Investigating Language Attitudes

Social Meanings of Dialect, Ethnicity and Performance

Peter Garrett, Nikolas Coupland, Angie Williams

• Publisher: University of Wales Press
• Year: 2003
• ISBN:
  9781783162086
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This is a systematic and critical appraisal of the variety of ways in which people's attitudes to language have been researched internationally over recent decades. The authors explain this complex field through clear reviews and commentary on previous work, while also offering a demonstration of language attitude research in one specific and important context, the English language in Wales. In addition to discussing different ways of expressing attitudes, from teenagers' and teachers' attitudes to regional and subcultural variation in attitudes, the book also considers issues such as degrees of authentic Welshness, the impact of rapid social change in Wales.

1

Introduction: the scope of language attitudes

We begin this book with a critical review of the main methods employed in language attitudes research, in order to discuss their various strengths and weaknesses. Some of the methodological issues raised in this review we then explore and develop in a series of investigations that we have conducted into language attitudes in Wales over recent years, focusing mainly on how the main regions of Wales and their associated patterns of English speech are characterized and evaluated. Within this structure, the book has three parallel aims. The first is to provide an overview of approaches to investigating language attitudes. The second is to introduce a range of linked empirical studies, focusing on the Welsh context, demonstrating two broad methodological approaches. The third is to develop a dialogue between these first two aims, to explore how sociolinguistic interpretations are both guided and constrained by the different empirical approaches. Through this, we will address the issue of, and indeed demonstrate, how different research methods produce different insights into language attitudes and sociolinguistic structure, contributing to a multi-faceted account of the ‘subjective life’ of language varieties.

In this first chapter, we begin by considering the nature of language attitudes, since it is their complex and rather elusive nature that brings to the fore the methodological issues considered in this book. We then move on to consider why, for sociolinguistics, it is necessary to study language attitudes, and so why it is necessary to grapple with these methodological problems. We then introduce the main approaches to studying language attitudes, as they have developed mainly since the 1960s. Finally, we set out the main research questions to be addressed in this book, and provide a plan of the book as a whole.

The nature of language attitudes

Let us begin by considering the concept of ‘attitude’ generally, without being concerned too much at this stage about whether it relates to language or to other objects, processes, or behaviours. Despite attitude being one of the most distinctive and indispensable concepts in social psychology (Perloff, 1993: 26), and, indeed, a pivotal concept in sociolinguistics ever since Labov’s (1966) pioneering work on the social stratification of speech communities, defining the concept is by no means straightforward. Researchers have offered a number of definitions. The difficulty undoubtedly stems from the latent nature of attitudes. Allport’s work in the 1930s commented on this hampering characteristic of attitudes research in the following way: ‘Attitudes are never directly observed, but, unless they are admitted, through inference, as real and substantial ingredients in human nature, it becomes impossible to account satisfactorily either for the consistency of any individual’s behaviour, or for the stability of any society’ (1935: 839).

Some authors settle for brief and somewhat general definitions. For example, Henerson, Morris and Fitz-Gibbon (1987: 13) write: ‘In this book, the word “attitude” will be used quite broadly to describe all the objects we want to measure that have to do with affect, feelings, values and beliefs.’ Others offer more elaborate definitions. Oppenheim (1982) includes in his definition some of the many outcomes, including behaviours, from which people try to infer other people’s attitudes. For him, an attitude is:

a construct, an abstraction which cannot be directly apprehended. It is an inner component of mental life which expresses itself, directly or indirectly, through such more obvious processes as stereotypes, beliefs, verbal statements or reactions, ideas and opinions, selective recall, anger or satisfaction or some other emotion and in various other aspects of behaviour. (Oppenheim, 1982: 39)

For our present purposes, we will follow the practice of Cargile, Giles, Ryan and Bradac (1994: 221), albeit with more elaboration, and take a general and simple ‘core’ definition that has an adequate basis of agreement for proceeding, and then establish some of the qualities of attitudes on which there is considerable consensus. Sarnoff’s (1970: 279) statement can be used as a starting point, that an attitude is ‘a disposition to react favourably or unfavourably to a class of objects’. We take it as axiomatic, then, that an attitude is an evaluative orientation to a social object of some sort, but that, being a ‘disposition’, an attitude is at least potentially an evaluative stance that is sufficiently stable to allow it to be identified and in some sense measured.

Beyond this, it is widely claimed that attitudes have a tripartite structure, in that they are said to have cognitive, affective, and behavioural components (for example, Edwards, 1982). They are cognitive in that they contain or comprise beliefs about the world (for example, that learning the Welsh language will help me to get a better job in Wales). They are affective in that they involve feelings about an attitude object (for example, enthusiasm for poetry written in the Welsh language). And they are systematically linked to behaviour, because they predispose us to act in a certain way (for example, to learn Welsh).

In language attitudes, cognitive processes are likely to be shaped by the individual and collective functions arising from stereotyping in intergroup relations. Linguistic forms, varieties and styles can set off beliefs about a speaker, their group membership, and can lead to assumptions about attributes of those members. This sort of categorization is said to serve a number of functions (Tajfel, 1981). At the individual level, the complex social world is made more orderly, and so more manageable and more predictable. Whether they are favourable or prejudiced, attitudes to language varieties and their users at least provide a coherent map of the social world. One way in which this is achieved is through stressing similarities within a category and differences between and amongst categories, simplifying the complex array of individual experiences in social life. At the intergroup level, stereotypes can serve two major social collective functions: a social-explanatory function and a social-differentiation function. The former is the creation and maintenance of group ideologies that explain and defend relations between groups, in particular evaluations and treatment of members of outgroups. The latter concerns the creation, preservation and enhancement of favourable differentiations between the ingroup and relevant outgroups. The contents of stereotypes vary from one intergroup context to another, and are defined by which group function or functions they fulfil in any specific social context. Hence it is possible for people to construct almost any evaluation of a speaker to fit their collective cognitive needs. That is, we have a situation where social stereotypes tend to perpetuate themselves, acting as a repository of ‘common-sense’ beliefs or filters through which social life is transacted and interpreted. In summary, ‘stereotypes constitute a crucial aspect of intergroup communication’ (Hewstone and Giles, 1997: 278).

The affective component of attitudes can sometimes appear to determine an attitude, to the exclusion of the cognitive component (Mackie and Hamilton, 1993). For example, a person may hear a language or linguistic variety which they are unable to identify, but may nevertheless consider it ‘pleasant’, or ‘ugly’, and this may affect their response during the encounter (van Bezooijen, 1994). In contrast, however, Cargile et al. (1994) consider it rare for the cognitive component to evoke judgements that are devoid of affective content, and indeed most would claim that attitudes always have a strong affective component (Perloff, 1993: 28). The third component – behaviour – is where much controversy lies in the study of attitudes, and this issue is dealt with separately in the next section.

Although most theorists appear to agree that there are affective, cognitive and behavioural aspects to attitudes, a number of models thread these together in different ways. The simple ‘tripartite model’, which is so often referred to in language attitudes work, was outlined by Rosenberg and Hovland (1960), and supported by subsequent studies by Ostrom (1969), Kothandapani (1971) and Breckler (1984). This model claimed that affect, cognition, and behaviour emerge as separate and distinctive components of attitude, and it has been criticized for prejudging a relationship between attitude and behaviour (Zanna and Rempel, 1988). For many professional persuaders, most notably advertisers and politicians, this is the relationship by which much of the justification for the study of attitudes stands or falls (Perloff, 1993; 79); will surveys of attitudes allow them to predict actual behaviour?

Where do attitudes come from?

When we talk about attitudes, we are talking about what a person has learned in the process of becoming a member of a family, a member of a group, and of society that makes him [sic] react to his social world in a consistent and characteristic way, instead of a transitory and haphazard way. (M. Sherif, 1967: 2)

This view locates attitudes as a fundamental part of what is learned through human socialization. It also emphasizes the durable qualities of attitudes as socially-structured and socially-structuring phenomena. However, there is by no means unanimous agreement on these points. The persuasion literature (for example, Sears and Kosterman, 1994: 264) points to differing levels of commitment in attitudes. Some attitudes are superficial and less stable, and others are more enduring. Evaluative responses may be so superficial and unstable that they might be labelled ‘non-attitudes’ (Ostrom et al., 1994), where people might just make up an evaluation on the spot, perhaps as a first-reaction phenomenon to a new topic, or to one that is too complex to evaluate fully. On the other hand, attitudes that are enduring are sometimes seen as being acquired early in the lifespan and unlikely to change much in later life (Sears, 1983). As we shall argue later, there is evidence that, like language itself, some language attitudes are acquired at an early age, and so, following the point above, are likely to be relatively enduring.

Nevertheless, the claim that attitudes can even potentially be stable and enduring is itself by no means uncontroversial. Potter and Wetherell (1987), for example, arguing for a discourse analytic perspective, paint a picture of individuals’ evaluative stances unfolding in social interaction, and changing from moment to moment, demonstrating considerable variability and indeed volatility. They claim that traditional attitude measurement misses this dynamic and constructive process. We address the link between attitude and discourse in the series of investigations we report in the second part of this book. We certainly agree with Potter and Wetherell that discourse – in the sense of spontaneous face-to-face social interaction through language – is a rich and dynamic locus for doing social categorization and social evaluation. More than that, what we are calling ‘language attitudes’ can themselves be stereotyped responses to community-bound ways of speaking, to discourse styles as well as to dialect varieties in the conventional sense. On the other hand, we see no value in restricting the study of social evaluation to the qualitative analysis of talk in interaction, as Potter and Wetherell imply we should. These arguments are caught up in much wider debates about quantitative/qualitative and empiricist/interpretive designs for research, and we return to some aspects of them in later chapters. But our starting point for the volume is an open stance on method and interpretation, and one that includes attempts to generalize about community-level phenomena, including subjective phenomena. The methodological concerns of this book are anchored more in the group-focused empirical work in sociolinguistics and the social psychology of language. The research discussed in the latter half of the book is concerned with attitudes of groups about other groups. In particular, we aim to construct a geolinguistic atlas of attitudes in regional Welsh communities, rather than to conduct an in-depth investigation of individuals’ attitudes and how these may be variably constructed in social interaction. The theoretical issues of attitude stability/ephemerality, and of context-dependent versus context-independent attitudes, are nevertheless important. Even when social evaluations can be shown to be variable across or within social situations, this does not preclude the existence of stable subjective trends existing at higher levels. In much attitudes research, a degree of variability or ‘systematic variation’ (Potter and Wetherell, 1987: 45) is not seen as seriously prejudicial to the notion of durability, any more than systematic language variation in the speech of an individual severely problematizes the notion of someone ‘having a dialect’, sharing features with others at the level of the community. Social judgement theory (for example, C. Sherif and M. Sherif, 1967), which is often employed as an explanatory framework for attitude change, suggests that people operate with an ‘anchor’ position, but will tend to move comfortably within a finite latitude of evaluations that they find acceptable. This issue is revisited when considering attitudes and behaviour below.

Another generally accepted characteristic of attitudes is how they function as both input to and output from social action. This is of particular importance in educational research, and in areas such as language planning, but it can also be invoked to explain the role of attitude in both the reception and production of language. For example, Baker (1992: 12), focusing on Welsh-language education, sees attitude towards Welsh as an important input factor. A strongly favourable attitude towards Welsh may provide the impetus to high achievement in a Welsh-language programme. Conversely, success in a Welsh-language course for beginners may foster a more favourable attitude towards the language. Educationists and language planners often work with such issues in the hope that attitudes will ultimately serve a double function, as both a presage (input) and a product (output) ingredient. Beyond the educational context, in terms of the everyday language use of individuals, since language attitudes and the sociocultural norms that they constitute are an integral part of communicative competence (Hymes, 1971), they would be expected not only to affect our responses to language users around us, but also to allow us to anticipate others’ reactions to our own language use. So we may modify our speech in an attempt to gain from others particular reactions that we seek (for example, to be seen as trustworthy, educated, from a particular region, competent, an ideal person to employ, or to gain approval from the teacher, etc.). Here too then, attitudes may be seen in terms of input and output, completing a cycle of influence between language variation and social cognition. Indeed, it has been argued from this dynamic relationship between ‘language’ and ‘language attitudes’ that the two need not be separated conceptually (Giles and Coupland, 1991: 59). However, when attitudes are considered in terms of input and output in this way, they again are being considered in relation to (as input to, or output from) behaviour.

Attitudes are also seen as complex phenomena in the sense that they can have many facets and manifestations. For example, if we wanted to investigate ‘students’ attitudes towards their Spanish-language lessons at school’, we would need to identify the relevant facets of such attitudes: what do we mean – and what do the students mean – by ‘Spanish-language lessons’? Facets are likely to include a host of components of communicative events, such as teachers, classmates, teaching methods, course materials, perhaps even the room in which the lessons are held, quite apart from the Spanish language itself. Manifestations concern how we think these attitudes will reveal themselves: that is, what will we look for empirically in our studies and try to assess? We might interview the students individually or in pairs or groups. We might get them to write essays for us about their Spanish-language lessons. We might ask them to circle numbers on attitude-rating scales. We might try to infer their attitudes from their productivity in these lessons. We might try to assess their level of attention in the lessons (for example, by counting how many times students put up their hands to ask relevant questions), and infer attitudes from this behaviour. We might want to assess a number of different manifestations to see if they tell us the same story, or to see if they seem to tell us different things. Comparing different manifestations is a central concern of this book.

The problematic relationship between attitudes and behaviour

A common-sense view about the relationship between attitudes and behaviour can lead people to assume that if they are able to change someone’s attitude towards something, they will also change that person’s behaviour. It can also lead people to assume that they can confidently infer someone’s attitudes from the way that that person behaves. In addition, the assumption is sometimes made that if we can get someone to behave in a certain way, their attitudes will ‘look after themselves’.

Much advertising and marketing, in fact, bases itself on such assumptions. To take the first of the above, an advertiser might, for example, try to get men to associate a certain make of car with speed and masculinity, on the assumption that such changes in attitudes towards the car will lead to more men buying that make. To take the last point in the previous paragraph, marketing managers are keen to get us to try out free samples, on the assumption that, having tried out the product, we will then develop favourable attitudes towards the product. Indeed, Festinger’s (1957) theory of ‘cognitive dissonance’ proposes that we prefer to keep our beliefs, attitudes and behaviour aligned.

However, there is considerable evidence from attitudes research that attitudes and behaviours may at times be far removed from such alignment (for example, Wicker, 1969; Hanson, 1980). Various explanations are provided for this. Many of these reflect the method or context of attitude measurement (the main focus of this book). For example, Ajzen and Fishbein’s (1980) ‘theory of reasoned action’ stresses the social context within which any individual operates, and ho...