Recursive Frame Analysis

by Ronald J. Chenail

The Qualitative Report, Volume 2, Number 2, October, 1995

To learn more about how others and I use and understand language in therapy and other such conversations, I have helped to lead a qualitative research project dedicated to the construction of a practical and useful way to "get closer to the talk." The system which has evolved in our project is called Recursive Frame Analysis (RFA) (Chenail, 1991). Created by Bradford Keeney (1991), RFA is a method for understanding and presenting conversations. Later I developed RFA into a qualitative research method for the analysis of narratives, conversations, and other forms of discourse. In this fashion, RFA is used as a type of sequential analysis which helps researchers and therapists to note their perceptions of semantic shifts in a conversation.

I have used this method to research a variety of conversations. These studies include an examination of parents' conversations about their children's heart murmurs (Chenail, 1991), a description of family therapist-supervisor talk behind the one-way mirror in a therapy session (Chenail & Fortugno, 1995), an analysis of divorce mediator-disputants discourse in child custody dispute resolution (Chenail, Itkin, Bonneau, & Andriacchi, 1993, October; Chenail, Zellick, & Bonneau, 1992, October), and an in-depth look at systemic family therapy discourse (Rambo, Heath, & Chenail, 1993). Through each of these studies I was able to learn something different, something new, that can happen when two parties sit down to discuss how to solve a problem.

RFA Assumptions

The roots of Recursive Frame Analysis can be traced back to the work of Gregory Bateson (1972) and Erving Goffman (1974). Bateson and Goffman understood frames as being our conceptual or cognitive views of particular situations. For instance, do we perceive a story we hear from a client to be a tale of problems or of solutions? Our choices of frames help us to hear certain aspects of the talk, while not helping us to hear other parts of the conversation.

Within this system, frame is synonymous with context: "that which leads up to and follows and often specifies the meaning of a specific expression" (The American Heritage Dictionary of the English Language, 1970, p. 288). For an observer to comprehend a particular behavior or action, that observer must identify or construct a context or frame for that event. With RFA, Keeney and Chenail have adapted this general usage of frame and have applied it to the understanding of talk. An etymology of the word "context" helps to illustrate just how an observer constructs these frames or contexts in discourse.

Context comes from the Middle English word "contextus," which means "coherence" or "sequence of words." The past participle of contextus is "contextere:" "to join together" or "to weave" (The American Heritage Dictionary of the English Language, 1970, p. 288). Contexts in conversation are built by the joining or weaving together of words: Knowing the meaning of a word or phrase is the result of joining them together with other words and phrases. From this perspective we become weavers of words as we attempt to fabricate meaning and create coherence for ourselves. It is important to remember that context is created from the weaving together of words because it is on this very matter which many become confused: Context does not cause that which is contextualized to have meaning; meaning is produced when the two--the context and the text, are brought together.

In a sense, the way meaning is produced, by juxtaposing text and context, is similar to the way water, yeast, flour, sugar, salt, and milk become bread dough. The aforementioned ingredients do not cause bread to be made. Water, yeast, flour, and such become dough through their active relationships with each other. As they are kneaded together, unfolding over and over again, back upon each other, the meaning of the ingredients become clearer: Water, yeast, flour, sugar, salt, and milk, taken together, become dough. In other words, the meaning of these ingredients in relationships to each other is dough.

In RFA terms, the process of kneading dough is expressed by the use of the word "recursive" to describe the way talk unfolds upon itself as we attempt to make sense of conversations: Context and text are in turn contextualized by other contexts and text. New meanings are pondered, considered, and re-considered as the relationships of texts and contexts are compared again and again. The form or shape of a conversation comes from this arrangement of the parts of the conversation: Words are woven together to create contexts and then these frames are configured to create a shape or contour to the conversation. RFA then becomes a way to "figure out talk," and an RFA presentation, or more correctly, an RFA re-presentation of a conversation would be a "figure of speech."

Maybe the most important point to RFA is that we actively help to create meaning through our play with words. We greatly figure in our sense making in language. When we talk and write we have to play and to struggle with the arrangement of our words. We arrange and rearrange our words as we strive to express our ideas as best we can in the words we have. The same holds for our listening and reading. We hear sounds and see symbols, and we create meaning from these signs as we actively connect these elements into words, phrases, sentences, paragraphs, and discourse. As we do this, we begin to ascribe meanings to the words. Our choices of meaning for any one word are many and varied and fluid. How we arrive at "a meaning" or "one meaning" for a group of words depends a great deal on our experience with a language in general and our contextualizing of words in a particular conversation.

RFA is an approach to talk which reminds us that we do this contextualizing thing all the time. A word as understood in a dictionary may not be the same as the meaning or meanings a word takes on in a certain conversation at a particular time and place for a specific person. The difference between the two meanings has much to do with this process of contextualizing or weaving of words. RFA can help us both to construct these patterns of words for ourselves and to appreciate how others create their own figures of speech.

To understand the notion of frame as word, Recursive Frame analysts also employ what Deborah Tannen and others (Tannen & Wallat, 1993; Putnam & Holmer, 1992) call "interactive frames." By interactive frames, we mean those linguistic patterns through which we create meaning in our conversations. We build our conversations word-by-word. We understand the words we hear and use in conversation by how we contextualize them. To contextualize or to frame a word is to connect it with other words. Context is built by the ways we connect words with other words in conversations.

For example, each sentence I have used so far in this essay has been built by connecting words. For you the reader, each word I use already comes with a "dictionary" meaning. Each word has its own meaning for you prior to my using it in a sentence. When you come across a word like "frame," you have to look around at the other words I have used along with "frame" in order to construct how I am using and not using "frame" in a particular sentence. "Frame" understood in context means that you construct the meaning of "frame" by understanding it along with the other words around it.

The notion of recursion becomes a bit more complex with Recursive Frame Analysis in that the cognitive frames we think are in recursive relationship with the linguistic frames we speak and hear. Our understanding of a situation helps us to grasp "what is going on," and at the same time, as we experience "what is going on" in a situation, our understandings can be re-shaped or reframed.

For example, if a therapist understands therapy as a "teaching" situation, he or she organizes therapy into "lessons" and "evaluates" how well the client has "learned." If the client does something in therapy which the therapist has never experienced before, the therapist may then see therapy as a "learning" opportunity and begin to appreciate what can be learned from the client-as-teacher.

Also, in conversation, there is a recursive relationship between text and context. A particular piece of text contextualizes other text, and in turn, is also contextualized by the other surrounding bits of text. If someone talks about success in business in terms of "scoring big with a contract" or "slam-dunking the competition," we can hear that this person contextualizes business in a sports frame. At the same time, these juxtapositions can lead us to think of sports in business terms too. Professional athletes sign contracts which pay them for scoring big, and basketball players compete against each other for money in slam dunk competitions during all-star games.

A Musical Interlude

In his 1991 book, Improvisational Therapy, Keeney called RFA a method for scoring conversations. This metaphor was not chosen by accident. Keeney is an accomplished pianist and composer. He has a keen ear for understanding music and in creating RFA, he wanted to apply certain notions from musical theory to the understanding on conversations. He felt that if we could hear conversations in terms of "notes," "phrases," "chords," "melodies," "movements," "tempos," and "rhythms," then we would have a better feel for how conversations are put together.

As a family therapist, Keeney had a language orientation to his clinical practice. He felt it was important for clinicians to get closer to the talk in the room. He wanted therapists to learn how to hear "stuck patterns" in clients' talk and to see that these repetitive conversational patterns of talk reflected how the clients were mired in on-going life problems. From this perspective, therapy was conducted by identifying repetitive patterns in clients' talk and by engaging in conversation with the clients so as to help them to find solutions to their problems. For Keeney, change in therapy meant a change in the ways clients talked, and most likely thought, about their problems. If he could hear changes in how clients talked about their problems in sessions, then he assumed that there would be corresponding changes happening in the clients' lives outside of therapy.

To help his students to develop an ear for hearing these patterns in clients' talk, Keeney taught them how to score conversations, just like a musician would score musical compositions. In this way, RFA became a "listening aid" for his students. Just as a hearing aid helps a person to hear sounds which would have otherwise "fallen on deaf ears," RFA as a listening aid would help a therapist to recognize meanings or differences in a conversation which would have otherwise have been meaningless or just noise to the therapist's "untrained ears."

In this RFA as listening aid process, therapists were trained to listen to therapy sessions in terms of notes, phrases, chords, and melodies. They were taught to hear each word in the talk as a "note." The therapists would listen to these individual words-as-notes and examined them to listen how they were in accord, or "in-a-chord" with the other surrounding words. When the flow of conversation continued on the same subject or topic, to the therapists ears, they talked about the words as being a musical phrase or melody.

As time went on, Keeney moved from this adapted musical notation system. He felt that there were difficulties in adapting a musical scoring approach to the spoken and written word. He believed that the complexities of music theory and musical notational systems made the whole process of following conversational flows too difficult. In addition, Keeney felt that the basic differences between systems of notes and systems of words precluded futher development of the musical metaphor. He pointed out that wherein the physics of music made it possible to clearly identify a particular note as being in a particular key and harmonics allowed for the accurate construction of notes into chords, the multiple interpretation found in conversations did not seem to conform to the particularities of scoring music, or in other words, a word is not quite the same as a note.

As a result of this new thinking, Keeney began to develop the frame metaphor for understanding conversations. For Keeney, the notion of frame was a bit looser than note and that flexibility was a better fit with how he conceptualized words, meanings, and conversations. In this new notational system, words went from notes to frames. Contextual groupings of words went from phrases to galleries. And lastly, the creation of meaning went from harmony to recursion.

Despite this shift from musical notation to framing distinctions, a very important concept remained in the mix--noting. In RFA, noting is a two-step process by which a listener "notes" differences in the talk and then makes personal renderings of these notings. According to Bateson (Harries-Jones, 1995, pp. 203-204)

...difference enters twice into the process of perception.... In the first instance (in time), we subjectively percieve difference and differences that make a difference. In the second instance, a perception of change in the pattern of differences becomes the distinction on which percepts and premises are constructed. Through this "product" of difference, the "given" distinction enters into an aesthetic sensibility. Making visible these differences requires investigation of what sort of "product" of interactions we sense through our aesthetic sensibilities.

Although this all sounds difficult, this two-step process being described is well-known to us, especially if we have ever had to "take notes" from an article or from a lecture in class. As we read the words written by the author or as we hear the sounds produced by the speaker, we begin to note differences in the flow of the words or sounds. Next, we begin to record what has now become "note-worthy" to us in our notebooks or in the margins of our books. The words written on the pages of these notebooks and in these margins are our own personal renderings of our notings of what we heard in the lecture or read in the articles--notable differences or differences of note which have made a difference to ourselves.

In RFA, we are reminded that what we "hear" in a conversation is only the first step. This "perception of difference" must be followed up with a "listening" or a making sense of what we have heard or what we thought we have heard. RFA aids us in this listening or making sense process by asking that we make visible or audible these patterns of differences as we create these personal aesthetic sensibilities--our own figures of speech.

Creating Figures of Speech

In RFA, figures of speech are created when listeners begin to organize the flow of conversation so it coheres or makes sense to them. In this light, RFA can be seen as a way to present the "logic" of a narrative or conversation at hand. Recursive frame analysts attempt to seriously play with the wordplay so as to create meanings or interpret the talk. This wordplay begins with analysts' careful listening to a conversation. This listening can be done in real-time, as the event is unfolding, or the listening can be done out-of-real-time, from audio or video tapes, from transcripts of conversations, or from a variety of combinations.

In listening to the talk, recursive frame analysts become sensitive or curious to differences in the conversation. For example, where and when does the content being discussed by the participants change? Are questions answered? Are answers questioned? How does it make "sense" that one speaker says "X" after another speaker says "Y," or how does it make sense that after one speaker says "Y," that another speaker does not say "X"? Somehow, in some way, recursive frame analysts have to try to make sense of the "what" and the "how" of conversations. What text seems to go with what context? What new meanings are generated if the text becomes the context and vice versa?

In a face-to-face conversation, I, as an RFA-informed listener, try to name the talk I am hearing. By naming I mean that I consciously try to characterize to myself what I think is being said. I always want to have an idea where I think I am in a particular conversation at a particular time. This is very important because what I think is happening in a conversation will contextualize or frame what I will and will not say in a conversation. Depending on how I characterize a particular utterance in a conversation, I have to decide whether I will continue in that line of conversation, or whether I will open up a new line of talk, or whether I will participate in the talk by remaining silent.

For instance in a therapy conversation, I try to identify for myself if I think the client and I are talking about "problem-talk" (e.g., "My husband thinks it's okay for our two-year old to jump in bed with us when we're sleeping, but I don't think that's right), or "solution-talk" (e.g., "What would be the first sign for you all that your relationship was getting better?"), or some type of "non-therapy talk" (e.g., "The rest room is down the hall and to your left.). From my model of therapy, these framing or naming activities are crucial for the performance of therapy. I organize my listening to hear possible solutions. Given this stance, I judge every utterance I hear as to its solution potential. If I or no one in the room hears some talk as being a possible solution, then there is no solution talk in our therapy conversation. If, on the other hand, I think a particular bit of discourse uttered by one of us in the room sounds like a solution, then I try to pursue that line of inquiry in the conversation.

For physicians, I think, a similar process would hold for doctor-patient interactions. I would think that doctors would track patients' conversation to hear whether or nor their patients are telling stories about diseases, illnesses, life's problems, or the weather. It would be important for a doctor to track distinctions in the talk like "compliance-talk" (e.g., "I've taken my pills every day."), "quality-of-life-talk" (e.g., "I can walk to the store again."), "patient-history-talk" (e.g., "I had the measles when I was seven years old.), and other such biopsychosocial threads that make up doctor-patient interactions.

As these musings take shape over the course of a conversation, a listener can begin to construct how they understand a conversation to be patterned. If time permits or if the listener wants to conduct some research, then the recursive frame analysts can draw up their re-presentations or figures of speech to visualize how the conversation is taking shape for them. Recursive frame analysts can choose to show how context and text relate to each other in a variety of ways:

From this figure, frame is to be understood within an RFA context as contrasted with the meaning of frame within a picture context.

RFA Practice

With RFA, researchers listen or watch a recording of a conversation while reading and re-reading a transcript of the discourse in question. As mentioned above, RFA is a type of sequential analysis which helps researchers note semantic shifts in a conversation. In RFA, observers listen or watch a recording of a conversation while reading and re-reading a transcript of the discourse in question. They discuss (a) how they observe the subject matter of the conversation being developed (i.e., an emphasis on content or what is being said) and (b) when they observe that there are shifts from one subject to another in the course of a conversation (i.e., an emphasis on process or how are things being said by the speakers). This method is especially useful when there is an interest on the part of researchers to chart observations of how they see the talk of an interaction unfold.

Both of these practices are based on the notion of difference. As we listen to someone we discern that the speaker is saying a variety of words. We can identify one word from another word in many ways. There are phonetic differences (i.e., differences in pronunciation), semantic differences (i.e., differences in meaning), and pragmatic differences (i.e., differences in how words are used). As we listen and mark differences between the words we hear, we are creating patterns in the talk. In RFA terms, a basic pattern of meaning created by a listener is called a frame, a bit of talk marked as being distinctly different from its surrounding words.

An RFA analysis proceeds as follows: After the recording had been perused numerous times, the team members noted instances when there were repetitive use of certain words on the parts of the speakers and then the team began to "chunk" these instances into informal groupings. In RFA, chunking is the process by which an observer or team of observers makes sense of a collection of data by gathering together those discourse examples which seem to the observer(s) to have some characteristics in common with each other. In RFA lingo, we say that we chunk these frames into galleries.

For example, in a therapy conversation, one gallery that can usually be constructed is a Problem Gallery. A problem gallery is a chunking that would contain all those frames uttered by the client(s) that the therapist or researcher understands as "problems." Another gallery commonly chunked by therapists would be a Solution Gallery. Again, this gallery would be a chunking of frames all understood as being solutions or possible remedies by the therapist or researcher. In both cases, the therapist's or researcher's chunking of the frames may or may not be the same as how the client understands the conversation. In addition, other therapists or researchers may also differ on how they chunk the talk.

With RFA, researchers can also conduct pragmatic analyses (Haslett, 1987; Nofsinger, 1991). Pragmatic analysis with RFA is when researchers focus on how speakers used their language in an attempt to shift the flow of the talk. In listening to the talk, recursive frame analysts become sensitive or curious to differences in the conversation. For example, where and when does the content being discussed by the participants change and who helps to make that shift? Are questions being answered? Are answers being questioned? How does it make "sense" that one speaker says "X" after another speaker says "Y," or how does it make sense that after one speaker says "Y," that another speaker does not say "X"?

Somehow, in some way, recursive frame analysts have to try to make sense of the "what" and the "how" of conversations. What text seems to go with what context? What new meanings are generated if the text becomes the context and vice versa? As these musings take shape, recursive frame analysts draw their re-presentations or figures of speech to visualize how the conversation is taking shape.

RFA-Informed Therapy: Two Examples

To help the readers understand how RFA research has informed my practice of therapy, I would like to discuss briefly two in-the-room-processes which I first noticed as an RFA researcher, and then subsequently used as a therapist, are torqued talk and opening up closings. Most of the therapy I practice and supervise nowadays is organized by these two patterns of talk. As a result, I find that by concentrating on whether I think the talk seems to be "standing still" (torqued talk) or whether I gather that the talk is "moving from one" (opening up closings), I can better gauge how I want to participate in the conversation at a particular moment in time.

Torqued Talk

Most clients come to herapy because they feel that they are stuck. They do not know what to with a life situation, or if they do know what to do, they are unable to accomplish their goal(s). By tracking the frames in therapy with RFA, I can usually notice how tightly the talk can be for clients in therapy. They will repeat their stories, sometimes word for word, again and again in the sessions. The rigidity or tightness I experience in these conversations led me to describe this talk as being torqued. By torqued, I mean that the talk seems twisted tightly for me and that I am having trouble hearing any change in wording in the conversation.

In a recent study (Rambo, Heath, & Chenail, 1993), I examined a full-length family therapy session. The family had come to therapy to discuss how a son, "Randy," could move into his father's ("Ted") home to live. The son's mother and father had been divorced a number of years and during that time, he had lived with his mother and her new husband, his father, and also with an uncle.

During the session, I had chunked a number of frames into a gallery I called "Randy and Ted Getting Together Talk," a solution gallery. The talk in the conversation returned to that gallery nine times. Seven of those times the conversation turned to that topic, it was followed by another gallery, "Ted's Problem with Randy Talk," a problem gallery. To me, that seemed to be representative of torqued talk.

Many therapy sessions can be seen as having a pattern of torqued talk similar to the one I experienced in studying the case with Randy and Ted. Until some new bit of talk can be introduced into the conversation, both the talk in the therapy room and situation outside of the room will remain stuck. One technique I have learned that can be helpful in untorquing talk is the opening up closing.

Opening Up Closings

RFA can be used for conducting a sequential analysis of discourse. As in the case above, the recursive frame analyst charts the flow of conversation and marks when conversations shift from one chunking to another. For instance, the analyst may mark when the conversation shifts from talk about the children's school problems to talk about the children's problems at home. The talk may then shift from talk about children's problems to talk about husband and wife problems. In each instance, the researcher or therapist would mark or take note of when they would notice one of these shifts.

Along with charting changes in meaning or semantic shifts in these conversations, a recursive frame analysts may also take note of who is initiating these shifts and how the particular speaker is able to successfully move the talk from one gallery to another. The term I use to note this shifting phenomenon is called opening up closings, a term I have borrowed from conversation analysis (Schegloff & Sacks, 1973). With an opening up closing, the speaker uses certain words which allows for the opening up of a new line of conversation while simultaneously closing down the current topic of talk. In RFA terms, one gallery is opened up as another is closed down.

In studying divorce mediation discourse, a group of colleagues and myself (Chenail, Itkin, Bonneau, & Andriacchi, 1993, October; Chenail, Zellick, & Bonneau, 1992, October) became quite curious how divorce mediators were able to help disputing parties come to resolution regarding child care, custody, and support. In many of these cases, the ex-husband and ex-wife had not had much success in getting along with each other, much less working out complex agreements over their child or children.

When we examined the transcripts of over 30 divorce mediation cases from an RFA perspective, we noted a number of times that the mediator was able to open productive resolution talk, while at the same time, closing down unproductive fighting talk. As we took a closer look at this gallery transition moments, we noticed that the talk took its turns when the mediator reminded the parties that they were both at the mediation sessions for the best interests of their child or children. In most of the cases we observed, that move on the part of the mediator helped to open up a new line of talk different from the preceding line of conversation.

As result of this and other work I have done on conflict resolution, I have become more sensitive to how speakers open new galleries in other conversations. This research turn towards pragmatics, the study of how people do things with words, has helped me to create new ways of loosening up previously torqued talk in therapy sessions. This work has also helped me to notice another interesting speech act, the closing down opening. This type of talk is occurs when one speaker appears to offer an opening such as, "I really think that that might work..." and then follows it up with a closing down ending such as, "...but not with this situation." Closer attention to the uses of "Yes, but" in therapy has helped me understand how seemingly promising lines of therapeutic talk can be quickly shut down, and how I can possibly change the situation without becoming part of the torqued talk.

Some Practice with RFA

A couple of examples are presented over the next few pages. In the first one, a physician and a patient are just beginning an office visit and the patient is telling why he came in for that particular visit. In the second example, a physician is talking with a mother whose daughter has a heart murmur. Read both examples over a few times.

Example One
1.1 Physician 1: Well John, how are things going?
1.2 Patient: I've been having a lot of pain in my back I can't mow the lawn anymore or pick up anything around the house (pause) my wife and I have been having quite a few arguments lately we don't get along like we used to do my kid is going off to college and that's bringing on a lot of pressure (pause) I didn't realize how expensive they've become (pause) my knees hurt too especially with all this wet weather we have been having I've been having problems sleeping through the night my sinuses have really been kicking up I get stuffed up at night and I'm snoring more or at least that's what my wife tells me (pause) anyway, when I get up I don't feel rested and I've been really tired at work for the past month or so what do you think's wrong with me?
Example Two
2.1 Physician 2: When Dr. Ramos was listening today, he heard the heart murmur in its said location and I came back and listened to it and sure enough, there is a murmur there which we hadn't heard before this and of course, you had known about a heart murmur when he was born? What did they tell you about it then?
2.2 Mother: They just told us he had a little murmur.
2.3 Physician 2: Okay and that it would probably go away?
2.4 Mother: Um-huh
2.5 Physician 2: And we can hear it now but he seems to be growing okay?
2.6 Mother: Um-huh
2.7 Physician 2: And he seems not (pause) he doesn't turn blue or anything like that he has the cough but that has been improved since he started taking the Ventolin of course you have asthma have you started taking Ventolin every day?
2.8 Mother: For Jacob?
2.9 Physician 2: Yeah
2.10 Mother: Three times a day
2.11 Physician 2: And he seems to not cough when he runs with it he doesn't squat or do anything like that? He doesn't have to squat?
2.12 Mother: No
2.13 Physician 2: He doesn't tire when he's feeding or tire excessively? [Mother nods] okay Uh (pause) as you know (pause) well (pause) you may not know (pause) a heart murmur is just an extra sound that the heart makes when it contracts there's a normal sound closing of the valves but sometimes between them you can hear a "clicking" sound he's got what we call a systolic murmur that is when the heart is pushing blood out (pause) you hear a squish coming up that way

After reading the examples a few times, draw boxes around all the different frames you observe in the patient's words. Remember, a frame would be the smallest grouping of words that have a coherent meaning for you. A frame could be just one word or it could be a number of words such as a phrase or a sentence.

For instance, in Example One, I framed much of the patient's talk like "I've been having a lot of pain in my back," "I can't mow the lawn anymore," "pick up anything around the house," "my wife and I have been having quite a few arguments lately," "my kid is going off to college and that's bringing on a lot of pressure," "my knees hurt," "problems sleeping through the night," "my sinuses have really been kicking up," "I'm snoring more," "I get up I don't feel rested."

Re-read the examples and pay close attention to the words you framed. Can you chunk any of these frames into any semantic groupings or galleries? If you can chunk any frames into a semantic grouping, what would you call this gallery? For my analysis, I chunked all of the patient's frames under one gallery and called it "How things are going" Gallery.

For Example Two, I framed the following Physician 2 talk: "the heart murmur," "growing okay," "turn blue or anything like that he has the cough but that," "started taking the Ventolin," "of course you have asthma," "he seems to not cough when he runs with it," "he doesn't squat or do anything like that," "He doesn't tire when he's feeding or tire excessively," "a heart murmur is just an extra sound that the heart makes when it contracts there's a normal sound closing of the valves," "but sometimes between them you can hear a "clicking" sound," "he's got what we call a systolic murmur that is when the heart is pushing blood out (pause) you hear a squish coming up that way." As for galleries, I chunked a number of frames into a "Heart Murmur" Talk Gallery and I also grouped some other frames into an "Asthma" Talk Gallery.

I also took notice how Physician 2 moved the conversation between the two galleries. Physician 2 started the talk in a "Heart Murmur" Talk Gallery (see Turn 2.1) and then switched galleries in Turn 2.7 as he went from "Heart Murmur" Talk to "Asthma" Talk. I thought there was another switch in Turn 2.13. This time I marked that Physician 2 moved the talk from "Asthma" Gallery talk back to "Heart Murmur" Talk. In contrast to the Physician 2 switches, I did not notice that the mother made any semantic shifts in the conversation. I also did not note any new frames contributed by her in the conversation. It seemed to me that her participation was mostly of a "answering" posture in that she seemed to participate only as a respondent to Physician 2's inquiries.

With RFA, there is a variety of ways to re-present conversations on paper. Earlier I described a way to show galleries and frames using an indentation-style of re-presentation. The following gallery is listed with its corresponding frames in such a fashion:

Gallery 2: "Asthma" Talk

Frame 2.1: "started taking the Ventolin"
Frame 2.2: "of course you have asthma"
Frame 2.3: "he seems to not cough when he runs with it"
Frame 2.4: "he doesn't squat or do anything like that"
Frame 2.5: "He doesn't tire when he's feeding or tire excessively"

Indentation is used to show that the Frames 2.1 through 2.5 are to be understood as being embedded in Gallery 2. In addition, the frames can be annotated by showing the numbered lines or turns where the frames appeared in the original transcript.

Another option would be to draw boxes in order to re-present how the researcher constructed the frames and galleries as can be seen in Figure 1.
Figure 1
Gallery I Gallery II Gallery III
"Heart Murmur"
to "Asthma"
to "Heart Murmur"

I encourage Recursive Frame Analysts to create whatever re-presentational schema which best helps them to show how they "see" the talk being organized. Also, the type of RFA re-presentation should also cohere with the research question which the researcher is working. For instance, if the researcher's question was about doctor-patient relationship and how each party contributes to the construction of knowledge, each gallery could be divided into a physician side--patient side. Frames in a gallery would be placed on either the physician or patient side as depicted in Figure 2.
Figure 2
Gallery I: "Asthma" Talk

In this way, at a glance, a reader could read how much and what is being contributed by whom in a particular conversation. For example, an RFA done in this fashion with the two examples in this paper in the split-gallery fashion would show some interesting contrasts between the active patient in Example One and the active physician in Example Two.


From the years of analyzing sessions from an RFA perspective as a researcher, I have a feeling of operating from within the talk in therapy. As I am engaged in the clinical hour as a therapist, I am also participating as a discourse analyst. I find this reflective stance helps me to stay closer to the process within the therapy room both as therapist and as researcher.

As a researcher, I feel that Recursive Frame Analysis provides me with the flexibility I require to do the variety of research projects I need to pursue. I can focus on sequential, semantic, or pragmatic analysis. If I choose, I can combine these analyses and examine sequential-semantic analysis or pragmatic analysis of semantic sequences, and so forth. Whatever the choice of Recursive Frame Analysis, what remains common is that RFA allows me an easy way to mark the talk the way I hear and understand it.


The American heritage dictionary of the English language (second edition). (1970) New York: American Heritage Publishing.

Bateson, G. (1972). Steps to an ecology of mind. New York: Ballentine Books.

Chenail, R. (1991). Medical discourse and systemic frames of comprehension. Norwood, NJ: Ablex.

Chenail, R. J., & Fortugno, L. (1995). Resourceful figures in therapeutic conversations. In G. H. Morris & R. J. Chenail (Eds.), The talk of the clinic: Explorations in the analysis of medical and therapeutic discourse (pp. 71-88). Hillsdale, NJ: Lawrence Erlbaum.

Chenail, R. J., Itkin, P., Bonneau, M., & Andriacchi, C. (1993, October). Managing solutions in divorce mediation: A discourse analysis. Paper presented at the Twenty-first Annual Society of Professionals in Dispute Resolution International Conference, Toronto, Canada.

Goffman, E. (1974). Frame analysis: An essay on the organization of experience. Cambridge, MA: Harvard University Press.

Harries-Jones, P. (1995). A recursive vision: Ecological understanding and Gregory Bateson. Toronto: University of Toronto Press.

Haslett, B. J. (1987). Communication: Strategic action in context. Hillsdale, NJ: Lawrence Erlbaum.

Keeney, B. P. (1991). Improvisational therapy: A practical guide for creative clinical strategies. New York: Guilford.

Nofsinger, R. E. (1991). Everyday conversation. Newbury Park, CA: Sage.

Putnam, L. L., & Holmer, M. (1992). Framing, reframing, and issue development. In L. L. Putnam & M. E. Roloff (Eds.), Communication and negotiation (pp. 128-155). Newbury Park, CA: Sage.

Rambo, A. H., Heath, A. W., & Chenail, R. J. (1993). Practicing therapy: Exercises for growing therapists. New York: W. W. Norton.

Schegloff, E. A., & Sacks, H. (1973). Opening up closings. Semiotica, 8, 289-327.

Tannen, D., & Wallat, C. (1993). Interactive frames and knowledge schemas in interaction: Examples from a medical examination/interview. In D. Tannen (Ed.), Framing in discourse (pp. 57-76). New York: Oxford University Press.

Author Note

An earlier version of this paper was presented at the Family in Family Medicine Conference, March 2, 1995, Amelia Island, FL. and a shorter version of this paper will appear as part of the following publication:

Gale, J., Chenail, R., Watson, W., Wright, L., & Bell, J. M. (1996). Research and practice: A reflexive and recursive relationship--three narratives, five voices. Marriage and Family Review, 24(1).

Ronald J. Chenail is the Dean of Nova Southeastern University's School of Social & Systemic Studies and an Associate Professor in the School's Department of Family Therapy. His e-mail address is

Return to the Table of Contents