he evolution of two basic premises in 20th-Century structural linguistics created the conditions for the emergence of variationist sociolinguistics as an interdisciplinary field. These premises are cultural relativism and orderly linguistic heterogeneity. Cultural relativism is an anthropological tradition inherited by linguistics, according to which no culture or language of a speech community is classified as inferior or underdeveloped irrespective of the level of Western technology that the speech community has achieved. Under the influence of the principle of cultural relativism, linguists posited the ‘functional equivalence and essential equality of all languages, and rejected mistaken evolutionary stereotypes’ (Hymes, 1974, p. 70). At a first stage the relativistic premise applied to comparisons across languages but when the orderly heterogeneity premise was postulated by variationist sociolinguists in the late 60’s, it evolved to comparisons across different varieties or styles of a same language. No variety within a language would be considered inherently superior to the others as concerns its structure even though functional distinctions among them would be acknowledged.

These new assumptions have motivated a group of linguists to describe the orderly differentiation in many languages. They developed empirical studies of language samples recorded in natural settings and started to correlate linguistic variation with social factors such as ethnicity, regional background, socioeconomic status, age brackets, gender, formality of the interaction and so on as well as with structural linguistic factors. The data so gathered received a quantitative treatment.This treatment was made possible with the introduction of the concept of variable rules, that are designed to meet a general principle of accountability, as Labov (1972, p. 94) explains: ‘Any variable form (a member of a set of alternative ways of “saying the same thing” should be reported with the proportion of cases in which the form did occur in the relevant environment, compared to the total number of cases in which it might have occurred.’ The quantitative treatment was mathematically refined by David Sankoff and associates (see Sankoff, 1988). For each area of variation in a given language it is necessary to identify the constraints that are determinants, in a statistical sense, of the relative proportions of each variant of the variable rule, i.e. of each possible realization of the variable rule. The constraints (also referred to as factors) can be defined as forces that operate simultaneously to make the application of the rule more or less probable (Naro, 1981). There is presently a computer program with several versions, known as VARBRUL, that operates this quantitative analysis with much refinement (Pintsuk, 1988).

EARLY DEVELOPMENTS

From its outset in the 1960’s, variationist sociolinguistics argued in favor of the essential equality of the varieties of any language and had to deal with correlations between a child’s dialect with her educational achievement. In fact, the reading achievement of children from nonmainstream families or from ethnic minorities was in the 1960’s - and still is - well below the national norms in several countries. An example of this is given by Stubbs (1980, p. 141) who reproduces the results of a reading survey carried out by Kellmer-Pringle et al., in 1966, in the United Kingdom. They grouped I 1,000 seven-year-olds into three categories: good, medium and poor readers, using as a parameter their performance on the Southgate word-recognition test. The percentage of poor readers in the upper classes was 7.1 per cent; in the middle class it was 18.9 per cent; in the lower classes, the percentage went up to 26.9 per cent.

Stubbs points out that the study may have a bias toward the overestimation of the reading abilities of upper class youths but nonetheless results like these were consistent with nationwide school achievement of minority children and many scholars tended to explain them with the hypothesis derived from Basil Bernstein’s theory of social codes. Children from the lower classes were speakers of what Bernstein (1971) called restricted code and were deprived at home of the linguistic skills necessary to successful learning. This fact, some influential educators argued, had a negative effect on their cognitive abilities. This explanation came to be known as the deficit model. The sociolinguists found this hypothesis not convincing at all and offered an alternative explanation that was known as the difference hypothesis. They argued firstly that 0 languages and dialects were systematic and showed an equivalent structural complexity, and secondly that the dialect speakers possessed the same capacity for conceptual learning as speakers of more prestigious varieties. As they discussed the differences between nonstandard varieties and the school language, they started to search for other explanations for the reading failure of economically deprived children. Nothing was intrinsically wrong with the children’s linguistic forms, they argued, but with society as a whole, and that included the teachers, who regarded these forms as a sign of stupidity. As Wolfson (1989) observes, the notion behind this attitude was that nonstandard speakers were attempting to speak English (the standard variety) but were too lazy or sloppy and ignorant to succeed. A seminal article written with the purpose of demonstrating the verbal and cognitive capacity of ghetto children was Labov’s (1969) The Logic of Nonstandard English (see the review on non-standard varieties by Corson in volume 1).

The variationist sociolinguists’efforts in regard to the AfTican American Vernacular (AAV) at this point followed two related trends: 1) to reject vehemently the deficit model and the remediation educational policy that it gave rise to; and 2) to explore the relationship of language to reading and new approaches to teaching reading. This is well documented in a collection of papers edited by Joan Baratz and Roger Shuy in 1969 in the Center for Applied Linguistics of Washington, D.C. (Baratz & Shuy, 1969). The general assumption shared by the authors in the volume was that reading problems originated from differences in the linguistic systems of AAV and Standard English as well as in the mismatch between the African-American children’s cultural orientation and the school’s expectations. Labov (1969) made a clear distinction between structural conflicts stemming from differences between linguistic structures, and functional conflicts, which are a cross cultural phenomenon. For him, ghetto children’s reading problems were rooted in a situation of reciprocal ignorance where teachers and students were ignorant of each other’s system. He strongly recommended that teachers learn to make a distinction between differences in pronunciation and grammar and mistakes in reading.

Several authors of the volume proposed the use of dialect readers whereby children would be taught to read first in their dialect and then transfer the reading skills to the standard variety. They believed that if the primers were written in the dialect, the child would be spared the double load of learning to read and also learning a new variety of English. As Shuy (1969, p. 117) put it ‘ the child would read adequately if the material and method were consistent with his linguistic behavior patterns.’ This dialect reading method was identified by Stewart (1969) as a native-to-foreign approach to literacy. There was a strong and a weak version of it. The former advocated the use of a special orthography which would reflect more closely the dialect pronunciation. The weak version accepted Fasold’s contention that alphabetic symbols represent phonological segments on a more abstract level and ‘therefore the main, conventional English orthography is as adequate for Black English speakers as it is for Standard English speakers’ (Fasold, 1969, p. 85). This is presently a well established point (see Stubbs, 1980).

Despite the initial academic enthusiasm, the dialect reader proposal did not prove to be a solution to the reading underachievement of socially marginalized children as the pioneer variationists expected it to be (see Toohey, 1986). Fasold (1990) points out that the idea of teaching children in a stigmatized variety was distressing to society at large, which traditionally holds the standard written variety in high regard. Indeed, such proposals raised much controversy in the U.S. and Europe. Cheshire et al. (1989, p. 7) for example refer to the pamphlet The Language Trap written by John Honey in 1983, which claimed that the doctrine of equality spoused by linguists and sociolinguists was contributing to the declining moral and educational standards in British schools.

There are however reports of good results of the use of spoken dialect in the classroom. Edwards (1989) argues that such use represents a move to a child-centered philosophy that builds on existing knowledge. Accordingly, Bull (1982) claims that Norwegian dialects have been acknowledged in Norwegian schools for over 100 years and this approach has proved successful. Nonetheless the prevailing position nowadays is stated by Fasold (1990, p. 279): ‘At the moment, it seems that language differences are not the greatest problem in teaching reading. Labov would agree, although he is convinced that black and white vernaculars are diverging.’ For Labov (1987, p. 10) ‘the primary cause of educational failure is not language differences but institutional racism.’ But he acknowledges that linguists have yet to make a significant contribution to better school curriculum. According to him (p.c.) the main issue to be investigated is whether or not the nonstandard features in the reading of a dialect speaking child preclude hisher understanding of the reading content. On this issue, Fasold (1990) argues that a teacher can be even more confident that the child understood what she is reading when she translates a standard form of the text into a dialect one, rather than simply decoding the standard forms.

MAJOR CONTRIBUTIONS AND FUTURE DIRECTIONS

We have seen that over the past 30 years the use of dialect in the classroom has been a very debatable issue. As Edwards (1989, p. 317) notes, ‘the co-existence of standard and dialect grammar is now widely recognized as giving rise to problems in educational systems.’ Wolfram & Christian (1989) deal specifically with these problems and emphasize that the notion of grammmatical correctness is always a subjective judgment. Every student should be taught the written standard, as well as the regional oral standard. But they recommend that teachers should take as parameter the social acceptability and not the linguistic acceptability of any particular form. In order to do this teachers must be aware of linguistic variation and information concerning dialect diversity must be part of the school curriculum (see the reviews by Corson in Volumes 1 and 6).

Thus, the current trend is not to choose between the standard and the dialect as a medium of instruction, but rather to consider both varieties as functional components of the teachers’ and the pupils’ repertoire. In a study that combined ethnographic observation and quantitative analysis in a rural school in Brazil, Bortoni-Ricardo (1996) identified four types of teacher-led events according to their interactional configurations, ‘marked by ways of speaking, ways of listening, ways of getting the floor and holding it, and ways of leading and following’ (Erickson & Schultz, 1977, p. 6).

The first type of event was made up of highly context dependent short utterances. The second one consisted of longer stretches of instructional conversation. They were both typical events of orality. The third type was the three-part exchange structure made up of teacher initiation (I), pupil response (R) and teacher evaluation (E) (see Sinclair & Coulthard, 1975). The fourth type of event can be classified as an event of secondary orality, as it consisted of readings or of expositions based on a written plan. The events were divided into utterances and a VARBRUL analysis was carried out. The dependent variable was the utterance and two variants were postulated: utterances produced in colloquial standard Portuguese and utterances produced in popularrural Portuguese. The analysis showed that literacy-related events were carried out mostly in the standard whereas orality-related events were performed in a non-standard variety of Brazilian Portuguese.The continuum of events from type 1 to type 4 runs parallel to a continuum of standardness. The results also showed that classrooms are multidimensional sociolinguistic domains where teachers intuitively monitor their speech according to a widespread system of beliefs regarding school literacy.

This evidence hinges on the debatable issue that basic linguistic traits are not transmitted across group boundaries simply by exposure to other dialects in the mass media or in schools (Labov & Harris, 1986). Of course it is not expected that schools will influence the students’vemacular. Rather the purpose of language teaching in schools is the development of monitored styles that are superimposed on the vernacular. And in fact there is much evidence that the school contributes to the enlargement of the pupils’ stylistic repertoire.

A variationist study of noun phrase agreement in interview style carried out with 4th graders and their parents in a school in Brasilia brings striking evidence of the role of schooling in imparting the linguistic resources that are necessary for the performance of monitored styles (Freitas, 1996). The frequencies of noun phrase agreement in the mothers’ speech were 0%; 26%; 32%; and 84%. In their children’s speech the frequencies were, respectively, 94.4%; 100%; 85.7%; and 100%. A similar pattern of differences in the speech of literate youths and that of their illiterate parents was described in a low-income community in Brazil (Bortoni-Ricardo, 1985), as well as in bilingual contexts.

The influence of school on language acquisition is well demonstrated in a survey using variationist methodology carried out among students enrolled in Ontario’s French-language schools, with the twofold purpose of 1) assisting in the evaluation of the students’proficiency in French and 2) providing data which could be used to develop materials for the teaching of French (Mougeon & Beniak, 1991). Percentage of francophones in the communities where the students came from, use of French at home and grade level were the independent variables in the study. The dependent variable was mastery of the reflexive pronouns in French. The results showed that the students that were predominant users of English at home did not reach the level of mastery of reflexive pronouns in French achieved by grade 2 students from French-speaking homes and led to the argument that it is unrealistic to expect that schooling in a minoritysecond language setting will on its own ensure full development of skills in that language. On the other hand, however, it is noteworthy that there was an increase from 35% to 75% in French proficiency of the students from non-French speaking homes throughout their schooling years. This is evidence of the success of the French program in the schools.

In a recent volume dedicated to the description of variation in the speech of Rio de Janeiro (Silva & Scherre (eds.), 1996) Silva and Paiva discuss the influence of the factor of education on language variation and give a detailed description of 22 variationist studies in Portuguese, in English, in French and in Spanish, in which the lower frequency of the stigmatized variant correlates with the higher levels of schooling.

Finally, there is a trend toward application of variationist findings to the development of teaching materials. Mollica and associates (1996) have been conducting school experimental research in which the experimental treatment is the training of pupils with exercises based on results of variationist studies. The observed outcomes of their work show that this type of teaching can be very effective. Three basic principles are followed in the designing of the exercises: 1) going from larger units to smaller ones (e.g. from discourse to the sentence; from word to phoneme etc.); 2) going from most frequent to least frequent forms; and 3) going from most probable to least probable occurrences.

Over the thirty years of its existence, variationist sociolinguistics has made many contributions to education and has offered quantitative tools that can be used in combination with other approaches in the effort to better the school performance of nonstandard speaking pupils. It has also been the target of much criticism and has certainly overcome the stage of naive belief that the study of variation by itself could solve all educational problems. The main problem, as Labov (1987) has claimed, is not language difference but institutional racism (see the reviews by Baugh in this volume and in Volume 1). However, information on language heterogeneity is a valuable instrument to help fight social inequalities and ethnic prejudices. In order to treat children equally their cultural and linguistic differences must be considered.

University of Brasília

Brazil.

REFERENCES

Baratz, J.C. & Shuy, R.W. (eds.): 1969, Teaching Black Children to Read, Center for Applied Linguistics, Washington, DC.

Bernstein. B.B.: 197 1, Class, Codes and Control (vol. 1), Paladin, London.

Bortoni-Ricardo, S.: 1996, ‘Codeswitching in a bidialectal school’, in J. Arnold et al. (eds.),

Sociolinguistic Variation, Data, Theory, and Analysis, Selected Papers from NWAVE 23 at Stanford, CSLI Publication, Stanford, 377-386.

Bortoni-Ricardo, S.: 1985, The Urbanization of Rural Dialect Speakers: A Sociolinguistic Study in Brazil, Cambridge University Press, Cambridge.

Bull, T.: 1982, ‘Mother tongue teaching and learning: Teaching beginners to read and write in the vemacular’, mimeo.

Cheshire, J., Edwards, V., Miinstrermann, H. & Weltens B.: 1989, ‘Dialect and education in Europe: A general perspective’, Journal of Multilingual and Multicultural Development 10, 1-10.

Edwards, V.: 1989, ‘Dialect and education in Europe: A posscript’, Journal of Multilingual and Multicultural Development 10, 317-323.

Erickson, F. & Shultz, J.: 1977, ‘When is a context? Some issues and methods in the analysis of social competence’, The Quarterly Newsletter of the Institute for Comparative Human Development 1(2), 5-10.

Fasold, R.W.: 1990, Sociolinguistics of Language, Basil Blackwell, Oxford.

Fasold, R.W.: 1969, ‘Orthography in reading material’, in J.C. Baratz & R.W. Shuy (eds.), 68-91.

Freitas, V.L.: 1996, ‘A Variação Estilística de Alunos de 4ª Série em Ambiente de Contato Dialetal’, M.A. unpublished thesis, University of Brasilia.

Hymes, D.: 1974, Foundations of Sociolinguistics, University of Pennsylvania Press, Philadelphia.

Labov, W.: 1987, ‘Are black and white vernaculars diverging?’, American Speech 62(l), 5-12.

Labov, W.: 1972, Language in the Inner City, University of Pennsylvania Press, Philadelphia.

Labov, W.: 1969, ‘The logic of nonstandard English’, Georgetown Monograph on Language and Linguistics, Center for Applied Linguistics, Washington, DC (Reprinted in Labov, 1972).

Labov, W. & Harris, W.A.: 1986, ‘De facto segregation of black and white vernaculars’, in Sankoff, D. (ed.), Diversity and Diachrony, John Benjamins B.V., Amsterdam, 1-24.

Mollica, M.C.: 1996, ‘Sociolingüística e Ensino de Língua’, Paper read at the SBPC Annual Conference.

Mougeon, R. & Beniak, E.: 1991, Linguistic Consequences of Language Contact and Restriction, Clarendon Press, Oxford.

Naro, A.J.: 1981, ‘The social and structural dimension of a syntactic change’, Language 57(l), 63-98.

Pintsuk, S.: 1988, VARBRUL programs, unpublished manuscript.

Sankoff, D.: 1988, ‘Variable rules’, in U. Ammon, N. Dittmar & J. Klauss (eds.) Sociolinguistics, Academic Press, New York, 119-126.

Shuy, R.W.: 1969, ‘A linguistic background for developing beginning reading materials for black children’, in Baratz & Shuy (eds.), 117-137.

Silva, G.M. de O. & Paiva, M.: 1996, ‘Visão de Conjunto das Variáveis Sociais’, in Silva & Scherre (eds.), 335-378.

Silva, G.M. de O. & Scherre, M.M. (eds.): 1996, Padrões Lingüísticos, Tempo Brasileiro, Rio de Janeiro.

Sinclair, J. McH. & Coulthard, M.: 1975, Towards an Analysis of Discourse, Oxford University Press, London.

Stewart, W.A.: 1969, ‘On the use of negro dialect in the teaching of reading’ in Baratz & Shuy (eds.), 156-219.

Stubbs, M.: 1980, Language and literacy, Roulledge & Kegan Paul, London.

Toohey, K.: 1986, ‘Minority educational failure: Is dialect a factor?’ Curriculum Inquiry 16(2), 128-145.

Wolfram, W. & Christian, D.: 1989, Dialects and Education: Issues and Answers, Prentice Hall, Englewood Cliffs, NJ.

Wolfson, N.: 1989, Perspectives - Sociolinguistics and TESOL, Newbury House Publishers, New York.