Abstract
The research reports on a study investigating the cognitive benefits of bilingualism in children who speak the minority languages of Sardinian and Scottish Gaelic, in addition to their respective ‘national’ languages of Italian and English. One hundred and twenty-one children, both bilingual and monolingual, were administered a series of standardised cognitive ability tests targeted at the four areas that have been previously shown to be advantageous to bilingual children in the literature, namely, cognitive control, problem-solving ability, metalinguistic awareness and working memory. The bilingual children significantly outperformed the monolingual children in two of the four sub-tests, and the Scottish children significantly outperformed the Sardinian children in one of the sub-tests. The differences found were largely due to the superior performance of the Scottish bilingual children who receive a formal bilingual education, in contrast to the Sardinian bilingual children who mostly only speak the minority language at home. The implications of the results are discussed.
Introduction
The specific benefits of bilingualism in cognitive and language development are still being explored. Since the 1970s and 1980s, the early research into bilingualism (e.g. Barke & Perry-William, 1938; Carrow, 1957; Harris, 1948) has been much criticised for its methodology and dubious interpretation of results (for an extensive critique, see Cummins, 1976; Hakuta & Diaz, 1985; Lee, 1996). These early studies outlined the potential harm there was in raising children as bilingual (Darcy, 1953; Macnamara, 1966). Instead, more recent research evidence indicates that there are many possible benefits from being bilingual, often grouped into the following four areas:
Inhibitory control and selective attention (cognitive control)
Bialystok (2001) outlined the advantage that bilingual children have in enhanced inhibitory control for ignoring perceptual information, or in other words, that they are better than their monolingual peers at selectively attending to important information and ignoring misleading cues (known as cognitive control). Bialystok and Martin (2004) describe a process whereby bilingual children are continually sorting out extra perceptual information. For every object or action, bilingual children assign two words or labels: one in each language that they speak. Depending on the context they find themselves in, they have to develop the ability to be able to choose the appropriate label. Thus, bilingual children nurture their capacity to focus on relevant and appropriate information and inhibit their attention on the information unrelated to the context.
The process of inhibition has been shown in research into adult cognition and neuropsychology to be ascribed to the prefrontal lobe and thus has a key role in general cognitive functioning (Hasher & Zacks, 1988; Kimberg, D’Esposito, & Farah, 1997). Moreover, inhibition has been shown in research in young children to be a key factor in the ability to perform a number of different cognitive tasks (Dagenbach & Carr, 1994; Dempster, 1992; Diamond & Taylor, 1996).
Other research studies have demonstrated the usefulness of this key cognitive process in bilingual children when compared with monolingual peers. For example, Bialystok (1999) demonstrated that bilingual children were more able to perform a card-sorting task involving various dimensional changes. This research was followed up by Bialystok and Martin (2004), which excluded the possibility that the differences found in the original study were due to the bilingual children’s superior representational abilities, as indeed no differences in such abilities were found. Instead, the authors concluded that the bilingual children’s superiority in the test was due to their enhanced inhibitory control.
Bialystok and Majumder (1998) demonstrated bilingual children’s ability to perform two specific cognitive tasks (the Block Design sub-test of the Wechsler Intelligence Scale for Children (WISC) (Wechsler, 1974) and the Water Level task (Pascual-Leone, 1969), both of which require high levels of inhibitory control. The bilingual children scored significantly higher than the monolingual children on both tasks and thus provided evidence of the bilingual children’s enhanced ability to perform non-verbal tasks.
More recent research has provided further evidence of the inhibitory skills of bilinguals over monolinguals (Treccani, Argyri, Sorace, & Della Sala, 2009). In an elegantly designed study, involving a task where the inhibitory skills would prove disadvantageous, Treccani et al. demonstrated that the bilinguals’ efficiency in inhibiting distracting information resulted in specific non-linguistic benefits of bilingualism that are not always reflected in cognitive advantages. They conclude that their results ‘represent evidence in support of a substantial effect of bilingualism on executive control processing’ (p. 326). However, they caution that these advantages will only be established depending on the characteristics of the task being undertaken.
Bialystok and Codd (1997) investigated bilingual and monolingual children’s development of cardinality and found that the bilingual children were more advanced in performing the number tasks, especially on those that required high levels of selective attention. Moreover, research by Clarkson (2006) has shown that the specific cognitive benefits of bilingualism can extend to ability in mathematics. He found that English–Vietnamese bilinguals outperformed monolinguals in a series of tests assessing mathematical ability. He concluded that it was not only the metalinguistic skills of the bilingual children that enabled them to perform better, namely, their ability to self-correct, but also that they were ‘perhaps more confident in their approach to solving difficult problems’ (p. 191). Such a statement bridges two other potential benefits of bilingualism: problem-solving and metalinguistic awareness. It is to these that we now turn.
Problem solving and creative thinking
One of the earliest studies to indicate the possible benefits of bilingualism on problem solving and creativity was conducted by Peal and Lambert (1962). They found that bilingual children performed significantly better on both verbal and non-verbal measures, especially on those tasks that required mental or symbolic flexibility and concept formation. The authors argued that the key aspect of their research was that the bilingual children involved in the study were ‘balanced bilinguals’, that is, they were able to speak both languages (French and English) to a similar degree of proficiency. A consistent criticism levied against previous research that demonstrated the disadvantages of bilingualism (that is, the early studies referred to above) was that the bilingual children were not suitably proficient in both languages and that the tests conducted by the researchers were in the least proficient language.
Ricciardelli (1992) found that Italian–English bilinguals equally proficient in both languages (balanced bilinguals) outperformed monolinguals on several tests of creativity. Interestingly, Ricciardelli found that there were no differences in the tests of creativity when monolinguals were compared with those bilinguals who were proficient in only one language (partial bilinguals). Hakuta and Diaz (1985), in a study involving more than 300 Spanish–English bilingual children, found that non-verbal intelligence scores were positively related to the degree of bilingualism, that is, the more ‘balanced’ bilinguals (more proficient in both languages) scored higher in the non-verbal intelligence tasks. Diaz (1985) specifically reported the performance of Spanish–English bilingual children’s performance on a task requiring them to demonstrate their ability to reason by analogy. In this longitudinal study, Diaz found that those children with stronger proficiency in both languages demonstrated stronger analogical reasoning ability.
A significant contribution to the debate on the problem-solving and creativity skills of bilinguals was recently undertaken by Kharkhurin (2010). He administered two different kinds of problem-solving tasks to 150 participants (103 Russian–English bilinguals and 47 English monolinguals). The two tasks involved creative thinking but one of them required the use of verbal skills, while the other required non-verbal skills. Kharkhurin found that the bilinguals significantly outperformed the monolingual participants on the task requiring non-verbal skills, while the reverse was true for the task involving verbal abilities (i.e. the monolinguals significantly outperformed the bilinguals). Kharkhurin (2010) concluded that the results offer important implications for the understanding of the domain specificity of the problem-solving and creative-thinking abilities of bilinguals, namely, that the advantages they hold appear to be restricted to those tasks requiring non-verbal skills.
Metalinguistic awareness
Metalinguistic awareness, that is, the ability to use knowledge about language, has been shown to be better among bilingual children. Various studies have investigated the ability of bilinguals over monolinguals in different aspects of metalinguistic awareness, namely, word awareness (Bialystok, 1988; Edwards & Christopherson, 1988), syntactic awareness (Galombos & Hakuta, 1988) and phonological awareness (Campbell & Sais, 1995). However, the results have not always been conclusive (e.g. see Bialystok, Majumder, & Martin, 2003) and may depend on the degree of similarity between the two languages spoken in terms of assisting the transferability of the phonological skills. Jessner (2008), however, clearly delineates the advantages of bilingualism and multilingualism on metalinguistic awareness, which she argues includes the enhancement of highly developed skills in pragmatic competence, communicative sensitivity and flexibility as well as skills in translation and the acquisition of further languages.
It is important to highlight here that there is much overlap between metalinguistic ability and cognitive control. In fact, Bialystok and Codd (1997) cite ‘control of processing’ (p. 85) as an important metalinguistic skill. It has also been argued that metalinguistic awareness overlaps with skills that develop ‘at the higher level of creativity and reorganisation of information’ (Jessner, 2008, p. 277). It is clear, therefore, that the potential benefits of bilingualism cited in the literature have common characteristics.
Working memory
While it could be argued that the advantages of bilingualism in cognitive control and problem-solving ability are fairly well established, there is much less known about the relationship between bilingualism and memory (Kormi-Nouri, Moniri, & Nilsson, 2003). Research has demonstrated that the capacity for digit span (as a measure of working memory) differs according to the language spoken (Hoosain, 1984; Naveh-Benjamin & Ayres, 1986; Stigler, Lee, & Stevenson, 1986). Thus, bilinguals may perform better at recall of digits in one language rather than another. One explanation for this hypothesis has been that the speech or articulation time for digits is shorter in some languages than others, thus making fewer demands on the brain’s capacity for working memory. However, Da Costa Pinto (1991), in his research of Portuguese–English bilinguals, did not find this relationship between articulation time and digit span (the children had a higher digit span when performing the task in Portuguese, even though the Portuguese numbers have more syllables and phonemes than the same numbers in English). Da Costa Pinto instead suggested that there was a ‘familiarity effect’, such that the bilingual child’s higher digit-span performance will be in the ‘mother tongue’ of the bilingual, that is, the language in which the bilingual child is most proficient.
There remains some uncertainty, however. Chincotta and Hoosain (1995) found that Spanish–English bilinguals, who were more proficient (self-declared) in Spanish, obtained a higher digit-span performance in English. They highlighted the possibility that the bilingual children in Da Costa Pinto’s study may not have been ‘balanced’ bilinguals. In other words, the children used in the Da Costa Pinto study may have been significantly more proficient in Portuguese than English, and therefore, it was difficult to make a fair comparison of their performance in the two languages. Research into the verbal working memory of bilingual children has shown that there were no significant differences in the performance of bilingual (Spanish and English) and monolingual (English) children on two measures of verbal working memory, specifically, language processing and comprehension (Gutierrez-Clellan & Weismer, 2004). In contrast, a series of research studies by Bialystok and colleagues (Bialystok, 2006; Bialystok et al., 2005; Bialystok, Craik, Klein, & Viswanathan, 2004) have demonstrated that bilingual children perform significantly better on a task assessing working memory (the Simon task, in which subjects have to respond to a stimulus flashed onto a computer screen, e.g. a coloured square, by pressing the correct response key). Similarly, the positive effects of bilingualism on episodic and semantic memory were found in a study involving 120 children (60 bilingual and 60 monolingual) on tests of verbal tasks and word fluency tests involving both free recall and cued recall. The authors conclude that the bilingual children’s superiority is due to their ability to ‘integrate and/or organise the information of the two languages’ (Kormi-Nouri et al., 2003, p. 47).
Bilingualism and minority languages
The research findings presented thus far suggest that it is crucial that in order to gain from the advantages of bilingualism, the speaker should be equally proficient in both languages. It is noteworthy, moreover, that most of the research conducted has been with children who are proficient in two well-recognised international languages (e.g. French, English, Spanish, Italian, Russian, Portuguese). There has been much less research on the cognitive benefits of bilingualism where one of the second languages is a so-called ‘minority’ language (e.g. Catalan, Basque, Gaelic, Sardinian). It is in this area that the current research study was conducted.
In recent years, there has been a cultural and political willingness to revitalise minority languages since they have been in danger of becoming obsolete (e.g. see the European Charter for Regional or Minority Languages). It was the objective of the current research to explore whether the same cognitive benefits that have been shown to exist in bilingual children of two ‘widely recognised’ languages would also exist when one of the two languages spoken is a ‘minority’ language.
The research was conducted in Sardinia (Italy) and Scotland, contexts that are not too dissimilar in nature. Both have devolved government. Both have surviving cultural traditions that are extremely similar in nature (especially with regard to music, the launeddas and the bagpipes, and dress, the traditional Sardinian costume and the kilt), and both have a surviving minority language spoken by an increasingly small proportion of the respective populations (Sardinian in Sardinia; Gaelic in Scotland). Moreover, both regional governments have recently introduced legislation to promote and help preserve their respective minority languages (the Sardinian Regional Law of 15 October 1997, n.26 Promozione e valorizzazione della cultura e della lingua della Sardegna, which translates as ‘The promotion and development of the Sardinian culture and language’ and the Gaelic Language (Scotland) Act, 2005, which gave effect to the principle that the Gaelic and English languages should be accorded equal respect). It seemed reasonable and logical to conduct research in these two locations.
The research was intended to investigate whether the same cognitive benefits found in bilingual children who speak two internationally recognised languages could also be extended to children who speak minority languages (Sardinian and Gaelic). Given the situation in Sardinia and Scotland, where there is a determined focus on the revitalisation of these minority languages, it seems essential to cast further light on the possibility that being bilingual provides cognitive advantages, or otherwise, especially given the respective governments’ positions on the promotion of these minority languages. Specifically, the aim of the research was to administer four cognitive ability tests that would reflect the four areas currently highlighted in the literature as being advantageous to bilinguals, namely, those areas discussed above: cognitive control, problem solving, metalinguistic awareness and working memory. The cognitive ability tests chosen will be described in more detail below.
Method
The research took place in Dorgalì, Sardinia, and Stornoway, Scotland. A school in each setting was chosen that had sufficient numbers of bilingual (who spoke either Italian and Sardinian or English and Gaelic) and monolingual children (who spoke either Italian only or English only). The data collection took place in the period November 2009–February 2010.
Participants
One hundred and twenty-one children were involved in the research: 62 bilingual children (32 in Sardinia and 30 in Scotland) and 59 monolingual children (29 in Sardinia and 30 in Scotland). In Scotland, the bilingual children were classified as bilingual as they were undergoing Gaelic-medium education and spoke Gaelic and English fluently. In Sardinia, the bilingual children were classified as bilingual by Sardinian-speaking teaching staff according to whether they were fluent in Sardinian in school and if they spoke Sardinian at home. In both settings, the bilingualism of the children was verified by asking parents in the consent form whether they considered their child to be fluent in the minority language, and whether they spoke the minority language at home in an everyday context. For those parents who replied negatively to this question, their children were eliminated from the research.
There were no significant differences in the sex of the children involved in the research, according to their assigned group. This is shown in Table 1.
Sub-division of children according to sex.
There was a significant difference in the average age of the participants in Scotland, namely, that the bilingual children were younger than the monolingual children (t-test; t = 3.81, df =58, p < 0.001). In Sardinia, there were no differences in the age of the participants. The means and standard deviations (SDs) of the ages can be viewed in Table 2.
Means and standard deviations of age according to group.
Socio-economic status was controlled by asking parents to provide information on their occupation. An adapted version of Thompson and Hickey’s (2005) classification system was used, with four categories: (a) professional and managerial (examples as written by parents: teacher, businessman, lecturer, engineer and IT management), (b) skilled manual work (mechanic, chef, joiner, butcher and electrician), (c) semi and unskilled manual work (sales assistant, labourer, delivery driver, waitress and childcare worker) and (d) unemployed. Statistical analysis found no significant differences between the bilingual and monolingual children’s parental occupational group.
Procedure
Four sub-tests of the WISC (version III in Italy, version IV in the UK) (Wechsler, 2006a, 2006b) were used, each designed to reflect the four areas already highlighted in the literature as being of potential advantage to bilingual children. The WISC is a standardised test, with established norms, which means that raw scores are transformed into scaled scores, thus comparisons across age groups can be made. The average scaled score for a child of any age is 10, with an average range of 8–12. In other words, if a child scores in the range of 8–12, then they have performed within the average expected for their age. For the purposes of the current research, the use of the WISC was ideal, as comparisons could be made across groups, even when there was a difference in age (as was found for the Scottish children – see in Table 1) and also across countries (i.e. between UK and Italian children).
The four sub-tests of the WISC were as follows: (a) Block Design: in this sub-test, children are shown patterns that they have to reproduce through the manipulation of several coloured blocks. It is a test of non-verbal, problem-solving ability and requires cognitive control; (b) Digit Span: children have to repeat orally a series of numbers, first in the order presented (Digit Span Forwards) and then in the reverse order (Digit Span Backwards). It is a test of working memory; (c) Vocabulary: children have to provide clear definitions of words provided to them by the assessor. Points are awarded depending on the clarity and detail provided. It is a test of verbal ability and word awareness, representing metalinguistic skills and (d) Arithmetic: children have to resolve mentally a series of arithmetical problems, increasing in difficulty as the test proceeds. It is a test of mathematical problem solving and working memory, with some problems requiring selective attention (cognitive control). The four sub-tests were chosen to represent the spread of cognitive skills as outlined in the literature on the cognitive benefits of bilingualism.
Results
Analysis of the four sub-tests of the WISC (cognitive ability)
The results from the four sub-tests of the WISC are shown in Table 3, where the means and SDs (scaled scores) of each of the groups can be read. One can see that the bilingual children (that is, when the bilingual children from both countries are considered together) have performed better than the monolingual children in each of the four sub-tests. These differences are highlighted in bold for ease of reference.
Means and standard deviations of scaled scores in WISC sub-tests.
WISC: Wechsler Intelligence Scale for Children.
To explore whether these differences represent statistically significant differences, inferential analysis was undertaken using a multivariate analysis of variance (MANOVA). The MANOVA was performed on the data using the two factors ‘bilingual group’ and ‘country group’ and the four dependent variables ‘Blocks’, ‘Digits’, ‘Vocabulary’ and ‘Arithmetic’, representing the four sub-tests of the WISC. The results from this statistical analysis indicated that the factor of ‘bilingual group’ was statistically significant for the Block Design (F = 4.57, df = 1, 117, p = 0.035, η2 = 0.037) and Vocabulary (F = 4.19, df = 1, 117, p = 0.043, η2 = 0.033) sub-tests. The factor of ‘country’ was statistically significant for the Arithmetic sub-test (F = 11.8, df = 1, 117, p = 0.001, η2 = 0.088). The effect sizes through use of the η2 values represent medium effects according to Cohen (1988). No other significant differences were indicated.
Follow-up tests were conducted in order to specify where the group differences lay. Post hoc multiple comparisons (using the method of least-significant difference pairwise comparison) revealed that the Scottish bilingual children significantly outperformed the other three groups on both the Vocabulary and Arithmetic sub-tests of the WISC and did significantly better than the two monolingual groups (but not the Sardinian bilinguals) on the Blocks sub-test. The list of significant differences identified by the post hoc multiple comparisons can be found in Table 4. There were no other significant differences found other than those reported in Table 4.
Identification of significant differences using post hoc multiple comparisons.
In summary, the results indicated that the bilingual children scored significantly higher than the monolingual children in the Block Design and Vocabulary sub-tests and also that the Scottish children scored significantly higher than the Sardinian children in the Arithmetic sub-test. Follow-up multiple comparisons revealed that the differences were largely due to the performance of the Scottish bilingual children, who significantly outperformed the other groups in three of the four sub-tests administered.
Discussion
The results highlighted that the bilingual children (who speak Italian and Sardinian or English and Gaelic) scored significantly higher than the monolingual children (who spoke Italian only or English only) in two of the four cognitive ability sub-tests (Block Design and Vocabulary). In the other two sub-tests, the bilingual children scored higher but not to a level that was statistically significant (Digit Span and Arithmetic). However, in the Arithmetic sub-test, the Scottish children significantly outperformed the Sardinian children, and moreover, this difference was considerably greater for the Scottish bilingual children, who scored highest out of the four groups. Indeed, the Scottish bilingual children scored highest of the four groups in three of the four sub-tests (in all but Digit Span) and post hoc multiple comparisons revealed that they significantly outperformed the other groups in these three sub-tests (aside from the Blocks sub-test where they significantly outperformed the two monolingual groups but not the Sardinian bilinguals).
The results represent clear evidence regarding the cognitive advantages of being bilingual and reinforce previous research that has outlined the potential cognitive benefits to speaking two languages. An original aspect of the current research, however, is that the bilingual children were speakers of a minority language (either Sardinian or Gaelic) as well as a well-known and established international language (Italian and English respectively). While it is clear that the bilingual children as a whole group outperformed the monolingual children, a supplementary question remains: why did the Scottish bilingual children significantly outperform the Sardinian bilingual children?
It could be argued that the Scottish children’s level of bilingualism was stronger than the Sardinian children, in that they were benefiting from receiving formal education in the minority language, a provision that is unavailable to children in Sardinia. Sardinian as a language is predominantly transmitted orally, with very little communication done in the written form. This is mainly due to the diversity of forms of Sardinian that are used throughout the island: it is known that Sardinian speakers from two different parts of the island do not always understand each other when communicating in Sardinian and therefore often have to resort to Italian in order to be understood (see Virdis, 1988). Clearly, such a position will not help the diffusion of both the oral and written form of the language, and in fact, the Sardinian language is mostly transmitted in private and familial settings.
In contrast, Scottish Gaelic has a formal and legal status in Scotland and its written form is well established. Moreover, it is still taught both orally and in written form, albeit in a minority of schools. If we consider Cummins’ (1976) claim that there is a critical threshold of L1 and L2 ability that must be attained before the positive effects of bilingualism can be observed, it may be the case that the Sardinian children, while considered by their parents and teachers to be able to speak the Sardinian language fluently, did not have a similar level of command of the Sardinian language in both oral and written form, when compared with Italian. Thus, in Peal and Lambert’s (1962) words, they were perhaps not ‘balanced bilinguals’ (i.e. where the language proficiency of the two languages is roughly equal), or certainly not as ‘balanced’ as the Scottish children, therefore resulting in the differences found. Peal and Lambert outlined the importance of being a ‘balanced bilingual’ in order to demonstrate the cognitive benefits from speaking two languages.
However, it remains true that the Sardinian bilingual children’s scores did contribute to the bilingual factor being found as significant in the statistical analysis. The scores of the Sardinian bilingual children in the sub-tests largely supported the differences found when the Scottish bilingual children’s scores were considered. They scored higher than the Scottish monolingual children in three of the four sub-tests, albeit that this difference was not at a statistically significant level. In other words, it is noteworthy that the Sardinian bilingual children’s performance did not eliminate the superior scores of the Scottish bilingual children but instead extended the differences between the bilinguals and monolinguals further.
The lack of significant differences in the working memory sub-test (Digit Span), while on the surface does not reflect the results of the differences found in the other sub-tests, does in fact mirror similar conclusions from a recent study by Bialystok, Craik, and Luk (2008), involving 96 adults, grouped as ‘younger’ and ‘older’. They found that there was little difference in the performance of bilinguals and monolinguals on two tests of working memory, although the ‘younger’ participants did perform better than the ‘older’ group. It could be argued that the Arithmetic sub-test did require a degree of working memory in order to resolve the more difficult problems presented, however, the factor of ‘country’ was significant in the analysis of this sub-test and not ‘bilingual group’. Reflecting on the present results, we can only support Bialystok et al.’s conclusion that ‘The more general question of whether bilingualism affects working memory performance in any task or situation remains somewhat open’ (p. 870).
The results also appear to support the recent research by Kharkhurin (2010), who argued that bilinguals perform better on problem-solving tasks that require non-verbal skills. The Block Design sub-test from the WISC is a test of non-verbal reasoning, and the bilinguals significantly outperformed the monolinguals on this task. However, the significant differences found in the children’s performance on the Vocabulary sub-test (where again the bilinguals significantly outperformed the monolinguals) would appear to contradict Kharkhurin’s conclusions that ‘bilingualism impedes verbal performance’ (p. 223). Nevertheless, we must highlight that the Vocabulary sub-test, while predominantly a test of verbal ability, also draws upon the ability of the children to provide sufficient information in their explanations of the objects and concepts provided. It may be that the bilinguals’ enhanced cognitive control and metalinguistic skills could have played an important role in their superior performance.
Despite the research evidence presented, some methodological considerations should be made. There is the question of causality: does bilingualism enhance cognitive development or is it that more intellectually gifted children become higher-proficient bilinguals? Similarly, Hakuta, Ferdman, and Diaz (1986) raised the issue that bilingual and monolingual groups may not always be comparable due to the difficulties there are in true random assignment. In other words, any differences found between bilinguals and monolinguals may not be due solely to the factor of ‘bilingualism’, but instead, one must consider the variations there will be in socio-economic status, cultural and educational background (including the value placed upon education by the parents of the children) and ethnic variables, all of which may contribute to differences in the performance of children on cognitive ability tests.
The current research aimed to counter the issue of socio-economic status in two ways: (a) by asking parents to communicate their respective professions in the consent form and conducting analysis on the data using a standard internationally renowned classification system and (b) by conducting the research in the same school, thus aiming to combat the issue of the bilingual children coming from a more affluent area than the monolinguals. There were no ethnic variables to consider: all Scottish children were of the same ethnic group, as were the Sardinian children. The bilingual status of the children involved in the study was not related to ethnicity. Thus, while it is recognised that one can never be absolutely certain that the groups were matched perfectly on the factor of socio-economic status and the other variables noted above, it is argued here that a considered attempt was made to ensure that the comparative groups were as closely matched as possible.
Another counter-argument regarding the emerging differences between the bilingual and monolingual children might be that the advantages found in the bilingual children are only temporary and that the monolingual children will ‘catch up’ their bilingual peers within the next year (Bialystok et al., 2005), therefore rendering the benefits rather meaningless. While this may indeed be true, it is still likely that the benefits acquired by the bilingual children, for example, cognitive control, could transfer to other areas of cognition or be developed further: ‘the processes responsible for this advantage with [bilingual] children may be more permanently enhanced, making problem solving that requires control over attention easier for bilinguals at all stages of development’ (Bialystok et al., 2005, p. 106). Moreover, the linguistic benefits for bilingual children for learning further languages cannot be discounted. However, it is acknowledged that further research is required in this area.
Conclusion
The results in the current study provide supportive evidence that there are clear cognitive advantages of being bilingual: in the current research, specifically, the situation where one speaks a minority language as well as an internationally recognised language. Moreover, the results would seem to support the conclusion from Peal and Lambert’s seminal study, which outlined the importance of having ‘balanced’ bilinguals in order to demonstrate any cognitive benefits from speaking two languages (i.e. that the language proficiency of bilinguals has to be roughly equal), or indeed, Cummins’ (1976) claims that there exists a critical threshold of L1 and L2 ability that must be achieved before the positive effects of bilingualism can be perceived.
However, the results presented here can also lead to the conclusion that, even in cases where this threshold of equal proficiency is not achieved (as in the Sardinian sample), the development of two languages, including situations where one is a minority language, is certainly not a disadvantage with regard to the cognitive development of the child. We discovered that the Sardinian bilinguals scored higher, albeit not at a level of statistical significance, than the Sardinian monolinguals. Such a result is significant when one considers the historical perspective of bilingualism bringing potential harm to a child’s development; a view that, while outdated, unfortunately does still exist in some parts of Scotland and Sardinia today and was probably influenced by the early, flawed research into bilingualism, where speaking the minority language at home was discouraged. However, this view is already changing and parents are beginning to recognise the potential benefits of bilingualism and the growing body of evidence to support such a view (see O’Hanlon, McLeod, & Patterson, 2010). It is clear from the current research that the speaking of these minority languages, whether it be at home or in a school setting, but preferably both, should be encouraged. Otherwise minority languages, such as Sardinian and Gaelic, and many others like them, will struggle to survive well into the 21st century.
Footnotes
Funding
The research was carried out while the first author undertook a Visiting Professor post at the Università di Cagliari, funded by the Regione Autonoma della Sardegna (the Sardinian Regional Government).