ElAnt v1n1 - The Teaching of Latin in a Multicultural Society: Problems and Possibilities
THE TEACHING OF LATIN IN A MULTICULTURAL SOCIETY: PROBLEMS AND POSSIBILITIES ( 1 )
Jo-Marie Claassen, Department of Latin, University of Stellenbosch, Stellenbosch 7600, South Africa.(first published Scholia 1, 1992)
Abstract. This article traces some of the problems relating to the politicising of mother-tongue education in South Africa, looks at multilingualism in other countries, and discusses language learning theories in the context of the statutory requirements for admission to the Bar and the teaching of Latin to non-Indo-European mother- tongue speakers.
Because of the many languages spoken in South Africa, medium of instruction in the schools is problematical. Educational problems in Southern Africa are not as great as in India, which has ten writing systems and about 1652 different mother tongues.( 2 ) With changing migratory patterns, even British education has to cope with the needs of pupils from over one hundred different linguistic backgrounds. Medium is now an issue even in previously monolingual and monocultural England.( 3 ) The issue of mother- tongue education, like so many other social issues, has in South Africa unfortunately been politicized and is consequently more problematical than elsewhere.
This paper sets out to explore the issue as it relates to Latin teaching at university level in South Africa. Successful students derive more from their compulsory Latin studies than a facility to decipher phrases in their Roman law course books. Some students are particularly disadvantaged by these Latin studies not because Latin per se is 'too difficult', but because of the basic problem of medium of instruction, exacerbated by the fact that Latin is a so- called 'dead language' that is taught in a severely codified form.
Bicultural Alienation
The phenomenon of bicultural alienation amounting to almost total loss of identity is frequently documented. The problem of multilingualism has different aspects: to be able to join the mainstream of modern African urban life, native African speakers need to acquire at least one, often two, European-based languages. To promulgate an own, African education that would equip the native speaker with the tools to adapt to an 'Africanized' technological and media- controlled era would require the adaptation in South Africa alone of about eleven African languages to this era and its specialized vocabulary.( 4 )
A solution most generally adopted has been to accept English, the international lingua franca of the northern world, as the language of education. This 'solution' is not without its own pitfalls, since it leads to further disorientation and alienation, particularly of intellectuals, as graphically described by Ngugi wa Thiong'o: 'The African Prometheus had been sent to wrest fire from the gods, but instead became a captive contented with warming himself at the fireside of the gods. Otherwise he carried the fire in containers that were completely sealed and for which the majority had no key'.( 5 )
South Africa: The Background
Teaching medium has been a burning issue in South Africa for the last century or longer. Adoption of English as the medium of instruction in schools and African first language teaching cause problems.( 6 ) In the struggle between contending 'colonial settlers', the issue was solved variously in the course of power struggles between Boer and Brit. In the nineteenth century Transvaal, Boer reaction to the suspect 'free thinking' of President Burgers and the teachers he imported from Holland led to a rejection of Dutch as the teaching medium and the preference of English for their children by Afrikaans-speaking parents.( 7 ) Latin was consistently taught through the medium of English.( 8 )
Ironically the imposition of British rule and Anglophile emphasis in the schools, particularly before 1900, evoked fierce pro-Dutch reaction and the final victory during the first quarter of this century of Afrikaans as the general medium of instruction. Even up to 1948 and beyond many Afrikaans-speakers still favoured at least dual- medium instruction for their children, but a severe curtailing after 1948 of such dual-medium schools for white South Africans led to virtual monolingualism in the schools.( 9 )
Paternalism and the Bantu Education Act
Much that is wrong on the educational scene may be imputed to the well-meant paternalism of the traditional wielders of power. The ills of black education do not all stem from Afrikaner actions since 1948. The English language as the 'colonial tool of oppression' carries its own burden of guilt.( 10 ) C. T. Loram, the influential educator of the 1920s and 1930s, first suggested that primary syllabi for black pupils, even those within urban areas, should reflect a 'rural' and 'subservient' culture. Primary classes were consciously conducted in the vernacular, with the best interests of the pupils within their culture, as seen by educators, at heart.( 11 ) The notorious Bantu Education Act of 1954 was built on an established system.( 12 ) The ideological and content switch was now carried through to the high school level. Latin as subject in black schools was thus unthinkable. The medium of instruction generally remained English, although Afrikaans in some cases was used and even imported as a subject for the first time. Writing some fifteen years later, after the inadequacies of the system had become glaringly apparent, the architect of this scheme refused to acknowledge these and reiterated the need for the black pupils' 'suit to be cut according to an "African-type" cloth' ( 13 ), but with the powers that be deciding on the type of cloth. Consciousness of their struggle against linguistic imperialism underlay much of the Afrikaners' insistence on education in the vernacular for black South African pupils.( 14 ) Publications outlining studies in Wales, Belgium and elsewhere propagated 'mother-tongue education' for young Afrikaners and by implication for young Africans.( 15 )
Recent studies have shown that in South Africa, where the switch from the vernacular is implemented during the fifth year of school, some teachers find the medium of instruction an almost insuperable obstacle in understanding and teaching 'content subjects'.( 16 ) The problem of medium is pervasive and has a deleterious effect on the whole educational system. Yet a vast corpus of international publications extols the value of mother-tongue education, particularly in the first grades.( 17 ) It would appear that the Bantu Education Act and more recently the De Lange Commission( 18 ) have laid emphasis on an important but delicate and difficult principle: the literature indicates that preparation of a pupil in his mother tongue for eventual progression to a different teaching medium is a complex and specialized process, involving initial parallel instruction by skilled bilinguals, and then gradual progression, over some years, to the second languagefrom object of instruction to instructional medium.( 19 ) The hermetic sealing off of social contact and cutting-off of black pupils' contact from 'western' (or 'northern') culture, except as mediated by a generation of teachers, themselves deprived of wider cultural contacts, resulted in many cases in sadly stunted linguistic development, which is termed the 'ghettoisation' of linguistic skills. As bilinguals straddling two different and vibrant cultures, pupils should have had the opportunity of blossoming and thriving as functional multi- culturals.( 20 ) Restriction of access to the language of the dominant power led to further handicaps, particularly in legal and administrational access.( 21 ) In South Africa politics have therefore exacerbated the problem of medium of instruction. The ghettoising effect of 'Bantu Education' mother-tongue education is a sad impediment that will take generations to overcome.( 22 ) Problems are the restricted skill, and sometimes inadequate degree of bilinguality, of teachers, cultural isolation of schools, and the perceptions of parents and pupils that mother-tongue education (seen as 'additive' in most cultures) is harmful and inferior (or 'subtractive').( 23 ) This perception is based on solid fact. A new move further to amend the language policy is therefore being opposed.( 24 )
Language medium in schools of the apartheid-based 'House of Representatives' also offers problems. Afrikaans or Xhosa-speaking parents, perhaps regarding English as less 'political', enrol their children in 'English-medium' classes.( 25 ) A situation of potential cultural enrichment is often traumatic for children on the lower end of the socio- economic scale. Insufficient qualification of teachers, consequent on 'ghettoisation,' remains a problem. Of course there are notable exceptions: at university level we deal with exceptional people.( 26 )
In the rest of Africa the medium of instruction in 'western-type' schools has traditionally been that of the locally dominant colonial power. After liberation, awareness of the educational value of mother-tongue teaching was offset by the fear that Africans should be 'imprisoned in their own vernaculars'.( 27 ) Only recently have guidelines for the vernacularisation of primary education been compiled, advocating 'contextual teaching' in the vernacular, that is, the integration of the cultural habits of a particular community. No linguistic hierarchy is suggested; nor teaching solely in the vernacular. Dual-medium presentation, on lines followed elsewhere in the world, is outlined.( 28 )
The Problem of Cultural Illiteracy
Some of our students' problems do not stem from 'Bantu Education' as such but, as with the 'cultural illiteracy' diagnosed as the root of the American educational malaise, relate to an almost universal educational policy, developed by Dewey from the ideas of Rousseau.( 29 ) 'Skills' are elevated above 'content' as the aim of elementary training in reading. Kindergarten pupils from a disadvantaged background progressively fall behind through lack of exposure to culture: what they bring to their education is too meagre to enhance what little they derive from it and invalidates the skills they have acquired. It is even worse when South African schools restrict pupils' contact with western (or 'northern'-type) thought, and subject urban children to a sometimes inappropriate 'rural skills training' that is unrelated to their daily lives. Many system (F. Esterhuyse, 'Saunders Suggests a Bold Initiative', UCT News 19.1 [1992] 3-5).
South African pupils transcend these profound handicaps, achieve Matriculation exemption, and enter the universities to read for degrees in engineering, medicine, the arts and law. Latin is unfamiliar to almost all. Some universities, as in Europe and the United States, have standard programmes to help non-native speakers to perfect their command of English as teaching medium.( 30 )
When English has been a medium of instruction since their fifth year at school, disadvantaged students resent the imputation of inadequacy and dislike being compelled to overload their timetables with remedial classes. These students are great achievers in the eyes of their friends and family, not defective underachievers. Remedial courses are often viewed with suspicion. University teachers should appreciate such students' remarkable accomplishment, inter alia by a stronger valorisation of the students' home languages. This is not a plea for extended monolingualism in the schools or the vernacularisation of university courses, but for a new look at the fact of the bilingualism or multilingualism of South African students. Subtractive bilingualism can be converted to additive bilingualism, particularly in Latin teaching.
Latin for Law in South Africa
The issue of medium of Latin instruction has great local significance. Southern Africa (i.e., South Africa, Zimbabwe and Botswana) is unique in its application of the Roman- Dutch legal system.( 31 ) An ability to interpret correctly the Latin terminology of the legal profession is therefore required of aspirants to the Bar. The South African statutory minimum has recently been relaxed to either matriculation level or a single year at university. No one denies that Latin is a valuable ancillary to legal studies, but some see this as a ploy of 'apartheid-minded' legislators to exclude non-Indo- European first language speakers.
The desirability of Latin studies at South African universities may be taken as given.( 32 ) Latin and Roman studies offer 'cultural literacy' and provide South African law students with more than the ability to grope at the meaning of legal authors. In a multilingual social and legal system Latin-based terminology codifies the intricacies of legal thought. There is a large body of evidence showing that Latin enhances intellectual profi- ciency.( 33 ) Furthermore, study of the classical world as a closed microcosm gives students a standard whereby to compare and evaluate their own environment and a metaphor for the dispassionate discussion of delicate political issues such as democracy, disenfranchisement, liberty, prejudice and power.
Theories of Multilingual and Multicultural Education
There is very little literature( 34 ) on the problem of Latin teaching in a non-Indo-European setting.( 35 ) There is, however, a growing body of informative literature, some extremely idealistic, on 'multicultural education'. The American 'melting-pot' theory was based partly on an erroneous assumption that bilingualism is an undesirable cultural defectthe so-called 'myth of the bilingual handicap'.( 36 ) British experience has shown that a teacher's positive attitude to the pupil's mother tongue and minority culture leads to an enhancement of learning in a bilingual situation.( 37 ) A very idealistic M.Ed. dissertation happily postulates a South African school situation where half a dozen mother tongues are adopted in a multicultural school setting.( 38 ) More than another forty-odd years outside the wilderness of Bantu Education is needed for the training of enough multiculturally proficient teachers.
In the South Africa of 1992 we cannot speak of 'minority education' nor imply that black South African culture, either traditional or urban, is inferior or 'minor'. But for many reasons it lies outside the main stream which our potential lawyers must enter, for which they need a general South African 'cultural literacy'. This real minority, a select group of highly intelligent students, is subjected to the compulsory study of a foreign language from an alien past. University teachers of Latin must acknowledge the great linguistic potential of these multilingual students and hone their teaching strategies to meet their students' needs. Language is the object and the means of their teaching.
Linguistics, Learning Theory and Teaching Theories
Language learning theory derives from both psycholinguistics and sociolinguistics. Consensus is an unattainable ideal. All teaching, including language teaching, is partly an art, partly a science and partly 'inspired'. Teaching constitutes at its best an almost indefinable interaction between teacher and learner that involves the personalities of both. 'Theory' and 'prejudice' converge. Traditional Classics teaching is based on the prejudgment that synthesis of prescribed analytical norms will enable the learner both to reproduce and to decipher Latin texts. Such teaching often results in a loss of awareness of the 'living reality of the language' with an adherence to the 'stereotyped patterns of the textbook' that 'acquire the force of an absolute truth'.( 39 ) Furthermore, in most Latin courses it is not 'Latin as a language' that is taught, but rather 'Latin grammar as an organized system aimed at unravelling linguistic intricacies'. The extremes of the 'direct' or 'natural' method versus the 'grammar- translation' method, between which lie the 'structural' and 'audio- lingual' methods, have given way to a compromise weighted in the direction of the 'grammar-translation' method. We Classicists are once more in line with modern linguistic teaching theory, where the 'cognitive approach' underlies 'communicative language teaching'. We may, however, disregard theories of language acquisition, knowledge and competence a la Krashen. At the level of compulsory elementary Latin at university the question may be begged by stating that students of Latin, as most other students, wish to pass their examinations and that university Latin examinations generally tend to test what the students have been taught, whether it be reading competence (synthesis of grammatical or structural analysis) or grammatical insight (analysis of form for the sake of deciding function).
Grammar, Latin and Instructional Medium
Teaching and testing are therefore not so much of competence in Language 2 (Latin) in relation to Language 1 (English/Afrikaans) but rather of understanding of Language 2 (Latin) in terms of a specialized code (Latin grammar as formulated in English or Afrikaans) in order to attain a non- language-specific grasp of the information conveyed (in transformational grammar terms, the 'deep meaning', also known as 'interlingua') that can generate a surface meaning in any languagen. Whether this process is or can be effective is beside the point: the teacher's or tester's prejudgment that it will be effective directs both teaching and testing strategy.
Language can feature in three ways: as mother tongue, second language (both as object of instruction) or medium of instruction.( 40 ) These categories are not always distinguished. In addition, when teaching occurs with a second language as the medium, the student oscillates between the language in question as medium and as subject matter, that is, between the meaning of the words in which information is conveyed and the content of what is conveyed.( 41 ) This applies equally when a second language is employed as medium to convey a third language (Latin).
Research in Sweden has shown that Finnish trilinguals (Finnish- Swedish-German) with an active knowledge of their non-Indo- European mother tongue have greater problems in learning English (another Indo-European language) than those with only a passive familiarity (understanding but not speaking). Both fare better than monolinguals. From this it is deduced that passive bilingualism facilitates second language learning, whereas active bilingualism delays it. The potential for 'interference' is greater.( 42 ) In the case of non-Indo-European first language speakers the learning process cannot merely be linearly represented as:
(a) L1 <----> (b) L2
Nor can it be represented as:
(a) L2 <---- (b) Interlingua ----> (c) Deep meaning in own (Gramm. code) language L1.n
Where the medium of instruction is neither the learner's mother tongue nor from the same language system, one should postulate a triangular system as follows:
(b) Gramm. code in L3 (or 4)
(a) Latin (L2) (c) Own language (L1.n)
Given the prominence of grammatical terminology in Latin teaching, a four-cornered scheme is most suitable for portraying the relationship of the codes with which the learner is con- fronted:
(a) Latin (L2) ----> (b) Grammatical code
(d) Own language (L1.n) <---- (c) Translation medium (L3 [or 4])
It is (d), the 'fourth corner' of this imaginary rectangle, and consequently any direct connection between it and (a), which are missing in Latin teaching strategy to non-Indo-European language speakers in South Africa.
This does not imply that Latin should be taught though the vernacular. In an African context use of non-mother tongue as a teaching medium is (as shown above) the norm. Attempts to foster vernacular education in South Africa are resented and there are many arguments against further ghettoisation of black education. Just as, politically, the elusive ideal is for the New South Africa to be non- racial rather than multiracial, so the didactic ideal is elusive: a virtually non-linguistic approach to Latin teaching, that is, practical exploitation of students' multilinguality in an attempt to reach students' level of non-verbal conceptualisation (or image-formation) as postulated by Chomsky and the generativists.( 43 ) Effective contrastive analysis, showing similarities and differences between the Latin and African language systems, will help students to understand and master the grammatical code on which university Latin teaching and testing is focused.( 44 ) For Latin grammar teaching in an African context, what should be learnt must clearly be the point of departure in contrastive analysis. Illustration of similarities and contrasts must proceed in various African languages.
It is said that both great divergence and virtual convergence of two languages can impede foreign language learning, the first being crucial to language learners as speakers and the latter to hearers.( 45 ) One may add 'or to readers and interpreters'. Consequently, points of similarity between Latin and the African language system should receive initial stress, to form a base for the non-Indo-European first language speaker, before points of contrast are subsequently highlighted. Divergence of opinions indicates that the 'scale of difficulty' is almost impossible to gauge and is seldom an indicator of preferable didactic precedence. 'Difficulty' is perhaps item- specific rather than overall. A different order may be required for encoding than for decoding a language.( 46 )
Contrastive analysis has many levels. For the purpose of contrast with a so-called 'dead language' the phonological level may be set aside, which leaves the lexical, morphological, syntactic and contextual levels to be dealt with in turn, with increasing complexity and greater pitfalls. Context, which appears as the most complex level, is also involved in every other level. At the lower end of the scale of 'complexity', context involves the familiarity of lexical items within the learners' culture;( 47 ) at the upper end, the legal ambience of Roman Law.
The presumption of one-to-one equivalence of lexical items in two different languages must be dissipated in every fresh learner. At the basic level of foreign language learning it is not a very serious issue. Approximate verbal equivalents for words may be given by means of elaborations or circumlocutions. When vocabulary is presented in context, either contrast or similarity between contextual usage can pinpoint 'meaning' in the target language.( 48 ) Translators engaged on a multilingual vocabulary for a South African edition of Nepos' Vita Hannibalis point out that some words require circumlocution and others need translation into a different part of speech. Because the African language system is adnominal in its syntactic structure and Latin is largely adverbial, the functional appearance of conceptual equivalents often differs.( 49 )
Morphological and syntactic levels can scarcely be separated in any effective Latin teaching. The interrelation of the two levels influences students' perceptions of each when contrasting two language systems. On the morphological level Latin and the African language system appear superficially to be strikingly similar: both make use of inflection to indicate changes in meaning, syntax, aspect and mood. The learner may initially conclude that the greatest point of contrast is that Latin inflects word endings, whereas agglutinative prefixes predominate in the African language system. This morphological contrast is, however, merely superficial; more importantly, in Latin the verb is the focal or 'growth' point in a sentence, whereas in the African language system syntax is regulated by nominal connection.( 50 ) This syntactic contrast must form the basis of contrastive teaching of Latin grammar and grammatical terminology.
Speakers aware of the so-called 'universal rules' of their own language will recognize these rules in another language.( 51 ) Native speakers are often unaware of them in their own mother tongue and must therefore be trained to see them at work in familiar syntactic structures.( 52 ) Ideally students should be made aware of either similarity or contrast between Latin structure and the structure of their mother tongue, without intermediation of yet another structure in another language. The teaching language should act only as the means of communication between teacher and student. In practice the best one can hope for is that both learner and teacher will be aware that there are similarities and differences. A Latin teacher can make students more linguistically aware by asking them to think of the corresponding structure in their mother tongue when a new structure is introduced.
At the contextual level contrasts in idiom may be illustrated as they arise, but these contrasts are minor, once the learner has mastered the difference in focus between the two systems, as mediated by whatever modern Indo-European language is used as the teaching medium. English and Afrikaans have far fewer inflectional forms, but the genius of the two languages lies closer to that of Latin in the sense that their verbal focus is paramount, with the addition of word order as a determinant of meaning.
Frequency is a decisive factor in choice of material. Low- frequency forms and constructions have in the past made up too large a proportion of Latin teaching syllabi: in the limited time of a compulsory one-year university Latin course only the most frequent forms and syntactic usages should receive attention. Here another modern teaching strategy comes into play: error analysis will show common problems of non-Indo- European home language speakers (as opposed to those problems common to all) that require different remedial strategies.( 53 ) The existence of idiosyncratic errors implies interference of the student's mother tongue, an assumption that is often contested. Some research appears to indicate that all foreign language learners make similar mistakes, related to the developmental stage in the learning process.( 54 ) This negates the concept of 'interference'. Yet Latin teachers can attest to the observably 'Germanic' word order in English of Afrikaans- speaking students and obvious 'mother-tongue interference' when both language groups apply word-order rules to decipher Latin sentences.( 55 )
Practical Application of This Research
Theoretical study should relate to practice, that is, to both the normal and the remedial teaching strategies of a particular educational institution. At the University of Stellenbosch pioneering research in efficient Latin teaching was initiated by the late Professors Smuts and Bruwer and since then has been perfected by a team of Latinists.( 56 ) A system of computer-aided Latin learning (CALL) programmes was initiated. It is now in its fifth year of implementation. Students using the system are consistently more effective than those who do not.( 57 ) Perhaps they simply work harder, but this observation seems consistent with some psycholinguistic research that shows computer aid involving the right hand ('creative') side of the brain, which is apparently also important in second and third language acquisition in adults.( 58 )
It is hoped eventually to adapt these CALL programmes to include contrastive material in the various African languages. Contrastive research should therefore ideally be set out in such a way that it will complement existing CALL material.( 59 ) On a simpler level the provision in the computer-aided learning context of multilingual word lists is proposed. Such lists must offer equivalents for the most frequent Latin words without which no student can make sense of even the simplest passage of legal Latin.( 60 ) An experimental list has already been instituted within the extant CALL programme at Stellenbosch. As yet there is no statistical evidence of the efficacy of these word lists for linking the Latin material to be learned to images existing at students' conceptual level; even anecdotal evidence is at present meagre. Some non-Xhosa-speakers consciously choose the 'multilingual' mode when practising vocabulary, apparently because it gives a 'new dimension' to what they are learning. An Owambo- speaking Namibian student presently enrolled avers that his home language helps him in Latin learning, as he can discern similarities in the systems of the two languages. When in class reference is made to similarities between Latin and Xhosa (in the gradation of demonstratives, for instance) this student's face lights up with interest. Even the Afrikaans and English speakers, who predominate at this stage, evince great interest in the existence of different language systems and seem to enjoy having parallels demonstrated. Some English speakers say that they prefer to work in Afrikaans once they have become familiar with the Afrikaans grammatical terminology. One student explained that in 'English First Language' at school he did not learn any grammar. So formal grammar as a linguistic codification is being learned as a new system and then being applied to students' extant language systems.
Education is a pragmatic discipline that relies largely on trial-and- error methods. It sometimes strays into error through the application of fallacious or conflicting theories. The theories postulated above are still purely hypothetical. In correspondence with a Latin teacher from the famous Kamuzu Academy in Malawi on the topic of language medium, I was told that a particularly articulate pupil once explained that she simply had 'two domains' within her thoughtShona and Englishand that she located Latin within her 'English domain'.( 61 ) This makes nonsense of the postulation of the mother-tongue as the essential 'fourth corner' in the 'Latin learning rectangle,' but in the case of this pupil, further questioning would arguably have shown her to be a so-called co-ordinate bilingual, whose 'mother tongue' is Shona-English bilingualism, rather than compound, with English bedded over a Shona base. As a co- ordinate bilingual (English-Afrikaans) whose baby years were spent thinking that English was a women's language and Afrikaans a men's language, I find that my own perception of Latin has entered the specific-language-free conceptual level. I experienced medium crossover at the seventh school year and, being bilingual, I found movement from the first medium (English) to the second (Afrikaans) easy. Having subsequently been taught Latin through the medium of Afrikaans, I now have no difficulty in translating the language or its grammatical descriptions into either language.
If this paper does no more than alert fellow Latinists to the complexity of language learning processes and their obligation to exploit their students' extant bilinguality in a positive way, it will have achieved its major purpose. Admittedly much more remains to be done.
FOOT NOTES
(1) Delivered during the University of Natal Roman Studies Conference at Durban, July 1992.
(2) D. P. Pattanayak, Multilingualism and Mother-Tongue Education (Delhi, 1981) 42, 95.
(3) R. Todd, Education in a Multicultural Society (London, 1991) 17-35, 53.
(9) J. A. Marcum, Education, Peace and Social Change in South Africa (Berkeley, 1982) 19.
(15) E.g., Schmidt [11]; Basson [7]; cf. Williams [7].
(21) Cf. Pattanayak [2] 74, 110, 154f.
(24) The Argus (11 May 1992) 4.
(27) A. R. Thompson, Education and Development in Africa (Basingstoke, 1981) 309-13.
(31) H. J. Erasmus, 'Roman Law in South Africa Today', SA Law Journal 106 (1989) 666-77.
(39) J. Higgins and T. Johns, Computers in Language Learning (London, 1984) 44.
(44) D. H. Harding, The New Pattern of Language Teaching (London, 1967) 51-67.
(51) R. D. Gumb, Rule-Governed Linguistic Behavior (The Hague, 1972) 17-23.
(58) See P. Chary, 'Aphasia in a Multilingual Society: A Preliminary Study', in Vaid [35] 183-97.
Jo-Marie Claassen,
Electronic Antiquity Vol. 1 Issue 1 - June 1993 edited by Peter Toohey and Ian Worthington antiquity-editor@classics.Server.edu.au ISSN 1320-3606