1 lists the sentences and questions that were used. The test sentences reflect various syntactic points of the relative clause in English. Some con­trast relative clauses that qualify the Subject of the sentence with those that quality the Object; sentences 3 — 9 qualify Subjects, sentences 1015 Ob­jects:

        4. The food that is cooking is nice,


        14. This is the horse that the boy rides.

Others contrast the relative pronoun (which was always "that") being present and being absent (8, 9, 15):

        7. The ball that the girl is bouncing is small.


        8. The lady the boy is drawing is funny.

The function of the relative pronoun in the relative clause also varied; in sen­tences 3—6, 9—13 it is the Subject, in 7, 8, 14, 15, the Object;

        10. This is the man that drives the bus.


        14. This is the horse that the boy rides.

Two sentences had Noun Phrase Objects in both the main clause and the re­lative clause (6, 13);

        13. The girl is watching the boy that is climbing the tree.

Finally some of the sentences differed in terms of case grammar. In sentences 3, 6, 10, 13, the surface subject of the relative clause was derived from an underlying Agentive;

        3. The girl that is washing the dog is little.

In sentences 4 and 11, the surface subject of the relative clause comes from an underlying Objective;

        11.This is the door that opens.

In sentences 5 and 12, the surface subject of the relative clause comes from an underlying Instrumental:

        5.The hammer that is breaking the cup is big.

Comprehension questions were asked after sentences 3, 5, 7, 8. The form of the question (3. "Is it a little girl or a little dog?") was intended to elicit whether the syntactic status of the relative clause had been appreciated, i. e., whether the subject realised that the sentence contained an embedded sentence. The order of the items in the question alternated between that in the model sentence and the reverse, that is to say 3. Is it a little dog or a little girl?

Let us now look at the results. Taken overall both children and adults performed rather badly. Counting silence as an attempt, the C group had word-perfect imitations only 25 times out of 31 2 attempts (8 per cent), the F group 69 out of 260 (26.5 per cent), (p = < .01). Even if the substitution of "what" for "that" is counted as word-perfect imitation, this only improves to 16.6 per cent for the C group and 26.9 for the F group, (p = < .01). The types of error could be categorised fairly easily. One common error was the omission of "that" from the sentences in which it occurred. In those sentences where "that" is the Object of the relative clause, its omission still leaves a grammatical sentence. This occurred twice in the C group as in

        (C) 14 this is the horse the boy rides

and five times in the F group,'

        (F) 7 the ball the girl is bouncing is small.

Where "that" is the Object of the relative clause, its omission leaves an un-grammatical sentence. This happened fourteen times among the C group, (C)5 the hammer is breaking the cup is big and twenty four times among the F group, (F) 12 this is the knife cuts the meat

Another frequent type of error was to substitute another relative pronoun for "that". The C group substituted "what" 22 times,

        (C) 10 this is the man what drives the bus

but the F group did so only once,

        (F) 12 This is the knife what cut the meat.

In some of the sentences "that" could be replaced by grammatical alternatives.

The C group did so 24 times,

        (C) 12 this is the knife which cuts the meat

        (C) 6 the boy who's helping the man has got a shirt

and the F group used grammatical alternatives 8 times. The C group, though not the F group, also replaced "that" by ungrammatical alternatives on 5 oc­casions,

        (C) 5 the hammer who's breaking the cup is big.

Three of the sentences could also have a relative pronoun added to them (8, 9, 1 5); the C group added one 8 times, the F group 5.

Table 2 compares the performance of the groups on 'Subject' relative clauses (3—7) in terms of word-perfect answers, answers with substituted "what", deleted "that" and with other substitutes;

Table 2. Comparison of C and F groups on 'Subject' relative clauses (3 - 7)

Correct "What" substitution "that" deletion Other substitution
2 5 14 15
20 0 19 5

 Table 3 does the same with the 'Object' sentences (10—14).

Table 3. Comparison of C and F groups on 'Object' relative clauses (3 - 7)

Correct "What" substitution "that" deletion Other substitution
15 17 2 14
30 1 10 3

Both groups tend to imitate 'Subject' relative clauses better than 'Object' ones, in terms of word-perfect imitations, less deletion of "that" and less substitution of other items, except for the C group's use of "what". The comparison of C group 3 — 7 with C 1 0—14, of C group 3—7 with F group 3 — 7, and of C group 1 0—14 with F group 10—14 are all statistically significant (p = < .01); only the comparison of F group 3 — 7 with F group 10—14 does not reach the required level of statistical significance. This finding that 'Object' clauses are easier than 'Subject' clauses is not surprising in the light of experiments that have shown that syntactic complexity is more confusing near the beginning of the sentence than near the end.

Table 4 compares the groups on sentences differing in terms of case grammar.

Table 4 Comparison of C & F groups on 'Agentives' (3,10), 'Objectives' (4,11) and 'Instrumentals' (5, 12)
    Correct "What" substitution "that" deletion Other substitution
'Agentives' C 5 4 0 6
  F 16 0 1 0
'Objectives' C 5 7 3 3
  F 10 0 7 1
Instrumentals' C 3 4 4 4
  F 12 1 5 2

Little differences emerged between the groups or between the sentences. Again the figures comparing the C and F groups are statistically significant (Agentives, p = < .01; Objectives and Instrumental, p = < .05) but com­parisons of the sentence pairs inside the groups are not.

Another type of error consisted of 'recording' the syntactic structure of the sentence while preserving its meaning. Sometimes this might be recoding it as a sequence of two sentences or clauses. The C group did this 9 times, sometimes keeping a syntactic link between the clauses,

        (C) 4 the food's cooking and it's nice

         (C) 5 the hammer is breaking the cup 'cos it's big

sometimes simply juxtaposing them

        (C) 9 the doggie's eating the bone it's black

The F group recoded 8 times in this manner as in

         (F) 4 the food's cooking that's nice

        (F) 14 this is the horse and the boy rides

A second type of recoding consisted of shifting the Subject of the relative clause to the main clause,

        (C) 7 the girl is bouncing the ball is small

        (C) 14 this is the boy what rides

This occurred 11 times with the C group but didn't occur among the F group.

There was, however, an error made by the foreign adults in sentence 13 that could be related. This did not preserve the meaning but interchanged the Subject and Object of the main clause,

        (F) 13 the boy is watching the girl that is climbing the tree.

This occurred in one form or another 6 times from the F group but not at all from the C group.

Another difference between the groups was that the C group 43 times re­peated only the last word or two of the sentence,

        (C) 3 is little

and a further 11 times they repeated the last 3 or 4 words,

       (C) 12 cuts the meat.

This never happened with the F group. However, what might be considered the opposite strategy, namely repeating the first 1—4 words of the sentence, was used by both groups, the C group 8 times, the F group 1 3 times.

Let us now turn to the answer to the questions asked after sentences 3, 5, 7 and S. The C group were right 46 times, wrong 42; the F group were right 53 times, wrong 22, (p = < .05). There was a slight tendency, statistically non­significant, for Case to influence comprehension, the 'Agentive' sentence 3 being the most difficult for both groups and the 'Objective' sentence 7 the easiest. To correct one possible misapprehension, omitting the relative pronoun in sentence 3 did not necessarily mean that the comprehension question would be answered wrongly, i.e. that the sentence was being perceived as a succession of two sentences; of the 3 C group who deleted "that" only 1 got the answer wrong; of the 7 F group only 3 got the answer wrong.

To sum up, both groups seemed to have tackled the tasks of imitation and comprehension of relative clauses in much the same manner. The similarities are: (i) the low proportion of word perfect imitations, (ii) the omission of "that" even when required grammatically, (iii) the replacement of "that" by grammatical alternatives; (iv) the addition of relative pronouns to sentences where grammatically possible; (v) the tendency to find clauses qualifying the 'Object' more difficult to imitate than those qualifying the 'Subject'; (vi) the fact that neither group showed a markedly different pattern of imitation with sentences differing in case grammar; (vii) the 'recoding' of syntactic structure was found in both groups; (viii), the low proportion of correct answers to the comprehension questions in both groups. On the other hand the following differences occurred: (i) the C group used much more substitution, particularly of "what"; (ii) only the C group shifted the Subject of the relative clause to the main clause; (iii) only the C group repeated the last few words of the sentence.

The trend of these results does not confirm the teacher's belief that foreign adults approach language in ways fundamentally different from native children. So far as the imitations were concerned, adults made much the same kind of alteration as the children. Indeed it was surprising to find that many of the mistakes one had long accepted as typically foreign were also made by the children; one such example is the omission of Subject relative pronouns; another is the omission of "s" from the third person singular form of the verb. With comprehension questions also, the foreign adults met nearly as much difficulty as the children, in spite of the fact they already knew one language.

It must now be considered how serious the differences are between the groups. The substitution of "what" by the children probably reflects the va­riety of English spoken around them, which allows "what" as a relative pro­noun; the variety of English the foreigners were taught in the classroom does not permit "what" as a relative. Hence it would be surprising if the children and the adults did not differ in their use of "what". The recoding of sentences by shifting the Subject from the relative clause to the main clause, done by the children but not the adults, cannot be explained in the same way. Possibly it may be the by-product of the audio-visual teaching method.

The children's imitation of the last few words seems a more important difference. It is possible to explain this in terms of mental differences that are not strictly speaking language differences. The capacity of the child's primary memory increases as he grows older; the manner in which he stores information also changes; though not much is known about this area as yet, some research has suggested that there is a correlation between short term memory capacity and the ability to comprehend different types of syntax. It may be then that, when the child is faced with a sentence that exceeds his memory capacity he adopts the strategy of repeating only the last few words he has heard; the adult would only have to fall back on this strategy with sentences much greater complexity. Thus this difference between children and adults may simply show that the child can remember less than the adult. Further research is being carried out on the relationship between the ability to comprehend relative clauses and short term memory capacity. So far as the language, teacher is concerned, this line of research casts an interesting sidelight on the teaching technique of repetition. The teaching of both pronunciation and model dialogues has relied on the students repeating a model as accurately as possible. It has been clear for some time that this form of imitation is inadequate as an explanation of how children learn language. When children repeat, they do not make parrot-like imitations; instead they process what they hear in terms of their own competence, altering both pronunciation and syntax. It now seems that foreign adults are also far from passive and that they too adapt what they hear to fit their own competence. Hence the second language teaching technique of repetition may prove more effective if it takes advantage of this rather than suppresses it, possibly 'permitting the student to be much freer in his repetitions than is at present the custom.

2. A comparison of the comprehension of "The duck is happy to bite" and The duck is hard to bite"

The first comparison described above had the disadvantage that it compared something about which little is known — the comprehension of relative clauses in children — with something that is almost totally unknown — the comprehension of relative clauses in foreign adults. Hence the two groups could not be broken up into developmental stages. The comparison to be described now took an area of syntactic development in the child where the stages are known and tried to see if foreign adults passed through the same stages, essentially it consists of replicating an experiment performed with English children with a group of foreign adults.

The experiment that was used was devised by R. F. Cromer. He tested children's comprehension of sentences such as "The duck is happy to bite" and The duck is hard to bite". While these sentences appear to have the same surface structure, the native speaker of English knows that in the First case the duck was biting something, in the second the duck was being bitten, terms of the 'Aspects' model of transformational grammar, the sentences have the same surface structure but different deep structures. The deep structure of "The duck is happy to bite" can be crudely shown as "The duck is happy (the duck bites something)"; "the duck" is the Subject not only of the sur­face structure but also of the deep structure sentence enclosed in brackets. In the same way the deep structure of "the duck is hard to bite" could be shown as "the duck is hard (something bites the duck)"; "the duck" is the Subject of the surface structure but the Object of the deep structure sentence. Cromer set out to test when children became aware of the deep structure difference between the sentences; his method was to ask the child to act out sentences with two puppets of a wolf and a duck. The sentences he used are given in Table 5, and were used also in the present comparison.

Table 5.
Test Sentences for Comparison 2.

The wolf bites the duck.
The duck bites the wolf.
The wolf is happy to bite.
The duck is anxious to bite.
The wolf is tasty to bite.
The duck is easy to bite.
The wolf is willing to bite.
The wolf is hard to bite.
The duck is glad to bite.
10.The duck is fun to bite.
The wolf is bad to bite.
The duck is horrible to bite.
The wolf is nice to bite.
The duck is nasty to bite.
The wolf is bitten by the duck.
The duck is bitten by the wolf.
The wolf is risp to bite.
The wolf is larsp to bite.

Some of the sentences followed the pattern of "The duck is happy to bite" (3, 4, 7, 9) in which the surface structure Subject is also the Subject of the embedded deep structure sentence. Some followed the pattern of "The duck is hard to bite", (5, 6, 8, 10) in which the surface structure Subject is the Object of the embedded deep structure sentence. Some could be interpreted in either way (11—14); two were Passive (15, 16). The last two sentences contained nonsense adjectives. Before the child heard No. 1 7, he was shown a picture of a dog carrying a bone and told "Someone gave this dog a bone and so he's feeling very risp. le's feeling very risp." Before No. 16, the child was shown a picture of a cat with a rose in its mouth and told "This cat climbed up and picked a rose and he found that eating the rose was larsp. Eating the rose was larsp". Thus the context was intended to bias 17 towards "the wolf" being Subject of the embedded sentence and 18 in the opposite direction. The results showed that the children could be divided into three groups corresponding to stages of syntactic development:

(a) Primitive Rule Users. These interpreted the surface structure Subject as being identical with the Subject of the embedded sentence, regardless of the deep structure of the sentence (henceforth the 'surface Subject' interpretation).

(b) Intermediates. These gave mixed answers, sometimes interpreting the sur­face structure Subject as the Object of the embedded structure (the deep Ob­ject interpretation) but not consistently or correctly.

(c) Passers. These distinguished the two deep structures correctly, did not al­ways interpret the ambiguous sentences in terms of surface structure, and interpreted the nonsense word sentences according to the preceding context. This division into three groups corresponded also to a difference in Mental Age. The Primitive Rule Users were all less than 5:7; the Intermediates less than 6:6; the Passers more than 6:8.

The present comparison essentially repeated this experiment with foreign adults. These numbered 67 and were all attending day classes in English for Foreigners at Ealing Technical College. The average period they had spent learning English in England was 8 months; the average period they had learned it in their own countries was 3 years 8 months. Their native languages were mixed: Spanish (15), German (13), Persian (9), French (7), Italian (6), Polish (7), Arabic (2), Indonesian, Danish, Serbo-Croat, Chinese, Greek, Portuguese, Hebrew, Armenian (1 of each). The experiment was adapted to make it usable with adults in groups rather than with individual children. Each foreigner was given a form that asked for his name, mother tongue and time spent learning English both in England and in his own country. Below this was a number for each sentence followed by the words "duck" and "wolf". In a short demonstration on the blackboard using the sentences "The dog bites the cat", "The :at bites the dog", and "The dog the cat is biting is big", the adults were shown low to ring round one word for each sentence "to show who is doing the biting". The sentences were then read out and the pictures and preceding context given before 17 and 18.

The results showed that the foreign adults could be divided into three groups in terms of their answers to sentences 3—10.

(i)   Primitive Rule Users. 7 adults interpreted sentences 5, 6, 8 and 10 as having the surface structure Subject the Subject of the deep structure.

(ii) Intermediates. 45 gave mixed answers, choosing erratically between both interpretations.

(iii). 14 were completely correct on sentences 3—10. (One set of answers had to be discarded because the task had been misunderstood).

The average time the Primitive Rule Users had been learning English was 2 months in England and 2 years 2 months in their own countries; the Intermediates 7 months in England, 3 years 5 months at home; the Passers 1 year in England, 4 years 8 months at home. Thus success in understanding the sentences was linked to the amount of time that had been spent learning English, as for the children it was linked to mental age. A principal components analysis of the data confirmed this grouping. The major component accounted for 40.7 per cent of the variance; when ordered in terms of this component the Passers clustered at one end of the scale and the Primitive Rule Users at the other. A second factor that emerged was a clear separation of sentence 1 8 from the rest (Sen­tence 17 was not scored for the computer as it did not distinguish between groups); this accounted for 14.6 per cent of the variance. Sentence 18 contained the nonsense word "larsp"; this result will be discussed below.

Table 6 shows the interpretation that the groups gave to the ambiguous sentences 11—14.

Table 6. Comparison of answers to ambiguous sentences (11— 14).

Surface Subject=Deep Subject Surface Subject=Deep Object
Primitive Rule Users 17 8
Intermediates 37 136
Passers 2 54

This shows that the Primitive Rule Users usually interpreted the sentences in terms of surface Subject, i. e. that the surface Subject was also the deep subject; the Intermediates preferred the deep Object interpretation in which the surface Subject was the deep Object; the Passers almost invariably chose the deep Object interpretation. These results are statistically significant (p = < .01). They are very similar to those for the children; child Primitive Rule Users also preferred surface Subject interpretations; child Passers deep Object interpretations. The chief difference seemed to be one of consistency: while child Primitive Rule Users interpreted all four sentences as surface Subject, adults were inconsistent, only three of the seven giving surface Subject interpretations to all four.

The passive sentences 15 and 16 did not cause much difficulty; 6 were wrong on 15 and 5 or. 16.; Cromer also found only three children that did not understand passives. Table 7 gives the results for sentences 17 and 18, the ones containing nonsense words.

Table 7 Answers to sentences 17 and 18

S= Surface Subject Interpretation (17)

D=Deep Subject Interpretation (18)

Primitive Rule Users 4 2 2 4
Intermediates 22 22 10 34
Passers 10 3 2 11

These show a slight tendency for all groups to give surface subject interpretations to 17 and deep Object interpretations to 1 8, particularly the Passers. These results do not, however, reach the required levels or statistical significance. They nevertheless contrast with those for the children: nearly all the child Primitive Rule Users preferred surface Subject interpretations for both sentences; nearly all the child Passers interpreted 17 in terms of surface Subject and 1S in terms of deep Object.

There seem then to be similarities in the ways that native children and foreign adults perceive this structure. Both start with the strategy that the surface structure Subject is the Subject of the deep structure; both go on to a period in which they interpret deep and surface structure on a hit-or-miss basis; both finally enter a period when they are fully aware of deep and surface structure. There are dangers in generalising from these results and those in the previous comparison. Both comparisons have dealt essentially with how people understand sentences, not how they learn them, — with decoding rather than code-breaking. Though it is possible to argue that the perceptual strategies of the adult are the same as the learning strategies of the child, this has to be proved rather than assumed in advance. Even if both native children and foreign adults showed a point for point identity in the way in which they understood sentences, this would not necessarily prove that they learned how to perceive in the same fashion. Nevertheless in spite of this warning it at least seems likely that different learning strategies would be revealed by different perceptual strategies at different stages of development; the absence of such differ is at least some negative evidence for underlying similarity of the processes. This distinction between perceptual processes and learning process seems particularly crucial to the nonsense word sentences in the second comparison. There we saw that a principal component analysis showed that the foreign adults treated sentence 18 differently from the others; it emerged from the answers to 17 and 18 that the adult groups, who behaved in every other way like the children, did not treat the nonsense words in the same way. This difference is important since it could be argued that 17 and 18 were the only parts of the test where something approaching 'learning' took place; the subjects had in fact to 'learn' from the preceding context the classes of adjec­tive to which "risp" and "larsp" belonged. So while the present paper seems to show similarities between the ways that native children and foreign adults understand sentences at different stages of development, it must be interpreted with caution as evidence for similarities of learning. Further research is needed to establish whether the learning strategies as well as the perceptual strategies are similar.

Nevertheless this line of inquiry starts to raise some problems for the critical period theory of language acquisition. This states that there is a definite moment at which the child becomes capable of language acquisition and a definite moment at which he loses this capacity. This theory has not dealt very adequately with the fact that adults are capable of learning a second language and that, in some respects, they do so better than children. E. Lenneberg claims, for instance, that "This does not trouble our basic hypothesis on age limitations because we may assume that the cerebral organization for lan­guage learning as such has taken place during childhood and since natural languages tend to resemble one another in many fundamental aspects . .. the matrix for language skills is present." We would then suppose that in childhood the adult had learned the principle that sentences can contain other sentences embedded within them, regardless of the language. When he learned a new language, he would be on the look-out for this and his learning of embedded sentences would be facilitated. But this does not happen with relative clauses; the adult falls into the same trap as the child on sentences such as "The lady the boy's drawing is funny" in that he does not realise there is embedding. His learning has not been greatly facilitated and he appears to go through the same stages as the child.

If the adult does not learn by facilitation, how can he learn a second language if the critical period theory is correct? One possibility would be to accept Chomsky's division of the mind into faculties. Normal language acquisition takes place through the faculty of language learning which atrophies at a certain age; languages can still be learned after that age by making use of other faculties of the mind such as the mathematical or logical. Thus first language acquisition and second language learning in young children exploit the language faculty, second language learning in adults exploits some other faculty. If this were the case, one would expect to find some clear differences between children and adults in the way that they tackle language. Recent research tends to suggest the opposite. D. Palermo and H. Howe demonstrated that adults approached an experimental learning situation in the same way that children learn the past tense inflections in English; W. Stolz and J. Tiffany showed that the characteristic differences between the word associations of children and adults could be cancelled out by giving the adults unfamiliar words; the present research shows similarities between syntactic comprehension in foreign adults and native children. At the moment there seems to be no certain evidence to show that adults are different from children in language learning, once the other attributes of the adult such as increased memory span have been cancelled out.

V. J. Cook

Dept. of Arts and Modern Languages
North East London Polytechnic
Longbridge Road



1 V. J. Cook, 'The Analogy between First and Second Language Learning,' IRAL, VII/3 (1969).

F. M. Holley and J. K. King, 'Imitation and Correction in Foreign Language Learning,' Modern Language Journal, LV, No. 8 (1971).

J. F. Torrey, 'Second Language Learning,' in C. Reed (ed), The Learning of Language (1971).

2 D. P. Dato, American Children's Acquisition of Spanish Syntax in the Madrid Environment: Preliminary Edition, U.S. Dept. Health, Education, and Welfare Final Report Project 3036 (1970).

3. H. Sinclair-de-Zwart, Acquisition du langage et developpement de la pensee (Paris, 1967).

4. I would like to thank Mrs. Ziranek, the headmistress of Grove House Nursery School, Southall, and B. Abbs, Senior Lecturer in charge of English to foreigner at Ealing Technical College, for letting me carry out these tests I am also extremely grateful to R. Baldy, of Brunel University, for carrying out the statistical calculations

5. D. I. Slobin and C. A. Welsh, 'Elicited Imitation as a Research Tool in Developmental Psycholinguistics,' in C S. Lavatelli (ed.), Language Training in Early Childhood Education (ERIC, 1971).

6. R. C. Troike, 'Receptive competence, productive competence, and performance,' in J. Alatis (ed.), Twentieth Annual Round Table Meeting (Georgetown, 1970).

7. C. J. Fillmore, ‘The Case for Case’, in E, Bach and R. Harms (ed.), Universals in Linguistic Theory (1968)

8 E. Clark, 'How young children describe events in time,' in G. H. Mores d'Arcais and W. J. M. Levelt (eds.), Advances in Psycholinguistics (1970)

9. D. J. Murray and B. Roberts, 'Visual and Auditory Presentation, Presentation Rate, and Short-term Memory in Children,' British Journal of Psychology, 59; No. 2 (1968).

10. J. Mehler, 'Studies in Language and Thought Development,' in R. Huxley and E. Ingram (ed.), Language Acquisition: Models and Methods (1971).

11 N. C. Graham, 'STM and Syntactic Structure in Educationally Subnormal Children,' Language and Speech, Part 4 (1968).

12 V. J. Cook, 'Relative clauses, and STM in children and adults' (in preparation).

13 R. F. Cromer, '"Children are nice to understand": surface structure clues for the recovery of a deep structure," British Journal of Psychology 61, No. 3 (1970).

14 N. Chomsky, Aspects of the Theory of Syntax (1965).

15 Cook 'Speech perception and language teaching' in preparation

16 T. Bever, 'The cognitive basis for linguistic structures', in J.R. Hayes (ed) Cognition and the Development of Language, 1970

17 E. L. Thorndike, Adult Learning 1928

18 E. Lenneberg, Biological Foundations of Language, 1967, p. 176

19. Chomsky, 'Noam Chomsky and Stuart Hampshire discuss the study of language,' The Listener, Vol. 79, No. 2044 (1968).

20. S. Palermo and H. E. Howe, Jr., 'An Experimental Analogy to the Learning of Past Tense Inflection Rules," Journal of Verbal Learning and Verbal Behavior, 9 (1970).

21 W. Stolz and J. Tiffany, 'The Production of "Child-like" Word Associations by Adults to Unfamiliar Adjectives,' Journal of Verbal Learning and Verbal Behavior, 11 (1972).