Language and content in bilingual education

Linguistics and Education 16 (2005) 238–252 Language and content in bilingual education Constant Leung King’s College London, United Kingdom Abstract This article suggests that there is a tendency to argue for or against bilingual education in terms of productivity (student attainment expressed as test scores), and that productivity is discussed in terms of division of time, curriculum and speakers. Although this orientation has produced some valuable macrolevel accounts, it does not address the need for close-up interaction data showing how language(s) are used by teachers and students in classroom activities. It is argued that such data is vital for understanding language and curriculum content learning in specific local contexts, which in turn can be fed into wider discussions on pedagogy and policy in bilingual education. © 2006 Elsevier Inc. All rights reserved. Keywords: Bilingual education; ESL; Content-based language learning; Content-language pedagogy; Second/additional language learning 1. Researching bilingual education—diverse interests and perspectives The term bilingual education normally signals the promotion and use of more than one language for teaching and learning curriculum subjects. Bilingual education is, however, manifested in a variety of curriculum configurations and serves many different purposes in a range of contexts. Consider the following examples: • Two-way bilingual programmes in some parts of the United States, for example, use two curriculum languages, English and a local linguistic minority language. These programmes are designed to engender high levels of bilingualism in both linguistic majority and minority students who attend the same classes. • In transitional bilingual programmes, the use of linguistic minority students’ first language is phased out after the first two or three years. Such programmes are designed to promote the E-mail address: constant.leung@kcl.ac.uk. 0898-5898/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.linged.2006.01.004 C. Leung / Linguistics and Education 16 (2005) 238–252 • • • • 239 development of the majority language (e.g. English in the United States) by linguistic minority students (for further discussion, see, for example, Thomas & Collier, 1997, 2002). In Canada, immersion French programmes have been designed for Anglophone majority students to achieve a high level of French, a national minority language. The different varieties of content and language integrated learning (CLIL) programmes in European countries have multiple language and cultural aspirations in relation to a non-first language target language (Maljers, Marsh, Coyle, Hartiala, Marsland, & Pérez, 2001). Some elite International Schools, in different parts of the world, provide ‘a type of dual track programme which enable[s] children of the host country to learn about the language and cultures of other countries, while speakers of other languages, who are temporarily resident, can learn the language of the host country’ (de Mejı́a, 2002, p. 16). Thomas and Collier (2002) report that one of the goals of the French–English bilingual programme in Northern Maine (USA) is to revitalise the local language, French. These examples suggest that bilingual education, in its many manifestations, can be used to serve a number of educational and social goals which include: • • • • • promotion of a majority language in a linguistically diverse society; promotion of a minority language in a linguistically diverse society; promotion of both majority and minority languages in a linguistically diverse society; revitalization of a local minority language in a linguistically diverse society; promotion of foreign language in a foreign language learning context. The diversity of programme types, educational aims and socio-political projects can be seen as a major reason for the broad endorsement of bilingualism by different educational jurisdictions across the world. But this diversity has also been the source of misunderstanding, confusion and contestation. In the United States, for instance, parents, teachers and policy makers have seen competing claims about bilingual education from advocates and opponents in recent times, in relation to California’s Proposition 227, which sought to reduce the state’s support for bilingual education (for further discussion, see Cummins, 2000; Dicker, 2000; Ricento & Burnaby, 1998). Proponents and critics of bilingual education have argued over its merits on the grounds of humanistic values, linguistic and cultural rights and entitlements, and educational outcomes. In this paper, I focus on two issues that are perhaps less commonly discussed: (a) the ways in which the notion of language as medium of instruction is abstracted in scholarly discussions and research; and (b) pedagogic integration of curriculum learning and language learning, foregrounding the need to attend explicitly to issues of language learning, particularly second language/additional language learning in bilingual education. It is argued here that both of these issues should be at the heart of any pedagogic discussion on bilingual education. Given the multi-valency of bilingual education in relation to educational policies in many societies, it is no accident that some recent discussions have been couched in broad social and ideological terms. A key point to bear in mind is that, irrespective of the concerns of the moment, be it social integration, nation building, language and curriculum learning or language revival, the claims for or against bilingual education of any form ring hollow when there is not a clear sense of what happens inside the classroom. We need to know not just whether a particular bilingual programme has ‘delivered’ the desired outcomes; we also need to know how it works at the points where social and curriculum content meanings are communicated, and where opportunities for language development for the individual student may take place. This kind of close-up knowledge 240 C. Leung / Linguistics and Education 16 (2005) 238–252 will enable us to make better sense of the banner headlines of ‘success’ and ‘failure’. It will, perhaps more importantly, also better enable us to examine pedagogic issues in a situated and contextualised way. The myriad ways in which languages are used in the classroom can, in essence, define the characteristics and nature of a bilingual programme. In a critical sense, then, knowledge of the uses of the languages involved is an indispensable aspect of any study or debate in bilingual education. I will first examine some of the ways in which the notions of language(s) and language use have been treated in the published debates on bilingual education. I will then discuss the importance of a detailed understanding of how the concept of bilingual education is enacted in the classroom in diverse circumstances. Next, I will suggest a possible expansion in the range of questions that can be addressed by bilingual education research with reference to some of the work that has emerged from the field of English as a Second Language (ESL). In the classroom, ‘language’ both represents and constitutes curriculum content-based (and wider social) meanings in particular ways. This dual role for language can introduce specific demands as well as affordances for language learning. I will suggest that curriculum content learning and language learning, which are still generally seen as two separate pedagogic issues, should be consciously taken into account in an integrated way in classroom-based bilingual research. Throughout this discussion my principal concern is with how research may enrich our understanding of pedagogic possibilities in bilingual education. 2. Bilingual education: terms of conceptualisation In August 2004, education officials in Arizona published school student performance data, which, they argued, showed that their Structured English Immersion programmes were producing higher levels of achievement in English language, reading and mathematics than bilingual programmes within the state. Structured English Immersion is ‘a form of English-only education that allows for the smallest amount of native-language instruction necessary to supplement an English-only curriculum’ (Arizona Department of Education, 2004). The press release stated that: Students in structured English immersion programs outperformed students in bilingual programs in that they were anywhere from one to four months ahead between second and fourth grade, as much a six months ahead in fifth grade, and over a year ahead from sixth grade on. This means that for students in sixth grade and above, students in structured English immersion programs were over a year ahead of students in bilingual programs. (Arizona Department of Education, 2004) The explicit criterion of evaluation shown in this statement is educational productivity or outcome in terms of student attainment. The measurement of educational productivity is expressed in terms of time efficiency (based on converted test scores, see Appendix A). This approach to the evaluation of bilingual education is by no means exceptional. Indeed a good deal of published research in bilingual education is oriented towards demonstrating the productivity of the use of one or more languages for instruction and school communication. The comparison and evaluation of different bilingual programmes have tended to be couched in terms of how much students have gained (test scores) and how long it takes to make the gain (time in a particular educational programme). Thomas and Collier’s (1997, 2002; also Collier, 1989, 1992 and other earlier papers) large-scale, long-term and methodical analyses of student attainment data from different schooling programmes are a good example of this kind of work. For instance, Thomas and Collier (2002) report that: C. Leung / Linguistics and Education 16 (2005) 238–252 241 English language learners immersed in the English mainstream because their parents refused bilingual/ESL services showed large decreases in reading and math achievement by Grade 3, equivalent to almost 3/4 of a standard deviation (15 NCEs) [normal curve equivalents], when compared to students who received bilingual/ESL services. 50-50 Two-way bilingual immersion students who were former ELLs [English language learners] attending a high-poverty, high-mobility school: 58 percent met or exceeded Oregan state standards in English reading by the end of 3rd and 5th grades. (Two-way is two language groups receiving integrated schooling through their two languages; 50-50 is 50 percent instruction in English and 50 percent in the minority language.) (pp. 2–3) Here the comparative effectiveness or productivity of the two different school programmes is accounted for in terms of language(s) used officially in the curriculum and the amounts of time involved. The findings reported by the Arizona Department of Education, and Thomas and Collier are clearly very different and contradictory. One way of dealing with such contradictions is to attribute the diverse findings to ‘unsound’ data collection design and technical ‘problems’. For instance, it is possible to suggest that, because the Arizona study did not appear to distinguish the different types of bilingual education programmes (one-way, two-way, transitional and so on), and since there is a body of research literature which has reported their differential educational outcomes, the findings of this study are untrustworthy and invalid. Furthermore, one might say that the statistical analyses used by the Arizona study to establish the relationship between the mean scores (representing educational attainment) of the two student populations (in Structured English Immersion and bilingual education) cannot be technically interpreted to show a causeand-effect relationship between students’ participation in a particular educational programme and their scholarly attainment. There are many other possible intervening variables. In one way or another, this kind of methodological dispute is not uncommon. Indeed the handling of data and technicalities of analysis are key points of contestation for proponents and opponents of bilingual education in recent debates (see, for example, Cummins, 2000, chapter 8; Teachers of English to Speakers of Other Languages [TESOL], 2004). Even if, however, all technical problems could be ironed out, it would still not be possible to get closer to an understanding of ‘how’ and ‘why’ a particular bilingual education programme may or may not raise attainment with any degree of confidence. This is because the two studies concerned here, and other similar studies, have relied largely on what Baetens Beardsmore (1996) has referred to as macro-level situational variables (e.g. student population make-up, languages involved, programme types and language policy) and outcome variables (e.g. language and curriculum attainment). The situational and outcome data collected and analysed are abstracted indices of contextual and implementational features, and summary measures of student learning; they are very distant from classroom activities in which the languages concerned are experienced and used, and through which language learning opportunities occur. 3. Dividing bilingual education: time, curriculum and speaker Some of the discussion in the bilingual education literature conceptualises bilingual education in terms of time division of language use. For instance, in an early account of the bilingual education in the United States, with particular reference to Hispanic students, Cohen (1975) discussed the different programmes as follows: 242 C. Leung / Linguistics and Education 16 (2005) 238–252 One model calls for repeated teaching of the same subject matter in both languages – e.g. math in English in the morning, math in Spanish in the afternoon. Another model, alternate days bilingual education, involves teaching a subject matter in one language on one day and continuing on to new content in that subject matter in the other language on the following day . . . (p. 18) In a rebuttal of the claim that bilingual education is not conducive to English language acquisition, Cummins (2000) puts his argument in this way: . . . overwhelming amounts of data on bilingual education programs internationally should have permitted the authors [opponents of bilingual education] to state definitively that the major theoretical argument underlying the push for all-English programs is without merit. The time-on-task (maximum exposure) hypothesis predicts that any form of bilingual education that reduces the amount of instructional time through the medium of English will result in academic difficulties in English. As a theoretical proposition, this hypothesis is refuted by the outcomes of countless bilingual programs evaluated in countries around the globe which demonstrate that students suffer no adverse effects in their mastery of the majority language (English in North America) as a result of spending significant instructional time through the minority language . . . (pp. 208–9) Urow and Sontag (2001) also focus on division of time when they describe a two-way immersion programme at the Inter-American Magnet School, Chicago: . . . although all teachers are bilingual, they take great care to use only the target language (i.e., the designated language of instruction for the time or subject). This means that, for example, for the 80% of the day devoted to Spanish in kindergarten, the teachers give directions, speak, read books, and sing songs in Spanish only. When it is English time, the teachers speak only English. In the early childhood grades, children are encouraged, but not obligated, to speak the target language. They receive positive feedback when they respond with the appropriate language . . . (p. 15) The work in the trilingual schools in Ladin Valleys of South Tyrol, Italy has been reported in a similar vein: Because of the difficulties the students are likely to encounter if faced with second languages (Italian and German) from the beginning, in the first two levels primary school (ages 6–7), parents choose either German-Ladin or Italian-Ladin classes for their children. In the German-Ladin classes, German is the main language of instruction (21 hours per week); the children have 1 hour every day (6 hours per week) of conversation in Italian. In the Italian-Ladin classes, the situation is the reverse. This arrangement means that the children learn to write in one of the two principal languages, German or Italian. In its role as an auxiliary language, Ladin is particularly important during the first two grades of primary school because it supports learning through the medium of German or Italian . . . (Egger & Mclean, 2001, p. 64) In addition to time division, bilingual education is also discussed in terms of curriculum division of language use, i.e. different teaching and learning activities are language-tagged. Urow and Sontag (2001) when discussing the bilingual education programme for elementary students in their school (the Inter-American Magnet School) state that: C. Leung / Linguistics and Education 16 (2005) 238–252 243 Science is an area that teachers use to combine content area study with language development. For example, second graders are given the opportunity to get to know an animal by raising it in their classroom. This firsthand study and observation is conducted in Spanish as students generate observations, questions, and experiments. The students learn the Spanish vocabulary specific to the animal as well as the vocabulary necessary to conduct experiments . . . The focus of the unit then shifts to research, and the language of instruction becomes English. Skills needed in the search for answers to their questions, such as using the parts of a book (e.g., table of contents, index, glossary), are introduced. In the process, students learn science and research vocabulary in English . . . (p. 23) In a description of the two-way bilingual immersion programme at Hueco Elementary School, El Paso, Texas, Calderón and Slavin (2001) state that: At the heart of [a literacy programme] is 90 min of uninterrupted daily reading instruction at each grade level. Beginning in Grade 1 (ages 6–7) the students are grouped across classes and grades by reading level, giving most teachers the opportunity to work intensively with students at one reading level in the language that the students know best. In Grades 2–5 (ages 7–11) reading instruction in the L2 is introduced using the same language-reading-writing methods and materials that are used to teach the students in their Ll . . . (p. 29) Saunders and Golderberg (2001) describe the transitional bilingualism curriculum of two schools in southern California as follows: The general guidelines followed throughout the district for transitional bilingual education programs are as follows. Students are taught literacy/language arts and academic subjects in Spanish while they acquire oral English proficiency (20–45 min of daily oral English language development [ELD] instruction). This program continues until students demonstrate approximately second-grade-level proficiency in Spanish reading and writing and basic oral English communicative skills, as measured by district-developed assessments. When students demonstrate these proficiencies, they qualify to transition and begin English reading and writing instruction; they continue receiving Spanish literacy instruction during this period. (p. 43) In addition to time and curriculum divisions, there is another kind of division, speaker division. In a wide-ranging discussion on the Spanish-English programme of Oyster Bilingual School (Washington, DC.) Freeman (1998) provides an example: There are two full-time teachers in each class; the English-dominant teacher ideally speaks and is spoken to only in English and is responsible for teaching half of the curriculum in English, and the Spanish-dominant teacher ideally speaks and is spoken to only in Spanish and is responsible for teaching the other half of the curriculum in Spanish. (p. 152) These accounts suggest that language division in terms of time, curriculum activities and speakers appears to be common in the practice and study of bilingual education. The use of such abstracted macro-level indices is clearly useful for many descriptive and comparative purposes. However, because of the wide empirical distance between macro-level indices and classroom activities and processes, attempts at pedagogic discussions based on macro-level data have themselves tended to be quite abstract. I will look at this issue next. 244 C. Leung / Linguistics and Education 16 (2005) 238–252 4. Beyond description The use of macro-level variables is sometimes extended beyond description and classification of different variants of bilingual education programmes. Some very ambitious generalizations have been built on such data. Thomas and Collier (1997), for instance, draw on their analyses of a robust database of 700,000 sets of student records of test scores with reference to the types of school programmes these students attended (e.g. two-way bilingual, transitional bilingual and traditional English-medium programmes) over a period of 14 years. They propose a model of language acquisition for school which comprises sociocultural processes, language development, academic development and cognitive development. This is an ex post facto model drawing support from a good deal of interesting and pedagogically helpful research and theorising in adjacent fields, such as the transfer of knowledge from L1 to L2 within a person’s cognitive and intellectual functioning. The model makes good general educational sense. However, because the modelling was based on macro-level analyses some of the generalizations can be quite difficult to interpret at a classroom level. For instance: Academic knowledge and conceptual development transfer from the first language to the second language. Thus it is most efficient to develop academic work through students’ first language, while teaching the second language during other periods of the school day through meaningful academic content. (Thomas & Collier, 1997, p. 43) This recommendation can give rise to a number of pedagogic questions. From the point of view of moment-by-moment interaction in the classroom, does the recommendation mean that any use of students’ first language would be effective? It is understood from classroom research that, while language(s) are clearly an integral part of instructional, regulative and other social activities, they are not always clearly separated in terms of moment-by-moment unfolding activities (see, for example, Bloome, 1994). Students and teachers may also operate quite different interpretive frames in apparently shared activities (e.g. Rampton et al., 2002; Leung et al., 2004). So what counts as academic work for the teacher may not be regarded in the same way by the student. Other pedagogically oriented questions include: would it be necessary to tune the teacher’s use of students’ first language to their current repertoire or competence? How does one go about teaching the second language through academic content? The point is that studies based on macro-level data, irrespective of the technical analyses performed, do not lend themselves to showing how teaching and learning have been organised and enacted in classroom interaction. The reliance on this kind of data necessarily restricts their usefulness for classroom level pedagogic discussions. Conceptualising curriculum and pedagogy in bilingual education in terms of language division across time, curriculum and speaker, and not in terms of what is done with the languages involved in the classroom, can lead to dissonance between principles and practice. For instance, two-way bilingual education programmes have been generally regarded as an effective programme option for language enrichment and all round academic achievement (Thomas & Collier, 2002). According to Roberts (1995), the implementation of two-way bilingual education can be characterised in different ways: . . . classes taught in the morning might be taught in one language, while classes taught in the afternoon might be taught in the other . . . Another possibility is to teach one content class such as math in one language, and then teach the next math class in the other language the following semester . . . A second approach is known as concurrent, in which classes are C. Leung / Linguistics and Education 16 (2005) 238–252 245 simultaneously taught in both languages in a team teaching approach, where one teacher represents English and the other represents another language. In the preview-review technique of concurrent language teaching, one teacher previews the lesson in his/her language, the other teaches the lesson in the other language, and the first reviews the lesson in the first language . . . (pp. 375–376) However, at the shop floor level these descriptions can often be much less clear-cut. For instance, in Wiese’s (2004) ethnographic study of teachers’ understanding of what the official two-way immersion model meant to their school (in California), two staff members offered their thoughts: Yeah it is dual language . . . in the process . . .. We haven’t really defined it like that. It is bilingual education but basically I would say, 1 don’t know if the word is correct, like catered to the need of each child. It is . . . biliteracy’. [2nd grade bilingual teacher] ‘How to describe the program? We are calling it a two-way immersion and I don’t know . . . I think programs evolve and over time they become well founded, where we know this is where it is going. But right now, we are still in that gray area. Are we . . . a two-way immersion? It is kind of looking year to year who the kids are. [Principal of the school] (p. 69) French immersion programmes, as described by Swain and Johnson (1997), are built on principles that include using students’ second language as a medium of instruction for curriculum subjects taught at the same curriculum level as that used in the local first language. De Courcy, Warren, and Burton (2002), in an attempt to explore why ESL children appeared to do well in mathematics in a French immersion programme in Australia, receive the following plausible but somewhat speculative comment on the use of French in lessons by the programme coordinator: [T]he maths [teachers’] association is really delving into quality teaching in maths, with maths language. I think that’s something that we’re addressing because the children need to learn the language to be able to do the maths. So they’re actually learning in a rich way the language of the maths which gives them a very strong foundation whereas [when they’re teaching] in English people assume children understand words because they can use them. But they might not be using them with the depth of knowledge they need to be using them. The teacher assumes that a child understands what ‘between’ means/cause they can say it/and the child actually doesn’t. But if we’re using it, we give them concrete examples/we demonstrate/they really understand it. (p. 117) The teachers in these studies appeared to have some difficulty in articulating their practice. It may be that they were simply insufficiently tutored in specialist ways of talking about their work, although this is unlikely given the seniority of at least two of the staff involved. Another possible explanation for these somewhat hesitant and vague accounts is that in the classroom the use of languages in bilingual programmes is a complex phenomenon which defies simple and easy description. It may be that these teachers are aware of the official models of bilingual education in which they work, but that their classroom experience has diverged from the official rubric and recommended practice, so causing their hesitance. In the next section, I look at examples of close-up classroom research to highlight some of the diverse practices that can exist in bilingual programmes. 246 C. Leung / Linguistics and Education 16 (2005) 238–252 5. Languages in the classroom Discussions on bilingual education have not, of course, been exclusively conducted in terms of language division across time, curriculum or speaker. Issues of language use in classrooms have been addressed extensively. For instance, Brisk (1998) observes that in thematically organised curricula which alternate languages within the thematic unit: Activities are done in one or the other language, depending on students’ language level and availability of materials . . . Yet another approach to language use in content classes is to teach bilingually and allow students freedom of language choice. (p. 108) Freeman (1998), whose description of the programme at Oyster Bilingual School I mentioned earlier, reports that from the school’s standpoint ‘there is no problem if the students need to use their first language to aid each other’s comprehension’ (p. 152). In a description of the European Union-funded multilingual European Schools, Baetens Beardsmore (1993) notes that teachers are expected to exercise their judgements when producing teaching materials: Only for the teaching of languages themselves or for subjects taught through the medium of the Ll are standard text books used . . . For all other subjects teachers . . . are expected to devise much of their own material in order to meet the requirements of their mixed language populations. Whatever materials are used require (sic) slight modifications at this level and teachers may provide multilingual glossaries of the specific subject-matter terminology to ensure that the student has the requisite vocabulary in both the Ll and the L2. (p. 132) In a discussion on how late English immersion (in secondary education) in Hong Kong and Canadian French immersion may differ, Johnson and Swain (1994) comment: In Hong Kong insistence on the need to use the Ll for some purposes, including interpersonal communication, was almost unanimous. In Canada, it seemed to us that there was a division in teaching styles between those who could tolerate a great deal of partial understanding, and those who found this unacceptable. Many teachers show great skill and flair in improvising the means for getting meaning across without using the Ll, but research is needed before it can be concluded that total avoidance of the Ll is the only or the optimal approach in late immersion programmes. (p. 222) Other studies that attempt to look at language use in a more close-up way point to a high level of variation even within a single programme type. I will cite three such accounts covering three contexts: late English immersion in Hong Kong, French immersion in Canada and Malay-English bilingual education in Brunei. Lin’s (1999) descriptions of the uses of English and Cantonese in late English immersion in Hong Kong, where Cantonese-speaking students are required to carry out their secondary education through the medium of English, is a good example of local variations within the same bilingual programme. She finds very different patterns of language use in the English lessons in the four schools in her study: • Classroom A: a Form 3 (third year in secondary school) class of 33 girl students. The teacher appeared to be fluent in English and seemed to be at ease with everyday use as well as with using English for teaching purposes. The students seemed to be comfortable with English. The use of English as a medium for classroom teaching and interaction appeared to be working C. Leung / Linguistics and Education 16 (2005) 238–252 247 well, both in whole-class talk and group discussion sessions. The school is located in a middle class/professional neighbourhood. • Classroom B: a Form 2 class (second year in secondary school) of 42 students (boys and girls). The teacher seemed to speak in English only. The students did not seem to be cooperative in class and tended to speak in Cantonese except when being told to do a specific task in English. The school is located in a government subsidised housing estate. The students are reported to speak only Cantonese at home. • Classroom C: a Form 2 class of 39 students (boys and girls). The students had limited English proficiency for their grade level. The teacher would ask task-related questions in English first but often she had to repeat or elaborate on her questions in Cantonese to get responses from students. When an acceptable answer in Cantonese was offered she would then rephrase the student’s response in English. The school is located in an industrial area and the parents of the students are from manual/service work backgrounds. • Classroom D: a Form 1 (first year in secondary school) class of 20 boys and 10 girls. Of the four classes studied by Lin, the teacher of this class used the most Cantonese. She explained vocabulary, gave directions, made the English texts come alive, explained grammatical points, and interacted with students in Cantonese most of the time. The school is on a public housing estate with a similar socio-economic profile to those in Classrooms C and D. Fazio and Lyster (1998) studied the use of language in what they refer to as submersion French classes (originally designed for mother-tongue French speakers, but also attended by a large number of linguistic minority students whose first language is not French) and immersion French classes (French classes attended by mother tongue speakers of English in English-medium schools) in Canada. They report that: . . . the language arts classes in the submersion context . . . were almost exclusively analytic in their approach to language teaching: the content focus was primarily on language form and most materials entailed only minimal discourse. On the other hand, the French immersion context in the English-language schools proved to be varied in its integration of analytic and experiential instructional options in the language arts classes . . . (p. 314) Baetens Beardsmore (1996) reports that, in his observation of three English-medium lessons in a Brunei context, where a policy of bilingual education in Malay and English was in place, there were different teacher practices at work. In one of the lessons, the teacher encouraged group oral chanting by the students and there was no opportunity for sustained language use during the lesson. In another lesson, a different teacher generated opportunities for the students to actively engage in negotiation of meaning and the teacher focused on meaning as well as on language form. In the third lesson, the teacher prioritised the presentation of the content of a history lesson and used mainly display questions to check students’ factual recognition. These accounts point to the diverse and complex ways in which languages can be used in a range of different bilingual classrooms. They suggest that macro-level factors such as socioeconomic circumstances and local factors such as participants’ language competences, teaching approaches and teacher-student relationships all bear on the ways languages are used. Lin’s observations of late-immersion Classrooms B, C and D signal that, even in schools which can be described as sharing a similar socio-economic status, classroom practices can vary substantially. Any attempt to address the pedagogic value and/or development of bilingual education will benefit from classroom research focussed on this aspect of language use in specific contexts. 248 C. Leung / Linguistics and Education 16 (2005) 238–252 6. Focus on content-language pedagogy In addition to the need for greater understanding of the use of languages in classroom activities, it would be a good idea to explore the demands and affordances of language learning in different curriculum areas. The issues of language development and language pedagogy in bilingual education have tended to be glossed over in much of the macro-factor oriented discussions (see Jacobson & Faltis, 1990). But these issues are at the heart of the enterprise of bilingual education. Language learning, especially second/additional language learning, is not an automatic and universal process for all learners: the vast field of work in Second Language Acquisition (SLA) testifies to this. In bilingual education programmes, students are expected to develop their knowledge and skills in two or more languages through the study of curriculum subjects, not just language. In the ESL pedagogic literature, there is body of work that is concerned not just with language learning in a general sense, but also with more specific questions such as what kinds of language expressions and forms are associated with curriculum materials and activities in different subject areas. This latter focus is equally applicable to bilingual education. Many students in bilingual education programmes are second/additional language learners of one (or more) of the languages involved. My argument is that it is not enough to expect students to learn German and English in a bilingual German–English programme because they have had the opportunity to use these languages in some of their lessons. It is important to relate the general idea of language learning to more specific demands, opportunities and affordances of content-related language learning. Seen in this light, there is every reason to explore the ideas and principles of content-language integration first developed in ESL pedagogy, for use in bilingual research. I will mention some key ideas found in this body of work, which may serve as points of departure for a more detailed discussion. (For further discussion on pedagogic application, see Snow & Brinton, 1997; Snow, Met, & Genesee, 1992). The work of Crandall and her colleagues (Crandall, 1987; Crandall, Spanos, Christian, SimichDudgeon, & Willetts, 1987) can be seen as an early example of curriculum content-oriented second language teaching. Their approach is built on the idea that if school-aged ESL students are to participate in mainstream subject classroom learning, then it makes sense to focus ‘on the ways in which the language is used to convey or represent particular thoughts or ideas’ (Crandall, 1987, p. 4). Subject-specific uses of vocabulary and discourse expressions are identified and classroom strategies are built around these in order to promote both understanding of the subject content and learning of English at the same time. For example, it is pointed out that mathematics uses English language vocabulary and structures in particular ways. The notion of subtraction, for instance, can be expressed by ‘subtract from’, ‘decreased by’, ‘less’, ‘take away’ and so on; language expressions such as ‘If a is a positive number, then −a is a negative number . . .’ are used to elaborate on the axioms of opposites (Dale & Cuevas, 1987, p. 17). Classroom activities designed to promote ESL development are built around the identified content-language. Working within the systemic functional linguistics perspective, Mohan (1986, 1990, 2001) proposes a content-language integration approach which ties language expressions and curriculum content together via a set of underlying knowledge structures. These knowledge structures, such as description and sequence, are found in different subjects across the curriculum. So one finds sequence in narratives, in ordering historical events and when following procedural steps in science experiments. This framework of knowledge is intended to help teachers analyse the key knowledge structures in different subject areas and tasks, and identify appropriate language expressions for teaching and learning for students at different stages of ESL development. Mohan also suggests that knowledge structures can be visually represented in images and graphic forms such as charts C. Leung / Linguistics and Education 16 (2005) 238–252 249 and diagrams. Thus, the use of visual representations such as picture sequences and other forms of graphics can be used to promote understanding of the key content meaning and associated language expressions. Cummins (1992, 1996, 2000) suggests that language proficiency within a curriculum and school context can be analysed in terms of basic interpersonal communicative skills (BICS) and cognitive/academic language proficiency (CALP). BICS is understood to mean ‘the manifestation of language proficiency in everyday communicative contexts’; CALP is conceptualised as ‘manipulation of language in decontextualised academic ‘situations’ (Cummins, 1992, p. 17). BICS tends to occur in situations where the meanings communicated are broadly familiar to the participants and/or the immediate context or action provides supportive clues for understanding; greeting friends and getting food in a student canteen are examples of context-supported BICS; a class discussion on the merits and demerits of the use of pesticide in farming, without any supporting print, sound, visual or video materials, is an example of context-impoverished CALP. These two conceptual categories do not yield precise linguistic descriptions nor do they map on to any specific area of the curriculum directly. But they can be used to estimate the language and cognitive demands of a variety of communicative situations in school. The predicted language and cognitive demands have to be worked out with reference to the content and language learning needs of specific students. In general, second language students tend to acquire BICS relatively quickly whereas the development of CALP used in decontextualised situations is a more complex and long-term process. Pedagogically, second language students, particularly those in the beginning stages, benefit from context-embedded communication, such as learning new information and language expression through hands-on activities and/or with the support of visuals or realia. The work of the above researchers has clearly gone beyond conceptualising language instruction and learning in terms of language per se; they have tried to tackle the issues of how the language of instruction is to be analysed and used in the classroom for teaching and learning purposes (see Leung, 1996, for a further discussion). The attention to the different ways in which language use may be considered in relation to curriculum content and communication processes is clearly relevant to bilingual education. This kind of analysis can be conducted in respect of each of the languages involved in a bilingual programme. For instance in an English–Spanish bilingual programme, in addition to conceptualising bilingual education in terms of how much time or in which curriculum areas English or Spanish should be used, we may now add two further issues. Firstly, how should language-in-use in the curriculum be conceptualised? Secondly, what key target language should be prioritised in teaching and developing appropriate classroom communication, if students’ language and curriculum learning needs are taken into account? The kind of second language pedagogic approaches cited here can be recontextualised and adopted in classroom and curriculum research (including curriculum development and teacher preparation) in bilingual education. They would also add new dimensions to discussions on programme development. In addition to detailed observation and analysis of teacher and student interaction and the attendant uses of languages (as discussed in the preceding section), examination of how the chosen medium of instruction has been used and exploited in teaching materials and classroom processes would open up new angles of research. Combining these two aspects of language use is likely to enrich the classroom research agenda within bilingual education. 7. Concluding remarks Padilla (1990) argues that, in the United States context: 250 C. Leung / Linguistics and Education 16 (2005) 238–252 We should not ask the question: “Does bilingual education work?” Rather, “How can instruction be designed and implemented that maximises the linguistic, cognitive, and social exchanges between students who come from different home language backgrounds?” (p. 22) Bilingual education will play an increasingly important part in both national and international education systems. Research in bilingual education appears to have shifted from classroom level pedagogic enquiries to wider socio-political evaluations (Tsui & Tollefson, 2004). In this paper, I have sought to emphasise the importance of looking at how teachers and students use their languages in teaching and learning activities so that we have a better understanding of what goes on in bilingual education classrooms in different world locations. More specifically, I have argued that bilingual education research should pay greater attention to two related pedagogic issues: the ways languages are actually used in classroom interaction and activities, and the demands and affordances of language learning in the context of curriculum subject learning. While it is important to study the effectiveness of different approaches to bilingual education in terms of macro-level variables such as the choices of languages and the curriculum time officially assigned to each of the chosen languages, there is a continuing need to pay close attention to how languages are used in the classroom. Existing classroom studies show that there is a good deal of variation in the ways that teachers and students use their languages for curriculum and social purposes. Macro-level and classroom level descriptions and analyses should be seen as complementary; both are required for an in-depth understanding. Furthermore, a good deal of the discussion in bilingual education has tended to present language learning itself as a relatively non-contentious issue, once the important programme-wide questions such as which languages are to be used as mediums of instruction have been settled. However, it should be recognised that students in bilingual programmes are also second/additional language learners of one (or more) of the languages used as mediums of instruction; as a consequence, successful outcome of second/additional language learning cannot be taken for granted. Some researchers in second language education have long been concerned with issues of language learning in the context of curriculum subject learning. Some of this work, first developed in ESL can be recontextualised and made relevant to bilingual education research. The classroom research agenda in bilingual education can thus be extended in pedagogically relevant ways. Appendix A Extracts of SAT9 Achievement Data reported by the Arizona Education Dept (2004). (The full data set covers 2nd grade through 8th grade.) Mean scale score n Reading Math 2nd grade Language SEI Bilingual SEI Bilingual SEI Bilingual Bilingual Bilingual Grade equivalent difference 10739 1271 11461 1329 11342 1318 1410 1026 563 551 564 560 545 539 564 595 2 months** 1 month** 1 month** 251 C. Leung / Linguistics and Education 16 (2005) 238–252 Mean scale score n Reading 8th grade Math Language SEI Bilingual SEI Bilingual SEI Bilingual Grade equivalent difference 6402 692 6490 697 6452 700 669 650 674 659 628 610 1 year and 5 months** 1 year and 2 months** 1 year and 3 months** Note: “n” refers to the number of students tested. ** p < .001. References Arizona Department of Education. (2004). The effects of bilingual education programs and structured English immersion programs on student achievement: A large-scale comparison. Retrieved 3rd September 2004. From http://www.public.asu.edu/∼macswan/ade/ade document.pdf. Baetens Beardsmore, H. (1993). The European school model. In H. Baetens Beardsmore (Ed.), European models of bilingual education (pp. 121–154). Clevedon: Multilingual Matters. Baetens Beardsmore, H. (1996). Reconciling content acquisition and language acquisition in bilingual classrooms. Journal of Multilingual and Multicultural Development, 17(2–4), 114–122. Bloome, D. (1994). Reading as a social process in a middle school classroom. In D. Graddol, J. Maybin, & B. Stierer (Eds.), Researching language and literacy in social context (pp. 100–130). Clevedon: Multilingual Matters. Brisk, M. E. (1998). Bilingual Education: From compensatory to quality education. London: Lawrence Erlbaum Associates. Calderón, M., & Slavin, R. E. (2001). Success for all in a two-way immersion school. In D. Christian & F. Genesee (Eds.), Bilingual education (pp. 27–40). Alexandria, Virginia: Teachers of English to Speakers of Other Languages, Inc. Cohen, A. (1975). A sociolinguistic approach to bilingual education: Experiments in the American Southwest. Rowley, Massachusetts: Newbury House Publishers. Collier, V. (1989). How long? A synthesis of research on academic achievement in a second language. TESOL Quarterly, 21(4), 509–531. Collier, V. P. (1992). A synthesis of studies examining long-term language minority student data on academic achievement. Bilingual Research Journal, 16(1–2), 187–212. Crandall, J. (Ed.). (1987). ESL through content-area instruction: Mathematics, science, social studies. Englewood Cliffs, NJ: Prentice Hall Regents. Crandall, J., Spanos, G., Christian, D., Simich-Dudgeon, C., & Willetts, K. (1987). Integrating language and content instruction for language minority students. Washington, DC: National Clearing House for Bilingual Education. Cummins, J. (1992). Language proficiency, bilingualism, and academic achievement. In P. A. Richard-Amato & M. A. Snow (Eds.), The multicultural classroom: Readings for content-area teachers (pp. 16–26). New York: Longman. Cummins, J. (1996). Negotiating identities: Education for empowerment in a diverse society. Ontario: California Association of Bilingual Education. Cummins, J. (2000). Language, power and pedagogy: Bilingual children in the crossfire. Clevedon: Multilingual Matters. Dale, T. C., & Cuevas, G. J. (1987). Integrating language and mathematics learning. In J. Crandall (Ed.), ESL through content-area instruction: Mathematics, science, social studies (pp. 9–54). Englewood Cliffs, NJ: Prentice Hall. de Courcy, M., Warren, J., & Burston, M. (2002). Children from diverse backgrounds in an immersion programme. Language and Education, 16(2), 112–127. de Mejı́a, A.-M. (2002). Power, prestige and bilingualism: International perspectives on elite bilingual education. Clevedon: Multilingual Matters. Dicker, S. J. (2000). Official English and bilingual education: The controversy over language pluralism in U.S. society. In J. K. Hall & W. G. Egginton (Eds.), The sociopolitics of English language teaching (pp. 45–66). Clevedon: Multilingual Matters. Egger, K., & Mclean, M. L. (2001). Trilingual schools in the Ladin Valleys of South Tyrol, Italy. In D. Christian & F. Genesee (Eds.), Bilingual education (pp. 57–67). Alexandria, VA: Teachers of English to Speakers of Other Languages. Fazio, L., & Lyster, R. (1998). Immersion and submersion classrooms: A comparison of instructional practices in language arts. Journal of Multilingual and Multicultural Development, 19(4), 303–317. Freeman, R. D. (1998). Bilingual education and social change. Clevedon: Multilingual Matters. 252 C. Leung / Linguistics and Education 16 (2005) 238–252 Jacobson, R., & Faltis, C. (Eds.). (1990). Language distribution issues in bilingual schooling. Clevedon: Multilingual Matters. Johnson, R. K., & Swain, M. (1994). From core to content: Bridging the L2 proficiency gap in late immersion. Language and Education, 8(4), 211–229. Leung, C. (1996). Context, content and language. In T. Cline & N. Frederickson (Eds.), Curriculum related assessment, Cummins and bilingual children (pp. 26–40). Clevedon: Multilingual Matters. Leung, C., Harris, R., & Rampton, B. (2004). Living with inelegance in qualitative research on task-based learning. In K. Toohey & B. Norton (Eds.), Critical pedagogies and language learning (pp. 242–267). Cambridge: Cambridge University Press. Lin, A. M. Y. (1999). Doing-English-Lessons in the reproduction or transformation of social worlds? TESOL Quarterly, 33(3), 393–412. Maljers, A., Marsh, D., Coyle, D., Hartiala, A., Marsland, B., & Pérez-Vidal, C. (2001). CLIL: Compendium. Retrieved 2004-01-05, from http://www.clilcompendium.com/index.html. Mohan, B. (1986). Language and content. Reading, MA: Addison-Wesley. Mohan, B. (2001). The second language as a medium of learning. In B. Mohan, C. Leung, & C. Davison (Eds.), English as a second language in the mainstream: Teaching, learning and identity (pp. 107–126). London: Longman. Mohan, B. A. (1990). LEP students and the integration of language and content: Knowledge structures and tasks. In C. Simich-Dudgeon (Ed.), Proceedings of the first research symposium on limited English proficient students’ issues (pp. 113–160). Washington, DC: Office of Bilingual Education and Minority Languages Affairs. Padilla, A. M. (1990). Bilingual education: Issues and perspectives. In A. M. Padilla (Ed.), Bilingual education: Issues and strategies. London: Sage Publications. Rampton, B., Roberts, C., Harris, R., & Leung, C. (2002). Methodology in the analysis of classroom discourse. Applied Linguistics, 23(3), 373–392. Ricento, R., & Burnaby, B. (Eds.). (1998). Language and politics in the United States and Canada: Myths and realities. London: Lawrence Erlbaum Associates, Publishers. Roberts, C. A. (1995). Bilingual education program models: A framework for understanding. The Bilingual Research Journal, 19(3/4), 369–378. Saunders, W. M., & Golderberg, C. (2001). Strengthening the transition in transitional bilingual education. In D. Christian & F. Genesee (Eds.), Bilingual education (pp. 41–56). Alexandria, Virginia: Teachers of English to Speakers of Other Languages, Inc. Snow, M. A., & Brinton, D. (Eds.). (1997). The content-based classroom: Perspectives on integrating language and content. Reading, MA: Addison Wesley Longman. Snow, M. A., Met, M., & Genesee, F. (1992). A conceptual framework for the integration of language and content instruction. In P. A. Richard-Amato & M. A. Snow (Eds.), The multicultural classroom: Readings for content-area teachers (pp. 27–38). New York: Longman. Swain, M., & Johnson, R. K. (1997). Immersion education: A category within bilingual education. In R. K. Johnson & M. Swain (Eds.), Immersion education: International perspectives (pp. 1–16). Cambridge: Cambridge University Press. Teachers of English to Speakers of Other Languages (TESOL). (2004). TESOL Response to the Arizona Department of Education Study on Student Achievement in Structured English Immersion Programs. Retrieved 2005-01-05, from http://www.tesol.org/s tesol/bin.asp?CID=1&DID=2627&DOC=FILE.PDF. Thomas, W. P., & Collier, V. (1997). School effectiveness for language minority students. Washington, DC: Center for the Study of Language and Education, The George Washington University. Thomas, W. P., & Collier, V. P. (2002). A national study of school effectiveness for language minority students’ long-term academic achievement. Retrieved 4th September 2004. from http://www.crede.ussc.edu/research/llaa/1.1 final.html. Tsui, A. B. M., & Tollefson, J. W. (2004). The centrality of medium-of-instruction policy in sociopolitical processes. In J. W. Tollefson & A. B. M. Tsui (Eds.), Medium of instruction policies: Which agenda? Whose agenda? (pp. 1–18). London: Lawrence Erlbaum Associates, Publishers. Urow, C., & Sontag, J. (2001). Creating community – Un Mundo Entero: The Inter-American experience. In D. Christian & F. Genesee (Eds.), Bilingual education (pp. 11–25). Alexandria, Virginia: Teachers of English to Speakers of Other Languages, Inc. Wiese, A.-M. (2004). Bilingualism and biliteracy for all? Unpacking two-way immersion at second grade. Language and Education, 18(1), 69–92.