NOU (eh) wil ik- eh 't kan me niet schele,Utterances produced by informants in interview situations can be said to have a triple orientation. They are, firstly, descriptive or demonstrative of the life world of the speaker. They re-tell his or her experiences or re-express his or her viewpoints. That is their primary purpose. Secondly, they have a local relevance, within the interview situation, i.e. as answers to questions. And thirdly, they are used as input in a research project, as materials for analysis. As generally conceived, the actual interview situation is a strategic site, designed to have the informant produce statements which are both real or natural in relation to his or her life world, and useful or relevant in terms of the current research project. There is a large and varied literature which collects strategies, recipes and advice on how to conduct interviews. Writers who put the research project in first position, generally promote pre-designed, standardized interview schedules, while those who stress a natural expression of the life world favor open or semi-structured interviews. In both, the interview itself is seen strategically, as a designable course of events (either pre-designed or locally steered), serving transcending purposes. But, whatever the strategy chosen, on the ideological or the practical level, some tensions between the three worlds - life world, interview situation and analytic framework - seem to remain.
In the present paper, we will consider some examples from qualitative, open or semi-structured interviews. Although this type of interview is often chosen because it allows for a less imposing, more natural kind of information gathering, our examples can still be seen to display some of these tensions. By way of introduction, we will first discuss two contrastive types of interview organization, which differ in terms of how these tensions are managed, i.e. in terms of control. Secondly, we focus on negotiations concerning the level of detail at which some answers are to be given. Detailed answering seems to fit with life-worldly preoccupations, while answering in a summarizing fashion pre-serves the analysis that is to follow. Interviewers in open interviews seem to take an ambivalent stance in these negotiations, on the one hand calling for a free and natural telling, while on the other often displaying a preference for a summarized answer, that can be easily processed in terms of the research project. We will consider some examples in which we can see how participants in such interviews deal with these essential tensions of that particular interaction type. Finally, we will consider ways in which informants' answers are taken up by interviewers, especially in the form of repeats and formulations. We will discuss how such answer receipts function at the local level of post-answer negotiations, but also try to show the ways in which they anticipate the later professional analysis of the data, by moving away from the details of the lifeworld.
The perspective we take when we are discussing these examples is an ethnomethodological and especially a conversation analytic one. What we are trying to do is to analyze the ways in which interviewers and interviewees in situ negotiate what is talked about and how that is done. What we are hunting is the actual negotiation of local and extra-local rationalities, i.e. the life world story and the research interests. We treat these tensions or controversies with what could be called 'ethnographic respect', taking care to be 'symmetrical' in the ways in which we discuss the various stances involved. At the request of the editors of this volume, however, we have added a section in which we discuss some implications of our analyses for the theory and practice of interviewing and of interviews as data-gathering devices in the social science research process. This means that we have to leave our comfortable position of 'ethnomethodological indifference' (Garfinkel & Sacks, 1970) to join the debate among practitioners - reluctantly.
12 A: u begon al even te vertelle, e:h
you started already a bit to tell uh
13 (.)
14 A: wat °voor KLACHTe u had h?
what kind of complaints you had huh
15 (.)
16 B: ja:h,=
yes
17 A: =MAAR, (0,5) °dat we: hebbe 't nog niet
but that we did not discuss it
18 A: uitvoe:rig:e:h over gehad, ·hh maar:,
extensively but
19 A: kunt u me daar wat meer v('r) v::-ertelle,
can you tell me some more about that
20 A: want is dat ook waar u
because is that also for which you
21 A: die medicijne voor gehad he[bt?
had that medication
22 B: [ja:.=
yes
23 A: =ja.
yes
24 B: °inderdaad.=
indeed
25 A:
well I want uh I don't care
26 A: hoe dokters dat noeme,
how doctors call that
27 A: ·hh maar: e:h ik wil wel graag we:te,
but uh I would like to know
28 A: ·h nou: hoelang u dat nou hebt,=
well how long you have that uh
29 A: =e::h wanneer u dat voor 't eerst hebt gehad
when did you get that for the first time
30 A: ·hh en: (.) ho(e-)
and how-
31 A: wat voor klachte precies
what kind of complaints exactly
32 A:
in your own words
33 (0,9)
34 B: NOU: 't begon met-,
well it started withThe interviewer indicates that there already has been some exchange of information before she switched on her tape recorder (line 12, 14). She mentions that this previous discussion has not been very extensive (lines 17-8), making it clear that she would prefer to have such an extensive coverage now (line 19), referring to her research interest and her reason for selecting this informant (lines 20-1). Then, she formulates a lengthy introduction (lines 25-32). She not just explains her interests positively, in terms of what she would like to know, but she interrupts herself soon after she has started (25) to formulate what she is not interested in, i.e. medical jargon, how doctors call it (lines 25-6). She wants to know how long the informant has suffered from the complaint, when it started, and how he would describe the complaints, and adds: in your own words (lines 27-32). So the introduction of the topic invites the informant to tell his story at length and it is embedded in encouragements to do so from his own perspective, first negatively (25-6), then positively (32). Following that, the informant starts a long story, marking that he picks it up at the beginning (34).
This type of organization implies a loose kind of control by the interviewer on the informant's talking. During his DU the latter is largely self-directing, choosing his own relevancies, categories, and level of detail. The interviewer can try to control the style and direction of the informant's talk by including specific elements in the introductory DU (as in lines 28-31 above), by showing selective attention to what is being told(4), by follow-up questions or by starting a new DU, redirecting the topic.
The other type of organization we found in our corpus is characterized by a turn-by-turn allocation of speakership, so we use TBT-interviews as a shorthand for this type. These interviews mainly consist of an alternation of relative short speaking turns, in this case mainly questions, answers and acknowledgement tokens or similar objects. This TBT-format is, of course, the one many people expect in interviews, in research settings as well as elsewhere(5).
Extract (2) provides an illustration of such a TBT-format.
72 A: e:hm: (.)
73 A: heeft u misschien 'n: composthoop of zo:
do you perhaps have a compostheap or something like that?
74 (0,3)
75 B: neeh::.
no
76 (0,3)
77 A: n[ee.
no
78 B: [neh:
no
79 (0,9)
80 A: e:hm-
81 (0,8)
82 A: dan 't glas h
then the glass, you know.
83 (.)
84 A: ·hh wat doet u met uw glas:?
what do you do with your glass?
85 (.)
86 B: (nou) dat gaat naar de glasbak, (·hh:)=
(well) that's going to the bottlebank,
87 A: =jah, en: [waar is die?
yes, and where is that?
88 B: [jah,
yes
89 B: ·hh (.) staat daar op 't e:h hoekje,
standing over there on the u:h corner
90 (0,4)
91 A: oh bij dat pleint[je daar [zo
oh near to that little square over there
92 B: [ja, [bij dat pleintje,=
yes, near that little square
93 A: =ja,
yes
The questions in (2) are rather direct. They allow short answers, although some could be used as starting points for more elaborate descriptions. As it happens, the answers are rather short ones, after which the interviewer acknowledges their receipt in a way that either allows for a further elaboration, should the informant elect to do so (77-9), or initiates such an elaboration (87, 91). In other words, the extendibility of answers is continuously established in local negotiations.
Although the interviews we examined tend to fall mainly in either the DU- or the TBT-type, we do find TBT-episodes in DU-interviews and the other way around. For instance, an interview may start with a series of factual questions, after which DU's are used to deal with larger topics, explanations of complicated circumstances or stories, etc.
These negotiations and their structure have been described extensively elsewhere (Mazeland, 1989, 1992). It was found that both interviewers and informants display an orientation to the overall relevance as well as to the local fit of the answers to the questions that asked for them. In their negotiations, the interactants search for a formulation that is acceptable on these counts to both parties. The correctness, in the sense of veracity, of the answers cannot, of course, be established in the context of an interview. It seems that the answer's completeness functions as a kind of substitute for such considerations. The interviewers often tended to provide a maximum opportunity for extensions of the answer, leading to a succession of completions and post-completions.
It is in these environments of a post-answer trajectory that we found
instances of the phenomenon we want to focus on in this paper, one specific
type of post-answer negotiation, the one dealing with the level of detail
of the descriptions contained in the answer. It is that type of negotiation,
we think, that most clearly displays the essential tensions that were discussed
in the first paragraph of this paper.
The problems which surface in this extract follow from the interviewers questions in lines 163-5 of fragment (3). He asks for the location in Amsterdam to which the informant's family moved at a specific point in his life. In his reformulation of the question-word phrase, which he adds immediately to his question (which (...) neighborhood roughly, lines 163-5), he displays a preference for a specific kind of description: he wants to know the neighborhood to which his informant moved at the time (lines 161-3). With the addition roughly he makes it clear that this description does not have to be very exact: what he wants is a rough, for current practical purposes sufficient kind of information.
157 A: eve kijke je ging:- (0,4) tweenzestig
let's see you went- sixty two
158 A: dus met [vier jaar,
so at the age of 4
159 B: [otweeenzestig (j[ah)
sixty two (yes)
160 A: [verhuisden jullie naar Amsterdam:.=
you moved to Amsterdam:.
161 B: =°ja,
yes
162 (0,2)
163 A: waar kwame jullie toen te wone,
where did you start to live then
164 A:
which eh
165 A: buurt onge°ve[er
neighborhood roughly
166 B: [in de [Steen]-straat,=
in the [Stone]street,
167 A: =jah, ik weet nie waar dat °is:,=
yes, I don't know where that is
168 B: =°das de[: (i-)
that's the (i-)
169 A:
in sou-
170 B: [(in de) [Museum] buurt,
(in the) [Museum]-quarter
171 (0,2)
172 A: oh jah, (.) °zu[id.
oh yes, south
173 B: [dat is vlak [bij 't e:h
that's near to the u:h
174 A: [°jah
yes
175 (0,4)
176 B: [Maleisia]plein:,
[Malaysia]square,
177 (0,3)
178 A: jaA:,j[a.
yes, yes
179 B: [h (.) (en vl[ak dan)
you know? (and then near)
180 A: [jah,
181 (0,5)
182 B: en waar ik op school zit.
and where I am going to school
183 (.)
184 A: ja:,ja.
yes yes
185 (0,2)
186 A: ja:. °dan weet 'k 't wel.
yes. then I know it indeed
187 (0,3)
188 A: ·hhh en e:hm:- (0,4)
and u:hm:
189 A: nou: (d)van die tijd tuurlijk=
well from that period of course
190 A: =herinner je je waarschijnlijk °meer van
you probably remember more of that?Although the interviewer, in his instruction contained in the reformulation of his question, made it clear what kind of answer he preferred, his recipient does not answer with a description organized in terms of a category from the collection of neighborhood names, which the questioner had indicated. By using a street name, he chooses a member from a collection of categories which is more specific than the one for which the interviewer has just displayed a preference, i.e. Streets versus Neighborhoods(6).
The interviewer received the answer in a contradictory manner. At first he seems to let the answer pass with a Yes-receipt (line 167; Mazeland, 1990). One might say that he concedes that the informant has given an answer that is at least relevant. He immediately, however, adds that he is unable to process the information given (I don't know where that is, line 167). The analysis of the membership knowledge used (Schegloff, 1972), which the informant seems to make, i.e. that co-members of the category people living in Amsterdam would have more or less the same topographical knowledge, proves to be false in this case. The name of the street is not a recognitional for the interviewer (Sacks & Schegloff, 1979).
The informant treats this claim of ignorance on the interviewer's part as a correction invitation (Schegloff, Jefferson, Sacks, 1977: 379). He starts his next turn in a format which marks it as an explanation (that's the, line 168). The moment, however, this turn's type can be recognized, the interviewer interrupts him. But although the informant does not continue his explanation, the interviewer also breaks off his interruption shortly after its beginning (line 168). The interviewer's turn is just recognizable as a formulation of a city district (in sou-, line 169, i.e. in (district) south); a member of a type of description that is even more encompassing than the one earlier indicated in the addition to his question.
The fact that the interviewer breaks off his turn at the point at which this type of formulation is just recognizable is probably not accidental. By producing a word just up to its recognition point, a speaker can show what he was about to say without really saying it (Jefferson, 1973, 1983). In this case, one might think that the interviewer points out the kind of information he is seeking, without correcting his interaction partner openly.
Because the interviewer does not finish his turn, the informant gets another opportunity for self-correction. He uses that opportunity to produce a neighborhood term that should certainly be a recognitional for a co-resident of the city (in the Museum quarter, line 170). In this fashion, he tries to make a previously given description referentially adequate by embedding it in a more global one. But he also chooses a term which is compatible with the category instructions given by the interviewer.
The interviewer immediately displays his recognition (oh yes, line 172; cf. Heritage, 1984 b). But he also demonstrates this recognition by transforming the information given in a term from a collection of even broader locational categories (south, line 172).
But the sequence isn't finished yet. The informant keeps pressing for recognition on a level that is more specific than the one to which the interviewer displays to be oriented. On the one hand, he varies the types of collections that he selects. In line 182, for example, the informant chooses a relational description based on autobiographical information (and where I'm going to school). He seems to appeal to the interviewer to be treated at least in this respect as no longer a stranger. On the other hand, he tries to bring about recognition of more specific information by relating the apparently unknown to presumably known marks in the city landscape (that is near by the (...) Malaysia Square, line 173-6).
The interviewer, however, demonstrates several times that he has been informed to his satisfaction. One of the ways in which he does this, is by producing double yes's (lines 178, 184). In this way, a recipient claims to be familiar with the information that has been provided (cf. Komter, 1991). It is only when the has stated that he indeed knows it now (line 186), which explicitly revokes the assertion which led to the informant's repair work, that the sequence can be closed and the interviewer can start his preparations to initiate a new one (lines 188-91).
At the start of the sequence analyzed, one could still suppose that the informant has taken the categorical instructions, given by the interviewer, insufficiently into account. Alternatively, one might think that he has made an inadequate membership analysis of his partner. But later in the sequence, it is hard to avoid the conclusion that the speakers are interested in two entirely different identifications of the location under discussion. The descriptions provided by the informant seem to be designed to enable the researcher to make a locational analysis as part of a reconstruction of the specifics of the informant's living conditions at the time. The interviewer, however, seems just to be interested in an overall formulation of that location.
In the next sequence, it becomes clear retrospectively that, in terms of interview design, the interviewer wanted to use the information as the starting point for a subsequent question (well from that period of course you probably [don't] remember much, lines 190-1). The question about the location of residence was designed as the start of a pre-sequence (Schegloff, 1980), which prepared for a subsequent question which initiates the kernel-activity of remembering. The expansion of the pre-sequence postpones the planned subsequent activities. The discussion of the details of the informant's residence seem to have been, for the interviewer, just a preparatory activity, part of a strategy of what has been called getting through (Mehan, 1979: 111-4).
As concerns the interviewer's interests in the contents of the information provided, these might be limited to a formulation of the location of the informant's residence which would allow a certain way of professional generalization. The name of the neighborhood is the kind of information that makes possible an immediate sociological categorization in terms of socio-economic status, lifestyle, etc. Of course, the name of a street also allows such operations, but only after having been subsumed under a category of the type for which the interviewer has expressed a preference in his reformulation of his question. The informant is requested, so to speak, to deliver his information on that categorical level which makes possible an immediate sociological processing as a de-contextualized datum. That level, however, does not a allow a consideration of the details of the informant's life-world in itself and for itself, which seems to be the informant's motive. What we may have, then, is a complicated conflict between the informant's and the researcher's interests, both in terms of organizing the course of the interview and in those of the overall relevance of levels of information.
As a preparation for subsequent remembering of other details of the informant's life at the time, the evocation of the details of his place of residence makes sense. It's their recognition which may be considered superfluous. The researcher makes it clear that for him an overall locatability or sociological analyzability is enough, while for the informant a more detailed sharing of the information is desirable(7). In and through the way in which the interviewer moves away from the informant's detailing, and - reversely - the informant tries to raise the researcher's interest in the details which constitute his life-world, these interactional partners are showing the positions from which they are ultimately interacting, as members of the relationally paired categories (professional, lay person) (cf. Sacks, 1972 a).
This material demonstrates the existence of complex problems for the participants in interviews. There is the problem of choosing the right level of description. We see negotiations going on concerning detailed description versus overall categorization. The informant tries to bring the interviewer into his world, but the latter seems unwilling to do this to a larger extent than appears necessary for his purposes. All this demonstrates that for participants the interview, as it is enacted on the spot, is a practical compromise between the interests in the informant's life world and the researcher's conceptual world of sociology.
One of the reasons why qualitatively oriented social scientists use open interviews is that they do not want to impose a professionally motivated conceptual framework on their informants' descriptive activities. An interviewer, however, is continuously confronted with the strategic problem that s/he would, in a really natural conversation, have to immerse her/himself in the informant's life world, in all its detail. That seems an impossible ideal. In her/his daily practice, s/he compromises, by avoiding too much details.
In the extract discussed above, we can see that the interviewer operates very cautiously. His preference for more encompassing categories is expressed as in an afterthought to his question and - after he has claimed ignorance of the street's name - he produces a series of acknowledgments and recognition marks. His interest in the location as just a preliminary only becomes clear at the start of the next sequence, while the idea that his interest in the location itself was limited to it being material for a sociological reading of the informant's socio-cultural environment was our inference, which we could not support with specific data in the extract itself. In the next section, we will consider an other aspect of post-answer negotiations, interviewers' reception of answers, for which similar remarks concerning practical compromises can be made.
In our corpus, this tendency is also strongly observable, exemplified in extracts (2) and (3). Minimal responses, as frequently produced by our interviewers, seem mainly to serve local organizational purposes. They mark receipt of the preceding and function as a continuer (Schegloff, 1982)(8). Sometimes, however, a response is given that does some more elaborate interactional work, for example a display of understanding, in contrast to the claim of such understanding made by acknowledgements. Oh at that small square there (in (2), line 91), would be an example.
In our corpus we specifically found examples of two different types of more elaborate response, repeats and formulations. Repeating the answer just given is most frequently done in (phases of) interviews that are organized in a turn-by-turn fashion, especially when the questions project answers that offer short descriptions of facts; compare lines 10, 32, 39 and 58 in (4).
6 A: ·hhh hoe oud ben je
how old are you?
7 (0,3)
8 B: * zeventien
seventeen
9 (0,4)
10 A:
seventeen
(...)
27 A: wat voor school?
what kind of school?
28 (0,4)
29 B: e::h effe kijke ze zit op de:: (°nah)
u::h let's see, she's still going to (nah)
30 B: * lagere school zit ze nog, vijfde klas.
elementary school still, fifth grade.
31 (0,4)
32 A:
fifth grade elementary school
33 (.)
34 A: ·hh en je va:der,
and your father,
35 A: waar wer:kt die?
where's does he work?
36 (.)
37 B: * (th) [Rabo]bank,
(th) [RAbo]bank,
38 (.)
39 A:
(the) [Rabo]bank
(...)
54 A: heeft ie daarvoor gelee:rd
did he do a training for that?
55 (0,4)
56 B: * nee.
no
57 (0,6)
58 A:
no
When the informant packs his answers in longer Discourse Units repeating these answers is not a viable option. In such cases, interviewers tend to use formulations, which offer a summarizing interpretation of the locally relevant tenor of the just provided information (cf. Garfinkel & Sacks 1970; Heritage & Watson, 1979, 1980; Heritage 1985). In interviews organized on a turn-by-turn basis, formulations are also used, not so much to summarize the preceding answer, which is already rather short, but to provide an interpretation of it, that is, to help establish a shared sense of what the answer really means, as will become clear in our following discussion.
The relationships of formulations to the preceding answers can be quite complex, as becomes clear from the following list of aspects (cf. Heritage 1985: 101 a.f.):
Heritage (1985), in his analysis of news interviews, broadcast on radio and television, distinguished three types of formulation use. Similar types are found in research interviews, although their distribution is different from that in news interviews. The first type concerns relatively neutral and precise summaries of what has been said; Heritage (1985: 104) calls these prompts. In a second type, the formulation also presents a rather precise summary of the informants position, but in a way that is based on a more extensive inferential process. Heritage (1985: 106 etc.) calls these cooperative recycles: the tenor of the preceding utterances is expanded in a way that probably is agreeable to the informant; the recipient just assists the speaker in making his or her point utterly clear. The third type involves a more provocative use of formulations, when a far reaching upshot of the preceding is formulated that is bound to be denied by the informant. In this way the limits of his or her position are tested. Heritage (1985: 108 etc.) calls this type non-cooperative; the informant is not so much assisted as tested, forced by the interviewer to take a clear stand on issues or aspects selected by him or her.
These types differ not just in the content of the relations taken towards the preceding utterances, but also in their subsequent sequential effects. Formulations of the first and second type tend to engender positive decisions, agreements, which they clearly project (Heritage & Watson, 1980). But in so doing they also have a sequence terminating effect. When the parties agree on a formulation, no further talk on that topic is needed, so a next topic can be initiated, which is often done by the interviewer just after the agreement token. A non-cooperative formulation, on the other hand, very often leads to a negative decision, a disagreement, which will be explained and elaborated in more or less extensive fashion. So the third type generally has a sequence extending and topic elaborating effect(10).
Although they could function to stimulate the more extensive provision of information, non-cooperative formulations are rarely used by interviewers in research interviews, in contrast to, for instance, cross examinations in judicial settings (Atkinson & Drew, 1979; Drew, 1985) or critically toned news interviews (Heritage, 1985). As an interactional genre, research interviews seem to share the preference for agreement that is also generally found in ordinary conversations (Sacks, 1987; Pomerantz, 1978, 1984). This may be related to the fact that informants in research interviews do not have an intrinsic interest in the interview, so interviewers are under a constant pressure to keep them motivated to continue their participation, which an open disagreement might endanger. On the other hand, a research informant does not have to save a public face, as is the case with a politician being interviewed on television before an overhearing audience (Heritage, 1985). The previously mentioned distribution of types of formulations, then, fits with the tendency, earlier noted, for recipient reactions by interviewers to be rather neutral, displaying just a registration of the information provided.
Cooperative formulations and prompts, then, are well-designed objects to operate as a neutral representation of information provided by the informant. Below we summarize our analyses of formulations' properties which supports these functions:
In this way, interviewers' use of formulations not only contributes to the neutralizing, registering properties of the interview as a specific kind of locally organized speech exchange system, it also provides a locally produced anticipation or preview of what is to follow, a generalizing processing of information in which the particulars discussed serve as exemplars of types, processes, etc.
Consider the next extract (5), in which the interviewer formulates the reasons his informant might have for joining a project of separate garbage collection as an environment motive (line 878).
867 A: e:hm:- (°ng) (0,4) °wat-
u:hm:- (ng) what-
868 A: wat zijn de redenen om 'r aan mee te doen.
what are the reasons to participate
869 (.)
870 A: °voor u,
for you
871 (1,3)
872 B: nou jah:, (w)wille allemaal ons steentje
well yes, we all want
873 B: wel bijdrage, natuurlijk h=
to do our share, of course, isn't it?
874 B: =(°anders) 't gaat verkee:rd,
otherwise things go wrong,
875 (0,5)
876 A: °jah.
yes
877 (0,5)
878 A:
that are those environment things
879 B: [.hh [ja:h, °dh[at e:h
yes, that u:h
880 A: [°jah,
yes,
881 (0,3)
882 B: dat spreekt me [wel aan: ja,
that appeals to me yes,
883 A: [°jah,
yesIn one of the research reports from the project of which this interview was a part, a quotation is given that is partially composed from the utterances in lines 872-4. The quote serves to illustrate the position of that part of the population that is called the environment conscious ones.
The category that was used in the interview situation to formulate the information provided by the informant, is the same as the category used later in the analysis to typify the position of its speaker. The interviewer, then, is in his formulation already doing some processing work anticipating the later analysis. During the interview itself, the life world information is already cut in pieces that fit the conceptual scheme that will be used later, when the interview is over.
These features involve, on the one hand, a tension between the frames of relevance of the life world and the research project, in terms of, among other things, the different importance of details and the use of concepts, and, on the other, the practical need for the parties to manage their local interactions in a viable way. More research is needed to describe the strategies and mechanisms that are involved in this management. Our research has focussed on some exemplary aspects, i.e. the negotiations concerning detailing and the use of repeats and formulations.
These analyses will have to be deepened and extended to cover other aspects as well(11). They will have to deal in any case with what seems to be a basic constraint in the open research interview as such, the preference for agreement (Sacks, 1987). The open interview format shares this preference with ordinary conversation, in relative contrast to some other kinds of speech exchange systems like interrogations or debates. It seems that interviewers have such a feature in mind, when they argue for interviews to be like conversations. In other ways, however, the research interview is markedly different from conversations, as the previous analyses have shown.
It may well be the case that the combination in one format of features which are conversational with some that are not, contributes to the awkwardness many semi-structured interviews seem to have. Participants are under pressure to cover over their difference of perspective, as when informants shape their expressions as answers which are at least in formal agreement with the suggestions contained in preceding questions, while the interviewers proffer displays of at least formal acceptance of those expressions.
The semi-structured interview, then, may be in and for itself a problematic convergence of interactional features, (Cf. Jefferson & Lee, 1981; Ten Have, 1989). It is designed as an asymmetrical information-gathering encounter, in the sense that one party questions, while the other answers. But in the ways in which the interviewer - in contrast to what his or her survey colleague is supposed to do(12) - adapts his or her questioning to the recipient and the flow of the interaction, and in accepting the answers gives a demonstration of understanding, a more symmetrical format, designed for the sharing of life-worldly informations and interests, is suggested. In their situated interactions with one another, interviewers and informants devise more or less elaborate and noticeable ways of dealing with these tensions which we think are essential to the format.
Informants are in the business of remembering, expressing and narrating their own personal experiences, while researchers are interested in using those expressions and narratives as data to be summarized, categorized, in short analyzed in terms of a research question which should make sense in some scientific debate. In short, informants and researchers tend to have different agendas. The extent to which this difference is a problem for the researcher will depend on the actual agenda he or she is pursuing. Consequently, researchers use rather different strategies to deal with the essentials tensions of the interview format.
Interviews may be analyzed with diverse kinds of questions in mind. As our paper demonstrates, interviews can even be analyzed from a perspective that is more or less completely at odds with the 'original' purposes of the interview. For our particular purposes, the tensions were not a problem at all, but an interesting phenomenon(13). While ordinarily only informants are considered as research subjects, for us both interviewers and respondents were cast in that role. Since most interviews are not done in order to study interviews, this may seem to be a marginal case. We do recommend, however, the analytic mentality displayed in it.
A more popular strategy of avoiding the tensions, however, widely recommended in the literature, is to try to minimize the interactional demonstrations of the analytic interests and personal evaluations by the interviewer. For instance, in interviews in which the research is focussed on the actual telling of a story, the interview may be designed to leave out any conceptualizing suggestions by using a vague opening statement setting the theme in everyday terms. Furthermore, the interviewer may be asked to restrict his or her questioning and reactions to the emic(14), i.e. non-analytic, terminology previously used by the informant. From a conversation analytic perspective, one would suggest that such restrictions, being 'unnatural' for the interviewer may be difficult to apply in actual practice. Furthermore, the informant may not expect the interviewer to speak from any other position than his or her own, i.e. that of an interested outsider. An informant may even make an effort to speak the interviewer's 'language', as he/she sees it, rather than his or her own, and expect the interviewer to display his or her understanding in that same 'language', not just repeating the informant's words.
The literature on interviewing especially urges interviewers to act in a 'neutral' fashion by refraining from evaluative comments and restricting reactions to minimal responses. Acting on this advice, however, may work to elicit more elaborate answers and more strongly dramatized stories, than would have been produced if the interviewer had acted otherwise (c.f. Rombouts, 1984). Even the type and intonation of minimal responses will be taken as indications of interviewer's interest and evaluation, and will, for that reason, influence the way answers are elaborated and stories told (c.f. Mazeland, 1990, 1992). Interview statements are interactional products, inevitably.
Another way in which the tensions can be managed to a certain extent, is to phase the interview in terms of varying interviewer control of the conversation. We observed, for example, that it is common practice to start interviews with a series of factual questions. This leads to what we have called a Turn-by-Turn format. We suggested that answer repeats in such an environment serve to fix the facts as interactionally established. Such a format, then, presupposes a factual world which can be unproblematically become known to a non-participant.
When such a phase is finished, the alternative DU-format can be initiated explicitly, explaining its intent as was done in extract (1). In such a format, a difference of 'world' and even categories for understanding worlds, is often recognized. Here, the interviewer's interactional influence tends to be more subtle. Researchers can opt for a restriction of interviewer pre- and post-structuring by avoiding etic terminology in questions and answer-receipts, in order to get the story on tape in the most natural way possible. In that case, it is only after the interview, that they try to make sense of what was said in terms of their research questions. Alternatively, Researchers can invite explicit negotiations about the meanings of previous statements, within the interview context itself. That would involve the interviewer in an explanation of his or her interests and basic concepts, and engage the informant as a sort of research assistant. Again, a phase differentiation is possible. One can first ask the informant to give an extended answer or tell a story, without bothering the process of remembering and telling with analytic clarification, and then start a discussion of its meaning in emic or even etic terms. This way of working would overtly recognize the interview as a situated, collaboratively achieved re-constitution of the part of the informant's world in which the interviewer is interested.
Research on structured survey-interviews has demonstrated that many respondents give an 'instance' when a 'category' is asked for. In many cases, it is only after the interviewer has provided some exemplary 'candidate answers', that the respondents are able to give an answer at all, often in line with, or in contrast to, the example given (c.f. Houtkoop, 1990, 1995). A similar phenomenon was found in our research on semi-structured interviews (c.f. our discussion of extract (3) above, note 7, and Mazeland, 1992). When the interviewer does not have an extensive pre-knowledge of the topic under discussion, it is probably inevitable that questions are formatted as categorical ones. This observation should not be taken, however, as suggesting that informants do not use categories themselves. In fact, a case could be made that the interesting phenomenon is how emic categories are used as part of native practice, turning the interview into an occasion to have such practices demonstrated on tape. The problem, then, is how to get a faithful demonstration of native usage, rather than a translation exercise.
A number of writers have stressed the moral character of interviews (Locker, 1981; Baruch, 1981; Silverman, 1985). This relates specifically to the fact that an interview inevitably is a meeting of two strangers, a confrontation of two worlds. The interviewee, who is invited to speak, is under pressure to not just express his or her life experiences and viewpoints, but also to account for these, to demonstrate their rationality. Locker (1981) stresses, for instance, that the mothers he interviewed were strongly motivated to convince him that their decisions regarding the health and illness of their family members were reasonable, given the circumstances. And Baruch (1981) reports a marked difference in first and second interviews with parents of handicapped children. Moral accounting and complaining about medical personnel was much more frequent in the first interview, as a conversation with a stranger, than during the second one, when the researcher was already an acquaintance of sorts. Interviews, then, offer useful materials for the study of 'accounting practices'.
The crux of our argument regards the analysis of interview 'data'. We think that interview statements should always and explicitly be seen as actions-in-contexts. This fundamental fact should be basic to the analysis. The statements can be seen as part of an overall stance that the interviewee is trying to bring across. But one should also take the details of the local interactional context - the 'sequential environment' - into account. It is in response to that environment, which is continuously rebuilt during the interview, that interview statements have been designed and should be analyzed.
As a practical consequence, we think that for an interview to be used as more than a source of overall, easily summarized factual knowledge, it should be taped and analyzed in a detailed fashion. Statements from interviews should always be quoted 'in sequence', with both preceding questions and following elaborations and receipts - the full natural unit. Furthermore, knowledge of sequential structures and membership categorization devices - in short Conversation Analysis - should be applied to interview materials of various kinds in order to yield generalized procedural insight into interview formats and to stimulate attention to interactional details among interview researchers.
Interviews are used to constitute sociological objects, such as 'lay medical systems', 'youth cultures' or 'environmental consciousness'. A reflexive sociology should investigate how this is done. We recommend the analytic mentality displayed in this paper as a starting point for such an enterprise.
2. The present paper was written by Paul ten Have, based on the analyses made by Harrie Mazeland as part of a conversation-analytic investigation of social-scientific interviews supported by the "Netherlands organization for scientific research" (NWO; projectnr. 500-278-005 Stichting Sociaal-Culturele Wetenschappen, S.S.C.W.; Cf. Mazeland, 1989, 1992).
3. We collected one or more recordings of the semi- or unstructured interviews made by researchers as part of 14 different research projects. Detailed analyses have been based on the opening 15 - 20 minutes of 13 different interviews, which were transcribed using a variant version of the conventions established by Gail Jefferson (cf. Atkinson & Heritage, 1984: IX - XVI).
4. Cf. Ten Have (1991 a) for the use of silence by physicians in similar circumstances.
5. I.e. examinations in court, Atkinson & Drew, 1979; oral examination in ambulatory care encounters, Frankel, 1984; news interviews, Heritage, 1985, Greatbatch, 1988; etc.
6. This analysis is inspired, of course, by Sacks' work on the organization of Membership Categorization Devices (1972 a & b; 1992). See also Schegloff's (1972) related analysis of locational formulations.
7. Such an orientation on a non-liveworldly level of description on the part of interviewers can also be seen in the distribution of question terms like wat voor (what sort of) and welke (which); examples from the first category are twice as frequently used in our corpus as those from the second (see, for example extract 1, line 13 and 37, extract 4, line 27).
8. The ways in which these items function in our corpus have been analyzed elsewhere in a more extensive fashion (c.f. Mazeland, 1989, 1992). A separate paper (Mazeland, 1990) presents a parallel analysis based on a small corpus of medical consultations.
9. It may be the case, in this fragment as well as on other occasions where repeats are produced, that the interviewer writes down the answers he repeats. This just adds to the fact that, in so doing, he constitutes himself on those moments as being just a (neutral) recording device.
10. This is a general property of dispreferred alternatives: they are to be accounted for, which involves further interactional work. Cf. Sacks (1987); also: Levinson (1983: 334-5), Heritage (1984 a: 265-80) and Heritage (1988).
11. This discussion is based in part on ongoing research by the authors.
12. But see Houtkoop-Steenstra's (1990, 1995) studies of survey interviews, demonstrating the many ways in which such interviewers actually adapt their phrasings and receipts to the interviewee.
13. C.f. Silverman (1985: 156-76) on different uses of interview materials, depending on different theoretical and methodological frameworks.
14. The concept of emic stems from anthropology
and refers to native terminologies and categorizations, in contrast to
etic which refers to the cognitive and linguistic apparatus which
an outsider, like a researcher, tends to use when considering a native
culture.
Atkinson, J.M., P. Drew (1979) Order in Court: The Organisation of Verbal Interaction in Judicial Settings. London: Macmillan
Atkinson, J.M., J. Heritage, eds. (1984) Structures of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press
Baruch, G. (1981) 'Moral tales: parents' stories of encounters with the health professions', Sociology of health and illness 3: 275-95
Drew, P., (1985) 'Analyzing the use of language in courtroom interaction'. In: T.A. van Dijk, ed. Handbook of discourse analysis. 4 vols. London: Academic Press: vol. III: 133-47
Frankel, R.M. (1984) 'From sentence to sequence: Understanding the medical encounter through micro-interactional analysis', Discourse Processes 7: 135-70
Garfinkel, H., H. Sacks (1970) 'On formal structures of practical action'. In: J.C. McKinney & E.A. Tiryakian, eds. Theoretical sociology: perspectives and developments. New York: Appleton-Century-Crofts: 338-66
Greatbatch, D. (1988) 'A turn-taking system for British news interviews', Language in Society 17: 401-30
Have, P. ten (1989) 'The consultation as a genre'. In: B. Torode, ed. Text and Talk as Social Practice. Dordrecht/Providence, R.I.: Foris Publications: 115-35
Have, P. ten (1991a) 'The doctor is silent: observations on episodes without vocal receipt during medical consultations'. In: B. Conein, M. de Fornel, L. Quéré, Les formes de la conversation, vol 2. Issy-les-Moulineaux: CNET, 1991: 55-76
Have, P. ten (1991b) 'Talk and institution: a reconsideration of the 'asymmetry' of doctor-patient interaction'. In: D. Boden & D.H. Zimmerman, eds. Talk and social structure: studies in ethnomethodology and conversation analysis. Cambridge: Polity Press: 138-63
Heritage, J. (1984a) Garfinkel and Ethnomethodology. Cambridge: Polity Press
Heritage, J. (1984b) 'A change-of-state token and aspects of its sequential placement'. In: J.M. Atkinson & J. Heritage, eds.: 299-345
Heritage, J. (1985) 'Analyzing news interviews: aspects of the production of talk for an overhearing audience'. In: T.A. van Dijk, eds. Handbook of discourse analysis. London: Academic Press: vol. III: 95-117
Heritage, J. (1988) 'Explanations as accounts: a conversation analytic perspective'. In: C. Antaki, ed. Analyzing everyday explanation: a casebook of methods. London: Sage: 127-44
Heritage, J.C., D.R. Watson (1979) 'Formulations as conversational objects'. In: G. Psathas, ed. Everyday language: studies in ethnomethodology. New York: Irvington: 123-62
Heritage, J.C., D.R. Watson (1980) 'Aspects of the properties of formulations in natural conversations: Some instances analysed', Semiotica 30: 245-62
Houtkoop, H., H. Mazeland (1985) 'Turns and discourse units in everyday conversation', Journal of Pragmatics 9: 595-619
Houtkoop-Steenstra, H. (1995) 'Meeting both ends: standardization and recipient design in telephone survey interviews'. In: P. ten Have, & G. Psathas, eds., Situated order: Studies in the social organization of talk and embodied action. Washington, D.C.: University Press of America: 91-106
Houtkoop-Steenstra, J.P. (1990) Normative and analytical perspectives on interviewing techniques. In: H. Pinkster & I. Genee (eds.) Unity in Diversity. Dordrecht, Foris Publications: 133-151
Jefferson, G. (1973) 'A case of precision timing in ordinary conversation: overlapped tag-positioned address terms in closing sequences', Semiotica 9: 47-96
Jefferson, G. (1983) 'On exposed and embedded correction in conversation'. Studium Linguistik, 14: 58-68
Jefferson, G., J.R.E. Lee (1981) 'The rejection of advice: managing the problematic convergence of a "Troubles-Telling" and a "Service Encounter"', Journal of Pragmatics 5: 399-422
Komter, M.L. (1991) Conflict and cooperation in job interviews; a study of talk, tasks and ideas. Amsterdam/Philadelphia: John Benjamins
Levinson, S.C. (1983) Pragmatics. Cambridge: Cambridge University Press
Locker, D. (1981) Symptoms and illness: the cognitive organization of disorder. London: Tavistock
Mazeland, H. (1989) "Interviews als gesprekken: gespreksanalytisch onderzoek naar sociaal-wetenschappelijke interviews" [Interviews as conversations: conversation analytic research into social scientific interviews]. Eindrapport SSCW-onderzoeksproject [End of grant report, SSCW research project] 500-278-005. Sociologisch Instituut, Universiteit van Amsterdam
Mazeland, H. (1990) '"Yes", "no" and "mhm": variations in acknowledgment choices'. In: B. Conein, M. de Fornel, L. Quéré, Les fornmes de la conversation, vol 1. Issy-les-Moulineaux: CNET: 251-82
Mazeland, H. (1992) Vraag/antwoord-sequenties.[Question-answer sequences]. Amsterdam: Stichting Neerlandistiek VU (also: PhD Dissertation, University of Groningen)
Mehan, H. (1979) Learning lessons: social organization in the classroom. Cambridge, Mass.: Harvard University Press
Pomerantz, A. (1978) 'Compliment responses: notes on the co-operation of multiple constraints'. In: J.N., Schenkein, ed., Studies in the organization of conversational interaction. New York: Academic Press: 79-112
Pomerantz, A. (1984) 'Agreeing and disagreeing with assessments: some features of preferred/dispreferred turn shapes'. In: Atkinson & Heritage, eds.: 57-101
Rombouts, R. (1984) 'Verhalen vertellen over etnische minderheden: hoe bevooroordeeld zijn we eigenlijk?' [Telling stories about ethnic minorities: How prejudiced are we really?], Tijdschrift voor Taal- en Tekstwetenschap 4: 133-46
Sacks, H. (1972a) 'An initial investigation of the usability of conversational data for doing sociology'. In: D. Sudnow, ed. Studies in social interaction. New York: Free Press: 31-74
Sacks, H. (1972b) 'On the analyzability of stories by children'. In: J.J. Gumperz, & D. Hymes, ed. Directions in sociolinguistics: the ethnography of communication. New York: Rinehart & Winston: 325-45
Sacks, H. (1987) 'On the preferences for agreement and contiguity in sequences in conversation'. In: G. Button, J.R.E. Lee, eds. Talk and social organisation. Clevedon: Multilingual Matters: 54-69
Sacks, H. (1992) Lectures on conversation. 2 vols. Edited by Gail Jefferson with introductions by Emanuel A. Schegloff. Oxford: Basil Blackwell
Sacks, H., E.A. Schegloff, G. Jefferson (1978) 'A simplest systematics for the organization of turn taking for conversation'. In: J.N., Schenkein, ed., Studies in the organization of conversational interaction. New York: Academic Press: 7-55 (1974)
Sacks, H., E.A. Schegloff (1979) 'Two preferences in the organization of reference to persons in conversation and their interaction'. In: G. Psathas, ed. Everyday language: studies in ethnomethodology. New York: Irvington: 15-21
Schegloff, E.A. (1972) 'Notes on a conversational practice: formulating place'. In: D. Sudnow, ed. Studies in social interaction. New York: Free Press, 1972: 75-119
Schegloff, E.A. (1980) 'Preliminaries to preliminaries: "Can I ask you a question?"', Sociological Inquiry 50: 104-52
Schegloff, E.A. (1982) 'Discourse as an interactional achievement: some uses of "uh huh" and other things that come between sentences'. In: D. Tannen (ed.), Analyzing discourse: text and talk. Washington, D.C.: Georgetown University Press: 71-93
Schegloff, E.A., G. Jefferson, H. Sacks (1977) 'The preference for self-correction in the organization of repair in conversation', Language 53: 361-82
Silverman, D. (1985) Qualitative methodology and sociology: describing the social world. Aldershot: Gower
Suchman, L., B. Jordan (1990) 'Interactional
troubles in face-to-face survey interviews', Journal of the American
Statistical Association, 85: 232-41