Volume 1, No. 3, Art. 3 – December 2000
Innocence and Nostalgia in Conversation Analysis: The Dynamic Relations of Tape and Transcript
Malcolm Ashmore & Darren Reed
Abstract: This paper attempts an analysis of some of the methodological practices of Conversation Analysis (CA); in particular, tape recording and transcription. The paper starts from the observation that, in the CA literature, these practices, and the analytic objects they create (the tape and the transcript), are accorded different treatment: simply put, for CA the tape is a "realist" object, while the transcript is a "constructivist" one. The significance of this difference is explored through an analysis of the dynamics of CA practice. We argue that the "constructivist transcript" is premised on an understanding of CA as predominantly concerned with maximising its "analytic utility": a concern of one distinct temporal stage of CA work: that of the "innocent" apprehension of objects in the "first time through". The "realist tape", in contrast, is based on a different aspect of the work of CA: its quest for greater "evidential utility", achieved by the "nostalgic" revisiting of previously produced objects for purposes of checking them against each other; work done in the "next time through". We further argue that both the ontology and the epistemology of CA's objects are changed in any next time encounter. We conclude with a cautionary speculation on the currently-projected, transcript-free, digital future of CA.
Key words: conversation analysis, tape recording, transcription, rhetoric, epistemology, phenomenology, realism, constructivism
Table of Contents
2. The Realist Tape and the Constructivist Transcript
2.2 The tape
2.4 What counts as data: The mutual elaboration of tape and transcript
3. The General Epistemo-Phenomenological Schema
4. The General Schema Applied to CA
4.1 Nostalgia dynamics
4.2 Reverse checking: Doubting the tape
5. Innocence and Nostalgia: First Time, and Next Time, Through
5.1 First time through—Innocence
5.2 Next time through—Nostalgia
6. Concluding Discussion
In this paper, we examine certain of the methodological practices of Conversation Analysis (CA). Thus, this paper joins a number of recent discussions of CA which are variously concerned with CA's foundational concepts (ten HAVE 1997, LYNCH & BOGEN 1994), its relationship to ethnomethodology (CLAYMAN & MAYNARD 1995, LYNCH 1993, p.203-264), the proper range of its empirical materials (McHOUL 1987, MOERMAN 1988, NELSON 1994), and with analyses of its empirical practices (ANDERSON & SHARROCK 1994, BOGEN 1992). All of these texts are authored by people who can and do claim some sort of membership in the ethnomethodological/conversation analytic community. Though our membership is much less secure, we wish the following text to be taken as similar in spirit to the above, i.e. as "insider critique" motivated by a deep agreement with the broad aims and substantial achievements of this intellectual movement. 
CA is a "unique" (ten HAVE 1990) form of qualitative social research, both in its restricted topical orientation and its rigorous methodological procedures. CA is exclusively concerned with the analysis of "talk-in-interaction"; usually, but not always, of casual, or mundane, conversation. According to CA, this is the primordial stuff of social interaction: at once the most mundane and the most consequential of all social phenomena. As a practice, ordinary talk is not considered by its practitioners to be particularly skilled (presumably because it is so basic, so pervasive, so ordinary); yet CA shows it to be a precision instrument, wielded by maestros. Subtle, nuanced and highly sensitive; yet structured, normative and accountable; it displays "order at all points" (SACKS 1984, p.22), yet is entirely improvised. Moreover, the doing of talk produces and reproduces all the supposedly "external" phenomena of the socio-psychological sciences: persons, interaction, groups, membership categories (class/gender/ethnicity), the "sense of social structure" and ultimately society itself (BODEN & ZIMMERMAN 1991, SILVERMAN 1998, HUTCHBY & WOOFFITT 1998). 
CA is done in a very specific way: starting with audio, and sometimes video, recordings of "naturally-occurring" verbal interactions; transcribing them in a manner designed to show the "hows" of talk rather than just the "whats"; and subsequently developing highly detailed analyses of the ethnomethods of talk. We are interested in this paper in how conversation analysts work with, work on, and work up the series of "analytic objects" produced in and as the course of their research. Put schematically, these objects are the Event, the Tape, the Transcript, the Analysis and the Article.1) How does the apparently fixed, temporal, and linear relationships between these objects change at different points of the research process? How does the "evidential utility" of these objects vary with respect to their "analytic utility"? What is the difference between seeing (say) a transcript, and reading it; between hearing a tape and listening to it? At what points during the dynamic process of research are objects encountered by the epistemic mode of seeing/hearing as opposed to that of reading/listening? 
In developing our understanding of these dynamic "epistemo-phenomenological" processes, we focus on the Tape and the Transcript as mutually elaborative analytic objects. We examine how the practices of tape recording and transcribing are described in the CA literature, and how this literature variously formulates the practical and epistemic relations between the analytic objects these practices produce—the Tape and the Transcript. We then introduce an abstract and highly general "epistemo-phenomenological schema" which, when subsequently applied to CA, orientates our explication of the process we call the "general and evidential nostalgic dynamic", including its epistemic functions, and the varying and relative roles played within it by the set of analytic objects. 
Specifically we reveal the relative reification of analytic objects such that the original phenomenon, the tape, the transcript (and the analysis) are actively construed differently depending on whether the object is apprehended in a state of "innocence" or "nostalgia". This is pursued through a "first time through" and "next time through" trope. 
2. The Realist Tape and the Constructivist Transcript
The first thing to notice is that taping as an activity receives much less explicit discussion than does transcribing. While transcription (in CA and other forms of research on talk) has been the specific topic of at least one edited volume (EDWARDS & LAMPERT 1993), several research articles (OCHS 1979, JEFFERSON 1985, PACK 1986, COOK 1990, PSATHAS & ANDERSON 1990, MISHLER 1991, O'CONNELL & KOWAL 1994, JEFFERSON 1996, GREEN, GRANQUIZ & DIXON 1997), and is routinely and extensively discussed in recent CA introductory textbooks (PSATHAS 1995, HUTCHBY & WOOFFITT 1998, SILVERMAN 1998, ten HAVE 1999), recording has received less attention. This is not, however, because the Tape is considered less important; indeed, in many ways, as we shall detail, it appears to carry more weight than does the Transcript. 
When the practice of recording is discussed, what is addressed tends to be its technical aspects (GOODWIN 1994, ten HAVE 1999, Ch.4, MODAFF & MODAFF 2000).2) Moreover, most discussions concentrate on the relatively novel technology of video recording. A plausible reason for this neglect, is that audio recording (and its results) has become so culturally naturalised, that, like photography, it is extremely difficult to problematise, to loosen the hold of its stubborn realism (cf. MISHLER 1991, pp.255-259). Though many, perhaps most, (Western) children have had the vertiginous experience of hearing and not recognising the sound of their own recorded voice, the "doubt" that this experience potentially raises is quickly reframed as normal and explicable. 
2.2 The tape
A typical statement of the importance of the Tape in CA is from POMERANTZ and FEHR (1997, p.70): "Conversation analysts strongly prefer to work from recordings of conduct". The reason given for this preference is four-fold:
"certain features of the details of actions in interaction are not recoverable in any other way"
"a recording makes it possible to play and replay the interaction, which is important both for transcribing and for developing an analysis."
"a recording makes it possible to check a particular analysis against the materials, in all their detail, that were used to produce the analysis."
"a recording makes it possible to return to an interaction with new analytic interests." (1997, p.70) 
Notice first that these remarks address the tape as a "found object": they are concerned with what can be done with a recording, not with the activity of recording itself. The origin of the Tape—its relation to any particular Event—is not of specific interest. This lack of interest in the process of recording (whether as a technical or conceptual issue) is an important first step in the "naturalising" of the Tape. In effect it provides for the "forgetting" of the Event, and its wholesale replacement by the Tape. 
The first desideratum attends to the technical necessity for this replacement. Neither the Event-in-itself nor any other procedure for its "reconstruction"—ethnographic observation, memories, post hoc inquiries, intuitive invention—is adequate. The second reason introduces the idea of "replayability": the Tape, as opposed to the Event, can be encountered more than once; it can be re-heard, can be subject to repeated listening. Reasons three and four specify uses for the Tape's replayable character: the Tape can be used as the standard against which the Analysis can be checked; and it can be revisited to produce a new Analysis, i.e. the Tape can be the source of more than one series of analytic objects. 
"[I]t is a truism to note that all transcription is in some sense interpretation ..."
(COOK 1990, p.12)
"A first observation is that there is not, and cannot be, a 'neutral' transcription system." (PSATHAS & ANDERSON 1990, p.75)
"[A] transcript is a text that 're'-presents an event; it is not the event itself. Following this logic, what is re-presented is data constructed by a researcher for a particular purpose, not just talk written down." (GREEN et al. 1997, p.172)
"Transcription as theory" (OCHS 1979, Title)
These statements are typical of the "sophisticated" understanding of transcription in CA and other work on discourse. It is routinely understood as a craft process, as itself a part of the practice of analysis, as conventional and constructive. Debates about "how much" to transcribe (O'CONNELL & KOWAL 1994), or about the consequential features of particular systems and designs (HOPPER 1989, EDWARDS & LAMPERT 1993) are commonplace. Students (and other researchers; see GREEN et al. 1997) are regularly warned not to fetishize the transcript, nor to treat it as the data, as we shall shortly see. 
Of course, we are very far from objecting to this approach to transcription. What we are interested in is the contrast, in CA, between this "constructivist" approach to transcription and the "realist" approach to recording and the tape. As we will detail below, this contrast is itself consequential for the practices of CA, and particularly for how the mutually elaborative relation between the transcript and the tape is understood. 
2.4 What counts as data: The mutual elaboration of tape and transcript
There are various formulations of the roles of tape and transcript in CA, some of which are set out in the following five quotations. They are ordered in an array from the most to the least "transcript-friendly".
"Audio recordings, while faithfully recording what the machine's technology allows to be recorded, are not immediately available, in a sense. The details that the machine records have to be remarked by the listening analyst and later made available to the analyst's audience. It is the activity of transcribing the tapes that provides for this, that captures the data, so to speak." (HAVE 1999, p.6)
"The best way to develop analyses is to use both a tape and a transcript. It is harder to isolate and study phenomena when working only with a tape, and much information is lost when working only with a transcript. Also, without hearing/seeing the tape from which a transcript was derived, one cannot know how much confidence to have in a transcript." (POMERANTZ & FEHR 1997, p.70-71)
"The transcription of data is a procedure at the core of analysis ... It is important to stress that, for CA, transcripts are not thought of as 'the data'. The data consist of tape recordings of naturally occurring interactions ... Given this conception of the data, the aim in CA is not simply to transcribe the talk and then discard the tape in favour of the transcript ... Conversation analysts ... do not analyse transcripts alone: rather, they aim to analyse the data (the recorded interaction) using the transcript as a convenient tool of reference. The transcript is seen as a 'representation' of the data; while the tape itself is viewed as a 'reproduction' of a determinate social event." (HUTCHBY & WOOFFITT 1998, pp.73-74)
"... any claim made about the interaction is answerable, not to a transcript, but in the final analysis to the recording" (WILSON 1999)
"... from bitter personal experience I have learned never to trust a mere transcript too much. The time is soon approaching when transcripts (at least in the traditional sense) will seem far too crude for our analytic purposes—a hopeless attempt to fix on paper what is, in its deepest sense, dynamic." (CARROLL 2000) 
These five quotations encapsulate rather different understandings of the relative place of tape and transcript in CA, which of the two should be given more weight, and which should count as "the data". The first quote from ten HAVE expresses some doubt about what we will be calling the "analytic utility" of the tape. It cannot be used on its own: it is not "immediately available"; it has to be "remarked" (transcribed); and it is this transcription process that "captures the data". Here, then, it is the transcript which is the more valuable of the two objects. POMERANTZ and FEHR put forward a "middle ground" position in which both objects are equally valuable and should be used in conjunction with one another. Like ten HAVE, they note the tape's relative lack of analytic utility, but they also point to the transcript's inability to engender "confidence": in our terms, its relative lack of "evidential utility". The third quote from HUTCHBY and WOOFFITT shifts the objects' relative value in favour of the tape. While acknowledging the importance of transcription as a "procedure at the core of analysis", the thrust of their remarks is an insistence on the superiority of the tape. The tape has this greater status because it embodies the data "(the recorded interaction)" whereas the transcript is merely "a convenient tool of reference".3) These valuations are neatly encapsulated in HUTCHBY and WOOFFITT's final ontological distinction between the tape as "reproduction" and the transcript as "representation". The fourth contribution from WILSON attends succinctly to what accounts for the superiority of the tape. The tape is a better "spokesperson" for analytic claims than the transcript: the evidential buck of "answerability" stops at the tape. Finally, CARROLL projects a desired (digital) future when the "mere transcript" can be abandoned entirely in favour of analyses that work directly on and from recordings. For CARROLL the transcript is both too untrustworthy and too "crude": it not only lacks evidential utility (as noticed by POMERANTZ & FEHR and by WILSON), its analytic utility is also in doubt. 
What we want to argue here is that the disparity between these versions of the mutual relations of tape and transcript can be accounted for by recognising a corresponding distinction in the "phase" or "stage" of the CA research process that these authors implicitly are addressing. A positive evaluation of the transcript (as most clearly evidenced by ten HAVE) corresponds to a dominant concern with what we will call the stage of First Time Through, while those accounts that find the transcript wanting (especially WILSON & CARROLL) are more concerned with the later stage of Next Time Through. Another way to put this is that the value of the transcript makes itself felt most clearly in the business of building the series of analytic objects that make up the "material" of any CA research project and thus in the search for analytic utility. On the other hand, when the tape appears as the "better" object of the two, what is being alluded to is its value in strengthening the evidential utility of the already-produced objects. 
Before we can fully explain these concepts we must first acquaint you with our "General Epistemo-Phenomenological Schema" (see also Figure 1) which provides the essential, if inelegantly-titled, framework for our discussion.
Figure 1: The General Epistemo-Phenomenological Schema 
3. The General Epistemo-Phenomenological Schema
We are describing here a kind of dynamic continuum, whose poles are mythic: on the left, we have "reality", or the "natural attitude", and on the right, "fiction", or the "mediated attitude".4) These are the pure, and impossible, states in which there is no admixture of the other. On the left, a "state of nature", apprehended instinctively by totally embedded "members" devoid of intentionality and sense-making ability. On the right, a "state of mind", in which all actions are self-started by radically self-conscious agents with no direct, nor even mediated, relation to an external world. 
In between these poles is our world; and perhaps any conceivable world.5) Let us map out some of the relations we see obtaining in this space. First, note that it is divided diagonally into two phenomenological areas: "Life", which dominates the left hand side, and "Work", which governs the right. "Life" labels a phenomenology of Being; "Work", one of Becoming. Movement along the horizontal plane (which is also a timeline; see below) reduces the influence of one while correspondingly increasing that of the other. 
Second, there are two corresponding epistemologies, or modes of world-encountering, that are, relatively, more leftward or rightward leaning. Hearing/seeing is leftward leaning while listening/reading tends toward the right. Hearing and seeing have in common a certain automatic quality: routine yet spontaneous, as if called up by the apprehended object as a response to a stimulus. They are the unreflective modes of the natural attitude, our unremarkable ethnomethods for living our lives. The knowledge required to hear and to see is atheoretic: a matter of "knowing how" rather than "knowing that". The effect on the world of engaging in such activities is minimal: the world stays much as it was. 
When encountered through the right hand side epistemic modes of listening and reading, the world is interpreted rather than apprehended, and represented instead of simply responded to. Rightward leaning epistemic activities, then, are forms of productive work. Engaging in them results in the addition of new objects to the world—objects that, as representations or interpretations, change what the world is taken to be. Note, however, that we are not dealing directly here with the products themselves (talk, texts, images, etc.) nor with their "modes of production" (speaking, writing, imaging, ...). We are limiting ourselves to an examination of what, in vocabularies we do not wish to buy into, might be called "responses" or "receptive behaviours".6) 
There is a kind of exchange of qualities in play too. As we move rightwards or leftwards we suffer losses yet achieve compensating gains. Moving right entails a loss of "original detail": of the complexity and richness of "the world as it is" in all its "blooming, buzzing confusion". In compensation, rightward-tending epistemic activities add their own products, their own "novel detail"; which, the further to the right one moves, themselves gain in richness and complexity. And as the reverse movement is made, so the opposite exchange takes place. 
Finally, the left-to-right dimension marks a temporal ordering, on many conceivable scales (species time; a lifetime; even, as we shall see in the next section, the temporality of conversation analytic research). Movement from the left to the right involves a change from a "pre" time to a "post" time. However, this timeline is not a one way street. Shifts from the right to the left, from the "post" to the "pre", can occur. Our name for what may occasion such shifts is the "nostalgia dynamic". By this, we mean to indicate that desire, regularly felt (and encountered in others), born, perhaps, of a SARTREan mauvais fois, for greater simplicity, authenticity, and directness, after what (is deemed to be, at this juncture) a surfeit of sophistication and irony, an overdose of ramified interpretations; simply, just too much rightward-ness. On a larger and more exalted scale, this kind of nostalgia motivates that ever-popular social scientific classic, "The Critique of (Post) Modernity". In Figure 1, we have marked the operation of this dynamic as producing a single shift all the way back from the Fiction mythline to the Reality mythline. However, smaller leftward movements from any point along the timeline are possible; indeed, more common.
Figure 2: The General Schema Applied to CA 
4. The General Schema Applied to CA
We now move from the abstractions of the general schema to its concrete application in the case of CA. Note, to begin with, that in Figure 2 we have inserted the five analytic objects used/produced in CA—Event, Tape, Transcript, Analysis, Article—ordered from left to right.7) Obviously, this is a temporal order of production; but it is more than that as we shall see. Second, the phenomenological division of the general schema, with "Life" at the left and "Work" at the right, has been augmented with, respectively, "Evidential Utility" and "Analytic Utility". A further change is a specification of (some of) the ways that the "nostalgia dynamic" operates in CA, motivating, for example, a shift "back" from Transcript to Tape. Finally, we have labelled the left-hand mythline (which marks off the impossible "reality" pole of the continuum), as, additionally, a "modal-line" marking the shift in modality from the Event to its recording as the Tape; and added a second which marks the modal shift from the Tape to the Transcript, Analysis and Article. The first of these shifts is the more radical: crossing the modal-line from Event to Tape is permanent; there is no going back. Indeed the Event is, in relation to the practices and objects of CA, effectively mythic; as its position to the left of the mythline suggests. The second modal shift—from the aural to the textual—is less irreversible, but it still has significant consequences. 
When we say that the left-right ordering of the analytic objects is more than simply a temporal one, we mean that as each object is produced, one after the other, changes of a phenomenological and epistemic character occur too. For example, while the Event is full of Life, it does not present itself successfully as an object suitable for Work. It has, in its state of brute Being, minimal Analytic Utility. On the other hand, the Event possesses a great deal of Evidential Utility. That is, by pointing to the Event as the natural origin of its endeavours, CA can claim an evidential grounding in the real. (CA is neither fiction nor "merely interpretation" because it is "about" real interaction, done by real people in real settings in real time, as evidenced by the Event.) Note, however, that this pointing is both post hoc and other-directed: it can only occur after the production of other, "later", analytic objects; and it is useful mainly as a rhetorical device to justify CA to non-members.8) We have said that analytic work cannot be easily done on the Event-as-such. In fact the only Work generated by the Event is that of its radical transformation into something other than itself through the process of recording and the production of the Tape. 
The reason the Event is so unsuitable for analytic purposes, is that it is apprehendable only transiently, in the course of its flow, and only "from within". As a part of the scene, and by a member of the setting (who may of course be a researcher), an Event can be, in our terms, seen and heard. But it cannot be read nor listened to. 
With each move to a new analytic object further to the right, the level of "original detail" is reduced and replaced by a corresponding level of "novel detail". Though in the move from Event to Tape, much of the lived complexity of the former is erased, something new takes its place: the Tape introduces novelties that were not there in the "original"; chief among these being "replayability". At the other end of our scale, the shift from Analysis to Article involves a loss of some of the complexity (and certainly a lot of the detail) of a specific analytic exercise, yet compensates for this by the addition of new connections and the refinement of the overall network of conversation analytic results. 
4.1 Nostalgia dynamics
The various backward shifts diagrammed in Figure 2 (from Article to Analysis, from Analysis to Tape, etc.) can be understood to be doubly motivated. First, there is the operation of the general nostalgia dynamic, prompting returns to earlier, and thus more "actual", more "lifelike", stages of the analytical process. What is sought for, in effect, is the recovery of some level of "original detail". Closely connected is the desire to revisit the past for purposes of strengthening the evidential adequacy of the analysis, by checking (say) the Analysis against the Transcript, or the Transcript against the Tape. In each backwards shift motivated in this way, the "earlier" object is treated as (relatively) fixed with respect to the "later". Indeed, on such occasions, the former acts as an unquestioned standard with which to assess the fidelity of its translation into the latter. On these occasions, then, the leftward analytic object is reified. That it is possible to treat objects in this way, however, is the result, we will argue, of how they are treated (worked on and worked up) at the stage of the research process we call First Time Through. 
4.2 Reverse checking: Doubting the tape
But before we advance this argument, we will deal briefly with an apparent exception to our claim that it is the earlier, leftward, object that is reified in the activities of evidential checking we have been describing. 
In one of the few articles we have found which address the activity of audio recording (MODAFF & MODAFF 2000) we find an example of evidential checking in which an object to the right is used to check one to its left. At one point in their paper, MODAFF and MODAFF discuss the differential quality of recordings of each party's contribution to telephone conversations depending on the location of the recording device:
"As a rule, the interlocutor with the recording device attached to his or her phone will be louder and more predominant on the tape. For example, we compared two tapes of the same interactional moment made with two Marantz audio recorders set to the same recording settings but at different locations. Fragments 1A and 1B show the difference in what was hearable on the two tapes:
As these fragments are (of course) transcripts, the work MODAFF and MODAFF's reader has to do in order to find "louder and more predominant" recordings of Jay in fragment 1A and of Dee in fragment 1B, is one of seeing, immediately and imaginatively translated into hearing. Thus, in 1A (Jay's), line 001, we see Jay doing "uh::m (.)", whereas in 1B (Dee's), line 001, we see Jay doing "uh:: (0.2)". With our instructions to see/hear Jay as "clearer" in 1A than in 1B, we can duly do so by finding both additional detail (the "m") and greater refinement (the micropause "(.)"). Similarly, we can find a superior Dee in 1B. The weak and imprecise hearing of "(yeah)" in 1A, line 003, can be seen, in Dee's 1B (line 002, "=Okay") to be seeably/hearably more accurate in three distinct ways. First, the hearing is disambiguated, as visually evidenced by the absence of parentheses; consequently, we know that "yeah" is an artefact, and therefore, that "Okay" is not. Second, then, we know that Dee said "Okay". And finally, we see that its correct sequential placing is at the beginning of line 002, as an utterance latched to the end of Jay's turn on line 001; the addition of the latching symbols ("=" "=") acting as our visual evidence for this third form of increased accuracy.9) 
So what we have here is a rare example of the Transcript being used as the standard with which to check the Tape. This "reverse checking" is explicable in that the only object "to the left" of the Tape is the Event; and, being mythic, the Event is unavailable for this kind of work. Moreover (however "merely technical" it may be—MODAFF and MODAFF's treatment being entirely within the frame of technical adequacy) the examination of recording as a practice disturbs the unproblematic treatment of the Tape as immutable "data". Suddenly, the Tape is as "insecure" and "untrustworthy" as all the other, more obviously crafted, objects. And also, of course, the Transcript in such a procedure takes on the qualities of immutability more frequently attributed, in CA, to the Tape. The contingencies of its production have to be erased, just as the local histories of particular recordings routinely are in CA's rhetoric of method, so as to enable the Tape to act as the data: the object to which claims about the interaction are "in the final analysis" (WILSON 1999) answerable. 
5. Innocence and Nostalgia: First Time, and Next Time, Through
We wish to argue that the mode of epistemic apprehension of CA's analytic objects (whether aurally accessible objects are heard or listened to, and textual ones, seen or read) varies systematically with reference to the temporal "stage" of the research. We conceive of these stages in terms of "innocence" and "nostalgia". When the objects are encountered for the first time ("First Time Through"; cf. GARFINKEL, LYNCH & LIVINGSTON 1981) they are subjected to activities done in the name of the rightward-leaning epistemic of listening/reading. This frame of innocence is a one time process—it cannot be repeated—motivated by the move to greater analytic utility. However, when they are apprehended subsequently ("Next Time Through") they are treated according to the leftward epistemic of hearing/seeing. Next Time encounterings, in our terms, are accomplished through the workings of the nostalgia dynamic, as moves to gain evidential utility.
Figure 3: Epistemic Modes: First Time Through and Next Time Through 
5.1 First time through—Innocence
Let us unpack and elaborate this rather opaque formulation, using the Tape as our example (see Figure 3). A recording of a piece of talk has been made (whether audio or video is immaterial to our analysis). A conversation analyst coming to it for the first time will use the tape to make another analytic object: the Transcript.10) This is accomplished through listening to the tape. Note that there is no Transcript there already; in fact (of course) there are as yet no analytic objects to the right of the Tape on our timeline: they are only projected. 
As we have noted, CA's rhetoric of method generally understands the relationship of Transcript to Tape in terms of "representation" (HUTCHBY & WOOFFITT 1998, p.74) or translation from one modality (aural) to another (textual). In order to achieve this translation as "faithfully" as possible, the Tape undergoes intense and focussed listening. The interpretative and productive act of listening changes the Tape's status from an unknown to a known, from an object that is radically unstable to one which is relatively fixed. Listening polices the Tape. The "rules for hearing" distilled from this process are articulated in and as the Transcript. Thus, at this stage, the Transcript appears not so much as the Tape's translation, but as its caption. In Bruno LATOUR's terms, the "coupled object" of Tape & Transcript, bound together as "image & caption", have begun to take on the character and utility of an "immutable mobile" (LATOUR 1987). As we will see, the immutable character of the Tape is strengthened in subsequent turns (Next Times) which construe it as heard rather than listened to. 
Let us follow this First Time Through trajectory a little further to the right on our timeline. The analyst now has two objects: the Tape and the Transcript. To make the third in the sequence (the Analysis), the Transcript is subjected to intense and focussed reading, with the objective of producing another set of "interpretation rules"; this time, for "what there is to see" in the text of the Transcript. Again, these rules are written in and as the next analytic object, the Analysis, which, once more, acts as the caption to the Transcript's image. And to repeat, what is produced in this process is not just a new analytic object: the "old" one, the one being currently acted on interpretatively and productively (i.e. read), is thereby strengthened, shored up, made more immutable.11) As this reiterated process proceeds rightwards, the result is the production of a series of new, rightward, objects in tandem with a set of worked-on, and thus stronger objects to the left of the current work site. The rightward production of analytic objects in the move towards greater analytic utility thus leaves in its wake a series of changed objects. 
5.2 Next time through—Nostalgia
The work of Next Time Through typically occurs at moments of doubt or distrust. For example it occurs when accounts are called into question, or when a transcript is "checked", "refined" or "revised". In CA "re-listening" (to the Tape) and "re-reading" (the Transcript) are two exemplary moments of the Next Time Through. On these occasions what is re-experienced, according to CA's rhetoric of method, is an unchanged analytic object: each return is construed as though it were the first time the object had been encountered, as though it were through a frame of innocence. This orientation ignores the reflexive effects of returning to an object built through the search for greater analytic utility. 
We understand re-listening and re-reading differently. Motivated by the nostalgia dynamic we have outlined above, the return to and re-working of analytic objects on a second, or subsequent—any Next Time—occasion, is not, for us, strictly speaking, a re- anything. Next Time work is done on a different character of object. The objects worked on in the First Time Through are novel and thus fluid and indeterminate. The actions of listening and reading make them less so, give them a known quality, increase their (potential) evidential utility. As we have argued, it is this relative strength of the "older", "left-from-here" objects that prompts nostalgic moves "back". The operation of the nostalgia dynamic draws upon this relative immutability, this relative reification of CA's analytic objects. 
It is not only the character of the objects that is different in the Next Time Through. The mode of epistemic apprehension changes too. For example, returning to the Tape does not involve listening to it to find out what there is to be heard in it. That work has already been done and its results have been enshrined as and in the Transcript, as the caption to the Tape's image. But once the nostalgia dynamic starts to operate, objects can come "under suspicion". Perhaps there is "something wrong" with the Transcript. If so, we can find out by checking the suspect item against an (older, stronger) item to its left. So whereas in the First Time Through, the Transcript acted (briefly) as the warrant for the Tape, in the Next Time Through it is the other way round. And, as we have mentioned above, the activity of checking item A against item B involves holding B constant. In order to hold the Tape firmly in place while the Transcript is checked against it, it needs to be heard. Hearing, you will recall, is an epistemic action that leaves its object as it found it. Hearing the Tape takes for granted the already-known reality of "what there is to be heard". 
6. Concluding Discussion
This "insider critique" of CA has elaborated some of the epistemic ethnomethods of the analytic practices found in the discipline, focussing on the mutual relations of tape and transcript. Our realisation of the significance of the range of positions on this relationship motivated our elaboration of an epistemo-phenomenological schema of evidential and analytic utility. By formulating the two distinct temporal processes we denote First Time and Next Time Through, we have been able to understand how CA's analytic objects are experienced differently relative to these "stages". First Time Through is engendered by the move to analytic utility, that is the requirement to produce worked up and work-able analytic objects. The second activity of Next Time Through is begot by the search for increased evidential utility: the need to "prove" those objects' adequacy, reliability and mutual fidelity. And it is our contention that the majority of activity in CA is best characterised as Next Time Through. 
While CA's rhetoric of method includes, at times, hints of the recognition of the Next Time Through dynamic, it is in large part focussed on an understanding characterised by innocence. That is, it views its various objects "naïvely": as though they were unchanged on each occasion of apprehension. The trajectory of innocence holds only one object—the Event—genuinely immutable by rendering it, through its recording, forever out of reach of the researcher and the reader. However, as we have argued, this is not the case for CA's other analytic objects. When returning to the (apparently unchanged) tape and "re-listening" to it, this activity, being Next Time Through, reflexively formulates it, not as a listenable object, but as a hearable one. Similarly when a transcript is "re-read", it is formulated as seeable. 
Let us add two complications to this picture. Firstly, the hearing and seeing of a relatively reified analytic object is itself transient; a brief moment of "proof". To question what is heard or seen is to immediately render the object once more as produced, crafted, worked-on-and-worked-up. The evidential adequacy of any object is momentary, contingent and fragile. 
Secondly, on each occasion in the Next Time Through where an evidentially "weaker" object is checked against a "stronger", the reification of the latter is balanced by an equivalent "de-reification" of the former. Evidential weakness and strength are characteristics mutually achieved for the particular practical purpose at hand. Also, which object plays which role in the activity of checking one against the other cannot always be "read off" from CA's self-understanding. For CA, the Tape stands as its strongest object and, as we have shown, the studied lack of interest in its production and unconcern with its origin is one of the things that contributes to this standing. But, as we have seen in the case of MODAFF and MODAFF's (2000) paper on the technicalities of audio-recording, when an interest in the Tape's production is evidenced, it is the Transcript that is treated as "stronger" than the Tape. 
The idea that objects are apprehended differently in First Time and Next Time Through has implications for a significant plank of CA's claim to superior scientificity. Harvey SACKS, the founder of CA, claimed that, in the sociology he was trying to develop, "the reader has as much information as the author and can reproduce the analysis ... I'm showing my materials and others can analyze them as well ..." (SACKS 1995, vol.1, p.27). MAYNARD strengthens and "scientizes" this claim by noting that: "In a sense, it is possible to obtain independent verification of interactional patterns because those who hear or read a researcher's report can themselves analyze the data" (MAYNARD 1989, pp.130-131). 
These positions are framed by an innocent, First Time Through perspective. The claim is that other researchers can also experience the data—whether understood as Tape or Transcript—as naïve observers. They can return to the starting point and simply repeat the process. But, in our understanding of this projected task, the reader is confronted by a set of different objects than were available to the innocent researcher in First Time Through. In an important sense, the reader has much more information than the author had on the equivalent occasion: they have the author's analysis which instructs the reader in what to hear on the tape and what to see in the transcript. They are simply not in a position to approach the task of reanalysis with the requisite innocence. 
Alternatively, the reader of a "researcher's report" always has much less information than the author. When a written piece of analysis is presented to a wider audience, it is fragments of transcript that evidence the author's analytic claims. Should the reader wish to "go further" than these texts allow, s/he will have to "go back", on our nostalgic trajectory, to a more complete set of materials—the (whole) Transcript, the Tape—which is always somewhere else. It is this problem which motivates analysts like CARROLL to advocate a digital solution: a transcript-free hypertext linking the Analysis directly to the Tape. It is our speculative view, however, that in this situation the Tape would cease successfully to play the role of "data": the place where the buck of "answerability" stops. At present the Tape's fulfilment of this role is dependent on its being "unavailable for questioning". Making it available for routine inspection in and as the course of reading CA would provoke the "invention" of a new, "stronger" analytic object (to its left in our schema) to which the Tape would relate, as, currently, does the Transcript to the Tape. In effect, the Tape would become a new form of Transcript. 
1) Our analysis is largely schematic and "conceptual", rather than "empirical". This may be taken to be a fault by those who argue, as conversation analysts tend to do, that it is only "in the presence of data" (ten HAVE 1997) that sensible and relevant analyses of such matters can be had. While we have some sympathy for this view in general, it is problematic for a project which, among other ambitions, wishes to address the question of what counts as data, when it so counts, and for whom does it so count. <back>
2) An exception is LOMAX and CASEY's (1998) non-CA reflexive analysis of the constructive and constitutive effects of video recording. They argue, and we agree, that "the activity of data collection is constitutive of the very interaction which is then subsequently available for investigation" (1998, abstract). This is not a view for which it is easy to find support in the CA literature. <back>
3) PSATHAS and ANDERSON concur: "It should be noted ... that the status of the transcript remains that of 'merely' being a representation of the actual interaction—i.e., it is not the interaction and it is not the 'data'" (1990, p.77). <back>
4) SCHUTZ (1972) and GARFINKEL (1967) use the term "scientific attitude" here. We prefer "mediated" as this term can account for all activities of formulation, understanding, representation, performance, whether done in the course of scientific work or not. <back>
5) Though a kind of left-hand world is sometimes said to have existed in the past and/or in simpler societies; a nostalgia for this "world we have lost" can be seen to have animated certain founding sociological contrasts such as TOENNIES' Gesellschaft and Gemeinschaft. Similarly, we are sometimes said to be entering a form of the right-hand world, such as BAUDRILLARD's postmodern dystopia of inauthentic simulacra. <back>
6) We have sketched an account of the "machinic-productive processes" of CA elsewhere (REED & ASHMORE 2000). <back>
7) We limit ourselves here to the objects in play in any single piece of CA research. Further analytic objects (including text books, reviews, critiques, bibliographies, web sites, email lists, courses, conferences, ...) constitute CA as a discipline, a culture and a literature. <back>
8) However, it may be used too for self-justification and encouragement: "If I didn't think that CA was about understanding reality, there would be no point in doing it" (paraphrase of personal communication: Charles ANTAKI to Malcolm ASHMORE, August 2000) <back>
9) Actually, though, there is a problem here. The technical aim of this comparative exercise is to recommend the use of doubled recordings of telephone conversations (MODAFF & MODAFF 2000, p.111). But an even closer comparative analysis of the lines we have just been looking at suggests that having two recordings may produce radical kinds of undecidability. Note that in fragment 1A (better for Jay), line 002, Jay continues his line 001 turn with "and". However, as we have seen, Dee's version (1B) has the first syllable of Dee's latched utterance "Ok[ay]" at this precise point. So what is being done here? Jay's "and"? Dee's "Ok"? Or, perhaps, both, overlapped? And how could this be decided? <back>
10) We are ignoring here the idea, increasingly mooted these digital days, of an Analysis being made directly from the Tape. We do, however, address this possibility below. <back>
11) The process can be likened to the building of an escape tunnel, as featured in Second World War movies like The Great Escape and The Wooden Horse. As the tunneller (the Listener, the Reader) digs the formless earth, behind him the tunnel is shaped, strengthened, shored up, given a definitive form. <back>
Anderson, Robert J. & Sharrock, Wesley W. (1984). Analytic work: Aspects of the organization of conversational data. Journal for the Theory of Social Behaviour, 14(1), 103-124.
Boden, Deirdre & Zimmerman, Don H. (Eds.) (1991). Talk and Social Structure: Studies in Ethnomethodology and Conversation Analysis. Cambridge, UK: Polity Press.
Bogen, David (1992). The organization of talk. Qualitative Sociology, 15, 273-296.
Carroll, Donald (2000). Re: [languse] transcribing digitized recordings. Private e-mail message sent to Malcolm Ashmore on Sat, 27 May 2000 21:36:09 +0900.
Cook, Guy (1990). Transcribing infinity: Problems of context presentation. Journal of Pragmatics, 14, 1-24.
Edwards, J.A. & Lampert, M.D. (Eds.) (1993). Talking Data: Transcription and Coding in Discourse Research. Hillsdale, New Jersey: Lawrence Erlbaum.
Garfinkel, Harold (1967). Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
Garfinkel, Harold; Lynch, Michael & Livingston, Eric (1981). The work of a discovering science construed with materials from the optically discovered pulsar. Philosophy of the Social Sciences, 11(2), 131-158.
Goodwin, Charles (1994). Recording human interaction in natural settings. Pragmatics, 3, 181-209.
Green, Judith; Franquiz, Maria & Dixon, Carol (1997). The myth of the objective transcript: Transcribing as a situated act. TESOL Quarterly, 31, 172-176.
Have, Paul ten (1990). Methodological issues in conversation analysis. Bulletin de Méthodologie Sociologique, 27 (June), 23-51 (original publication). Updated version available on the World Wide Web at www.pscw.uva.nl/emca/mica.htm (Broken link, FQS, June 2003).
Have, Paul ten (1997). In the presence of data: Conversation-analysis as "empirical philosophy". Paper read at the conference on "Ethnomethodology, an improbable sociology?", Cerisy-la-Salles, France, June 1997. Available on the World Wide Web at www.pscw.uva.nl/emca/presence.htm (Broken link, FQS, June 2003).
Have, Paul ten (1999). Doing Conversation Analysis: A Practical Guide. London: Sage.
Hopper, Robert (1989). Conversation analysis and social psychology as descriptions of interpersonal communication. In D. Roger & P. Bull (Eds.), Conversation (pp.48-65). Clevedon: Multilingual Matters.
Hutchby, Ian & Wooffitt, Robin (1998). Conversation Analysis: Principles, Practices and Applications. Oxford: Polity Press (UK and Europe), Blackwell (USA).
Jefferson, Gail (1985). An exercise in the transcription and analysis of laughter. In T. Van Dijk (Ed.), Handbook of Discourse Analysis, Vol. 3: Discourse and Dialogue (pp.25-34). London, UK: Academic Press.
Jefferson, Gail (1996). A case of transcriptional stereotyping. Journal of Pragmatics, 26, 159-170.
Latour, Bruno (1987). Science in Action. Milton Keynes: Open University Press (UK), Cambridge, Massachusetts: Harvard University Press (USA).
Lomax, Helen & Casey, Neil (1998). Recording Social Life: Reflexivity and Video Methodology. Sociological Research Online, 3(2). http://www.socresonline.org.uk/socresonline/3/2/1.html
Lynch, Michael (1993). Scientific Practice and Ordinary Action: Ethnomethodology and Social Studies of Science. Cambridge & New York: Cambridge University Press.
Lynch, Michael & Bogen, David (1994). Harvey Sacks' primitive natural science. Theory, Culture & Society, 11, 65-104.
Maynard, Douglas W. (1989). On the ethnography and analysis of discourse in institutional settings. In J.A. Holstein & G. Miller (Eds), Perspectives on Social Problems, Vol. 1 (pp.127-146). Greenwich, Connecticut: JAI Press.
McHoul, Alec W. (1987). An initial investigation of the usability of fictional conversation for doing conversational analysis. Semiotica, 67, 83-104.
Mishler, Elliot G. (1991). Representing discourse: The rhetoric of transcription. Journal of Narrative and Life History, 1(4), 255-280.
Modaff, John V. & Modaff, Daniel P. (2000). Technical notes on audio recording. Research on Language and Social Interaction, 33(1), 101-118.
Moerman, Michael (1988). Talking Culture: Ethnography and Conversation Analysis. Philadelphia, Pennsylvania: University of Pennsylvania Press.
Nelson, Christian K. (1994). Ethnomethodological positions on the use of ethnographic data in conversation analytic research. Journal of Contemporary Ethnography, 23, 307-329.
Ochs, Elinor (1979). Transcription as theory. In E. Ochs & B. Schieffelin (Eds.), Developmental Pragmatics. New York: Academic Press.
O'Connell, D.C. & Kowal, S. (1994). Some current transcription systems for spoken discourse: A critical analysis. Pragmatics, 4, 81-107.
Pack, Christopher (1986). Features of signs encountered in designing a notational system for transcribing lectures. In Harold Garfinkel (Ed.), Ethnomethodological Studies of Work (pp.92-122). London: Routledge and Kegan Paul.
Pomerantz, Anita & Fehr, B.J. (1997). Conversation analysis: An approach to the study of social action as sense making practices. In T.A. Van Dijk (Ed.), Discourse as Social Interaction. London: Sage.
Psathas, George (1995). Conversation Analysis: The Study of Talk-in-Interaction. Thousand Oaks, California: Sage (Qualitative Research Methods 35).
Reed, Darren & Ashmore, Malcolm (2000). The naturally-occurring chat machine. M/C: A Journal of Media and Culture 3(4). http://www.media-culture.org.au/0008/machine.html [Broken link, FQS, December 2004].
Sacks, Harvey (1984). Notes on methodology. In J. Maxwell Atkinson & John Heritage (Eds.), Structures of Social Action: Studies in Conversation Analysis (pp.21-27). Cambridge, UK: Cambridge University Press.
Sacks, Harvey (1995). Lectures on Conversation. 2 Vols. Edited by Gail Jefferson with introductions by Emanuel A. Schegloff. Oxford: Basil Blackwell.
Schegloff, Emmanuel A. (1979). Identification and recognition in telephone conversation openings. In George Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp.23-78). New York: Irvington Publishers.
Schutz, Alfred (1972). The Phenomenology of the Social World. London: Heinemann.
Silverman, David (1998). Harvey Sacks: Social Science and Conversation Analysis. Oxford: Polity Press.
Wilson, Thomas P. (1999). Re: [languse] transcribing digitised recordings. Posted to the "LANG-USE" (Language Use Discussion) e-mail list <LISTSERV@VM.TEMPLE.EDU> on Wed, 9 Jun 1999 16:50:41.
Malcolm ASHMORE is the author of The Reflexive Thesis (Chicago: The University of Chicago Press, 1989) and, with Michael MULKAY and Trevor PINCH, Health and Efficiency (Buckingham: Open University Press, 1989. His main research interest is in the social analysis of science and expertise. He is currently researching the recovered memory/false memory controversy.
Department of Social Sciences
Loughborough LE11 3TU
Darren REED is a postgraduate student in the Department of Social Sciences. His research is an investigation of sequential practices in Internet newsgroup messages.
Department of Social Sciences
Loughborough LE11 3TU
Ashmore, Malcolm & Reed, Darren (2000). Innocence and Nostalgia in Conversation Analysis: The Dynamic Relations of Tape and Transcript [45 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 1(3), Art. 3, http://nbn-resolving.de/urn:nbn:de:0114-fqs000335.