Language Attrition Research Archive
                                  PacSLRF'98 Second Language Research Forum of the Pacific. Tokyo, Japan. March 26, 1998. Lynne Hansen and Machiko Tomiyama (Chairs). Explaining language attrition: Predictors from language acquisition.

l. to r. Emiko Yukawa, Lynne Hansen, Machiko Tomiyama, Sumiko  Nagasawa, Robert Russell.
Papers presented at the symposium appear in Peter Robinson (ed.). Representation and process: Proceedings of the 3rd Pacific Second Language Research Forum, Volume 1. Tokyo: The Pacific Second Language Research Forum (Aoyama University).

  This colloquium explores the influence of variables in the language acquisition process on subsequent attrition of the language. Questions and issues concerning relationships between language learning and loss addressed by the presenters include the following:

  • To what extent does the regression hypothesis- the notion that the process of language loss is a mirror-image of the process of language learning- predict the sequence of attrition?
  • Is it the case that "the more you learn, the less you lose?" Is there a critical threshold of acquisition which, once attained, significantly increases the liklihood of language stability and maintenance? Do longer periods of L2 exposure predict better retention of what is learned than shorter periods?
  • What is the influence of classroom learning on language retention? How does formal instruction (vs lack of it) at various stages of acquisition affect language maintenance?
  • Does the attainment of literacy skills during acquisition provide an anchor for the subsequent retention of the spoken language? 


Hansen, Lynne & Ching-Fen Chantrill (Brigham Young University, Hawaii). Literacy as a second language anchor: Evidence from L2 Japanese and L2 Chinese.

Nagasawa, Sumiko (Kurume University). The effects of initial achievement, learning experiences, and classroom instruction on adult attrition/retention of L2 Japanese.

Russell, Robert (Brigham Young University). Measuring attrition in L2 Japanese syntactic competence.

Tomiyama, Machiko (International Christian University). The later stages of natural L2 attrition.

Yukawa, Emiko (Stockholm University, Center for Research on Bilingualism). L1 Japanese attrition and regaining: the age and pre-attrition variables.



Literacy as a second language anchor: Evidence from L2 Japanese and Chinese
Lynne Hansen & Ching-Fen Chantrill

To what extent does learning to read and write anchor a second language after the termination of instruction or exposure to that language? The present study addresses this question by examining L2 loss among 511 English-speaking adults whose literacy attainment had varied dramatically as they had learned Japanese or Chinese. Since learning their L2 while working in Japan or Taiwan there had been a hiatus in regular use of their second language from 1 to 42 years.

Several measures of current L2 ability are the dependent variables in the study. In a series of factor analyses which included language proficiency and literacy variables as independent variables, the one that is singled out as a consistently robust predictor of language retention is the number of kanji or Chinese characters that had been learned. These findings suggest that literacy acquisition, at least in the context of logographic scripts, relates significantly to the maintenance of a second language over several decades of disuse.

The Loss and Regaining of Japanese by Bilingual Children
EmikoYukawa, Stockholm University, Center for Research on Bilingualism

        This paper reports on three case studies of L1 Japanese loss and relearning. The purposes are to confirm our intuition that small children, especially early bilinguals, are truly quick in losing (relearning) their languages once they stop (resume) using them often (Kuhberg, 1992; Olshtain1986, 1989), to describe in detail the phenomena of the loss, and to examine the nature of the loss of these young bilingual children. While addressing these issues, factors of age (Olshtain, 1986 1989) and pre-attrition proficiency (Bahrick, 1984) are taken into consideration as important variables.

        The subjects are two siblings (S and H). One was studied twice (addressed as H1 and H2). They were raised in Japan to be English and Japanese bilinguals from birth. H (H1) was taken to Hawaii for 5 months when he was 5;5-5;10, and both S and H (H2) were taken to Stockholm when S was 3;10-5;5 and H2 was 7;0-8;4. While abroad, they spoke English at home and at (pre-)school, and thus the lack of input and output opportunities caused attrition of their Japanese.

        The three cases were compared based on lexical and syntactic analyses of spoken data and observation of comprehension and repetition data. It was confirmed that the two younger subjects (S and H1) were quick (within a few months) in losing the ability to produce sentences in Japanese. As for the nature of attrition, analysis of the data revealed that there were two qualitatively different attrition processes going on, one for the two younger subjects (strong processing failure for production) and the other for the oldest one (mild change in peripheral grammar and lexicon).

The Later Stages of Natural L2 Attrition
Machiko Tomiyama, International Christian University

    This paper examines the later stages of natural L2 attrition in the L1 environment as observed in a Japanese male returnee child. It documents the period between two to four years after his return to Japan, the period which still remains unexplored under a longitudinal-study framework.

    The subject was 8;0 when he returned home. He acquired English as his second language during a seven-year residence in the United States. He was a subject of high-proficiency and literacy in L2 at the onset of the study. The data consisted of spontaneous and elicited speech collected at regular intervals.

    During the period of observation the deterioration in the subject's overall command of English became evident with the gradual decline in his ability to coordinate various linguistic skills. Specifically, the decline in his conversational discourse skills was observed, including such skills as initiating a new topic and volunteering additional information. Furthermore, his utterances became shorter and syntactically fixed using a few formulaic patterns reflecting L1 influence. His productive vocabulary achieved stability in contrast to the rapid decline observed during the initial stage of attrition. On the other hand, his phonological and receptive skills showed no apparent erosion even in the later stages of the attrition process.

    Comparing these attrition data of later stages with those of earlier stages of acquisition, issues surrounding the regression hypothesis will be discussed. In addition, the factors of age and critical threshold in proficiency will be considered by examining the present subject's profile against those from previous research.

The Effects of Initial Achievement, Learning Experiences, and Classroom Instruction
on Adult Attrition/Retention of L2 Japanese

Sumiko Nagasawa, Kurume University

    A multiple-case study of the attrition of adult L2 Japanese speakers in the U.S. was conducted. Subjects were seven MBA students who were required to attain or maintain advanced speaking skills for graduation. The study focused on whether their initial achievement, learning experience, and classroom instruction were predictors of attrition/retention. It was different from other attrition studies in that the subjects had non-intensive classroom instruction, instead of non-use, during the academic year (seven months) after returning from a summer immersion in Japan.

    Subjects initial achievement, the major predictor, was measured by the ACTFL oral proficiency scale, and its predictive power was tested against several dependent variables (Japanese particle use, sentence structure, and fluency), features analyzed from the OPI tapes. The length and type of classroom instruction prior to the study was determined through a questionnaire and the researchers personal interview with the subjects.

    Students who had experienced initial exposure to the L2 without instruction in a natural milieu experienced greater language attrition during the observation period than those who had been given instruction from the beginning. The findings revealed that the initial achievement of Advanced High on the ACTFL scale was a strong predictor for retention of speaking skills.

Measuring Attrition in L2 Japanese Syntactic Competence
Robert A. Russell, Brigham Young University

    This paper reports the results of a detailed syntactic analysis of oral monologue samples elicited from 20 speakers of Japanese as a second language three times over a period of two years. The speakers had all recently returned to the US from approximately two years' residence and intensive, informal JSL acquisition experience in Japan. During the two years of observation following their return, the subjects had relatively little opportunity to speak Japanese, but 10 of the subjects had taken eight or more credit hours of formal instruction, mainly reading-oriented, following their return from Japan, while the other 10 had had little or no formal instruction in the language following their return.

    The goals of the syntactic analysis were: (1)to describe changes in case and focus particle usage over the two years of relative disuse of the language, (2) to analyze particle errors and changes in error rates for particular particles, (3) to measure changes in syntactic complexity over time, using both T-Unit-based quantitative measures as well as qualitative analyses of specific syntactic constructions (including, for example, complement clauses, conditionals, modal constructions, reason/purpose clauses, time and other adverbial clauses, relative clauses, etc.), and (4) to compare the two groups of subjects to see whether the formal instruction group suffered less loss of syntactic competency than the group without instruction.

    The analysis is still in progress, but the prediction is that the subjects' speech samples will exhibit, over time, decreasing levels of syntactic complexity and increased rates of particle and other syntactic errors. It is further predicted that any decrease in syntactic competence will turn out to be less pronounced in the formal instruction group than in the group without formal instruction. This paper will also address the question of whether different initial levels of proficiency (to the extent that these can be measured with the data available) are correlated with different degrees of retention or loss of specific syntactic features.