CHAPTER 3

 

METHODOLOGY

Ethnomethodologists and conversation analysis seek to discover the interpretive practices through which interactants produce, recognize, and interpret their own and others’ actions. (Pomerantz, 1988, p. 361)

 

3.1 Introduction

This study is informed by the theoretical principles of ethnomethodology and a conversation analysis (CA) approach. Both are outlined initially in this chapter, together with procedures for doing CA work. Since this study examines the social organisation of institutional activity, and explicates aspects of interaction as constitutive of the setting, the chapter outlines considerations relevant to undertaking CA work. The chapter concludes with an examination of a number of specific studies that have informed the data analysis in this study. These include studies of ordinary conversation (Sacks, Schegloff & Jefferson, 1974) and of classroom talk.

 

3.2 Theoretical position of ethnomethodology

Ethnomethodology is a theoretical position (Freiberg, & Freebody, 1995) and a sociological approach (Watson, 1992). It takes as its focus the study of “naturally occurring ordinary activities” (Cuff, Sharrock & Francis, 1990), and the common-sense knowledge, procedures and consideration by which “the ordinary members of society make sense of, find their way about in, and act on the circumstances in which they find themselves.” (Heritage, 1984, p. 40). Ethnomethodology is interested in the way in which social order and social organisation are constituted (Goodwin & Duranti, 1992, p. 27). It maintains an insistence on “a praxiological understanding of meaning and structure in situ” (Macbeth, 2001, p. 49).

 

Ethnomethodology was developed by Harold Garfinkel (1972; 1991) as an alternate sociology. According to Garfinkel the “ethnomethodologicalness” of the phenomena of interest in ethnomethodology involves:

1.      being available to policies of ethnomethodology but not to classic sociological approaches e.g. ethnomethodological indifference rather than a priori determinations

2.      in general found uninteresting by social science (and ignored)

3.      not reducible in the conventional social science sense without loss of the phenomenon itself

4.      “inspectably the case” (not available through generalizations, nor typifications nor metaphoric representation)

5.      discoverable only (not imagined)

6.      specifying foundational issues for ethnomethodology as a sociology concerned with “issues of produced order in and as practical action”

7.      produced locally, organised within, and reflexively accountable

8.      every topic of logic, order, meaning, or method is eligible for respecification as only locally and reflexively produced, naturally accountable phenomena of order.” (Garkinkel, 1991, pp. 16-17)

 

Garfinkel’s comments on ethnomethodology encapsulate his critique of mainstream sociology and the ways he proposes that ethnomethodology provides an alternative. For example, he poses that most sociological studies are driven a priori by theory and typifications that obscure the very phenomena that they seek to understand. He positions ethnomethodology as an alternate approach to understanding the central concern of sociology – the social order. Respecification is proposed as the way in which ethnomethodology undertakes to examine the social order.

 

3.2.1. Respecification

Central to the development of ethnomethodology was Harold Garfinkel’s intention to provide a methodological approach that would challenge conventional sociology. Although other “new moments” in sociology have sought to do this, frequently they buy into sociological ways of addressing core issues and in this way contribute to the maintenance of the very aspects they would seek to change (Button, 1991). Ethnomethodology has avoided this, since from its beginnings it was committed to the respecification of what counts as doing sociological research. Respecification is evident in all aspects of ethnomethodology, from its approach to topics of study, methods of research and through to the place of theory in research.  Ethnomethodology claims to be an alternate sociology precisely because it respecifies how to go about understanding fundamental concerns and issues of sociology.

 

A central issue for sociology, for example, concerns the relationship between social structures and the agency of the individual (Barnes, 2001). For some sociologists a fundamental belief is that the individual has little control over what are the wider and more powerful influences and affects of society. Social structures determine social reality for individuals. The influence and effect of the individual is constrained by wider forces. Garfinkel gave attention to what he believed to be the source of social order: ethnomethods or the ways in which people on an everyday basis make sense of their world through actions. He gave his focus to interactions in order to examine the ways in which mutual activities talk social realities into being (Freiberg & Freebody, 1995). In this way, Garkinkel positioned ethnomethodology as an alternative to functionalism.

 

Parson’s Structure of Social Action (Parsons, 1937; cited in Hilbert, 1992) laid the foundations for functional sociology which came to dominate sociology in America until the late 1950s (Hilbert, 1992). According to functionalism, the factual, or behavioural, order is produced by members of society due to their socialization into the normative order (Hilbert, 1992, pp. 2-3). Factual order is the way that members behave, regularly and in orderly ways. In Parson’s view normative order is moral and prescriptive, and involves a system of norms, values, roles and statuses internalised through the socialization process. Values are reproduced through actions and the orderly nature of society carried on or maintained. Such a perspective can be said to be one in which people are ironicized (Garfinkel, 1967). That is, sociologists must work out how social processes work because people themselves do not really know, even though they might think that they do. The problem of order, according to Parson’s, is the “problematic relationship between factual order and normative order” (Hilbert, 1992, p. 19). Functional theorizing involved “figuring out what kinds of norms, processes and so on are necessary for socialization to produce order” (Hilbert, 1992, p. 3).

 

Garfinkel, a student of Parsons, questioned the view that social order came about through norms and rules of institutionalized systems (Holstein & Gubrium, 1994) and the shared ideas of people concerning what is morally right. Garfinkel took a different tack to Parsons. This eliminated a priori theorising about norms and used investigation instead. Garfinkel’s approach to social order takes a local approach which focuses on the everyday, the ordinary and mundane. Conceptually Garfinkel positioned ethnomethodology in opposition to the focus for social inquiry on a macro level. Garfinkel looked to the everyday, arguing that the site for inquiry was the local and its focus should be members’ practices. He claimed the local order had its own accountability evidenced in its self-organisation and management.

 

Garkinkel was critical not only of structural approaches to sociology but also of survey research and quantitative analyses that examined aspects of it. Although earlier sociological work had drawn on anthropologically inspired ethnographies, sociological inquiry in the 1950s and 60s involved a central concern with variable analysis. Conceived by Lazarsfeld (Benson & Hughes, 1991), it involved data collection methods such as questionnaires and interviews, and statistical analyses of information from these. An ethnomethodological critique of these included that individual cases were irretrievable, that coding obscured individual details, and that categories used were those of researchers rather than those of the researched.

 

Most importantly, the use of variable analysis reflected sociology’s concern with “providing a social basis of dealing with the problem of evidence and inference …” (Benson & Hughes, 1991, p. 110). Garfinkel argued that ultimately researchers engaged in statistical analysis failed to perceive or acknowledge the interpretive processes that it involved. In seeking to be objective, researchers overlooked the role played by their own common knowledge and interpretive processes. In a study of clinicians’ work in an outpatient psychiatric clinic Garfinkel determined a range of ad-hocing processes central to categorising patients. Garfinkel showed the essential nature of ad-hocing to categorising as an activity and illustrated the way in which social scientists treated social categories as static (Brandt, 1992, p. 315). Garfinkel has maintained a program of research that has highlighted the interpretive processes involved in social and scientific research.

 

Respecification of the focus for sociological analyses has resulted in understandings that inform numerous aspects of sociology. For example, Schegloff’s (1992) study of repair in ordinary conversation enabled consideration of intersubjectivity as the practical accomplishment of what is taken to be going on during social activity.  That is, it informed how people accomplish their shared understandings of what is going on through the sequential updating of their accounts in their turns at talk. Respecification has been applied to approach old problems with “fresh” eyes in areas such as education.  For example, Freebody and Freiberg (2001) considered how reading education has been dogged by psychological accounts of reading instruction. They determined that descriptions of reading pedagogy draw on accounts that involve taken-for-granted idealisations that inform the basis for pedagogical debates concerning reading instruction. They respecified reading instruction as practical accomplishment in everyday reading lessons in classrooms, and detailed the ways in which current approaches to reading instruction bear little resemblance to current reading practices.

 

3.2.2 A phenomenological sensibility

Garfinkel perceived that the emphasis on administrative structures or demographic and family background variables treated people as cultural or “judgemental dopes” (Sharrock, 2001). Ethnomethodologists instead have “sought to exhibit the phenomenal complexity and material embeddedness of conventional understandings.” (Lynch & Peyrot, 1992, p. 118). The approach developed in ethnomethodology expresses a “phenomenological sensibility” (Heap & Roth, 1973, p. 363-65; cited Maynard & Clayman, 1991, p. 388).

 

Phenomenology then undertakes to see how members experience “the world and all its manifestations as real through identifiable acts of consciousness.” (Maynard & Clayman, 1991. p. 389). As a philosophical school of thought, phenomenology can be traced back to Hegel’s Phenomenology of Mind. However its current expression dates from the work of Husserl. Husserl maintained that objects and scientific achievement rest on “a vast prescientific foundation of unclarified presuppositions” (Maynard & Clayman, 1991. p. 389). To discern and elucidate these is to reach the ultimate grounds of knowledge (Maynard & Clayman, 1991, p. 389). Husserl took a descriptive orientation to his approach.

 

Phenomenology’s influence on Garfinkel was most obvious through the work of Husserl’s student, Alfred Schutz. The phenomenological notion of Lebenswelt, “the whole universe of life” (Schutz, 1970, p. 259) or “the mundane world of lived experience” became the starting point for the phenomenological sociology of Schutz (Heritage, 1984, p. 44). For Schutz, the social world in the first instance is experientially interpreted as “meaningful and intelligible” through the social categories and constructs evidenced in language. Researchers also interpret using their own knowledge of these things. This is referred to as second order constructs or “constructs of the second degree” (Schutz, 1970, p. 273).  Schutz maintained this was so even in studies of natural science (Heritage, 1984).

 

3.2.3 Intersubjectivity

The notion of intersubjective understanding is central to Schutz’s work. He proposes the “reciprocity of perspectives” (1970, p. 183) to explain the way in which members interact and, therefore, overcome differences inherent in their own uniqueness. It involves two basic idealizations: that of the interchangeability of standpoints and the idealization of the congruency of the system of relevances. The former involves the taken-for granted assumption that if the ego and alter ego were to change places things would be experienced and perceived in the same way. The congruency of the system of relevances takes it that individuals assume that differences inherent in the uniqueness of their physical situations are irrelevant in their interactions and, that they have “selected and interpreted the actually of potentially common objects and their features in an identical manner” (Schutz, 1970, p. 183). Private experience according to Schutz is thus transcended by a common world (Schutz, 1962a, cited Heritage, 1984, p. 55).

 

For Schutz the world of shared experience is created or brought into being through language. He perceives that individuals’ knowledge is only gained in part through personal experience, with the rest being socially derived. Schutz focused on language to explain this. He asserted that “the typifying medium par excellence by which socially derived knowledge is transmitted is the vocabulary and the syntax of everyday language.” (Schutz, 1970, p. 96). Inherent in the naming of things is typification and generalization. Garfinkel’s consequent focus on language, in the development of ethnomethodology, was ground breaking for sociology and a source of controversy. Central to his focus were the notions of reflexivity and indexicality.

 

3.2.4 Reflexivity and indexicality

The ethnomethodological “version of reflexivity” (Lynch, 2000, p. 26), was developed by Garfinkel and predates many current usages and definitions (Lynch, 2000, Macbeth, 2001). It takes reflexivity to be integral to all social action, including that of analysts. To Garfinkel and ethnomethodologists, reflexivity refers to “accounting practices and accounts” (Lynch, 1991) or to the “embodied practices through which persons singly and together, retrospectively and prospectively produced account-able states of affairs” (Lynch, 1991, p. 73). In other words, members display their understandings of the actions of others through actions. These in turn are available as an account for others. According to Garfinkel (1967, p. 4), “Members’ accounts are reflexively and essentially tied for their rational features to the socially organised occasions of their use for they are features of the socially organised occasions of their use.”. The “essential reflexivity of accounts” (MacBeth, 2001) is a taken-for-granted aspect of social interaction and in that sense unremarkable. This is shown in the way that members treat each others’ accounts as “the most passing matter of fact” (Garfinkel, 1967). 

 

Language is indexical. That is, its intelligibility is tied to the circumstances for its use (Maynard & Clayman, 1991, p. 392). As Garfinkel demonstrated (1967) it is impossible to produce a linguistic utterance that does not make use of ‘et cetera’. Rather than treating this as a problem, ethnomethodology takes indexicality to be a resource for members in the organisation of local action, and makes it a topic for investigation by analysts. In this sense then, Garfinkel, (1967, p. 11) describes “ethnomethodology” as the term used to refer to “the investigation of the rational properties of indexical expressions and other practical actions as contingent ongoing accomplishments of organised artful practices of everyday life.”

 

3.3 Conversation analysis

This section outlines a CA approach. It considers its development, including the influence of ethnomethodology and the relationship between the two approaches. Other influences on CA are also outlined. The distinctiveness of CA is established and central facets of CA work are introduced.

 

3.3.1 Defining the approach

CA takes the analysis of talk as “situated social practices” (Goodwin & Duranti, 1992) as its focus. It describes “the procedures and expectations in terms of which speakers produce their own behaviour and interpret the behaviour of others.” (Heritage, 1984, p. 241). Central to this is the analysis and understanding of conversation as being sequentially organised as turns at talk (Heap, 1997). CA aims to reveal or make explicit the “lay analysis” (Watson, 1992, p. 263) or “tacit reasoning procedures” (Hutchby & Wooffitt, 1998) used by members in the production of everyday talk-in-interaction. Its focus of study is the “interactional organization of social activities.” (Hutchby & Wooffitt, 1998, p. 14).

 

The approach was developed initially by Harvey Sacks, and later in collaboration with colleagues Schegloff and Jefferson (Psathas, 1995), during the mid 1960s. It involved the detailed examination of recorded conversations so as to describe the structural organisation of everyday language use and the social order that it revealed (Hutchby & Wooffitt, 1998). Specifically Sacks posited a turn-taking system, also called a speech exchange system. His work involved the analysis of members’ management of the system (Watson, 1992, p. 263), using recordings of naturally occurring talk. Sacks rejected the use of recollections of actual talk, and the use of artificial talk constructed for the purposes of analysis. The use of tape recordings by Sacks was innovative. It was part of his attempt to deal with “specific, singular events of human conduct” (Heritage, 1984, p. 235) rather than type constructs. In this way he strove to develop an approach firmly grounded in the lived social world, and in the process “established nothing less than an alternative sociology of an analytically revolutionary kind.” (Watson, 1992, p. 263).

 

Sacks’s development of CA evidences wide-ranging influences. Theoretically it was informed by ethnomethodology. It drew on Goffman’s work on the interaction order. Other non-sociological influences on Sacks included: ordinary language philosophy, Chomskian generative grammar, Freudian psychology and anthropological fieldwork (Sacks, 1995). The influence of ordinary language philosophy (Austin, 1970; Wittgenstein, 1958) is evidenced particularly in the way in which CA sees the “identity of action as dependent on community conventions and practices applied in ways sensitive to local settings of interaction.” (Heap, 1997, p. 221). Sacks’ studies themselves are “demonstration of what the idea of ‘language game’ might stand for” (Lee, 1991, p. 223).

 

3.3.2 Conversation analysis and ethnomethodology

CA emerged from its ethnomethodological roots to become a “prominent form of ethnomethodological work” (Heritage, 1984, p. 233) and a “distinctive field of research” (Heritage, 1984, p. 234). Despite its distinctive work, CA retains its relationship to ethnomethodology through shared theoretical perspectives and a common interest in social order. Ethnomethodology and CA are therefore, a “family of related and overlapping empirical approaches" (Watson, 1991, p. 268) or "cognate sociological approaches" (Watson, 1992). Some would suggest that CA is “perhaps the most visible and influential form of ethnomethodological research” (Maynard & Clayman, 1991, p. 396).

 

The concern with order and the local accomplishment of it is arguably the most salient evidence of the ethnomethodological roots of CA. According to Psathas (1995), CA’s basic assumptions are that order is produced, situated, oriented to by parties rather than conceived by analysts, repeatable and recurrent. The task of the analyst is to discover, describe and analyse the structures of social action (practices and procedures and so on), to put aside issues of frequency and applicability and to produce analysis that is “formal, that is, structural, organizational, logical, atopically, contentless, consistent, and abstract, terms.” (Psathas, 1995, p. 3). Pomerantz determines minimally three central things that ethnomethodologists and CA researchers agree on: the focus on “how participants themselves produce and interpret each other’s actions.”; desire to treat “ordinary events as worthy of serious analytic attention” and preference for analysing naturally occurring interactions (Pomerantz, 1988, pp. 360-361)

 

According to Taylor and Cameron (1987) the central principles of CA are accountability, normativity and intersubjectivity. These show indisputably the ethnomethodological underpinnings of the approach. In fact, they argue, it is “Garfinkelian principles which give life to the distinctive methodology characteristic of CA.” (Taylor & Cameron, 1987, p. 99). Due to this, it is not possible for conversation analysts to divorce themselves from ethnomethodology. Although the relationship between the two is contentious (Sharrock, 2001) reference in this research to CA or to a conversation analytic approach encompasses the understanding that the approach is ethnomethodologically informed.

 

Although CA owes much to ethnomethdological understandings and shares many of its theoretical perspectives, it is obvious that the work of Sacks brought substance to Garfinkel’s aims for ethnomethodology.

The discovery of structure in interaction sequences proved to be an important finding because it confirmed what had been proposed in ethnomethodology from the outset, namely, that there is an order to be found in the most mundane of interactions, and that close examination of actual occurrences would enable the analyst to discover, describe, and analyze that orderliness. (Psathas, 1995, p. 17)

 

A number of central ethnomethodological concerns are brought to fruition through CA. These include its rigorous attention to detail through the analysis of the transcript data, and its approach to collection studies where all similar cases of a social activity are accounted for in the analysis (ten Have, 1990). CA has been able to provide the attention to detail and order that ethnomethodologists aspired to. Some have even argued that CA is the fruition of ethnomethodology’s ideals and is now its central arm.

 

A clear example of the way in which CA has been able to inform ethnomethodology is provided through Heritage’s account of the “architecture of intersubjectivity” provided by sequential organisation (1984a, p. 254ff). That is intersubjectivity can be said to be sequentially constructed so that members display their understanding and correct/confirm those of their interactants (Taylor & Cameron, 1987, p. 104). In this way conversation, as seen through turn-taking, ensures that:

 a context of publicly displayed and continuously updated intersubjective understanding is systematically sustained. It is through this “turn by turn” character of talk that participants display their understandings of the state of “the talk of one another” … Mutual understanding is thus displayed, to use Garfinkel’s term, “incarnately” in the sequentially organized details of conversational interaction. Moreover, because these understandings are publically produced, they are available as a resource for social scientific analysis. (Zimmerman & Boden, 1991, p. 259).

 

So the “problem of intersubjectivity”, or the “how” of members’ shared understandings, is shown to be an “ordinary practical accomplishment” (Taylor & Cameron, 1987, p. 104) in which members’ display their understanding and correct and/or confirm those of their interactants. (Taylor & Cameron, 1987, p. 105). According to Schegloff, “particular aspects of particular bits of conduct that compose the warp and weft of ordinary social life provide occasions and resources for understanding, which can also issue in problematic understandings.” (1992, p. 1299). Schegloff outlines the way in which turn taking works to maintain members’ shared understandings of ‘what is going on’ in a study of misunderstandings and repair.

 

3.3.3 Goffman and the interaction order

The work of Goffman was another influence on Sacks. In the 1950s Goffman focused on the “ritual nature of face-to-face interaction” (Hutchby & Wooffitt, 1998, p. 27). He viewed our social selves as performed and thus affecting how others orient towards us. Although in his later work, Goffman showed an increasing focus on language, his interest in the ritual aspect meant that he attended to ‘ceremonial’ aspects of interaction. The use of ceremonial as distinct from basic intelligibility meant that for Goffman there was a distinction to be made between the focus of the interaction order and that of CA on the orderly nature of talk-in-interaction. Sacks would not have maintained this distinction.

 

Erving Goffman contributed significantly to the analysis of face-to-face interaction for example, through developing analytic frameworks for describing and analysing multi-party interaction and through investigating framing invoked by a single speaker in a strip of talk (Goodwin & Duranti, 1992, pp. 24-25). Goffman’s work showed the value of the focus on the everyday interactions of people, including the social significance of it (Psathas, 1995). However, methodologically, Goffman’s data collection and analyses lacked the systematic methods employed by CA. Goffman eschewed recordings. Instead he used observation field notes and even constructed examples. Data was often used to illustrate Goffman’s concepts rather than to develop them. This method of using data contrasts with the approach taken by Sacks, who sought to develop theory that was data driven. Goffman also used concepts which “glossed” many of the situational particulars which had been his original focus. Finally and ultimately, some would argue his interests were with the individual, and the psychological, rather than with considerations of the structuring of interaction (Schegloff, 1988).

 

3.4 Conversation analysis and institutional talk

This section introduces and discusses methodological considerations for analyzing and explicating the social organisation of institutional activity.  Many conversation analysts have examined talk-in-interaction within institutions, creating what has become a particular and dominant “arm” of the approach (Hester & Francis, 2001). The discussion draws out areas of contention relevant to this study.

 

One particular arm of conversation analytic work is the study of institutionalized talk in order to describe the machinery of the talk and the “recurrent and relatively specialized sets of situated identities” Zimmerman & Boden, 1991, p. 13). For some researchers this will involve comparisons with ordinary mundane conversation as a method of explicating the features of talk that are “specialized and respecified in various ways” (Heritage & Greatbatch, 1991) within the individual institutionalized setting.

 

A key argument in the CA approach to institutional talk is that rather than people talking as they do because they are in a particular institutional context, “people’s ways of managing their talk itself constitutes the ‘institutionality’ of such settings” (Hutchby, 2001, p. 76). Studies of institutionalized talk have included, for example, talk in classrooms (McHoul, 1978), courtrooms (Atkinson & Drew, 1979), news interviews (Heritage, 1985), doctor-patient interaction (Heath, 1986), talk radio (Hutchby, 1996) and emergency services (Whalen & Zimmerman, 1989; Zimmerman, 1984).

 

Garfinkel’s early ethnomethodological work examined such things as juries. His studies maintained a focus on ‘the methods used by societal members to render their circumstances orderly and intelligible” (Maynard & Clayman, 1991), and he employed ethnographic methods of examination. The early studies of Sacks and Schegloff sought to “explicate the endogenous organization of the talk-in-interaction” rather than to detail the relationship between the talk and its setting (ten Have, 2001, p. 3). Much later work by conversation analysts shifted the focus from ethnographic data to the sequential analysis of talk. Their work can be considered along several lines or forms of work (Maynard & Clayman, 1991, p. 406). One is explicating the turn-taking that constitutes particular activities, such as job interviews (Button, 1987). Another is describing the turn-taking systems that organise aspects of institutional settings (Mehan ,1979) and similarly, how modifications of the ordinary conversation speech exchange system serve to accomplish certain functions such as neautrality in new interviews (Clayman,1988). Another approach has been to examine how sequence types exhibit reductions of conversations’ “options and opportunities for action”, and “specializations and respecifications” of the remaining conversational functions (Heritage & Greatbatch, 1991, p. 95) that carry over from ordinary conversation. For example, Heritage (1984) examined third turn responses in question-answer sequences.

 

Heritage and Greatbatch (1991) draw out salient methodological aspects central to many conversation analytic studies that examine institutional talk. Integral is an understanding of aspects of it as different from ordinary conversation, though both involve “an ordinariness”. Sacks, Schegloff and Jefferson (1974) established the practice of taking ordinary conversation as “primary and as collectively constituting a fundamental matrix through which social interaction is organized.” (Heritage & Greatbatch, 1991, p. 94). This view of ordinary conversation invites comparative studies (ten Have, 2001). Comparing the talk of ordinary conversation and that of a particular institution is a methodological device used by some conversational analysts to explicate features of institutionalized talk. The employment of comparison as an analytic method adds to understandings of talk and the institutionalized “whatness” of it. Comparative studies make it possible to identify both distinctive and generic qualities of interaction in institutions (Maynard & Clayman, 1991).

 

Drew and Sorjonen (1997) consider that all analytic approaches that examine institutionalized talk or dialogue involve some comparative aspect. They cite sociolinguistics, ethnography of communication, microethnography of face-to-face interaction, discourse analysis, and CA as examples. Drew and Sorjonen claim that a necessary condition of establishing the distinctiveness of institutionalized talk is comparison with ordinary conversation. Further, features characteristic of certain settings “are associated with distinctive patterns of sequences in institutional interactions, as compared with, for example, ordinary conversation.” (Drew & Sorjonen, 1997, p. 106). Differences often involve the third position response in sequences involved in questioning and answering (Drew & Sorjonen, 1997). An example would be the evaluation act, in third position, in the classroom teacher-led IRE sequence. As well, some expected response types may be missing. For example, “oh” is not used for news receipt in news interviews (Heritage, 1985). Other features commonly occur in institutionalized talk and less frequently in ordinary conversation, for example ‘and’ (Heritage & Sorjonen, 1994).

 

According to Drew and Heritage (1992, p. 22) there are three main differences between institutional and ordinary talk.  At least one participant is oriented to the core “goal, task or activity” of the institution and talk is informed by “goal orientations of a relatively restricted conventional form”; “special and particular” constraints are involved in contributions; and talk is associated with “inferential frameworks and procedures” (cited in Cameron, 2001, p. 101). Cameron takes talk in magistrate’s court (Harris, 1984, cited in Cameron, 2001, p. 101) to illustrate these. It is: directed towards a specific goal reflected in the many questions and answers about defendant’s resources; involves in particular a constraint on defendant’s rights to ask questions; and exhibits inferences regularly made by defendant’s concerning magistrates’ intentions (Cameron, pp. 101-103).

 

In a consideration of ‘talk and social structure’, Schegloff (1992) draws out two issues: the problem of relevance and the issue of procedural consequentiality. The former involves the need to show members’ orientations to aspects of context. Consideration of the problem of relevance provides a way to approach analysis so as to enrich through further detail. It may also be used to establish how social structure is “confirmed, reproduced, modulated, neutralized, or incrementally transformed in that actual conduct to which it must finally be referred.” (p. 110).

 

The second problem or issue is that of procedural consequentiality. Schegloff asserts that an examination of talk in a setting or context must show how its occurrence in that place is procedurally consequential. Schegloff asks:

How does the fact that the talk is being conducted in some setting (e.g. “the hospital”) issue in any consequences for the shape, form, trajectory, content, or character of the interaction that the parties conduct? And what is the mechanism by which the context-so-understood has determinate consequences for the talk? (1992, p. 53).

 

The challenge, then, as stated by Schegloff, is for the analyst to show a direct connection between the setting and the talk that constitutes it. One way to do this is though detailing the speech exchange system (Schegloff, 1992, p. 112).

 

Hester and Francis (2001) take issue with the institutional talk ‘arm’ of CA as it is conceived by Drew & Heritage (1992), and Schegloff (1992). Specifically, they disagree with what they describe as presuppositions of this work; most notably that it is the properties of talk-in-interaction which are “foundational for the recognizable production of given institutionalized settings.” (Hester & Francis, 2001, p. 207). The focus in particular on the sequential involves two claims. The first is that sequential forms are what compromise “institutionally specific and distinctive adaptions” of the speech-exchange system of ordinary talk (Hester & Francis, 2001, p. 210). The second claim is that the distinctive forms of turn-taking constitute what is the recognizable institutional character of a setting.

 

Hester and Francis argue that these claims involve a linear fallacy. They propose that it is not the sequential organisation of talk but, rather, particular membership categories and their related rights and responsibilities that comprise the accomplishment of everyday activities. Further, Hester and Francis argue that the claim that members orient to their institutional identities (e.g. Drew & Sorjonen, 1997, p. 111) ignores their situated work and the situated “recognizability” (Hester & Francis, 2001) involved in knowing an institutional setting as just that setting.  In this way, Hester and Francis explicate what they regard as a fatal flaw in conversation analytic work. If CA studies adopt the view that it is the sequential nature of talk that constitutes the distinctiveness of institutional talk, they compromise the ethnomethodological underpinnings of their work (Hester & Francis, 2001, p. 271).

 

In a discussion of the issues related to ‘applied” CA, ten Have suggests that it is a matter of where the ‘centre of gravity’ is located in the research; either in “the institutionalized power of one category of participants over another” or in the “local interaction and its procedural infrastructure itself, in the general institutional arrangements” (ten Have, 2001, p. 5). ten Have indicates, that from an ethnomethodological position, the focus must be on the in situ work of members rather than the mechanics of their turn-taking.

 

Sharrock argues that to attempt to make CA responsive to the institutional setting (Sharrock, 1989, p. 657) is methodologically problematic and represents a misunderstanding of turn-taking studies themselves. Sharrock’s cites his own research on ‘police work’ to illustrate his claim that the explication should be about how talk accomplishes the work rather than about the turn-taking itself. In a discussion of interactional asymmetries in institutional settings, Drew and Heritage (1992, p. 53) claim that researchers who examine institutional talk must do one of two things. They must show the ways in which participants talk is oriented to “role-related asymmetries” and the consequences of this for its outcomes. In other words they must consider Schegloff’s problems of relevance and procedural consequentiality. Or, they must demonstrate by comparative analysis that the features of talk that “embody” asymmetrical relationships do not usually exist in ordinary conversation.

 

3.5 Conversation analysis and classroom interaction

This section outlines key features (Austin, Dwyer & Freebody, 2003) of CA. It incorporates CA studies of classroom interaction, selected according to their relevance to the analyses and findings of the present study. Discussion also draws on some studies of ordinary conversation, since that work is central to CA. Much of the influential CA work has followed on from a study by Sacks, Schegloff and Jefferson (1974) in which they articulated the rules of turn-taking, and introduced features of ordinary conversation such as trouble, repair, and recipient design. These have been explored in numerous CA studies since and are essential to undertaking CA work.

 

3.5.1 Turn- taking

In a groundbreaking study, Sacks, Schegloff and Jefferson (1974) established the rules of turn-taking that locally manage ordinary or mundane conversation. The account involved the positing of a continuum that encompassed free-flowing conversation on the one hand and pre-determined talk on the other. The model consists of a turn construction component and a turn distribution component (Hutchby & Wooffitt, 1998). The former involves the property of projectability, whereby during its construction members make evident what kind of unit it is, when it is probably going to end and the position of the transition-relevance places. The turn distribution component involves the rules to which members orient. It therefore has a normative aspect (Hutchby & Wooffitt, 1998). The rules are:

(1) For any turn, at the initial transition-relevance place of an initial turn-constructional unit:

(a) If the turn-so-far is so constructed as to involve the use of a ‘current speaker selects next’ technique, then the party so selected has the right and is obliged to take next turn to speak: no others have such rights or obligations, and transfer occurs at that place.

(b) If the turn-so-far is so constructed as not to involve the use of a ‘current speaker selects next’ technique, then self-selection for next speakership may, but need not, be instituted; first starter acquires rights to a turn, and transfer occurs at that place.

(c) If the turn-so-far is so constructed as not to involve the use of a ‘current speaker selects next’ technique, the current speaker may, but need not continue, unless another self-selects.

 

2. If, at the initial transition-relevance place of an initial turn-constructional unit, neither 1a nor 1 b has operated, and, following the provision of 1c, current speaker has continued, then the rule-set a-c re-applies at the next transition-relevance place, and recursively at each next transition-relevance place, until transfer is affected (Sacks et al., 1974, p. 704)

 

In summary, as Cameron (2001, p. 90) neatly put it, the rules for turn taking are:  “1 Current speaker selects next speaker or if this mechanism does not operate, then … 2 Next speaker self-selects or if this mechanism does not operate, then … 3 Current speaker may (but does not have to) continue”. These rules take account of related aspects of turn-taking which continue to be, or have since become, the focus for attention by analysts. These include recipient design; adjacency pairs, repair and so. The rules also provide “proof procedure”, or the “display of understandings in the talk of subsequent turns” to which analysts can turn (Sacks et al., 1974, p. 729) for confirmation of their own interpretive work. As well, the turn-taking system provides for comparative analysis of turn-taking systems within speech exchange systems.

 

Sacks et al. (1974) discussed self-selection techniques in their explication of turn-taking. They considered the need to “begin with a beginning” (1974, p. 719) since beginnings with self-starters may be subject to overlap and potential impairment of analyzability. They suggested that appositional beginnings (e.g. ‘and’, ‘so’, ‘well’) satisfy the need to have a beginning but do so without “revealing much about the constructional features of the sentence thus begun. That is, it is possible for the speaker to start without having a plan in hand as a condition for starting.” (Sacks et al., 1974, p. 719). This allows the speaker to “start fast”. As Sacks et al. put it, in their outline of appositional beginnings and self-starting, “their overlap will not impair the constructional development or the analysability of the sentence they begin.” (1974, p. 719). They referred to the use of appositional beginnings, used in this way, as “turn-entry devices’ or “pre-starts”.

 

The rules of turn-taking in ordinary conversation provide a base for describing turn-taking features in other systems and a methodological approach to the analysis of institutionalized talk (Psathas, 1995). McHoul (1978) proposed modifications to the speech-exchange system of ordinary conversation in his study of the “feelings of formality” (p. 186) in teacher-led whole class instruction. The speech-exchange system for whole-class instruction involved the following rules:

(I)For any teacher’s turn-so-far, at the initial transition-relevance place of an initial turn-constructional unit:

If the teacher’s turn-so-far is so constructed as to involve the use of a ‘current speaker selects next’ technique, then the right and obligation to speak is given to a single student; no others have such a right or obligation and transfer occurs at the transition-relevance place.

If the teacher’s turn-so-far is so constructed as not to involve the use of a ‘current speaker selects next’ technique, then current speaker (the teacher) must continue.

(II) If 1(A) is effected, for any student-so-selected’s turn-so-far, at the initial transition-relevance place of an initial turn-constructional unit:

If the student-so-selected’s turn-so-far is so constructed as to involve the use of a ‘current speaker selects next’ technique, then the right and obligation to speak is given to the teacher; no others have such a right or obligation and transfer occurs at the transition-relevance place.

If the student-so-selected’s turn-so-far is so constructed as not to involve the use of a ‘current speaker selects next’ technique, then self-selection for next speaker may, but need not, be instituted with the teacher as first starter and transfer occurs at that transition-relevance place.

If the student-so-selected’s turn-so-far is so constructed as not to involve the use of a ‘current speaker selects next’ technique, then current speaker (the student), may, but need not, continue unless the teacher self-selects.

(III) For any teacher’s turn, if, at the initial transition-relevance place of an initial turn-constructional unit either I(A) has not operated or I(B) has operated and the teacher has continued, the rule-set I(A)-I(B) re-applies at the next transition relevance place and recursively at each transition-relevance place until transfer to a student is affected.

(IV) For any student’s turn, if, at the initial transition-relevance place of an initial turn-constructional unit neither II(A) nor II(B) has operated, and, following the provision of II(C), current speaker (the student) has continued, then the rule-set II(A)-II(C) re-applies at the next transition-relevance place and recursively at each transition-relevance place until transfer to the teacher is effected. (McHoul, 1978, p. 188)

 

McHoul (1978) argued that the “rule-modifications” for the management of turns in the classroom form the basis for the distribution of differential participation rights. McHoul found that teachers took long pauses during their turns without others taking a turn. This implied relaxed responsibilities as “hearers” in the classroom; “they only need attend to others’ utterances which they (teachers) have called for.” (p. 192). McHoul illustrated the rules in terms of features of classroom talk-in-interaction that differed from ordinary conversation. These included maximisation of gap and pause, minimization of potential for overlap, and minimization of permutability of turn-taking (McHoul, 1978, p. 189). The latter refers to the potential for open-endedness which is strongly evidenced in conversation. McHoul claimed that “in classrooms, no other parties than teachers have the right to self-select as first-starters.” (p. 192). It needs to be noted that McHoul was referring implicitly to all teacher-led instruction in classrooms, rather than to interaction during time of individual activity.

 

A rule-of-thumb for when classroom talk breaks down, and the teacher needs attention to return to order, is to “Get back to normal procedure as soon as possible by doing something normal (rule guided) rather than by invoking the rule(s) explicitly. The invocation of the rule(s) can be done as a ‘reminder’.” (McHoul, 1978, p. 199). One way that teachers’ maintain attention is to have students answer questions on the lesson’s topic(s). This distributes “the burden of ‘discovering’ knowledge” and requires that parties need to listen in order to answer questions that might potentially come their way (McHoul, 1978, pp. 200-201).

 

Heap (1992) made a distinction between “conversational machinery” and “discourse-action machinery” in order to explicate goal-oriented talk where classroom members are “working interactively toward the visible completion of a task” (Freiberg & Freebody, 1995, p. 198). Ordinary conversation is task independent, primarily involves speech acts, makes relevant turns at talk and can be expanded through talk. Discourse machinery (Heap, 1992), on the other hand, is task oriented. Its speech acts are used to accomplish non-verbal acts and may be expanded by them. Turns-at-talk may be oriented to prior non-verbal turns and to successive non-verbal acts through directives followed by compliance. One consequence, then, is that directives may be followed by non-verbal actions that replace turns at talk. The distinction between conversation machinery and discourse-action machinery is useful, although discourse-action machinery may occur in other settings. For example interactions involving parents and children completing homework may involve directive and compliance pairs involving non-verbal acts (Freebody, et al., 1995).

 

3.5.2 Trouble and repair

Turn taking involves a moment-by-moment show of understanding however misunderstandings can occur. Repair occurs in talk when they do (Schegloff, Jefferson and Sacks, 1977). So, repair is “addressed to “troubles’ that emerge in interaction” (Hutchby and Wooffitt, 1998, p. 215). These vary from trouble at a word and pronunciation “level”, through to trouble related to problems of understanding (Schegloff, 1992, p. 1341) or even hearing (Bilmes, 1992). Trouble can involve person reference or speaker selection (Hutchby and Wooffitt, 1998).

 

Repairs occur close to their trouble source. This eliminates organisational difficulties that would result if repair involved substantial backtracking to previous talk. Repair works to eliminate escalation of trouble (Hutchby and Wooffitt, 1998). In conversation the preference is for self-correction (Schegloff, Jefferson and Sacks, 1973) and a speaker may address potential trouble in a turn, or just after. A hearer may initiate repair in the next turn after the turn in which the source of trouble occurs. Some trouble may extend beyond this turn if the hearer does not recognize the previous turn as problematic in any way. In that case, the original speaker treats the hearer’s response as problematic and addresses the trouble source “by engaging in some operation on the source of trouble,” (Schegloff, 1992, p. 1303).

 

This type of repair, termed third position repair, involves four main types of components and is highly recurrent in form (Schegloff, 1992, p. 1304). One component, A, consists usually of a “no” utterance, multiple “nos” or “oh no”. B takes the form of agreement/acceptance with a prior turn that has treated the trouble source turn as a complaint but then goes on to deny that the trouble source turn involved complaining. The C component rejects the understanding inherent in the prior turn, and D does something to an aspect of the prior turn concerning the trouble source turn in order to address the trouble. The D component is the one most likely to be present in the repair turn. As the particular use of “turn” and “position” suggests here, although the third position repair usually occurs in the third turn it need not necessarily be the case.

 

Fourth position repair is rare. It draws on two conversational components: one is “oh” and a recharacterization of the trouble source, the other is “oh” plus a revised response to the turn involving the trouble (Schegloff, 1992, p. 1323). It is carried out by the recipient of the trouble source turn rather than the speaker of it.

 

There are also alternatives to repair. A ‘next turn’ display of misunderstanding can be heard as a joke for example. Or, a misunderstanding can be ignored and then redone at a later stage as if it were occurring for the first time so “adopts nonrepair ways of dealing with it.” (Schegloff, 1992, p. 1331). The absence of repair has been identified in studies of classroom interaction. Ridley, Radford and Mahon (2002) examined the management of topic and repair in the interactions between a ten-year old child with her mainstream teacher, specialist teacher and a peer. The researchers found that the mainstream teacher frequently made no attempts at repair or failed to follow through with clarifications when repair was initiated (Ridley, Radford and Mahon, 2002). The researchers suggested that failure to repair risked confusion and ultimately represented lost opportunities for support that would lead to learning.

 

Freebody (2003) examined trouble and repair in a classroom reading lesson. He established trouble in a sequence of talk that involved a teacher questioning students during whole class activity. Although students were unable to provide the answer that the teacher’s questions required, the teacher did not repair the talk. Instead she provided clues which required that students attended to those to work out the answer. In this way, she “showed the students that they need to attend to the often subtle cues that marked the sustained relevance at work in teachers’ talk.” (Freebody, 2003, p. 116).

 

3.5.3 Parties-at-talk

The organisation of talk is “sensitive to the number of participants because those participants can and do design their conduct and understand one another’s conduct as shaped in part by reference to numbers of participants.” (Schegloff, 1995, p. 31). Talk tends to “schism” with greater number of participants. In CA studies, a distinction is made between participants and parties in talk. Importantly, the seminal description of the turn taking system (Sacks, Schegloff & Jefferson, 1974) concerned parties rather than numbers of participants.

 

Multi-party conversation is a “distinct phenomenon”, not just “a variant off two-party conversation.” (Sacks, 1995, p. 523). Therefore rules that members orient to, in settings involving more than two parties, are of interest. Sacks considered non-directed questions. He made the point that in two-party conversation every utterance is for the other. In multi-party conversation this need not be the case; a party may choose to act as though a certain utterance were directed at a present other. So an utterance in a two-party conversation may do or mean something quite different in a multi-party conversation (Sacks, 1995, p. 668). Although two-party talk may contain more than two people, Sacks discussed the notion of “overhearers” in relation to multi-party conversations. He stated that in the case of some talk between A and B, C should still listen because talk might be directed to C at some time in the future. Therefore the particular case of overhearers in multi-party conversation requires a different order of analysis (Sacks, 1995, p. 531). “So one gets, then, a set of intrinsically two-party activities which, when done in multi-party situations, can be different sorts of things.” (Sacks, 1995, p, 532).

 

The notion of “overhearers” or “witnesses” is relevant to understandings of classroom interaction given the two-party nature of whole class or small group talk led by the teacher. The dyadic nature of talk in classrooms is evidenced most particularly during teacher-led activities involving the whole class. The teacher controls the speaker turns and talk exhibits a two-speaker pattern, although large numbers of potential speakers are usually present. This pattern of interaction has been found to characterize teaching activity in much of the research examining talk in classrooms (Edwards & Westgate, 1987; Mehan, 1985). Macbeth referred to the two-party organisation of the lesson as “our familiar sense of classroom order and instruction” (2001, p. 61). During teacher-led instructional whole-class setting, most of the students are “overhearers” to the two-party talk that occurs and yet are always potentially the next party to be called on to speak.

 

The multi-party nature of teacher-led whole class instruction requires talk that will ensure shared attention (Atkinson, 1982, p. 97). Teacher-led talk during whole-class instruction is two-party talk. So turn-taking in teacher-led instruction involves teacher control of talk. The teacher is able to take many turns, to nominate individuals to speak, and to evaluate the responses of students. Thus, talk in classrooms is found to differ from that of mundane conversations in ways that enable the accomplishment of activities particular to the classroom, and constitute classroom talk as institutional talk.

 

Research during times of individual activity in classrooms suggests differences between it and teacher-led whole-class instruction. Szymanski (1999) applied Goffman’s notion of copresence (1963, cited Szymanski, 1999) to describe the interactional nature of the group work setting. In her study of bilingual children working in groups in a language arts classroom, the physical proximity of the children resulted in ever present opportunities for interaction or access to a number of “interactionally relevant items” (Szymanski, 1999, p, 2). Children did not talk all the time but rather came in and out of conversations. Szymanski examined children’s methods for re-engaging and dis-engaging in talk.

 

3.5.4 Turn design

Lerner’s (1995) study of students’ participation opportunities included examination of teachers’ turn designs and opportunities for interaction afforded students by these.  One such design incorporated unfinished turns that invited completion by students. Two forms of these were found to be turns designed as elicitation questions and turns designed as lists when involving an “unfinished turn-constructional component” (Lerner, 1995, p. 115). In the case of the latter the turn performed at least two types of work. It offered a place where a student might take a turn and gave a deal of information about what that answer might be. For example the empty (third) slot in a list needed to be filled by items of the sort already named by the teacher.

 

Lerner described the elicitation question (Mehan, 1979; Sinclair & Coulthard, 1975), as initiating a “small sequence” consisting of elicitation-reply-evaluation. According to his analysis, “Asking an elicitation-type question made an answer specially relevant and provided the sequential possibility of an evaluation.” (Lerner, 1995, p. 115). An answer that was not acceptable might have resulted in a “sequence expansion” (Mehan, 1982) where withholding a completing act afforded possibilities for continued student participation (Lerner, 1995). Lerner established that students oriented to elicitation questions as demonstrations of knowledge rather than as requests for information (Lerner, 1995, p. 112).

 

Lerner explicated the way that the variations of the elicitation questions gave a focus to particular information, and so he established how teacher known-answer questions took account of student needs. For example the list construction pattern might have increased the predictability of students’ attempts. At the same time unfinished turns still appeared to be another form of “guess what the teacher is thinking”, only with narrowed options for possible answers.

 

Lerner (1995) examined how a teacher’s turn design worked to provide a resource that enabled collaborative construction of written answers in a word definition writing task. The teacher instructed students through her talk, although she didn’t tell. The teacher used an interrupted turn involving the beginning of the form of writing to be used. The next turn, by a student, required the completion of the teacher’s utterance and hence the sentence itself. So the teacher “made it clear in the manner in which she has designed her utterance and gestures that she could complete the turn unit, but that it is the students’ (instructed) task to do so.”(Lerner, 1995, p. 122). Lerner found the turn to be “a vehicle for instruction”. This finding, however, cannot make clear whether students oriented to the talk as instructional.

 

Lerner’s work illustrates that talk, as it is shaped by the teacher, exhibits features particular to the classroom and to the work of instruction. For example, it demonstrates that instruction is carried out largely through questioning of students. Further, Lerner’s work shows that successful instruction in the classroom requires competent use of particular interactional formats, such as, question-answer-comment sequences (Heap, 1992). Teachers draw on interactional formats to lecture, ask questions, accept feelings, praise, use student ideas, give directions, and justify authority. Students draw on relevant interactional formats in a classroom to respond appropriately to the teacher, or initiate talk with the teacher.” (Heap,1992, p. 23).

 

McHoul noted ‘now’ as a “classic teaching technique” that teachers used to return to the “main events” of lessons (1978, p, 205). This occured, for example, when students introduced a new topic or attempted to change the topic ‘we’re on” during teacher-led instruction. Although referred to “in passing” in some of the literature, the role of “now” in classroom interaction appears not to have been well explored. Schifflin’s (1987) analysis of “now”, used as a discourse marker, is the most extensive study of it, although Schifflin does not employ CA.

 

Schifflin found that the use of ‘now’ “marks a speaker’s progression through discourse time by displaying attention to an upcoming idea unit, orientation, and/or participation framework.” (Schifflin, 1987, p. 230), and occurs for example, in comparisons, lists and arguments. The word ‘now’ may also mark a switch in modes, from declarative to interrogative turns, or a speaker shift from narrative or evaluative. In its use as a discourse marker, ‘now’ “provides a temporal index for utterances with the emerging world of talk” (Schifflin, 1987, p. 245), looks ahead or operates catephorically. The temporal use of ‘now’ locates an utterance in an “ego-centred space” (Schifflin, 1987, p. 245) and gives a focus to what the speaker is about to say. As well, ‘now’ can mark a speaker shift in orientation to interpretation of own talk. It also may involve a struggle for turn taking, when a speaker needs to negotiate the “right to control what will happen next in talk” (Schifflin, 1987, p. 241). Young children evidence greater use of ‘now’ in their role plays of teachers than in their ordinary conversations (Hoyle, 1994).

 

3.5.5 Adjacency pairs

The early work of conversation analysts established the primacy of adjacency pairs as an aspect of talk-in-interaction. The discovery of this unit was important in that “for the first time in the study of social interaction, sequential structures of actions were discovered in naturally occurring situations.” (Psathas, 1992, p. 20). The adjacency became one of the most widely known concepts of CA (Hutchby, 2001), perhaps “the linchpin” of the approach (Taylor & Cameron, 1987) and “the most powerful device for relating utterances (Sacks, 1995, p. 554; cited in Silverman, 1998, p. 104).

 

Adjacency pairs are sequences of at least two turns in length, and they consist of at least two parts. The first part is produced by one speaker and the other by a second speaker. The parts are relatively ordered and discriminately related such that the second part must be an appropriate pair to the first (Silverman, 1998); the sequence involves immediate next turns where the parts are conditionally relevant to each other (Psathas, 1995, p. 18). Or rather more simply put, an adjacency pair involves “a pair of utterances in which the second is functionally dependent on the first.” (Cameron, 2001, p. 96). Examples of adjacency pairs include greetings and return greetings and summons-answer pairs.

 

Sacks examined and detailed the phenomenon he termed adjacency pairs or “utterance pairs” (Sacks, 1995, p. 104). Adjacency pairs were shown to be ordered such that the two parts normally occur next to each other, and the use of the first part of the pair is regularly followed by the second (Sacks, 1995, p. 667). Adjacency pairs involve the display of mutual understanding such that the absence of the second part is relevant and subject to inference.  Thus adjacency pairs exhibit conditional relevance (Schegloff, 1968, cited in Hutchby & Wooffitt, 1998, p. 42). This involves a normative aspect in a number of senses. Motivational inferences may be drawn in the absence of the second pair part, the second part need not occur immediately following the first part and dispreferred responses are marked (Hutchby, 2001). The first part of the pair as well as the second may be viewed as constrained in that to get the second you may need to initiate the first (Silverman, 1998).

 

A complementary notion is that of preference. Preference refers to the fact that there may be alternative forms of second part replies. Some adjacency pairs with preferred and dispreferred second-pair parts are: offer-accept/decline, accusation-rebuttal/justification, assessment-agreement/disagreement, self-deprecation-disagreement/agreement; accusation-denial/admission and request-acceptance/refusal. Preference is a controversial notion (Taylor & Cameron, 1987) best conceived of as being achieved without ‘markedness’ (Heritage, 1984a; Levinson, 1983), or as “ways that things are typically done” (Cameron, 2000, p. 97). Dispreferred actions are problematic (Silverman, 1998) and their status, as such, is marked.  Taylor and Cameron, drawing on Levinson (1983, pp. 334-335), listed the following markers that occur as part of disprefered utterances: pausing, prefaces (‘uh’, ‘well’), token agreements, appreciations and apologies, qualifiers, use of accounts and use of a ‘declining component’. The features are important since their use displays users’ awareness of the relevance of the rule (Taylor & Cameron, 1987).

 

Wootton’s (1981) examination of children’s request sequences illustrated some aspects of adjacency pairs and preferences. Grantings and rejections by parents showed distinct differences on several points. Rejections were dispreferred and sometimes involved accounts or “softenings”. There was a greater likelihood of a certain type of rejection leading to extended sequences of talk. When delay techniques were employed children were likely to persist with the request. Deliberate/rapid topic switches could act to terminate a particular sequence (Wootten, 1981, p. 76).

 

Notwithstanding the integral work of adjacency pairs in talk-in-interaction, it is possible for talk to extend beyond the adjacency pair. For example, three part sequences may occur when someone asks for the time and thanks the respondent as an acknowledgment of the answer (Sinclair & Coulthard, 1975, p. 37 cited in Mehan, 1985, p. 126).  A question followed by a question extends the adjacency pair. Psathas discerned this pattern in his study of direction giving. He termed it a conversion sequence. In direction giving the question-question turns worked to establish, for example, the starting point for giving directions after the question that asks for directions. A question followed by a question may involve two adjacency pairs, one inserted inside the first. The insert sequence need not disrupt or cause trouble (Schegloff, 1968) but results in the separation of the two pair parts of the first adjacency pair. In the classroom if the first question should come from the teacher then it also involves two adjacency pairs. The conclusion to the first is tied to the completion of the second. The second is completed by an evaluation act which itself rests on the correctness of the answer to the first pair part of the first adjacency pair initiated by the teacher (Mehan, 1979).

 

Sacks (1972) showed how children’s use of the question “You know what?” serves to produce a question answered by a question. The use of “What?” in response to the original question enables the child to tell the story. This is because the use of “What?” obliges the child to answer and thus, to tell the information or story. Sacks claimed that in this way the child is provided with the opportunity to say what he intended in the first place only it becomes a “a matter of obligation.” (Sacks, 1972, p. 344). The question enables children to get the floor, and perhaps to have a longer turn despite their usually restricted talking rights.

 

Four-part structures are similar to adjacency pairs but consist of four turns (Psathas, 1992). Psathas examined the example of an invitation-acceptance adjacency pair which has a prefactory sequence (or presequence). For example, the question “Are you doing anything tonight?” may lead onto an invitation to go out, depending on the response to the question. The outcome of the presequence is relevant to what occurs next. Similarly, closings might involve a sequence that prefaces the actual terminal sequence. So the four-part sequence involves “two ordered adjacency pair structures in which the first adjacency pair implicates what could be a relevant second adjacency pair (Psathas, 1992, p. 19).

 

Sacks referred to chains of adjacency pairs (1995, p. 529). He considered the example of a question-answer followed by an invitation-response with the first adjacency pair, for example, working as a pre-signal that an invitation is to follow (Silverman, 1998, p. 107). Another example might be a series of question and answers since once an answer is given the first party has the right to speak again and therefore to ask another question. This is termed the ‘chaining rule’ and it operates in certain kinds of professional-client settings” (Silverman, 1998, p. 108). Use of the chaining rule constrains the options open to speakers in ways that make it different to the turn-taking of casual conversation.

 

Merritt (1976) examined questions that followed questions in service encounters. She found for example that a second question-answer pair could be embedded within a first question- answer pair such that the answer to the first question is contingent on the answer to the second question.  So, a request for cigarettes might generate a question regarding details of the brand of cigarette. She also described another question-question sequence, she termed “elliptical coupling” (Merritt, 1976, p. 348). This consisted of a first question-answer “pair” where the answer was ellipsed and a question filled the slot such that the sequence became question-question-answer.

 

There are other variants of the chain rule (Frankel, 1990). A side sequence occurs where “large scale activities can warrantably be interposed between the initiation of a two-part sequence such as question and answer and its conclusion.” (Frankel, 1990, p. 236). A distinction between side and insert sequences is that insert sequences are restricted exclusively to question-answer pairs (Frankel, 1990, p. 236). According to Jefferson (1972) side sequences occur within some on-going activity. They represent a break in the activity but at the same time are relevant to it. She examines the example of a counting mistake in a children’s game. Repair is accomplished to the error, in a side sequence, and then the game continues.

 

A question powerfully requires an answer and gives the asker rights to control the next cycle of conversation. So, to respond to a question with a question is to exert countercontrol and may require that the responder has greater or equal social power (Mishler, 1975, p. 106 cited in Frankel, 1990, p. 260). Frankel (1990) established doctors’ dispreference for patient initiated questions during medical consultations. Frankel also noted that jargon in medical encounters generates side and insert sequences, and that the use of “now” generated a “projectably” new sequence.

 

3.5.6 Extended sequences

Although extended sequences are discussed in some of the CA literature, generally much less attention is given to longer sequences of talk, in particular to those that extend over many turns. Early on Sacks (1995, p. 355) identified “long sequences of talk” as an appropriate area for investigation, though one that was only in its initial stages of development. In his discussion of long sequences of talk, Sacks made a distinction between talk that has a known pre-organisation (as in certain games for example) and talk that does not.

 a basic sort of investigation is that of long sequences as a coherent matter as compared to simply studying, utterance by utterance, a long sequence of talk which you then have as an in-some-way connected series of small fragments. (Sacks, 1995, p. 355.)

 

According to Sacks, the analysis of long sequences of talk is not an additive process and treating it as such may not inform understanding of the whole. Psathas also adopted the analytic perspective that extended sequences, as he termed them, “need to be analyzable and understandable as whole units” (1992, p. 99). Psathas (1992) claimed that extended sequences are more than just a string of sequences. So they need to be considered and examined from the perspective of their whole, and complex, structure rather than through a sequence by sequence analysis strung together (Psathas, 1992, p. 100). He considered extended sequences to be longer than four turns.

 

According to Psathas, this type of sequence includes types of lessons (Mehan, 1979; Psathas, 1992), stories (Sacks, 1970) and direction giving (Psathas, 1986a; 1986b; Psathas & Kozloff, 1976, cited in Psathas, 1992). Sacks analysed a competition sequence (Sacks, 1995) and stories (Sacks, 1995) as examples of long sequences. The latter involved his analysis that a story preface asks for the right to take an extended turn in order to tell the story (Sacks, 1995, p. 226). According to Sacks, the story preface also contains within it the ‘seeds’ of the closing of the story since it gives information about what will be in the story for it to be over.

 

Psathas approached the analysis of extended sequences from three points: how do parties enter into the activity; how do they exit, and what is the internal structure of the activity. He used this analytic approach to examine direction-giving (Psathas, 1986a, 1986b, 1990). Psathas worked from a gross characterization of the features of direction sets and determined that directions appeared as “coherent conversational units” (1995, p. 23) that involved methods for entering, proceeding with directions through a sequence of talk, and closing the direction giving. Directions were shown to be collaboratively produced and to consist of “multi-turn extended sequences of talk” (Psathas, 1991, p. 196).

 

According to Psathas (1991), entry into direction giving is usually initiated by a request and sometimes this will be answered by a request for a starting point (Psathas, 1991). This conversion sequence functions to establish the starting point for directions. Apart from destination and starting point, initial talk establishes time and mode of travel and membership categorization of the parties involved in the talk. As well, direction-giving involves insertions (Schegloff, 1972), or suspensions from the on-going main activity. These result from orientations to repair or requests related to further detail. They can be initiated by either party and do not disrupt the overall function of the activity (Psathas, 1991, p. 205). The activity is closed in two parts. The talk hearably “arrives at” the destination, and the first party produces an acceptance and a positive assessment. Once this is accomplished the conversation moves into closure or shifts to other topics.

 

The organisation of giving and receiving directions was found to be both context sensitive and context free. It was both responsive to particular parties involved in it and showed orderly and patterned ways that could be found across any number of instances (Psathas, 1991, p. 214) of direction sets. Internal features of directions sets included: named and shared understanding of destination; presumption that movement to the destination was possible; presumption that there were “recognizable-locatable” sets of operations related to movement; naming or ordering in sequence; and the sequence of operations that involved a sequence of utterances or single utterance of talk (Psathas, 1991, p. 215).

 

Schegloff (1990) referred to talk occurring in “clumps”. He took the coherence of these as an analytic focus rather than their topic since determining topic can be problematic. He determined that the sequential analyses of “extremely long spates of talk” (Schegloff, 1990, p. 72) provided understanding of its coherence, as it does in the analysis of adjacency pairs. Schegloff argued that sequential analysis of the sequence is the proper unit of analysis of an extended sequence of talk. Schegloff’s argument highlights the issue of coherence across stretches of talk and is taken up by other researchers. To illustrate, he analysed a telephone conversation involving many turns at talk. Schegloff established that attention to the sequential structure of the talk shows talk-in-interaction that is coherent (Schegloff, 1990, p. 66) and cohesive, despite what appeared to be a lack of topical coherence.

 

Schegloff examined presequences, or sequences specifically designed to come before some other turn type. For example, preinvitations come before invitations and preannouncements before announcements. Presequences act to find out whether a next action will not be dispreferred. Schegloff examined prerequests and three types of responses. These were: talk that acts as a go-ahead by moving the action forward to the request; one which blocks the talk from proceeding to the request; and a preemptive response in which the one addressed offers what they think will be requested next (Schegloff, 1990, p. 61). A prerequest may also involve an action-projection through its format. Schegloff delineated two types of these. One is a predelicate where the utterance makes it known that what is coming up may be delicate. The other type is prepres, or preliminaries to preliminaries. An example of the latter would be “Can I ask you a question?”. Prepres are frequently not followed by the action projected but by preliminaries.

 

Goodwin (1990) examined stories within the structure of argumentative sequences involving children. She analysed one story in particular which began with a sequence of paired exchanges, the challenge and threat/ counter to it, termed paired counters, and then the introduction of a story. The latter enabled other participants to enter the interaction. Goodwin noted the role of “ratified overhears” in the initial part of the dispute (1990) and their later role as invited participants in the story. Audience requests for expansion or repetition illustrated that oral stories consist of more than pre-determined sequential structures.

 

In arguably the most notable studies of classroom interaction, Mehan (1979) established many distinctive features of classroom talk. Mehan described his research as constitutive ethnography but drew heavily on conversation analysts and on Garfinkel’s work. Mehan examined the internal structure of classroom lessons, in particular their instructional phase. His study was of a teacher’s lessons with first, second and third graders in a culturally diverse classroom.

 

The characteristic interactional units in lessons were “elicitation sequences” that involved the exchange of academic information. These were found to be jointly produced, sequential and consisting of three parts described as the initiation act, reply act and evaluation act (IRE). The three-part sequence consisted of two “coupled” adjacency pairs, where the second part of an initiation reply pair became the first part in the second adjacency pair that was completed by the evaluation act. The end of the sequence was marked by positive evaluation. Mehan established that negative evaluations did not appear at the ends of sequences, only in their interior since they work to “keep the interaction moving until symmetry between initiation and reply acts is established.” (Mehan, 1979, p. 65). Mehan’s definition of an extended sequence involved the absence of that completing turn.

 

According to Mehan, extended sequences occur in classroom talk when the expected reply does not occur immediately. This may be because of partial, incorrect or no answers. The initiator employs strategies to obtain the necessary reply. Strategies include prompting, repeating initiation acts or simplifying. The completion of an extended sequence is marked, like three-part sequences, with “positive evaluation of the content of students’ replies, slowed cadence, and manipulation of educational material.” (Mehan, 1979, p. 122).

 

Mehan took up the issue of how talk operates and is understood in classrooms, where it frequently occurs across considerable stretches of interaction and involves numbers of potential speakers. He posited extended sequences, together with adjacency pairs, as reflexively tied structures (Mehan, 1979, p. 76). Mehan noted that reflexive tying occurs also at increasing levels in the lesson, contributing to the social organisation of interactional sequences and providing “the glue that binds entire interactional events together.” (Mehan, 1979, p. 76).

 

Mehan characterised the instructional phrase of lessons as “a progression of topically related sets” (Mehan, 1979, p. 122). Topically related sets consisted of a basic sequence and then often, one or more conditional sequences. The latter built on the topic introduced through the basic sequence; they never appeared alone and they appeared only after a basic sequence has been completed. Therefore their appearance was “dependent upon the prior appearance of a basic sequence.” (Mehan, 1979, p. 65). Hence conditional sequences were tied to basic sequences.  Mehan found that topically related set beginnings were often signaled by particular paralinguistic, kinesic and verbal behaviour for example, “uh now” or “now”, together with changes in tone and cadence of speech. These markers were “a closed set” –they didn’t appear at other times in the lessons (Mehan, 1979, p. 66).

 

Mehan’s analysis discerned sequential and hierarchical structures in lessons. He also established that students showed their orientation to the components of sequences and the hierarchical structure of lessons (Mehan, 1979, p. 35). This was a critical finding since it supported his claim that interaction in lessons was tied by more than individual sequences of talk. Cohesiveness in lessons was achieved across extended sequences of talk, and across topically related sets involving numbers of three-part turn sequences.

 

McHoul (1990) examined talk aimed at repair (Schegloff, Jefferson & Sacks, 1977) that potentially involved a number of turns. He termed these “expanded sequences of talk”. Expanded sequences are distinctive because they involve trouble and repair, although they involve IRE sequence, so have features in common with Mehan’s notion of extended sequences.

 

McHoul (1990) considered repair trajectories in the classroom. One type of trajectory, where students made corrections to their own trouble source in response to a teacher initiation, was in greater evidence than teacher corrections of talk or self-corrections without initiation by the teacher. Classroom repair was seen to be different from ordinary conversation since it was marked by a preference for self-correction following initiation of repair by the teacher.

 

McHoul explicated three related aspects of repair: clueing, “recycling” in the form of expansion sequences and “withholding”. Withholds include clueing, reformulations and redirections of questions to other students during whole-class instruction. Although withholding occurs in ordinary conversation (Schegloff et al. 1977) it appears differently in classroom talk. Rather than withholding the initiation of a correction after trouble has occurred in order for self-repair to occur, the teacher initiates correction and withholds other-correction. McHoul also noted the frequently tentative nature of other-initiation. This “modulation” (Schegloff et al. 1977, cited here p. 367), is integrally tied to the dispreferred states of other–corrections in the classroom.

 

A consequence of this repair is that its trajectory can extend beyond the pattern found in ordinary conversation where repair space is frequently limited to a maximum of three turns (Schegloff, 1992). McHoul suggests that expanded other-correction withhold sequences (McHoul, 1990, p. 364) may be features of “learning”, “socialization”, or “competent/precompetent interaction genres.” (McHoul, 1990, p. 367). The absence of direct repair by teachers has been noted in other studies in the classroom (Freebody, Ludwig & Gunn, 1995; Ridley, Radford & Mahon, 2002) and found to be potentially confusing for students. Freebody, Ludwig and Gunn refer to extended sequences of talk in the classroom as requiring “trial and error, guessing, or other indirect reasoning practices to complete the task” (1995, p. 343).

 

According to Mehan an extended sequence resulted from the absence of a completing turn in classroom talk initiated by the teacher. McHoul (1990), on the other hand, described expanded sequences which act to allow students to self-repair their own trouble source in talk and result in talk trajectories that go beyond patterns found in ordinary conversation. There is overlap between the definitions used by Mehan and McHoul. Although Mehan did not refer specifically to trouble, the reasons given for the employment of strategies that expands sequences clearly involve trouble. For example, silence in response to a question could be trouble. Similarly, the strategies that Mehan outlined fit with McHoul’s notion of clueing, reformulation and so on. For Mehan, the sequence was completed when a positive evaluation was given. For McHoul it was when repair occurred.

 

3.5.7 Applications

Within the field of education ethnomethodology and CA have been applied in areas as diverse as special education (Radford & Tarplee, 2000); early childhood (Danby, 1998; Leiminer, 2000); and music education (Roulston, 2000). Reading instruction has been a strong focus and boasts an impressive body of work (see Baker, 1991a, 1991b; Baker & Freebody, 1993, Freebody & Freiberg, 2000, 2001; Heap 1980, 1985; McDermott, 1976, cited in Mehan, 1985). Writing classrooms in comparison have received limited attention.

 

Freebody and Freiberg (2001) framed a cogent argument for what they termed, an applied ethnomethodological approach to examining reading instruction. They claimed that although an extensive body of research has examined aspects of reading instruction, much of it actually misses everyday teaching and learning practices. In particular, they argued that all sides of debates involving effective reading instruction draw on terms that are merely glosses for practices that continue to be unexplicated and taken-for-granted.

 

Baker (1991) explicated three benefits to be gained from the application of CA to everyday practice in education. First, it makes it possible to study events as they are observed, unlike the case with unseen psychological or sociological processes. Second, it focuses on instances of observable actions rather than on how they relate to educational theories themselves. Third, it results in “principled, publicly available methods” (Baker, 1991, pp. 163-4; cited in Freebody, 2003, p. 97) for examining actual events and their consequences.

 

Baker (1991) argued that pedagogical descriptions and accounts, and conventional reading psychology accounts miss what is actually accomplished in reading lessons.  She showed that it is through the use of analysis of interactional and situated reading events that we can understand and describe the ways in which classroom relations are construed as part of reading in the early years. Baker examined the ways in which question and answer sequences within whole class reading events were “foundational in the social construction of classroom literacy and in the social production of differences among students as classroom-literate speakers. “ (1991, p. 180). Baker established that reading lessons accomplish “institutionalized ways of reading and talking about texts with teachers in classrooms, and simultaneously the assembling of social relations and social order for classrooms and for schooling.” (1991, p. 161).

 

Much of the CA research in classrooms has given an emphasis to teacher-led whole class instruction, or small group instruction led by the teacher. It establishes that:

In addition to accumulating a stock of academic knowledge, students need to accumulate a stock of social knowledge. Learning that there are appropriate ways to cast academic knowledge, that certain ways of talking and activity are appropriate on some occasions and not on others, learning when, where, and with whom certain kinds of behaviour can occur are some of the essential constituents of the stock of social knowledge that accompanies academic knowledge; the intertwining of the two is relevant for effective participation in the classroom community.” (Mehan, 1985, p. 119)

 

As section 3.5 has illustrated, an extensive literature demonstrates what is known about the ways in which students and their teacher accomplish aspects of their social activity, particularly in whole-class settings. This thesis examines the “stock of social knowledge” required to accomplish the social organisation of independent writing, a time of individual activity rather than whole-class instruction.

 

3.6 Summary

This chapter has been used to outline theoretical, methodological and conceptual aspects of ethnomethodology and CA.  Integral to these is the notion that both provide an alternate sociology, an alternate way of examining and explicating sociological phenomenon. The application of CA to studies of institutional activity involve a number of questions and issues, hence considerations for this study. The analytic approach taken in this study involves consideration of Schegloff’s (1992) problem of relevance and issue of procedural consequentiality. It also involves comparisons using ordinary conversation and with interaction that accomplishes teacher-led whole-class instruction.

 

Many distinctive features of classroom talk have been explicated in CA studies however most give focus on teacher-led talk during whole class or small group instruction. Much less is documented about sequences of interaction between students, or where the teacher is absent from the interaction entirely. Hence descriptions and understandings of classroom talk frequently encompass interaction involving the teacher. Expanded or extended sequences provide a case in hand. Both the terms “extended sequence” and “expanded sequence”, used to describe this in classrooms, define the sequences in relation to the role of the teacher and her instructional goals. This suggests an issue, to be considered later in this thesis, as to whether these terms can be applied to interaction when the teacher is not party to the talk yet numerous turns result.

 

The next chapter of this thesis outlines the study design. This involves the broad application of ethnomethodology and CA to the study of writing, and the development of a focus on the social organisation of independent writing. The chapter will detail considerations relevant to ethical clearance and the research site, observation and collection of data, and the analysis of it.