Background

Research on language and identity has experienced an unprecedented growth in the last ten years. The time when scholars in the field needed to advocate for the centrality of language in the study of identity (see for example, Benveniste 1971 in linguistics or Bruner 1990 in social psychology) seems far away indeed. Research in fields as diverse as anthropology, linguistics, psychology, sociology, history, literature, gender studies, and social theory, among others, has now firmly established the fundamental role of linguistic processes and strategies in the creation, negotiation and establishment of identities. It is impossible to give a comprehensive view of the theoretical work in all of these areas and of how it has shaped identity studies. Our aim with this introduction is more modest: we want to briefly discuss some of the approaches and concepts that have had the greatest impact on current visions of identity, beginning with background perspectives and then turning to central constructs underlying the chapters in the volume. We then present an overview of the volume and a conclusion recapitulating some of the common ground among the contributors.

Background perspectives

Here we describe several approaches to the study of discourse and identity that pervade the chapters in the volume. We begin with those that have become widely accepted in research on discourse and identity and conclude with some that produce potential divisions in the ways scholars examine discourse and identity.

Introduction

The metaphor of the ‘double arrow of time’ points to the problem I wish to explore in this paper, namely, the implications for research and theory about identity development of two contrasting approaches to the functions of temporal order in narratives: clock/chronological vs. narrative/experiential models of time. In discussing this problem, I will draw primarily on my studies of life history interviews with craftartists and sexual abuse survivors. I will argue that more is at stake for narrative theory and research in the choice made between these two models of time than the specification of narrative structure. And, as well, that the differences are consequential for theory and research on human development, raising questions, for example, about research strategies that rely on a clock-time view of causality and on prediction as a criterion of adequacy for such studies. Further, the alternative narrative-time model underlines the importance of context in the production of narratives by showing how temporal ordering is a function of both cultural preferences for well-formed stories and the situated nature of storytelling, for example, whether elicited in interviews or expressed in the course of naturally occurring conversations. Thus, addressing problems of time and temporal order leads us to general and complex issues about relationships between context, structure, and function in narrative studies of identity.

Introduction

For most of the twentieth century, the dominant North American view of mothering was that one woman – the biological mother – should take almost exclusive responsibility for taking care of children during their formative years, and conversely that children need “constant care and attention from one caretaker, their biological mother” (Glenn 1994: 3). This view is called “the ideology of intensive mothering” by Sharon Hays (1996). It “declares that mothering is exclusive, wholly child centered, emotionally involving, and time-consuming. The mother portrayed in this ideology is devoted to the care of others” and self-sacrificing (Arendell 2000: 1194). An intensive mother is held and holds herself accountable for keeping her children fed and housed and “for shaping the kinds of adults these children will become” (Hays 1996: 108). Intensive mothering ideology is dominated by and exhibits a logic of family and private life that requires a moral commitment to “relationships grounded in affection and mutual obligations” (Hays 1996: 152). Among other things, intensive mothering ideology assumes a seamless progression from conception to birth. However, some women's abilities to manage this trajectory is fragile and unpredictable; an easy pregnancy followed by the birth of a healthy infant cannot always be assumed.

Introduction

Contemporary approaches view identity as a discursive construction that can be conceptualized further through a number of different data analysis frameworks. In relation to the understanding of teacher identities, the more familiar analytical approach has been generated from different versions of critical discourse analysis. Accordingly, the intention is to explore teacher talk as a space for the articulation and the repression of “voice” (Bloomfield 2000; Britzman 1992; Ellsworth 1989; McWilliam 1994). This can be considered as a “top-down” approach because the voices expressed and withheld – in interview talk, for example – are embedded in the wider ideologies and discourses of power that constitute educational and other cultural institutions. From this perspective identity is represented and shaped through the social and discursive practices that are available to individuals and groups at particular moments. Consequently, as members of particular discourses, individuals are positioned to speak, think and act in particular ways and are able to take up or refuse that positioning (Davies and Harré 1990; Gee 1996).

A much less familiar approach to teacher identity analysis is generated from ethnomethodology and conversation analysis (Baker 1983, 1984; Johnson 2002a, 2002b; Paoletti 2000, 2001, 2002). Using this “bottom-up” approach, researchers have become engaged in conceptualizing identity as “the set of verbal practices through which persons assemble and display who they are while in the presence of, and in interaction with, others” (Hadden and Lester 1978: 331).

Introduction

In this chapter I explore the role of narrative in the construction of hegemonic identities by analyzing two narratives told by members of a college fraternity who are male, heterosexual, white, and middle class. I show that these men rely on shared cultural knowledge, or Discourses (as discussed by Fairclough 1992; Foucault 1980a, 1988) to structure the crucial context for the interpretation and performance of their local narratives, a practice which also reinforces these Discourses. I also explore how the narratives help the speaker create a local stance with respect to his hearers, and then how the local stances, the narrator's “story self,” and Discourses interact to create identity and provide material for later identity performances. It is through this interaction among stances, identities, and Discourses, I argue, that hegemony works and Discourses circulate throughout society.

Hegemonic identities

I understand identity as a person's relationship to her/his social world, and I use the term “hegemonic identities” rather than “powerful identities” because people in hegemonic positions do not always feel powerful, and they in fact may not directly dominate anyone. Hegemony, as described by Gramsci (1994), involves maintaining dominant social positions through less obvious but more basic means than direct coercion: for example, by controlling the basic ideologies in a society rather than ruling by force.

Introduction

The focus on narrative studies across the humanities and social sciences reflects a shared concern with the interpretation of subjective experience. Life stories may be, as Mark Freeman asserts, our best “inroad into the phenomenon of self-understanding and selfhood” (1993: 6). Indeed, by listening closely to talk, how tellers describe who they are and where they come from, life stories allow us to explore subjective understandings in great complexity and draw interpretations about how persons make sense of self and world. If we accept the premise that narratives allow us special insight into the process of identity, then, a more adequate understanding of life narratives should help us to better understand the character of identity. We might, therefore, begin by asking, what is a life story? How are we to read and interpret life stories? Are life narratives the product of the person, the situation of telling, or something else?

When interpreting a life story, it is common practice to consider the individual level of analysis. After all, there is a real person who is sitting before us, describing the details of her/his past. Tellers, typically, use the first-person singular and speak with their own distinct voice. Elements of personality and our unique life experience are important contexts in understanding a life story (McAdams 1996). Such a tradition, beginning with Freud's case studies, and on through Henry Murray and Robert White's work, allows us to see how the contours of the past are shaped through the eyes of single persons with their unique developmental history and life circumstances (Runyan 1982).

In the general introduction we talked about the interconnectedness of individual and social processes in the formation and presentation of identities. The chapters in Part II look closely at ways in which social processes, ideologies and institutions interact with individual histories, behaviors and needs in the discourse construction of identity in different contexts. The discursive configuration of the self can take shape at many levels and in many ways. It may result, for example, from direct use of categorization devices through which people assign themselves and others to different social groups or sets of social networks. However, often identity claims are made indirectly, for example through the careful insertion and management of stories or through recourse to shared assumptions and social knowledge about the meaning of words used to describe self or others. Also, importantly, identities projected and constructed in interactional situations are the result of reciprocal positionings by the interlocutors and of their negotiations over the pertinence of roles, actions, attitudes and behaviors in certain social situations. Finally, identities interact with ideological prescriptions about roles and relationships in specific domains of social action that assign preferred properties, needs and desires to individuals.

The pressure of ideologies and social conventions on identity production is at the center of Robin Lakoff's contribution (Chapter 5).

Introduction

There is much more to late modernity than the comfort it has brought to our daily lives: this age of rapid and sometimes radical change in which we live has also rendered more complex the interactions between human beings. One example of this complexity is the bureaucratization of interpersonal relations, brought about by the expanding commercialization of services. Modern society all too often trivializes the contact between service providers and clients, and this has resulted in efforts to make direct contact between individuals and organizations more frequent. In past decades, high-quality relationships with customers were considered a key element of success in bureaucratic organizations (Dubar 2000: 113). Yet the focus on the customer has further problematized the question of communication in individual–organization interactions. The client's evaluation of the service is directly related to his/her expectations, which are invested with subjectivity. The success of the company's engagement in providing a service that will satisfy the client depends, therefore, on communication that favors a better reciprocal understanding and a shared sense of the actions of the participants (Zarifian 2001).

The focus of this chapter is on communication in a health insurance service. One relevant aspect of this communication is the negotiation of the sense of health and the contractual conditions that regulate the client/company relationship. In this context, the processes by which the participants construct their identity become crucial to mutual understanding and to the consequent validation of the service provided.

Introduction

Many have argued that narrators can partly construct themselves when they tell autobiographical stories. For this reason, autobiographical narrative has been proposed as a therapeutic tool (Anderson 1997; Cohler 1988; White and Epston 1990), as a means to critique unjust social orders (Personal Narratives Group 1989; Rosenwald and Ochberg 1992; Zuss 1997), and as an educational tool (Cohen 1996; Witherell and Noddings 1991). This body of work makes at least two important points. First, the “self” is not an unchanging entity beyond the reach of everyday human action, but is something that can under some circumstances be changed with effort. Second, changing the self can happen through the social practice of narration, not just through the activity of an isolated individual.

Although this work on narrative self-construction promises both theoretical insight into the processes of self-construction and practical tools for changing the self, most of it has failed to provide a comprehensive account of how autobiographical narration can actually construct the self. A full account would require three components: a linguistically sophisticated account of how narrative discourse creates relevant patterns; an account of the mechanism through which these discursive patterns influence social and psychological processes; and a theory of what the self is, such that it can be partly constructed through some narrative mechanism. Most existing work on narrative self-construction includes only one or two of these components.

Part III of our volume consists of three contributions to the field of identity studies concerned with the more specific business of gender identity. Even more concretely, within the field of gender studies, all three chapters speak to what it means to be or become a young man. While other contributions in this volume similarly show how participants orient toward issues of gender (cf. Georgakopoulou's analysis of 17-year-old girls' identity formation in projective narratives, or Bell's study of a woman who is trying to make sense of herself as mother and as DES daughter), we decided to give these chapters their own categorical umbrella. They all show how male identities are formed discursively vis-à-vis particular hegemonic discourses of masculinity, but also vis-à-vis discourses of heterosexuality and whiteness – in positions of complicity but also in positions of resistance towards these discourses.

All chapters use narratives-in-interaction as their database to explore these identities as constructed and emergent, but the data collection techniques are different in each case. While Kiesling made use of the ethnographic interview as a story-elicitation technique with fraternity students, Moita-Lopes relied on a focus-group interview between three researchers and seven young adolescents. Wortham and Gadsden had young African-American researchers interview lower-class, urban African-American men who had become fathers as teenagers. Their semi-structured interviews resemble more the traditional biographic interview style used typically by sociologists.

Now that you have some basic understanding of how the variable rule program works, let us now turn to the step-by-step procedures involved in performing an analysis. I will begin with distributional analysis.

FUNDAMENTALS

All too often when students first set out to do a distributional analysis, they do it the wrong way round. In order to do it right, distinguish between the roles of the dependent variable and the independent (explanatory) factors. Recall that in every variation analysis the focus is the tendency for the dependent variable to occur in a series of cross-cutting independent factors: ‘The essence of the analysis is an assessment of how the choice process is influenced by the different factors whose specific combinations define these contexts’ (Sankoff 1988c: 985).

THE WRONG WAY TO DO DISTRIBUTIONAL ANALYSIS

Many students make the mistake of reporting how the variants are distributed across the explanatory factors. Consider the results file for variable (t,d) in (1).

Now that I have explained the history, development and nature of variable rules and the variable rule program, let us now work directly with the program. The prototype program, Varbrul 2 S, was written in Fortran by David Sankoff (1975). It was revised by researchers at the University of Pennsylvania, including Don Hindle and Susan Pintzuk. Refinements to the program were also conducted by Pascale Rousseau at the University of Quebec in Montreal. When Goldvarb 2.0 was released in 1988, it was based on these versions of the program, but was partially reprogrammed in Pascal by David Rand (1990). Other versions of the program also exist, including MacVarb (Guy 1993) and R-Varb (Paolillo 2002). Goldvarb 2001 provides users with a Windows version of the original Macintosh application. Goldvarb X has been available since October 2005. It is an update of Goldvarb 2.1 for Macintosh or Windows in which the entire program has been translated into C++ (Sankoff et al. 2005).

In this chapter, I will abstract away from the various individual attributes of these different packages. Details of their workings can be found in the users manuals, documentation and online help menus which come with various applications.

Chapters 1 to 3 have focused on methods for collecting optimal data for analysis. Now it is time to learn what to do with data once you have it. This chapter focuses on data handling and, in particular, techniques for representing speech data in writing.

When faced with a collection of dozens upon dozens of audio-tapes, minidisks or sound files, what do you do next? How can you make the invaluable data contained within maximally accessible and useful?

In this chapter, I focus on tried-and-true procedures from my own experience. I build on the foundations of earlier corpus-building projects (Poplack 1989, Poplack and Tagliamonte 1991). However, I also focus on data arising from fieldwork conducted in the British Isles between 1995 and 2001 (e.g. Tagliamonte 1998, Tagliamonte et al. 2005).

THE CORPUS

The components of a corpus, at least in my own research, are listed in (1):

Components of a corpus

1. (1)

1. a. recording media, audio-tapes (analogue, digital) or other

2. b. interview reports (hard copies) and signed consent forms

3. c. transcription files (ASCII, Word, txt)

4. d. a transcription protocol (hard copy and soft)

5. e. a database of information (FileMaker, Excel, etc.)

6. f. analysis files (Goldvarb files, token, cel, cnd and res)

The basic substance of a language corpus is the data. Most of my corpora have been collected on audio-tapes and represent one to two hours of conversation between a single interviewer and an informant.

In the last two chapters, I have focused on setting up a research project, entering the speech community, and fieldwork ethics. Now, I turn to the question of how to collect appropriate data.

THE ‘INTERVIEW’

The basic tool for recording conversation in sociolinguistic variation is referred to as the ‘sociolinguistic interview’. In fact, this is a misnomer; a sociolinguistic interview should be anything but an ‘interview’.

MODULES

Labov (1984: 32) defines the sociolinguistic interview as ‘a well-developed strategy’ that is defined by a number of goals. The most important of these is to record one to two hours of speech and a full range of demographic data for each speaker within one's sample design. In Labov's (1984: 33–4) early formulation of the sociolinguistic interview, it was defined as a series of hierarchically structured sets of questions, what he refers to as conversational modules or ‘resources’ (Labov 1973).