Language attrition and theories of forgetting: a cross-disciplinary review.

Abstract

My purpose in this paper is to examine the psychological nature of forgetting (parts of) a language by individuals involved in language change across the lifespan. The objectives of the article are (1) to summarize psychological theories and hypotheses about forgetting and memory failure, and (2) to evaluate their relevance for the explanation of individual language attrition. By reviewing linguistically and psycholinguistically oriented language attrition studies that appear to implicate mechanisms of forgetting, I seek to contribute to the interdisciplinary study and discussion of language attrition phenomena.

Key words

cross-disciplinary

language attrition

theories of forgetting

1 Introduction

The study of individuals' language attrition has become a vibrant subfield of applied linguistics (Ammerlaan, Hulsen, Strating, & Yagmur, 2001; de Bot, 1996; de Bot & Weltens, 1995; Lambert & Freed, 1982; Weltens & Cohen, 1989). Language attrition in individual speakers may contribute to language change at the community and global level (cf. Meisel, 2001; Seliger, 1996). Some languages are gaining influence as first languages (L1) and second languages (L2) in a world characterized by globalization and migration, whereas other languages (often the L1 of minorities) and dialects are in decline and in danger of becoming forgotten (Crystal, 2000; Dorian, 1989; Grenoble & Whaley, 1998; Hyltenstam & Obler, 1989; Munstermann, 1989; Wong Fillmore, 1991). The knowledge and use of language(s) by individual speakers is in continuous flux. It is dynamic in nature and subject to change, that is, both to the acquisition of (novel) language structures and the attrition/loss of (obsolete) structures (Herdina & Jessner, 2002; Hyltenstam & Viberg, 1993). Whereas the growth of language skills has been a subject of scientific inquiry for a long time, the decline of language skills has been studied systematically only for about 20 years, mostly by applied linguists (see Hansen, 2001; Lambert & Freed, 1982; Schmid & de Bot, in press; Seliger & Vago, 1991; Weltens, 1987; Weltens & Cohen, 1989). Only a limited number of studies have investigated language attrition with reference to psychological or psycholinguistic theories (cf. Ammerlaan, 1996; de Bot, 1999; de Bot & Stoessel, 2000; Kenny, 1996; Pan & Berko Gleason, 1986). On the other hand, psychological studies about the remembering and forgetting of verbal and nonverbal information have devoted little or no attention to the forgetting of language in healthy individuals across the lifespan (e.g., Ashcraft, 1989; Golding & MacLeod, 1998; Schacter, 1996). Increasingly, however, also psychologists and speech scientists are becoming interested in the fields of bilingualism and language attrition (e.g., Kohnert, Bates, & Hernandez, 1999; McElree, Jia, & Litvak, 2000; Yeni-Komshian, Flege, & Liu, 2000).

The present article is an attempt to review psychological assumptions about forgetting and research into language attrition that may be of common interest to psychologists and linguists. It is intended as a positive response to the question asked by Ammerlaan et al. (2001, p. 3): "... should linguists employ insights and notions from decades of psychological research on memory performance in their work, or merely dismiss the findings as being based on experimental work on a limited area of vocabulary?" Applied linguists have been skeptical with respect to the potential contribution of psychological theories to the study of language attrition. Weltens and Grendel (1993), for example, came to the conclusion that psychological theories of forgetting have relatively little to offer for the study of L2 lexical attrition (also Weltens, 1987). The present discussion represents a more optimistic view and intends to demonstrate that part of the general psychological insight into forgetting may be of value for hypothesis formation and explanation of language attrition, which in turn can reveal important insight into the functioning of the human language faculty (Slobin, 1977).

In the context of the present discussion, language attrition refers to the decline of any language (L1 or L2), skill or portion thereof in a healthy individual speaker. The review will take into account attrition at the phonological, morphological, lexical, syntactic, and semantic levels. Since most research has been conducted on the lexicon, however, lexical loss will somewhat dominate the present discussion. Because of space limitations, language loss due to brain injury (aphasia) or severe pathological changes due to aging (dementia) will not be included in the review (see, e.g., Hyltenstam & Stroud, 1993; Paradis, 1995, 2001). Studies of language shift (e.g., Fishman, 1972), that is, the gradual change of language use in generations of a community will only be considered if they encompass information on individual language loss.

The first part of the paper explores the phenomena of forgetting and language attrition in relation to each other. The second part summarizes main assumptions about forgetting advocated in psychological theories as a foundation for the paper's third and main section that reviews findings from language attrition studies that seem consistent with or related to psychological explanations of forgetting.

2 Forgetting and Language Attrition

Forgetting may result from failure in one of three basic components of remembering: encoding (the capture and acquisition of novel information), storage (the integration and permanent representation of information), and retrieval (the access to information when it is needed by the speaker). Two different types of memory have been assumed to carry out these functions. Short-term (or working) memory is a limited computational space that temporarily stores, monitors, and manipulates information. It is of crucial importance to any mental activity extended in time, including the acquisition and processing of speech, although certainly in some interaction with long-term memory, the memory system that is responsible for the permanent storage of information (cf. Baddeley, 1986, 1999; Spear & Riccio, 1994). In studies of language attrition, it is usually assumed that the encoding component had been intact, and that only what has been acquired can be forgotten. This implies that the basic problem of forgetting resides in either the storage or the retrieval of structures that had been acquired and retained at some point in the past and that were "lost" under certain conditions. However, attrition researchers have also acknowledged the need to take into account the possibility that linguistic structures have not been acquired (encoded) completely and that the instable storage of these structures may have contributed to poor performance and a sensation of forgetting (e.g., Isurin, 2000; Levine, 1996; Montrul, 2002; Saville-Troike, Pan, & Dutkova-Cope, 1995). This holds in particular for L1 attrition in bilingual children and L2 loss in post puberty L2 learners.

Four important question complexes appear of relevance for investigations into language attrition as well as for educational efforts to preserve and revitalize languages, L1 as well as L2:

(1) What is forgetting? Is it the loss of information (the decline of competence) or is it manifested in the impaired access of information (a performance/processing problem)?

(2) What are the causes of forgetting? Or formulated differently, what are the internal and external factors that contribute to the forgetting of language structures?

(3) What are the linguistic effects of forgetting? Which linguistic levels are affected by forgetting, to what degree and in what sequence? How does forgetting in one area (e.g., lexis) affect competence and performance in another (e.g., syntax)? How does the individual compensate for linguistic deficits?

(4) What are the social-psychological consequences of forgetting? How does forgetting affect the individual's life, identity, or self?

In practical terms, the answers to question (1) will have implications for attempts to recover or relearn (parts of) a language since they address the potential success of treatment of individual language attrition: Under what circumstances, at what cost, and to what degree is regaining the language possible? The responses to questions (2) and (3) will be of relevance with respect to the prevention of language attrition. What is necessary to maintain continuous language use and development? Insight into questions (4) could help educators, parents, and language planners make decisions as to whether and under what circumstances the forgetting of language should be prevented or resisted.

3 Psychological explanations of forgetting

Seven theories/hypotheses about the forgetting of (verbal as well as nonverbal) information will be discussed concerning their potential relevance for research on language attrition: Repression/suppression, distortion, interference, decay, retrieval failure, cue dependency, and dynamic systems theory (1). Other recent explanations of forgetting, for example, "the seven sins of memory" discussed in Schacter (2001), will be subsumed under the more familiar theories and conceptions above.

3.1

Repression and suppression

This explanatory account of forgetting originated in the area of psychoanalysis and the works of Sigmund Freud. According to Freud, unpleasant or traumatic memories are repressed deliberately and deplaced from consciousness by the individual who intends to avoid recalling displeasing, negative, or traumatic experiences (see Freud, 1899/1961; Jones, 1993). Repression also has been used widely to denote a massive "repression mechanism that operates unconsciously and defensively to block out traumatic experiences" (Schacter, 1996, p. 255). The main difference between suppression and repression lies in the possibility to consciously retrieve the target information, which is presumed possible in the first case, whereas it is expected impossible in the second (Golding & Long, 1998). While some cognitive psychologists dispute the possibility of repression, most agree that intentional suppression (also directed forgetting) exists as a commonly used, essential human defense mechanism (see Golding & MacLeod, 1998). In this discussion, I will refer to repression/suppression as an (at least initially) intentional mechanism in which the individual refuses recalling/using a memory structure. Such refusal may entail the withdrawal of attention from an initially highly active memory structure by deliberately remembering another structure (cf. Bjork, 1998; Wegner, Eich, & Bjork, 1994).

3.2

Distortion

The concept of distortion emphasizes the subjectivity and permeability of personal memories. Individual experiences, current knowledge, beliefs, values, and attitudes affect and change what a person remembers. The information retained in memory is being altered and restructured unconsciously (see Schacter, Coyle, Fischbach, Mesulam, & Sullivan, 1995). Unconscious restructuring may be caused or affected by internal factors as well as external factors. Distortion research has become important, for example, to evaluate eyewitness testimonies in court processes. Certain questioning techniques, information from third parties, and the rehearsal of testimonies, among other factors, were shown to lead to incorrect (distorted) recollections of to-be-recalled events (e.g., Wells, 1993).

3.3

Interference

According to interference theory, prior and posterior learning, retention, and processing of segments of information that compete with each other are responsible for the forgetting of target information. On the one hand, information acquired at a later point in time can inhibit or block the recall of information that was acquired earlier (retroactive inhibition). On the other hand, information that was acquired in the past can interfere with the learning and recall of more recent or novel information (proactive inhibition). See Anderson, Bjork and Bjork (1994), Keppel (1968), and Postman and Underwood (1973). Concepts analogous to interference are "mental blocking" (Jones, 1989) and "competition" (Bates & MacWhinney, 1987). Contrary to the decay and transmission deficit hypotheses discussed below, proponents of blocking assume that previously activated competing items inhibit the retrieval of a target item.

3.4

Decay

Decay is probably the oldest of all approaches to the explanation of forgetting. It assumes that information "evaporates" or declines gradually in memory through lack of use (Thorndike, 1914). The assumption implies that the frequency and recency of use of the structure (i.e., its continuous activation) is crucial for the maintenance and access of the information in memory. The lack of use of the information results in the dissipation of a "trace" that has been imprinted for a piece of information, represented in the brain.

3.5

Retrieval slowdown and failure

Forgetting does not have to represent the total loss/elimination of information from memory (Ashcraft, 1989, 1998; Spear & Riccio, 1994), its principal problem may lie in achieving access of the desired information. As in decay theory, frequency of use is a crucial variable in models of information retrieval. In (modular) search models, differences in retrieval speed are figuratively explained by means of autonomous storage bins (e.g., Forster, 1992, on lexical access). Frequently and recently used information is stored on top of the bin and rapidly accessible whereas infrequently used items are located at the bottom of the bin and require more time and effort to be accessed. Parallel search of various bins can be assumed for bilingual lexicons.

In neural network and spreading activation models, retrieval slowdown and failure are seen as a consequence of decreasing activation levels in neurons and processing units, and weakening connections between neurons and nodes respectively. The lack of activation through infrequent use and the aging of connections in older adults cause deficits in the transmission of information (e.g., from the semantic level to the phonological level in word production) resulting in a decrease of retrieval speed and an increase of retrieval failure rates (e.g., Burke, 1999).

3.6

Cue dependency

Advocates of "cue dependent retrieval" argue that forgetting is neither satisfactorily explainable as the decay of information nor as a consequence of interference. Higbee (1996) describes memory search as analogous to reviewing an extensive file cabinet. In order to quickly find the desired information in the cabinet, labels (cues) are needed to indicate the address of the location of the information. The recognition and use of cues is crucial for successful retrieval. Cue availability depends on the consistency of internal states of subjects, their feelings and mood (Bower, 1981) as well as the external environment in which learning and recall take place (Tulving & Madigan, 1970). Changes in context can reduce the availability of retrieval cues (Spear & Riccio, 1994). When the internal or external cues that were available during memorization are unavailable during recall attempts, the information is more difficult to retrieve.

3.7

Interaction and dynamic systems

It is likely that various approaches offer relevant explanation for certain types of language attrition and that the described phenomena overlap and interact collectively in the complex cognitive systems of the bi- or multilingual speakers. In emergent, dynamic, self-organizing systems, "... formal structures of language emerge from the interaction of social patterns, patterns implicit in the input, and pressures arising from general aspects of the cognitive system" (MacWhinney, 1998, p. 199). Herdina and Jessner's (2002) Dynamic Model of Multilingualism is an attempt to relate variables and phenomena, such as, language acquisition, language maintenance effort, transfer/interference, and loss of languages to each other. Learning an additional language is achieved only in competition to and at the cost of limited cognitive resources already used for the maintenance of the other language system(s). According to the model, the acquisition of the novel language negatively affects the other language system(s) over time resulting in less acquisition and attrition of the previously acquired systems (Jessner, 2003). Only the speaker's metalinguistic awareness and/or language learning aptitude counteract the decline of resources, use, and competence to keep the system at equilibrium with balanced environmental (communicative) demands and cognitive resources.

In the following, cases of language attrition will be reviewed that appear consistent with the psychological assumptions presented above. An attempt will be made to assess the significance of the theories of forgetting for language attrition research.

4 Forgetting Language

4.1

Linguistic repression/suppression

Various language attrition studies suggest that individuals can suppress the use of a language because of social-psychological reasons (cf. Pavlenko & Lantolf, 2000). Cases that seem consistent with language suppression have been reported in studies of children who had been removed from their native country to an environment where the L1 was not spoken. Three studies investigated L2 acquisition and L1 attrition in orphan children from Russia, the Ukraine, and China who had been adopted by U.S. families (Isurin, 2000; Nicoladis & Grabois, 2002; Saville-Troike et. al., 1995). After a short time of initial resistance to the L2, the children progressed rapidly in the acquisition of English, but also in the loss of the L1. The children started to refuse speaking in the L1, possibly because of their desire to assimilate in the new culture, acculturate in the terms of Schumann (1978) or simply because they associated the …