This chapter aims to promote dialogue between discursive-oriented researchers (e.g., Potter & Hepburn, 2007) and therapists (e.g., Lock & Strong, 2012). Both share an interest in therapy talk as “reflexive”: they approach discourse as a form of joint action rather than as information that goes through a tube (e.g., Edwards, 1997; Tomm, 1988). At the same time, they clearly differ in aims and “methods” of inquiry. Discursive therapists seem to be interested in optimizing therapy and, therefore, deliberately using reflexivity in client-responsive ways. They do it by theorizing about therapy practice using (theory-charged) concepts such as “clients’ preferences.” They may talk of “absent but implicit” values (White, 2003); “preferred” identities and other developments in clients’ lives (e.g., Freedman & Combs, 1996); collaborative or subject-subject type of therapeutic relationships (e.g., Weingarten, 1992); and clients’ autonomy (Tomm, 1988). On the other hand, discursive-oriented researchers are less concerned with abstract concepts depicting therapy talk and its “outcomes” and more with how participants in a specific interaction use language , in observable ways, and for what local social purposes.

In this chapter, we explore how drawing on both discursive (therapeutic and research) perspectives may help deepen understanding and enrich practice of discursive therapy . Our focus is on a specific discursive therapy practice—reflexive questions (Tomm, 1987b, 1988). Reflexive questions can be used to constitute “preferred” descriptions and understandings of clients’ relationships, identities, and actions. Drawing from discursive researchers, specifically discursive psychologists, we suggest that reflexive questions invite conversational “attribution work.” Discursive psychologists approach attributions as talk-in-action rather than as individual cognitive processes. In particular, they investigate how psychological matters (e.g., emotions , cognitions, identities) are formulated, ascribed, and resisted by people engaged in social interaction (te Molder, 2015). Re-conceptualized in discursive terms, this attributting to people of traits, motives, affective states, resposibility, etc. (i.e., “attribution work”) becomes empirically accessible to discursive therapists interested in co-constructing clients’ identity and experience in client-responsive ways. In other words, a discursive-oriented lens on reflexive questions may enhance therapists’ capacity to learn from and teach therapy practice: to learn, because it helps understand how discursive therapists use reflexivity therapeutically by asking reflexive questions and subsequently responding to clients’ answers to accomplish clients’ preferred versions of their experiences. To teach, as these investigations may help offer some ideas about how to make the use of such discursive reflexivity a more deliberate practice.

Enacting discursive reflexivity  by asking reflexive questions invites a heightened ability to notice and orient to how the co-construction of alternative (deemed as preferred or helpful) meanings gets done, allowing therapists to participate in the therapeutic endeavor both responsively and collaboratively. Our main suggestion is that reflexive questions invite preferred attribution work. First we will elaborate on this theoretically, by unpacking the notion of reflexivity within both discursive practice and discursive research traditions. Then we will analyze segments of therapy talk to explore empirically the same idea.

Therapeutic Perspective on Reflexivity and Reflexive Questions

For therapists, everything they do in relation to clients is typically seen as a form of communication. Therapists cannot help but behave; therefore, they cannot help but communicate. This is, as Watzlawick, Beavin, and Jackson (1967) put it half a century ago, an “axiom” not just for therapy interaction, but for all human communication. As therapists, we may either consider seriously or ignore this axiom but, arguably, our stance will affect how the therapeutic conversation unfolds. This idea aligns with Karl Tomm’s (1987a, 1987b, 1988) notion of “reflexivity” in his interventive interviewing framework for therapy and the first quote at the beginning of this chapter. Tomm proposed that “everything an interviewer does and says, and does not do and does not say is … an intervention that could be therapeutic, nontherapeutic, or countertherapeutic” (Tomm, 1987a, p. 4). For Tomm, even questions intended as mere information-gathering tools are interventions and will influence how the conversation unfolds, including the topics, activities, and presuppositions pursued and advanced in therapy. Regardless of intentions, every question a therapist offers favors and subtly advances particular understandings (e.g., ideas, actions, ways of being and relating). As a result, only certain topics, issues, or concerns get talked into being, and not others. We see meanings or understandings not as lifted from one person’s mind and transferred to another person’s mind but as jointly and interactionally worked up. As we discussed elsewhere (Sutherland, Sametband, Gaete, Couture, & Strong, 2013), such understandings are discursively accomplished as people produce, coordinate, and negotiate meaning through their interactions (see also, e.g., Berger & Luckmann, 1991; Shotter, 2008).

Tomm (1987b) suggested that therapists who seek to bring forth alternative possibilities for understanding and action may find reflexive questions useful. He defined reflexive questions as:

Questions asked with the intent to facilitate self-healing in an individual or family by activating the reflexivity among meanings within preexisting belief systems that enable family members to generate or generalize constructive patterns of cognition and behavior on their own. (p. 4)

Some examples of reflexive questions include: “What are you worried might happen when your daughter stays out late?… What is the worst thing that comes to mind?” or “If you did raise these worries with her, do you think she would take it as a lack of trust? … As an intrusion into her privacy? … Or as an indication of your caring as a parent?” (Tomm, 1987b, p. 5). Reflexive questions may invite the construction of potentially useful understandings and courses of action previously not considered (e.g., parents disclosing their worries to the daughter rather than being critical of her behaviors). They can be used therapeutically to both modify (unhelpful or problematic) and/or stabilize (preferred) understandings, which in turn may modify/stabilize the practices and relationships sustained by such understandings (Tomm, 1987b).

Reflexive questions may help clients change their views and everyday practices and relationships (Tomm, 1987b). Linguistic distinctions embedded in questions and taken up in answers are, in this sense, political. As co-created, mutually acceptable conversational realities, they have power over those who generate and accept them as apt to understand or act within their world (see also Strong, 2007). Eventually, most therapists hope that clients will engage with these powerful distinctions (e.g., endorse, live by, apply) beyond the therapy room: this is how the constructive discursive process of therapy is typically expected to be influential in clients’ extratherapeutic lives (Dreier, 2015), and this is how we interpret Tomm’s idea of reflexive questions as facilitating “self-healing.” For Tomm (1987a, 1987b), then, the associated therapeutic “effects” in these questions lies in two complementary sources of reflexivity, namely, (a) in how therapists’ talk inevitably shapes clients’ responses (and vice versa) and (b) in the shifting in meaning these questions trigger within clients’ “belief systems.”

Two Therapeutic Effects: Preferred Descriptions and Clients’ Autonomy

With respect to the effects of reflexive questions, Tomm (1988) argued that both therapists and clients are likely to experience these questions as generative. Reflexive questions can help open space for new possibilities and help generate meanings /accounts that are deemed preferred (e.g., helpful, fitting, important) than previously held views (Tomm, 1988). These “preferred” realities include preferred descriptions of identities, practices, relationships, values, or emotions (e.g., Andersen, 1987; Combs & Freedman, 2012; Freedman & Combs, 1996; White, 2007; White & Epston, 1990). Accordingly, reflexive questioning can be seen as discursively accomplishing “preference work” in the sense that they help craft and invite continued co-construction of distinctions and meanings regarded as more preferable or helpful, as compared to prior meanings (Sutherland et al., 2013).

Reflexive questions may also have an effect of fostering client autonomy (Tomm, 1988). Proponents of various approaches to therapy increasingly strive to promote clients’ participation, active engagement, self-exploration, and agency or capacity to self-heal or generate change on their own (see Bohart & Tallman, 2010; Gendlin, 1996; Gonzalez, 2016; Greenberg & Pinsoff, 1986; Orlinsky, Ronnestad, & Willutzki, 2004; Prochaska, DiClemente, & Norcross, 1992). Similarly, but invoking different reasons, discursive therapists have depicted therapy as a collaborative activity, respectful of clients’ knowledge and agency , something that is done with rather than to clients (e.g., De Jong & Berg, 2002; Paré, 2012; Weingarten, 1992). Be it for pragmatic, ideological, or ethical reasons, there seems to be a growing consensus concerning the importance of relating to clients as subjects/agents (rather than objects/patients) (Buber, 1970; Gadamer, 2004) and treating clients as individuals whose “lay” accounts on therapy and experience are comparable in legitimacy to professional explanations. Arguably, reflexive questions can be used to invite this type of subject-subject relationship between clients and therapists, wherein clients’ knowledge is invited and treated as relevant and consequential and clients are seen as agents and active contributors of ideas rather than passive recipients of expert knowledge.

Tomm (1987b) suggested that therapists cannot determine clients’ responses, because clients inevitably respond out of their own discursive capacities and background of meanings . Later Tomm (1988) contended that reflexive questioning invites clients to “make pertinent discoveries on their own” (p. 2). As a consequence, therapists may become not just more influential or effective, but more collaborative and ethical or, as Tomm (1988) put it, “their influencing intent is moderated by respect for the autonomy of clients” (p. 8). Tomm thus portrayed therapists’ reflexive questions as deliberatively facilitative, enabling, or triggering rather than instructive, linear, or causally deterministic of clients’ responses. The conceptual picture of reflexive questions clearly sets the stage for subject-subject therapeutic relationships and interactions to flourish.

To summarize, “reflexivity ” in Tomm’s therapeutic conception of reflexive questions could be described as having two key effects (which purportedly optimize the therapy process): they help generate alternative (preferred) understandings and they enable client autonomy or self-exploration. Let us now examine these effects from a discursive psychological research perspective. We will argue and demonstrate that discursive psychology can help bring to light features of therapy discourse and questioning practices that may be therapeutically relevant.

Discursive Psychological Perspective on Reflexive Questions and Their Effects

Discursive psychology (DP) (Edwards, 1997; Edwards & Potter, 1992; Hepburn & Wiggins, 2005; Potter, 2010; Tileagă & Stokoe, 2016; Wiggins, 2017) is an approach to the study of psychological phenomena (e.g., attribution , emotion , agency ) from the perspective of participants in social interaction . DP scholars also see language use as “reflexive”: not as transmitting information about people’s mental life or the world but as a tool to accomplish social action. Claims of uncertainty (“I don’t know”), for example, could be used in certain contexts to play down one’s stake in a specific description (Potter, 1996). Complaints can be done in ways that manage inferences about the complainer (Edwards, 2005).

Drawing from DP , we approach the study of reflexive questions as interactional accomplishments. Descriptions are designed with the audience “in mind” or to be acceptable by their recipients. Discursive researchers can help understand how, specifically, affiliate with certain descriptions and how certain descriptions or “versions” of reality are locally established as more or less (mutually) acceptable. They can shed light on the rhetoric involved in issuing and responding to accounts offered in therapy, envisioning preferred or more acceptable meanings as a practical matter and discursive achievement of therapists and clients. From this perspective, the attention shifts away from preference as a matter that is internally formed (cognitive) and then outwardly expressed to preference as an interactional accomplishment.Footnote 1

For the purposes of this chapter, we focus on attribution . Attribution is a form of describing concerned with how people make causal explanations of actions and events. Participants may invoke intents, interests, and motives or other explanatory devices of action as resources to accomplish social activities, such as blaming, inviting, or defending (Edwards & Potter, 1992). Attributional explanations do not have to be uttered directly (e.g., “he is to blame for our marital problems”) and can be accomplished implicitly, for example, implied through factual descriptions of actions and events (e.g., “he’s been working 24/7”). Speakers can attribute or ascribe certain traits or characteristics, sometimes in subtle or implicit ways, to themselves or others (te Molder, 2015). Issues of blame and accountability can be highly relevant in therapy, particularly in couple and family therapy where family members may have different perspectives on who is responsible for the “problem” and who needs to change (e.g., O’Reilly, 2014; Patrika & Tseliou, 2016; Stancombe & White, 2005). Applying a DP perspective to reflexive questions, such questions may be seen as ways to accomplish important attribution work—ascribe certain attributes to people or attribute causal sources of actions and events to certain sources (e.g., external or internal, situational or personal, temporal or permanent) in an effort to perform therapy-relevant social actions (e.g., blaming, justifying, excusing, requesting, complementing). We will suggest discursive-oriented therapists often see this type of work as co-constructing “preferred meanings” of their clients’ life circumstances and identity (Tomm, 1987b, p. 4).

Let us revisit Tomm’s (1987b) therapeutic conception of reflexive questions in an effort to recast it discursively. Tomm defined reflexive questions as:

Questions asked with the intent to facilitate self-healing in an individual or family by activating the reflexivity among meanings within preexisting belief systems that enable family members to generate or generalize constructive patterns of cognition and behavior on their own. (p. 4)

DP scholars may argue that there are several features in this definition, such as the therapist’s “intent” or clients’ “self-healing” or “preexisting belief systems,” that might be inconsistent with a DP perspective focused on social interaction rather than individual cognitive processes. As aforementioned, we see reflexive questions as opening space for co-construction and negotiation of “new” and mutually acceptable/preferred meanings and descriptions. Thus, the “reflexivity” of reflexive questions is not intrinsic to these questions, as suggested by Tomm (1987b), but is interactionally achieved. If we look back to our initial quotes, this conceptualization falls more in line with the second quote where reflexivity is accomplished as therapists become more responsively involved in the back and forth invited by these questions. Reflexive questions, as we will show, may embed “optimistic” presuppositions (MacMartin, 2008) or views of clients as positive people—resourceful, competent, and agentic. In their responses to reflexive questions, clients may take up (endorse, unpack, extend, etc.) such presuppositions. Accordingly, we would like to reformulate reflexive questioning in discursive terms as questions that elicit, clarify, and unpack clients’ reasoning—their explanations of and reflections on their own experiences. As aforementioned, reflexive questions can (a) make available and ascribe to clients (through referring, hinting, evoking, etc.) certain psychological characteristics, or (b) explain clients’ experiences and actions in particular ways (e.g., as caused by internal or external factors)—attributions which can then be take up or resisted by clients in their responses to questions.

We contend that our discursive-oriented reformulation helps make reflexive questions’ constructive potential open to view, “since everything lies open to view there is nothing to explain” (Wittgenstein, 1953, p. 126). The “hidden” magic of the meaning-making process gets observable and understandable and, hence, teachable and learnable. Rather than speculating upon therapists’ intent, as discursive researchers, we prefer to look for what observably happens in interaction as a consequence of therapists asking certain questions. We are interested in what happens after questions are posed, including how clients take up (or not) therapists’ presuppositions and how therapists respond to clients’ “resistant” responses to questions (de Shazer, 1984) observable, for instance, in clients’ non-answers (e.g., MacMartin, 2008). From a micro-discursive perspective, we should be able to observe how clients and therapists mutually orient to the relevance of interactionally “centering” clients’ reflections and explanations. Moreover, we should be able to see how alternative understandings of clients’ experiences and identities—theorized effects of reflexive questions—are generated through social interaction.

Analyzing the Effects of Reflexive Questions

The exemplars below illustrate how reflexive questions can be examined discursively to better appreciate how discursive therapists and clients manage their responses to each other to stay in an intended zone of mutual preference. Exemplar 1a was taken from a family therapy process with Antonia, José, and their two children, Alex (17) and Sofía (12).Footnote 2 Exemplar 1a comes from a session with the parents alone. They had previously stated they wanted to address “violence issues” between them, to better address disruptive behaviors in their children, which was the original complaint. Prior to Exemplar 1a, the therapist (T) and José (J) had been talking about his “violent behavior.” J shared an account for his violent behavior toward Antonia (A), stating that A usually insults him when she drinks too much alcohol, and that her alcohol consumption makes him “angry,” makes him “lose control,” and that he “reacts” with violence toward A. In line 114, T starts a sequence of reflexive questions about J’s reaction. What makes these questions reflexive is that they elicit further explanations and reflections from the client. We can also observe how alternative (positive or non-pathologizing) distinctions or constructions of the client as a person and partner are produced through question-answer sequences of actions. In particular, we can see how T and J attribute certain attitudes and predispositions to J. Reflexive questions are highlighted (bolded) in each exemplar.

Exemplar 1a

114 T:

did you ever try to do something to not have the reaction you have

115 :

when Antonia drinks?

116  J:

yes

117 T:

what did you try?

118  J:

stop the fight and go away

In lines 114–115 and 117, T issues reflexive questions that elicit information from J concerning his prior attempts to react non-violently toward A. T’s questions in lines 114–115 and in 117 are presuppositional, that is, they tacitly advance certain assumptions about J’s actions or attribute certain motives or intentions to J, namely that on rare occasions (“did you ever”) he may have been doing something to prevent his violent reactions. In seeking, with a yes/no question (Raymond, 2003), J’s confirmation that he has made attempts to abstain from violence (lines 114–115), T implicitly proposes that J’s violent reactions are inappropriate or problematic (consider that any direct answer to a question treats the question’s presuppositions as valid; Hayano, 2013; MacMartin, 2008). The use of vague or neutral language (“reaction”), as opposed to more specific, morally charged terms (e.g., “violence” or “abuse”), may be a way for T to attribute responsibility to J without being heard as critical or judgmental.

J’s reactions are constituted as temporary and event-driven (“when Antonia drinks”) rather than dispositional (e.g., J being an angry or violent person) (Edwards, 1999), which may be a way to present the “problem” as changeable and amenable to a therapeutic intervention. J’s “reactions” are also presented as controllable or preventable (Edwards, 1999). J is distinguished not as a passive, non-agentic man who can do little about his reactions, but as a someone who has tried reacting non-violently. Once J confirms that he has attempted to act non-violently, T’s information-seeking wh- (what) question (line 117) elicits details about J’s prior attempts. In response, J endorses and exemplifies the proposed construction of him as an agentic subject—someone who has made efforts to act non-violently.

We can see how T and J orient to issues of blame and responsibility. This is evident in T advancing the notion that J is responsible for violence or that violence is morally problematic without sounding critical or judgmental, or in J joining the construction of himself as actively seeking to change his behavior. Responsibility and agency are attributed to J, that is, he is implicitly presented as responsible for his reactions and as actively attempting to act non-violently. J endorses the attribution of the motive to treat his partner in more caring, non-violent ways by supplying evidence of his attempts. Let us explore, in Exemplar 1b, how this interaction unfolds. Here, we see how an alternative description of the client’s identity is offered (and endorsed) using a reflexive question.

Exemplar 1b

119 T:

and how did Antonia respond?

120  J:

she follows me

121 T:

she follows you?

122  J:

((nods))

123 T:

so what happens to you, what happens that makes you want to leave,

124 :

what is it that happens when you want to leave?

125  J:

so that I don’t hit her

126 T:

that seems important to you not hitting her?

127  J:

yes ((nods))

128 T:

why is it important, José?

129  J:

because I think I still love her

T issues reflexive questions that elicit explanations (why… because…; line 128) and information from J concerning his intentions and values (line 123–124 and 126). J is mutually treated as more knowledgeable than T (Heritage, 2013) concerning J’s mental life (motives, values, attitudes, etc.); it is his reflections and explanations that are sought and unpacked in this exchange. T’s first question (line 124) recycles the information previously supplied by J (“stop the fight and go away,” line 118). J is treated as more knowledgeable as compared to T about what promoted him to leave. T’s question is reformulated mid-turn (lines 123–124) from “what happens that makes you want to leave” to “what is it that happens when you want to leave,” which may be a way to constitute J as an agent with positive intentions, rather than a passive victim of his circumstances. T’s question is formulated as variably interpretable: as eliciting information concerning, among others, Antonia’s actions, other contextual factors, or J’s emotional or mental states (e.g., feelings, intentions, thoughts, wishes, needs, etc.). Analytically, T’s “actual” intentions in asking the question are irrelevant, as our analysis focuses on interactionally displayed interpretations of conduct—how T’s question is “heard” or interpreted by J, which is observable in how J responds to the question. Line 125 shows that J hears T’s question as being about intentions behind J’s “wanting to leave” (“so that I don’t hit her”): J’s leaving is constituted as an outcome of his (caring) intent to not hit A.

T’s next declaratively formulated yes/no question (line 126) seeks confirmation from J of T’s proposition that the intent to “not hit” is of significance to J, making available the inference concerning J being a certain kind of person (someone who is kind and concerned about his partner), the construction confirmed by J in the next turn (line 127). That is, the question implicitly assigns positive attributes to J. The question is mitigated with “seems” to downgrade T’s knowledge of J’s inner experience, namely, his values.

In lines 114–129 T issues a series of five reflexive questions. After accomplishing a description for a past event where J tried to behave differently (“stop the fight and go away”), T’s questions invite J to account for this “exception,” which they do by making relevant his “inner experience” or “values”: both T and J accept this version where he tried, where he “wanted to leave,” because he had “motive” (“so that I don’t hit her”). T’s last question (line 128) displays acceptance of J’s ongoing version by inviting further accounting work (T elicits an explanation from J of the importance of not hitting A). J cites his continued (still) love for A as the reason, downgrading the certainty of his claim with I think. Issuing a non-committal or uncertain claim of love presents J as someone who has just realized that his efforts to not be violent might be motivated by his love for A. If he claimed his love for A more confidently, he may be accused of acting inconsistently with his claimed feelings for A. J’s responses to T’s questions help extend the construction of himself as a good person and partner and, perhaps, block the need to defend himself against the accusation that he is a violent person and “bad” partner. J’s responses in lines 127 and 129 mark his affiliation with or endorsement (“yes,” nod, unpacking with “because…”) of T’s claim that being non-violent carries importance for J. Here again, the therapist and the client co-construct (and implicitly accept, validate) an alternative account of J’s identity through the use of reflexive questions.

We use Exemplar 1c to focus on the last two turns of Exemplar 1b (lines 128–129) in order to illustrate how reflexive questions may elicit “emotionally charged” accounts , which may in turn accomplish important attributional work (e.g., attribute certain attitudes or predispositions). This may be discursively “done” through very subtle features of talk, such as recognizable displays of upset. We use a more detailed transcription system here to include paraverbal and non-verbal aspects of talk (see the footnote for transcription notation).

Exemplar 1c

128 T:

why is it important, José? [(10)          ]

129  J:

        [(smiles, deeply inhales and exhales)]

130 :

because I think I still love her = (wobbly voice, aspiring, sniffingFootnote 3)

131 T:

=Mh hm (6) how important (nods).

132  J:

(nods) (3)

133 T:

°Mmh° (3)

After T’s question in line 128, we witness an enormous gap in the dialogue of ten seconds. In response, J performs several features of crying: whispering (note the ° ° sign), wobbly delivery, sniffing, deep breaths (Hepburn & Potter, 2012). T exhibits interest in J telling more with pauses and acknowledgment tokens (lines 131 and 133) (Jefferson, 1984). This “interested” response seems to validate (accept) J’s performed, “emotive” explanation in line 130. T seems to recognize or “hear” (and accept) J’s performance of “true pain,” so to speak, the pain one feels when realizing and publicly acknowledging that one has hit and hurt his beloved. Displays of upset make certain inferences available about what kind of a person J is—someone who sincerely repents and cares. The dialogue continued with additional reflexive questioning (see Exemplar 1d).

Exemplar 1d

134 T:

and if there was a change José that you would like to do,

135 :

a minimal change, that you would like in your relationship with

136 :

Antonia, what would it be?

137  J:

that I would have to do?

138 T:

you in your relationship with Antonia

139  J:

to control myself more

140 T:

control myself in which sense?

141  J:

like this quick exploding, like yelling right away, swearing

142 T:

is this something important to you, José, being able to change that?

143  J:

yes, because maybe this way we can keep going

Exemplar 1d contains three reflexive questions. The first wh-question seeks information from J about relational changes he “would like to do.” The second question elicits clarification of the meaning for J of “self-control.” The last declaratively formulated with a rising intonation question seeks confirmation from J that enhancing self-control is important to J. Let us look more closely at these question-answer sequences.

In lines 134–136, T issues a hypothetical question (Peräkylä, 1995; Speer, 2012). The question describes a scenario in which J wishes to make changes in how he relates to A. The wh-question (“what would it be like,” line 134) asks J to specify changes he “would like” to make. An extreme case formulation (Pomerantz, 1986) (“a minimal change”) may be used to present T’s proposal that J changes as reasonable and not excessive. J is implicitly asked to assume a small portion of responsibility for the couples’ marital distress, rather than the entire responsibility. This may be done to undermine potential accusations of T being partial and unfair (siding with one partner against the other) and to enhance the likelihood that J endorses the idea of changing himself. Line 139 is hearable as J’s eventual answer to T’s hypothetical question (“to control myself more”), in which he endorses the attribution of responsibility for change to him. “More” is added to possibly imply that J has already exercised some degree of control over his actions.

With the second question T elicits clarification from J of the meaning of self-control (“control myself in which sense?”). With a three-part list (“quick exploding, yelling right away, swearing”) (Jefferson, 1991) he exemplifies what he would not be doing if he had more self-control. T’s use of an indexical (“this”) in the next question treats the previous description as sufficiently clear/acceptable for the issue at hand. This last interrogatively formatted question (is this …?) seeking a yes response helps further center the discussion on the topic of J’s changing himself and establish the relevance of self-change for J. The idea that J should change (i.e., enhance his self-control) is implicitly presented as stemming from J rather than as coming from T’s agenda. J is discursively constructed as an agent and source of self-growth and self-improvement—as someone who is motivated to change , who cares, and who responds out of what is “important” to him. On line 143, J endorses (“yes”) the notion that enhanced self-control is important to him and accounts for his agreement (“because…”). He attributes responsibility to himself for maintaining the relationship and implicitly assigns to himself the concern for his marriage and the attitude of caring for A. On line 143 we again see evidence of J affiliating with T’s claim that self-change is important to J (“yes” and elaborating or unpacking “because…”).

In the last exemplar below, the construction of J as a caring father, someone who seeks to instill certain values in his daughter, is “thickened” conversationally. Sara (11-year-old daughter) reacts by crying when she witnesses her parents fighting and asks J to not yell, which worries J.

Exemplar 1e

190 T:

why do you worry José when Sara tells you that?

191  J:

because I don’t want her to think that

192 :

it is okay that they talk to you this way

193 T:

so you don’t want her to learn

194  J:

the same pattern

195 T:

and instead of that José what is important to you that Sara learns?

196  J:

it is about being respected, that nobody can disrespect her

197 T:

it seems that that’s important

198  J:

yes

The reflexivity of reflexive questions in lines 190 and 195 lies in the subsequent attributional work they observably elicit. The wh-question (line 190) elicits an account or explanation (“why… because…”) from J of his emotional state (“worry”) in response to his daughter’s request that J does not yell at her mother. J provides such explanation in the next turn (lines 191–192). “They” in J’s response presumably references either people in general or men. In the next turn (line 193), T formulates or summarizes (“so…”) J’s explanation, subtly reshaping its content from Sara thinking to Sara learning, highlighting J’s fatherly role of a teacher and role model. Once again, J is attributed certain parenting attitudes and intentions; he is constituted as a father motivated by care and concern for his daughter. J completes T’s utterance (line 194), which can be seen as a sign of his affiliation with or endorsement of T’s claim that J “wants” to be a positive model to his daughter.

T’s subsequent wh-question (line 195) builds on the material that has been previously established (“and instead of that”) and elicits further information from J regarding the specifics of what J wants his daughter to learn from observing him, provided by J in the next turn. T’s question once again attributes care to J and highlights the relevance and significance of positive fatherly intentions for him. J joints this construction by supplying information about values he seeks to instill in Sara (line 196). And once again, T’s final turn (line 197) marks his limited access to J’s mind by downgrading his statement with “seems.” That is, T advances his ideas about J’s experience while honoring J’s superior knowledge of his own experience.

Both question-answer sequences work to attribute causality, presenting J’s emotions and actions as stemming from his underlying positive fatherly intentions to teach his daughter good values. J may be accused of being a bad father who exposes his daughter to violence. We would argue that attributional work is done in service of defending against potential accusations that J exposes his own child to domestic violence. Indeed, we see that while T’s first question more broadly elicits from J an explanation of his emotional state, J attributes positive intentions to himself in his response, arguably as a way to defend against potential accusations and attributions of blame. This highlights the importance of examining causal attributions while attending to the broader interactional context and social activities (e.g., defending, blaming, praising, complaining) oriented to and managed in interaction.

Discussion

In this chapter, we hoped to encourage a dialogue between discursive-oriented researchers and therapists. Our focus was on reflexive questions as constructive interventions. We offered a discursive psychological perspective on reflexive questions to better understand their potential effects for the therapeutic process: they invite attributional work. In doing so, they (a) help constructing mutually acceptable accounts ; and (b) they foster the discursive construction of clients’ autonomy (or capacity to generate preferred “patterns of cognition and behavior on their own”; Tomm, 1987b, p. 4).

In the exemplars we presented, T’s questions opened space for the co-construction of alternative meanings and depictions of the client’s actions and identity, namely, the construction of a morally acceptable (non-violent, kind, caring) partner and father identity. We have shown that questions were used to manage issues of blame and responsibility for change and accomplish attributional work—attribute agency , responsibility, and certain attitudes to the client. The client’s actions were presented as causally linked or driven by his underlying values and intentions. The questions embedded and subtly advanced optimistic presuppositions about the client being a person motivated by “good” intentions and values (MacMartin, 2008). The client’s positive intentions were topicalized (Vehviläinen, 2008) or presented as a relevant topic to explore in therapy and constituted as significant for the client. The client consistently endorsed the therapist’s propositions concerning him being an agentic person, and the therapist systematically backed this up. Interestingly, the therapist treated the client as more knowledgeable than her in the accounting business (i.e., granting higher epistemic rights about clients’ “inner” experience; Heritage, 2013), which some discursive-oriented therapists have long described as the enactment of a “not-knowing stance” (Anderson, 1997).

Tomm (1987b) offered a rich taxonomy of reflexive questions. What seems missing from his description is the back-and-forth conversational work involved in co-constructing alternative, client-preferred meanings . Rather than solely fixing our attention on carefully crafted questions, as therapists, we are able to orient to how alternative meanings is a collaborative endeavor: they are discursively generated in ongoing interaction. We highlighted that the micro-details of language use are not only relevant for therapists to pay attention to but are also consequential for how a conversation unfolds and which versions of clients’ identities and experiences are advanced and constituted and which other possible versions are passed over. This perspective directs therapists’ attention to the social interaction and discursive production of meaning in therapy (Avdi, 2005; Strong, 2007). As suggested in our second initial quote, attention to this ongoing production heightens therapists’ responsive involvement or discursive reflexivity —the discursive cousin to Tomm’s notion of reflexivity suggested in our first initial quote (i.e., activating the reflexivity among meanings within preexisting belief systems).

Tomm (1987b) portrayed reflexive questioning as an aspect of interventive interviewing , an empowering conceptual “posture” oriented toward fostering clients as autonomous agents of change —toward “enabling clients to generate new patterns of cognition and behavior on their own” (p. 4). Arguably, fostering clients’ autonomy and self-determination is not just more effective but also more ethical, as compared to therapists persistently advancing their meanings and proposals when clients consistently display or convey reluctance to take them up (see, e.g., CPA, 2000, Principle I). It is our hope that our brief discursive analysis helps make this reflexive conversational process more visible—it is an observable, researchable, teachable, and learnable activity. We offered an alternative, discursive conception of client agency , presently marginalized in the field, which can complement a more conventional understanding of agency as a key variable accounting for positive therapy outcomes (e.g., Asay & Lambert, 1999; Bohart & Wade, 2013; Gonzalez, 2016). A conception of client agency and other preferred client accounts as discursively or interactionally constituted may be useful to practitioners, particularly discursive, social constructionist practitioners who see social interaction and language use in therapy as constitutive (rather than reflective ) of clients’ subjectivities and relationships. Rather than seeing clients’ agency (or other preferred client self-understandings) as internal and stable, we offered a perspective on agency as a practical, situated accomplishment.

Our hope in writing this chapter was to highlight the discursive basis of reflexive questions: how lines of talk (attributional work) are invited through reflexive questions that subsequently involve responsive attributional work to arrive at mutually preferred answers to those invitations. We encourage therapists to adopt a discursive orientation in their work with clients that envisions therapy as discourse . Through discursive psychological inquiry, the sequentiality and nuance of therapeutic actions can become more available to awareness to be employed in the service of therapeutic goals.