Language Matters : Studies in the Languages of Southern Africa - latest Issue

It has long been an accepted tenet of language teaching and language learning that teachers are key linguistic 'gatekeepers' (alongside dictionaries, textbooks and the media and so on) in entrenching the approved or 'standard' model of language that is passed on from one generation to the next. In South Africa, though, there is more than one model of English that is taught in the schools, given the complex socio-political history of the country: traditional Standard English, and various Black South African Englishes - generally regarded today as the varieties of English commonly used by mother-tongue speakers of South Africa's indigenous African languages in areas where English is not the language of the majority. This article concerns itself with a particular sub-type of Black South African English, Xhosa English, and with the nature of patterns of interaction in English classrooms where teachers are Xhosa speakers. The article is based on analyses of two separate corpora: one is a corpus of the spontaneous spoken English of mother-tongue Xhosa speakers (De Klerk 2006 provides an overview and a summary of various papers resulting from research into this corpus), and the second is a (smaller) corpus of the teacher talk of mother-tongue Xhosa speakers in the classrooms in which they teach (in schools in Grahamstown). The article explores how the norms presented by these urban, Eastern Cape teachers compare with general characteristic features of Xhosa English (as described in de Klerk 2003), as well as whether the teachers follow the same discourse patterns which characterise L1 classrooms more broadly.

This article presents a system designed for automatic detection of errors in second language learners' writing. The system contains two modules, one aimed at detecting and correcting of spelling errors, and a second module providing a broad detection of higher level language errors. The article describes extensions to existing error detection algorithms and introduces a novel method of context-based ranking of spelling correction candidates based on probabilistic context-free grammars. The performance of the system is evaluated on real data with manually marked and classified errors.

Previous research on student writing in South Africa, particularly Black South African English (BSAE) speakers, tends to focus on differences between BSAE writing and Standard varieties of English, and to fail to notice similarities. There is also a trend to focus on multiple linguistic features, but usually in isolation, without attempting an analysis of features as coherent subsets or subsystems. This article adopts the multifunctional-multidimensional approach of Biber, examining a set of linguistic features that form a coherent set for the expression of involvement in different registers. The Tswana Learner English Corpus (TLEC) and the Louvain Corpus of Native English Student Essays (LOCNESS) are used for the analysis. Results indicate that the TLEC, compared to LOCNESS, is characterised by writing that is less formal and more colloquial; it exhibits more reduction phenomena typically associated with conversation; is less integrated and more fragmented in terms of information presentation; uses more general and potentially more ambiguous cohesive devices; and is more cautious and polite when claims are made. Using Correspondence Analysis, the comparison is extended to include five other registers from the original study by Biber (1988). The analysis indicates that the differences between the two student writing corpora are relatively slight in terms of this overall perspective. As a set, the TLEC and LOCNESS exhibit many similarities to academic writing, but the student writing differs in that it reveals also similarities with spoken registers, rather than other written registers.

The fact that corpora have hitherto been concerned, in the main, with the sampling of the general rather than the specific is only now beginning to be perceived as a weakness both of intention and of design. Except where the assistance of collocation is enlisted (Louw 2003), corpora of natural language often tell us more about language than they do about the functioning of institutions. Institutional corpora do not exist formally. Informal institutional collections of interaction are gradually being pressed into service as our first institutional corpora. Two such collections are brought into contact in this paper. Huthwaite, a firm specialising in business communication and formerly based in Sheffield, United Kingdom (UK) collected recordings of authentic negotiations in the early 1980s with a view to discovering exactly which types of language and behaviour distinguished skilled negotiators from unskilled ones. At about the same time, the first institutional 'corpus' was unwittingly being created. It was a record of a public inquiry at the Sizewell B nuclear power station in the UK. In a machine-readable form this corpus totals almost 17 million words of running text. By searching the Sizewell Corpus for the same features identified by Huthwaite as indicators of the behaviour and language of skilled negotiators, the possibility of data-assisted training in negotiating is opened up for the first time. However, such techniques may finally prove too powerful for use in settings where an ideology of institutionalised insincerity remains unchallenged by human intuition. Access to authentic data and bottom-up training may well unravel the ideology of the boardroom and threaten the institutions themselves. A new class of negotiators and a more manifestly honest world may be temporary products of this uncharted revolution.

South Africa is anecdotally known for its complex system of speech varieties correlating with variables such as ethnicity, first language, class and education. These intuitions (e.g. Lass 1990) require further investigation, especially in the context of a changing South Africa where language variety plays a key role in identifying social, economic and ethnic group membership. Thus, in this research, the extent to which these variables play a role in variety is explored using a corpus approach (the nature of class and race in the corpus is discussed more fully later in the article). The corpus project, focusing primarily on accent, has been undertaken by members of the Department of English Language and Linguistics at Rhodes University in South Africa, collaborating with staff from the Department of Electrical and Electronic Engineering from Stellenbosch University, South Africa. A corpus (the first of its kind) is being compiled, comprising the speech of educated, white, mother-tongue speakers of South African English (as distinct from Afrikaans English, Indian English, and the second language (L2) varieties of English used by speakers of indigenous African languages), and data collection is well under way. This short article aims to describe the aims of the project, and the methodological approach which underpins it, as well as to highlight some of the more problematic aspects of the research.

This article reports on a project on the development of language resources and linguistic tools for Afrikaans. In the first part of the article, the multilingual parallel corpus that was created in the project is described. The corpus consists of aligned translations of the Bible in three languages, English, Dutch and Afrikaans. The second part of the article discusses one of the applications of a multilingual parallel corpus: induction of linguistic tools for a resource-scarce language.

Traditionally, the classification of word categories (or parts of speech) in African languages has been modelled on Western European languages without taking cognisance of the fact that the European languages and the African languages belong to different language types. Moreover, the classification of word categories in the European languages has been typically based on the highly questionable classical view of categorisation in general. These historical factors have had rather adverse implications for the various word categories which have been distinguished in the African languages up to this very day. In this paper, we critically explore these issues, and propose a rather revolutionary framework for the treatment of word categorisation in African languages. This new framework, as we shall point out, has significant implications for the units to be distinguished in a tagset as well as for the tagging of corpora in these languages.