The Bilingual Mental Lexicon

Interdisciplinary Approaches

Aneta Pavlenko

• Publisher: Multilingual Matters
• Year: 2009
• Print ISBN: 9781847691248
• eBook ISBN: 9781847698926
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

How are words organized in the bilingual mind? How are they linked to concepts? How do bi- and multilinguals process words in their multiple languages? The first aim of this volume is to offer up-to-date answers to these questions. Its second aim is to provide readers with detailed step-by-step introductions to a variety of methodological approaches used to investigate the bilingual lexicon, from traditional neurocognitive and psycholinguistic approaches to the more recent ones that examine language use in context.

Chapter 1

Neurolinguistic Contributions
to Understanding the Bilingual
Mental Lexicon

RENATA MEUTER

Introduction

Many bilinguals will have had the experience of unintentionally reading something in a language other than the intended one (e.g. MUG to mean mosquito in Dutch rather than a receptacle for a hot drink, as one of the possible intended English meanings), of finding themselves blocked on a word for which many alternatives suggest themselves (but, somewhat annoyingly, not in the right language), of their accent changing when stressed or tired and, occasionally, of starting to speak in a language that is not understood by those around them. These instances where lexical access appears compromised and control over language behavior is reduced hint at the intricate structure of the bilingual lexical architecture and the complexity of the processes by which knowledge is accessed and retrieved. While bilinguals might tend to blame word finding and other language problems on their bilinguality, these difficulties per se are not unique to the bilingual population. However, what is unique, and yet far more common than is appreciated by monolinguals, is the cognitive architecture that subserves bilingual language processing. With bilingualism (and multilingualism) the rule rather than the exception (Grosjean, 1982), this architecture may well be the default structure of the language processing system. As such, it is critical that we understand more fully not only how the processing of more than one language is subserved by the brain, but also how this understanding furthers our knowledge of the cognitive architecture that encapsulates the bilingual mental lexicon.

The neurolinguistic approach to bilingualism focuses on determining the manner in which the two (or more) languages are stored in the brain and how they are differentially (or similarly) processed. The underlying assumption is that the acquisition of more than one language requires at the very least a change to or expansion of the existing lexicon, if not the formation of language-specific components, and this is likely to manifest in some way at the physiological level. There are many sources of information, ranging from data on bilingual aphasic patients (Paradis, 1977, 1985, 1997) to lateralization (Vaid, 1983; see Hull & Vaid, 2006, for a review), recordings of event-related potentials (ERPs) (e.g. Ardal et al., 1990; Phillips et al., 2006), and positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) studies of neurologically intact bilinguals (see Indefrey, 2006; Vaid & Hull, 2002, for reviews). Following the consideration of methodological issues and interpretative limitations that characterize these approaches, the chapter focuses on how the application of these approaches has furthered our understanding of (1) selectivity of bilingual lexical access, (2) distinctions between word types in the bilingual lexicon and (3) control processes that enable language selection.

1. Methodological Issues

Studies focusing on the neurophysiological correlates of bilingual language processing often are more concerned with localization and lateralization than with the consideration of how behavioral and physiological data converge to inform our understanding of the bilingual lexicon. (For reviews of select areas see Abutalebi et al., 2005; Fabbro, 2001b; Hull & Vaid, 2005; Vaid & Hull, 2002.) Important insights have been gleaned from the closer observation of bilingual aphasics (Fabbro, 2001a; Paradis, 2004). In addition, a number of studies have provided PET and hemodynamic evidence, as well as ERP correlates, for lexical organization and access (e.g. De Bleser et al., 2003; Phillips et al., 2006), semantic organization (e.g. Halsband et al., 2002) and issues of control (e.g. Rodriguez-Fornells et al., 2006) in the bilingual lexicon. It is important to be mindful of the strengths and limitations of each approach and how they may be optimally applied.

1.1. Bilingual aphasia

Early forays into the bilingual brain intrigued us with descriptions of the many different recovery patterns observed in multilingual aphasia. For example, some patients demonstrated selective recovery of either the first learned or most familiar language (Pitres, 1895; see also Albert & Obler, 1978; Paradis, 1977, 1997). Others showed pathological switching between languages (Fabbro et al., 2000) and selective, antagonistic recovery patterns (Aglioti et al., 1996; Aglioti & Fabbro, 1993). Inferences drawn from aphasic deficits involve a reverse extrapolation to a(n assumed) premorbid state of intact linguistic functioning. However, often little is known about the patient’s language history and consequently a retrospective comparison is potentially flawed. Such information is critically important because age and manner of acquisition, and premorbid language use and proficiency would have affected the patient’s performance in each language post-injury.

Those concerns aside, the seductive aspect of this approach is that patterns of language performance (both intact and impaired) suggest selective damage to (or preservation of) one or more language components or processing pathways. Following a modular approach to cognitive architecture (cf. Fodor, 1983; Shallice, 1988), dissociations in performance can be taken to reflect an underlying cognitive architecture characterized by subcomponents that can be impaired selectively. However, it is difficult to determine the source of the cognitive deficit and which task components are adversely affected, because it is unclear whether the cause is a circumscribed lesion or more extensive damage, or whether the deficit results from abnormal functioning elsewhere or even reflects intact residual functioning in the lesioned area (see also Green & Price, 2001).

When testing bilingual aphasics, the degree of similarity between the languages spoken is often ignored and there is a tendency to use concrete items. Yet we know that concrete and abstract items are processed differently (e.g. De Groot & Nas, 1991; Dong et al., 2005). An explicit comparison of bilingual aphasic performance on such items would yield important insights, and so would contrasting the processing of cognates and noncognates, also known to be differently represented (Sánchez-Casas & García-Albea, 2005). Cognates have been used in attempts to improve language functioning in bilingual aphasics, on the assumption that reinforcing cross-linguistic form and meaning overlap should encourage cross-linguistic generalization (e.g. Kohnert, 2004).

While studies of bilingual aphasia have provided valuable insights into localization of language function, they have been less immediately helpful in informing models of bilingual language processing. Case studies tend to be descriptive and the patterns of language performance observed often are not interpreted within an existing processing framework nor used to test the validity of any such framework (for a more detailed discussion, see Gollan & Kroll, 2001). Green and Price (2001) recommended combining the traditional clinical neurolinguistic approach with functional imaging, thus enabling a direct comparison of normal and aphasic bilinguals. Importantly, the use of neuroimaging techniques reveals not only the regions involved in the execution of components of a task (without, however, being able to determine which regions are essential), but also the interaction between them. When applying this approach to patients, it becomes possible to ascertain (for example) whether the observed deficit is due to differences in neuronal areas recruited for the task or to the use of different cognitive strategies, or whether control of language performance is compromised because of deficits in neural areas believed to mediate such control (such as the anterior cingulate). Limitations aside, several case histories can be interpreted in the light of existing models of the bilingual lexicon and inform those same models.

1.2. ERPs and imaging studies

Valuable contributions have been made by imaging and ERP studies of bilingual language processing. Their comparison, however, is often limited by the heterogeneity of participants (in terms of proficiency, and age and manner of acquisition), the wide range of tasks used, the lack of appropriate comparison groups and the variety of languages spoken. Inconsistencies between PET and fMRI findings in particular may be due also to differences in technology.

In PET studies, data is averaged across participants. Findings are therefore more difficult to interpret and detail is lost in reduced spatial resolution. Thus, if regions of activation potentially correlated with distinct language components are anatomically close, it may not be possible to distinguish between them using PET. With fMRI, on the other hand, individual data can be considered and spatial resolution is superior (see Abutalebi et al., 2005, for a review). A difficulty affecting hemodynamic methods generally is how to interpret increased (differential) activation. Is it just that, activation, or are inhibitory processes occurring elsewhere contributing to the pattern? (See also Fabbro, 2001b.) Furthermore, activation in a certain region does not ipso facto imply that the area is necessary for task execution nor does lack of (differential) activation imply lack of involvement. Subtraction methods may simply obscure the possibility that the same region might be implicated in several of the task components (cf. Green & Price, 2001). However, careful comparisons across tasks can reveal areas that are differentially active. Given existing knowledge about cognitive processes believed to be subserved by those areas and knowledge of the task components, converging evidence may be obtained for the existence of those components.

The particular strength of ERP scalp recordings is that they are online measures of brain activity that is directly (and temporally) linked to cognitive processes and allow the various properties of linguistic stimuli to be indexed. One primary language indicator is the N400, a negative ERP component with a central scalp distribution and a peak amplitude at approximately 400 ms after the appearance of a word. The N400 is used to index aspects of semantic processing in sentence context as well as that of individual words, within and between languages. In response to semantic violations, a larger (i.e. more negative) N400 is typically observed (e.g. Ardal et al., 1990). Lexical aspects, such as word frequency, can also be indexed by N400, which is larger for words of lower frequency (Kutas & Van Petten, 1990). A repetition priming effect is captured in reduced N400 amplitudes (e.g. Rugg et al., 1995). Another measure sometimes used is the phonological mismatch negativity (PMN), an ERP component with an earlier peak between 250 and 350 ms post stimulus-onset and elicited when phonological processing is carried out. Its distribution is more fronto-central (cf. Phillips et al., 2006).

Form and meaning can be varied both within and between languages, and the careful evaluation of ERP components can determine whether these linguistic variations are processed differently. A phenomenon such as adaptation can be applied embedded in repetition priming within and between languages, allowing inferences to be drawn about the level at which words are processed by common or shared cognitive processes. This methodology is predicated on the demonstration that repetition of information results in attenuated neuronal responses. For example, neurons in the visual system reduce their response to repeated occurrences of the same stimulus (adaptation), but might fire as strongly to a repeat stimulus of which just one aspect has been changed (Grill-Spector & Malach, 2001). Useful insights have been gained using this technique (e.g. Phillips et al., 2006).

1.3. Summary

Optimal information comes from converging evidence from different methodologies, especially where cognitive deficit due to a prescribed brain injury dovetails with activation in the same, undamaged region when the task is performed successfully in an intact person. But some caveats remain. Importantly, for these techniques to be maximally effective they should be combined with a thorough understanding of the cognitive underpinnings of the tasks to be performed and the research tools used. Careful evaluation of aphasic performance and use of ERP recordings and imaging techniques within a well-articulated cognitive framework can provide supporting evidence for the existence of distinct cognitive processes underlying bilingual task performance and further our knowledge regarding lexical access, word type distinction and control in the bilingual lexicon.

2. Selectivity of Bilingual Lexical Access

There is general consensus that bilingual lexical access is characterized by non-selectivity (e.g. De Groot et al., 2000; Dijkstra & Van Heuven, 1998, 2002). Such non-selectivity has been found to operate for orthographic (e.g. De Groot & Nas, 1991) and phonological codes (e.g. Duyck, 2005; Jared & Kroll, 2001; Nas, 1983). More strikingly, even when the other language is not mentioned or used in the task, behavioral responses are affected (Dijkstra et al., 2000).

One model that incorporates non-selective access is the Bilingual Interactive Activation (BIA) model (Dijkstra & Van Heuven, 1998). A model of bilingual word recognition, it assumes rich interconnectedness between lexical representations within and across languages, with bidirectional inhibitory and excitatory links between feature, letter, word and language levels. Candidate words are activated independently of language membership – the embodiment of non-selective lexical access. In an extension of the BIA model, BIA + (Dijkstra & Van Heuven, 2002), control over language selection is implemented via a task/decision system, a function previously assigned to language nodes. Instead, their function – together with orthography, phonology and semantics – is to indicate language membership. Integration across the differ...

Table of contents

• Coverpage
• Titlepage
• Copyright
• Contents
• Contributors
• Preface: Time for New Metaphors?
• 1 Neurolinguistic Contributions to Understanding the Bilingual Mental Lexicon
• 2 The Bilingual Lexicon and Bilingual Autobiographical Memory: The Neurocognitive Basic Systems View
• 3 Audio-visual Integration During Bilingual Language Processing
• 4 An Overview of Semantic Processing in Bilinguals: Methods and Findings
• 5 Lexical Transfer
• 6 Conceptual Representation in the Bilingual Lexicon and Second Language Vocabulary Learning
• 7 Why Gestures are Relevant to the Bilingual Lexicon
• 8 The Tip-of-the-Tongue Phenomenon as a Window on (Bilingual) Lexical Retrieval
• 9 L1 Attrition and the Mental Lexicon
• Author Index
• Subject Index