Vivian Cook   Writing System Topics   Online Writings

The consequences of bilingualism for cognitive processing

Vivian Cook

 In A. de Groot & J.F. Kroll (eds.), 1997 Tutorials in Bilingualism: Psycholinguistic Perspectives, Lawrence Erlbaum, 279-300

Perennial questions about bilingualism concern the relationship between the two language systems in the same mind. Do they form two separate systems for processing language or a single combined system? Does one language aid or interfere with the other? Are there differences in other areas of the mind than language between a person who knows one language and someone who knows more than one? The more complex system involving two languages may confer benefits or losses on areas of the mind other than language: bilinguals and monolinguals may think in different ways.

The term ‘cognitive’ has led to some confusion in this type of discussion. Linguists seem to assert both that linguistics is a branch of cognitive psychology, as it is concerned with the human mind, and that the language faculty is entirely distinct from all other faculties of the mind, i.e. different from ‘cognition’. Linguists furthermore often distinguish between the abstract knowledge of language called ‘competence’ and the ‘cognitive’ processes involved in the actual production and comprehension of speech, called ‘performance’. Psychologists, however, often explore the relationship between ‘cognition’ (i.e. the ‘rest’ of the mind) and ‘language’, for example Cromer (1974). This chapter employs the term ‘cognitive’ broadly within this range of possible definitions. It should also be noted that the term second language (L2) is here used to cover a wide range of knowledge of languages other than the native language; the term L2 user is also preferred as being a more encompassing category than L2 learner, including people who have finished learning as well as those still in the process.

This chapter restricts itself to the relationship of L2 using to cognitive processing rather than being concerned with language knowledge. Interesting as models of language competence such as parameter-setting may be, they treat language as a separate faculty of the mind and language competence as distinct from language performance, and so cannot in principle entertain the types of relationship discussed here even if some speculative links could be made. Furthermore the recent development of the Minimalist Program with its restriction of parameters to the lexicon has put much of the L2 research based on principles and parameters in jeopardy since it is closely tied to the now outdated principles and parameters model of the 1980s (Cook & Newson, 1996).

The way in which the relationship between bilingualism and cognitive processing is approach­ed depends upon the ideology of the questioner. The standard approach is the ‘monolingualist’ view that the norm for human beings is to know a single language; any deviation from the monolingual standard is going to have a cost. To the monolingualist the person who uses two languages is odd; it is natural to have a single nose and a single mouth, unnatural to have two of each or to be seven feet tall. A person who has two lang­uages is then strange in some sense, obviously different from the ‘normal’ person. Hence the questioner looks for the differences caused by this unnatural condition of knowing two or more languages, whether with using the first or the second language separately or with mental processes.

On the one hand this leads to the concept of L2 ‘cognitive deficit’ (Long & Harding-Esch, 1977): the mental activities that one carries out through a second language will be fractionally less efficient than in the first language. Section 1 will examine this ‘subtractive’ view of bilingualism. On the other hand some monolingualists see knowing a second language as a deviation from the monolingual norm in a positive direction. The L1 processing of L2 users may become richer, their mental processes more effective or their view of the world more balanced, than those who know only one language. This view is perhaps prevalent in education. The question is how the processes of the two languages relate, whether there are in fact definite positive effects on mental processing. Section 2 will report on some of the evidence for this ‘additive’ view of bilingualism.

Opposed to this general monolingualist approach in principle is the ‘multi­lingualist’ view that it is normal for human beings to know more than one language. It is the monolingual that is deficient, having been deprived of the natural human heritage. The perspective can be reversed; what matters is the cognitive deficiency suffered by monolinguals. The quest­ions are how much monolinguals lose overall from knowing only one language and what lacks they have in other mental processes. This ‘multilingualist’ alternative looks at the same evidence from another viewpoint; we will consider at the at the end whether this difference could ever be settled through research or is indeed an ideological, political, or moral commitment.

Section 1.  Subtractive views of Bilingualism: cognitive deficit

This section looks in general at the ‘subtractive’ view that learning a second language means subtracting something from the monolingual state, looking at deficiencies in the processing of the first and second languages as well as in memory systems.

Deficiency in processing the L2

It is perhaps blindingly obvious that L2 users are less efficient in the second language than native mono­linguals. Many of them are after all still learners and hence not fairly comparable with native speakers. Teachers often make the uncheckable assertion that 80% of the L2 learners of English in the world at any given time are beginners. It is hardly surprising that L2 users are less efficient at the L2 than are native speakers in terms of accuracy and of speed. A truer comparison might be with L1 children after a similar number of hours of using the language.

A typical effect with syntax can be seen in an experiment with the binding relationship between pronouns and their antecedents (Cook, 1990). Native speakers scored 93% correct in saying that him referred to ‘John’ in sentences such as John said Peter helps him, advanced Japanese users of English 92%; Yet timing measurements show native speakers took 4.6 seconds to decide, Japanese users 7.6 seconds; on John said Peter helps himself native speakers scored 91% correct in choosing himself as referring to ‘Peter’ versus Japanese 76%, timings 4.2 seconds vs 7.1 respectively. The comprehension times of foreign learn­ers of English were longer than those of native speakers even though the L2 speakers clearly knew the binding relationship. Lehtonen and Sajavaara (1988) summarize a series of acceptability judgement experiments with Finns learning English as an L2; visually presented sentences took an average of 5.1 seconds to judge, compared to native speakers 3.1, auditorily presented sentences 2.1 compared to native speakers 1.3. The short-fall on this acceptability task is similar to that on the comprehension task.

One issue is whether the few L2 learners that become balanced bilinguals, or ‘ambilinguals’ as Halliday, McIntosh, and Strevens (1964) call them, are actually processing the L2 in the same way as monolingual native speakers - now sometimes called the problem of ‘ultimate attainment’ in L2 learning (Birdsong, 1992): does the L2 learner ever become the same as the native speaker? The usual research cited is Coppetiers (1987) which demonstrated that successful bilingual American residents in France still differed in grammatic­ality judgements from native French speakers; while the native speakers varied from the norm between 5% and 16%, the non-native speakers varied between 23% and 49%; none of the 21 L2 users fell within the extreme bounds of the native group. In the earlier literature this difference was usually covered by the notion of the bilingual’s ‘dominant’ language, in which, inter alia, there were more word associations to a given stimulus (Lambert, 1955), and reaction times for words were faster (Lambert, 1956). The rarity of ambilinguals leads to the vast literature of individual variation in L2 learning, for example Skehan (1989); if people cannot learn an L2 as efficiently as an L1 then much may be learnt by studying their lack of success compared to the almost uniform efficiency of the L1 child.

Thus the overall point is that people process a second language less swiftly than their first. It is not that the L2 users are ignorant of the syntax or the vocabulary of the language, or even that they necessarily make more mistakes; it is just that they use it more slowly.

Deficiency in processing the L1

However, for the monolingualist, the consequences of knowing an L2 can be even more dire: L2 users may lose efficiency in their first language as well as speed. The presence of an L2 in the mind in some way detracts from the L1. This effect was established in a series of experiments carried out by Edith Magiste. One experiment (Magiste, 1979) studied German children aged 13-18 learning Swedish in Sweden, using a variety of tasks, such as timing how long the children took to name objects in the L1 and the L2; after about 5.5 years in Sweden the children were as fast in Swedish as in German and they gradually became faster still in Swedish. Yet, after 17 years, they were still 0.2 seconds slower than Swedish monolinguals on the naming task, supporting the idea of deficit. However their speed in German stayed about the same rather than improving, as might have been expected. So the children who had been in Sweden one year were faster in German than children who had been there ten years - their native performance was, if anything, declining. On decoding tasks such as “Mark the third letter from the left”, monolingual subjects of German and Swedish were also consistently faster than bilinguals, who were in turn faster than trilinguals, namely a reaction time of 0.8 seconds per item for monolinguals, 1.0 seconds for bilinguals and 1.3 for trilinguals. A second experiment (Magiste, 1986) used the same tasks with a younger age group aged 6-11 as well as 13-18s; results showed the same pattern of switch-over from L1 to L2, with slight loss in the L1 over time and with the L2 not reaching native speeds. A third experiment on the same lines (Magiste, 1992) found that younger children (average age 8) switched over to the second language at an earlier point than the older children (average age 14); younger children’s reaction times in the L1 German naming tasks started to decline after one year, older children’s after 2 years.

In general then these experiments support Magiste’s conclusion that ‘The very fact of having available more than one response to the same stimulus may lead to slower reaction times unless the two response systems are hermetically isolated from each other’ (Magiste, 1986, p.118). This was supported by Ransdell and Fischler (1987) who also found that adults who had been bilingual all their lives were slower at responding on list recognition and lexical decision tasks even if their accuracy was the same.

Short Term Memory (STM)

Short Term Memory (STM) has been used mostly as a general term for short-term storage of information, in the tradition of the classic article by Miller (1956) that showed that human information processing seemed to be limited in a capacity of 5-9 items, called the magic number 7 plus or minus 2. Early L2 research aimed to show that STM span was comparatively limited in a second language. Experiments asked learners to repeat or write down strings of words or numbers and then calculated the individual’s maximum span. Glicksberg (1963) measured an increase in L2 digit span in adult students from 6.4 digits at the beginning of a course in English to 6.7 at the end eight weeks later; there was a final deficit of 0.8 digits compared to their L1 scores. Cook (1977) contrasted the average span of 5.9 digits for adult beginners in English with the 6.7 of advanced learners. Cook (1979) linked span to age and found that, though digit span improved in both languages by 0.7 digits for English secondary school children learning French between 12 and 14, they had a shortfall of 2.8 digits in French at both ages. It seemed then that this type of task also revealed a cognitive deficit in STM in the L2.

However the source of this deficit was not clear. Why should the language employed make a difference to the span? Some early psycholinguistic experiments had attempt­ed to establish a correlation between memory processing and syntax; for example, Savin and Perchonock (1965). More recently linguists working with parsing postulated buffer memory systems capable of holding syntactic constituents for later processing; for example Marcus (1980).

The growth-point for STM research in psychology has, however, been the phonological system. Since the work of the 1960s with confusable letters (Conrad, 1964), it had been claimed that STM relies on sounds, not syntax; all types of language information coming in to the memory system have to be encoded in phono­logical terms. Cook (1977) compared L2 word spans for adult beginners and advanced students of English, taking two types of word: similar-sounding words such as cat, mat, bat and different-sounding words such as bus, spoon, train. Beginners had a span of 3.5 similar-sounding words compared with 4.5 different-sounding words; advanced learners had a span of 3.7 similar-sounding words and 5.0 different-sounding words. While the advanced learners were better than the beginners, both had a shortfall for similar-sounding words, showing that they were using phonological coding. The phonological basis of STM could be established on the same grounds in L2 processing as in L1 processing. This is in itself interesting in that some aspects of memory are developmental: L1 children up to the age of five remember objects through shapes and colours rather than sounds (Conrad, 1972) but adult L2 learners use labelling as a memory aid in Short Term Memory like native adults rather than returning to the state of young native children (Cook, 1981). In STM at any rate L2 adults are not recreating the system from scratch but employing an ‘adult’ phonological-based principle from the start.

Working Memory

During the 1980s the links between STM and language were developed through the working memory model of Alan Baddeley (Gathercole and Baddeley, 1993). Working memory is in a sense a complementary term to Short Term Memory, stressing the aspects of STM that are involved ‘in the temporary processing and storage of information’ (Gathercole and Baddeley, 1993, p.2). In the Baddeley model, working memory consists of a ‘central executive’, which controls how information is passed around the system, and of visual and phonological ‘slave systems’ which temporarily process and retain the information appropriate to their two modes. The phonological system has a phonological store, which can hold information for about two seconds, and a phonological loop that recycles information back through the store to extend its life, that is to say by repeating information over and over whether audibly or inaudibly.

Baddeley’s Working Memory Model (simplified)

‘Working memory’ capacity in this theory is linked to the articulation of sounds in a language and to the amount that can be said during the two seconds window. Hence more short words such as wit can be stored in working memory than long words such as university since short words take less time to articulate than long words (Baddeley, Thomson, and Buchanan, 1975); people who can articulate more sounds in two seconds will have larger spans. Digits with fewer syllables or shorter vowels can be said more quickly than those that are longer. Consequently speakers of languages with short words for digits such as English will have a longer digit span than those with longer words such as Arabic (Naveh-Benjamin and Ayres, 1986); speakers of Chinese which has ‘short’ digits have better digit spans than speakers of English (Stigler, Lee & Stevenson, 1986). There is also some interaction with the written mode of language; speakers of Hebrew, whose writing system is based on consonants, have a superior span for sequences of consonants of 7.3 letters compared to the 5.4 letters of English, whose writing system uses both vowels and consonants (Kinsbourne & Cohen, 1971). Though it is hard to compare the results of experiments carried out with diff­erent techniques, Chinese speakers seem to hold the digit-span record with 9.9 digits (Hoosain, 1979). One side effect is that people who speak faster will have higher scores on IQ tests, in so far as some IQ tests involve digit repetition, and hence reward people able to say more digits in two seconds than those who speak more slowly; Welsh children for example have smaller digit spans than US children because of the greater length of Welsh digits—average duration 385 milliseconds per digit compared to English 321 milliseconds (Ellis, 1992), contributing to poorer performance on the Weschler intelligence test (Ellis & Henneley, 1980) .

The source of the L2 deficit might be in the central executive, which decides how to handle information, or in the phonological store, which stores information provided to it through the articulatory loop, or in the articulatory loop itself, which recycles information contin­uous­ly in phonological form. Most L2 research to date deals with the relationship of the phonological system to learning rather than to processing, though the two strands of ‘code­breaking’ and ‘decoding’ are well-nigh impossible to disentangle (Cook, 1991). Service (1992) found that accurate repetition of English pseudo-words by Finnish children learning English was a good predictor of success at English two years later, showing the relevance of phonological encoding to L2 learning; she claimed that the actual encoding of language into sounds is a function of the central executive rather than the articulatory loop. Service and Craik (1993) extended this research to deal with age: old people (average age 72) relied more on phonology in learning L2 vocabulary than young people (average age 25.2). Papagno, Valentine & Baddeley (1990) used the task of articulatory suppression (muttering sounds while carrying out a memory task, thus occupying the articulatory loop) compared with tapping (a non-verbal task that does not tie up the articulatory loop); subjects had to acquire lists of paired words either in the L1 or in the L1 and the L2. After five trials with tapping, subjects succeeded in learning 7.5 pairs of words within the same language, their L1 Italian and 6.1 pairs that linked L1 Italian with L2 Russian; with suppression, they learnt 7.0 pairs in L1Italian, but only 4.2 pairs L1/L2 Italian/ Russian pairs. Thus, suppression disrupted the learning of L1/L2 pairs far more than same-language L1 pairs. The authors argue this demon­strates that the articulatory loop is involved in the acquisition of new words in an L2.

However the same effect was not found in a second experiment involving the learning of Russian L1/L2 words by English learners. So the articulatory loop is not the only factor. They argue that semantics also plays a role; this affects going from Russian to English more than going from Italian to Russian as shown by the greater number of word association by English subjects to Russian words than by Italians. Papagno and Vallar (1992) similarly found that the learning of novel words was affected for the worse by the length of the word and by phonological similarity, again implicating the articulatory loop. Ellis and Beaton (1993) looked at the keyword method of learning, which requires the learner to make an imaginative link between the foreign item to be learnt and the L1 translation, for example learning French hérisson (hedgehog) via the key word hairy son. They asked subjects to learn German vocabulary using this technique or a repetition task of saying each word aloud; performance turned out to be optimal if both techniques were involved, i.e. if the semantic keyword technique were combined with an articulatory loop task. Brown and Hulme (1992) also used the articulatory suppression technique with English learners of French: they found that it indeed cut the digit span from 4.8 to 2.9 digits in French and from 6.1 to 3.6 digits in English. But, as they point out, while this certainly shows the involve­ment of the articulatory loop in L2 processing, it does not explain why the L2 span is smaller in both L2 conditions than in the L1; they argue for a deficiency in the long term memory phonological represent­ations of L2 words. However, in an experiment varying word length with Welsh/ English bilinguals, Ellis (1992) found articulatory suppression brought the difference between 6.5 digits in English and 5.8 digits in Welsh down to non-significance. The L2 research on Working Memory is thus mostly concerned with how the different elements contribute to L2 learning rather than with deficiencies in the L2 processing itself.

Other memory and cognitive processes

There has not so far been the same amount of interest in extending L2 memory work into Long Term Memory, that is to say memory processes involv­ing storage for longer than about 30 seconds. One exception was research by Long and Harding-Esch (1977) who found that Cambridge under­graduates studying French not only remembered less information from talks given in their L2 French than in their L1 English but also had more false memories of things which the speaker had not actually said. Hummel (1986) hypo­thesised that the extra cognitive processing required in a second language would increase the amount that people could remember. Her design tested Canadian bilingual French/English speakers by varying the order of two passages in L1 and L2. Subjects remembered more inform­ation from the first passage when there was a switch of language between the two passages, whether the second was an entirely different text or a translation of the first. While this would appear to contradict the Long and Harding-Esch (1977) results, Hummel’s repeated passage design is not strictly comparable with the single text design.

Some work has also looked at L2 users in the light of the well-known schema theory experiments of Bransford and Johnson (1982). Adams (1983) tested American students of French with the same texts as Bransford and Johnson and found that knowing what the passage was about helped them equally in both languages. Hence activation of the approp­riate schema is relevant for processing in both languages. Carrell (1984) tried to see not only whether the presence or absence of context made a difference to how much L2 learners could understand, but also how important it was whether the text had precise words like clothes and washing machine, or vague words like things and facilities. Advanced learners and natives found lack of context affected their comprehension; intermediate L2 learners found the use of vague words was also a hindrance. Interestingly, while native speakers had a fair idea of how difficult the passages were for them to understand, non-natives did not. Later research by Roller and Matombo (1992) did not get the same results: speakers of Shona actually remembered more of the Bransford and Johnson texts in English than in their first language.

To conclude this section, some other cognitive tasks should be briefly mentioned in which people are more restricted in their second language than in their first. L2 users for example find it harder to count flashing lights silently in the L2 (Dornic, 1969). Mental arithmetic is also harder to do in the L2 (Marsh & Maki, 1976; Ellis, 1992); this may have more to do with the language in which one learnt to do arithmetic than whether it is the first or second language; L2 users of English educated through English as the medium of instruction may prefer English for such calcul­ations rather than their L1. Ellis (1992), however, claims that bilingual Welsh/English children prefer to calculate in English, since the shorter digit names are more helpful; in Welsh the time taken for multiplication sums was 37.2 seconds, in English 28.0 seconds; mistakes in Welsh averaged 1.6 out of 6 sums, while for English they are 0.96 out of 6. Peynircioglu and Tekcan (1993) investig­ated the ability of Turkish L2 users of English to recogn­ise words in 4X4 letter squares based on the game of Boggle. When performing only in the L1 Turkish, they could find 21.4 words per square, when performing in the L2 English they averaged 12.8, when performing in both languages at the same time they found 15.6 Turkish words and 8.2 English words. Their score of 21.4 in the L1 was virtually the same as the 20.4 of monolinguals’ and they were just as quick at identifying words in the display - 5.5 seconds compared to the monolinguals’ 5.7 seconds. Hence Peynircioglu and Tekcan (1993) claim there is no deficiency for bilinguals on ‘data-driven’ tasks that use recognition and lexical decision compared to ‘conceptually-driven’ tasks like free recall and object-naming; this challenges the finding of Ransdell and Fischler (1987) mentioned above that bilinguals were slower at the data-driven tasks of list recog­nition and lexical decision. It raises the perennial question of whether bilinguals have one memory or two, which strays outside our brief here; see Cook (1992) for a summary. Durgu­noglu and Roediger (1987) argue that the nature of the task, whether conceptually or data-driven, dictates whether information is stored in a language-specific mode; Spanish/ English biling­uals were better at completing fragments of English words like _l_i_a_ _r (alligator), when primed with English words than with Spanish words; it is the fact of having seen the word that counts, not having processed its meaning, a data-driven task, not a concept-driven one.

While it strays into the area of language knowledge, finally we should mention the vexed problem of the links between language and thinking often called the Sapir-Whorf hypoth­esis. Suppose that speakers of different languages think and perceive the world in different ways rather than simply talk about it differently. For example Bahasa Melayu, one of the languages of Malaysia, reflects Malaysian cooking in recognising several degrees of saltiness such as masin kitchup ‘salty like soy sauce’, masin ayer laut ‘salty like sea water’, masin garam ‘salty like salt’, and masin maung ‘horribly salty’; O’Mahoney and Muhiud­een (1977) tested the ability of speakers of Bahasa Melayu and English to distinguish solutions with different amounts of salt. Malaysians were indeed able to make much finer distinctions that English speakers. It is not just that their language is richer in vocabulary, it is that their perception of taste is different. In a second language then it would be interesting to see whether the speakers changed both as they acquire a new language and as they switch from one language to another. The only relevant L2 research on this is Caskey-Sirmons and Hickson (1977), who found that monolingual speakers of Korean use the word paran sekj (blue) to mean something greener and less purple than bilingual Koreans who know English; in other words something of the second language system is seeping through into their perception. Virtually all the research in this area has however been confined to the putative differences between monolingual speakers of different languages, surveyed for example in Hunt and Agnoli (1991) and Cromer (1991) rather than to effects on the L2 on L2 users.

This section has hinted that there may be an interaction between L2 learning and age. The cognitive processes that we have been talking are themselves subject to development; child­ren acquire both the adult working memory capacity and adult memory processes such as labelling over years. The ability to process the second language is to some extent depend­ent on this, as is the ability to process the first, as the research by Brown and Hulme (1992) suggests. On the other hand a critical period theory of language acquisition has sometimes been proposed (see chapters by Harley and Ellis in the volume) that argues that language acquisition is less efficient after the early teens. The major evidence for this has been from judgments of foreign accent (Oyama, 1976) and from grammaticality judgements (Johnson and Newport, 1989) rather than from the research into cognitive processing dealt with here, though there are hints of effects of age in the findings of Service and Craik (1993) that older people rely more on phonology and Magiste (1992) that younger children switch to the L2 sooner.

To sum up this section, there is evidence that L2 users are less effective; in speed of processing the L2 and the L1, in working memory processing in the L2, and in certain types of cognitive task in the L2. Most of these deficits are slight and have to be balanced by the gain that they are able to use two languages compared to the monolingual’s one.

Section 2. additive view of bilingualism

The alternative view is that knowing a second language extends rather than diminishes the individual’s capabilities. In one sense this too is obvious; a person who has two language has access to a range of situations and experiences that are not available to the monolingual, whether these are the minimal possibilities conferred by two years of school French or the maximal possibilities open to the long-stay immigrant to Paris. If a monolingual has 100% language capacity in a single language, a second language user who knows only 5% of another language has a total of 105%. This can be called the additive mono­lingualist view: learning a second language increases the normal capacity of the individual and so confers a benefit rather than creating a problem.

Metalinguistic awareness and L2 users

One of the virtues ascribed to L2 learning by educational policy is that it sharpens the indiv­idual’s aware­ness of the nature of language. The UK National Criteria for GCSE (General Certificate of Education) in Modern Languages claims one objective is ‘to develop an aware­ness of the nature of language and language learning’ as an end in itself. An illustrious list of writers who speak two languages can be cited, in English, including Conrad, Nabakov, and Beckett, suggesting their writing may have been facilitated by their L2 learning. Conrad is a fascinating example of someone with native Polish, French spoken with ‘elegance’ , ‘general ease’ and ‘no trace of an accent’, and English spoken like ‘a foreigner, talking broken English’, according to Virginia Woolf, yet one of the greatest writers of English prose of the twentieth century (Page, 1986). General sensitivity to language and a greater precision in the choice of words are often claimed to be spin-offs from the ability to use another lang­uage.

The attempt to give precision to the view that learning a second language has a cognitive advantage for the learner has led to the area of meta­linguistic awareness - the awareness of language itself independent­ly of the message it is conveying. However, it should be pointed out that, from a linguist’s perspective, the aspects of language involved in such tests are limited. This is partly due to the limited range of tasks that have been used, which fall into three main types: tasks involving phonological awareness of the sound system of language; tasks involving grammaticality judgements which can show the person’s underlying knowledge of the language; tasks testing whether the person can separate the form of language from the meaning, that is to say awareness of the arbitrariness of words. Let us then look at the contributions under these three heads.

phonology and metalinguistic awareness

So far as the phonological level is concerned, early work by Cohen, Tucker and Lambert (1967) found that bilingual university students were better able to reproduce sound sequences that did not occur in their first language than monolinguals. When a similar approach was applied to children, however, Davine, Tucker and Lambert (1971) showed that, while third and fourth grade children in bilingual French/English classes were better than monolinguals at perceiving initial consonant sequences with a French second compon­ent, for example /vr/ as in vrai which does not occur in English, they were not better at the more general task of perceiving sound combinations that occur in neither French nor English, say /bg/. They argued that the general bilingual advant­age in phonology may be slow to develop. Nevertheless Rubin and Turner (1989) more recently found that primary school children in French immersion programs were better at phonemic segmentation than monolinguals.

grammar and metalinguistic awareness

Grammatical awareness has been extensively tested in research by Galambos and Hakuta (1988) and Galambos and Goldin-Meadow (1990). The design of Galambos and Goldin-Meadow (1990) postulated three levels of  language awareness: at the first ‘content’ level speakers may be able to say only that a sentence is ungrammatical, at the next ‘correction’ level they may be able to correct what is wrong with it, at the third ‘explanation’ level, they may be able to describe the ways in which it is ungrammatical. Children between four and a half and eight years of age, who were either monolingual English or Spanish or bilingual English/Spanish, were asked to evaluate and correct sentences covering such grammatical points as noun/verb agreement Claire and Eleanor is a sister, gender agreement in Spanish El pescado es bien bonita (the fish (m.) is very pretty (f.)), and adverb/adjective confusions The smartly boy read very quickly. Bilingual children progressed through the early stage of content-oriented awareness more rapidly than monolingual children; they were not, however, better at the later stage of explaining the errors. Nor were different types of error noted for monolinguals and bilinguals. In other words, while knowing two languages made the children progress through the levels more rapidly, it did not give them different abilities from monolinguals. Galambos and Hakuta (1988) also found a positive link in Spanish-speaking children between language awareness and proficiency in the first language.

A slightly different approach has been pursued by Bialystok (1987, 1993). Her model of second language acquisition recognizes two different dimensions that add up to language proficiency: control of processes  - the selection of the information for use - and analysis of knowledge–the way in which the language is represented in the mind. An often-made analogy is with a library where, on the one hand the books are set out on shelves according to a definite scheme (analysis of knowledge), on the other readers have a laid-down procedure for obtaining a book (control of procedures) (Bialystok & Sharwood-Smith, 1985). For example English learners of German have a well-known difficulty with the difference between the Subject Verb Object order of main clauses and the Subject Object Verb order of subordinate clauses. This might be because imperfect L2 procedures are controlling a correct knowledge of German order or because correct German procedures are controlling an imperfect knowledge of German. Different uses of language vary in the contrib­utions they require from the two dimensions of analysis and control. Children’s convers­ation is low on both control and analysis; lecturing is high on control and analysis, a disc jockey’s patter is high on control, low on analysis, and so on. Different metalinguistic tasks also vary on the two dimensions: correcting sentences is fairly low on control, but fairly high on analysis, while making grammaticality judgements is low on both control and analysis. In Bialystok (1987) bilingual English/French children performed in general better than monolinguals on control-based tasks; bilingual children with literacy skills also did better on analysis-based tasks. Bilingual children are better at spotting the grammaticality of semantically anomalous sentences (Why is the cat barking so loudly?) (high control) than monolinguals, but slightly worse at evaluating ungrammatical but non-anomalous sentences (Why the dog is barking so loudly?) (high analysis); they are also better at judging how many words there are in a sentence or string of words (high analysis, high control). Bilingualism primarily affects the control dimension of language, though not without effects on the analysis dimension. Bialystok’s contribution puts the different aspects of meta­linguistic awareness within an integrated model

arbitrariness of the sign

Since at least de Saussure (1916), linguistics has emphasized the arbitrary connection between the sounds or letters that make up a word and its meaning. A rose is indeed a rose whatever its name. But do most monolinguals agree? A rural myth has a Wiltshire farmer saying ‘Rightly is they called pigs because they have such dirty habits’. Most people may indeed find it hard to separate the qualities of the object from its name. Logically it might follow that an insight into the arbitrariness of language will come with greater ease to bilinguals, since they can see that one object in fact has two names in different languages.

The classic experiment by Ianco-Worrall (1972) asked Afrikaans/English children aged 4-9 years whether, if names could be interchanged, a dog could be called cow and a cow called dog. Over twice as many English/Afrikaans bilingual children agreed that this was possible than monolingual children. Similarly Ben Zeev (1977) told children that objects had a new name and then asked them whether it now had the characteristics of the new name: if we call a bird spaghetti, can we say spaghetti flies? A more difficult task included breaches of selection restrictions; if spaghetti is the new word for children, how do you say They are good children? The aim was to test if children were capable of appreciating the arbitrariness of the sign. English/Hebrew bilingual children were better than monolingual children, despite the fact that on tests of spontaneous speech they made more mistakes and that on tests of vocabulary they were below the monolinguals.

Several other tasks have been used to support the general notion that bilinguals are better at separating language form from language meaning. A further element in Ianco-Worrall’s experiment involved giving the children three words, say cap, can and hat and asking them ‘Which is more like cap - can or hat?’ The child’s choice was between phonetic similarity (cap/can) and semantic similarity (cap/hat). Bilingual children at 4-6 years of age were more likely to choose the semantic option than monolingual Afrikaans or English children, though by 7-9 years this difference had diminished. That is to say, they were better able to separate meaning from form. A word counting task was employed by Bialystok (1986), who found French/English bilingual children were better than monolinguals.

A neat test of arbitrariness of form and meaning is the size of words: do big objects have big names? Bialystok (1986) asked children to say which was the biggest word in such pairs as hippopotamus and skunk; bilinguals were better able to keep the word size distinct from the object size and answer the question correctly. Yelland, Pollard and Mercuri (1993) employed all possible combinations of big and large objects with big and large words (ant, cater­pillar, airplane, whale). After five months of one hour a week of Italian, English-speaking children in the first class of primary school showed significant differences from the ‘monolingual’ children, though this had largely disappeared in the next class up. Moreover the ‘bilingual’ children were better at tasks of written word recognition; that is to say they were learning to read better. Indeed Yelland et al (1993)  claim that one of the main benefits of learning an L2 may be the help it provides with learning to read.

other cognitive advantages

It has also been claimed that L2 users have advantages in other cognitive areas. Landry (1974) found that L2 users scored higher on standard tests of divergent thinking which value flexibility, originality and fluency, though this advantage over monolinguals did not appear earlier than the sixth year of an FLES (Foreign Language in the Elementary School) program. Lambert, Tucker & d'Anglejan (1973) showed that children in Canadian immers­ion programmes scored better on the ‘unusual uses’ test of creativity than monolinguals. Peal and Lambert (1962) found that ‘bilinguals appear to have a more diversified set of mental abilities’, as seen in performance on both verbal and non-verbal IQ tests. Feldman and Shen (1971) showed advantages for bilingual five-year-olds on tests of ‘object constancy, naming and the use of names in sentences’. Diaz (1985) added to the list of bilingual advant­ages ‘measures of conceptual development’, ‘creativity’, and ‘analogical reasoning’. Accurate comparisons are necessarily hard to make between bilingual and monolinguals in view of all the other possible differences between the groups, such as socio-economic status. There is also a lingering suspicion that there is a horse and cart problem: do bilinguals gain cognitive skills or are the special skills they already possess the reason why they have become bilingual? This would not of course apply to those who are not selected by others or by themselves for second language learning.

Nevertheless the conclusion is often that L2 users do have advant­ages, even after a few hours of second language acquisition (Yelland et al, 1993). This contrasts with the well-known views of Cummins (1978) that there is a threshold, below which L2 learning is a hindrance, and above which it is a gain. Yet it is in a way meaningless to compare the double system of the bilingual with the single system of the monolingual in terms of efficiency. While there may be losses or gains in specific areas, the overall system of the L2 user is more complex and has a greater range of uses. The payoffs or losses in other areas of cognition are outweighed by the ability to use two languages, with all the benefits it can bring to the individual and to society.

To sum up this section, the additive effects of using a second language seem to be an increas­ed meta­linguistic awareness of phonology, syntax, and the arbitrary nature of meaning, and gains in cognitive flexibility. While none of these may be overwhelming, they certainly contradict the notion that L2 using has a decremental effect on the user’s cognitive processing in general.


In a sense the consequences of bilingualism for cognitive processing are a question of swings and round­abouts. The slight cost of bilingual deficit on language processing is offset by the slight gains on other cognitive processes, without mentioning all the other gains of bilingualism in people’s lives. But is this the right way of looking at it? Should we ever compare two types of people in terms of a book-keeping exercise of profit and loss?

Let us come back to the ‘multilingualist’ view that knowing two or more languages is the norm. While it seems hard to garner precise statistics, partly because of the range of meanings of the word bilingual, it may well be that most people in the world use more than one language. Take a typical multi­lingual situation such as West Africa. In the Cameroon, two official languages are spoken, four lingua francas and 285 native languages (Koenig, Chia & Povey, 1983). Nigeria has 400 languages including three main languages spoken by 50% of the population; children learn English and a second major language at school (Bamgbose, 1994). In these countries a person needs to use a variety of languages for everyday living. If the study of second language acquisition had started in Yaoundé or Ibadan, their capital cities, rather than in the West, psychologists would have interpreted everything from a different multilingualist perspective. The issue of subtractive elements might have been seen as a by-product of over-specialisation; if you restrict yourself to hopping on one foot, you are going to be much better at hopping than someone who walks on two legs. The issue of additive elements again would be reversed; cognitive flexibility, meta­linguistic awareness, and so on would be the normal human state; people would investigate the reasons why mono­linguals do not attain the normal potential of a human being rather than look at the deficiencies of bilinguals. Gains and losses in terms of cognitive processing are then measured against a norm. The question is: what is the norm?

Undoubtedly the norm in most psychological and second language acquisition research till now has been the monolingual, against which L2 users are measured and found wanting. Most second language acquisition research techniques for instance do not look at the learner’s language in its own right but categorize mistakes in ‘Error Analysis’, look at morphemes ‘missing’ from obligatory contexts (defined with reference to monolinguals), compare grammaticality judgements with those of native speakers, and so on. The term bilingual has itself contributed to this bias towards the native speaker monolingual in suggesting that the only successful L2 learners are those that acquire a knowledge of the L2 identical to that of an L1 speaker; the term ambilingual seems preferable for this restricted meaning. Other people are seen as deficient in not coming up to this virtually unattainable goal. The term second language learner as used in second language acquisition research is also misleading as it too implies that people never get there - they are perpetual learners; second language user seems preferable as referring to people’s ability, not to how far they have travelled on a never-ending stairway to an unattainable heaven.

When Chomsky sets the agenda for linguistics as finding out what constit­utes knowledge of language (Chomsky, 1991), this is interpreted as knowledge of a single language, not of one or more languages, which he justified by the argument that the pure idealized form of language knowledge should be the first object of study rather than the muddy waters of bilingualism (Chomsky, 1986). Some writers on bilingualism indeed have started to challenge this view; the title of Grosjean’s important 1989 article proclaims ‘the bilingual is not two monolinguals in one person’: Romaine (1989) points out ‘it is clear that a reasonable account of bilingual­ism cannot be based on a theory which assumes monolingual competence as its frame of reference’; Hoffmann (1991) says that ‘For the vast majority of bilinguals, “bilingual competence” is not measurable in terms of monolingual standards’.

The choice of the norm is then rooted in preconceptions about the nature of human beings and of human society; in England the outsider may be the person who speaks more than one language, in Cameroon the person who speaks only one. When linking features of language with the speaker’s sex, class, or race, linguists have agreed since at least Labov (1972) to talk in terms of difference rather than deficit. The same courtesy can be extended to L2 users; the cognitive processes of second language users are whatever they are, neither to be praised or condemned relative to native speakers. They should be studied in their own right, not as deviations from a monolingual norm. As Grosjean (1994, 1657) puts it, we should see the bilingual ‘as a specific and fully competent speaker/hearer who has developed a communicat­ive competence that is equal, but different in nature, to that of the monolingual’.


Adams, S.J. (1983). Scripts and Second Language Reading Skills. In J.W. Oller, Jr. & Richard-Amato, P.A. (eds.), Methods that Work. Rowley MA: Newbury House

Baddeley, A.D., Thomson, N., & Buchanan, M. (1975).  Word length and the structure of short-term memory. Journal of Verbal Learning and Verbal Behavior, 14, 575-589

Bamgbose, A. (1994). Nigeria’s choice. UNESCO Courier, February, 22-31

Ben Zeev, S. (1977). The influence of bilingualism on cognitive strategies and cognitive development. Child Development, 48, 1009-18

Bialystok, E. (1986). Children’s concept of word. J.Psycholinguistic Research, 15, 13-22

Bialystok, E. (1987). Influences of bilingualism on metalinguistic develop­ment. Second Language Research, 3, 2, 154-166

Bialystok, E. (1993). Metalinguistic dimensions of bilingual language proficiency. In E. Bialystok (ed.), Language Processing in Bilingual Children. Cambridge: CUP, 113-140

Bialystok, E., & Sharwood-Smith, M. (1985). Interlanguage is not a state of mind: an evaluation of the construct for second-language acquisition. Applied Linguistics, 6, 1, 101-117

Birdsong, D. (1992). Ultimate attainment in second language acquisition. Language, 68, 706-55

Bransford, J.D., & Johnson, M.K. (1982). Contextual prerequisites for understanding: some investigations of comprehension and recall. Journal of Verbal Learning and Verbal Behavior, 11, 717-726

Brown, G.D.A. & Hulme, C. (1992). Cognitive processing and second language processing: the role of short term memory. In R.J. Harris (ed.), Cognitive Processing in Bilinguals. Amsterdam: Elsevier, 105-121

Carrell, P.L. (1984). Evidence of a formal schema in second language comprehension. Language Learning, 34, 87-111

Caskey-Sirmons, L.A., & Hickerson, N.P. (1977). Semantic shift and bilingualism: variation in the color terms of five languages. Anthropological Linguistics, 19/8, 358-367

Chomsky, N. (1986). Knowledge of Language: Its Nature, Origin and Use, New York: Praeger

Chomsky, N. (1991). Linguistics and adjacent fields: a personal view. In A. Kasher (ed.), The Chomskyan Turn. Oxford: Blackwell, 5‑23

Cohen, S.P., Tucker, G.R., & Lambert, W.E. (1967). The comparative skills of monolinguals and bilinguals in perceiving phoneme sequences. Language and Speech, 10, 159-168

Conrad, R. (1964). Acoustic confusions in immediate memory. British Journal of Psychology, 55, 75-84

Conrad, R. (1972). Form and colour as short-term memory codes in pre-school children. Psychonomic Science, 27, 225-226

Cook, V.J. (1977). Cognitive processes in second language learning. Inter­nation­al Review Applied Linguistics, XV/1, 73-90. Reprinted in D. Nehls (ed.), Studies in Language Acquis­ition. Julius Groos, 1980

Cook, V.J. (1979). Aspects of memory in secondary school language learners. Interlanguage Studies Bulletin—Utrecht. 4, 2, 161-172

Cook, V.J. (1981). Some uses for second language learning research. Annals of the New York Academy of Sciences, 379, 251-258

Cook, V.J. (1990). Timed comprehension of binding in advanced L2 learners of English. Language Learning, 40, 4, 557-599

Cook, V.J. (1991). Second Language Learning and Language Teaching. Edward Arnold

Cook, V.J. (1992).  Evidence for multi-competence. Language Learning, 42, 4, 557-591

Cook, V.J. & Newson, M. (1996). Chomsky’s Universal Grammar. Oxford: Blackwell

Coppetiers, R. (1987). Competence differences between native and near-native speakers. Language, 63, 3, 545-573

Cromer, R.F. (1974). The development of language and cognition: the cognition hypothesis. In Foss, B. (ed.), New Perspectives in Child Language. Harmondsworth: Penguin

Cromer, R.F. (1991). Language and Thought in Normal and Handicapped Children. Oxford: Blackwell

Cromer, R.F. (1974), The development of language and cognition: the cognition hypothesis. In B. Foss (ed.), New Perspectives in Child Language. Harmondsworth: Penguin

Cromer, R.F., (1991). Language and Thought in Normal and Handicapped Children. Oxford: Blackwell

Cummins, J. (1978). Bilingualism and the development of metalinguistic awareness. Journal of Cross-Cultural Psychology. 9, 131-149

Davine, M, Tucker, G.R., & Lambert, W.E. (1971). The perception of phoneme sequences by monolingual and bilingual elementary school children. Canadian Journal of Behavioral Science, 3, 1, 72-76

de Saussure, F. (1916). Cours de Linguistique Générale. C. Bally, A., Sechehaye, & A. Reidlinger (eds.). Paris: Payot. Trans. 1959 as Course in General Linguistics. London: Peter Owen

Diaz, R.M. (1985). The intellectual power of bilingualism. Quarterly News­letter of the Laboratory of Comparative Human Cognition, 7, 1, 16-22

Dornic, S. (1969). Verbal factor in number perception. Acta Psychologica, 29, 393-399

Ellis, N. (1992). Linguistic relativity revisited: the bilingual word length effect in working memory. In R.J. Harris (ed.), Cognitive Processing in Bilinguals. Amsterdam: Elsevier, 137-155.

Ellis, N., & Beaton, A. (1993). Factors affecting the learning of foreign language vocabulary imagery keyword mediators and phonological short-term memory. Quarterly Journal of Experimental Psychology, 46A, 533-558

Ellis, N., & Henelley, R.A. (1980). A bilingual word-length effect: implicat­ions for intelligence testing and the relative ease of mental calculation in Welsh and English. British Journal of Psychology, 71, 43-52

Feldman, C., & Shen, M. (1971). Some language-related cognitive advant­ages of bilingual five-year-olds. Journal of Genetic Psychology, 118, 235-244

Galambos, S.J., & Hakuta, K. (1988). Subject-specific and task-specific characteristics of metalinguistic awareness in bilingual children. Applied Psycholinguistics, 9, 141-161

Galambos, S.J., & Goldin-Meadow, S. (1990). The effects of learning two languages on metalinguistic awareness. Cognition, 34, 1-56.

Gathercole, SE & Baddeley, A.D. (1993). Working Memory and Language, Erlbaum

Glicksberg, D.H. (1963). A Study of the Span of Immediate Memory among Adult Students of English as a Foreign Language. Ph.D. University of Michigan

Grosjean, F. (1989). Neurolinguists, beware! The bilingual is not two monolinguals in one person. Brain and Language, 36, 3-15

Grosjean, F. (1994). Individual bilingualism In The Encyclopedia of Language and Linguistics. Oxford: Pergamon Press, 1656-1660

Halliday, M.A.K., McIntosh, A., & Strevens, P. (1964). The Linguistic Sciences and Language Teaching. London: Longman

Hoffman, C. (1991). An Introduction to Bilingualism. London: Longman

Hoosain, G.J. (1979). Forward and backward digit span in the languages of the bilingual. Journal of Genetic Psychology, 135, 263-268

Hummel, K.M. (1986). Memory for bilingual prose. In J. Vaid (ed.) (1986). Language Processing in Bilinguals: Psycho­ling­uistic and Neuro­linguistic Perspectives. Hillsdale, NJ: Lawrence Erlbaum, p.47-64

Hunt, E. & Agnoli, F. (1991). The Whorfian hypothesis: a cognitive psychological perspective. Psychological Review, 98, 377-389

Ianco-Worrall, A. (1972). Bilingualism and cognitive development. Child Development, 43, 1390-1400

Kinsbourne, M, & Cohen, V. (1971). English and Hebrew consonant memory span related to the structure of the written language. Acta Psychologica, 35, 347-351

Koenig, E.L., Chia, E., & Povey, J. (1983). A Sociolinguistic Profile of Urban Centers in Cameroon. Kinsey Hall, UCLA: Crossroads Press

Lambert, W.E. (1955). Measurement of the linguistic dominance of bilinguals. Journal of Abnormal and Social Psychology, 50, 197-200

Lambert, W.E. (1956). Developmental aspects of second-language acquisition, Journal of Social Psychology, 43, 83-102

Lambert, W.E., Tucker, G.R., & d’Anglejan, A. (1973). Cognitive and attitudinal consequences of bilingual schooling. Journal of Educational Psychology, 85, 2, 141-159

Landry, R.G. (1974). A comparison of second language learners and mono­linguals on divergent thinking tasks at the elementary school level. MLJ, 58, 10-15

Lehtonen, J., & Sajavaara, K. (1988). Psycholinguistic testing of transfer in foreign-language speech processing. Folia Linguistica, XXII, 179-202

Long, J., & Harding-Esch, E. (1977). Summary and recall of text in first and second languages: some factors contributing to performance difficulties. In H. Sinaiko, & D. Gerver (eds.). Proceedings of the NATO Symposium on Language Interpretation and Communication. New York: Plenum Press

Magiste, E. (1979). The competing linguistic systems of the multilingual: a developmental study of decoding and encoding processes. Journal of Verbal Learning and Verbal Behavior, 18, 79-89

Magiste, E. (1980). Memory for Numbers in Monolinguals and Bilinguals. Acta Psychologica, 46, 63-68

Magiste, E. (1982). Automaticity and interference in bilinguals. Psycho­logical Research, 44, 29-43

Magiste, E. (1986). Selected issues in second and third language learning. In J. Vaid (ed.) (1986), Language Processing in Bilinguals: Psycho­ling­uistic and Neurolinguistic Perspectives. Hillsdale, NJ: Lawrence Erlbaum, 97-122

Magiste, E. (1992). Second language learning in elementary and high school students. European Journal of Cognitive Psychology, 4, 355-365

Marcus, M.P. (1980). Theory of Syntactic Recognition for Natural Languages. Cambridge MA: MIT Press

Marsh, L.G., & Maki, R.H. (1978). Efficiency of arithmetic operations in bilinguals as a function of language. Memory and Cognition, 4, 4, 459-464

Miller, G.A. (1956). The magical number seven plus, or minus two: some limits on our capacity for processing information. Psychological Review, 63, 81-97

Naveh-Benjamin, M., & Ayres, T.J. (1986). Digit span, reading rate, and linguistic relativity. Quarterly Journal of Experimental Psychology, 38A, 739-751

O'Mahoney, M., & Muhiudeen, H. (1977). A preliminary study of alternative taste languages using qualitative description of sodium chloride solutions: Malay versus English. British Journal of Psychology, 68, 275-278

Oyama, S. (1976). A sensitive period for the acquisition of a non-native phonological system. Journal of Psycholinguistics Research, 5, 261-285

Page, N. (1986). A Conrad Companion. Basingstoke: Macmillan

Papagno, C., & Vallar. G. (1992). Phonological short term memory and the learning of new words: the effects of phonological similarity and item length. Quarterly Journal of Experimental Psychology, 44A, 47-67

Papagno, C., Valentine, T., & Baddeley, A. (1990). Phonological Short-Term Memory and foreign-language vocabulary learning. Journal of Memory and Language, 30, 331-347

Peal, E., & Lambert, W.E. (1962). The relation of bilingualism to intelli­gence. Psychological Monographs: General and Applied, 76, 27, 1-23

Peynircioglu, Z.F., & Tekcan, A.I. (1993). Word perception in two languages. Journal of Memory and Language, 33, XXXXXX

Ransdell, S.E., & Fischler, I. (1987). Memory in a monolingal mode. Journal of Memory and Language, 26, 392-405

Roller, C.M., & Matombo, A.R. (1992). Bilingual readers use of background knowledge in learning from text. TESOL Quarterly, 26, 1, 129-140

Romaine, S. (1989). Bilingualism. Oxford: Blackwell

Rubin, H., & Turner, A. (1989). Linguistic awareness skills in grade one children in a French immersion setting. Reading and Writing, 1, 73-86

Savin, H.B. & Perchonock, E. (1965). Grammatical structures and the immediate recall of English sentences. Journal of Verbal Learning and Verbal Behavior, 4, 348-353

Service, E. (1992). Phonology, working memory and foreign-language learning. Quarterly Journal of Experimental Psychology, 45A, 21-50

Service, E., & Craik, F.I.M. (1993). Differences between young and older adults in learning a foreign vocabulary. Journal of Memory and Language, 33, 59-74

Skehan, P. (1989). Individual Differences in Second-Language Learning. Edward Arnold

Stigler, J.W., Lee, S-Y, & Stevenson, H.W. (1986). Digit memory in Chinese and English: evidence for a temporally limited store. Cognition, 23, 1-20

Vaid, J. (ed.) (1986). Language Processing in Bilinguals: Psycholinguistic and Neuropsychological Perspectives. Hillsdale, NJ: Lawrence Erlbaum

Yelland, G.W., Pollard, J., & Mercuri, A. (1993). The metalinguistic benefits of limited contact with a second language. Applied Psycholinguistics, 14, 423-444