Social categories and theoretical descriptions
Ethnicity and language varieties
A particularly striking example of this kind of conflation is provided in the various labels that have been assigned to a variety of American English, namely 'Black English', 'African-American Vernacular English' and, latterly, 'Ebonics,. l Wardhaugh (1992), for example, claims, ' ... we can be sure that there is such a variety of English as Black English in the United States' (p.326). In the preceding passage, he has described the linguistic features that contrast with standard English, providing empirical evidence for his claim that the variety exists. He continues, 'Those who speak it recognize that what they speak is something different from the varieties employed by most non-blacks' (pp.336-7).
This sentence provides examples of two of the problems we have been discussing. Firstly, the perceptions of the speakers themselves are presented as equivalent to the theoretical description of the variety by the linguist. As we have said, we believe speakers' perceptions to be important sources of sociolinguistic evidence (Preston 1999; Niedzielski and Preston 2000), but we would wish, to distinguish speakers' own accounts from those of the analyst, since the two kinds of description serve different purposes. The second problem is signalled by the need for Wardhaugh to qualify 'non-blacks' by the word 'most'. The passage continues:
Most Americans are prepared to categorize someone who contacts them by telephone as either black or non-black using speech alone as the criterion, and most such categorizations are correct. In cases of mis-categorization, it is usually because of special circumstances: a black person has been brought up very closely with non-blacks, or a non-black has been brought up very closely with blacks. (Wardhaugh 1992: 337)
Now, from our perspective, the alarm bells should begin to ring for the researcher, prompting a re-evaluation of the categories 'blacks' and 'non-blacks'. If the variety really is 'Black English', then all those speakers and only those speakers who belong to the category 'black' will use that variety. However, as both Wardhaugh and any number of studies and personal experiences will testify, there are exceptions in both directions (black speakers who don't use the variety and nonblack speakers who do). Fought (2002), for example, cites a study by Hall (1995) which 'found that some European-American women were more successful at performing a stereotyped "Black identity" on the phone than African-American women', while a manager told Hall that '''the best white woman we ever had here was Black'" (Fought 2002: 456). Such discrepancies mean that it behooves the researcher to analytically detach the variety from the speakers, and to seek a more satisfactory theoretical description of the relationship between agency (speakers) and culture (language varieties).
Let us be clear about this. Our objection to Wardhaugh's use of the term 'Black English' is not that it imperfectly captures the empirical usage of the language form denoted by the term. Few descriptive terms will have a nil leakage of this sort (Bell pers. comm.). Our objection, rather, is to some specific characteristics claimed for this kind of variety. Naming a variety 'Black English' ties together the terms 'Black' and 'English', thus defining a particular, empirically identifiable, language variety in terms of its use by a group of speakers. One problem with this is that the empirical identification of this group is not straightforward, in the way that, say, 'New Zealand English' could be defined by classifying the population of speakers to be investigated as 'every English speaker born in, and currently resident in, New Zealand'. A further problem is that the relationship between 'being black' and the empirical distribution of the language variety is unclear. Are 'black people' those who speak 'Black English'? No, because many of them do not speak it. Is 'Black English' defined as that language variety spoken by 'black people'? No, because many 'non-blacks' use it (witness the rap musician Eminem, for instance).
Part of the difficulty with 'speakers of Black English' as a social category for research purposes- social or linguistic- is that it conflates aggregative features (the number of people who happen to share particular somatic features, including skin color) and collective features (the extent to which individuals using the variation are 'party to' norms and conventions which define membership of that group).
In a color-racist society, to be dark-skinned may mean being predisposed towards using those distinctive language forms which are expressive- and constitutive- of particular forms of social solidarity (see Gilroy 1993). These include identification with various symbolic forms, such as music, for example, as well as ways of speaking. But, as we have said, predisposition is not entailment, so the researcher needs to identify the circumstances in which this mechanism (these structural conditions interacting with these cultural resources) is triggered to produce this particular outcome (these speakers utilizing this variety). The claim that 'Black English' is a description derived from the use by a majority of 'blacks' of a determinate language variety implies a priori a particular causal relationship between being 'black' and speaking in a particular way. 'Black-English' is defined as the language variety that (most) 'black' people speak, but this does not take us far beyond empirical description. What we want to know is why this is the case and, more interestingly perhaps, why some 'blacks' do not use it whilst some 'non-blacks' do. Clearly it is not the 'blackness' of the language that is the causal mechanism. We need, in other words, social categories that allow an exploration of the relationship between language variety, human agency and social relations (being of a certain skin color in a society in which this is perceived in certain, derogatory ways, for example).
One significant study in which these issues have been explicitly addressed is the research into Creoles in the Caribbean by Le Page and Tabouret-Keller (1985). These authors do distinguish between language and speaker, claiming, 'We should constantly remind ourselves that languages do not do things; people do things; languages are abstractions from what people do' (p.188). Le Page and Tabouret-Keller found that their informants were likely to hold stereotypical views about the various 'communities' found in their locality, and that their beliefs about both ethnic categories and language varieties were influential on their behavior. However, from the researchers' point of view, these beliefs were highly problematic, for many of the reasons we have outlined above (such as the variable salience given by different informants to skin color, family descent, religious affiliation and so on). Le Page and Tabouret-Keller's account, in contrast with the traditional correlational studies that they criticize, emphasizes the role of agency:
the individual creates for himself the patterns of his linguistic behavior so as to resemble those of the group or groups with which from time to time he wishes to be identified, or so as to be unlike those from whom he wishes to be distinguished. (p.181)
These 'acts of identity' assume a high priority, and individual agency is paramount: 'A community, its rules, and its language only exist in so far as its members perceive them to exist' (p.205). Now although we are sympathetic to several aspects of Le Page and Tabouret-Keller's account of 'acts of identity', we would wish to take issue with this particular claim on two counts. Firstly, it would appear to deny the World 3 status of language itself, by asserting that the existence of languages is dependent on the perceptions of their extant speakers. Yet, as World 3 objects, languages possess properties- such as their syntactical relations- which are independent of particular speakers. If this were not the case, the hermeneutic recovery of 'dead' texts would be impossible. Although exclusively oral languages do not leave the traces on which researchers must subsequently rely, there is no reason to assume that such languages lack the properties which would allow them to be 'recovered' if, for example, aural recordings were available after the deaths of all their speakers.
Our second objection to Le Page and Tabouret-Keller's claim is related to the first. To insist that languages only exist insofar as their 'members perceive them to exist' is to subject languages to 'the tyranny of the present tense' (Archer 1995). The corollary of their claim would seem to be that the structures of languages are the effect of members' beliefs about languages, of contemporary action and perception. It then becomes difficult, on both counts, to explore the interplay of language with agency, with the efforts of people to modify or maintain some aspect of social reality. To reiterate, we regard it as essential both to recognize and to maintain a distinction between a} social actors' own purposes and choices from among linguistic varieties, and how they perceive these from their own experience (agency); b) the structural enablements and constraints which condition the choices open to them (social structure); and c) the properties of language itself (culture).
