Aptitude Effects in Near-Native Second Language Acquisition

30 Oct

Abrahamsson, N., & Hyltenstam, K. (2008). THE ROBUSTNESS OF APTITUDE EFFECTS IN NEAR-NATIVE SECOND LANGUAGE ACQUISITION. Studies in Second Language Acquisition, 30(04), 481–509.

In their study “The Robustness of Aptitude Effects in Near-Native Second Language Acquisition”, Abrahamsson and Hyltenstam (2008) investigated individuals who had learned Swedish as a second language (L2) either in childhood or adulthood. Their aim was to find out whether a high level of language aptitude (i.e. language talent) affects how likely the two groups (early and late L2 learners) attain native-like levels of proficiency in their L2.

The researchers predicted that a high level of language aptitude would only be a predictor of success in Swedish (i.e. how native-like they were in their L2) in the late learners, but not in the early learners. The reason why they predicted that a high level of aptitude would only “help” the late L2 learners is that they work with the idea that early language learners are at an advantage when learning second languages anyway (because they are young). In other words, children don’t need “help” (here, in the form of language aptitude) like adults do: learning an L2 for children comes naturally, according to Hyltenstam and Abrahamsson. In order to investigate the effect of aptitude on early versus late learners, the authors used language aptitude tests (developed in Swansea, Wales, largely by Paul Meara), and various tests to examine proficiency levels in Swedish. In total, Abrahamsson and Hyltenstam investigated  42 highly proficient L2 speakers of Swedish (Spanish was their L1; 31 were early learners, whereas 11 were late learners). Fifteen native speakers of Swedish comprised the control group against which L2 proficiency was measured.

As predicted, their findings showed that the language aptitude scores had a greater effect on predicting the proficiency levels in the late L2 learner group than in the early L2 learner group. Surprisingly, their findings also showed a small but significant influence of aptitude on early L2 learners. In brief, the authors conclude that individuals who have a high language aptitude are better equipped to learn an L2 in adulthood than individuals who have less aptitude, and that aptitude might even “help” children who acquire an L2.

However, there are two ways in which their results can be viewed critically. Firstly, aptitude runs the risk of circularity. Perhaps the “best” L2 learners improved their aptitude because they had indeed become very proficient in their L2, rather than aptitude aiding them in becoming very proficient. Moreover, it is difficult to assess the causal role aptitude plays in L2 acquisition (whether it be in young or in late learners), when other potential predictor variables are not investigated. Factors such as motivation have often been reported to influence level of proficiency in an L2, and, if such factors are not likewise examined, it is essentially impossible to determine whether they too may have affected the proficiency levels reported in the late L2 learners; or, indeed, whether aptitude and motivation are highly correlated to begin with, and hence it may not be aptitude which is the driving force behind high proficiency levels in a late learned L2, but rather motivation (or something else).

All that said, the findings give compelling reason to further examine the role aptitude plays in predicting individuals’ success in learning languages both during childhood and in adulthood.

Mariana Esmeriz and Esther de Leeuw


Multilingualism and personality

24 Jun

Jean-Marc Dewaele and Li Wei. 2013. ‘Is multilingualism linked to a higher tolerance of ambiguity?Bilingualism: Language and Cognition 16 (1): 231-240

The personality trait ‘tolerance of ambiguity’ (TA) has been described as a tendency to see ambiguous situations as positive or desirable. Ambiguous situations might be ones that are difficult to categorise, that lack familiar cues, or that present a person with complex or contradictory signals. TA has been studied in organisations, businesses, and classrooms, and relates to flexibility, adaptability, learning, and positive experiences in diverse groups and situations.

Surveying over 2000 people, Dewaele and Li Wei found that monolinguals have the least TA by a significant margin, and multilinguals are significantly more ‘tolerant of ambiguity’ than both bilinguals and monolinguals. Knowing more than 3 languages wasn’t associated with greater TA, but greater proficiency did correspond to higher TA scores. Interestingly, even relatively short, intensive acquisition of another language correlated with TA, suggesting psychological benefits of language learning later in life.

How should we interpret this link? Is it greater TA that causes people to seek out new experiences like language learning? Or is it the learning of languages that creates greater TA in a person? The authors acknowledge that further work is needed to establish the causal relations involved, but previous research suggests that the second reading is certainly plausible. Here, I reflect briefly on two important implications.

A distinction is often drawn between ‘elite’ bilingualism (becoming bilingual by choice, e.g. for travel or cultural prestige) and ‘folk’ or ‘minority’ bilingualism (becoming bilingual as a consequence of one’s family or community situation). Elite multilinguals are often admired in the UK, whereas community multilinguals are often viewed as a costly burden. Dewaele and Li Wei’s research stands to bridge this artificial divide. In fact, both groups are likely to develop greater TA, which in turn corresponds to inter-personal flexibility, empathy, ease of cross-cultural communication, and desirable professional and leadership skills.

Dewaele and Li Wei’s work is also important in shifting discourses about multilingualism beyond just economic advantages (in public discourse) and cognitive advantages (in academic discourse), towards interesting questions of personality and psychology. Supporters of English monolingualism in the UK and in the US routinely assert that the ease of speaking English internationally means that teaching other languages is a waste of funds. Dewaele and Li Wei’s findings imply, however, that this reasoning could create nations of subtly less tolerant people, less psychologically able to cope with diversity, difference, and change.

Devyani Sharma

Multilingualism and Genesis

31 May
This is an Anglican Church in Gitwangax with the totem poles in the background.

This is an Anglican Church in Gitwangax (Kitwanga) in northern British Columbia, Canada with traditional totem poles in the background. UNESCO lists the indigenous language spoken here as “severely endangered“.

In Genesis, Sebastião Salgado has photographed the endangered, the threatened and the rare. Looking at his photographs is like looking through windows. Salgado has opened the blinds into a wild, disappearing, and otherwise hidden world. Here, I relate his photographs to language.

Languages carry social meaning, both for the group – or person – who produces the language, as well as for the group – or person – who perceives the language. With language, a culture can distinguish itself from others; a person can stand out as unique from those with whom he or she is conversing; and a passage of time can surface as different from all those which preceded it, and all those which are to follow. Language is intrinsic to our identities: we express and perceive identity through language. Crucially, when a language, or a form of a language, is lost, so too is an identity, or a form of an identity.

Just like the animals and landscapes Salgado photographed, many languages are endangered, threatened and rare. UNESCO has documented that, if nothing is done, half of the over 6000 languages presently spoken will disappear by the end of this century. These are the languages of the people Salgado captured in his photographs, but also those of the many other colonised areas of the world, such as the First Nations languages in northern Canada (see picture of Gitwangax, where Gitxsanimaax is endangered),  the aboriginal languages in Australia, and the indigenous languages of South America. We see these threatened cultures in Salgado’s windows, hovering on the brink of assimilation with our Western world, and we know, ultimately, they are who they are because of language.

And yet – Salgado can’t take a picture of language. Language is this enormous abstract entity, which essentially loses all meaning unless it is represented across the dimension of time. It is impossible to take a “snapshot” of language as Salgado does of a smile, a gesture, a glance. To understand language, to produce language – and all that which is intrinsic to our identities – one must, at least to a certain extent, acquire some sort of knowledge of that language. If language is recorded, and a slice extracted, as a visual moment is through a photograph, it loses its meaning.

Looking through Salgado’s windows, one cannot hear the voices of the people he has photographed. Language, something so all encompassing, so powerful, is silent. From my perspective, a way to give endangered languages a voice, in one’s own life, is to support multilingualism. The process of language loss is, according to UNESCO, “neither inevitable nor irreversible: well-planned and implemented language policies can bolster the ongoing efforts of speaker communities to maintain or revitalize their mother tongues and pass them on to younger generations“. Peter Austin of SOAS has coined the personal concept of a language footprint. To improve one’s own language footprint, one can, for example, learn a new (perhaps even a minority) language, avoid products and activities that give people no choice other than to use dominant languages, support increased language learning in one’s own country, and learn about the world’s diversity of languages.

For me, such small steps may be a personal way to help counter language loss, and therefore, ultimately, all that is lost therewith.

Esther de Leeuw

Multilingualism means business

30 Apr

There are many reasons as to why multilingualism is considered valuable. In terms of language policies (e.g. whether a government decides to endorse multilingualism), more often than not encouraging multilingualism needs to be financially justifiable. Only then will multilingualism be endorsed through policy.

But how does one know whether multilingualism is indeed more profitable than monolingualism? Is it really the case that a multilingual population will increase a country’s economic prosperity? Does a person who speaks more than one language earn more than a person who speaks only one language? These questions are difficult to answer, especially because multilingualism is closely tied to other factors, such as level of education. One way of examining the financial benefits of multilingualism is to ask employers whether they value multilingualism in their companies. If they do, the assumption is that this is most likely because they associate multilingualism with an increase in profit.

The CBI / Pearson education and skills surveys are large scale questionnaires which examine what employers are looking for in their workforce. In the present blog entry, some of the 2012 CBI / Pearson survey’s findings on multilingualism are reported. In brief, the survey suggests that employers consider a multilingual workforce to be beneficial. For example, “[n]early three quarters (72%) of businesses say they value foreign language skills among their employees, particularly in helping build relations with clients, customers and suppliers (39%)” (p. 55). The report also revealed that companies which are export orientated are more likely to value a multilingual workforce than companies which are less export orientated. “[O]ver half of firms in manufacturing (59%) and in engineering, hi-tech/IT and science (55%) see foreign language skills among staff as helpful in building relations with overseas contacts, while a third report foreign language skills as assisting staff mobility within their organisation (33% in manufacturing and 32% in engineering, hi-tech/IT and science)” (p. 56). Indeed, many employers are actively recruiting new staff which are proficient in foreign languages, whilst other employers are addressing the issue by “raising the language skills of their existing staff, using private providers (21%), FE colleges (6%) or in-house language training (15%)” (p. 57). European languages are considered to be the most useful (by those valuing staff with foreign language skills) with German, French and Spanish respectively rated as the most useful foreign languages. Moreover, “[l]anguage skills geared to business in China and the Middle East feature prominently (of those valuing staff with foreign language skills, 25% rate Mandarin as useful, 12% value Cantonese and 19% Arabic). Findings from the survey also reflect the need for language skills that can help to manage workforces in the UK from non-English (e.g. Polish) speaking backgrounds.

The findings from the CBI / Pearson education and skills survey most likely influenced Michael Gove, the Education Secretary, in his recent decision to make foreign language learning compulsory from the age of seven. This policy initiative was largely endorsed throughout the UK, and certainly is by me. Perhaps, however, in addition to this initiative to increase the learning of foreign languages in the UK, it would similarly be wise to harness the extant multilingual resources in the UK through encouraging the maintenance of immigrant languages. For example, under Michael Gove’s plans, primary schools will offer lessons in French, Mandarin and Spanish, amongst other languages. Indeed, these languages are already widely spoken in immigrant communities in the UK. According to the 2011 Census results, 147 000 people in England and Wales already speak French as their main language; 141 000 speak a form of Chinese as their main language; and 120 000 speak Spanish as a main language.

From my vantage point, it seems contradictory to encourage the learning of these languages in schools without encouraging their maintenance in immigrant populations. Moreover, I would predict that – all else being equal – employers value these languages as such in their companies.

Esther de Leeuw

Up Close: 2011 Census Data on Multilingualism

20 Mar
Street logo Brick Lane, London in English and Bengali

Street logo Brick Lane, London in English and Bengali

Since data from the 2011 Census were released, there has been a fair amount of discussion in the press about multilingualism. For example, in one analysis, the BBC accurately reported that of those surveyed in England and Wales, 138000 speak no English. In Languages Mapped, the Guardian displayed this in relation to the total population, showing that this is equivalent to slightly less than 0.3% of the total population. It perhaps goes less reported that in London almost 50000 people can speak no English (that’s 0.6% of the population of London). And that in Tower Hamlets, where Queen Mary, University of London is located, nearly 4000 people reportedly speak no English (that’s 1.6% of the population of Tower Hamlets).

For many people in the UK, it is these absolute numbers which are staggering.

Why is it that some people – whether it be nearly 138000 in England and Wales, 50000 in London, or almost 4000 in Tower Hamlets – speak no English? 

To answer this question, it is necessary to look at immigration patterns in the UK. Globally, the UK is known as what is called a “recipient country”. This means that, for various reasons, political policies presently promote overall movement to the UK from other countries, rather than overall movement from the UK to other countries.

Therefore, understandably, another question which was asked in the Census was country of birth. Here, the Census data tell us that 13% of the population in England was born either in an EU country, like Poland, France, or Spain, or in a so-called “other country”, like Afghanistan, Bangladesh, Canada, and so on. But – and this is worth emphasising – most likely in a country where English was *not* their main language. And, rather amazingly, 37% of London’s population was born in such an EU or “other” country. Even more striking, 42% of the population in Tower Hamlets was born either in the EU or in such an “other country”.

And – this is what the Census data also tell us – many of these immigrants haven’t been here that long. Almost 2% of the total population of England arrived from outside of the UK within the last 2 years; in London, 4.5% of the population arrived within the last 2 years; and, in Tower Hamlets, just over 7% of the population arrived within two years. Here my point is that it’s really not that surprising that, given such a high amount of people who’ve actually just arrived, some have not yet learned English. Think about this just for London: 0.6% of the total population reports to not speak English, and 4.5% of the population of London has arrived in London within the last 2 years. It seems like these new arrivals might actually be doing quite a good job at learning English in those two years.

Finally – the Census data also tell us how old immigrants were when they arrived to the UK. (The Census surveyed all those 3 years and over.) Indeed, the data revealed that 1.7% of the population in England and Wales arrived between the ages of 0 and 4; 3% of the population in London arrived between the ages of 0 and 4; and 4% of the population in Tower Hamlets arrived between the ages of 0 and 4. Some of these little ones most understandably won’t yet speak English because at home they speak the language of their parents. But they will learn English, when they go to school, and start to meet other little boys and girls around the country, in the capital, and in Tower Hamlets.

So, we have some tentative answers to the question: why is it that some people here speak no English? And these answers don’t have to do with motivation or attitudes. They simply have to do with population demographics. Some people don’t speak English because they’ve “just” arrived to the UK, and young children might not yet speak English because they’re too young to have learned it – they are learning the languages their parents speak to them. But they will learn English; it just takes a bit of time.

The Census data therefore suggest that immigrants to the UK are indeed learning English (and there will be another blog related to this soon). But the data also tell us something else. They suggest that immigrants are losing the language of their heritage. Like in so many other recipient countries, children from immigrant families at one point switch from the language of their parents to English. Although approximately 13% of the population in England was most likely born in a country where English was not their main language, just under 9% of the population reported a language other than English as their main language. Similarly, although 37% of London’s population was born abroad, only 22% of the capital’s population list a language other than English as a main language. And although in Tower Hamlets 42% of the population was born abroad, only 35% of its population list a language other than English as their main language. More people come here with a language other than English as their main language than people actually report having a main language other than English. This effect – and it is standard in recipient countries – is referred to as a lack of intergenerational language transmission. We see it in other recipient countries as well, like Canada and the U.S.

I believe that there was consensus in the panel yesterday at Multilingual Capital that, at the very least, all people who’ve moved to the UK should be, if so desired, supported in their wishes to maintain the languages of their heritage. The reasons for this support are vast, and in my view, ultimately amount to human rights, but there are also practical reasons, such as to support the economic prosperity of the UK, and to maintain a more healthy population, as there are links between health and proficiency in more than one language.

For me, these are all reasons to support multilingualism. Specifically in the capital, this can be achieved, for example, through ensuring support for the maintenance and development of heritage languages alongside the parallel acquisition of English – one in no way precludes the other.

Esther de Leeuw

Multilingual Capital: A Resource for London Communities

15 Mar


London is a multilingual capital. The most recent census results show that over 20% of Londoners use a main language other than English, far above the country’s average. Yet bilinguals face many challenges in using their languages.

Devyani Sharma and I are launching a new initiative called Multilingual Capital: A Resource for London Communities on 18 March 2013 at Queen Mary, University of London to support multilingual communities in the capital and beyond.

The primary aims of Multilingual Capital include:

  1. Engaging with London communities so that the rich linguistic diversity of the capital is accurately represented;
  2. Increasing the visibility of multilingualism in the capital so that groups impacted by multilingualism can share ideas and develop a common voice;
  3. Sharing research findings on multilingualism to ensure that relevant knowledge is disseminated to all groups impacted by, and interested in, multilingualism;
  4. Creating knowledge in the emerging research area of multilingualism through documenting language use and diversity in London and beyond.

This event will include a series of short talks by researchers in the field of multilingualism, with plenty of time for a question-answer session.

Find out more about this event here.

Esther de Leeuw

2011 Census results: How well do English speakers speak English?

14 Feb
Main Language?

2011 Census questions on language use (or, my first picture)

For the first time in its history, the 2011 Census posed questions related to multilingualism in England and Wales (see image). However, although this development as such is commendable, it is my view that there is room for these questions to improve in the 2021 Census. In this post, I will discuss an incorrect assumption which underlies the language related questions in the 2011 Census.

The assumption with which I take issue is evident as one progresses from question 18 to 19. Here, the assumption is that there is no variation in proficiency levels within the group of people who mark English as their “main language”, i.e. if English is your “main language”, you will speak it just as well (or just as poorly) as the next person, who likewise claims English to be his or her “main language”, so you can progress to question 20 without filling in question 19. Depending on how prescriptivist your take on language is, you might find this assumption questionable. Surely, it’s not that contentious to claim that some English speakers have a better command of the English language than others (e.g. they know more words, or are able to use those words more elegantly), or, alternatively (if you don’t accept the former), that one’s command of the English language develops over time (e.g. I certainly hope that my English monolingual students (many, but not all of them are monolinguals) have a better command of the English language after studying than before they started their degrees). To take an extreme example, I wouldn’t think twice if someone told me that Noam Chomsky has a better command of the English language than I do, and nor would I think twice if someone told me that my command of the English language is better now than when I was 18 years old (well, I hope they would tell me that – and, at any rate, I wouldn’t mind if someone compared my English with Noam Chomsky’s, however poorly I fair in the comparison). Professor Chomsky and I would probably both put down English as our “main language”, but how well we speak English varies. The point is that even amongst people who claim that English is their “main language”, there is very possibly variation in how well they do so. Moreover, if we draw bilinguals into the equation, who claim that English is their “main language”, but perhaps speak another, or even a few more languages on top of English, there might be even more variation in their proficiency levels in English. Essentially, it would be nice to know how well English speakers claim to speak English.

Furthermore, the questions in the 2011 Census imply that those who have English as their “main language” have a higher proficiency in English than those who don’t claim English to be their “main language”. Currently, “[o]f the eight per cent (4.2 million) of usual residents aged three and over who reported a main language other than English, 41 per cent (1.7 million) could speak English very well, 38 per cent (1.6 million) could speak English well, 17 per cent (726,000) could speak English but not well and the remaining three per cent (138,000) could not speak English at all”. I do wonder how the “main-language-English-speakers” would rate themselves had they been asked to answer this same question. Indeed, if we were asked to rate each other, we might score all in all quite poorly (think Emma Thompson’s review of school girl English). I would probably say “well” for myself (probably not “very well”) and it would depend on a lot of factors such as with whom I compare myself. If it were Noam Chomsky, I might go down to “not well”.

However, and this I want to emphasise, we really don’t know how to interpret the ratings from the “non-English-main-language-speakers” unless we have ratings from the “main-language-English-speakers”. And this we would only get if we asked both groups how well they speak English. I think that this should be done for the 2021 Census. Then we’ll know what those percentages – 41, 38, 17, and 3 per cent – really mean.

Esther de Leeuw

Individual-Level Relationships between Social Capital and Self-Rated Health in a Bilingual Community

31 Jan

Hyyppä, M. T., & Mäki, J. (2001). Individual-Level Relationships between Social Capital and Self-Rated Health in a Bilingual Community. Preventive Medicine, 32(2), 148–155.

Previously, I have written about how bilingualism affects individual health. For example, research from Canada has shown that there is a delay in the onset of the symptoms of dementia in individuals who have spoken more than one language for the majority of their lives; and that in indigenous communities where the indigenous language is maintained, or revived, there is a significantly lower youth suicide rate in comparison to similar Canadian communities where the indigenous language is not maintained. Such studies confirm that, for the individual, learning to speak another language – and in the latter case, the language of one’s heritage – can have far-reaching positive effects on health.

The present study by Hyyppä and Mäki, examined how health and language interact in Finland. For starters, it is important to note that Finland is officially bilingual; most Finns are Finnish native speakers, but there is a small minority of native Swedish speakers (there are also other minority languages in Finland, such as Saami). One of Hyyppä and Mäki’s claims is that variables associated with health would differ across the two language groups even though all the individuals who responded to their large-scale survey (284 Finnish-speaking and 271 Swedish-speaking men and 374 Finnish-speaking and 355 Swedish-speaking women) came from the same bilingual Ostrobothnian municipalities in Finland, i.e. everyone had access to all the same health resources. This is different from other studies in which different language groups were geographically separate and hence might not have been exposed to all the same health resources.

Before embarking on the results of their own large-scale questionnaire, Hyyppä and Mäki provide some rather shocking results from similar studies in Finland, which are worth repeating here (references have been omitted, but are present in their original publication).

“Ever since epidemiological health surveys have been published in Finland, total mortality rates have favoured the Swedish-speaking minority. Significant disparities have been established in the annual suicide rates, violent and accidental death rates, and especially in cardiovascular mortality. The life expectancy of the Swedish-speaking people living in the Aland Islands and in Ostrobothnia ranks among the highest in the whole of Europe. Through 1991-1996, the Swedish-speaking men lived on average 8.7 years longer than their Finnish-speaking compatriots. Swedish-speaking women died at the age of 82.9 years and Finnish-speaking women at the age of 78.1 years. […] [W]e conducted surveys […] and found that the age of disability retirement fell at 48 years among the Swedish-speaking men and at 36 years among the Finnish-speaking men. For women, the corresponding ages were 53 and 45 years.” (p. 149)

To a certain degree, the results from Hyyppä and Mäki’s study corroborate these findings. For example, they found that Finnish speakers were more often unemployed than Swedish speakers; the Finnish population was more likely to drink until drunk than their compatriots; the Swedish speaking women were on average more educated than Finnish speaking women; and the Swedish speaking men were also more likely to participate in community events than the Finnish speaking males. Interestingly, the Finnish speaking population also showed more distrust. (Here, the specific questions were “Generally speaking, would you say most people can be trusted”? and “Do you think most people would try to take advantage of you if they got a chance?”). However, Hyyppä and Mäki continue that in their study good self-rated health, disability, and diagnosed long-term diseases were equally frequent in both language groups, which is perhaps due to the fact that, indeed, all their participants had access to the same health resources given that they were all from the same geographical area.

Nevertheless, to explain the apparent disparities between the two language groups within the same bilingual Ostrobothnian municipalities in Finland, Hyyppä and Mäki propose that the Swedish-speaking group holds more social capital than the Finnish-speaking group. The term social capital originates from the French sociologist Pierre Bourdieu, who put forth the idea that material wealth is only one form of capital; according to Bourdieu, and subsequent linguists, language is a form of capital and people can be at the wealthier ends (here, speak Swedish) or at the poorer ends (here, speak Finnish) of the language capital’s scale.  For Hyyppä and Mäki, social capital refers to “the properties of the individuals who adopt positive feelings toward others and belong to voluntary associations, which have been shown to reduce psychological stress, and therefore may promote good health” (p. 149). Their suggested rather direct relationship between membership in voluntary associations and positive feelings towards the group is perhaps somewhat speculative. However, their more general claim that membership to a specific language group within a bilingual community may have social advantages over membership to the other language group – and that this membership can impact health – indeed appears to be validated by both the results from their own and colleague’s research.

That speaking one language over another can be socially advantageous is not really surprising. Some dialects are more “respected” than others, and likewise, some languages give rise to greater opportunities than others. Hyyppä and Mäki’s research is novel because it suggests that via social capital the use of one language over another links to the health of the speakers. To gather more information on this, it might be worthwhile to investigate bilingual families. What do the health prospects look like for Finnish-Swedish bilingual children in Finland? Do they diminish cross-generationally if Finnish becomes the dominant family language? It would also be worthwhile to investigate other bilingual communities (I’m thinking of cities which are officially bilingual, such as Montreal and Brussels, as well as cities which have high immigrant populations, such as London and Lyon). Here, it would be reasonable to assume that as the immigrant group acquires the language of the socially advantaged group, the health of the immigrant population might also improve. However, such speculations would need to be viewed in light of other research, such as that of Hallett, Chandler, and Lalonde, which suggests that maintenance of the heritage language improves health (i.e. lowers youth suicide rates in indigenous populations).

Perhaps it is both – acquiring the language of the majority alongside maintenance of the language of one’s own identity (i.e. bilingualism) – which is most beneficial to an individual’s health.

Esther de Leeuw

Preserved Implicit Knowledge of a Forgotten Childhood Language

31 Dec

Bowers, J. S., Mattys, S. L., & Gage, S. H. (2009). Preserved Implicit Knowledge of a Forgotten Childhood Language. Psychological Science, 20(9), 1064–1069.

Parents who want their children to become highly proficient bilinguals might, unfortunately, view their children’s potential lack of what they consider to be perfect proficiency in one of the languages as a failure on their own account. For example, I’ve had conversations with parents who worry that they didn’t expose their children to enough of one of the languages, or weren’t strict enough in their expectations of which language should be spoken back to them by their children. Alternatively, the possibility alone that both their children’s languages might not become as proficient as fellow monolingual children’s languages might serve as reason to decide against supporting a bilingual upbringing.

To a large extent, this line of thought is grounded in the comparison of bilinguals with monolinguals (i.e. Is my bilingual child performing in both of his or her languages as two separate monolingual children would in those same languages?). Instead, there are many reasons as to why – instead of comparing bilinguals with monolinguals – bilingual children should be compared with bilingual children.  As such, children raised with two (or more) languages are not thought to be inadequate when compared to monolinguals indeed because the bilingual child does have an additional language (taken to an extreme, such a line of thought would actually consider monolingualism to be a form of “inadequacy” – if using this word at all – due to the complete lack of an additional language). This idea (that bilinguals serve as their own yardstick) is sometimes referred to as the “bilingual mindset”. So – it doesn’t really matter how proficient the bilingual child is in language A and language B because the bilingual child indeed has language A and language B, which, aside from the additional language, itself comes with many advantages (see previous blogs on e.g. dementia, bilingualism in Canada, and bilingualism in aboriginal communities).

But what if a child acquires only very little of a language in the early years (or potentially only hears this language), and thereafter contact with this early language is severed? This might be the case, for example, of children who are adopted by parents who speak a language other than the country where they were born; and of children who have caretakers who speak a language they do not speak, but, after a certain young age, the caretaker and family go their separate ways. Such early exposure might also occur in the case of divorce, if, for example, one parent spoke a particular language with the child early on, but then contact with that parent unfortunately stopped after separation. In my own case, I was exposed to Dutch through my Oma (grandmother) and Opa (grandfather) until the age of 4, because we lived very close to them, but when we moved to another town in Canada, far away from them, my exposure to Dutch dropped dramatically. Comparable situations to my own might occur in aboriginal communities where the grandparents speak with the grandchildren in the aboriginal language, but when the grandparents pass away, the grandchildren no longer receive input in that language because – for various reasons – the parents don’t speak with the children in the aboriginal language to the same extent that the grandparents did. Here, in terms of bilingualism, the question is whether this early exposure has lasting benefits in terms of acquiring that language later on. Or, put differently, do individuals who were exposed to a language in their early years have an advantage over those who were not exposed to this same language – when both groups attempt to acquire that language later on in adulthood?

Bowers, Mattys, and Gage (2009) investigated this very question. They examined whether native English speakers living in the United Kingdom, “who were exposed to either Hindi or Zulu as children due to their parents’ work abroad” (p. 1064) could learn to hear differences between sound contrasts which were unique to those languages faster than fellow native English speakers living in the United Kingdom who had never been exposed to those languages in early childhood. Crucially, they also wanted to know whether those who had been exposed to Hindi could acquire the sound contrasts of Hindi faster than those who had been exposed to Zulu, and vice versa. As such, they wanted to find out whether exposure to a particular language in early childhood aids (re?-) learning of that specific language in adulthood, rather than exposure to a particular language in early childhood aiding in the learning of any language in adulthood.

In their study, they tested all of the participants (the completely monolingual Brits, who had never been exposed to any language other than English; the Brits who were exposed to Hindi early on; and the Brits who were exposed to Zulu early on) to see whether they could hear sound contrasts of those languages. In this initial test, all of the participants scored equally poorly, i.e. they could all not hear the “foreign” sound contrasts. Thereafter, all the participants underwent sound contrast training, i.e. they were asked to determine whether the two sounds they heard were the same or different and they “were given feedback after each response (correct vs. incorrect)” (p.1066) over a series of training sessions which lasted approximately 30 days. Again, the idea was to determine whether or not the training which they underwent would help e.g. those who were exposed to Hindi early on to acquire the Hindi sound contrasts faster than the complete monolinguals and faster than those who were exposed to Zulu early on.

Their results showed that although the early exposure individuals “showed no preserved knowledge of their childhood language on initial testing, after practice, a subset of them (participants under the age of 40) regained sensitivity to a [sound] contrast from their childhood language. By contrast, when [this contrast] was unknown in childhood, no or minimal learning was observed after extensive practice for both young and old participants” (p. 1066).  They summarized that their findings “provide clear evidence of preserved implicit knowledge of a forgotten childhood language” (p. 1066); but, it appears, only if that language is re-acquired before the age of 40. After the age of 40, they suggest that disuse for such an extended period of time may indeed cause complete forgetting.

Accordingly, it seems that, even if children who are raised in an environment with more than one language do not acquire both languages completely and fully (in comparison to how monolingual children would have acquired those languages), their perhaps even very minimal acquisition (through e.g. simply early exposure to that language) puts them at an advantage in acquiring that same language later on in life, if they are so inclined to do so.

In other words, parents who decide to create an environment with more than one language for their child, might consider this itself to be a success, regardless of how “perfectly” the child learns the language. This is akin to viewing the glass half full, rather than half empty.

Even a little bit of water in the glass is something which has lasting benefits.

Esther de Leeuw

The effects of contact on native language pronunciation in an L2 migrant setting

31 Oct

De Leeuw, E., Schmid, M. S., & Mennen, I. (2010). The effects of contact on native language pronunciation in an L2 migrant setting. Bilingualism: Language and Cognition, 13 (Special Issue 01), 33–40.

De Leeuw, E., Schmid, M., & Mennen, I. (2007). Global foreign accent in native German speech. In J. Trouvain and W. Barry (eds.) Proceedings of the 16th International Congress of Phonetic Sciences, 1605–1608.

This research addressed the question of whether it is possible for a native language (L1) to become foreign accented when a new language (L2) is acquired in adulthood.

More specifically, de Leeuw, Schmid, and Mennen (2010) recorded 57 German L1 speakers (who grew up in a monolingual German environment), and in late adolescence or adulthood moved abroad to either Canada or the Netherlands where they acquired respectively English or Dutch to a high level of proficiency. These bilinguals (the speakers) moved abroad at an average age of 27 and had resided in their country of choice for an average of 37 years. The speakers were recorded in their German L1 describing a silent Charlie Chaplin film they had just seen and, thereafter, the edited recordings were played to another group of Germans in Germany (the listeners), who had a limited knowledge of other languages. Only grammatically correct utterances were included in the recordings because the researchers wanted the listeners, who rated the bilinguals on how native-like they were in their German, to base their ratings solely on their actual pronunciation, and not on any potential grammatical oddities.

Strikingly, the results from the study revealed that the German listeners were significantly more likely to perceive a foreign accent in the German speech of the bilinguals in Canada and the Netherlands than in the speech of the control group (made up of German native speakers with a limited knowledge of other languages, who were recorded in Germany performing the same task as the bilinguals). Indeed, 14 of the German native speakers residing in either Canada or the Netherlands were clearly perceived to be non-native speakers of their native German, more-or-less across the board by all listeners.

In another analysis of these same bilinguals (de Leeuw, Schmid, & Mennen, 2007), results indicated that contact with the native German language had a more significant effect on predicting foreign accented native speech than age of arrival or length of residence in the recipient country. In particular, for both English and Dutch L2 groups, foreign accent was more likely to be perceived in the native speech of the bilinguals who had less contact with their native German language than in those who had more contact, although this effect was clearer in the bilinguals who moved abroad after 22 years of age.

In many ways, this research challenges our understanding of what native speech actually is. On the one hand, it seems that native speech is malleable across the lifespan – and not “anchored in the brain” with no potential for it to change once acquired up to adulthood. On the other hand, the results raise questions with regard to who defines a native speaker – is it the speaker of the language? Or is it the listener of the language? Or, are other parameters involved in determining native speech, and if so, what are they?

These are questions which I will attempt to answer in future blog entries.

Esther de Leeuw