Home / Articles / Volume 10 (2013) / Formulations in ‘Trouble’ chat sessions
Document Actions



In this article, we examine the organizational and interactional tasks of formulations in chat counseling, as these have been found to play an important interactional role in face-to-face counseling. The data consist of 53 individual chat sessions of approximately 25 minutes each between a coach and a client with moderate symptoms of depression and/or anxiety. Using conversation analysis (CA) as a method, the study finds three uses of formulation: 1) organizing formulations, 2) clarifying formulations, and 3) ‘therapeutic’ formulations. The first and third types are used in ways comparable to face-to-face therapy. The second type is a strategy to deal with trouble in the interaction and aims at re-establishing intersubjectivity. The analysis reveals the important role of question marks and reduced conditional relevance of a response to a formulation in chat. This study contributes to the understanding of institutional interaction online and the interconnection between medium structure and social interaction.


Trouble is increasingly discussed online in both lay and institutional settings. A number of studies have found that online interventions are as effective as traditional face-to-face interventions: Both traditional face-to-face counseling and online counseling help about 80% of those seeking psychological help (Riper et al., 2007). Currently, online health promotion and counseling are being strongly pushed by health policies in the Netherlands (Gezondheid dichtbij; landelijke nota gezondheidsbeleid [Health closeby; national report health policy], 2011). This study uses conversation analysis (CA) to investigate chat sessions in a counseling setting. It analyzes the sequential organization of the chat interactions between clients and coaches. While the label “conversation analysis” might imply that the method is only suited for the analysis of routine conversation, it is in fact used to study a wide variety of interactions, including interactions in institutional settings (Drew & Heritage, 1992; Heritage & Clayman, 2010). Moreover, CA is increasingly used for the analysis of computer-mediated communication (CMC), ranging from asynchronous forum interaction (e.g., Lamerichs & Te Molder, 2003; Stommel & Koole, 2010; Vayreda & Antaki, 2009) to synchronous chat interaction (e.g., Danby, Butler, & Emmison, 2009; Nilsen & Mäkitalo, 2010; Schönfeldt & Golato, 2003).

A fair amount of research has been done on interactions in face-to-face counseling and therapy (e.g., Davis, 1984; Peräkylä, Antaki, Vehviläinen, & Leudar, 2008; Vehviläinen, 2003). However, little is known about interactions in the domain of online counseling and other online psychological interventions (but cf. Danby et al., 2009). Research on CMC suggests that there is an intricate relationship between structures of social interaction and the medium of interaction (e.g., Herring, 1999; Nilsen & Mäkitalo, 2010; Schönfeldt & Golato, 2003). At the same time, these studies have shown that even text-based CMC is “conversational” (Herring, 2010) in many respects.

In this article, we examine a form of computer-mediated counseling to explore structures of social interaction that may be related to the medium of interaction. Our data are chat conversations between coaches and people with moderate symptoms of depression and/or anxiety. Specifically, we examine the role of formulations (Heritage & Watson, 1979), an interactional tool that has been found to be frequent in face-to-face counseling interactions.

Formulations in Counseling Settings

The term “formulation” was first introduced by Garfinkel and Sacks (1970).1 In research following this often-cited paper, the term is most frequently used to label talk that somehow reformulates and negotiates the meaning of preceding talk:

[O]ur interest is in formulations which characterize states of affairs already described or negotiated (in whole or in part) in the preceding talk. (Heritage & Watson, 1979, p. 126)

Formulations are known to be produced routinely in institutional interaction, whereas they are rather uncommon in everyday conversations (Drew, 2003). In various institutional settings, it has been found that professionals rather than lay participants do the formulating, e.g., in radio interviews (Heritage, 1985) and in courtrooms (Van der Houwen, 2009).

Formulations have been studied frequently, especially in general practice consultations (Beach & Dixson, 2001; Gafaranga & Britten, 2004; Hak & de Boer, 1996) and, even more relevant for our analysis, in counseling settings (Antaki, 2008; Antaki et al., 2005; Davis, 1984; Hutchby, 2005; Vehviläinen, 2003). In the context of an analysis of formulations in psychotherapy, Antaki (2008) states that “psychotherapists want to find out how clients see things, and then, at some point, get them to see things differently” (2008, p. 26). Formulating is one way of presenting an alternative view of the client’s account, and it can thus be considered an important conversational tool for counselors. Antaki (2008) subscribes to Heritage and Watson’s (1979) understanding of formulation in their insistence on the intimate relationship between what someone has just said and the formulator’s account of it. He states:

That allows us to see three specifically directive editorial features: the formulator’s claim to find the new description in the very words of the previous speaker; their transformation of those words; and the presumption that this new description is to be agreed with. […] Such locally implicative force is a powerful resource for the speaker who wants to get an alternative, and possibly rival, version of events on record. (p. 33)

This means that the examination of formulations involves not only the formulation design (in relation to the words of the client) but also the client’s uptake against the background of a preference for agreement with the formulation.

Various uses of formulations in counseling can be discerned (Antaki, 2008). First, formulations can be used as a way of interpreting the client’s words psychologically, or preparing such an interpretation (e.g., Davis, 1984; Hutchby, 2005; Vehvilianen, 2003). Thus, formulations help the counselor to make the conversation therapeutic in orientation. To give an example, a formulation can reshape the client’s talk to refer to the client’s feelings, and thus to feelings as a therapeutically relevant topic (Hutchby, 2005). Second, formulations can be used to manage the interactional progress of the interview, for instance by closing down certain avenues of talk (Antaki et al., 2005). This is an organizational use of formulations: formulating what the client has said as not being psychologically relevant, for the moment, allows the counselor to move on to other business while also being attentive to something in the client’s words. Third, formulations in history taking can be used to cast the client’s symptoms into a shape more suitable for later interpretative work (Antaki et al., 2005).

Danby et al. (2009) have shown that formulations do not only occur in spoken counseling, but also in online counseling. In the opening sequences of online counseling, formulations are the first thing done by counselors after the client has presented his/her reason for contacting the service. The authors claim that in chat counseling the formulations are more ambiguous than in spoken counseling, because “it may not be as evident whether the formulation is being used to summarize the client’s problem, to prompt a continuation, or to invite the client to offer their perspective on the problem” (Danby et al., 2009, p. 109). It remains to be seen whether this suggestion holds for formulations in chat generally, not just the opening sequences. Moreover, the claim that formulations in chat are more ambiguous has not been related yet to clients’ responses.

With our analysis of formulations in chat sessions in the domain of psychological intervention, we aim to expand knowledge about formulations in counseling settings, but we also intend to explore how an interactional tool frequently used in spoken counseling may have specific characteristics related to the chat setting. Thus, the questions leading the analysis are: what uses of formulations can be found in counseling chat sessions? And which uses and features seem specific to the chat setting?


The data for the analysis we present here consist of 53 logs of chat sessions of approximately 25 minutes each between a client and a coach. The logs were obtained from the Department of Clinical Psychology, VU University Amsterdam. Each client was offered one 25-minute chat session a week during five weeks. The coaches and clients had no personal information about each other. The coaches received instructions on how to do the chatting. The five-week chat service is meant as a form of support rather than treatment based on psychotherapeutic techniques. Coaches should ask about the client’s wellbeing, ask about what s/he did during the past week, and give advice. In other words, they are supposed to provide a ‘listening ear’ like a neighbor or friend would do. According to the instructions, this listening, via email or chat, can be ‘enacted’ by referring to the examples the client has given and by repeating, paraphrasing, or summarizing the words of the client. Furthermore, the coaches are instructed to avoid the use of difficult words and a focus on psychological themes, negative thoughts, or psychological interpretations. On the basis of these instructions, one may expect that this chat support is more like mundane conversation (with a neighbor/friend) than psychotherapeutic conversation. However, a preliminary analysis of our materials revealed that formulations are ubiquitous in the chat sessions. Since in mundane talk formulations are extremely rare (Drew, 2003), this indicates that the participants treat the sessions as institutional interaction.

The clients whose chat sessions we have drawn on for this article are 5 men and 10 women with moderate symptoms of anxiety and depression. The three coaches are female psychology students in their third year of study for the bachelor’s degree. They worked with a maximum of two clients during a five-week period.

Ethical considerations are important when studying Internet communities (Flicker, Haans, & Skinner, 2004); this is even more true when studying online counseling interactions. Since we did not collect the data ourselves, we depend on the ethical research design of our colleagues, a design which is in line with most of the guidelines proposed by Flicker, Haans, and Skinner (2004). Most importantly, the clients were fully informed about the study by regular mail at least a few weeks in advance, and in order to participate, they were required to sign and return a consent form. By being part of the study, the clients did not have to pay for this chat service.

Two implications of the type of data we are analyzing concern 1) the chat system used and how turns appear on the screen and 2) how the chat excerpts compare to transcripts of spoken interaction. The chat program used was MSN or Skype (the client had to choose). The MSN and Skype chat systems differ in certain technical respects. First, MSN does not offer a time registration of the placement of turns (Skype does), so the MSN logs lack time stamps. Second, if the co-participant does not produce a next turn, the latest participant’s next turn appears directly under his/her previous turn, while in Skype the new turn would appear as a completely new turn, preceded by the participant’s name. For the MSN logs this means that every ‘enter’ in fact constitutes a unit of typed text. Moreover, in MSN (but not in Skype) a symbol indicates if the co-participant is typing. A coach reported that due to this feature the participants sometimes waited for the co-participant to finish typing before typing a next turn. Unfortunately, we can only speculate whether and when these technical aspects influenced the interaction, since we only had the logs and no registration of the participants’ actual typing.

The second implication is the difference between intonation in spoken interaction and the use of punctuation in chat interaction. Our chat data are not research products (like transcripts), but participant products. This means that, for instance, a question mark in chat interaction is an instance of punctuation as used in written language, while a question mark in a transcript of spoken conversation is a convention to indicate rising intonation. This implies that a participant-typed question mark makes a response relevant in a different way than rising intonation would. In the analysis we elaborate on the implications of such medium characteristics in the use of formulations.


Our method of analysis is Conversation Analysis. A main contribution of CA to the analysis of CMC is its focus on sequential aspects of the interaction (Schönfeldt & Golato, 2003; Vayreda & Antaki, 2009) such as adjacency pairs (Schegloff, 1968).2 Initially, the two authors of this article studied the data by closely reading the interactions. A recurrent sequence was found in which coaches, especially, would summarize preceding interaction; these sequences have been found in other institutional setting as well and have been labeled formulations, as discussed above. We then collected formulations and examined the sequential context for: 1) whether the utterance formulated something that was typed before (either in one of the immediately preceding turns, or in the previous session when it included a reference to “last time/week”), and 2) the uptake by the co-participant, using the conditional relevance of formulations to receive a response (preferably a confirmation) as a point of departure. This resulted in a collection of approximately 100 formulations produced by coaches and very few by clients. Therefore, we decided to focus on coach formulations. Our next step was to collaboratively examine six chat sessions in depth in order to analyze the formulations in the local context of the interaction. In this phase, we also did a data session with colleague interaction analysts to further refine the analysis.


We found three uses of formulations in our materials. The first was formulations that appear to have an organizational function. These formulations occur at various points in a chat session but are most prominent at the beginning and end. With these formulations the coach summarizes the previous session or the current session, respectively. Second, we found formulations that function as clarifications of an ambiguity in what the client has just said. This use is sometimes related to the medium of interaction. Third, we observed formulations that are found in face-to-face counseling as well, namely ones that propose an alternative view of the client’s account and thus make the conversation therapeutic.

1. Formulating to Organize Chat Sessions

The first type of formulation, which was easily identifiable, are formulations that function at the organizational level of the interaction. Examples are given in (1a-c). The formulations occur at the beginning and end of almost every chat session and seem highly ‘scripted.’ Rather than orienting toward attaining mutual understanding concerning the directly preceding interaction, these formulations appear to set the agenda, prompt the client to elaborate on a certain topic, or summarize and close the session as a whole.

Extract (1a) is taken from the second session of coach (CO) Eva and client (CL) Simone, and extract (1b) is taken from the second session of coach Eva and client Lisa. The example shown in excerpt (1c) is from coach Karin and client Miriam; this example will be analyzed in more detail. All excerpts are translations of the original Dutch chat interaction made by the authors of this article. The translations are as idiomatic as possible, but it should be noted that at times the original Dutch is not standard or grammatical written Dutch.3,4

Excerpt 1a. (1247-2)

Excerpt 1b. (1008-2)

Excerpt 1c. (1844-3)5

In excerpt 1c, which comes at the end of chat session 3 with Miriam, coach Karin summarizes the various things that they have talked about. The formulation is offered to the client as a summary of the chat session (with some evaluative comments woven into it) and put to the client explicitly, to assess its accuracy (line 9). Differently from formulations that formulate the immediately preceding interaction, these formulations formulate the chat session as a whole. The formulations in excerpt 1b and 1c function to wrap up the chat session and are reminiscent of formulations in other settings that fulfill a similar role. In medical consultations, for instance, ‘action formulations’ have been identified. In doing action formulations the doctor orients “towards acceptance of the agreed upon future action” (Gafaranga & Britten, 2004, p. 147). Similarly, in the courtroom, ‘judgment formulations’ have been found which formulate the main issues that are relevant to the judgment and thereby serve to draw a case to a close (van der Houwen, 2009). The closing formulations that we find in the chat sessions are similar to those found in the institutions mentioned in that they also bring the interaction to a close, but differ in that they do not project action or judgment. Rather, they appear to function as ‘minutes’ of the interaction which the client may sign off on and that can be drawn upon in the next session (see excerpt 1a and excerpt 2 for examples of this).

These organizing types of formulations may also occur at other points in the chat interaction. The next excerpt comes from the chat session following the one discussed above. In the remainder of this article, we show two other examples from this session, because it allows the reader to get an idea of how formulations are widely dispersed in the chat sessions. The formulation in excerpt 2 occurs early on in the session.

Excerpt 2. (1844-4)

We see how coach Karin draws on the closing formulation and the client’s response to it in the last session. The client had not fully agreed with part of the closing formulation in responding that “'getting out' is not the only measure for my depressive feelings” (excerpt 1, line 12). The coach now reintroduces this topic (although she changes “not the only measure” to “not a good measure”). In doing so, the coach does not orient to reaching mutual understanding but uses the formulation as a prompt (Heritage, 1985) for the client to start telling more about “measures for how she is doing.”

2. Formulating to Clarify an Ambiguity

The second type of formulation we encountered in the data functions to clarify an ambiguity. This type does not occur very frequently, i.e., not in each session. Excerpt 3 is an example of such a ‘clarifying formulation,’ and it comes from the beginning of a chat session. Coach Mieke introduces the first topic, namely the emetophobia (a fear of being sick or other’s being sick) client Irma is suffering from, and which they had already talked about in the previous session.

Excerpt 3. (1503-3)

The interaction in excerpt 3 shows an interesting ambiguity; interesting in the sense that it appears to be related to the medium. The first turn by the coach starts the two lines of talk that ensue: Where the coach appears to be oriented to the topic of emetophobia, the client appears to be oriented to talking about her past week. Both these lines of interaction are introduced with the first turn: What appears to be a topic nomination (Button & Casey, 1984, 1985) to talk about emetophobia is followed by a question about how the client has been this past week. The coach does not explicitly connect these two parts of her turn, but if her question about how the client has been this past week was intended as a general question, the first part seems to have little function. In other words, the juxtaposition of the two utterances implies a connection. However, it is this second part of the coach’s turn the client takes up. The coach treats the client’s response “yes quite okay, but i now have more problems with nausea” as causing trouble in understanding – first in line 6, where the but-prefaced question (“But you say that it went quite okay this past week?”) marks the preceding turn as not unproblematic (Mazeland & Huiskes, 2001, p. 147), and again in line 9 when the coach formulates the client’s words in such a way that her turns in line 3-4 and 8 are now causally connected. The utterance “you mean” and the use of a question mark make a response explicitly relevant, thus aiming at re-establishing intersubjectivity. Both lines 6 and 9 suggest that the coach’s turn was intended to mean something like “how has your week been with regard to your emetophobia?” The formulation in line 9, then, reestablishes shared meaning between the coach and the client.

While in excerpt 3 the source of ambiguity was located in the coach’s turn, the next excerpt illustrates how the coach’s formulations orient to an apparent contradiction between what the client told her last week and what she says this week.

Excerpt 4. (1680-3)

In excerpt 4 the source of ambiguity is of two kinds. First, it is in the description the client gives as to whether or not she takes medicine. The client claims that she does not take medicine (line 2), a description that is not consistent with what had been discussed previously. This causes a breakdown in understanding for the coach, as shown in the clarifying formulation in line 6. A second source of ambiguity is the kind of medicine the client refers to in line 3. Apparently the client was talking about a different kind than the coach was. The client has noticed the ambiguity and makes another attempt to explain in line 8. The coach’s trouble in following the client has not been resolved, however, and she formulates the client’s words (line 9) and adds a question mark (line 10), making a confirmation by the client a necessary next action. Furthermore, the coach emphasizes her trouble at establishing intersubjectivity with the client in line 11 and explicitly mentions the medium as causing trouble.

This type of clarifying formulation is used frequently in our data. It is used to solve an ambiguity (e.g., the relation between the two parts of the coach’s turn in excerpt 3, or the description the client gives as to whether or not and which kind of medicine she is taking as shown in excerpt 4) and thus to re-establish intersubjectivity. The clarifying formulations in excerpt 4 appear to be similar to ‘checking formulations’ that have been found in courtroom proceedings (Van der Houwen, 2009). Judges may encounter discrepancies between what they already know from the case file and what they now hear in the courtroom. In doing checking formulations they orient to bridging a prior understanding with incongruent information coming out in court. In excerpt 4, too, the coach uses a formulation to adjust her prior understanding (from the previous session) to what the client now tells her. In excerpt 3, however, the formulation functions to clarify misunderstanding in interpreting turn adjacency; a misunderstanding which appears to have been caused, or at least exacerbated, by the medium. For chat interaction, it has been found that turn adjacency is sometimes disrupted when unrelated turns appear on the screen prior to the second pair part (Garcia & Jacobs, 1999; Herring, 1999; Panyametheekul & Herring, 2003). This has to do with the fact that unlike in spoken conversation, in chat interaction the production and the transmission of utterances do not occur simultaneously. So, the adjacency in chat is an achievement of the participants “reading and selecting from the stream of messages addressed to them” (Schönfeldt & Golato, 2003, p. 251). Previous studies of chat interaction have found that participants use various means to deal with discontinuities in the sequential organization: repair and sanctioning (Schönfeldt & Golato, 2003), or accounting, addressing recipients, recycling, and reformulating (Nilsen & Mäkitalo, 2010). Through such actions, intersubjectivity is (re-)established in the ongoing interaction.

3. Articulating a Particular View of the Client’s Account

A third type of formulation we found in our data appears to be used to make the conversation lean towards the therapeutic or propose an alternative view of the client’s account (cf. Antaki, 2008). This type occurs probably at least once in each chat session, and sometimes up to 10 times. Two subtypes can be distinguished: 1) formulations that emphasize the positive aspects of a client’s account and 2) formulations that formulate the client’s troubles in terms of feelings.

Formulating the Positive Aspects of the Client’s Account

Excerpt 5 is taken from the fourth session out of five with client Miriam (see also excerpt 1 and 2). In the previous session the client reported that she sometimes stays in people’s homes to take care of their cats while they are on vacation. The letters (A-C) and numbers (1-3) in the third column are used to specify turn adjacency for ease of reference in the analysis below.

Excerpt 5. (1844-4)

Prior to where the excerpt above begins, the participants have exchanged opening greetings. The coach initiates the first topic by showing an interest in the client’s past week (line 1).6 The client’s answer is moderately positive (“quite ok”)7 and is expanded with a description of a short stay in Amsterdam. The expansion after the evaluation “quite ok” suggests that this stay is explanatory of the mild improvement that the client has experienced. The coach’s next question selects the stay in Amsterdam as a topic (line 3). In response, the client produces a description including a positive evaluation (“I am more active there than at home”), feeding the suggestion that the improvement is indeed caused by the stay in Amsterdam. However, in line 5 she mitigates the positive meaning of “more active” with the description “there at least I get out,” suggesting that more active should be interpreted as going out (for some fresh air, groceries, or some such), but maybe not as doing all kinds of things in the city.8 Thus, the client portrays her stay in Amsterdam as a minor improvement only. This means that the client tones down the positive implication of “more active.”

Then (line 6), the coach formulates the client’s activity in Amsterdam: “done more things?,” not acknowledging the down toning of “there at least I get out.” The Dutch original of “done more things” implies visiting museums or other sites, taking long walks, etc. – at least more activity than “there at least I get out.” So, the formulation underlines the positive aspect (implied by line 4) to put it “on record,” despite the client’s suggestions that this aspect is not that positive (implied by line 5). We may call these formulations ‘optimistic formulations’ in reference to what has been described as ‘optimistic questions,’ i.e., questions about clients’ strengths, abilities, and successes with regard to the issues that brought them to therapy (MacMartin, 2008).9 This optimism lies in the emphasis on the client’s agency to achieve improvements, which is a presupposition of the formulation and made explicit in the assessment in line 7.

Excerpt 5 displays how the medium plays a recognizable part in the use of the formulation. As we have argued above, the question mark at the end of the formulation makes a confirmation relevant by the coach, and thus it seems designed to check her understanding of line 4. However, the coach does not wait for the client’s confirmation, because she produces a response to line 5. It seems that the sequential implication of the formulation (the preference for an agreement) is weakened as a result of the “virtual-adjacency” (Schönfeldt & Golato, 2003) of actions that characterizes chat. We will explain this using the letters A-C in the excerpt. The formulation is a second pair part (A2) to the utterance in line 4 (A1). Probably, the client was typing the extension to A1 (line 5) while the coach was typing the formulation (A2, line 6). After the coach has read this extension, she assesses it and thus treats line 5 as an independent action (B1) that calls for a second pair part (B2). So, after line 7, the client has two sequences to respond to, sequence A and B. Before the client produces a next turn, however, the coach initiates a new sequence with a question (C1), which leaves the client with three sequences to which to produce a relevant next action. Although Schönfeldt and Golato (2003, p. 252) have suggested that the obligation to complete a first pair part seems to be as strong in chat as it is in mundane conversation, we see a counterexample here. Sequence B and C receive second pair parts in lines 9-11, but the formulation does not receive a second pair part. Apparently, the virtual adjacency of turns in chat shapes a context in which the absence of a second pair part to a formulation can remain unproblematic. In chat, it is thus possible that the general requirement for formulations to be confirmed is disregarded (although we should note that this does not occur frequently in our data). We lack interactional evidence for analysis of the interactional consequences of this, but the absence of a response to the formulation suggests resistance, which would be in line with the disagreement displayed to the optimism in lines 9-11 (“well, yeah..,” “it’s more”).

Another example of the use of an optimistic formulation in chat is excerpt 6, taken from the first chat session between coach Mieke and client Jacob.

Excerpt 6. (1290-1)

*unrelated utterance, referring to previous topic

In excerpt (6), the client describes his problem (lines 3-5) and makes a distinction between his problems more generally (“just about everything bothers me and often ramble on”) and his feelings now while chatting with the coach (“i just meditated and have less a feeling that i feel bad”). In other words, as in the previous fragment, the client has presented his current state as relatively positive, while his general situation/state is negative. In line 7, the coach formulates this positive aspect only: “So meditating does you good?” The formulation not only ignores the negative general state of the client and emphasizes the positive aspect (from “have less a feeling that i feel bad” to “does you good”), but also implies that meditation (like “getting out” in excerpt 5) is a solution, a strategy that works. By formulating it, she puts it on record. As in excerpt 5, a question mark makes a confirmation by the client relevant. While in excerpt 5 the client neither affiliated nor disaffiliated, in extract 6 the client resists the formulation (line 8); extensive hedging (“well,” “actually,” “not really”) indicates the dispreferred-ness of this reaction. In line 9, the client modifies his previous turn by nuancing the positive effect of meditation to having an occasional positive effect. Then, the coach uses an acknowledgement token (“hmhm”), thus aligning as a recipient and inviting a more extensive account (Jefferson, 1985). So, the client’s resistance (lines 8-9) has indirectly offered him the interactional space to further elaborate on his problems that persist despite his incidentally successful coping strategies (lines 11-14). By displaying resistance, the client implies that his problems are serious enough to legitimate seeking help from the coach (Heritage & Robinson, 2006). In other words, the client resists the optimistic formulation, as it threatens the legitimacy of his receiving help.

Formulating the Problem in Terms of Feelings

Another context in which formulations are used to make the conversation incline towards the therapeutic or to propose an alternative view of the client’s account is by formulating the client’s trouble in terms of feelings. These formulations have also been found by Hutchby (2005) in a study of child counseling. By formulating the client’s trouble in terms of feelings, the counselor foregrounds therapeutically relevant matters. Unlike the optimistic formulation that indirectly sidestepped the seriousness of the problem (excerpt 6), this kind of formulation endorses the problem by treating it as therapeutically relevant. The following case shows such a formulation (excerpt 7). Prior to the question by coach Mieke in line 1, the client reported that she had a second job interview at a hospital, but was not selected.

Excerpt 7. 1008-4

Excerpt 7 starts with a question about the second job interview. The interaction then develops towards discussing a recurrent problem of insecurity as a result of someone’s “bitchiness” (lines 1-10). It is the client in her answers in lines 2-3 and lines 5-6 who initiates feelings-talk by describing herself as “off balance” and “insecure.” In line 11 the coach formulates “off balance” and “insecure” as “being blown off,” thus picking out feelings-talk as the relevant issue for discussion. In response, the client agrees and offers an elaboration that refers to another situation (“there is laughing”) and a related type of feeling: “i feel i’m almost made fun of” (line 12). The coach aligns with the client’s description, confirming that this is a terrible feeling (line 13), and continues to reformulate it as the feeling of “not being taken seriously.”

The design of the formulation in line 11 includes the written acknowledgement tokens “hmm, yes” (Jefferson, 1984). Rather than inviting further elaboration from the client, they are “thinking tokens” that, together with the formulation, function as a display of active listening (Danby et al., 2009). At the same time, however, the question mark following the formulation shows that in addition to displaying active listening, it also orients to the need for a confirmation by the client and thus to the client’s authority in being knowledgeable about her own feelings. In spoken interaction, an interrogative following a formulation can have this function (“Is that what you’re saying?” [Danby et al., 2009]), but in chat it can also be conveyed through a question mark.

In sum, in excerpt 7, the interaction centers on the client’s feelings, and the formulations underline the therapeutic relevance of this topic. The client has initiated the feelings-talk, the coach formulates it, and the client aligns with that. Thus, the participants follow a ‘counseling’ trajectory in the conversation that has to do with discussing feelings.

For our last example, we return once more to the fourth chat session between coach Karin and client Miriam. It illustrates the use of formulation, along with other practices that orient at reestablishing intersubjectivity. Excerpt 8 starts with the last six lines of excerpt 5.

Excerpt 8. (1844-4)

In extract 8, the coach and client appear to have different agendas. The client orients toward discussing the conviction “that it is not good to stay inside” (lines 11, 19, 22). The coach, in contrast, orients to an articulation of the client’s feelings (lines 10, 14, 16). We will analyze the excerpt sequentially to spell out how these conflicting agendas are enacted and how formulations play their part in it. In line 10, the coach initiates a new sequence with the question: “Does that also influence how you feel?” This shifts the topic from the degree of the client’s activity in Amsterdam to the client’s feelings. In response (lines 11-13), the client disaffiliates with the agenda set by the question (note the disagreement markers “well” and three dots in line 11):11 She thinks (rather than feels) that staying in is not so good and that her mind and ‘common sense beliefs’ are responsible for that. The reference to “what one is supposed to do” indirectly criticizes the conviction that “staying in is not so good.”

Subsequently (line 14), the coach produces another formulation with a question mark. It is designed as a clarifying formulation (see above), relating the client’s description that used “think” and “mind” (lines 11-13) to feelings, and thus clarifying how the client’s description could be understood as an answer to the question in line 10. However, the formulation leaves out the aspect of the client’s account that criticized the normative assumption about staying in. The design, including a question mark, makes relevant the production of agreement, but the client does not straightforwardly affiliate with the formulation (“less so” line 15, rather than “no”).

A further formulation by the coach in lines 16-17 characterizes a larger stretch of interaction and makes explicit the relation between the two conflicting components of the client’s account in lines 11-13: knowing and feeling.12 The formulation further establishes the positive effects of getting out as a fact and misrepresents the client’s doubts about this. This time (lines 18-20) the client uses her next turn to display some resistance (“something like that,” suggesting that the formulation was not precise), and extends the formulation using the conjunction “and.” In other words, she accepts the formulation and offers an addition that renders it a more accurate formulation of her account. However, the coach explicitly displays a lack of understanding of the reference “that” in the client’s extension of the formulation: “what are you resisting?” (line 21). The reference “that” (line 19) is unclear: It has no referent in the client’s turn (it could refer to the formulation in lines 16-17, but also to the client’s account in lines 11-13). This question is designed as an insertion sequence, not affiliating/disaffiliating with the formulation extension. The client’s answer to this insertion question constitutes another explanation of the client’s account (lines 22-23). In response, the coach makes a new attempt to capture it, this time in the form of an extension (line 24), which is another practice by which “therapists might display a grasp of, and present an alternative to, the client’s account of their experiences” (Antaki, 2008, p. 27). It transforms the client’s words (“imposed from outside”) into feelings-talk (“you have not yourself experienced it that way”), but designed as an extension, it adds to the client’s explanation and does not formulate it. The client’s next turn reveals that this extended account (lines 22-24) receives agreement, albeit restricted to “not at this moment”(lines 25-26).13

In sum, we have seen that the coach’s actions foreground the client’s feelings, an agenda which conflicts with the client’s agenda of discussing a conviction that bothers her. Among the actions the coach uses, formulations are a rather forceful tool that refers “the issue under discussion back to the [client]” (Hutchby, 2005, p. 321) and thus makes the interaction incline towards the therapeutic.

Who Does the Formulating?

Our analysis of formulations has focused on those done by coaches. In our corpus we have found that formulations are rarely done by clients. This is in line with what has been found in other institutional settings. One could conclude from these findings that it tends to be the professional or representative of an institution who does the formulating, and that formulations are an instrument of power (Hutchby, 1996). In general practice consultations, however, both patients and doctors do the formulating (Gafaranga & Britten, 2004). The act of formulating, they suggest, is not tied to an institutional role but rather to one’s interactional role, namely that whoever does the listening also does the formulating. We would argue that perhaps the types of formulation participants draw upon do show an orientation to one’s institutional role. The formulations we laid out in this study are all done by the coaches and are indeed displays of ‘active listening.’ One could wonder, though, if orienting to ‘being a listener’ can also be an orientation to power. While we would argue that this is not necessarily so with formulations that do clarifying work, it can be the case with, for instance, organizational formulations such as summarizing what participants have been talking about and introducing and closing topics, comparable to what chairs of meetings do (Huisman, 2001).

Conclusion and Discussion

In this article we have shown how formulations in the chat sessions we studied are used by coaches in three ways. First, formulations are used to organize the chat sessions as a whole and at a topical level. Second, formulations can be attempts to clarify ambiguity or an apparent contradiction. And last, formulations do work in articulating a particular view of the client’s account by either formulating the positive aspect of the client’s account (‘optimistic formulations’) or formulating the problem in terms of feelings (‘feelings formulations’).

Overall, we can conclude that formulations in counseling chat sessions are used in comparable ways to face-to-face therapy. The organizational type of formulation we found is similar to the use of formulation described by Antaki et al. (2005). Apparently, formulations can be used to manage the progress of the interaction in various directions: to open a certain topical direction or to close it. The type of formulation that articulates a particular view on the client’s problem is also found in face-to-face counseling (Antaki, 2008; Hutchby, 2005; Vehviläinen, 2003). It seems to depend on the model of counseling (psychoanalysis, constructive therapy) in which direction these go, varying from the verbalization of a layer of the subconscious meaning of the client’s talk (Vehviläinen, 2003) to an articulation of the optimistic aspects of a client’s account (cf. MacMartin, 2008). It appears that “feelings” formulations and optimistic formulations are frequently met with resistance, just as has been found with optimistic questions (MacMartin, 2008). However, in these cases it is hard to say whether it is the formulation itself that is resisted or rather the broader agenda (see also Hutchby, 2005). Overall, the similarities between face-to-face counseling and our chat data in the use of formulations suggest that online counseling is interactionally similar to traditional counseling, implying that the use of formulations endorses the conversational (Herring, 2010) character of chat.

There are, however, also some characteristics of formulations that we found that appear to be specific to the medium of chat. First, the type of formulation that clarifies an ambiguity in the immediately prior utterance (type 2 in our analysis) has not been identified in face-to-face counseling. It resembles practices found in other chat settings that deal with a discontinuity or trouble in understanding in the interaction (Nilsen & Mäkitalo, 2010; Schönfeldt & Golato, 2003). Thus, we may add formulations as another strategy that aims at solving trouble and re-establishing intersubjectivity in chat.

Second, we found that question marks are specific to the design of formulations in chat. We pointed out that punctuation marks in chat sessions are more akin to those used in written language (in contrast to punctuation in transcripts of spoken interaction). Question marks are visible to the participants in chat and, arguably, strong signals for the other participant to produce a next action. In our data, formulations tend to be followed by question marks. We would argue that this design feature makes formulations a step in the negotiation of meaning leading to intersubjectivity. Danby et al. (2009) found that formulations in their chat data were ambiguous. In our data the question marks appear to solve some of this ambiguity, since the question mark at least disambiguates whether or not a second pair part is expected. Future research should further examine the question mark and other punctuation marks as tools in the construction of actions and their sequential implications in chat interaction.

Third, we found that the disrupted adjacency of pair parts in chat may reduce the necessity for agreement as a second pair part to a formulation (cf. Schönfeldt & Golato, 2003), even when a slot for a second pair part, unlike in conversation, is in principle always available. This implies that the medium of chat, because of the ambiguous display of turn adjacency (Herring, 1999), is more permissive of ignoring and, arguably, resisting the conditional relevance of a formulation.

To conclude, the three types of formulation identified in this study seem to appeal to different levels of intersubjectivity. While organizational formulations set a certain agenda and thus structure the interaction, the second and third types of formulation we identified more explicitly negotiate meaning. Formulations that clarify ambiguities operate on the level of wording. They restore a discontinuity in understanding and are generally confirmed or followed by an explicit account that solves the misunderstanding. The third type, formulations that articulate a particular view of the client’s account, seems to function as perspective-taking, and is generally followed by elaborations or further negotiation. It is probably this type of formulation that most explicitly brings out the differing orientations of professional and lay participants to the interaction and that aims at resolving the interactional trouble the different perspectives may cause.


  1. “A member may treat some part of the conversation as an occasion to describe that conversation, or explain it, or characterize it, or explicate, or translate, or summarize, or furnish the gist of it, or take note of its accordance with rules, or remark on its departure from rules. A member may use some part of the conversation as an occasion to formulate the conversation” (Garfinkel & Sacks, 1970, p. 350).

  2. An adjacency pair consists of two turns by two different speakers. The turns are functionally related to each other in such a fashion that the first turn (also called ‘pair part’) requires a certain type or range of types of second turn (questions require answers, greetings require return greetings, etc.).

  3. The analysis was based on the Dutch original excerpts. Furthermore, the original “enters” that relate to the different chat programs (see the Data section) have been preserved and were not edited by us.

  4. For the original excerpts, contact the first author.

    While some of the logs have time indications, unfortunately, not all sessions do. Where time indications are available, these are shown in the fragments used to illustrate our analysis. See also the Data section.

  5. Note that this utterance is not in the form of a question. By emphasizing curiosity, the utterance displays informality and a certain urgency to know, as opposed to a “howareyou” utterance that is usually produced after greetings in mundane phone conversation (Schegloff, 1986).

  6. Similar to how “pretty good” has been described as a downgraded conventional response (Jefferson, 1980), “quite ok” is an upgraded response, at least when the answer should be considered an “update” answer, that is, when the person who asks the question is already acquainted with the responder’s troubled situation. In other words, “quite ok” constitutes a mild improvement.

  7. Furthermore, the present tense in lines 4-5 describes how the client usually acts when she stays in Amsterdam, not specifically this time. This also implies that at home she rarely gets out. From a membership-categorization perspective (Schegloff, 2007), this non-activity points to the category of “psychologically troubled persons,” which legitimates this client for the treatment she is receiving.

  8. MacMartin (2008) associates these optimistic questions with solution-focused therapy. This type of clinical model is also interwoven into the coaching instructions, which might explain the occurrence of these formulations in the chat sessions.

  9. The original Dutch utterance, En je dat niet zelf zo hebt ervaren?, is hard to translate into English. Originally, we translated ervaren as ‘experienced,’ However, on second thought we decided that the meaning of the Dutch ervaren is closer to the English felt than experienced. Therefore, we changed our translation.

  10. In the context of excerpt 5, we analyzed lines 12-13 as responding to the question in line 10, and line 11 as responding to line 9. However, the direct adjacency of the utterances in lines 10 and 11 and the fact that line 11 does not overtly relate to the assessment in line 9 means that line 11 can also be understood (by us as analysts, but also by the coach) as a response to the question in line 10.

  11. Note also that the formulation in lines 16-17 transforms the client’s description by using the word “knowing” rather than the client’s “think” (line 11) and “mind” (line 12). Thus, the formulation foregrounds the issue as factual knowledge, rather than an opinion/conviction (cf. Hutchby, 2005, p. 325).

  12. The client’s extended response (line 26) also displays some resistance to talking about feelings: “it seems that” in line 26, rather than a direct report of her experience “i feel better staying inside.”


Antaki, C. (2008). Formulations in psychotherapy. In A. Peräkylä, C. Antaki, S. Vehviläinen, & I. Leuder (Eds.), Conversation analysis and psychotherapy (pp. 26-42). Cambridge: Cambridge University Press.

Antaki, C., Barnes, R., & Leudar, I. (2005). Diagnostic formulations in psychotherapy. Discourse Studies, 7(6), 627-647.

Beach, W. A., & Dixson, C. N. (2001). Revealing moments: Formulating understandings of adverse experiences in a health appraisal interview. Social Science & Medicine, 52, 25-44.

Button, G., & Casey, N. (1984). Generating topic. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action: Studies in conversation analysis (pp. 167-190). Cambridge: Cambridge University Press.

Button, G., & Casey, N. (1985). Topic nomination and topic pursuit. Human Studies, 8, 3-55.

Danby, S., Butler, C. W., & Emmison, M. (2009). When 'listeners can't talk.’ Comparing active listening in opening sequences of telephone and online counselling. Australian Journal of Communication, 36(3), 91-113.

Davis, K. (1984). The process of problem (re)formulation in psychotherapy. Sociology of Health and Illness, 8, 44-74.

Drew, P. (2003). Comparative analysis of talk-in-interaction in different institutional settings: A sketch. In P. J. Glenn, C. D. LeBaron, & J. Mandelbaum (Eds.), Studies in language and social interaction: In honor of Robert Hopper (pp. 293-308). Mahwah, NJ: Erlbaum.

Drew, P., & Heritage, J. (1992). Analyzing talk at work; an introduction. In P. Drew & J. Heritage (Eds.), Talk at work: Interaction in institutional settings (pp. 3-65). Cambridge: Cambridge University Press.

Flicker, S., Haans, D., & Skinner, H. (2004). Ethical dilemmas in research on Internet communities. Qualitative Health Research, 14(1), 124-134.

Gafaranga, J., & Britten, N. (2004). Formulation in general practice consultations. Text, 24(2), 147-170.

Garcia, A. C., & Jacobs, J. B. (1999). The eyes of the beholder: Understanding the turn-taking system in quasi-syncronous computer-mediated communication. Reseach on Language and Social Interaction, 32(4), 337-367.

Gezondheid dichtbij; landelijke nota gezondheidsbeleid [Health closeby; national report health policy]. (2011). Retrieved May 5, 2012 from from http://www.rijksoverheid.nl/documenten-en-publicaties/notas/2011/05/25/landelijke-nota-gezondheidsbeleid.htmll

Hak, T., & de Boer, F. (1996). Formulations in first encounters. Journal of Pragmatics, 25(1), 83-99.

Heritage, J. (1985). Analyzing news interviews: Aspects of the production of talk for an overhearing audience. In T. van Dijk (Ed.), Handbook of discourse analysis (vol. 3, pp. 95-119). London: Academic Press.

Heritage, J., & Clayman, S. (2010). Talk in action: Interactions, identities, and institutions. Chichester: Wiley-Blackwell.

Heritage, J., & Robinson, J. (2006). Accounting for the visit: Giving reasons for seeking medical care. In J. Heritage & D. Maynard (Eds.), Communication in medical care: Interaction between primary care physicians and patients (pp. 48-85). New York: Cambridge University Press.

Heritage, J., & Watson, D. (1979). Formulations as conversational objects. In G. Psathas (Ed.), Everyday language (pp. 123-162). Hillsday, NJ: Lawrence Erlbaum.

Herring, S. C. (1999). Interactional coherence in CMC. Journal of Computer-Mediated Communication, 4(4). Retrieved March 12, 2013 from http://jcmc.indiana.edu/vol4/issue4/herring.html

Herring, S. C. (2010). Computer-mediated conversation: Introduction and overview. Language@Internet, 7, article 2. Retrieved March 12, 2013 from http://www.languageatinternet.org/articles/2011/Herring

Huisman, M. (2001). Decision-making in meetings as talk-in-interaction. International Studies of Management & Organization, 31(3), 69-90.

Hutchby, I. (1996). Confrontational talk: Arguments, asymmetries, and power on talk radio. Mahwah, NJ: Erlbaum.

Hutchby, I. (2005). “Active listening”: Formulations and the elicitation of feelings-talk in child counseling. Research on Language and Social Interaction, 38(3), 303-329.

Jefferson, G. (1985). Notes on a systematic deployment of the acknowledgement tokens "Yeah" and "Mm hm.” Papers in Linguistics, 17(2), 197-216.

Lamerichs, J., & Te Molder, H. (2003). Computer-mediated communication: From a cognitive to a discursive model. New Media & Society, 5(4), 451-473.

MacMartin, C. (2008). Resisting optimistic questions in narrative and solution-focused therapies. In A. Peräkylä, C. Antaki, S. Vehviläinen, & I. Leudar (Eds.), Conversation analysis and psychotherapy (pp. 80-99). Cambridge: Cambridge University Press.

Mazeland, H., & Huiskes, M. (2001). Dutch ‘but’ as a sequential conjunction. Its use as a resumption marker. In M. Selting & E. Couper-Kuhlen (Eds.), Studies in interactional linguistiscs (pp. 141-169). Amsterdam/Philadelphia: John Benjamins.

Nilsen, M., & Mäkitalo, A. (2010). Towards a conversational culture? How participants establish strategies for co-ordinating chat postings in the context of in-service training. Discourse Studies, 12(1), 90-105.

Panyametheekul, S., & Herring, S. C. (2003). Gender and turn allocation in a Thai chat room. Journal of Computer-Mediated Communication, 9(1). Retrieved March 12, 2013 from http://jcmc.indiana.edu/vol9/issue1/panya_herring.html

Peräkylä, A., Antaki, C., Vehviläinen, S., & Leudar, I. (2008). Conversation Analysis and psychotherapy. Cambridge: Cambridge University Press.

Riper, H., Smit, F., Zanden, R. v. d., Conijn, B., Kramer, J., & Mutsaers, K. (2007). E-Mental health: High tech, high touch, high trust programmeringsstudie e-mental health. Utrecht: Trimbos Instituut.

Schegloff, E. (1968). Sequencing in conversational openings. American Anthropologist, 70(6), 1075-1095.

Schönfeldt, J., & Golato, A. (2003). Repair in chats: A conversation analytic approach. Research on Language and Social Interaction, 36(3), 241-284.

Stommel, W., & Koole, T. (2010). The Online Support Group as a community: A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357-378.

Van der Houwen, F. (2009). Formulating disputes. Journal of Pragmatics, 41(10), 2072-2085.

Vayreda, A., & Antaki, C. (2009). Social support and unsolicited advice in a bipolar disorder online forum. Qualitative Health Research, 19(7), 931-942.

Vehviläinen, S. (2003). Preparing and delivering interpretations in psychoanalytic interaction. Text, 23(4), 573-606.

Biographical Notes

Fleur van der Houwen [f.vander.houwen@vu.nl] is Assistant Professor in the Department of Language and Communication at VU University Amsterdam, where she does research in the area of institutional interaction.

Wyke Stommel [w.stommel@let.ru.nl] is Assistant Professor at the Center for Language Studies, Radboud University Nijmegen. In her research she uses Conversation Analysis and discourse analysis to study interactional aspects of online health communication in institutional and lay settings.


Any party may pass on this Work by electronic means and make it available for download under the terms and conditions of the Digital Peer Publishing License. The text of the license may be accessed and retrieved at http://www.dipp.nrw.de/lizenzen/dppl/dppl/DPPL_v2_en_06-2004.html.