BILINGUAL MISSION OBJECTIVES
This approach offers our children a very unique opportunuty:
THINK AND PLAY IN ENGLISH!
Just as M. Montessori theorized, the child's absorbent mind soaks up everything around him with extreme rapidity and semplicity. Knoledge of a second language from the first years of a child's life stimulates creativity in his mind. Children are sensitive to linguistic learning from infancy they are capable of distinguishing sounds that adults cannot, and are not afraid to make mistakes. Also, they are able to learn languages by simply listen and interacting with their peers.
The more, the merrier: facts and beliefs about the bilingual mind.
University of Edinburgh
To appear in Della Sala, S. (ed.) 2006. Tall Tales about the Mind and Brain: Separating Fact from Fiction. Oxford: Oxford University Press.
Anyone who has seen a small child switching from one language to another is likely to be amazed – and perhaps envious – at how effortlessly they are able to do this. Stories of immigrant children interpreting for their parents are commonplace, and in some parts of the world it is quite normal for children to be exposed to two or even more languages right from birth. Yet in modern industrial societies growing up with more than one language is often regarded as ‘special’. Bilingualism is still surrounded by false beliefs and misunderstandings, even among the otherwise educated and scientifically-minded. Many people are ready to believe that handling two languages at the same time is too much of a burden for the infant’s brain, or that the languages compete for resources in the brain at the expense of general cognitive development. The contrast between and these false beliefs and the amazement often expressed by people at how easily children pick up two or more languages has been termed the ‘bilingual paradox’ (Petitto & Kovelman 2003).
Why are these beliefs so resilient? Their enduring popularity might have to do, at least in part, with the fact that many people find it difficult to think scientifically about language, and therefore everyone feels entitled to have strong opinions about it: the world is full of linguistics experts. With regard to bilingualism, opinions are unfortunately not restricted to the domain of academic discussion, but often inform decisions – by parents, professional educators, policy makers – that end up affecting children’s lives. Many new parents who want their children to speak two languages for family reasons are likely to have heard somewhere that exposure to two languages can cause problems, so they may abandon bilingualism before they even give it a try, or may plan to introduce one of the languages only after the other is ‘well established’ and then find to their regret that the second language never has a chance. If they successfully establish bilingualism in their pre-school children, they may well be made to feel that they’ve created a problem by well-meaning primary school teachers, who are often ready to blame bilingualism for any performance problems. In this situation the parents may abandon successful bilingualism and even make active efforts will be made to re-establish monolingualism to ‘cure’ the problem. Given the sociological repercussions of these folk linguistic beliefs, it seems valuable to bridge the gap between the scientific approach to the study of bilingual cognition and what many people believe about life with two languages. In this chapter, we try to dissect some particularly strong misconceptions that are still alive and well and affecting the daily lives of bilinguals.
Myth 1: bilingual children are less intelligent than monolinguals (or alternatively: bilingual children are more intelligent than monolinguals).
Does knowledge of more than one language make you smarter? Or is it rather a cognitive handicap, at least in the early childhood years? The answer is, quite simply, neither. There is no link between bilingualism and general ‘intelligence’. Early research in the 60s suggested that bilinguals have a cognitive handicap, while subsequent studies in the 70s seemed to find that bilinguals are more intelligent than monolinguals. More recently, however, both conclusions have been found to be marred by failing to take important sociological and cultural effects into account (Grosjean 1982). The fact appears to be that bilingual children are neither more nor less intelligent than their monolingual peers.
Nevertheless, the experience of dealing with two languages does seem to give bilingual children some cognitive advantages in several domains. Such advantages are particularly evident in tasks that involve cognitive flexibility and the control of attention (Bialystok 1991; 2001): bilinguals seem to be better at selectively paying attention, at inhibiting irrelevant information, and at switching between alternative solution to a problem. In contrast, bilinguals do not seem to have an advantage over monolinguals with respect to functions that depend on the way knowledge is represented. For example, they don’t seem to be any better and encoding problems, at accessing relevant knowledge, or at drawing logical inferences.
What is the link between enhanced cognitive control and bilingualism? Bilingual speakers must develop a powerful mechanism for keeping the two languages separate, so that fluency in one language can be achieved without intrusions from the unwanted language. Therefore, the bilingual child’s constant experience of having two languages available and inhibiting one when the other is activated (Green 1986; 1998) enhances their ability to multitask in other domains. And there is more good news for bilinguals: it’s been suggested that some of these cognitive advantages are maintained in old age (Bialystok et al. 2004). If these results are confirmed by future research, it will be possible to conclude that bilingualism provides a defence against the decline of general processing functions that is a feature of normal cognitive aging.
A further spin-off of bilingualism is higher awareness of language and greater ability to think about it and talk about it. Bilingual children have a greater ability to focus on the form of language, abstracting away from meaning. Parents of bilingual children often report that their children engage in ‘language play’ that may take the form of ‘funny accents’ or impossible literal translations between one language and another. Many parents also report that bilingual children have more precocious reading skills, and this has recently been confirmed experimentally (Petitto and Dunbar 2004). Bilingual children recognize symbolic letter-sound correspondences earlier than monolinguals, although this does not appear to be related to greater awareness of the sounds themselves and it is also a function of the specific languages acquired as well as of the level of proficiency attained (Bialystok et al. 2003).
Because of their experience of selecting languages according to the perceived linguistic competence of the person they are addressing, bilingual children have also been said to have an enhanced ‘awareness of the other’. This often goes under the heading of “Theory of Mind”, which is a term used to describe the ability to understand other people’s mental states, and more specifically that other people may have beliefs, desires and intentions different from one’s own. (Perner and Lang, 1999). In the classic “Sally-Anne’ Theory of Mind test, the researcher uses two dolls, "Sally" and "Anne.” Each of them has her own basket. Children watch Sally put a marble in her basket and then leave the scene. While Sally is away, Anne moves the marble from Sally's basket into her own. Sally then comes back and the children are asked where they think she will look for her marble. Children pass the test if they say that Sally will first look inside her basket before realizing that her marble isn't there; they fail if they say that Sally looks into Anne’s basket. The cognitive abilities involved in Theory of Mind normally emerge around the age of 4 in monolingual children; they are permanently impaired in autistic children. Bilingual children develop Theory of Mind, on average, a year earlier than monolinguals, so they succeed in classic false belief tasks at age 3 (Goetz 2003; Kovácz 2005). Theory of Mind has also been found to correlate with central executive functions (planning, problem-solving, inhibition of habitual responses), so bilinguals’ superior performance may be due to their greater ability to suspend their own irrelevant beliefs, rather than to an understanding of other people’s mental state. Still, it is remarkable that the experience of dealing with two languages may have such extensive repercussions in so many apparently unrelated domains of cognitive development.
Myth 2: bilingual children are slowed down in their general cognitive development by the burden of handling two languages.
The idea that learning two languages from birth represents a burden is based on the assumption that the brain is naturally predisposed to deal only with one language. However, research in psychology and neuroscience research indicates that there are no foundations to the belief that monolingualism is somehow the biological norm. While is it true that the onset of speech in bilinguals may be slightly later than average, both monolingual and bilingual children go through the same major milestones in language development at approximately the same time. Commonly recognized stages are: babbling (playing with sounds apparently without intending to convey meaning) during the period of roughly 6-12 months; the emergence of single words about the end of the first year; the “50-word stage” at 14-18 months, after which there is a sudden explosion in vocabulary size; the two-word utterance stage at 18-24 months; and the emergence of multiword utterances sometime around the end of second year. If the brain were set up to acquire only one language, bilinguals would be at a disadvantage: they might be expected to reach the milestones later, or at different times in their two languages. The fact that they follow the same developmental timetable as monolinguals points to the brain’s capacity to deal with multiple types of language input.
Very persuasive evidence that the brain can easily accommodate more than one language comes from studies by Petitto and her team (Petitto et al. 2001). Petitto compared the more common case of bilingual children acquiring two spoken languages with that of hearing children of deaf parents, who were acquiring both a spoken language and a signed language. Her reasoning was that, if exposure to two languages causes delays and confusion, these should be particularly apparent in the sign-spoken bilingual group. What she found was that, as in the case of two spoken languages acquired simultaneously, the signed language and the spoken language follow similar developmental timetables: such babies go through developmental milestones (including a phase of ‘babbling with the hands’) at about the same time. This shows that babies are not sensitive to speech or sound per se, but rather to the abstract patterns and regularities that are encoded by any language in either modality. Petitto’s proposal is that children are sensitive to distributional, rhythmical, and temporal patterns that uniquely characterize the structure of human language. These patterns are found in any language, whether spoken or signed: so sign-spoken bilingual children are not hampered by exposure to different modalities, and achieve distinct representations in both languages just like bilingual children acquiring two spoken languages.
Myth 3: bilingual children speak a ‘mixed’ language in their first years and end up not speaking either language properly.
A hopelessly mixed language is the thing that many parents in bilingual families typically fear. Early research on bilingual children actually seemed to show that children are unable to distinguish between the two languages to which they are exposed. The result – it was claimed – is a single unitary system in which both the vocabularies and the grammars of the two languages are fused. Language mixing – it was believed – was a telling sign of this lack of differentiation. Another sign was the fact that in some bilingual children the early words often involve a mix of words taken from both languages, with many referents named by only one word. So for example a German-Italian bilingual child might have either Apfel or mela for ‘apple’, but not both. This led to the hypothesis that there was a unitary lexicon, which could not contain two words, one from Language A and one from Language B, for the same referent.
More recent research has completely discredited the idea of the unitary system. First, there are new techniques for studying whether babies can tell the difference between one outside stimulus and another. If you show a child a picture of a ball it will eventually get bored and look away, but if you then show it a picture of a car it will look again. There are now experimental techniques that let us present pictures or sounds to a child until it gets bored, then present it with something subtly different and see if the child notices the difference (Jusczyk 1997). Using techniques like these, we have learned that monolingual babies’ perceptual abilities are remarkably fine-tuned very early on: they know a lot about what their language sounds like long before they start producing their first words, and even at the age of a few months will notice when someone who was speaking English switches to speaking, say, Japanese. This makes it very implausible that bilingual children do not realize that they are hearing two languages. Indeed, they seem to be even more sensitive than monolinguals to a wider range of phonetic contrasts, and they may retain this ability for a longer time than monolinguals, for whom the ‘window’ of highest perceptual sensitivity to contrasts that are not present in their own language closes around 14 months.
Second, research on ‘code-switching’ – swapping back and forth between languages – shows that bilingual children, like bilingual adults, often switch from one language to another in order to achieve particular communicative effects. For example, even if they are talking in Language A, they may switch to Language B to report something that somebody said, if the speech they are reporting was originally in Language B. Or they may switch because of the topic they are talking about, or simply to play games with their languages. Naturally, this kind of code-switching takes place most often when bilinguals are talking to other bilinguals – when they are in what Grosjean (1998) has called ‘bilingual mode’ – not when they are talking to monolinguals. Moreover, code-switching is not random but generally obeys a remarkably strict grammar. For example, a Spanish-English bilingual child is much more likely to say ‘La house’ than ‘The casa’ (Spanish article + English noun, rather than ‘English article + Spanish noun), apparently favoring the combination that is more informative in terms of grammatical features like gender and number. Far from producing random mixings due to confusion, in other words, bilingual children know when and how it is appropriate to mix their languages.
What about the grammars of the two languages? Do bilingual children import the structures of one language into the other? Sometimes they do (though it is difficult to know whether they are doing so deliberately or not), but most of the time they keep their languages separate. The most interesting counter-evidence to the confusion hypothesis comes from research that compares the order of acquisition of grammatical structures in monolingual and bilingual children. There is little evidence that the bilinguals’ languages affect each other – they neither speed up nor delay normal acquisition processes. For example, children acquiring a language with complex morphology (grammatical endings, etc.) such as Italian normally start using verb inflections earlier than children acquiring a language with relatively simple morphology such as English. At an age when the English child still says ‘Daddy eat cookie’ and ‘me eat cookie’, the Italian child says ‘Papà mangia il biscotto’, ‘Io mangio il biscotto’, with the appropriate verb endings. If the two languages affected each other in development, one might expect an English-Italian bilingual child to do one of two things: either to acquire morphology in English earlier than monolinguals (as an effect of Italian) or to acquire morphology in Italian later than monolinguals (as an effect of English). In fact, Italian-English bilingual children do neither: the acquisition of verb morphology happens first in Italian and then in English, following exactly the same schedule as in monolingual children.
Of course this doesn’t mean that the features of one language never show up in the other. One current hypothesis is that ‘leakages’ between languages occur with constructions where the speaker must both know the grammar and understand the contexts in which a given grammatical choice is appropriate (Müller and Hulk 2001). An example is the possibility of ‘dropping’ subject pronouns in Italian. In Italian, as in other languages that allow sentences without a subject, subjects can be dropped when it is clear in context who the referent is. So if I say Maria non c’e’, e’ andata a casa (lit. ‘Maria’s not here, went home’) I can omit the pronoun in the second clause because it is clear that I’m referring to Maria, who has been mentioned in the previous clause. In other cases, the omission of subject pronouns is blocked for contextual reasons. So if I say ‘Maria e Yuri non si capiscono: lei parla l’italiano, lui no’ (‘Maria and Yuri can’t understand each other; she speaks Italian, he doesn’t’), I have to use an explicit pronoun because I’m contrasting two different people. English is much simpler in a sense, because it always has a subject, regardless of whether the sentence is referring to an easily accessible referent or not.
Now, bilingual children who acquire Italian and English often use too many explicit pronouns where it would be contextually appropriate to drop them (Serratrice et al 2004). Why do they do this? If you are a child acquiring Italian, you need to acquire two kinds of knowledge: first, you have to learn that it is grammatical to omit subject pronouns, and second, you have to know about the contextual conditions that favor dropping or not dropping subject pronouns. These conditions are sometimes called ‘interface conditions’, because they sit at the boundary between the grammar itself and the wider discourse context in which language is used, as represented in Fig.1.
One possibility is that bilingual children have a problem operating with these interface conditions in ‘real-time’ communication, so they extend the less complicated English system to Italian. The other possibility is that they are more aware of potential miscommunication and want to make the sentence as explicit as possible. Bilinguals experience potential miscommunication more often than monolinguals because they learn to adapt their language choice to the person they are addressing. So this ‘redundant’ feature of bilingual speech may betray a greater sensitivity to potential ambiguity. It is interesting to notice that these interface ‘crossovers’ from one language to the other are not always from the ‘dominant’ language to the ‘non-dominant language (i.e from the language the child hears or uses more to the one it hears or uses less), but rather from the ‘simpler’ language to the ‘more complex’ language. In other words, if one of the two languages offers a simpler way of expressing the same meaning, bilingual children will sometimes import that into the other language.
All in all, then, it appears that children are well aware of the difference between their two languages from very early on, and that each language develops in more or less the normal way, independently of the other. Curiously, however, what is known about bilingualism and the brain looks at first sight inconsistent with this idea. The latest generation of neuroscience studies has taken a serious interest in bilinguals, and particularly in the way the representation of the two languages in the brain is affected by the age at which they’re acquired. Several modern neuroimaging studies (such as fMRI and PET) indicate that early or simultaneous exposure to two languages from birth results in both languages being represented in the same areas within the left hemisphere, which are normally associated with the native language of monolingual speakers (Wartenburger et al. 2003; Abutalebi et al. 2001). By comparison, exposure to one language and then subsequent exposure to a second results in more bilateral representation, and in the involvement of more distributed frontal lobe areas typically recruited in working memory and inhibitory tasks. According to some models (Ullman 2001), early bilingualism is neurally distinguished from late bilingualism by its reliance on procedural memory mechanisms, which are implicated in long-established skills, or habits and are not accessible to consciousness; late(r) bilingualism, on the other hand, would mainly be served by declarative memory mechanisms, which are typically used in learning facts and events, and may be explicitly recalled. The brain signatures of these different neural mechanisms are visible in the greater involvement of left frontal ganglia brain areas in early bilinguals, as opposed to the recruitment of temporal lobe areas in later bilinguals.
However, these signatures are not permanently fixed: proficiency in a second language can involve brain reorganization (Perani et al. 1998). The flexibility of the brain has also emerged from a study by Mechelli (2004), who analyzed structural differences between the brains of bilinguals and monolinguals. The participants in his study had started learning English at various ages and had reached varying levels of proficiency. What he found was that bilinguals have more gray matter than monolinguals – in other words, that they develop a bigger brain in response to exposure to two languages – and that the growth involves a portion of the cortex (the left inferior parietal cortex) that is involved in some verbal fluency tasks and in verbal short-term memory. This happens not only to people who are bilingual from birth, but also to a lesser extent to late bilinguals. So while it may not be justified to say that bilinguals are more intelligent, it does seem that they may literally have bigger brains.
Myth 4: people can’t learn languages properly after the ‘critical period’ for language acquisition.
Unlike the first three myths, this belief has some basis in fact, but it’s still worth taking a closer look. We’ve seen that simultaneous early exposure to more than one language seems to provide an effortlessly natural path to becoming bilingual. What about older learners? Can they aspire to become fully bilinguals? And if they achieve fluency in a second language (as many people do in our increasingly multicultural societies), are they really different from monolinguals? Are there aspects of language that simply cannot be acquired after a certain age?
A theory that has been around for some time (Lenneberg 1967) maintains that there is a special period for the acquisition of language – a so-called ‘critical period’ – such that humans are maximally predisposed to acquire language early in life, and complete success at acquiring language can be guaranteed only by exposure during this early period. This seems true enough for the acquisition of a first language: if children are not exposed to any linguistic input during the first few years of their life, their language abilities, and especially their grammatical abilities, are irrevocably compromised. We know this from the few ‘natural experiments’ involving so-called feral children who grew up in conditions of extreme isolation and deprivation; we also know this from recent studies on deaf children of hearing parents, who are often not diagnosed as deaf until two or three years of age and who therefore are unknowingly deprived of language input up to that point. So why is there a critical period? We still don’t have a precise answer to this question, although we have a few good hypotheses. The critical period may be related to the existence of a biological mechanism that is innately geared to the acquisition of language in our species. There may be evolutionary reasons why this mechanism is available in early childhood, when the ability to communicate confers particular advantages in terms of success at survival. In contrast, the advantages of learning languages later in life are not so conspicuous for our species, although they certainly are for us as individuals.
Yet the detailed reasons why the outcome of second language acquisition in adults is different from first language acquisition in children (and from second language acquisition in children, for that matter) are far from clear. Are these differences due only to a decline in specific linguistic abilities, or do they result from age-related changes in more general skills, such as memory, for example? After all, second language speakers have been through the process of acquiring language once: why couldn’t second language acquisition be helped by “transferable skills” that are already in place? Moreover, we shouldn’t just assume that adults can never attain the levels in a second language that they effortlessly achieved in their first: we don’t know exactly how good an adult can become at a second language, or, to put it differently, we don’t know what the actual limits of adult second language acquisition are.
One way of addressing this question is to study so called ‘near-native’ speakers, that is, people who have achieved the highest level of competence in a second language. These speakers often pass for native speakers, unlike the vast majority of second language speakers who remain recognizably foreign, and in this sense they may be regarded as exceptional. However, exceptions have to be explained: near-native speakers provide us with evidence of what can be attained and what cannot, because their learning capacities have been pushed to their limit. So if we compare a speaker who is bilingual from birth with one who has become a near-native speaker of a second language, what do we find? The most obvious difference is that near-native speakers typically differ from early bilinguals because they retain a foreign accent. As we have seen, early exposure to a language seems to play some role in attuning first language acquirers to the phonetic details of their native language. As for grammatical abilities, studies of near-natives consistently point to native-like knowledge of syntax and morphology, but that this knowledge can temporarily fail in real-time communication. Near-natives have slower reaction times than native speakers in psycholinguistic experiments, and their comprehension abilities are less tolerant of noisy surroundings. Interestingly, too, problems have been observed in the way near-natives produce and understand sentences that require an interface between grammatical and contextual knowledge (Sorace 2003). These are the same areas that, as we have seen, are prone to ‘leakages’ in bilingual children. The difference is that in bilingual children these problems are developmental, and disappear with more exposure; they tend to be permanent in near-natives. If there is a remaining deficit in second language learners who have reached a near-native level, it seems to be an the level of the processing abilities required to link different sources of information in fluent use of the second language (Clahsen and Felser 2006).
In any case, the comparison between people who are bilingual from birth and even near-native speakers of a second language shows the fallacy of one approach that parents sometimes take to establishing bilingualism – introducing one language first, and waiting until it is established before exposing the child to the other. In all kinds of ways this appears to be the worst mistake one could make. There’s no reason to think that children cannot handle acquiring two languages simultaneously, and some reason to think that early bilingualism confers various cognitive advantages. If one of the family languages is introduced after the first one is established, some aspects of it may not be acquired in a native-like way. This is particularly true of phonetic and articulatory aspect of a ‘native accent’, which seem to be best acquired within a narrower window of opportunity; but also of some grammatical features for which children are maximally sensitive in the earliest period of their childhood. While children can and often do acquire second languages in later childhood, and often reach native-like mastery in these, the best outcomes are guaranteed by simultaneous exposure. If this is possible, why not go for it? Perhaps equally important, depriving children of exposure to one of the family languages may have other, more sociological, consequences. Children in this situation may feel that the second language is ‘less important’ than the first, and not really worth speaking.
Children do not know about the cognitive benefits of bilingualism, but they are exquisitely sensitive to the status and prestige of each language within the family and in the outside world. Even if the cognitive windows are still open, closed attitudes may put bilingualism out of reach.