Charles Briggs

Learning how to ask

Cambridge: Cambridge University Press. 1986.

Thirty years later, it would be better, I believe to title the book something like "Possibilities and Constraints in Asking," leaving the issue of "learning" aside. One cannot learn how to ask but one should not be naive about the act of asking and the processes that organize the people that may get involved in the asking, not to mention the "things" (referents) that may trigger the asking.

p.2
We still know very little about the nature of the interview as a communicative event. Worse yet, because the interview is an accepted speech event in our own native speech communities, we take for granted that we know what it is and what it produces. One major problem is that the interview is most unusual, as communicative routines go...

This mystification of the interview emerges primarily in three ways. First, interviews provide examples of metacommunication, statements that report, describe, interpret, and evaluate communicative acts and processes. All speech communities possess repertoires of metacommunicative events that they use in generating shared understandings with respect to themselves and their experiences. As I argue in Chapter 4, these native metacommunicative events are rich in the pragmatic features that root speech events in a particular social situation and imbue them with force and meaning. Unfortunately, researchers seldom gain competence (in Hymes's [1974a:92-97] sense of the term) in these repertoires, relying instead on the metacommunicative routine that figures so prominently in their own speech community-the interview. This practice deprives the researcher of an adequate sense as to how the information she or he obtains fits into broader patterns of thinking, feeling, and speaking.

An even more serious problem is inherent in the structure of the interview. By participating in an interview, both parties are implicitly agreeing to abide by certain communicative norms. The interview moves the roles that each normally occupies in life into the background and structures the encounter with respect to the roles of interviewer and interviewee. Attention is concentrated on the topics introduced by the researcher's questions. Preliminary "small talk" may highlight the participants' present states of mind and body ("How are you?") and their relationship ("It's good to see you. I appreciate your letting me interview you again"). But the initial question then shifts the focus away from the interaction to another time, place, or process ("Now tell me about . . .").

The problem here is that this movement away from the interview as a speech event mystifies researchers to such an extent that they generally retain this focus in the course of their analysis. What is said is seen as a reflection of what is "out there" rather than as an interpretation which is jointly produced by interviewer and respondent. Since the context-sensitive features of such discourse are more clearly tied to the context of the interview than to that of the situation it describes, the researcher is likely to misinterpret the meaning of the responses.

A third difficulty arises because suppression of the norms that guide other types of communicative events is not always complete. Some potential respondents are drawn from communities whose sociolinguistic norms stand in opposition to those embedded in the interview. This is likely to be the case in groups that do not feature the interview as an established speech-event type. Lacking experience in this means of relating, such individuals are less likely to be able and willing to adhere to its rules. The farther we move away from home, culturally and linguistically, the greater the problem. This hiatus between the com- - municative norms of interviewer and researcher can greatly hinder research, and the problems it engenders have sometimes abruptly terminated the investigation. If the fieldworker does not take this gap into account, he or she will fail to see how native communicative patterns have shaped responses; this will lead the researcher to misconstrue their meaning.

It has not been possible to limit the discussion to a critique of interview methods alone, however, because broaching these methodological issues raises much broader questions. Why are interviews ubiquitous in the human sciences? Why is the nature of the interview process so poorly understood, and why has it not been more adequately researched? Why are we so reluctant to modify our research methodology, particularly in the light of theoretical advances? The answer is easy: Interview techniques smuggle outmoded preconceptions out of the realm of conscious theory and into that of methodology. Both our unquestioned faith in the interview and our reluctance to adopt a more sophisticated means of analyzing its findings emerge from the fact that the interview encapsulates our own native theories of communication and of reality.

p.23
A number of sources have appeared that provide us with important insight into the limitations of the interview as a research method. Their contribution lies in two key areas.

Reliability and validity. First, these works have clarified the thorny issues that surround the important problems of reliability and validity. These two concepts are discussed in most works on methodol ogy, because they provide the benchmarks by which data analysis and collection are measured. "Reliability" refers to the probability that the repetition of the same procedures, either by the same researcher or by another investigator, will produce the same results. "Validity" refers to the accuracy of a given technique, that is, the extent to which the results conform to the characteristics of the phenomena in question.
Hyman et al. (1954:20-1) pointed out that very few studies in the methodological literature were concerned with validity. Emphasis was rather placed on decreasing inter-interviewer variation, that is, in reducing the extent to which inter-interviewer differences affect the reliability of data. As Hyman et al. note, this presents a strong force for methodological conservatism. It also suggests that such efforts toward methodological reform may have no positive effect whatsoever on the degree to which procedures enhance our understanding of the questions they investigate.

The work of Hyman et al. and others has increased the degree to which questions of validity are taken seriously, although a bias toward studies of reliability is still apparent (Gorden 1969:6). The degree to which such work has affected interview practice is, however, far front clear. Cicourel provides us with insight into why this is the case. He argues that Hyman and his associates fell short of resolving the dilemma due to their failure to question the premise that underlies the reliability versus validity issue-the idea that procedures can be designed that will be both reliable and valid (Cicourel 1964:93).

Cicourel (1982a) raises the question of ecological validity. This concept pertains to the degree to which the circumstances created by the researcher's procedures match those of the everyday world of the subjects (cf. Neisser 1976). The problem is that competing demands are placed on the researcher. Standardization, a crucial device for promoting reliability, leads interviewers to attempt to present each question in exactly the same manner to each respondent. Bailey (1978:171) even suggests that "the interviewer's inflection and intonation should be the same for each respondent."

This raises two issues. First, true standardization would be achieved if the meaning of the question were the same for each respondent. This leads writers such as Gorden (1969:61) to suggest that differences in social backgrounds between respondents will force the interviewer to change the question wording in order to maintain the validity of quantitative measures. The problem runs deeper than this, however, because, as Cicourel (1974c:20) and Dexter (1970:144) note, interviewees respond not simply to the wording of the question but to the interview situation as a whole.

A second issue emerges from the fact that each interview is a unique social interaction that involves a negotiation of social roles and frames of reference between strangers. As Hyman et al. (1954:80-1) show, "bias" seldom appears as the interviewer's imposition of her or his own ideological slant on the respondent. "Bias," meaning interrespondent and inter-interviewer differences in the presentation of questions and the perception of responses, is rather an interactional resource that is used in accomplishing the task at hand.

As this task becomes more onerous-for example, if the schedule is long and complex, questions are repeated, or the material is socially sensitive-it becomes more difficult to obtain answers. The negotiation process becomes more critical at this stage, and the interviewer must draw upon a wider range of conversational resources in narrowing the gap between the standardized questions and the background knowledge and communicative norms of the interviewee. Interestingly, writers are divided on the issue as to whether or not this is a good thing. (Compare, for example, Bailey 1978 with Gorden 1969.) Cicourel's point is that such divergences emerge from the fact that reliability and validity are incompatible goals.

The concept of context and the nature of interpretation. This realization has led to a deeper understanding of the role of the context in interviews. As noted earlier in this chapter, many writers conceive of
"the context" as being the sum total of physical, social, and psychological stimuli that exist at the time of an interaction. This definition conveniently allows the analyst to decide what counts as "the context" on the basis of his or her own assessment of the situation. It also dichotomizes the analysis of "contextual variables" from that of the verbal components of the discourse. Given the fact that responses address the total situation, this dichotomization is methodologically problematic.

Brenner (1978, 1980, 1981a, 1981b), Cicourel (1974c, 1981, 1982a, 1982b, 1985), Dexter (1970), and Mehan (1979) have challenged this separation. They argue that the context is a phenomenological construct that is created jointly by the participants. Not only are contexts not simply situational givens, they are continually renegotiated in the course of the interaction. The words of the interviewer and interviewee do not simply occur within this frame; along with nonverbal components, they are the very stuff of which the context is constructed. Each utterance thus reflects this ongoing process, just as it contributes to it. As Cicourel (1974c:88) notes, "the actor's remarks in the interview, even when termed spontaneous, are often the product of a carefully monitored kind of presentation."

This immediately calls into question the naive concept of the interviewer - as the medium through which the respondents' attitudes and beliefs are conveyed to the reader. The interviewer rather stands as a co-participant in the construction of a discourse. This view also challenges received procedures for the interpretation of interview data. Researchers generally draw on their a priori, commonsense understandings of the meaning of the questions; these are then used as frames for the interpretation of responses. If questions and responses rather constitute small-scale models of interviewer-interviewee interaction, this mode of analysis will radically distort their significance.

p.50:
Reference
The procedural problems I have examined thus far relate primarily to the indexical or context-sensitive functions of speech. Both interviewer and respondent may share a common interpretation of the referential meaning of what is said and yet may differ widely on their interpretations of indexical meanings. This can create severe problems for researchers who have mastered the semantics and syntax of another language without gaining sufficient competence in subtle indexical functions. The consequences can be just as great, however, for researchers working in their own society who mistake a common syntax and lexicon for shared communicative competence.

I now wish to show, however, that reference is really not quite as simple and transparent as this discussion may have suggested. Even if the interviewee can identify the semantic content of the words contained in a given question, he or she may be unable to discern its referent(s). Similarly, the fact that a given reply follows a particular question does not necessarily mean that it constitutes an answer to that question. The interviewee may have misunderstood the referent provided by the question or may be purposely shifting the frame of reference. In either case, the danger is that the researcher will overlook this referential hiatus and thus misinterpret the response. The business of establishing and maintaining reference in interviews is accordingly no less in need of critical analysis.

Procedural problems of this type are sometimes corrected through the operation of such devices as "side sequences" (cf. Jefferson 1972):

Int.: What was your grandfather's name?

Resp.: Which one?

Int.: Your mother's father.

Resp.: His name was Jose Rafael Aragon.

The interviewer and/or respondent are not always aware, however, that the latter has failed to identify the referent. Such a response will prove misleading if it is interpreted as a reply to the question. If the researcher is repeatedly unable to establish a frame of reference that is intelligible to respondents, this may seriously undermine the interviewing process itself. Given the importance of the referential function to the interview, it is worthwhile to examine the business of establishing and maintaining reference in some detail.

Establishing a referential frame. Reference is both a creative and a powerful act, since it provides an intersubjective link between speaker and hearer. One or more entities, processes, imaginative con structions, and so on are selected by the speaker from an infinitude of referential possibilities and are re-created in the mind of the hearer. As Garfinkel (1967) has noted, vagueness is never entirely absent from an utterance, and this re-creation is always more or less approximate. Nevertheless, a useful answer is obtained only through the presentation of a question (of whatever form) that possesses sufficient specificity to permit the potential respondent to discern what the speaker is asking.

Identification of the proper referent is, however, only the first step. If I was asked, "Tell me about your mother," I would find it quite difficult to respond. Answering a question presupposes awareness not only of the referent but of the type and quantity of information about the referent that is being requested. This is particularly crucial in the interview setting, where a broad range of types of information and degrees of detail may be sought. An appropriate response to the preceding question might thus be "What do you mean? Who is she, what does she do for a living, how did we relate when I was a child, or what sort of person is she?" In other words, the question must not only provide a referent but situate that referent within its larger conversational context--if procedural problms are to be avoided.

p.119
Reflexivity
A second major obstacle to methodological progress is our implicit ontology, particularly our theory of social reality. Karp and Kend 982) have neatly characterized the degree to which our perspecti n fieldwork follows "from behaviorist, reductionist, and naturalist preemises, namely, that the object of anthropological inquiry is the stuff out there' and that that `stuff' has as two of its primary attributes stability and observability. " In other words, if social facts are,  as Durkheim (1938) would have it, like rocks, we need not concern ourselves with the effects of our actions (including the asking of questions) on our consultants' behavior, or with the fact that our perceptions are mediated by our own personal, cultural, and conceptual orientations. This conception neatly excuses the social scientist from the task of examining his or her own role in the process.

Articulating such claims has become highly unfashionable, and many practitioners have admitted the importance of reflexivity, or reflection on one's relation to the research situation (cf. Babcock 1980; Ru 1980, 1982). Theorists have similarly replaced the objectivity-versus-subjectivity question with an awareness of the fact that social scientific investigations entail intersubjectivity, the creation of a psychological link between two or more minds. The role of social-cultural, theoretical, and personal predilections in the perception, interpretation, a translation of data is now well known.

Karp and Kendall's (1982) discussion of anthropological research provides a case in point. They argue that fieldwork does not sim Involve learning to think like the natives. It is rather a means of learning to provide a rapprochement between native meanings and the requirements of anthropological discourse. The last two decades have witnessed the emergence of an anthropological genre in which ethnographers detail their actions in and reactions to the field situation (see. for example, Belmonte 1979; Berreman 1962; Crapanzano 1977, 1980. Dwyer 1982; Rabinow 1977, 1982; Ruby 1980, 1982). If discussion of one's relation to the fieldwork process is treated as a task that is independent of interpreting the data, however, the status quo of linguistic and ethnographic reporting will be preserved. The problem of translating awareness of the need for reflexivity into procedures for systematically analyzing the effect of the fieldworker's presence on the data has scarcely been discussed.

Our reliance on interviews and our uncritical approach to them is partly responsible for this lack of progress. When we begin an interview, our questions direct the attention of the participants away from the ongoing social situation. We collaborate in using the referential power of speech to transport researcher and interviewee(s) alike to another place and perhaps another time. Ordinarily, the researcher will have played no role in the events or processes under discussion. This facilitates our false consciousness of objectivity-the illusion that the object of scrutiny is "out there" and operates independently of our actions; we accordingly excuse ourselves from the need to consider our relation to it.

Nothing could be farther from the truth. This collectively created portrait of "the real world" is produced to satisfy the goals of an interaction in which we have a leading role. Given the fact that the researcher plays the dominant interactional role in interviews, her or his participation must be assessed in analyzing each datum that emerges from this setting. Decontextualizing what is said by focusing exclusively on the referential content of responses irrevocably separates the act of self-examination from the task of interpreting the meaning of what is said. As was the case with Humpty-Dumpty, no methodological sleight of hand can ever reunite them.