Language Processing in Bilinguals

Psycholinguistic and Neuropsychological Perspectives

Jyotsna Vaid

• Publisher: Routledge
• Year: 2014
• Print ISBN:
  9780415724005
• eBook ISBN:
  9781317933137
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

For decades, bilingualism has resisted definition. If bilingualism is defined as habitual, fluent, correct and accent-free use of two languages, few individuals would qualify as bilinguals. A more viable approach may be to concede that 'bilingual' can be seen instead as a range of points on a continuum that allows for differences. The psychological study of bilingualism encompasses a wide range of phenomena including the organization and representation of the grammar, the perception and production of language mixing, cerebral lateralization of language functions, and patterns of recovery of aphasic patients. This book collects together an international array of researchers in experimental psychology, linguistics and neuropsychology, who bring their expertise to bear on the critical issues that are raised by the bilingual phenomena.

I

PSYCHOLINGUISTIC PERSPECTIVES

1

Skilled Reading in the Second Language

Norman Segalowitz

Can bilinguals read their second language as skilfully as they read their first? Often they do not, and there is much evidence that second language reading is generally slower, even in fluent bilinguals (as we shall see) and more difficult. A common explanation for this is that the reader lacks some of the basic linguistic knowledge necessary for skilled reading; the reader may not have a full and accurate command of the vocabulary and syntax of the language, the stylistic conventions of paragraph structure, or the cultural assumptions underlying the text (see e.g., Alderson & Urquhart, 1984, for discussion of these and related sources of second language reading difficulty). For example, Alderson (1984) offers two hypotheses about what is responsible for second language reading being weaker than first. One is that poor second language reading is due to the use of incorrect strategies, strategies different from those employed in native language reading. The other is that poor second language reading is due to insufficient knowledge of that language and the resulting inability to employ good first language strategies while reading it. Both these hypotheses focus on the bilingual’s ability or inability to make use of linguistic knowledge while reading. A similar point has been made in some of the psychological literature in terms of the bilingual being less skilled in making use of redundancy in the second language to facilitate reading (Albert & Obler, 1978; Favreau, Komoda, & Segalowitz, 1980; Macnamara, 1970). While there is undoubtedly merit in attributing some first and second language reading differences to such sources, this chapter considers a different source of second language reading difficulty. It examines the idea that certain basic cognitive mechanisms underlying reading, not directly related to higher order syntactic, stylistic and rhetorical linguistic knowledge, function less efficiently when the bilingual is processing second as compared to first language material.

We examine two sets of studies we have recently completed that look at this issue in highly skilled or “fluent” bilinguals whose languages have relatively similar writing systems (English and French; for a discussion of the impact of different writing systems on reading, see Hung & Tzeng, 1981; Feldman & Turvey, 1983). By “fluent bilingual” is meant the person who has for all practical purposes rapid and accurate ability to use the vocabulary and syntax of a second language, at least when required to perform under normal speaking and listening conditions, and is also generally skilled at reading the second language. We are here considering people who can express most ideas equally well in each language, who demonstrate good mechanical fluency in the second language although they may possess a slight accent (we have no data on whether there are subtle differences in speaking rates) and whose reading rate when reading for general comprehension is within the normal range (say, about 200 words per minute or better) (see Favreau & Segalowitz, 1982, for details). Such individuals, while perhaps not comprising the majority of second language users, nevertheless represent a significant and growing number, including citizens of bilingual regions, many foreign students and scholars, diplomats, politicians, and workers in international agencies and companies. Although such people are characterized here as “fluent” bilinguals it cannot be suggested that they function like two monolinguals nor need they be so considered (see Grosjean & Soares, this volume). Most likely we can find subtle but important differences between first and second language reading when they are required to perform a speeded task, such as reading a long text within some time limit or making rapid judgments about written material. Such individuals are interesting from a theoretical point of view because they are both relatively fast and slow readers at the same time depending on the language being read. Their slow reading is not easily accounted for in terms of some general reading skill deficit (skills that would affect reading any language), nor in terms of some fundamental weakness in the second language such as unfamiliarity with vocabulary or syntax.

The question of when and why second language reading might differ from first language reading has practical interest (see, e.g., the papers in Alderson & Urquhart, 1984). For example, in the bilingual region in which our laboratory is found, there are hundreds of thousands of highly skilled bilinguals whose work requires them daily to read a great deal of text in both French and English. We have consistently found that the majority of fluent bilinguals who visit our laboratory read significantly more slowly in their second language compared to their first. Typically, the second language reading rate is of the order of 60–70% of the first language reading rate (Favreau & Segalowitz, 1982). This difference can place bilinguals at a considerable disadvantage with respect to their colleagues or coworkers of the other mother tongue by affecting their relative working efficiency and, consequently, their possibilities for advancement.

INTERACTIVE PROCESSES

To begin, it will be useful to give a brief overview of current thinking regarding the nature of the reading process. In recent decades the focus of theories concerning the psychological mechanisms underlying native language reading has shifted from so-called bottom-up models and top-down models to models allowing for the dynamic interaction of processes at different levels of the system. Bottom-up models (e.g., Gough, 1972: LaBerge & Samuels, 1974) emphasize the one-way flow of information from low level perceptual analysing systems (e.g., mechanisms dealing with orthographic levels of the stimulus) to higher level semantic mechanisms (e.g., those dealing with lexical lookup and integration of textual material). Such models are generally now considered to be inadequate because they fail to account for a number of important phenomena studied by reading researchers (Rumelhart, 1977). For example, in one widely studied phenomenon known as the word superiority effect higher levels of processing (e.g., word recognition) affect lower levels of processing (e.g., letter identification) (Reicher, 1969; see Favreau et al., 1980, for a study of this in relation to first and second language reading differences in bilinguals). Similarly, various top-down models of reading (Goodman, 1976; Hochberg, 1970; Kolers, 1972; Levin & Kaplan, 1970: Neisser, 1967; Smith, 1973) that emphasize the direction of flow of information from higher levels to lower ones also have serious shortcomings (see McConkie & Rayner, 1976; Mitchell & Green, 1978; Rumelhart, 1977; Stanovich, 1980).

The general consensus among researchers appears to be that reading involves interactive processing. Information at one level may affect processes at a second level, regardless of which level is higher (Perfetti & Roth, 1981; Stanovich, 1980, 1981). This situation is illustrated in Fig. 1.1 which depicts in a general way the main levels of processing thought to underlie reading. (The figure conveys a rather simplified version of the interactive model insofar as few constraints on the interactions are made explicit; see e.g., Perfetti & McCutchen, 1982; Perfetti & Roth, 1981, for fuller discussions.) Three main levels are shown in the figure: a visual level concerned with the analysis of the printed stimulus, a lexical level concerned with word recognition, and a textual level concerned with integrating information across words to permit understanding of sentences, paragraphs, and stories. The figure also specifies two levels of phonological recoding, one prelexical between the visual and the lexical levels, and the other postlexical between the lexical and textual levels. The prelexical level reflects the possibility that visual information is converted into some form of phonological code (not necessarily subvocal speech, but possibly some abstract phonemic representation) as part of the process of recognizing the word (Lesgold & Perfetti, 1981; McCutchen & Perfetti, 1982). Perfetti and McCutchen (1982) propose that this happens automatically even though the speech based code so generated may not be directly involved in lexical access. The postlexical level reflects the possibility that already recognized words are recoded phonologically to facilitate holding them in memory (Barron, 1981) and thereby make them available to higher level linguistic processes (for syntactic parsing, propositional encoding, etc.). Finally, Fig. 1.1 also reflects the assumption that the lexical store for the two languages is common rather than separated (Caramazza & Brones, 1980; Nas, 1983; Albert & Obler, 1978; Segalowitz, 1977; Segalowitz & Lambert, 1969; but see also Hummel, this volume; Kirsner, Smith, Lockhart, King, & Jain, 1984; Scarborough, Gerard, & Cortese, 1984).

FIG. 1.1. Overview of the reading process.

The studies reported here focus on just two aspects of reading as depicted in Fig. 1.1. One concerns whether there are first and second language differences in the degree to which a printed word automatically activates a meaning representation in the mental lexicon and differences in the extent and time course of that activation (see also Magiste, this volume). The second concerns the degree to which phonological codes are implicated in second language reading.

AUTOMATIC AND CONTROLLED PROCESSES

The first issue we looked at concerns the degree to which word recognition with visual input is automatic. A complex activity such as reading is likely to involve many highly practised and automatized activities (LaBerge & Samuels, 1974) and skilled readers are thought to be better in automatic word recognition than unskilled readers (Stanovich, 1980). Automatic processing in reading research is often conceptualized in terms of the Posner and Snyder (1975) dual process model. They postulated the existence of two independent types of spreading activation. Automatic spreading activation is fast, does not use attentional resources and is not inhibitory. Controlled spreading activation, on the other hand, is slow, does use attentional resources and can be both facilitatory and inhibitory (see Schneider, Dumais, & Shiffrin, 1984, for a fuller discussion). Presumably, early in reading development, the processes of word recognition require effort and attention and only after extensive experience do they become automatized. Individual differences in reading skill may be partly due, therefore, to the degree to which this automatization has occurred. This suggestion is supported by the results of Stanovich and his colleagues (Stanovich, West, & Feeman, 1981; West & Stanovich, 1978) who report that...

Table of contents

• Cover Page
• Half Title page
• Title Page
• Copyright Page
• Orginal Title Page
• Orginal Copyright Page
• Contents
• Foreword
• Preface
• I Psycholinguistic Perspectives
• II Neuropsychological Perspectives
• Author Index
• Subject Index