Introduction

In the Postscript to the Case of Dora, Fragment of an analysis of a case of hysteria, written shortly after The interpretation of dreams but published five years later in 1905, Freud (1959b, 1942a) considers the role of Dora's transference, and his failure to detect it or at least perceive its consequences in the course of her treatment and its abrupt termination. He notes that “Psychoanalytic treatment does not create transferences, it merely brings them to light, like so many other hidden psychical factors (1959b: 140) [Die psychoanalytische Kur schafft die Übertragung nicht, sie deckt sie blo², wie anderes im Seelenleben Verborgene, auf (1942a: 281)].” He stresses the extreme difficulty of detecting the transference.

Freud calls attention here to one of psychoanalysis' most ticklish problems at both the clinical and the theoretical level: how does one — a participant in the psychoanalytic encounter — come to understand the complex symbolic nature of that encounter?

Structural and generative linguists and conversation analysts have traditionally assumed that the essential properties of language and interaction are most clearly — if not uniquely — revealed in “ordinary” uses of language. This belief has gone hand-inhand with a focus respectively on isolated sentences and “everyday conversations” as primary sources of data and the conviction that the inaccessibility of much linguistic patterning to the conscious awareness of native speakers renders the latter's reflections on language, which Silverstein (1979) has termed “linguistic ideologies,” beneath contempt. Boas (1911: 71) was perhaps the clearest in his condemnation of “the misleading and disturbing factors of secondary explanations,” arguing that their appearance is a threat to the clarity of ethnographic analysis. Boas argued that linguistic inquiry by contrast offers a “great advantage” in that linguistic categories “always remain unconscious.” This stance, which itself is highly ideological, provides the linguist with a privileged vantage point from which to identify and illuminate fundamental truths that are invisible to non-initiates. The privileging of “ordinary” language use similarly echoes a Kantian aesthetics in its assertion that poetically structured texts reveal only a subjective experience of beauty and not either cognitive knowledge about the world and our means of perceiving it or norms of interaction between human beings.

Since the beginnings of the modern era, the trope of language has underlain the Western figuration of subjectivity. Behind Descartes' cogito argument is a dico variant (“I am, I exist [ego sum, ego existo] is necessarily true each time that I pronounce it, or that I mentally conceive it”), and the deconstructionist turn in the humanities and social sciences suggests that although the royal road to the unconscious may be through dreams, its bedrock is constructed out of the rhetorical structures of language. Not only are the terms that Freud uses for the various defense mechanisms a veritable catalogue of traditional rhetorical devices, but the concepts of what is translated into English as the ego, superego, and id are originally metaphorical extensions of the German pronominal system — das Ich, das Über-Ich, and das Es. The issues raised are deconstructionist in their insistence that our models of subjectivity come from the subtle interplay between rhetoric and reference, Bakhtinian in that all such models depend on genre-specific forms of narration and their metalinguistic relations, and Whorfian in the relativistic implications of the cross-linguistic variability of metalinguistic structures.

This chapter attempts to clarify some of these issues by first examining the metalinguistic structures available for representing speech and thought in what Whorf (1956) called SAE (Standard Average European — primarily the modern Romance and Germanic languages); these include direct and indirect quotation/discourse, and particular attention is paid to the interplay between expressive presentations of subjectivity characteristic of direct discourse and the more propositional representations characteristic of indirect discourse.

Introduction

Having examined syntactic and syntactic–pragmatic variation in some detail, it is time to turn briefly to other areas of SAIE structure that exhibit significant social patterning: morphology, phonetics and lexis. This chapter will also offer a brief comparison between SAIE and other varieties of English in South Africa.

Morphology

Bughwan (1970) and Crossley (1988) describe some of the salient nonstandard morphological features of SAIE. Bughwan provides a catalogue of ‘errors’ found in the writing of a sample of high-school students in Durban, but does not assess their extent intra- or inter-individually. Crossley attempts to determine the features of SAIE verb-phrase morphology by systematic comparison between SAIE-speaking university students and a control group of White English-speaking students, under test conditions. The test involved an adaptation of one devised by Greenbaum and Quirk (1970) that made deductions about non-standard usage from students' responses to sentences which they were asked to change in certain ways. For example, students were asked to change they to he in the sentence They criticise everything I do.

The hypothesis was that students would be less tempted to go beyond the instructions and make additional changes, to render the new sentence fully standard, (criticise to criticizes) if they had the non-standard form in their dialect. In this particular instance eight out of 122 students from the SAIE group did not make the additional change, while all members of the control group did. On the basis of a close statistical examination of the test results Crossley posits fourteen features of SAIE verb-phrase morphology, some of which are given below:

1. (a) zero endings on verbs which take [∂z] in standard English: for example, The boy watch the game.

2. […]

Introduction

South African Indian English (henceforth SAIE) spoken by about three-quarters of a million people, chiefly in the province of Natal, offers the linguist and sociologist an opportunity of examining the dynamics of language shift. This study will first focus on the means by which English has become established as the first language of this speech community, ousting in the process several ‘ancestral’ Indian languages. It is my chief aim, however, to examine the kind of variation that arises in such a situation, and the type of language acquisition involved. At the same time I will describe the more interesting features of the dialect and compare them with second-language varieties of English and (similar) former second languages turned first language.

‘New’ Englishes

English as a dominant world language assumes a variety of forms in various parts of the globe, so much so that it has been for some time now problematic to speak of the English language. Sociologically speaking, it might make greater sense to think in terms of an ‘English language family’ – a cover term for a number of varieties with very different histories, functions and structural characteristics. As Kachru (1988) has indicated, many of the ‘sacred cows’ of English are untenable in the latter half of the twentieth century. It is now the property of the world at large, with the number of interactions in English amongst non-native speakers probably exceeding that amongst mother-tongue speakers, and between mother-tongue speakers and non-mother-tongue speakers.

This is the first study of the syntax of South African Indian English, as spoken in its natural home and neighbourhood surroundings. It is my belief that language study of this sort cannot be divorced from a historical and social context. Earlier studies of the dialect and prescriptive judgements by educators suffer in this regard. Unlike most earlier commentators, I do not believe that the dialect is deficient in any way. It is as systematic and logical as any other. If it has evolved many rules of its own, we must seek to understand the nature of these rules and establish the reason for their existence, rather than condemning them by some simplistic comparisons with the formal norms of upper-middle-class speech and writing. This work is both an attempt at understanding and a celebration of those rules, many of which turn out to co-exist in (new and old) English dialects all over the world.

Pidginisation, creolisation, second-language acquisition

Whereas many students of second-language acquisition (for example, Valdman 1980; Andersen 1980) and some psycholinguists (for example, Slobin 1983) have been amenable to viewing it in terms of processes of pidginisation and creolisation, creolists generally have not. Sankoff (1983: 242), for example, has warned that ‘we have about reached the limit of the usefulness of the terms “pidginization” and “creolization’”. While Bickerton (1977a: 49) had once characterised pidginisation as second-language learning with restricted input and creolisation as first-language learning with restricted input, in a subsequent study (1983: 238) he was to paint them as widely divergent processes:

Slobin (1983: 252) holds the opposite view:

The gathering of data

In a panchronic analysis of the morphology and syntax of a dialect in the late stages of language shift, care has to be exercised in making the data base rich enough to glean all the nuances of natural speech. By panchronic is meant the study of language along both the synchronic and diachronic dimensions. I have already suggested in chapter 1 that there is much in the way of diachronic variation and change as the dialect has moved from rudimentary L3 to L2 to L1. On the synchronic level there is much fluctuation today according to the speaker variables of age, education, social class and ancestral language, and according to situational variables like purpose of the discourse, the nature of the audience/addressees, topic, etc.

A partial explanation for the enormous syntactic variety within SAIE is the uneven pace at which different speakers have moved towards English. As with Bickerton's well-known study of the Guyanese continuum, the movement can be shown to be ‘not without sidesteps and hesitations’ (1975: 17). That is, we are not dealing with a simple linear progression towards the standard, but with a complex system that allows for, and indeed covertly places high value upon, a measure of backsliding according to earlier norms of the L2. Bickerton (1975: 17) could well have been speaking of SAIE when he wrote:

Introduction

My aim in the final two chapters is to develop two perspectives which will assist in an understanding of many of the characteristics of SAIE. The dialect will be viewed not as a reified finished product, but as an evolving dynamic system. This chapter will examine the nature of second-language acquisition, while the next will explore the SAIE continuum in terms of notions fruitfully developed within the field of pidgin and Creole studies. The two views overlap to some extent, and the applicability of models constructed within creolistics is still a live issue in second-language acquisition (SLA) studies.

The view from language acquisition

We will be examining four themes within the field of second-language acquisition that have particular relevance to the development of SAIE: (a) transfer, (b) universals, (c) strategies of second-language learning, (d) parameter setting.

Recent linguistic studies have emphasised that the acquisition of a second language is not the piecemeal hit-or-miss affair that it often appears to be. Learners seem to go through stages in acquiring the rules of the target language. Learning a new language (especially in natural settings outside the classroom) is not done through imitations of surface structures alone, but through deducing (sometimes incorrectly) the rules that underlie the output. Many errors made by the adult language learner are indications of this process of deduction or hypothesis formation: errors may be – paradoxically – evidence of learning.

There are significant similarities between second-language acquisition and the way a first language is learned by children. Like adults learning a second language, children making deductions about the rules of the target language, often overgeneralise some of them (resulting in forms like corned, doed).

Introduction

My aim in this chapter is to examine in detail the kind of syntactic variation resulting from the language shift out of which SAIE arose, via analysis of a single phenomenon – the relative clause. Of a total of 15,530 sentences in the corpus, 543 involved restrictive relative clauses. A small number of non-restrictive clauses (six), infinitive relatives (three), and related constructions like clefts and pseudo-clefts were excluded from the analysis. The 543 relative clauses can be divided into nineteen categories, some startlingly different from options taken in standard English. The use of particular relativisation strategies will be shown to correlate with certain groups which can be characterised in social and linguistic terms, and according to the nature and level of language acquisition. A brief characterisation of the variation in relative-clause usage in British dialects will serve as a prelude to the diversity of SAIE forms.

Relative clauses in English dialects

The standard ways of forming relative clauses via appropriate relative pronouns (who for human NPs, with case variants whom and whose; which for non-human) and their alternatives with that (for any noun in principle) and zero for object relative NPs are too well known to describe in detail here. Information on other dialects is less easily forthcoming. Trudgill (1983a: 41) provides a list of possibilities within regional dialects of England, the first two being standard (and hence non-regional):

1. (1) He's a man who likes his beer.

2. (2) He's a man that likes his beer.

3. (3) He's a man at likes his beer.

4. (4) He's a man as likes his beer.

5. (5) He's a man what likes his beer.

6. […]