Presentations:

• Agnes Bolonyai and Lida Dutkova-Cope, East Carolina University. "L1 attrition of verbal morphology in bilingual children."

This paper examines the attrition of L1 verbal morphology in Hungarian-English and Czech-English bilingual children. Are all grammatical features encoded in verbal markers equally affected in the L1? We address this question by comparing effects of attrition in L1s with complex verbal agreement systems. In contrast with English, verbal agreement in both Czech and Hungarian occurs across person, number, and gender (Czech) or definiteness (Hungarian). Previous studies demonstrate that changes in the waning language are selective. However, our understanding of mechanisms explaining why and how particular features are lost or restructured is still limited. We examine effects of L1 attrition in contact with L2 using a lexically-based framework (e.g. Levelt 1989, Myers-Scotton & Jake 1999) and explain outcomes based on how morphemes are projected from the mental lexicon. Functional elements (here, system morphemes) are considered in terms of two properties: (1) morpheme status in production, and (2) morpheme participation in co-indexing relations. System morphemes are projected at either the lexical-conceptual or the morphosyntactic level. Further, some system morphemes require cross-checking of grammatical features across lexical heads, while others do not. We predict that these differences affect the stability of morpheme types in attrition. Data (20 hours) include spontaneous conversations recorded from six Hungarian-English bilinguals and elicited narratives from two Czech-English bilinguals (ages 5-9) living in the U.S. The results show that gender and definiteness are less accurate than person and number in the unstable Czech and Hungarian, respectively. Confirming our prediction, features that are absent in L2 and require more complex morphosyntactic feature checking are less accurate. This study supports previous claims that (1) heaviness of functional load in L1 vs. L2 affects L1 attrition (Andersen 1982), and (2) more complex outer layer mappings from morphology to phonetic form are likely to cause greater difficulties in L2 acquisition (Lardiere 2000).

• Lynne Hansen, Brigham Young University, Hawaii. "Forgotten but not Gone: Extending the Savings Paradigm in the Recovery of Lost L2 Vocabulary and Syntax."

This paper extends the line of research which has recently applied the savings paradigm from cognitive psychology to the relearning of second language lexicon (de Bot & Stoessel, 1998, 1999; Hansen, Umeda & McKinney, 2000). It is a continuation of the work reported by Hansen et al.(2000) which demonstrated a significant savings advantage in the regaining of L2 Japanese and L2 Korean vocabulary by English speaking attriters. The 281 subjects, having learned their second language during two years in Japan or Korea, had been back in North America for periods ranging from 1 to 45 years. Through analyses of collateral data collected from the same subjects six months later, evidence is provided on long-term savings effects in L2 vocabulary relearning. In tracing the savings benefits longitudinally, the presentation will also consider the influence of word = class in the retention and regaining of a second language. In addition to the data on the recovery of lost lexicon, the second interview provides the first evidence so far on the applicability of the savings approach to the recovery of lost syntactic knowledge, specifically to grammatical particles in the two East Asian languages.

• Georgette Ioup, University of New Orleans. "Age Differences in Forgetting a Language."

It has been observed that children lose language much more rapidly and completely than adults (Pan & Gleason, 1986; Yukawa, 1997). Several explanations have been offered: i. level of fluency (critical mass) attained (Pan & Gleason, 1986); ii. literacy skills in the attriting language (Olshtain, 1986); iii. stability of the linguistic subsystems in memory (Neisser, 1984). These explanations have been called into question by various research studies: Bahrick (1984) of the critical mass hypothesis, Ousman (1992) of the literacy hypothesis, deBot & Stoessel (1998) of the stability hypothesis. A second question has confounded the investigation of age. Is the young attriter's language actually lost from memory or simply unable to be recalled. deBot & Stoessel (1998) investigated this question using the savings paradigm based on the work of Ebbinghaus (1885). Brief training in the attrited language was able to elicit language recognition from their subjects who had lost a childhood language they once had literacy in. The present study investigates the question of age and literacy by examining two male siblings who began attrition at very different ages: 6;9 and 13;1. At the time of testing subjects were aged 53 and 59 respectively. Neither subject had ever been literate in the attrited language, nor had either had contact with the language during the intervening 45 years. At the onset of attrition the two siblings had achieved equivalent levels of fluency. The study attempts to answer the following two questions: i. Is there an age difference in the amount lost? ii. Is the language lost or just inaccessible? The savings paradigm was used in an attempt to elicit language from memory. Results indicate that even in the absence of literacy skills, the older learner retained more of the language before training began, and recalled more after the training. Even more interesting is that the younger sibling, who initially recalled almost nothing of the attrited language, was able to remember quite a bit after the brief training exercise.

• Dorit Kaufman, State University of New York at Stony Brook. "L1 attrition and narrative structure."

L1 attrition has most commonly been characterized as occurring after an incubation period that begins once exposure to the native language ceases. L1 attrition among pre-puberty children in immigrant communities often diverges from this process. In this case, acquisition of the native language continues and progression of attrition is offset by renewed opportunities for the rejuvenation of the native language. These include sustained exposure to L1 in the home and within the immigrant community and periodic immersion in the native language environment during visits to the homeland. This study explores L1 attrition in this context and investigates L1 production among pre-puberty children who are native speakers of Hebrew residing in the United States. Discussion focuses on the production of oral and written narratives. Languages frame narratives in uniquely language-specific ways and offer narrators formal linguistic devices and expressive options that provide linguistic and rhetorical scaffolding. Bilingual narrators are cognizant of formal devices available in each of their languages and appropriately draw upon linguistic and expressive resources in their respective languages for their narratives. This study examines the linguistic and rhetorical features that characterize the narrators' oral and written recounting of events. Discussion includes the extent to which fragmentation in the linguistic structures and word-formation paradigms of L1 constrain the narrators' lexical and structural choices and lead to reformulation of L1 word-formation rules and synthesis of the L1 and L2 in their narratives.

• Robert Russell, Brigham Young University, Utah. "Attrition in English Native Speakers' L2 Production of Japanese: 10 Years Later."

This paper will report on a 10-year follow-up study of L2 attrition among L1 English learners of Japanese as a second language (JSL). The subjects of the study had acquired their Japanese skills while residing in Japan for approximately two years, in an intensive, largely informal setting. The subjects then returned to an L1 environment, where opportunities to use their JSL skills were relatively limited. Earlier studies analyzed changes in oral monologue data elicited from the subjects over a two-year period of observation, revealing somewhat lower-than-expected levels of attrition in lexical and grammatical skills and in fluency. (Russell, 1996; 1999a; 1999b) Evidence of attrition reported earlier included decreased vocabulary size overall, a decline in accuracy in a few T-unit-based measures of lexical accuracy, an increase in the proportion of L1 English words used, an increase in the percentage of hesitation and pause time relative to total speech time, and changes in the number of different types of subordinate clauses used. This paper will describe changes that have occurred over an additional 10-year period of relative disuse of JSL by the same subjects, in terms of particle usage, syntactic complexity, and types of subordinate clauses used. In addition to revisiting quantitative syntactic measures addressed earlier (Russell 1999b), this paper will discuss aspects of the acquisition and loss of specific particle functions and subordinate clause structures. For example, orders of attrition of functions of the particles -WA and -GA will be compared with orders of acquisition reported for the same functions (e.g., in Russell, 1985; Sakamoto, 1993). The particle and subordinate clause data give evidence bearing on hypotheses of syntactic reduction in language attrition (cf. Andersen, 1982) and on the regression hypothesis relating to the order of loss of linguistic structures.

• Saskia Stoessel, Tufts University. "Multilingual social networks: A sociolinguistic perspective on first language attrition in immigrants."

In this presentation of a multiple case study, I will investigate the structure of social networks of ten immigrant women in the US. I will look at their social networks with regard to numeric distribution of L1 and L2 speakers, multiplicity of ties, and frequency and geographic vicinity of contacts. I will determine whether these can be a factor linked to eventual language attrition. I will also present data from in-depth semi-structured interviews discussing the importance and meaning of the L1 for the individuals within their new linguistic environment. At the time the study was performed, these immigrants still claimed full mastery of all four skills in the L1. Many of them did, however, express frustration when speaking about their L1 skills, because they were showing subtle signs of deterioration. I will try to find a partial answer to the question which of the immigrants might be more prone to language attrition than others and for what reason. The presentation will focus on two of the ten women who seemed least interested in language maintenance and therefore most likely to show early signs of language attrition. Rather than looking at linguistic features of attrition in detail, I will highlight the importance of earlier life-experiences these women had had with their L1 which may have influenced their present attitudes and patterns of language use. The results suggest that women whose home-country networks were relatively weak can be assumed to be earlier language attriters. Reasons could be found in issues like the linguistic presentation of self and of social identity related to the L1 in the home country. These findings are interesting in light of the fact that studies on immigrants' language use sometimes only focus on the here and now rather than also take into account the earlier history of the individuals involved.

• Kamal Sridhar, State University of New York at Stony Brook. "Patterns of Acquisition and Attrition in a transplanted language: The case of Thanjavur Marathi in India."

It is often claimed that a distinctive feature of bilingualism in India is its stability, i.e., speakers of Indian languages tend to maintain their languages over generations and centuries, even when they live away from the region where it is dominant. It is further claimed that maintenance is possible because of (1) the ethnic separateness of home life and (2) majority community's attitude towards the minority speakers' language and cultural maintenance. Speakers of Marathi, an Indo-European language spoken in the state of Maharashtra (North), have lived in a predominantly Tamil-speaking Dravidian language area (South) for over 300+ years. The ancestors of the present-day Marathi speakers came to Tanjore between 1638 and 1680. Like most Indians, the members of this community are trilingual. They use Tamil and English outside the home, and Marathi at home and as a community language. A brief historical overview, followed by a sociolinguistic profile of the community will be presented, outlining the context(s) in which Marathi is acquired and used. Data in the form of analysis of recorded speech from adults as well as young children will be used to answer following questions about: (1) maintenance or attrition of Marathi in Tanjore, and (2) examine if the data supports the above two claims often cited as factors accounting for stability bilingualism in India.

• S.N. Sridhar, State University of New York at Stony Brook. Discussant.

Back to SLRF 2000 Colloquia