The language ecology of bilingual memory.

Abstract

The present research examined language(s)-memory relations among compound bilingual children. Findings revealed the importance of language socialization with respect to language behavior exhibited by participants at recall. Bilinguals exhibited significant gains, with increased language experience, in communicative and conceptual linguistic competence, and metalinguistic competence. Results suggest that bilingual memory is not a dormant, cognitive state but a dynamic mosaic of reciprocal relations between individual, cognitive, social, contextual, and behavioral factors.

Introduction

The focus of the present study was on the relationship of how language(s) interact with the cognitive processes of Long Term Memory among compound Greek-English bilingual children. Compound bilinguals (also referred to in the literature as simultaneous bilinguals) are developing both their languages, L1 and L2, concurrently through daily use (Ervin & Osgood, 1954; Grosjean, 1982). In the present research, bilingual participants, while learning two languages at the same time at home, were educated in two different contexts; learning two languages simultaneously in school daily versus receiving more limited (2 afternoons per week) formal education in Greek. Therefore, the aforementioned relationship was examined as a developing function of bilinguals' language socialization.

The dynamics of bilingual memory are illustrated within the construct of language ecology (Creese & Martin, 2003), which acknowledges the importance of both the language speaker (the individual) as well as the language setting (the environment). Thus, language is not viewed as a separate or static entity but rather as an integral part of the speech community in which it resides and is sustained (i.e., born, develops and is used). In this regard, language ecology embraces both the relationship between one's languages as well as the relationship between one's languages and one's cognitive development. Consistent with language ecology, it follows that a comprehensive examination of bilinguals' language behavior, including how they use their languages to perform different cognitive functions, such as memory, necessitates consideration of one's prior language experiences. These experiences, which are both cognitive and social in nature, include how bilinguals acquire, develop, use, and think about their languages, language attitudes and preferences, as well as prior social interactions, where bilinguals negotiate meaning and acquire knowledge about the social and cognitive demands of a given learning task.

If we examine the bilingual memory literature through the lens of language ecology it becomes evident that many of its branches have not been made readily visible and/or systematically controlled for by researchers and, as a result, the dynamics of bilingual memory has yet to be fully realized. More specifically, much of the literature has emphasized the cognitive aspect of bilingualism, namely L1 versus L2 proficiency among coordinate bilinguals. Coordinate or consecutive bilinguals have acquired their second (subordinate) language, L2, much later than their first (dominant) language, L1, sometimes well into adulthood (Grosjean, 1982; Weinrich, 1953). More specifically, for coordinate bilinguals, L2 is acquired subsequent to the linguistic and conceptual development of L1. Consistent with the focus on coordinate bilinguals are linear depictions of L2 acquisition as a cognitive burden that raises the difficulty level by a factor of x+1 (Mahon, Crutchley, & Quinn, 2003) and unidimensional characterizations of bilinguals as static language processing vessels and/or storage containers. Bilingual memory models, also based on coordinate bilinguals, have followed suit in this regard, reducing bilingual memory to a simple dichotomy (i.e., language specific versus language neutral memory) whereby language(s)memory links are based solely on L1 versus L2 proficiency and/or dominance.

Even at the cognitive level, debates have ensued with respect to the latter issue, highlighting the heterogeneity of the coordinate bilingual experience and the importance of individual and contextual variables in this regard. For example, although the hierarchical model of bilingual memory (Dufour & Kroll, 1995) grants E1 dominant memory access, denying L2 equitable access until it reaches L1 proficiency levels, a more recent model (DeGroot & Poot, 1997) grants L2 comparable memory access from the early stages of its development. By contrast, more recent research (Heredia & Altarriba, 2001) has proposed that the more dominant (i.e., more frequently used) language, be it L1 or L2, has greater memory accessibility. Notably, this debate may be a moot point given the lack of consistency across autobiographical memory research, based on coordinate bilingual adults who have acquired L2 as adults, over whether or not cognitive dominance in a language dictates memory dominance in that same language. For instance, while some research has argued that events experienced in coordinate bilinguals' dominant language (L1) are better recalled and thus more accessible in that same language (Marian & Neisser, 2000) other research has reported that memories encoded in L1 are readily retrievable by coordinate bilinguals in L2 (Schrauf & Rubin, 2000).

Given that the aforementioned suppositions have been based on a template of coordinate bilinguals is somewhat problematic for compound bilingual participants of the present study, who are developing and using both their languages daily and therefore, by definition, lack a dominant language. It has been proclaimed that bilingual assessment must progress beyond proficiency and dominance and embrace a more multidimensional approach, where language behavior is viewed as a product of bilinguals' previous socialization experiences as well as their given environmental context (Cummins, 2000; Hamers, 2004), highlighting the usefulness of the language ecology framework.

In the present study, the emphasis was on bilinguals' development of different language behavior competencies and the role of bilinguals' language experiences in this regard. These competencies include communicative linguistic competence, i.e., the ability to remember information in each of one's languages, conceptual linguistic competence, namely, the use of language as a decontextualized cognitive organizer of knowledge (i.e., organizing one's memory via language(s), and metalinguistic competence, i.e., bilinguals' cognitive awareness of their manipulation and use of language. The evolvement of language from a communicative tool to that of a cognitive organizer is said to be more enhanced among compound bilingual children that valorize both their languages and use them equally for the same cognitive and social functions (Hamers, 2004). This claim may be less true for coordinate bilinguals that are learning L2 subsequent to the development of L1, presumably in different contexts, resulting in the development of translation equivalents in two languages that correspond to two different sets of representations (Hamers & Blanc, 1990), yielding less linguistic overlap and requiring, perhaps, less cognitive organization.

Method

Participants were comprised of two cohorts (daily and weekly) of compound bilingual first graders, with a mean age of 5.8 years and two cohorts of third grade compound bilinguals, with a mean age of 7.7 years. There were 32 participants; 16 per grade and 8 per cohort. The two daily cohorts of participants were selected from the same Greek-English Parochial School, located in an additive bilingual community, where both languages (English and Greek) were socially and cognitively valued and used. They received 3 hours of daily academic instruction in Greek that consisted of reading and listening comprehension, writing, history, and religion. The remainder of the academic day consisted of instruction in English. The two weekly cohorts of bilinguals were selected from the afternoon program of the same Greek-English parochial school attended by the daily cohorts. They received 6 hours (two afternoons per week) of academic Greek instruction, comparable in content to that of the daily cohort and taught by the same Greek instructors. The weekly cohorts attended the same public school, located in the same additive bilingual community as that of the daily cohorts.

Procedure

Bilingual participants were presented with a long-term memory (3-day) story recall task. Bilinguals were instructed, in English, that they would hear two audio-taped stories over two days and to try to remember them because they would be asked to recall the stories on the third day. Each story consisted of two events (breakfast and party) each presented in a different language. A bilingual experimenter elicited recall for each participant in both English and Greek. Bilinguals were subsequently asked to explain the strategies they used to remember (metamemory). Language questionnaires were administered to all participants to obtain information on home language use, language attitudes and self ratings of language adeptness. Given researchers' acknowledgement that equal L1 vs. L2 proficiency is more of a theoretical, cognitive ideal versus a reality, bilingual studies have come to consistently rely on self-adeptness ratings and language use (Bialystok, 1997; Francis, 1999; Hakuta, Ferdman & Diaz, 1987).

Design

Bilingual participants were randomly assigned to one of eight possible experimental orders. The design of the present study was a 2(story order) X 2(language of event) X 2(language at recall) between subjects design. See table, issue website http://rapidintellect.com/ AEQweb/fal2005.htm

Results

Story recall was analyzed quantitatively, in terms of mean number of phrases recalled from each story, each event, and each language. Qualitative analyses of individual recall protocols examined to what degree bilingual memory was assembled by story (1 or 2), by general event (breakfasts (B) and parties (P)), by specific event (B1, B2, P1, P2), and by language of presentation (Greek and English).

How much did bilinguals remember? Group mean comparisons were tested for statistical significance through analyses of variances via the "F" statistic with two degrees of freedom (a,b). The first number, "a", is the number of groups in the comparison minus one and the second number, "b", is the total number of participants in the comparison minus the number of groups. The subsequent "p"(probability) value indicates the degree of likelihood that the finding occurred due to chance/sampling error (e.g., the .01 level indicates a less than 1 percent chance). The Main effect of grade revealed that third graders remembered significantly more phrases that first graders, F(1,30) = 166.7, p < .01. Furthermore, each cohort (daily and weekly) of bilingual third graders remembered significantly more phrases than either cohort of first graders (all p's < .01). Notably, memory comparisons of daily versus weekly cohorts, among first graders and among third graders, revealed no significant differences with respect to the amount of phrases recalled from any given story, breakfast event, and/or party event.

What did bilinguals remember? Communicative Linguistic Competence was evidenced by comparable memory for events (B&P) heard in English (E) versus Greek (G) among first (3.19 vs. 3.62) and third graders (7.62 vs. 8.25). There were no significant differences with respect to memory for BE vs. BG events and PE vs. PG events among daily and weekly cohorts within each grade. Analyses of first grade recall protocols revealed a variety of memory preferences. First, the majority (75%) remembered a breakfast and party event, from either the same story or from different stories and 25% recalled the last party event heard. Second, significantly more phrases were remembered from party versus breakfast events, overall, F(1,14) = 19.81, within the daily cohort, F(1,6) = 12.45, and within the weekly cohort, F(1,6) = 13.22, (all p's<.01). Third, a recency effect revealed that participants remembered significantly more phrases from the second story versus first story heard, F(1,14) = 6.70, p < .01, and the second party versus first party heard, overall, F(1,14) = 17.43, within the daily cohort, F(1,6) = 7.18, and within the weekly cohort, F(1,6) = 9.31 (all p's, <.01). Lastly, significantly more phrases were remembered from the B2 vs. B1 event, overall, F(1,14) = 7.26, p = .01, in the daily cohort, F(1,6) = 11.25 (p < .01), and in the weekly cohort, F(1,6) = 5.11, p < .05. Unlike first graders, the majority (68%) of third graders remembered a comparable number of phrases from all four story events, while 32% remembered a comparable number of phrases from three story events.

How did bilinguals remember? The nature of the story recall task of the present study afforded bilingual participants the opportunity to organize their memory by story (1 or 2), event [breakfasts (B's) and parties (P's)], specific event (B1, B2, P1, P2), and language of presentation (LP) [English(E) and/or Greek(G)]. Recall analyses of individual protocols revealed that the majority of all bilingual participants organized their memory via specific event, both across and within stories and in a variety of event and story orders (i.e., B1-P2, P2-B1, B1-P2-B2-P1, respectively.) and language orders (E-G, G-E, E-G-E-G, respectively), illustrating strong L1-L2 relations and L1-L2 mobility. [Refer to Figure One] First grade recall was organized three different ways: l) through a breakfast(B) and party(P) event, mainly in the same story and therefore heard in a different language (P2-B2/B2-P2/B1-P1), 2)through a B & P event across stories, via events heard in English (B2-P1) or via events heard in Greek (B2-P1), and 3)through the last party event heard (LP = E or G). Third graders' recall was organized through multiple breakfast and/or party events, across and within stories and in a variety of event & story orders, often crossing languages more than once [i.e., B1(E)-B2(G)-P2 (E)-P1(G)], illustrating stronger L1-L2 relations and L1-L2 mobility, compared to first graders. Notably, for all participants, although memory was comparable for events heard in English versus Greek, all recall was stated in English, even when initiated by the experimenter in Greek.

Conceptual Linguistic and Metalinguistic Competence Analyses of story, event, and language order of third grade recall suggested the use of language as a cognitive organizer of memory. For example, the majority(75%) of protocols revealed memory for a B event and then a P event (or vice versa) successively, across stories [i.e., heard in the same language- B1, P2 (P2,B1) or B2, P1 (P1,B2)], overriding the initial story and event order in which such remembered events were initially presented (i.e., B1P1 or B2P2). In fact, some third graders (25%) engaged in the aforementioned memory strategy twice within a given recall protocol [i.e., B1(E), B2(E), B2(G), P1(G)]. Likewise, for the majority (75%) of third graders, the presentation language(LP) of the initial event remembered was consistent with the language used by the bilingual experimenter to initiate recall (i.e., if recall was initiated in Greek (English), the first event remembered by the participant was initially heard in Greek (English). Cognitive awareness of the language of presentation (LP) of remembered events was revealed through third graders' narrative discourse and memory explanations as illustrated by the following excerpts:

"In the Greek part on Wednesday ... and then in the Greek part from yesterday....", "Yesterday's breakfast was in English ... the day before in the English party....", "The first part was ill English ... the other part was Greek", "In the Greek breakfast ... and then at the Greek party ...", and "Yesterday the party part was in English.... but the day before the party part was in Greek...". "I remembered two parties and two breakfasts-one was in English and the other was in Greek-I practiced it", "I remembered the English parts and the Greek parts", "It was easy-each story had a breakfast and a party-one was in English and the other one was in Greek, so I said them over and over".

The metamemory strategies of first graders consisted of brief memory explanations, mostly praising themselves, ("I am just good at remembering") and/or indicating practice ("I practiced the story in my head" or "I made a picture of it in my head over and over").

Discussion

The fact that all participants' memory was comparable for events heard in English versus Greek but all remembered information was stated in English, even when initiated by a bilingual experimenter in Greek, suggests two possible memory strategies with respect to bilingual memory: A)Bilinguals translated all story events heard in Greek into English during encoding, internally, illustrating language neutral memory or B)Bilinguals encoded, organized, and stored all information from story events in the language in which events were initially presented, suggesting language specific memory, and then translated all remembered information heard in Greek into English at recall, externally. Several key findings of the present study appear to refute memory strategy "A" and support strategy "B". First, memory was differentiated for bilinguals, i.e., recall was organized by specific event (B1, B2, P1, P2) versus general event (breakfasts or parties in general) or story (1 or 2). Second, for the majority of third grade bilinguals(75%), the presentation language of the initial event remembered was consistent with the language used by the bilingual experimenter to initiate recall. Third, all bilingual third graders and some bilingual first graders made some reference to language of presentation (LP) of remembered events in their verbal explanations of how they remembered and in their recall narrative illustrated in the following experimenter(E)-participant(P) commentary:

E: (experimenter elicits recall in English)

P: "Just the English parts?"

E: (experimenter elicits recall in Greek)

P: "Do you want only the Greek or the English parts too?"

E: (upon completion of recall) "Why did you say it all in English?"

P: "Isn't that what you wanted?--I can do it again and say the Greek parts in Greek if you want.

Implications & Conclusions

The language behavior competencies exhibited by present study participants are best understood via language ecology, incorporating personal factors, such as prior cognitive and social language experiences, and environmental variables, including the nature and context of a given recall task as well as bilinguals' perceptions of the cognitive and social demands of that task. The communicative competence exhibited by all participants, which denoted a high degree of L1-L2 mobility and strong L1-L2 relations, was consistent with their prior socialization experiences, i.e., consistent cognitive and social use and valorization of English and Greek at home and in school. Notably, communicative competence was not affected by the weekly amount of Greek formal instruction, as evidenced by comparable memory among daily versus weekly cohorts in each grade, in terms of amount of information recalled, memory assembly, and memory strategies. Language questionnaire analyses revealed that these cohorts were quite similar in terms of dual home language usage by parents, siblings, and grandparents as well as high self-ratings of language adeptness and positive attitudes towards both their languages, illustrating the importance of examining the nature and quality of all aspects of language socialization, and the futility of reducing language learning to quantitative amounts of prescribed teaching.

The conceptual linguistic and metalinguistic competence exhibited by third graders, (i.e., cognitive awareness of LP and the possible evolvement of language from a communicative tool to a cognitive memory organizer), facilitated by their significant gains in communicative competence, was not only developmentally age appropriate (Janowsky & Carper, 1996) but congruent with their augmented language experiences, relative to first graders, and claims (Hamers, 2004) that such abilities are enhanced among compound bilinguals that valorize and use both their languages for the same cognitive and social functions. The impact of participants' task perceptions on language behavior highlighted the interplay of cognitive and social factors at the personal and contextual level. That is, what participants "thought they were supposed to do", revealed through recall discourse, may have led them to state all remembered information in English, thus influencing their memory strategies; A)translating Greek story events into English internally during encoding and/or B)translating language specific (Greek) memories into English externally at recall. Consistent with constructivism, bilinguals' understanding and ability to adapt to such environmental task demands(e.g., bilingual intelligence in Piagetian terminology) is linked to their prior knowledge and social experiences, where they learn language appropriate behavior, namely, when, how and with whom to use their language(s), i.e., speak only Greek(English) in Greek(English) class. For Piaget, such learned behavior illustrates the synchronous advancement of cognitive development and language development, whereby children consider the social norms and the perspective of others in their own thinking and speech, i.e., enter concrete operations(age 7) and become less egocentric and more socialized (Chapman, 1988). The capacity of bilinguals to internalize such learned behavior for subsequent individual use, as evident in the present study, constitutes the advancement of their memory functions (Vygotsky, 1986). In the social context of the present study, for example, the fact that participants received all initial instructions by the experimenter in English may have influenced their language behavior, at recall, in this regard.

These findings suggest that bilingual memory is not a static, uniform, cognitive state, where L1 & L2 usage are dictated solely by proficiency and/or dominance, but rather a mosaic of socialization experiences consisting of reciprocal relations between individual, cognitive, social, contextual, and behavioral factors at the internal and external level. The dynamics of such relations are evidenced by the fact that although present study participants may not have been acting bilingually (i.e., speaking in L1 & L2 at recall) on an external level, they may have been thinking bilingually (i.e., using L1 & L2 to help them remember) on an internal level. It is quite probable that some bilinguals may have been more adept at verbally stating their memory strategies than others, due to heterogeneity of prior socialization, making it unlikely to identify what and how each participant was remembering internally and/or externally, further illustrating the dynamic role of language(s) with respect to bilingual memory. Given the fact that memory is an integral part of children's daily learning experiences, the language ecology of bilingual memory should be embraced when attempting to assess bilinguals at a particular point in time on any given task. It is probable that their observed language behavior may not always be an accurate reflection of their language capability or state of mind but a reflection of their adaptability, that is, how they have chosen to adapt to present situational task demands. Perhaps the assessment focus should shift from quantitative matters, such as whether or not bilinguals do use or can use both their languages equally, to more qualitative endeavors, such as what determines bilinguals' language(s) of choice in a given context, and how educators can facilitate children in this regard to capitalize on their learning experiences.

References

Bialystok, E. (1997). Effects of Bilingualism and Biliteracy on Children's Emerging Concepts of Print. Developmental Psychology, 33, 429-440.

Chapman, M. (1988). Constructive Evolution: Origins and Development of Piaget's Thought. New York: Cambridge University Press.

Creese, A. & Martin, P. (2003). Multilingual Classroom Ecologies: Inter-relationships, Interactions and Ideologies. International Journal of Bilingual Education and Bilingualism, 6, (3&4), 161-167.

Cummins, J. (2000). Language, Power, and Pedagogy: Bilingual Children in the Crossfire. New York: Multilingual Matters.

De Groot, A. M. B. and Poot, R. (1997). Word translation at three levels of proficiency in a second language: The ubiquitous involvement of conceptual memory. Language Learning, 47(2), 215-264.

Dufour, R. and Kroll, J.F. (1995). Matching words to concepts in two languages: A test of the concept mediation model of bilingual representation. Memory and Cognition, 23(2), 166-180.

Ervin, S. and Osgood, C. (1954). Second language learning and bilingualism. Journal of Abnormal and Social Psychology, 49, 139-146.

Francis, W. S. (1999). Cognitive integration of language and memory in bilinguals: Semantic Representation. Psychological Bulletin, 125(2), 193-222.

Grosjean, F. (1982). Life with Two Languages: An Introduction to Bilingualism. Cambridge, Mass: Harvard University Press.

Hakuta, K., Ferdman, B. M., & Diaz, R. M. (1987). Bilingualism and cognitive development: Three perspectives. In S. Rosenberg (Ed.), Advances in applied psycholinguistics, Vols. Reading, writing, and language learning. Cambridge: Cambridge University Press.

Hamers, J. F. (2004). A Sociological Model of Bilingual Development. Journal of Language and Social Psychology, 23(1), 70-98.

Hamers, J.F. and Blanc, M. (1990). Bilinguality and Bilingualism. Cambridge, Mass, Harvard University Press.

Heredia, R. R. and Altarriba, J. (2001). Bilingual Language Mixing: Why Do Bilinguals Code-Switch? Current Directions in Psychological Science, 10(5), 164-168.

Janowsky, J. S., & Carper, R. (1996). Is there a neural basis for cognitive transitions in school-age children? In A. J. Sameroff & M.M. Haith (Eds.), The five to seven year shift: The age of reason and responsibility (pp. 33-56). Chicago: University of Chicago Press.

Mahon, M., Crutchley, A., & Quinn, T. (2003). New Directions in the assessment of bilingual children. Child Language Teaching and Therapy, 237-243.

Marian, V., and Neisser, U. (2000). Language-dependent recall of autobiographical memories. Journal of Experimental Psychology-General, 129 (3), 361-368.

Schrauf, R. W., and Rubin, D. C. (2000). Internal languages of retrieval: the bilingual encoding of memories for the personal past. Memory & Cognition, 28 (4), 616-623.

Vygotsky, L.S. (1986). Thought and Language. Cambridge, Massachusetts; MIT Press.

Weinrich, U. (1953). Languages in contact. The Hague: Mouton.

Calliope Haritos, Hunter College School of Education, NY

Haritos, Ph.D., is an Associate Professor in the Department of Educational Foundations.