Neurolinguistic aspects of bilingualism

Volume 6 Number 4 December 2002, 411– 440 Neurolingustic aspects of bilingualism

411

Neurolinguistic aspects of bilingualism* Mira Goral 1, Erika S. Levy 2 and Loraine K. Obler 1,2 1 2

Boston University School of Medicine, VA Boston Healthcare System City University of New York Graduate School and University Center, Program in Speech and Hearing Sciences

Acknowledgments* An earlier version of this paper was presented at the conference La Langue Maternelle/ The Mother Tongue, held at the Université de Paris, March, 1999. That version is in the Conference Proceedings: Levy, E. S., Goral, M., and Obler, L. K. (1999). Neurolinguistic perspectives on mother tongue: Evidence from aphasia and brain imaging. La Langue Maternelle Cahiers Charles V, 27. Paris: Publication de l’Université de Paris 7, 141– 157. We thank three anonymous Bilingualism: Language and Cognition reviewers for their suggestions concerning the revision of that conference presentation, Nick Miller for his suggestions on this paper, and Jyotsna Vaid for providing us with recent references.

Key words

Abstract

To study the brain regions and networks that underlie knowledge of more bilingualism than one language, neurolinguists have traditionally compared what is impaired with what is spared in the language disturbance of aphasia. The multilingualism sizable literature on polyglot aphasia suggests left-hemisphere dominance for all the languages of most polyglots. Supporting evidence comes from the neurolinguistics literature studying lateral dominance in non-brain-damagedbilingual participants and studies of crossed aphasia. Imaging techniques such as cortical stimulation, Positron Emission Tomography (PET), and functional Magnetic Resonance Imaging (fMRI) offer the possibility of interspersed networks for multiple languages in the left hemisphere that are largely but not entirely overlapping. The Evoked Response Potential (ERP) literature, moreover, concurs with the finding of overlap in processing for proficient bilinguals and greater differences for less proficient bilinguals. Directions for future research such as documenting the underpinnings of recent behavioral findings on polyglots’ lexicons, individual variability, and language attrition are outlined.

1 Introduction The goal of neurolinguistic study is to determine how language is organized and processed in the brain. Three bodies of research contribute to our current understanding of how the brain subserves bilingualism: We begin with a discussion of aphasia, the language deficit resulting from damage to the language centers of the brain, in order to evaluate how research on bilingual and polyglot aphasic individuals has contributed to our knowledge of the representation of language and languages in neurologically intact humans’ Address for correspondence

Mira Goral, Boston University School of Medicine, Department of Neurology, VA Boston Healthcare System, 150 S. Huntington Ave., Boston 02130, U.S.A.; e-mail: < [email protected]>.

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

412

M. Goral, E. S. Levy, and L. K. Obler

brains. We then consider the corpus of literature treating lateral dominance for language in bilinguals. Finally, we evaluate the more direct observations of language processing in the brain, such as cortical stimulation and other forms of brain imaging. We discuss converging, diverging, and problematic evidence from these various strands of neurolinguistic research and their implications for language representation and processing in bilinguals and polyglots. We conclude with mention of topics yet to be addressed in exploring the brain bases for knowledge of more than one language. (Until the penultimate section of this paper, Future Research, we use the terms bilingual and polyglot interchangeably.)

logic of aphasia studies of language 2 The organization For over a century, the primary technique for inferring how language is organized in the brain was the study of aphasia acquired after injury from war, strokes, and tumors that damaged the language areas of the brain. The logic of the classic neurolinguistic argument is the following: In the normal brain, language operates in so complex a manner that it is difficult to extricate its component units or processes. However, when language breaks down after brain damage, many aphasiologists have maintained, the componential nature of language may be revealed. Language breakdown is generally selective, with some components spared, and others impaired. A subtractive methodology permits us to deduce the structural components. That is, we assume that brain damage does not create new ways of processing language but rather reveals, through a contrast between what is destroyed and what is spared, components of language processing. When researchers observe that certain structural components of language are consistently impaired following damage to particular brain areas, they conclude that those brain regions are responsible for, or at least involved in, those language abilities in the healthy brain. It was this logic that led Paul Broca (1861, 1865) to propose that the third frontal convolution of the left hemisphere was involved in deficits in spoken language, and Carl Wernicke (1874) to conclude that an area at the posterior limits of the Sylvian fissure in the first left temporal gyrus was responsible for comprehension of spoken language. While modern connectionist critiques of this logic point out that the dissociations between spared and impaired language abilities can be simulated under nonmodular assumptions (e.g., Farah & McClelland, 1991; Plaut & Shallice, 1993) and that a one-to-one relation between lesion site and linguistic deficit may not be present for many aphasic individuals (e.g., Goodglass, 1993), the study of bilingualism in the brain has been undertaken from the assumption that functional components must be instantiated in separable, albeit potentially interwoven, brain systems.

3 Polyglot aphasia The earliest neurolinguistic studies, at the end of the 19th century, were based in France and Germany, where many individuals were multidialectal if not multilingual. As a result, most instances of aphasia observed probably occurred in people who were at least bilingual or bidialectal. However, this facet of their language abilities was not mentioned in these early reports. In the late 19th century, several cases of polyglot aphasia did begin to The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

Neurolingustic aspects of bilingualism

413

pique the interest of scientists when unexpected patterns of language breakdown and recovery in patients who had been multilingual were revealed. For over a hundred years of research into polyglot aphasia, a few questions have been the primary focus of discussion. The question most frequently addressed in the literature has been the following: When differential recovery occurs (i.e., when one of the languages undergoes recovery and others do not or not to the same extent), what factors explain which language returns first? Secondary to this question have been the questions of the frequency with which all languages are affected equally in aphasia, and whether crossed aphasia (i.e., aphasia caused by right-hemisphere lesion in a right-handed individual) is seen more frequently among bilinguals and polyglots. Answers to such questions should provide insight into the premorbid (i.e., before the aphasia-producing incident) representation and processing of a polyglot’s languages. 3.1 Differential impairment and /or recovery

It is generally the case that multiple languages are lost to the same extent, and are regained simultaneously (e.g., Charlton, 1964; Fabbro, 1999; Obler & Albert, 1977). This suggests that for most bilinguals and polyglots, the areas involved in processing language may be the same; a lesion to those areas has resulted in parallel problems and parallel recovery. Nonparallel damage and / or restitution is the more intriguing pattern of recovery, however, wherein one language is restored before the rest, and may recover to a fuller extent than the other languages (e.g., Albert & Obler, 1978; Paradis, 1977). Although such cases are more frequent in the literature, in sequential studies of all bilingual patients seen, for example in our clinical observation and in Charlton’s 1964 report, they prove to be rare. For the cases of differential recovery, neurolinguists have investigated characteristics that might predict which of a patient’s languages will be the first to recover. Some of the factors that have been hypothesized to explain which language returns first include the status of the language, that is, whether a language was the mother tongue (Freud, 1891; Ribot, 1881), the language most used around the time of the lesion (Pitres, 1895), the language that is most practical at the time of recovery (Bay, 1964; Geschwind, personal communication), or the language motivated by unconscious affective factors (Minkowski, 1927). The first two factors, order of acquisition and language-familiarity at the time of the accident, have received the most attention and support in the literature. What has come to be known as Ribot’s rule (1881) predicts that the language that was learned first will be the first to recover and will be less impaired than the other languages. Von Mundy’s (1959) report of the case of JP illustrates this rule: JP was born and raised in Slovenianspeaking villages until he was recruited for military service in the German-Austrian infantry, where he learned German. After 12 years of military life, he settled down in the Austrian town of Admont, and, according to his son, had no opportunity to speak Slovenian for 40 years. After a stroke in the left cerebral hemisphere, JP began speaking only his mother tongue, Slovenian; that is, his first language was first to recover. Findings from this case are consistent with the views of Freud, who trained with the French neurologist Jean-Martin Charcot. Speaking of the saliency of the mother tongue in his monograph On Aphasia (1891), Freud asserted: “It never happens that an organic lesion causes an impairment affecting the mother tongue and not a later acquired language” The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

414


(1891, p. 24). According to Freud, the superimposed associations of a second language were more vulnerable to damage than was the primary language. However, as the polyglot aphasia literature grew, cases were reported for which Freud’s assertion was false; that is, cases in which the mother tongue was more affected and slower to recover than other languages. Pitres (1895) tried to reconcile this seemingly contradictory evidence in the literature, and proposed that polyglots’ premorbid familiarity with languages was more useful in predicting their recovery patterns than the order of acquisition of their languages. Pitres’ rule stated that the first-recovered language was the one that was the most familiar at the time of the accident. This may or may not be the mother tongue, of course. For example, among the seven cases described by Pitres (1895) was the case of Jean, whose mother tongue was Béarnais patois, but who moved from the BassesPyrénées to Bayonne at age 12, learned French, and thereafter spoke mostly French. After a stroke at age 48, he could express himself well in French, but found it impossible to express a thought or construct a sentence in his native Béarnais dialect. Pitres asserted that the reason the most familiar tongue was the first to recover was that it had the most solidly fixed associations. In recent decades, a few other questions have been addressed in the polyglot literature, and new patterns of recovery have been documented. Most pertinent to our discussion in this paper is a case Paradis and Goldblum (1989) reported of a 24-year-old righthanded multilingual man who had problems in only one language after surgery for a brain tumor affecting posterior areas of the right hemisphere. After surgery his naming and comprehension, as well as speech production, were impaired in Gujarati, the mother tongue he had still used at home, while no such problems were evident in his second and third languages, Malagasy and French. However, as his Gujarati improved over the following two years, his abilities in Malagasy, the language of the environment, decreased. When tested four years after the surgery, his language abilities in Malagasy were assessed to be normal as were those of Gujarati and French. Another pattern of interest is the one labeled “alternating antagonism” by Paradis, Goldblum and Abidi (1982). The case they presented was seen several days after a motorcycle accident. The patient’s skills in his two languages appeared to alternate in strength every day or every several days; on one day the patient could speak French, on the next day Arabic, and on the third French again. Interestingly, on days the patient could speak French, he could translate from Arabic to French but not from French to Arabic, and vice versa, on days he could speak Arabic he could not translate from Arabic to French. A similar pattern was described in Nilipour and Ashayeri (1989) for a Farsi-German-English trilingual. In addition to polyglots’ patterns of selective accessibility to their languages, several cases of brain damage resulting in inappropriate language choice and / or involuntary language mixing have been reported (e.g., Aglioti, Beltramello, Girardi, & Fabbro, 1996; Fabbro, 1999; Fabbro, Skrap, & Aglioti, 2000; Perecman, 1984; Pötzl, 1925). Fabbro et al. (2000) reported a case of a nonaphasic bilingual who suffered pathological switching between his two languages as a result of a tumor in his left hemisphere. Specifically, lesions in his left anterior cingulate and the left frontal lobe apparently yielded language production characterized by compulsive switching from his native Friulian (L1) to Italian, his second language (L2) and vice versa, even in conversations with monolinguals. No other notable language deficit was observed. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


415

Pathological language mixing in aphasic bilinguals and multilinguals is characterized by inappropriate mixing of elements (sounds, words, phrases) from two or more languages within an utterance, as demonstrated, for example, by the case reported in Perecman, 1984. Interestingly, many of the reported cases of excessive switching have been of patients with fluent aphasia (Fabbro, 2001). While increased language mixing (codeswitching) was not found to be associated with degree of severity of the Alzheimer’s Dementia (AD) patients reported in Friedland and Miller (1999), these AD patients showed little awareness of their inappropriate codeswitching and initiated little attempt to repair their production. The differential patterns of deficit and recovery in polyglot aphasia (see Fabbro, 1999, for a recent compilation of them) have opened interesting theoretical questions about the differentiability of “wiring” systems for the languages of the bilingual or polyglot individuals and of the resources available for controlling selection among them. For instance, if both languages are accessible, albeit in an alternating pattern, it cannot be assumed — at least for such cases — that the brain lesion destroyed or impaired one of the languages and spared the other. Indeed, Green (1986, 1998) has proposed a model of inhibitory control of resources that suggests that it is not the languages per se that are affected in cases of nonparallel patterns in bilinguals after brain-injury, but rather the ability to access and inhibit them appropriately. Green’s approach applies to verbal production in bilinguals and assumes two distinct but interdependent systems for the two languages. A language must be adequately activated to be selected and adequately inhibited to avoid interference. An additional component of this model is the “resource generator”: adequate resources must be available for a language to permit verbal production. Impairment of voluntary control over these resources and over the processes of activation and inhibition may account for various patterns of differential recovery in multilinguals while allowing varying degrees of spared abilities in each language. Paradis (1989, 1996) discussed the notion of language activation and proposed that a threshold of activation must be met for a language to be available for production; activation of an item in one language cooccurs with reduced activation of competing items from the other language(s). Alternatively, the intriguing cases of differential recovery suggest that, at least for some bilinguals and polyglots, different (or not fully overlapping) areas or subsystems of the brain were responsible for the representation and / or processing of the different languages. The fact that not all bilinguals evidence differential patterns, however, suggests marked individual differences in brain organization for the languages of the neurologically intact bilingual. Some factors that may contribute to this variability will be addressed further below. 3.2 Group studies of factors predicting differential recovery

In an early attempt to determine what features or characteristics may enter into predicting the pattern of language impairment and recovery in polyglot aphasia, Obler and Albert (1977) reviewed the 106 cases of bilingual / polyglot aphasia available in the literature, and statistically evaluated for each participant the extent to which the language that returned first followed the law of Pitres and / or the law of Ribot. Chance was calculated for each case based on the number of languages the patient had known premorbidly. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

416


They found that while Pitres’ rule obtained with significantly greater than chance accuracy (p < .025) for participants up to age 60, Ribot’s rule did not. For participants over age 60, interestingly, neither rule predicted recovery with greater than chance accuracy. Table 1 First-recovered language in polyglot aphasics: Cases consistent w ith Ribot’s rule and / or Pitres’ rule (analysis based on cases review ed by Obler & Albert [1977] )

Consistent with Ribot’s rule and Pitres’ rule Consistent with neither rule Consistent with either rule Of those consistent with either rule, cases consistent with Ribot’s rule Of those consistent with either rule, cases consistent with Pitres’ rule

16/49 4/49 29/49 7/29 22/29

(33%) (8%) (59%) (24%) (76%)

For several of the analyzed case studies, the language most familiar at the time of the onset of aphasia was also the mother tongue (with findings consistent with the rules of both Pitres and Ribot). Recently (Levy, Goral, & Obler, 1999), we have reanalyzed a subset of those data from a different perspective (as shown in Table 1), asking which language was more likely to be the first to recover when the mother tongue (when a clear first language could be identified) and the most familiar tongue were not the same. It should be noted that for a good number of the cases in the literature, only partial information is available about the patients’ history of language acquisition, use, and proficiency in their two or more languages. Out of the 49 cases of which information about both the mother tongue and language in use at the time of the aphasia was known, 16 cases (33%) are explained by the rules of both Pitres and Ribot and four cases by neither rule. The remaining 29 cases (59%) are predicted by one of the rules, but not by both. Only seven of these cases (24% of the 29 cases) follow Ribot’s law, whereas 22 cases (76%) follow Pitres’s law. A binomial test places the likelihood that such a distribution would have arisen by chance at less than .0001. Consistent with the Obler and Albert (1977) analysis, then, this suggests that a language learned later in life, but more used around the time of the stroke, is more likely to be the first recovered than even the mother tongue. This finding is, of course, inconsistent, at least with respect to language performance, with the notions of Ribot and Freud that early childhood plays a crucial role in later life. However, it is consistent with our experience, anecdotal reports, and the growing literature on language attrition that suggest that even a native language can become lost (in the case of children) or at least inaccessible (in the case of adults) if it is not used for a period of time while a second language is learned and employed (e.g., de Bot, 1996; Olshtain & Barzilay, 1991; Schmid, 2002; Seliger, 1991). The explanations of Pitres and others that such phenomena arise because of diminishing “associations” between the words of a lessused language are compatible with current behavioral models of multilingualism (e.g., Kroll & de Groot, 1997; Kroll & Dijkstra, 2002), including parallel-distributed processing models, although the brain-bases underlying them have yet to be studied. Of course, one must bear in mind the caveat that precisely because the rule of Pitres is more surprising than that of Ribot, more cases supporting the former may have seen their way to publication. As a result, Obler and Mahecha (1991) analyzed the first 156 cases of bilingual and polyglot aphasia in the literature to see what other factors predicted The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


417

recovery. In that study, power was sufficient to determine that patient handedness and hemisphere of damage were linked to whether or not the first-learned language recovered first when only one language recovered. Left-handers and those (left- or right-handers) with right-hemisphere lesions were significantly less likely than right-handers (p = .05) and than individuals with left-hemisphere lesions (p = .03) to show salient L1 recovery. We have learned from the case and group studies of aphasia in polyglots, then, that there is a substantial overlap in the inferred representation of languages for most patients. However, it is also true that individual differences in brain organization for different languages are revealed in some patients, perhaps particularly those who are left-handed or have otherwise anomalous dominance. Moreover, we conclude that ongoing language use is a better predictor for the pattern of language recovery than primacy of language acquisition for nonelderly aphasics (as well as neurologically intact bilinguals outside of their L1 environments). Aging-related changes in brain substrates underlying language representation and processing, however, seem to disinhibit L1 in demented patients and also in older aphasics more often than younger ones. We refer in later sections of this paper to the handedness-group differences in brain organization for bilingualism (e.g., Lamm & Epstein, 1999), and to the individual differences in right-hemisphere participation in L2 organization (e.g., Dehaene, Dupoux, Mehler, Cohen, Paulesu, Perani, van de Moortele, Lehericy & Le Bihan, 1997). 3.3 Crossed aphasia

Crossed aphasia refers to cases of aphasia resulting from damage to the right hemisphere of right-handed individuals. Though rare, it has piqued the interest of researchers who attempted to explain the anomalous hemispheric organization (e.g., Alexander & Annett, 1996). The discussion of whether crossed aphasia is more frequent among bilinguals and polyglots than among monolinguals was first raised by Gloning and Gloning (1965). They had observed four cases of aphasia in polyglots who had been right-handed yet had become aphasic as a result of damage to the right hemisphere. They suggested that there was more right-hemisphere involvement in polyglot aphasia, and hence, by implication, in healthy polyglots and bilinguals, than in monolinguals. The case-study literature on polyglot aphasics does not overwhelmingly support higher incidence of crossed aphasia in right-handed patients (than in monolinguals), and, as mentioned above, is quite skewed toward reports of the more unusual. While the possibility that crossed aphasia is more frequent in bilinguals than monolinguals remained unresolved in the late 1970s (Albert & Obler, 1978), April and Han (1980) systematically studied each case of aphasia in Chinese speakers entering their department. They expected that, particularly because of the ideographic representation of Chinese, more right-hemisphere organization might be seen in those Chinese-speaking bilinguals. Yet they found the usual low percentages of crossed aphasia seen in monolingual speakers of Western languages. Similarly, Chary (1986) reported equal incidents of crossed aphasia among monolinguals and multilinguals. However, Karanth and Rangamani (1988) provided evidence for possible differential language representation between monolinguals and multilinguals. They set out to systematically assess the prevalence of crossed aphasia in an unselected sample of monolingual and multilingual aphasics. They recorded the number of incidents of unilateral The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

418


aphasia, left-handedness, and crossed aphasia in several aphasic samples in India. They found some discrepancy among the samples they evaluated, with one sample showing greater percentage of crossed aphasia among the multilinguals and another showing greater percentage of crossed aphasia among the monolinguals. It should be kept in mind, however, that damage confined to the right hemisphere might not result in aphasia if the left hemisphere could appropriately take over. Furthermore, Karanth and Rangamani discuss the importance of the social attitude toward left-handedness and its effect on valid reports of left-handers among the studied populations. Specifically, if left-handedness is underdocumented due to social bias, the number of cases of aphasia resulting from right-hemisphere lesions in right-handed individuals may be inflated. Given these inconclusive findings, one cannot rule out the possibility of more bilateral organization for languages in bilinguals and polyglots; nevertheless, the fact that so many instances of aphasia in bilinguals or polyglots are seen as the result of left-hemisphere damage strongly suggests that it is the left hemisphere that is dominant for both or all languages.

4 Laterality research After the Glonings’ (1965) study of crossed aphasia in four right-handed individuals suggested that there might be more right hemisphere involvement in bilingual processing than in monolingual processing, two decades of laterality research on neurologically intact individuals ensued. Techniques that employed auditory stimuli (dichotic listening) and / or visual stimuli (tachistoscopic presentations) were used with numerous populations. Such techniques have been used to determine if, indeed, greater right-hemisphere lateralization, either for the second language compared to the first, or for the bilingual generally compared to the monolingual, can be demonstrated. In the dichotic listening technique, a list of several words is presented to one ear of the participant at the same time a different list is presented to the other ear. Participants are asked to recall all the words they hear; as a rule, they recall many but not all. Typically, some of the items from the middle of the list presented to the ear on the side of the dominant hemisphere cannot be recalled. This finding is taken to reflect the fact that the connections between that ear and the nondominant hemisphere are stronger than those between that ear and the dominant hemisphere. Thus, for most humans who are lefthemisphere dominant for language, the information delivered to the left ear travels predominantly to the right hemisphere first. Information delivered to the right ear arrives at the language processing areas of the left hemisphere first and may block left-hemisphere processing of some of the information that was received in the left ear. A comparison of the relative success with information that was delivered to left versus right ears, then, is a coarse measure of lateral dominance for auditory processing. Tachistoscopic presentation, by contrast, is employed to study hemispheric processing of visually presented materials. Written words are presented to the left or right of center at a sufficient angle so that in the brief period for which they are presented, one can assume that the information follows the strong pathways to the contralateral hemisphere first. As with the dichotic-listening technique, we assume that more correct information has been reported from the dominant hemisphere. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


419

These techniques would be ideal for measuring the relative dominance of the left hemisphere for each of a polyglot’s languages if they were more reliable (see the critique of the methods for monolinguals of Blumstein, Goodglass & Tartter, 1975), and if it were possible to pick perfectly matched sets of stimuli across languages. A sizeable number of such laterality studies of bilinguals was reviewed in Obler, Zatorre, Galloway, and Vaid (1982), as well as in Galloway and Scarcella (1982), Vaid (1983), Zatorre (1989), and Vaid and Hall (1991). In these reviews it became clear that substantial methodological issues arose in subject- and stimulus-selection and stimulus presentation. Among the participant factors that were focused on in some but not all studies were, as in the polyglot aphasia literature, age of acquisition of the L2, manner of acquisition, stage of acquisition, and proficiency. Problematic in stimulus selection, for example, are issues such as matching for word-frequency and length. For many languages, no word-frequency lists exist; for those for which they do, such lists have not been comparably constructed. With respect to word-length, one faces the question of whether words should be balanced for actual length (which may be measured in letters, phonemes, or syllables) or length relative to length-frequency in each language (e.g., two-syllable words are relatively more common in English than in Spanish, where three-syllable words are relatively more prevalent). Such considerations apply, of course, to studies of bilingual language representation generally. These reviews revealed that a sizeable number of the published laterality studies in normal bilinguals found no evidence for laterality difference either between the first and second languages, or between the language organization of the bilinguals and that of their monolingual controls. Another subset reported left hemisphere dominance but greater right hemisphere participation, either in the second language as compared to the first, or the bilingual as compared to the monolingual. In their review papers, Vaid (1983) and Zatorre (1989) and in the meta-analytic study of Vaid and Hall (1991) the authors conclude that the weight of the evidence is against enhanced right-hemisphere involvement in a second language or in bilinguals. Vaid and Hall indicate the need for more careful attention to the tasks employed, suggesting that tasks that require or permit semantic analysis seem to involve more right-hemisphere participation, perhaps particularly in early bilinguals, as compared to more strictly phonological or syntactic tasks. It appears, as demonstrated in Silverberg, Bentin, Gaziel, Obler, & Albert (1979), that in the early stages of learning a second language (and perhaps acquiring one as well), there is additional right hemisphere involvement. With greater proficiency and, presumably, automaticity, however, as Silverberg et al. demonstrated, one sees in bilinguals left-hemisphere dominance equivalent to that found in monolinguals. It is likely, we conclude, that the increased right hemisphere participation that has been evidenced in the early stages of acquiring a second language is due to the fact that the language is not being processed exclusively as “language” but perhaps as “novel material” requiring right hemisphere involvement, as, perhaps, in the early stages of L1 learning (see Saykin, Johnson, Flashman, McAllister, Sparling, Darcey, Moritz, Guerin, Weaver, & Mamourian, 1999, e.g., for a comparison of brain activations during processing old and new material). Such a developmental pattern would be analogous to musical representation in the brain. For most humans, the right hemisphere is responsible for musical processing; for professional musicians, the automatic and sophisticated processing appears The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

420


to be left-hemisphere dominant (Zatorre, 1984). Alternate explanations to be explored would be that early stages of L2 acquisition include relatively more concrete materials (known to have right-hemisphere representation in monolinguals) and / or that wholistic supersegmental “right-hemisphere-style” units, including intonation, are being disproportionately processed in the early stages of L2 learning. Similarly, a strong relation between changes in language processing and changes in language proficiency has emerged from recent brain imaging studies of bilinguals (e.g., Abutalebi, Cappa, & Perani, 2001) as well as psycholinguistic studies of bilinguals (e.g., Goral, 2001; Jared & Kroll, 2001). In sum, the laterality literature, to the extent it can be relied upon, suggests substantial overlap of processing of all languages and left-hemisphere dominance for them, perhaps particularly once speakers are proficient in them. However, the possibility remains of increased right-hemisphere participation, over that seen in monolinguals, in the early stages of L2 learning.

technologies for observing language 5 Imaging processing Cortical stimulation in the 1970s (and later), and brain-imaging techniques of the past two decades have permitted researchers to observe multilingual brains relatively directly as they process language. Consequently, we are getting closer to be able to answer questions of language representation in the brain. However, the results with respect to organization of language in the bilingual brain remain inconclusive. The field of neuroimaging is still developing and methods and interpretations of results can be problematic (See Aine, 1995, and Sergent, 1994, for accessible critiques of the neurological issues that arise in applying these techniques to monolinguals). Crucially for neurolinguists studying bilinguals, additional weakness lies in the paucity of critical information generally provided about participants’ proficiency (in the mother tongue and / or in the other languages), age and manner of acquisition of other languages, and general language-usage and language-disability history. In what follows we discuss individual cases and group studies reported in the literature on brain-imaging studies, including cortical stimulation, positron emission tomography, functional magnetic resonance imaging, and evoked related potentials. In this relatively small set of articles, all studies provide evidence for at least substantial common areas of representation of languages in the bilinguals’ cortices, and most report areas of separate representation along with the overlap. 5.1 Cortical stimulation

Cortical stimulation, introduced in the 1950s, may be considered one of the earlier brainimaging techniques in that investigators are able to employ it to map a patient’s language area. This technique is used primarily for patients who are preparing to undergo surgery for intractable epilepsy, in order to determine the brain regions involved in speech and other cortical functions. Since the brain has no pain receptors, the patient remains conscious as the surgeon opens the cranium and electrically stimulates areas of the cortex. Small voltages applied to the language area have typically caused patients to become temporarily incapable of naming items. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


421

In 1978, Ojemann and Whitaker adapted Penfield and Roberts’ (1959) cortical stimulation procedures to multilingual brains, mapping their bilingual patients’ languages. By way of example, consider Case 1 of the two cases that presented similar cortical patterns in their study. Case 1 was a 37-year-old right-handed patient who spoke only Dutch until age 25, when he moved to the United States and learned English. Ojemann and Whitaker’s modified technique included naming tasks in both Dutch and English. The patient was presented with line drawings of 45 common objects. Printed above each object was “This is a” and he was instructed to read the phrase and then to name the object. A preoperation baseline performance in each language was obtained. The task was performed in both English and Dutch. The assumption underlying this technique is that if a brain region involved in language production is stimulated during the naming task, the patient will have difficulty naming the item in that language. They found several sites at which both languages were affected — around Broca’s area as well as in the inferior parietal lobe. They also found several sites of differential localization — in the frontal lobe, and around Wernicke’s area in the superior temporal gyrus, where naming in one language only was affected (see e.g. Bhatnager, 2002, for the brain regions referred to here and in the following sections). This general pattern (i.e., some sites of overlap, some independent representation) seems to hold for sign languages, as well. Haglund, Ojemann, Lettich, Bellugi, and Corina (1993) reported the case of a native English-speaking participant who learned American Sign Language (ASL) in early childhood in order to communicate with her deaf sister. A cortical stimulation experiment using a naming task performed in English and ASL when she was 26 revealed the following: Partial overlap, but also sites with L1-only (English) around Broca’s area and along the superior portion of the temporal gyrus, and L2-only sites (several probably related to fingerspelling) in several portions of the temporal lobe. Rapport, Tan and Whitaker (1983) replicated much of Ojemann and Whitaker’s (1978) electrical stimulation technique on seven multilingual patients in Malaysia. For example, consider the patient they presented as Case 1. He was a mathematics teacher whose mother tongue was Cantonese, but whose dominant language was English. In the naming task, the patient was shown line-drawings of depicted objects headed by a phrase such as “This is a” and the patient would complete the sentence in English. The same procedure was performed in Cantonese, with the patient saying the introductory phrase and naming the object in that language. Stimulation in a site in Wernicke’s area consistently produced a complete inability to name in L2 only. Individual errors in one language but not in the other were found in several spots. Stimulation in certain sites produced consistent errors in one language but only a single error in the other. Although errors did not occur consistently in the same location for L1 (Cantonese), the fact that stimulation of certain sites did result in single errors in Cantonese and none in English suggests that there may be some independent representation in these sites. It must be noted that the findings from other cases in this study were less clear, and inconsistency between text and figures in that report renders results difficult to interpret. In considering the implications for organization and function of bilinguals’ brains, it is important to note that differences evidenced by cortical stimulation are subtle, with substantial overlap of languages, all within the general language area of the dominant The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

422


hemisphere. Of course, as a rule, only one hemisphere is tested via cortical stimulation, as the primary goal of this technique is to determine how to avoid language areas in conducting surgery to alleviate epilepsy, and the hemisphere where the epileptic focus lies is the one that must be tested. Thus we cannot know if stimulation of sites in the nondominant hemisphere would also result in naming problems in one language or more. In the past decade, virtually noninvasive technologies such as functional magnetic resonance imaging (fMRI), positron emission tomography (PET) scans, and evoked response potentials (ERPs) have been advanced to the point of being used for research purposes on neurologically intact individuals. Such techniques have, moreover, permitted neurolinguists to ask more refined questions about the organization of languages in the human brain. 5.2 Positron Emission Tomography (PET)

PET scans have been used as a tool for studying various cognitive processes, including the representation of languages. When part of the brain requires additional oxygen in order to perform an activity, glucose uptake increases in that area. PET scans show cerebral blood flow changes, permitting researchers to infer where processing occurs by averaging changes over numerous repetitions of a task (for a review and a critique of the PET method see, e.g., Aine, 1995). A PET study by Klein, Zatorre, Milner, Meyer, and Evans (1994) showed similar patterns of blood flow when proficient adult native English-speakers who had learned French around age seven repeated words in English and in French. Increased blood flow during production in L2 was found only in the region of the left putamen. In a second study from the same laboratory, Klein, Milner, Zatorre, Meyer, and Evans (1995) again found virtually no differential representation of bilinguals’ representation for first and second language rhyme- and synonym generation, translation, and repetition. However, they found no consistent activation in any particular locus for L2 processing across their participants, which led them to suggest that their findings of no differential representation could be due to intersubject variability. Nevertheless, Klein et al. concluded that a language learned later in life is represented in a common area with the mother tongue. This conclusion has raised criticism from researchers (e.g., Dehaene et al., 1997; Kim, Relkin, Lee, & Hirsch, 1997) because PET scans have serious limitations as tools for studies of language processing. In PET studies, intersubject variability may be lost because, as Klein et al. pointed out, all of the participants’ results are averaged (as opposed to the single-subject analyses available in fMRI studies). In addition, PET images are limited in their spatial resolution; thus if loci are distinct but close, PET may not be as useful a tool to detect differential areas of language processing (again, compared to fMRI scans, which offer better spatial resolution). In contrast to the results reported by Klein and her colleagues, in another PET study, conducted by Perani, Dehaene, Grassi, Cohen, Cappa, Dupoux, Fazio, and Mehler, (1996), some degree of nonoverlap was seen in a group of nine participants, between L1 (Italian) and an L2 (English) acquired after age seven. Specifically, listening to stories in L2 resulted in more extensive activation than listening to stories in L1, although, surprisingly, the areas activated for L2 were no different from those activated by a language not known by participants (Japanese). In line with Kim et al.’s comment on Klein et al.’s The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


423

results, Perani et al. concluded that intersubject variability, especially for L2, may account for this finding. The participants included in Perani et al. (1996) had all learned their L2 after age seven but, in contrast to those included in Klein et al. (1994, 1995), did not attain high proficiency in that language. In an attempt to dissociate these two variables — age of L2 learning and level of L2 proficiency — Perani et al. (1998) conducted two additional PET studies. In the first experiment, they asked one group of highly proficient bilinguals who learned their L2 after age 10 to listen to stories in Italian, their L1, English, their L2, and Japanese, an unfamiliar language. In their second experiment, they asked another group of proficient Spanish-Catalan bilinguals who learned both languages in early childhood to listen to a story in their dominant language (Spanish or Catalan), a story in the nondominant language, and to Spanish or Catalan played backwards as a control condition. Perani et al. (1998) found similar areas of activation for both groups of bilingual participants for their two languages. Based on these results, Perani and her colleagues concluded that regardless of age of L2 learning, the same regions are involved in processing both languages of proficient bilinguals and that differential areas may be involved in processing a second language at early stages of L2 learning. This account is also consistent with the findings of Klein et al. of their proficient bilingual participants. In an interesting case study using PET, Tierney, Varga, Hosey, Grafman, and Braun, (2001) reported a bilingual speaker who learned both English and ASL from infancy and had sustained left brain damage in childhood. When tested in adulthood, he demonstrated predominately right-hemisphere representation of both languages. Narrative production in both English and ASL by this early bilingual yielded activation in the right inferior frontal operculum (for English) and right posterior superior temporal gyrus (for ASL). These areas in the right hemisphere were homologous regions to those activated, in the left hemisphere, in the control group. Despite the fact that the area associated with ASL (e.g., in the control group in this study) was actually spared in the left hemisphere of this participant, his right, intact hemisphere appeared dominant for processing both languages. PET has also been used to study brain activation for translation and switching between languages by Price, Green, and von Studnitz (1999). Price et al. set out to compare the patterns of brain activation during word reading and word translation in two presentation conditions: single-language lists and mixed-language lists. Their participants were proficient German-English bilinguals who learned their L2 in late childhood. They were asked to read or translate written words in German-only, English-only, and mixed-list conditions. Greater activation of left supramarginal gyrus was seen for L1 than for L2 words in the reading task. The small difference Klein et al. (1994, 1995) had reported for left putamen activation for L2 speech production was not evident here. Interestingly, Price et al. reported specific activation for translation. Namely, greater activation during translation (compared to during reading) was found in the left anterior cingulate and in subcortical regions bilaterally (putamen and head of caudate), and decreased activation was found in the medial superior frontal gyrus, left middle temporal regions, left posterior temporoparietal region, and the posterior cingulate. The authors linked this finding to the necessity of greater coordination of processes such as inhibiting the reading of the written word aloud in order to speak its translation equivalent. Moreover, this finding The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

424


is consistent with the deficit in the ability to appropriately switch between two languages following a lesion in the left anterior cingulate reported in Fabbro et al. (2000). In a related study, Rinne, Tommola, Laine, Krause, Schmidt, Kaasinen, Teräs, Sipilä, and Sunnari (2000) used a PET procedure to assess brain activation during simultaneous interpretation. Their participants were highly proficient Finnish-English bilinguals who were actively working as professional interpreters at the time of the study. Rinne et al. found that the simultaneous translation task activated more extensive left frontal and left temporal regions than did repetition in either language. Differences were also found between the two translation directions, with more extensive activation in left frontal regions during translating from L1 to the non-native L2. Thus, while several studies found that bilinguals performing tasks in their two languages demonstrated similar brain activation for both languages (Klein et al., 1994, 1995; Perani et al., 1998; Tierney et al., 2001), others found some nonoverlapping activation (Perani et al., 1996; Price et al., 1999; Rinne et al., 2000). Some of these discrepancies may be accounted for by the different levels of participants’ proficiency (see below), however, additional factors should be considered. Price et al. (1999) pointed out that the task chosen for baseline (e.g., repetition in the case of Klein et al., 1994, reading in their study), as well as the modality of the task itself (e.g., auditory input in Klein et al., written input in their study) must be considered in interpreting apparently contradictory results. Indeed, in their comprehensive review of PET and fMRI studies of bilinguals, Abutalebi, Cappa, and Perani (2001) demonstrated the importance of the characteristics of the task used in these studies in accounting for differential findings (e.g., different patterns of results have been obtained for comprehension vs. production tasks), in addition to participant variables (e.g., different patterns of results have been obtained for more- vs. less-proficient bilinguals and possibly for languages learned early vs. later in life). Participant variables are particularly crucial in interpreting PET studies because individual data are averaged to yield the results; intersubject variability, as mentioned above, will greatly affect such results. 5.3 Functional Magnetic Resonance Imaging (fMRI)

In fMRI studies, magnetic resonance images of the brain are photographed while participants are performing a task. These pictures are taken rapidly enough to provide high-resolution images of the regions involved in the on-line language processing (see, e.g., Obler & Gjerlow, 1999, and Leblanc & Zatorre, 1997 for further explanation). In their report on their fMRI study of bilinguals, Kim et al. (1997) presented vivid images of common and differential language processing, as well as a strong case for the importance of age of acquisition for the representation of languages in bilinguals. Participants were selected to form two groups: early and late bilinguals. Early bilinguals were those who had been exposed to two languages in infancy and had spoken both of their languages continuously and regularly since acquisition. Late bilinguals had started learning their second language at a mean age of 11.2 years. Ten languages were represented in the study. Participants were asked to generate sentences describing events of the previous day. In order to avoid brain activation resulting from movement, participants were asked to generate their language production silently. The participant would be instructed in which language to perform the task, whereupon 30 images of different slices of the brain The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


425

were obtained. It should be noted that indeed the range of tasks used to date in brainimaging studies has been rather limited; because any movement in the head generates confounding images, most studies have employed silent tasks while only a few studies used noncovert tasks, comparing them to nonlanguage related tasks of equivalent motor activity. When the representation of languages within early bilinguals was compared to that of late bilinguals, the results were striking. In Broca’s area, if both languages were learned early, they were represented in common cortical areas, whereas if a second language was learned later, it appeared to be somewhat separate from the native language. The figures Kim et al. included of areas differentially involved in “production” of L1 and L2, moreover, evidence great intersubject variability. In Wernicke’s area, by contrast, regardless of when the language was learned, little or no separation of activity was found. We would add that since this was intended as a production task, substantial activity of Wernicke’s area (associated with comprehension tasks) should not be expected. In another fMRI study, Chee, Caplan, Soon, Sriram, Tan, Thiel, and Weekes (1999) reported two experiments of proficient English-Chinese early bilinguals asked to comprehend sentences written in English and Chinese. Chee et al. found no differences for areas activated in processing the two languages, despite the marked differences in orthographies and phoneme-grapheme relationships between the languages. In a second study, Chee, Tan, and Thiel (1999) studied Chinese-English bilingual speakers, this time 15 early bilinguals who had been exposed to both languages before the age of six, and nine late bilinguals who learned English after age 12. They employed a word-completion task with stimuli being presented on the screen and participants being asked to silently complete wordstems in English or compound words in Mandarin. Group results suggested great similarity between the brain activation seen when Mandarin words were the stimuli and when English stimuli were used. Substantial left hemisphere activation was reported, including interior frontal gyrus, supplementary motor area, as well as bilateral occipital and parietal regions; some participants showed posterior temporal or fusiform-gyrus activation. Again, the regions of peak activation were quite similar for both languages, coinciding completely for 20 out of the 24 participants. The size of the area activated, however, was greater for some participants in either the first or the second language, but age of acquisition did not predict which was greater. The authors emphasized the lack of differentiation evidenced in L2 among their participants despite the differences between the two languages used and the different ages of participants’ L2 learning. We note that individual differences may have been operating in that four participants showed somewhat different peak activation loci, and patterns of greater activation were seen for some participants in L1, and for others in L2. Results from an fMRI study by Dehaene et al. (1997) suggest that it is indeed intersubject variability that characterizes L2 processing. That is, L1 activates similar brain areas across listeners, whereas for the processing of L2, the cortical regions recruited vary considerably from participant to participant and are not restricted to the classic language areas. Dehaene et al. studied eight participants as they listened to excerpts from stories in French (L1), in English (L2), and in Japanese (unfamiliar language) played in reverse (so that the researchers could subtract out any brain activity that reflected the auditory processing of nonreal speech). French was the participants’ mother tongue, and they had The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

426


not been exposed to English until after age seven, when it was taught in school. By the time of the experiment they had achieved a moderate level of proficiency in English, as was verified in sentence comprehension tests, word translation tests, and in their responses to difficult factual questions about the stories they heard in English. The results demonstrated that listening to stories in L1 (French) activated sites near Wernicke’s area in the left hemisphere of all eight listeners. Listening to L2 activated an extremely variable network of right and left frontal regions. Six participants showed activation in the left temporal pole as well. However, it was more dispersed than the more concentrated localization that was found in the superior temporal sulcus for L1-listening. Two participants showed activation in the right temporal lobe with none in the left temporal region while listening to L2. For other participants, when there was left temporal activity for L2, the volume of activation was generally smaller than it was for L1. Thus, L1 consistently activated similar regions of the left temporal lobe, whereas listening to L2 activated a variable network of right and left frontal regions, varying from classic left lateralization to, surprisingly, exclusively right lateralization. The authors suggested that the frontal activation found in the L2-listening task might be due to the internal rehearsal of English words in working memory during processing. Also, they point out, some of the areas recruited for L2 only are necessary for the attention needed in order to process a second language. L1, which is more automatic, may not need to recruit these resources. Moreover, in the Dehaene et al. study, the results were remarkably consistent across participants for L1. All of the participants displayed activity along the left superior temporal sulcus and middle and superior temporal gyri, and most showed activity in the temporal pole and in the angular gyrus. Similar activity was sometimes found in the right temporal regions, though weaker and more variable. Outside of the temporal pole, the only consistent activity for L1 was found in the inferior frontal sulcus and precentral gyrus (found in the right hemisphere for three participants and in the left hemisphere for six participants). Like Chee and his colleagues (with proficient Chinese-English bilinguals), Hernandez, et al. (2001) found no significant differences between the patterns of activation during naming in the two languages of early English-Spanish bilinguals. Their fMRI study showed common activation for both languages in left dorsolateral prefrontal cortex and in Broca’s area. Furthermore, returning to the questions regarding the mechanisms that allow bilingual speakers to switch appropriately between their languages, Hernandez et al. also compared the patterns of activation when stimuli were presented in language-specific lists versus lists that contained items from both languages. They found significant increase of activation in the mixed-list conditions, particularly in the left inferior frontal gyrus as well as in the right dorsolateral prefrontal cortex. Focusing on semantic processing in bilinguals, Illes, Francis, Desmond, Gabrieli, Glover, Poldrack, Lee, and Wagner (1999) presented Spanish and English words to highly proficient Spanish-English bilinguals who had been exposed to their second language at age five or later (seven out of the eight participants had learned L2 after age nine). They asked them to respond via a pneumatic squeeze ball when they heard concrete or abstract words (the semantic task; half the group responded to concrete and half to abstract) and, in a second task, to respond to upper case or lower case words. While more activation was The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


427

seen for all eight participants on semantic tasks than on the nonsemantic tasks, no differences were evident between the languages of the bilinguals. A standard correction employed in fMRI studies to prevent overinterpretation is to focus only on specific hypothesized areas of interest. Such an ROI (regions of interest) analysis in the Illes et al. study verified their findings when they contrasted the inferior frontal gyri in the left and the right hemispheres. Illes et al. observe that their findings of no differences were consistent with those of the other semantic studies (e.g., Chee, Tan, & Thiel, 1999; Klein et al., 1995), while it is the studies of other language-processing tasks (such as silent production as in Kim et al., 1997) that evidenced differences between brain areas involved in the languages of bilinguals. With the increasing number of brain-imaging studies, researchers are not only asking whether languages in a polyglot may be partially separately represented, but also why L2 might recruit some different regions of the brain from those responsible for L1. The results from Dehaene et al. (1997) suggest that acquisition of L1 relies for the most part on dedicated processing areas in the left hemisphere, whereas late acquisition of L2 may not be associated with a predictable biological substrate. The authors hypothesize that as participants become more proficient in their L2, activation will become progressively more concentrated in the language network of the left hemisphere. In sum, the fMRI literature has yet to take full advantage of the possibilities the technique promises in order to document the course of language processing in bilinguals. While several studies found no differences between activation patterns in the two languages of bilinguals (Chee, Caplan et al., 1999; Chee, Tan, & Thiel, 1999; Hernandez et al., 2001; Illes et al., 1999), it appears that there is somewhat differential processing for L2 as compared to L1 for many tasks — perhaps less for semantic ones — and that there is more variability for L2 processing across participants than for L1 (Dehaene et al., 1997). In addition, level of L2 proficiency may interact with age of L2 learning and with task demands in determining the results observed. 5.4 Evoked Response Potentials

In Evoked Response Potential (ERP) studies, researchers associate changes in voltage of brain waves with linguistic activity; ERPs extracted from the continuous electroencephalogram (EEG) by averaging the signal during a stimulus presentation (see, e.g., Vaughan & Arezzo, 1988). This technique differs from the imaging techniques discussed above in that it primarily provides temporal information about brain electrical activity in response to stimuli over the time immediately ensuing after presentation, rather than specifying the location of processing (although more recent ERP techniques may, with appropriate statistical manipulations, provide localization information as well). For neurolinguists, who continue a history of over a century of looking for localization of function in the brain, temporal processing is more difficult to interpret; for those who worked with response-time measures in psycholinguistics, it may be more meaningful. The standard technique of taking ERP measures is like that of EEGs. Because cells communicate with each other via electrical changes that may be facilitated by a chemical environment, reading these electrical changes can tell us about intercommunication among brain cells. In an early ERP study of bilingual French and English speakers, Meuter, Donald and Ardal (1987) had participants read sentences presented in one language or the other, The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

428


one word at a time. The ERP technique often relies on presenting incongruent words at some point in the sentences, and contrasting participants’ electrical surprise-responses (when sentences were incongruent) to their responses to plausible sentences. A characteristic electrical response approximately 400 ms after the presentation of an incongruent stimulus item (N400) suggests that the incongruity has been appreciated. In this study, EEGs were obtained from six sites: two parietal, two frontal, and two over the midline between the two hemispheres. As in the fMRI studies from the lab of Dehaene discussed above, no differences were seen in response to first-language stimuli for each bilingual group, but differences did obtain for the second language. While N400 effects were obtained for incongruent sentences in participants’ L2, participants for whom L2 was English were more likely to show larger left-sided responses to incongruent words in English than were participants whose L2 was French. Meuter et al. speculated that this asymmetry between the groups was evident because participants were selected from a predominantly Englishspeaking area and thus, as a group, regularly used more English than French. In a follow-up study, Ardal, Donald, Meuter, Muldrew and Luce (1990) again observed N400 latencies among bilingual and monolingual speakers in response to incongruous sentence-final words. They collected data from 24 monolinguals and 24 proficient bilinguals of whom 12 were early learners and 12 were late learners, using congruent and incongruent sentences in participants’ L2 (English or French). Interestingly, despite the fact that the scores on recognition and recall tasks were no worse in the bilinguals than in the monolinguals, the monolinguals showed the earliest of the three N400 latency scores, on average. The bilinguals’ N400 latencies were latest for their second-language; first-language N400 latencies of the bilinguals were intermediate. Furthermore, there was overall a reduced left frontal negativity for the bilinguals’ L2 compared to their L1. No differences were seen in this study on the basis of the age of second language learning; early and late bilinguals did not differ in their N400 latencies in their L2 nor in their L1. Moreover, it is of interest that bilinguals were seen to respond more slowly to incongruence even in their first language as compared to monolingual controls, suggesting that becoming a bilingual may subtly modify processing of the first language. Whether bilingual participants should be compared to monolingual peers has become a matter of controversy with some scholars proposing that the representation of each of the two languages of a bilingual is different from the representation of that language in a monolingual speaker (e.g., Grosjean, 1998). Consistent with this approach, recent findings from psycholinguistic studies suggest that both languages may be activated even while bilinguals are engaged in tasks such as lexical decision in apparently monolingual conditions (e.g., Dijkstra, Timmermans, & Schriefers, 2000). In order to contrast such semantic measures to syntactic ones, Weber-Fox and Neville (1996) created stimuli that included syntactic anomalies as well as ones that included semantic ones. Because they were interested in age-of-L2-acquisition effects, they selected their Chinese-English bilingual participants in groups by their age of L2 acquisition: 1 – 3, 4 – 6, 7 – 10, 11 – 13, and 16 + . There was an age-of-L2-acquisition-related linear decrease in the response to syntactic incongruities, whereas for the semantic ones, all participants who had acquired their L2 before age 11 performed like the monolinguals. More recently, Hahne and Friederici (2001) also found differences in syntactic, but not semantic, processing between native and non-native speakers of German. They tested The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


429

native speakers of Japanese who learned German, their L2, in adulthood. Participants listened to German passive sentences in four conditions: correct, semantically incorrect, syntactically incorrect, and semantically and syntactically incorrect. Hahne and Friederici found that the ERPs collected during sentence processing in these late learners of L2 differed from those of native speakers of German (reported in an earlier paper), especially in the correct and the syntactic-violation conditions. Specifically, correct sentences elicited later activity and greater positivity for non-natives and, while semantic violations elicited similar N400 effect in both groups, syntactic violations elicited no significant effects for the non-natives as compared to correct sentences in contrast to the significant effects found for native speakers. The lack of significant effects in the syntactically incorrect condition for the non-native speakers was evident despite participants’ above-chance performance on the violation-detection (behavioral) task in that condition (note that only correct responses were included in the ERP averages). Comparing ERP patterns during semantic priming in L1 and L2, two additional studies reported no differences between the two languages of proficient bilinguals. Kotz (2001) collected ERPs from native speakers of Spanish who were near-native in their L2, English, learned in early childhood. She employed associated priming in within- and betweenlanguage conditions and found no differences in the ERP data between L1 and L2. Similar results were reported in Bruijn, Dijkstra, Chewilla, and Schriefers (2001) for native speakers of Dutch with English as their L2, who performed a visually presented word-recognition task in both languages. Bruijn et al. found the expected N400 effect for their priming condition in both languages and smaller N400 amplitudes (i.e., smaller electrical changes) for related items, regardless of the language of the prime preceding the target word. The ERP literature thus suggests that bilinguals do not differ from monolinguals in processing semantic information (Hahne & Friederici, 2001; Kotz, 2001; Weber-Fox & Neville, 1996), but show smaller or no effects during syntactic processing (Hahne & Friederici, 2001; Weber-Fox & Neville, 1996). In addition, language proficiency appears to affect latency and amplitude of ERPs obtained in bilinguals’ second language (Ardal et al., 1990; Weber-Fox & Neville, 1996). It is unclear whether the differences between semantic and syntactic conditions that emerge from the ERP studies described here are due to inherent differences between semantic and syntactic processing or to methodological differences (such as level of L2 proficiency and task demands). As in the fMRI and PET literature, variables such as task, modality, material, level of L2 proficiency, and current use of L2 may account for some inconsistencies in the results reported in the ERP studies. In summary, while the imaging results are not fully convergent, the overall conclusion one can tentatively make today is that second language representation and processing may well be somewhat differently organized across individual participants and across groups, more than are first language representation and processing. Variables such as language proficiency and the linguistic skill tested play a crucial role in the results obtained; based on the findings of recent studies it can be hypothesized that language processing and representation change with increased proficiency and use of and may differ for semantic versus syntactic processing. All of these imaging techniques present strikingly handsome pictures by computerized statistical manipulations of data. Despite the compelling nature of these pictures, and The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

430


the observation that more studies that seem consistent with what we already know about L1 organization in the brain tend to get published, it is worth remembering that there remain several problems with the imaging techniques and the interpretation of their data. For example, multiple manipulations of data are required to engender such pictures, and to decide what threshold above baseline to count as a meaningful difference. Also, researchers need to sort out responses to stimuli over potential artifacts and processes that may contribute to the results obtained. Moreover, the interpretation of increased and decreased levels of activation of a specific brain region is still controversial. These unresolved issues render strong interpretations of the findings from these brain-imaging studies premature at this stage of technological development.

6 Future research Despite the many years of neurolinguistic research on bilingualism, the phenomena that have been addressed are only a subset of the phenomena behavioral scientists have identified as being of interest. The aphasia case-study literature focuses primarily on the factors that predict which language will return first in differential recovery, with some discussion of hemispheric lateralization in the bilingual linked to the question of crossed aphasia, and some focus on resources available after brain damage precluding production of one or the other language. In the lateral-dominance literature studying neurologically intact individuals, the question has been entirely whether the left hemisphere is equally dominant for two languages, as a number of the studies find, or whether there is more right hemisphere participation, for bilinguals and / or for L2, at least in the early stages of second language learning, as other studies have reported. In the imaging literature, the focus has been on the degree of overlap versus separation of the two languages within the left hemisphere language areas, and where it occurs (e.g., whether and where in frontal and temporal lobes differences are seen while participants are listening to stories or producing internal speech in L1 or L2). In recent years, additional questions focusing on specific aspects of language processing, such as semantic versus syntactic processing, have been addressed. Several topics that warrant further research can be identified: 1.

The brain representations of the bilingual or polyglot lexicon(s) This has been a substantial focus of psycholinguistic research in the past decade, among such researchers as Kroll, de Groot, de Bot, Dijkstra, Schreuder and Weltens (e.g., de Bot, Cox, Ralston, Schaufeli, & Weltens, 1995; Kroll & Dijkstra, 2002; Kroll & de Groot, 1997; Nicol, 2001; Schreuder & Weltens, 1993), revitalizing an area of study that appeared to have reached a dead end in the 1960 s. Eventually we will want to have explanations for the research that suggests differential processing involved in accessing the structural lexicon and semantics of items in the first and second languages, and differences in directional translation between them (i.e., L1 to L2 translation vs. L2 to L1). As noted above, some recent imaging studies (e.g., Klein et al., 1995; Kotz, 2001; Price et al., 1999) have pursued this approach via comparison of activation during different lexical tasks.

The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


2.

431

Qualitative differences between bilingual and monolingual processing Indications that somewhat different systems are employed by the bilingual and monolingual have arisen across various kinds of behavioral research. For example, a voice-onset-time study of Hebrew-English bilinguals (Obler et al., 1982) suggested distinct systems for the two languages for production (but a single system for comprehension) that were, while in the directions of the monolinguals’ systems, not coterminous with them. Ransdell and Fischler (1987) have reported longer response times for proficient polyglots compared to monolinguals on list-recognition and lexical-decision tasks despite equally accurate performance. Ardal et al. (1990) also reported later ERP N400 latencies in proficient bilinguals than in monolinguals, even in the bilinguals’ L1. Bergman’s study of comprehension of distorted language in elderly participants, too, suggested that while proficient bilinguals could process speech of normal speed as well as monolingual elderly individuals, distortions introduced to the speech (both speeding it up and slowing it down) impaired their comprehension more than it did that of the monolinguals (Bergman, 1980). What these qualitative differences in processing tell us about brain representation and processing in bilinguals will be interesting to discover. Indeed, Grosjean (1989, 1997) emphasized the importance of not restricting bilingualism research to quantitative differences in comparing performance of monolinguals and bilinguals and exploring qualitative differences that may exist between monolingual and bilingual / multilingual language organization and processing.

3.

Language choice and codeswitching Codeswitching has been studied in bilingual populations and attempts have been made to account for bilinguals’ ability to separate their languages or switch between them appropriately (e.g., de Bot, 1992; Green, 1986; Hyltenstam & Stroud, 1989; de Santi, Obler, Sabo-Abrahamson, & Goldberger, 1990) as well as to account for the rules that might govern typical and acceptable switches (e.g., Poplack, 1980; Wei, Milroy, & Pong, 1992). Systematic evaluation of multilinguals’ ability to maintain rule-based language mixing, appropriate language choice, and appropriate language switching in the presence of aphasia, dementia, and other neuropathologies can employ discourse analysis methods such as the conversation analysis approach (e.g., Friedland & Miller, 1999), psycholinguistic experiments (e.g., Grosjean, 1997; Meuter & Allport, 1999), as well as neuroimaging techniques (e.g., Price et al., 1999). Such studies can shed light on the relation between these specific abilities and their underlying neural mechanisms.

4.

Differences in brain organization depending on how many languages one knows Throughout this paper we, like most authors in the literature, have been ignoring potential qualitative differences between bilinguals and trilinguals or polyglots. Yet a growing number of studies in trilingualism suggest that there are qualitative differences at least in the process of learning a second versus a third or fourth language, and possibly in the representation of additional languages (e.g., Cenoz & Genesee, 1998; Cenoz, Hufeisen, & Jessner, 2001; Hoffmann, 1999). Further research comparing bilinguals and multilinguals should contribute to our understanding of the representation of multiple languages in the brain. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

432

5.


Attrition We have little sense of the underlying brain bases of attrition, although, as mentioned earlier, it is certainly documented in children who stopped using a language before a certain period (Olshtain & Barzilay, 1991), as well as in adults (e.g., Pietilä, 1989; Schmid, 2002). Indeed the finding that the rule of Pitres predicts aphasia recovery at significantly greater than chance level for adult aphasics as a group (Obler & Albert, 1977) is also consistent with the notion of attrition of less-used languages. Yet, consistent with Ribot’s notion that a language an individual is exposed to early in life may reveal itself in adulthood even when that language has ceased to be spoken for several years, anecdotal evidence suggests that, at least across a substantial portion of the life-span, reimmersion in an attritted language can result in relatively fast reaccess to it. Further support of Ribot’s theory comes from the speech perception literature. Tees and Werker (1984) studied participants living in North America who were exposed to Hindi for only the first year or two of life and had no further exposure to it. When they were tested on a Hindi voicing contrast (unvoiced aspirate dental stop vs. breathy voiced dental stop) and a place contrast (unvoiced unaspirated retroflex vs. dental stop) in adulthood, their perceptual differentiation abilities were far superior to those of native English speakers with no exposure to Hindi at an early age, and superior to those of most native English listeners with training in Hindi. Thus it appears that contrasts acquired early in life are maintained, even when they are no longer relevant to the ambient languages. What underlying brain-based structures support these phenomena would be of substantial interest. The intensive nature of such longitudinal work, of course, has prevented it from being conducted using the current technologies. Determining which aspects of language-specific knowledge do attrite and which do not, and how these interact with the age at which the individual ceased to use a given language, remain to be explored.

6.

Cross-language differences in processing and organization There is a substantial literature in neurolinguistics looking at the differential impairment and processing of languages that have markedly more or less of certain linguistic structures (e.g., MacWhinney, Bates, & Kliegl, 1984; Menn & Obler, 1990). Potentially, knowledge of a pair of languages differing along certain linguistic features could be differently organized from that of a pair of languages differing on other features. Moreover, languages with fewer features in common could be differently represented and processed from languages that have more features in common. Fabbro (1999) concluded that the few studies demonstrating similar processing across structurally different languages argue against such a possibility. Along with MacWhinney, Bates, and Kliegel (1984) and Rosselli, Ardila, and Pinzon (in press), we would maintain that certain similarities will obtain but differences should well be found as well, analogous to those evidenced in cross-language studies of that form of aphasia called agrammatism in which the speech produced is “telegraphic” with functors and affixes omitted or substituted for.



7.

433

Individual differences in brain organization Using the fMRI technique, as mentioned above, allows researchers to observe individual performance in brain organization. Recall that Kim et al. (1997) found such differences for the second language only, in the earlier, less-proficient stages of second language acquisition. Little other neurolinguistic work has focused in this area, however. Yet a substantial body of literature in the discipline of second-language acquisition has considered individual differences in second-language-learning ability, investigating for differences beyond motivation and opportunity that may predict who will be talented and who will be less talented, or distinctly untalented, language learners (Obler & Menn, 1982). Carroll and Sapon (1959) in constructing the Modern Language Aptitude Test, derived factors such as syntactic sensitivity, phonological memory, and sound-symbol association that differ across individuals and are linked to success in L2 learning. Papagno and Vallar (1995) for example, confirm that good phonological memory is linked to good foreign language acquisition. A number of us have applied aspects of the Geschwind and Galaburda (1985a, b, and c) theory of general talent as it may be linked to dyslexia and other neurological, immunological, or endocrinological differences to explain talented or untalented second language acquisition. Humes-Bartlow (1989), for example, demonstrated that particularly poor second language-learning children are particularly good at arithmetic and visual construction. A number of other Geschwind and Galaburda factors have also been linked to particularly talented second language learners (e.g., Novoa, Fein, & Obler, 1988; Schneiderman & Desmarais, 1988). In a recent study of sizable populations, Lamm and Epstein (1999) report that left-handed adolescents are over-represented among poor L2 learners, consistent with this theory. Indeed, while studies of language development in monolinguals pay attention to the different abilities among monolinguals, particularly on the lower end of the scale where Specific Language Impairment (SLI) has been documented, there has been little carry-over of this approach in the neurolinguistic literature. Because SLI is presumably linked to subtle differences in the language learning and processing underlying neurological structure, it warrants extension to the L2 acquisition and bilingual development literature. Some work on dyslexia and difficulties with L2 learning has been undertaken by such authors as Ganshow and Sparks and their colleagues (e.g., Sparks, Ganshow, Artzer, & Patton, 1997). One wonders how the study of language differences relates to the controversial discussion of the phenomenon called “semilingualism.” Among those not exposed to high academic levels of language outside of school, “semilingual” bilingual individuals who are alleged to have equal functional communication in both languages but nothing more sophisticated might also be addressed from a neurolinguistic perspective. We recently learned of a related but somewhat different population: adolescent children from educated-middle-class-immigrant families, who have grown up immersed in a second language environment mastering their home language, yet mastering only the social-interactional uses of the language of school and environment but not the higher-level academic ones. Moreover, some retain their immigrant parents’ accent, unusually, despite growing up in the environment of the new language. Such exceptional individuals clearly warrant neurolinguistic study, in order to underThe International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

434


stand what is different about them from others in the same situations who acquire the second language with relative ease. These individuals may constitute an even more specifically language-impaired group than do SLI children.

7 Conclusion In addressing our questions about the organization of several languages in the brain, we have looked at findings in the literature of aphasia, lateral dominance, cortical stimulation, and brain imaging. From the aphasia literature we can conclude that polyglots’ first language may not be their most resilient language. In fact, it appears that when not all languages are accessible to the speaker post brain damage, the language most used around the time of the insult will be the first to recover, whether or not it is the mother tongue. However, these findings should be viewed with caution since so little information is available regarding the ways in which the languages have been used. It is interesting to speculate why “recency” seems better to predict recovery than “primacy” and why this appears not to be true for the oldest aphasics. We do suspect, as mentioned above, that the explanation is linked to the mechanisms underlying language attrition in non-brain-damaged individuals who no longer use a particular language. Lack of full information about language use, language proficiency, the material used, and long-term recovery patterns plagues most reports of polyglot aphasia and hinders our ability to find conclusive answers to questions of polyglot language representation based on the aphasia literature alone. It is now evident that most cases of polyglot aphasia are those in which patients show the same deficits and the same restitution patterns for multiple languages. These more typical case studies (which are generally less likely to get published than the dramatic cases of differential recovery) are consistent with the picture seen in the brain-imaging studies of bilinguals, that is, that of substantial overlap of the regions involved in processing the two languages (e.g., Abutalebi et al., 2001; Illes et al., 1999). It might appear that the numerous cases of differential recovery of languages in aphasics contradict these imaging findings. However, nonoverlap activation of L1 and L2 has been found in several imaging studies, particularly for syntactic processing, some language-comprehension tasks, and when the bilingual speakers were not equally proficient in both their languages (e.g., Chee, Caplan et al., 1999; Chee, Tan, & Thiel, 1999; Hahne & Friederici, 2001; Perani et al., 1998; Price et al., 1999). These reports of differential activation are consistent with cases of polyglot aphasia for which one language is impaired and others are not. The rarity of such cases probably results from the relatively small extent of nonoverlapping areas in gross brain substrates for the two languages, and the relatively larger size of lesions generally resulting in aphasia. An apparent contradiction might also be perceived between the findings of laterality literature and the imaging literature. While both sets of articles indicate substantial overlap of brain substrate for the bilinguals’ languages within the traditional left-hemisphere peri-Sylvian language area, the laterality literature hints at additional right-hemisphere participation. Often, the imaging literature considers only the left hemisphere; however, when right-hemisphere regions were examined, some speakers demonstrated right-hemisphere involvement in L2 processing (e.g., Dehaene et al., 1997). Moreover, the imaging literature is able to be more precise than the laterality literature in its intrahemispheric detail The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


435

and has revealed that different regions within the left hemisphere, particularly anterior areas, may be involved in processing L1 and L2. Clearly, we conclude, a critical phenomenon for investigators to pursue is the substantial variation across participants in the representation of L2 (e.g., Dehaene et al., 1997; Perani et al., 1996). We maintain that linking such differences to structures of the specific languages in question and to specific linguistic skills, considering proficiency, recent-use habits, and handedness-linked genetic differences in brain organization for language generally, will prove to be most productive. Our understanding of the neurolinguistics of bilingualism will benefit from converging evidence from case and group studies using careful language analysis, detailed language history, and brain imaging data. Received: July, 2000; revised: April, 2002; accepted: June, 2002

References ABUTALEBI, J., CAPPA, S. F., & PERANI, D. (2001). The bilingual brain as revealed by functional neuroimaging. Bilingualism: Language and Cognition, 4(2), 179 – 190. AGLIOTI, S., BELTRAMELLO, A., GIRARDI, F., & FABBRO, F. (1996). Neurolinguistic followup study of an unusual pattern of recovery from bilingual subcortical aphasia. Brain, 119, 1551 – 1564. AINE, C. (1995).A conceptual overview and critique of functional neuroimaging techniques in humans: MRI / fMRI and PET. Critical Reviews in Neurology, 9, 229– 309 ALBERT, M. L., & OBLER, L. K. (1978). The bilingual brain: Neuropsychological and neurolinguistic aspects of bilingualism. New York: Academic Press. ALEXANDER, M. P., & ANNETT, M. (1996). Crossed aphasia and related anomalies of cerebral organization: Case reports and a genetic hypothesis. Brain and Language, 55, 213 – 239. APRIL, R., & HAN, M. (1980). Crossed aphasia in a right-handed bilingual aphasic man. Archives of Neurology, 37, 341– 346. ARDAL, S., DONALD, M. W., MEUTER, R., MULDREW, S., & LUCE, M. (1990). Brain responses to semantic incongruity in bilinguals. Brain and Language, 39, 187 – 205. BAY, E. (1964). General discussion. In A. V. S. De Reuck & M. O’Connor (Eds.), Disorders of Language. Boston: Little Brown. BERGMAN, M. (1980). Aging and the perception of speech. Baltimore: University Park Press. BHATNAGER, S. (2002). Neuroscience for the study of communicative disorders. Baltimore: Lippincott Williams & Wilkins. BLUMSTEIN, S., GOODGLASS, H., & TARTTER, V. (1975). The reliability of ear advantage in dichotic listening. Brain and Language, 2, 226– 236. BROCA, P. (1861). Remarques sur le siègede la faculté de la parole articulé, suivies d’une observation d’aphémie (perte de parole). Bulletin de la Société d’Anatomie, 36, 330– 357. BROCA, P. (1865). Sur le siège de la faculté du language articulé. Bulletin d’Anthropologie, 6, 377–393. BRUIJN, E. R. A., DIJKSTRA, T., CHEWILLA, D. J., & SCHRIEFERS, H. J. (2001). Language context effects on interlingual homograph recognition, evidence form event-related potentials and response times in semantic priming. Bilingualism: Language and Cognition, 4(2), 155–168. CARROLL, J., & SAPON, S. (1959). Modern language aptitude test manual. New York: The Psychological Corporation. CENOZ, J., & GENESEE, F. (Eds.). (1998). Beyond bilingualism: Multilingualism and multilingual education. Clevedon: Multilingual Matters. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

436


CENOZ, J., HUFEISEN, B., & JESSNER, U. (2001). Toward trilingual education. International Journal of Bilingual Education and Bilingualism, 4(1), 1 – 10. CHARLTON, M. (1964). Aphasia in bilingual and polyglot patients— a neurological and psychological study. Journal of Speech and Hearing Disorders, 29, 307– 311. CHARY, P. (1986). Aphasia in multilingual society: A preliminary study. In J. Vaid (Ed.), Language processing in bilinguals: Psycholinguistic and neuropsychological perspectives. Hillsdale, NJ: Erlbaum. CHEE, M. W. L., CAPLAN, D., SOON, C. S., SRIRAM, N., TAN, E. W. L., THIEL, T., & WEEKES, B. (1999).Processing of visually presented sentences in Mandarin and English studied with fMRI. Neuron, 23, 127 – 137. CHEE, M. W. L., TAN, E., & THIEL, T. (1999). Mandarin and English single word processing studied with functional magnetic resonance imaging. The Journal of Neuroscience, 19(8), 3050 – 3056. De BOT, K. (1992). A bilingual production model: Levelt’s “speaking” model adapted. Applied Linguistics, 13, 1 – 24. De BOT, K. (1996). Language loss. In H. Goebl, P. H. Nelde, Z. Stary & W. Wolck (Eds.), Contact lingustics: An introductional handbook of contemporary research. Vol. 1. Berlin / New York: Walter De Gruyter. De BOT, K. COX, A., RALSTON, S., SCHAUFELI, A., & WELTENS, B. (1995). Lexical processing in bilinguals. Second Language Research, 11(1), 1 – 19. DEHAENE, S., DUPOUX, E., MEHLER, J., COHEN, L., PAULESU, E., PERANI, D., MOORTELE, P.-F. van de, LEHERICY, S., & LE BIHAN, D. (1997). Anatomical variability in the cortical representation of first and second language. NeuroReport, 8, 3809 – 3815. De SANTI, S., OBLER, L. K., SABO-ABRAHAMSON, H., & GOLDBERGER, J. (1990). Discourse abilities and deficits in bilingual dementia. In Y. Joanette & H. H. Brownell (Eds.), Discourse ability and brain damage: Theoretical and empirical perspectives. New York: Springer. DIJKSTRA, A., TIMMERMANS, M., & SCHRIEFERS, H. (2000). On being blinded by your other language: Effects of task demands on interlingual homograph recognition. Journal of Memory and Language, 42, 445 – 464. FABBRO, F. (1999). The neurolinguistics of bilingualism: An introduction. East Sussex, United Kingdom: Psychology Press. FABBRO, F. (2001). The bilingual brain: Cerebral representation of languages. Brain and Language, 79, 211 – 222. FABBRO, F., SKRAP, M., & AGLIOTI, S. (2000). Pathological switching between languages after frontal lesions in a bilingual patient. Journal of Neurology, Neurosurgery, and Psychiatry, 68, 650 – 652. FARAH, M. J., & McCLELLAND, J. L. (1991). A computational model of semantic memory impairment: Modality-specificity and emergent category specificity. Journal of Experimental Psychology: General, 120(4), 339 – 357. FREUD, S. (1891). Zur Auffassung der Aphasie. [Translated as On Aphasia.] Freud, S. (1891). In M. Paradis (Ed.), Readings on aphasia in bilinguals and polyglots. Montreal: Marcel Didier. FRIEDLAND, D., & MILLER, N. (1999). Language mixing in bilingual speakers with Alzheimer’s dementia: A conversation analysis approach. Aphasiology, 13, 427 – 444. GALLOWAY, L. M., & SCARCELLA, R. (1982). Cerebral organization in adult second language acquisition: Is the right hemisphere more involved? Brain and Language, 15(2), 10 – 233. GESCHWIND, N., & GALABURDA, A. M. (1985a). Cerebral lateralization: Biological mechanisms, associations and pathology: 1. A hypothesis and a program for research. Archives of Neurology, 42, 428 – 459. GESCHWIND, N., & GALABURDA, A. M. (1985b). Cerebral lateralization: Biological mechanisms, associations and pathology: 2. A hypothesis and a program for research. Archives of Neurology, 42, 521 – 552. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


437

GESCHWIND, N., & GALABURDA, A. M. (1985c). Cerebral lateralization: Biological mechanisms, associations and pathology: 3. A hypothesis and a program for research. Archives of Neurology, 42, 634 – 654. GLONING, I., & GLONING, K. (1965). Aphasien bei polyglotten. Wiener Zeitschrift für Nervenheilkunde, 22, 362 – 397. GOODGLASS, H. (1993). Understanding aphasia. Boston: Academic Press. GORAL, M. (2001). Language processing and language proficiency of trilingual speakers. Unpublished doctoral dissertation, The Graduate School and University Center of the City University of New York, New York. GREEN, D. W. (1986). Control, activation, and resource: A framework and a model for the control of speech in bilinguals. Brain and Language, 27, 210– 223. GREEN, D. W. (1998). Mental control of the bilingual lexico-semantic system. Bilingualism: Language and Cognition, 1, 67 – 81. GROSJEAN, F. (1989). Neurolinguists, beware! The bilingual is not two monolinguals in one person. Brain and Language, 36, 3 – 15. GROSJEAN, F. (1997). Processing mixed languages: Issues, findings, and models. In A. M. B. de Groot & J. F. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives. Mahwah, NJ: Lawrence Erlbaum. GROSJEAN, F. (1998). Studying bilinguals: Methodological and conceptual issues. Bilingualism: Language and Cognition, 1, 131– 149. HAGLUND, M. M., OJEMANN, G. A., LETTICH, E., BELLUGI, U., & CORINA, D. (1993). Dissociation of cortical and single unit activity in spoken and signed languages. Brain and Language, 44, 19 – 27. HAHNE, A., & FRIEDERICI, A. D. (2001). Processing a second language: Late learners’ comprehension mechanisms as revealed by event-related potentials. Bilingualism: Language and Cognition, 4(2), 123– 141. HERNANDEZ, A. E., DAPRETTO, M., MAZZIOTTA, J., & BOOKHEIMER, S. (2001). Language switching and language representation in Spanish-English bilinguals: An fMRI study. NeuroImage, 14, 510– 520. HOFFMANN, C. (1999). Trilingual competence: Linguistic and cognitive issues. Applied Linguistic Studies in Central Europe, 3, 16 – 26. HUMES-BARTLOW, M. (1989). Variation in children’s ability to learn second languages. In K. Hyltenstam & L. K. Obler (Eds.), Bilingualism across the lifespan. Cambridge, U.K.: Cambridge University Press. HYLTENSTAM, K., & STROUD, C. (1989). Bilingualism in Alzheimer’s dementia: Two case studies. In K. Hyltenstam & L. K. Obler (Eds.), Bilingualism across the lifespan: Aspects of acquisition, maturity, and loss (pp. 202– 226). Cambridge, U.K.: Cambridge University Press. ILLES, J., FRANCIS, W. S., DESMOND, J. E., GABRIELI, J. D. E., GLOVER, G. H., POLDRACK, R., LEE, C. J., & WAGNER, A. D. (1999). Convergent cortical representation of semantic processing in bilinguals. Brain and Language, 70, 347. JARED, D., & KROLL, J. F. (2001). Do bilinguals activate phonological representations in one or both of their languages when naming words? Journal of Memory and Language, 44, 2 – 31. KARANTH, P., & RANGAMANI, G. N. (1988). Crossed aphasia in multilinguals. Brain and Language, 34, 169 – 180. KIM, K. S., RELKIN, N., LEE, K. M., & HIRSCH, J. (1997). Distinct cortical areas associated with native and second languages. Nature, 388, 171– 174. KLEIN, D., MILNER, B., ZATORRE, R. J., MEYER, E., & EVANS, A. C. (1995). The neural substrates underlying word generation: A bilingual functioning-imaging study. Proceedings of the National Academy of Sciences, 92, 2899– 2903. KLEIN, D., ZATORRE, R. J., MILNER, B., MEYER, E., & EVANS, A. C. (1994). Left putaminal activation when speaking a second language: Evidence from PET. NeuroReport, 5, 2295– 2297. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

438


KLEIN, D., ZATORRE, R. J., MILNER, B., MEYER, E., & EVANS, A. C. (1995). The neural substrates of bilingual language processing: Evidence from positron emission tomography. In Paradis, M. (Ed.), Aspects of bilingual aphasia. NY: Elsevier Science. KOTZ, S. A. (2001). Neurolinguistic evidence for bilingual language representation: A comparison of reaction times and event-related brain potential. Bilingualism: Language and Cognition, 4(2), 143 – 154. KROLL, J. F., & De GROOT, A. M. B. (1997). Lexical and conceptual memory in the bilingual: Mapping form to meaning in two languages. In A. M. B. De Groot & J. F. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives. Mahwah, NJ: Erlbaum. KROLL, J. F., & DIJKSTRA, T. (2002). The bilingual lexicon. In R. Kaplan (Ed.), Handbook of applied linguistics. Oxford: Oxford University Press. LAMM, O., & EPSTEIN, R. (1999). Left-handedness and achievements in foreign language studies. Brain and Language, 70(3), 504– 517. LEBLANC, R., & ZATORRE, R. (1997). Functional magnetic resonance imaging. Journal of Neurosurgery, 86(2), 316– 317. LEVY, E. S., GORAL, M., & OBLER, L. K. (1999). Neurolinguistic perspectives on mother tongue: Evidence from aphasia and brain imaging. La Langue Maternelle, Cahiers Charles, Vol. 27. Paris: Publication de l’Université de Paris, 7, 141 – 157. MacWHINNEY, B., BATES, E., & KLIEGL, R. (1984). Cue validity and sentence interpretation in English, German, and Italian. Journal of Verbal Learning and Verbal Behavior, 23(2), 127–150. MENN, L., & OBLER, L. K., (Eds.). (1990). Agrammatic aphasia: A cross-language narrative sourcebook. Amsterdam: Benjamins. Three Volumes. MEUTER, R. F. I., & ALLPORT, A. (1999). Bilingual language switching in naming: Asymmetrical costs of language selection. Journal of Memory and Language, 40, 25 – 40. MEUTER, W., DONALD, M., & ARDAL, S. (1987). A comparison of first-and second-language ERPs in bilinguals. Electroencephalography and Clinical Neurophysiology. Supplement, 40, 412 – 416. MINKOWSKI, M. (1927). Klinischer Beitrag zur Aphasie bei Polyglotten, speziell im Hinblick aufs Schweizerdeutsche. Schweizer Archiv für Neurologie und Psychiatrie, 21, 43– 72. [Translated in M. Paradis (Ed.), Readings on aphasia in bilinguals and polyglots (pp. 205– 232). Montreal: Marcel Didier.] NICOL, J. L. (2001). One mind, two languages: Bilingual language processing. Malden, MA: Blackwell. NILIPOUR, R., & ASHAYERI, H. (1989). Alternating antagonism between two languages with successive recovery of a third in a trilingual aphasic patient. Brain and Language, 36(1), 23 –48. NOVOA, L., FEIN, D., & OBLER, L. K. (1988). Talent in foreign languages: A case study. The Exceptional Brain. New York: Guilford Press. OBLER, L. K., & ALBERT, M. L. (1977). Influence of aging on recovery from aphasia in polyglots. Brain and Language, 4, 460– 463. OBLER, L. K., & GJERLOW, K. (1999). Language and the brain. Cambridge, U.K.: Cambridge University Press. OBLER, L. K., & MAHECHA, N. R. (1991). First language loss in bilinguals with brain damage. In H. Seliger & R. Vago (Eds.), First language attrition: Structural and theoretical perspectives. Cambridge, U.K.: Cambridge University Press. OBLER, L., & MENN, L. (Eds.). (1982). Exceptional language and linguistics. New York: Academic Press. OBLER, L. K., ZATORRE, R. J., GALLOWAY, L., & VAID, J. (1982). Cerebral lateralization in bilinguals: Methodological issues. Brain and Language, 15, 40 – 54. OJEMANN, G. A., & WHITAKER, H. A. (1978). The bilingual brain. Archives of Neurology, 35, 409 – 412. OLSHTAIN, E., & BARZILAY, M. (1991). Lexical retrieval difficulties in adult language attrition. In H. Seliger & R. Vago (Eds.), First language attrition. NY: Cambridge University Press. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015


439

PAPAGNO, C., & VALLAR, G. (1995). Verbal short-term memory and vocabulary learning in polyglots. The Quarterly Journal of Experimental Psychology, 48A(1), 98 – 107. PARADIS, M. (1977). Bilingualism and aphasia. In H. Whitaker & H. A. Whitaker (Eds.), Studies in Neurolinguistics, Vol. 3. New York: Academic Press. PARADIS, M. (1989). Bilingual and polyglot aphasia. In F. Boller & J. Graffman (Eds.), Handbook of Neuropsychology, Vol. 2 (pp. 117– 140). Oxford: Elsevier. PARADIS, M. (1996). The cognitive neuropsychology of bilingualism. In A. de Groot & J. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives. Hillsdale, NJ: Erlbaum. PARADIS, M., & GOLDBLUM, M. C. (1989). Selective crossed aphasia in a trilingual aphasic patient followed by reciprocal antagonism. Brain and Language, 36, 62 – 75. PARADIS, M., GOLDBLUM, M. C., & ABIDI, R. (1982). Alternate antagonism with paradoxical translation behavior in two bilingual aphasic patients. Brain and Language, 15, 55 – 69. PENFIELD, W., & ROBERTS, L. (1959). Speech and brain mechanisms. Princeton: Princeton University Press. PERANI, D., DEHAENE, S., GRASSI, F., COHEN, L., CAPPA, S. F., DUPOUX, E., FAZIO, F., & MEHLER J. (1996). Brain processing of native and foreign languages. NeuroReport, 7, 2439– 2444. PERANI, D., PAULESU, E., GALLES, N. S., DUPOUX, E., DEHAENE, S., BETTINARDI, V., CAPPA, S. F., FAZIO, F., & MEHIER, J. (1998). The bilingual brain: Proficiency and age of acquisition of the second language. Brain, 121, 1841– 1852. PERECMAN, E. (1984). Spontaneous translation and language mixing in a polyglot aphasic. Brain and Language, 23(1), 43 – 63. PIETILÄ, P. (1989). The English of Finnish Americans with reference to social and psychological background factors and with special reference to age. Turku: Turun Yliopisto. PITRES, A. (1895). Etude sur l’aphasie chez les polyglottes. Revue de Medecine, 15, 873 – 899. [Translated in M. Paradis (Ed.), Readings on aphasia in bilinguals and polyglots (pp. 26 – 48). Montreal: Marcel Didier.] PLAUT, D. C., & SHALLICE, T. (1993). Deep dyslexia: A case study of connectionist neuropsychology. Cognitive Neuropsychology, 10(5), 377 – 500. POPLACK, S. (1980). Sometimes I’ll start a sentence in Spanish Y TERMINO EN ESPANOL: Toward a typology of codeswitching. Linguistics, 18, 581– 618. PÖTZL, O. (1925). Über die parietal bedingte Aphasie und ihren Einfluß auf das Sprechen mehrerer Sprachen. Zeitschrift für die gesamte Neurologie und Psychiatrie, 96, 100– 124. PRICE, C. J., GREEN, D. W., & STUDNITZ, R. von (1999). A functional imaging study of translation and language switching. Brain, 122, 2221– 2235. RANSDELL, S. E., & FISCHLER, I. (1987). Memory in a monolingual mode: When are bilinguals at a disadvantage? Journal of Memory and Language, 26, 392 – 405. RAPPORT, R. L., TAN, C. T., & WHITAKER, H. A. (1983). Language function and dysfunction among Chinese-and English-speaking polyglots: Cortical stimulation, Wada testing, and clinical studies. Brain and Language, 18, 342 – 366. RINNE, J. O., TOMMOLA, J., LAINE, M., KRAUSE, B. J., SCHMIDT, D., KAASINEN, V., TERÄS, M., SIPILÄ, H., & SUNNARI, M. (2000). The translating brain: Cerebral activation patterns during simultaneous interpreting. Neuroscience Letters, 294, 85 – 88. RIBOT, T. (1881). Les Maladies de la Mémoire (2nd ed.). Paris: Baillère. ROSSELLI, M., ARDILA, A., & PINZON, O. (in press). Alexia and agraphia in Spanish speakers: CAT correlations and interlinguistic analysis. In A. Ardila & F. Ostrosky (Eds.), Brain organization of cognitive activity: New perspectives of research. Plenum Press. SAYKIN, A. J., JOHNSON, S. C., FLASHMAN, L. A., McALLiSTER, T. W., SPARLING, M., DARCEY, T. M., MORITZ, C. H., GUERIN, S. J., WEAVER, J., & MAMOURIAN, A. (1999). Function differentiation of medial temporal and frontal regions involved in processing novel and familiar words: An fMRI study. Brain, 122, 1963– 1971. The International Journal of Bilingualism Downloaded from ijb.sagepub.com at Mina Rees Library/CUNY Graduate Center on February 17, 2015

440


SCHMID, M. (2002). First language attrition, use and maintenance: The case of German Jews in Anglophone countries. In Studies in bilingualism, Vol. 24. Amsterdam/Philadelphia: John Benjamins. SCHNEIDERMAN, E. I., & DESMARAIS, C. (1988). A neuropsychological substrate for talent in second-language acquisition. In L. K. Obler & D. A. Fein (Eds.), The exceptional brain: Neuropsychology of talent and special abilities. New York: The Guilford Press. SCHREUDER, R., & WELTENS, B. (Eds.). (1993). The bilingual lexicon: Studies in bilingualism. Amsterdam/ Philadelphia: John Benjamins Publishing Company. SELIGER, H. W. (1991). Language attrition, reduced redundancy, and creativity. In H. Seliger & R. Vago (Eds.), First language attrition. New York: Cambridge University Press. SERGENT, J. (1994). Brain-imaging studies of cognitive functions. Trends in Neuroscience, 17(6), 221 – 226. SILVERBERG, R., BENTIN, S., GAZIEL, T., OBLER, L. K., & ALBERT, M. L. (1979). Shift of visual field preference for English words in native Hebrew speakers. Brain and Language, 8, 184– 190. SPARKS, R. I., GANSCHOW, L., ARTZER, M., & PATTON, J. (1997). Foreign language proficiency of at risk and not-at-risk learners over two years of foreign language instruction: A follow-up study. Journal of Learning Disabilities, 30, 92 – 98. TEES, R. C., & WERKER, J. F. (1984). Perceptual flexibility: Maintenance or recovery of the ability to discriminate non-native speech sounds. Canadian Journal of Psychology, 38, 579–590. TIERNEY, M. C., VARGA, M., HOSEY, L., GRAFMAN, J., & BRAUN, A. (2001). PET evaluation of bilingual language comprehension following early childhood brain damage. Neuropsychologia, 39, 114– 121. VAID, J. (1983). Bilingualism and brain lateralization. In S. Segalowitz (Ed.), Language functions and brain organization. New York: Academic Press, 315– 339. VAID, J., & HALL, D. G. (1991). Neuropsychological perspectives on bilingualism: Right, left, and center. In A. Reynolds (Ed.), Bilingualism, multiculturalism, and second language learning: The McGill Conference in honor of W. E. Lambert. Hillsdale, NJ: Erlbaum. VAUGHAN, H. G., & AREZZO, J. C. (1988). The neural basis of event-related potentials. In T. W. Picton (Ed.), EEG Handbook. Amsterdam: Elsevier Sciences Publishers. Von MUNDY, V. G. (1959). A 94-year-old with one German language center and probably two Slovenian language centers. [Translated in M. Paradis (Ed.), Readings on aphasia in bilinguals and polyglots (pp. 624– 625). Montreal: Marcel Didier.] WEBER-FOX, C., & NEVILLE, H. (1996). Maturational constraints on functional specialization for language processing: ERP and behavioral evidence in bilingual speakers. Journal of Cognitive Neuroscience, 8, 231– 256. WEI, L., MILROY, L., & PONG, S. C. (1992). A two-step sociolinguistic analysis of codeswitching and language choice. International Journal of Applied Linguistics, 2(1), 63 – 86. WERNICKE, C. (1874). The aphasic symptom complex: A psychological study on a neurological basis. Breslau: Kohn and Weigert. Reprinted in R. S. Cohen & M. W. Wartofsky (Eds.), Boston Studies in the Philosophy of Science, 4. Boston: Reidel. ZATORRE, R. J. (1984). Musical perception and cerebral function: A critical review. Music Perception, 2, 196– 221. ZATORRE, R. J. (1989). On the representation of multiple languages in the brain: Old problems and new directions. Brain and Language, 36(1), 127– 147.
