Rutgers University School of Communication & Information Series; Zoom meeting on December 16, 2021
Introduction and transcription by Enhua Guo, Ocean University of China, and edited by Anita Pomerantz, University at Albany
The Rutgers University Conversation Analysis Lab has launched a series of online discussions with key figures in CA. On December 16, 2021, they interviewed Anita Pomerantz, Professor Emerita in the Department of Communication at the State University of New York at Albany. Anita’s talk, which is insightful and interesting as always, consists of two parts: (1) biography/history of her CA journey, and (2) art and science of CA. Below is only a transcription of the second part.
I was very honored to have joined this ZOOM-based interview and have transcribed the second part of it. I want to extend my sincerest gratitude to Anita for editing the transcript thoroughly. She has made detailed changes and has rendered the transcript more reader-friendly, systematic, and articulate. After her editing, the genre of the transcript resembles a cross between a talk and short paper. For those who unfortunately missed the interview and have no time to go over the recording, my suggestion is to just read through this transcript.
I would also like to thank the Rutgers CA Lab, especially Alexa Hepburn, for organizing the interview and allowing us to have it published here. We also want to thank the audience for raising such fascinating questions.
Question by Alexa Hepburn
In your discussion of the Asking and Telling collection for DARG, you noted in passing that when Sacks studied Yes/No, or polar, questions, he didn’t call them that – he named them in terms of their function rather than their grammatical structure (for example he called them a ‘correction invitation device’). I wonder if you’d like to say more about the naming of phenomena in CA.
Response by Anita Pomerantz
When we try to gain an understanding of the machinery of interaction, the apparatus, the resources, the practices, the strategies, or the devices, we want to know (1) how it works in interaction, and (2) how we can identify the discourse or conduct that is relevant to its operation. In starting to analyze a practice, we identify relevant interactional discourse or conduct, and we propose how it works or functions in interaction.
Let me provide two examples of names of practices that privilege the work/function over the discursive objects that are used to mobilize the work or function.
Example 1: correction-invitation
Sacks’ name, a correction-invitation device, identifies how the device works. On face value, yes-no questions ask the recipient to either confirm or disconfirm the candidate offered in the question. However, Sacks understood that such questions invite the recipient not to confirm or disconfirm, but to confirm or offer a correction of the candidate answer in the question. He named the device, “correction invitation,” in that the question invites a correction. Although he did not explicitly identify the types of questions (the discourse) that invite confirmation or correction, he gave various examples, and we can infer that they are yes-no questions.
Example 2: pre-closing
Schegloff’s research identifies how closings are accomplished. One way that participants start to move to closing an interaction is to perform what he called a pre-closing. The name “pre-closing” tells us how it functions. A pre-closing signals that a participant intends to move toward closing the interaction, yet it allows for co-participants to engage in further topical talk should they so choose. The name, pre-closing, privileges the working or function being performed. Schegloff identified common discursive forms, including “Well…”, “Okay…”, and “So…” yet he specified that the form, alone, does not do the work. It needs to be delivered with a specific identifiable intonation at certain moments in the course of sequences of actions. The same lexical items uttered in different sequential positions and/or with different intonation do different work. Schegloff’s analysis provides an excellent description of both discourse features (the common discursive forms and their placement in sequences) and their interactional function.
I have indicated in the past that I favor naming practices for the interactional work or function that is being carried out rather than for the lexical items and/or conduct mobilized to carry out the work. I have taken that view based on both my own research experience and the current prevalence of using lexical items or bits of conduct as names of practices.
Let me tell you why I think the name of a practice matters. What you name your practice seems to direct your search for additional instances of the phenomenon of interest. That’s important. If you think your phenomenon is “Well…” or “So…”, there is a good chance that you will collect them. However, we know that lexical items function differently in different sequential environments. When we analyze what we have collected and find that some of them work differently than what we initially had proposed, we end up writing papers with different sections dealing with the different uses. That kind of paper may be informative, but it’s a different direction than if you direct your search by saying “I’m interested in pre-closes,” or “I’m interested in inviting somebody to give information that I don’t want to directly ask for.” When you think of the phenomenon centrally as interactional work or function, you would say “What are the ways of doing it?” and “What are the alternatives for doing that kind of thing?”
The reason I like focusing on the interactional work or function and asking how it is accomplished rather than on a discursive form and asking what it does is because I think it is truer to what goes on in interaction, to the process of interacting. Of course, participants use lexical items, hand movements, eye gaze, etc. However when we interact, we perform actions, we respond to prior actions and to other aspects of the context, and we anticipate upcoming actions. We do things with one another when we interact. I believe those ‘doings’ are best studied by examining collections of some particular kind of work done in interaction. With that kind of collection, we would analyze the ways in which the work is being accomplished.
I would like to address a very different issue about naming practices. The issue came about in relation to a paper I published in 1980, “Telling my side: “Limited access” as a fishing device.” As I conceived it, the work or function being accomplished in and by the practice is ‘fishing’ for information, that is, indirectly seeking information. The two parts to the title identify two formulations of the discourse used to indirectly seek information: a ‘my side’ report versus a ‘limited access’ report. When I named the practice a “my side telling,” I picked that name because it suggests that there would be a ‘your side telling.’ The name suggests that the speaker expects the recipient to have a different side and possibly be motivated to report it.It turned out that the formulation of the discourse as ‘my side telling’ applies to more activities than just information-seeking. I came to realize that a better formulation for the discourse would be “offering a limited access report of the recipient’s activity”. The reason that I like the latter formulation better, even though it’s awkward, is because it better describes how the practice works. When I tell you a version of your own activity, a version that is recognizable as based on ‘limited access’ to your activity, that serves as an invitation for you to tell me a more complete version of that activity. That formulation provides a better understanding of the type of information-seeking I was investigating.
I am raising this issue here because we always have a choice of how we name or formulate conduct, actions, and practices. Sometimes we choose more general formulations and other time formulations that zoom in on the specific phenomenon. If you choose names that are more general, you are going to collect a lot of instances that you may not intend to collect. It is worth considering how we name, formulate, and describe phenomena of interest.
Question by Jonathan Potter
One of the things you are most well-known for is your work on assessments. While assessments are not claims about knowledge, there are plainly epistemic aspects to them. How does you see your work on assessments as relating to epistemics?
Response by Anita Pomerantz
When I did the research on assessment sequences, my thinking was fuzzy as to what I meant by “assessments.” The fuzziness had to do with sometimes thinking about them as a discursive form and other times thinking about them as an action, as in reporting or offering an assessment. Let’s consider an assessment such as “Oh, that’s beautiful!” On the one hand, I collected such assessments on the basis of the form: it has an evaluative term and a reference that indicated the apparent referent. On the other hand, I recognized that assessments were used to perform actions such as complimenting, deprecating, complaining, bragging, etc.
More recently, I started thinking about the broadest formulation for actions that involve a participant’s reporting or offering an assessment. I came up with the formulation “reporting a reaction.” If we think of offering an assessment as offering a reaction, then the connection between assessments and epistemics becomes quite clear. In offering a reaction to something, a speaker is understood to imply that he/she has had access to, experience of, and/or knowledge of that something. In reporting an assessment of something, the speaker is making an implicit claim of access to, experience of, and/or knowledge of the matter assessed.
The relationship between offering an assessment of a referent and making an implicit claim of having had access to and/or experience of the referent holds for initial assessments and second assessments. It also holds when people offer assessments in response to stories and news, responses such as “Oh, that’s wonderful!” With these assessments, the reference term (often “that”) refers to the news or story, and the evaluative term displays an understanding of that news or announcement. The speaker of the assessment is not claiming first-hand experience of the reported news or experience but rather assessing what has just been reported. In terms of epistemics, the implicit knowledge claim is that the speaker has access to, experience of, and/or knowledge of the report.
Across these different contexts, participants apparently understand, make use of, and hold each other accountable for honoring the relationship between offering an assessment and indirectly claiming knowledge of the referent assessed.
Question by Jonathan Potter
What are your thoughts about preference organization – how central should it be, what are its problems, where is it going, and what generates it?
Response by Anita Pomerantz
I have two comments in response to your question.
My first comment has to do with the various conceptions of preference. Sometimes, authors refer to preferred/dispreference as characteristic turn and sequence organizations. Using this conception, agreement is preferred because it is done with no gap, no prefaces, explicitly, etc. and disagreement is dispreferred because it is performed with delays, prefaces, mitigations, etc. Preference is no more than a pattern of interactional conduct. No more needs to be considered to assert that an action is preferred or dispreferred than the turn and sequence shape in which the action is performed. This is one way researchers treat preference organization.
Another conception of preference is to regard it as having to do with principles. This conception is described in the chapter on preference that was written by John Heritage and me that is included in The Handbook of Conversation Analysis. Using this conception, an action is not preferred or dispreferred because of how it is performed; how it is performed is a consequence of principles that are taken-for-granted and followed by participants. Participants follow a principle and that results in the particular turn/sequence shapes with which we are familiar. One of the examples we give in the chapter is that rejection is dispreferred inasmuch the participants accept the principle that rejections should be minimized. This is a second way that researchers treat preference organization.
To reiterate, one conception of preferred/dispreferred actions is to equate preference status with conduct, the performance of the action. Another conception of preference status is that participants orient to particular principles that result in the performance of the actions. Let me offer a third conception of preference.
A third conception is related to the second in that it treats conduct as a result of participants’ orientations. I think of this conception as participants’ orienting to cultural premises. A cultural premise in this sense is somewhat different from a principle as we wrote about in our article. A principle most directly concerns conversational actions, for example “Don’t do rejections if you can avoid it.” A cultural premise may concern other matters, such as assumptions about the value of friendship and how to maintain them. An example of cultural premises might be that friendships are important and that “To maintain a friendship, don’t outright reject a friend’s invitations.”
I’ve described three different ways of thinking about preference, and I’m guessing that there are other ways of thinking about it as well. While I would not argue that we need to come to an agreement, we do need to be clear about what we mean by preference and how we are using it in our analyses.
When I was working on developing the idea of preference, I had a conversation with Michael Moerman. He asked me to explain what I meant by preference and I essentially told him what I just described as the first approach. I said something like “This disagreement is preferred here because of the way it’s being done.” He responded by telling me that my argument was circular. “You’re saying it’s preferred or dispreferred because of how it’s done, but then you’re using that as evidence for the preference. You can’t play it both ways. You can’t say it is the phenomenon and then also offer it as evidence for the phenomenon.” I took what he said to heart, but I didn’t have the clarity to think through the problem and/or come up with a solution.
I currently believe that the weakest conception of preference is that it is a recognizable pattern of conduct. We need more than that.What are participants orienting to? I am fine with proposing that participants orient to specific discursive principles and/or specific cultural premises. In one way or another, we need to make account for why/how participants produce the conduct that we associate with preferred and dispreferred actions.
The vagueness of the term preference bothers me. When my students identified an action as preferred or dispreferred, I responded by telling them that they can start there but that they needed to do better than that. They needed to analyze what was happening in the interaction and how the discourse in question was working.
The issue of the various conceptions of preference was the first point I wanted to make. Let me turn to the second point. This point was developed in the Pomerantz & Heritage article that I referred to earlier. A problem that we wrote about is that we often generalize when it comes to preference. We propose that an action, in general, is preferred or dispreferred. We say “agreement is preferred” and “Disagreement is dispreferred.” It is simply not a good way to think about actions because there are a lot of circumstances in which those generalizations simply do not hold.
If we don’t talk in general about the preferences of an action, where does that leave us? Participants do a lot of sense-making work moment by moment. Some of that involves making sense of what the co-participant meant in the prior turn, relating the prior turn to the ongoing course of action, relating a current turn to possible upcoming actions, and so on. For example, when politicians apparently aim to show that they are different from their opponents, they seem to work to maximize their stated disagreements with their opponents. Rather than assert a general preference, as analysts we need to examine the conduct in context to arrive at defensible claims as to what the participants are doing and how they are doing it. Simply identifying the action as preferred or dispreferred does not give us sufficient understanding of the interaction.
Question by Jonathan Potter
Gloss: So you’ve emphasized a number of answers now: you focus on functions over pattern, and if you are simply spotting that you’ve got a preference, you are not fully doing an analysis. Would it be right to characterize your position as something like preference structure identification is a point along the way to doing full analysis and that as you’re getting to the specifics of the action formation, what you’re going to want to do is to look at how each element within this preferred or dispreferred form contributes to the function rather than just being satisfied with you’ve got an X you’ve got an Y. Is that too simple?
Response by Anita Pomerantz
I don’t think that’s the best starting place. I think the best starting place is to see what’s being said. I mean why start with preference? Start with the conduct and see what sense the participants seem to be making/conveying, what understandings seem to be displayed. Those are more fundamental issues. For example, let’s say a response is delayed. The timing of talk is of course important, but this one may not be associated with the action as dispreferred. Perhaps this delay is a product of the speaker’s needing to produce an appropriate argument in response to the prior claim. That may not involve the preference structure of actions. I would be concerned about starting with the notion that preference is operating and then examining each element rather than starting with a broad sense of ‘what’s going on here’ and developing that line.
Question by Steven DiDomenico
What paper are you proud of that hasn’t got as much attention as you would have liked?
Response by Anita Pomerantz
The paper I would choose is “Offering a Candidate Answer: An information-seeking strategy” published in 1988. There are two claims that I like but that haven’t gotten much attention.
First Claim: Currently we accept as given that different forms of questions convey different expectations about the recipient’s knowledge. Questions may be built to convey that the speaker expects the recipient to know the information, the speaker is guessing that the recipient knows the information, the speaker regards it as possible that the recipient knows the information, etc. While I wrote about that, the part that interested me more was my claim that candidate answer questions convey something about the speaker’s knowledge. I proposed that a candidate answer question may be understood as the speaker’s best guess. I suspect that candidate answers incorporated within questions are not always understood as the speaker’s best guess, but I was hoping additional research would be done on candidate answer questions to investigate the assumptions that are apparently made about the speaker’s knowledge in different contexts.
Second claim: I suggested that candidate answer questions that seek information carry a moral aspect with respect to the candidate answer incorporated in the question. In my corpus, the action or state that was referenced in the candidate answer was regarded as appropriate and legitimate and normal. That’s an interesting phenomenon that I don’t think has been picked up. Of course, if an illegitimate, inappropriate, and/or abnormal state is referenced in a candidate answer, it may well be heard as an accusation.
Question by Aleksandr Shirokov
I am curious about the relationship between Ethnomethodology and Conversation Analysis. Observing the emergence and development of both approaches, how did their relationship change over the years? How do you see their relationship in your studies?
Response by Anita Pomerantz:
From my point of view, Conversation Analysis has Ethnomethodology as one of its foundational influences. Both Ethnomethodology and Conversation Analysis look at the work that participants do and examine how they do it. However, there is no doubt that there have been disagreements, arguments, and politics that have gone back and forth. I have not been part of that because, personally, I don’t have time for that. I think there is good work being done in both areas. I have made my choices regarding my scholarship, choices based on both my training and my inclination.
However, I’ve learned from ethnomethodologists. I went back to look at titles of presentations that I’ve given, and it is clear that I am a Conversation Analyst who has been influenced by Ethnomethodology. One of my presentations was entitled “Meaning More than we Say in so Many Words” which is clearly an idea derived from Harold Garfinkel. The title of another presentation was “Notes on Intersubjectivity,” which is a concept I associate with Ethnomethodology. The title of a third presentation was “Negotiating What is Objective.” The idea that the ‘objective world’ is socially constituted is very much in line with Ethnomethodological thinking.
For me, Conversation Analysts and Ethnomethodologists ask complementary questions. Our respective approaches and analyses are potentially informative to each other. I don’t think it makes sense to spend time fighting with each other or putting each other down. I speak not as a purist but as someone attempting to look closely at interaction, to see what might be interesting, and to share what I find.