Introduction

This study investigates how therapists transition from the opening phase of family therapy meetings to the official business of the session and the topics that the family want to discuss. This is an important transition point as it orients the family from informal to formal business and also implies the therapist’s position in directing the session and expectations on how the family is to interact. This investigation uses conversation analysis (CA), an analytic approach that examines recordings of real-world interactions. CA pays particular attention to how people use conversational practices of word selection, intonation, and prosody to propose, negotiate, accept and resist social actions (Sacks et al., 1974; Sidnell & Stivers, 2013).

Family therapy approaches have long been shaped by postmodernist critiques about power and authority in therapeutic conversations (Flaskas, 2002). For example, Batesonian cybernetics rejected the notion of unilateral power, which was critiqued by Feminists working in child abuse and domestic violence (Flaskas & Humphreys, 1993) and the concept of neutrality has been criticised for taking a detached position that avoids therapist responsibility on important issues (Cecchin, 1987). Others have noted how the authority of the therapist, who despite their best intentions, may “colonize” the family by promoting and normalising particular structures and roles rather than understanding and promoting a family’s own understandings and resources (Rober & Seltzer, 2010). More recently, mental health services are increasingly shaped by the recovery model, which prioritises the increased involvement of services users and their support networks in collaborative interactions about their care (Australian Health Ministers’ Advisory Council, 2013; Waters et al., 2021). Consequently, there is a heightened interest in collaborative-dialogic practices (Anderson & Gehart, 2023), as either a complement or alternative to traditional mental health and family therapy interventions (Buus et al., 2021).

Collaborative-dialogic approaches to family therapy cover a potentially wide array of therapeutic models. One of the more well know approaches is Open Dialogue, which was developed in Finland through the 1980s and focussed initially on psychosis (Haarakangas et al., 2007). Open Dialogue centres around principles of providing immediate help; taking a social network perspective; the flexibility, mobility, continuity, and responsibility of clinicians; tolerating uncertainty; and promoting dialogical conversations. The approach has received growing international interest and developed a research base (Buus et al., 2017; Gromer, 2012; Pilling, 2018; Seikkula et al., 2003, 2006, 2011). Another aligned approach developed in the USA is the collaborative-linguistic approach of Anderson and Goolishian (1988). This approach emphasises the linguistic interactive element of therapy in constructing and also “dis-solving” problems (Anderson & Goolishian, 1988, p. 371), and has proposed the ideas of treating the “client as the expert” and therapists taking a curious “not-knowing” approach (Anderson & Goolishian, 1992, p. 25). While these are two of the more well-known approaches, these ideas have permeated other approaches to family therapy as well as individual clinicians in their applications of family therapy.

A prominent idea in collaborative-dialogic approaches is that therapists promote egalitarian relationships with clients and families and to maximise their agency within a meeting by promoting choice in what topics will be discussed. One way that therapists promote this agency is at beginning of the session when transitioning from greetings to the official clinical business. Therapists are advised to begin each meeting with an open question and then follow the lead of the family so that the client and family can set the clinical agenda (Haarakangas et al., 2007; Ong & Buus, 2021; Seikkula, 2011). For example, “who would like to start?”, “what would be best way to begin?”, or “how would you like to use this meeting?” (Olson et al., 2014, pp. 9–10; Seikkula & Trimble, 2005). However, it is not clear to what extent these recommendations are followed in a real-world therapeutic conversation; especially when there are multiple addressees, such as in a family therapy setting.

Conversation analysis research has shown that many routine clinical encounters between clients and health professionals follow a normative structural organisation (Robinson, 2013). This means that clinical sessions have predictable phases in which specific activities are accomplished (e.g., case history, diagnosis, treatment decisions). Transitions between these phases are marked by a deontic asymmetry (Ekström & Stevanovic, 2023; Stevanovic & Peräkylä, 2012), where professionals primarily take the lead in directing a session through changing topics and initiating transitions between phases of the session (Bercelli et al., 2008; Robinson, 1998). In settings like general practice or surgery consultations, this normative overall structural organisation contains phases of openings/greetings, establishing the presenting complaint or reason for the encounter, gathering information/examination, diagnosis delivery, treatment recommendations, and closure (Beach, 1995; Robinson, 1998, 2003; White & Stubbe, 2022).

Robinson (1998) demonstrates that, in the opening phase of their consultations, general practitioners greet clients, but their body orientation towards clinical notes shows that they are not yet ready to hear the client’s presenting complaint. The doctor’s transition to the official business of the session thus involves not only verbal communication such as “so, what’s the problem?”, but also a shift of gaze and body posture away from clinical notes and towards the client.

Conversation analytic research into psychotherapy sessions has tended to focus on later phases of the session rather than on openings or transitions. This former research suggests a similar, although less rigidly defined phase structure following a repeating structure of gathering information and enquiring into the client’s concerns through question and answer sequences (Bercelli et al., 2008), therapists providing formulations or reinterpretations about what the client has said (Antaki, 2008) or proposals for behaviour change (Ekberg & LeCouteur, 2014), which are then followed by client responses displaying various degrees of agreement with the therapists’ statements (Bercelli et al., 2008; Ekberg & LeCouteur, 2015).

Conversation analysis research on family therapy sessions has similarly focused on granular instances within specific phases of the therapeutic conversation rather than openings. For example, how therapists manage impasse (Couture, 2006; Couture & Strong, 2004; Muntigl & Horvath, 2016), display empathy (Schriver et al., 2022), and the structure of reflecting conversations (Schriver et al., 2019; Williams & Auburn, 2015). Two studies however, have specifically examined how therapists transitioned between the enquiry and formulation phases, into a reflection, which is a formal part of a session where two therapists discuss their thoughts and feelings regarding the prior parts of the session while the family are invited to listen and observe. Parker and O’Reilly (2013) describes how therapists pre-warned the family about the upcoming reflection and, when nearing the transition to the reflection, discouraged further contributions from the family while attempting to avoid ruptures to the relationship. Ong et al. (2023) described how, in an Open Dialogue approach to family therapy, therapists downgraded their authority when proposing transitions to a reflection through asking permission, providing a reason for the transition, and prefacing proposals with “I’m wondering”, which is an indication of low entitlement (Curl & Drew, 2008).

Little empirical attention has been directed to how therapists transition from openings and greetings to the “official clinical business” that will form the main focus of the meeting (Beach, 1995, p. 260; Robinson, 1998; White & Stubbe, 2022). This is particularly relevant for collaborative-dialogic approaches, as clinicians are tasked with a potential conundrum on how to direct the session into therapeutic topics while also following the lead of the client and family. Using conversation analysis, this study therefore aims to examine how therapists using a collaborative-dialogic approach to family therapy introduce the transition from openings to the official business of the session, how these transitions are responded to by their clients, and what these findings may contribute to collaborative approaches to family therapy more generally.

Method

The study site was a youth mental health service in western Sydney, Australia and was approved by the ethics committee in the area health service. Participants included 12 therapists who had completed either an intensive 5-day Open Dialogue foundation training (n = 10) or had an introductory orientation to Open Dialogue through their colleagues (n = 2). Seven of these therapists were also participating in part-time advanced two-to-three year Open Dialogue training. Therapists were male and female clinicians from various professional backgrounds including psychology, psychiatry, social work, and mental health nursing. Clinicians’ level of experience ranged from new graduates to over 10 years of professional practice.

Therapists who consented to participate nominated consumers and their families that they worked with whom they thought may be willing and able to participate based on their current presentation, relationships, and level of acuity. The first author then approached the potential consumer and family participants to explain the study and seek consent. The first author was not connected to the care of the consumer nor family and reassured them that participation was voluntary and a decision to not participate would not affect the care they received. Nine families totalling 36 consumers and family members consented to participate. All consumers were aged under 18 years of age and their family members included parents, guardians, and siblings of varying ages. Consumers had a range of possible diagnoses including depression, anxiety, psychosis, and at times self-harm and suicide attempts. If all participants provided written informed consent, then a family therapy session was recorded. The sessions were recorded with three video cameras placed around the meeting room to capture multiple aspects of this complex visual scene, and audio recorders and microphones were placed adjacent to participants. Ten sessions totalling 14 h of recordings were collected for analysis.

The composition of the meetings was varied. At a minimum, the client was present with at least one parent. In larger meetings, the client could be joined by both parents, siblings, and support workers from external organisations. The number of therapists also varied. Smaller teams included a minimum of two therapists who jointly facilitated the session and reflected with each other. In larger teams, there could be one therapist facilitating the conversation and a separate “reflecting team” of two to three therapists who only spoke when in a designated reflecting conversation with each other.

The analysis followed conversation analysis procedures (Sidnell & Stivers, 2013; ten Have, 2007). This included creating an initial verbatim transcription of talk in the recordings manually by the first author, followed by reviewing the recordings and transcriptions for interactions of interest. This research has investigated various aspects of agency and authority in Open Dialogue meetings (Authors, 2021, 2022, 2024). For this study, we noticed a similar question design when therapists began the session, which took the general form of “what would be useful/important to talk about today?” Intrigued by this repeated form, we turned the analysis towards identifying sequences containing these first questions that began the official business of the session. The identified sequences were transcribed in detail using conversation analysis conventions including word selection, intonation, prosody, overlap with other speakers, silences, and non-vocal behaviours, which was checked by the other authors (see appendix for transcription conventions, Hepburn & Bolden, 2017; Jefferson, 2004; Mondada, 2018). The analysis then looks at each speaking turn line-by-line with attention to the lexical construction and prosodic delivery to understand what functions they served and how they were responded to (Perakyla et al., 2008). These sequences included the target phenomena of the therapists’ question: “what would be useful/important to talk about today?”, the preceding turns before this question to provide an understanding of the context to which the question is delivered, and the subsequent responses to understand how the question was received and understood by the family. CA seeks to analyse and understand how conversational practices are utilised by the participants within a conversation. CA thus represents an emic approach that seeks to understand the internal sense-making of the participants themselves and can empirically demonstrate how theoretical therapeutic principles are actually achieved in real-world therapy sessions (Peräkylä & Vehviläinen, 2003). Validation of the findings is provided through a process of consultation between the authors and presentation of data extracts at group data sessions with other CA researchers. Validation is therefore a recurrent process of data observations and analytic insights and coherence with established CA principles and research (ten Have, 2007). As in all CA research, the analytic findings are grounded in the observed behaviours of the interaction itself though the next-turn proof (Potter, 1996), rather than through external theories about interaction or language. Therefore, validation is demonstrated through the interaction itself, i.e., any proposed phenomena is supported by how that phenomena is responded to by the other participants within the sequential context of the interaction (for more details on the next turn proof and validation procedures see Potter, 1996).

The first and third authors have experience as Open Dialogue practitioners and trainers as well as an applied understanding of CA. The second author specialises in CA and provided expertise in processes and validation. All authors contributed to the analysis and interpretation of the presented findings. Although we have experience in the Open Dialogue approach and this unavoidably influences our analysis, we intended to use CA as an external lens to provide different insights into how therapeutic practices are delivered. The discussion seeks to combine these different perspectives and how they may influence practice.

Findings

The opening phase of sessions included a range of actions and activities, such as introductions for new participants to the meeting, talk about participants’ roles in the organisation or relationships in a client’s support network, or a description of the expected structure of the session. If the participants were known to each other, then there could also be talk about what they had been doing in the time since the last meeting. While this “catch-up” talk may cover topics of therapeutic importance, at some point in the interaction there is a marked transition from this opening phase to the official business of the current session.

In this data, the official business of the session was introduced using an opening question by the therapist in eight of the 10 sessions, in another session it was raised by a family member without a question from a therapist, and in one session it was not raised at all. This last example was analysed as a deviant case because the absence of the first clinically oriented question was due to the consumer not wanting to participate in the session and the therapists then addressing this issue.

When the therapist introduced the official business of the session, we observed a similar sequence and turn structure across the eight sessions (see Fig. 1). This sequence begins after closure of the prior topic, which may include greetings, small/social talk, and/or description of the structure of the meeting, followed by a silence in the conversation. The therapist’s turn begins with a token (e.g., so, okay, alright) signalling a closing of the prior topic and/or an upcoming transition in the conversation (Bolden, 2009; Mondada & Sorjonen, 2021; Schegloff & Sacks, 1973). This is followed by a question or statement orienting to the official business of the session marked by pauses (short periods of silence), re-starting parts of the turn, and self-initiated self-repairs (self-corrections), described as dispreferred turn design features as they qualify or mitigate the therapist’s initiating action (Pillet-Shore, 2017). The therapists’ turns are addressed to the family as a whole and generally follow the form of “what would be useful/important to talk about?” This turn design makes relevant two tasks for the family; to select the next speaker and to propose a topic for discussion. This turn is followed by silence and a lack of direct uptake as the family members negotiate who will speak first. During this silence, the therapist may provide further specification or candidate areas that the family could select to talk about or nominate a next speaker, usually through gaze. After some negotiation within the family, one family member then progresses the conversation with elaboration on a selected topic.

Extract 1 illustrates this general form with the above phases marked. From lines 1 to 4, Therapist 2 is finishing a description of the reflecting conversation that will occur later in the session when the family quietly listens to the therapists talk about their observations (Ong et al., 2022; Schriver et al., 2019). He finishes this turn with a rationale for the reflecting conversation i.e., that by listening without having to think of a response, the family can “hear things differently”.

Extract¹

After a silence of 4 s (line 7), Therapist 1 begins the transition to the official business of the session with a shift marking “so” (line 8). This is followed by a preparatory statement at line 9, “I’m thinking what would be a good utilisation of this one hour” and then a question to the family, “what’s something you think we should talk about?” This turn includes a number of silences and sound stretches and can also be interpreted as tentativeness (Couture, 2006). It also does not specify a particular topic area nor a particular recipient. Consequently, this has to be managed by the family through a speaker self-selecting or some form of negotiation (Sacks et al., 1974). At line 14, there is a long silence as the Mother looks towards her daughter, the Client, with no verbal response from either of them. At line 15, Therapist 1 narrows the topic, adding an increment specifying “something that’s useful to you as a family at whatever point you are in”, suggesting that the family’s lack of response is due to a problem in topic rather than speaker selection. There is a further silence without a verbal response from the family at line 20 before the Mother proposes to the Client that they can talk about a recent incident, which is then accepted by the Client. At each point of possible turn completion there is a silence from the family. Therapist 1 then adds an additional element that the family can respond to. This addition also functions to re-complete the therapist’s turn and thus making relevant the family’s turn to talk, without directly indexing their lack of response. In addition, Therapist 1 introduces the transition to the official business of the session through an open question emphasising that the topic should be something that is useful for the family, presumably in contrast to what is useful to the therapists. This is consistent with a collaborative-dialogic approach where the family should determine the topics of conversation. This is followed by periods of silence or a disruption to the progressivity of the conversation (Stivers & Robinson, 2006) where the Mother looks to the Client before self-selecting and proposing a topic for discussion. Progressivity in this context can take various forms such as the delivery of a turn of talk and the movement towards relevant next actions, such as a therapist’s opening question receiving a fitted next response. In both instances, progressivity moves these interactional units, turns, sequences, and the conversation as a whole towards some projectable completion (Authors, under review; Stivers & Robinson, 2006).

Selection of Speaker Through Gaze

Extract 2 demonstrates the same structure of the therapist’s turn, including the appending of additional specifications to the end of the turn suggesting the qualities of topics that the family may elect to talk about. In addition, after the silence and lack for uptake from the family, the therapist selects the Sister as the next speaker through gaze. Just prior to this extract, the participants have each been introducing themselves and stating their leisure interests. The extract begins with Reflector 2 (an additional therapist) ending this topic and stating that she likes reading after work. At line 5, the Therapist facilitating the session, begins her turn orienting to the business of the session.

Extract²

At line 3, the Therapist’s “okay” displays receipt of the prior talk and closure of this phase of the interaction. After a silence at line 4, the Therapist self-selects and begins the turn with a lip smack and inbreath. This is followed by a silence and a more explicit marking of a change in the session with “so what we might do is just open it up” (lines 5–6). The preface “so” signals the “emerging from incipiency” of a pending agenda, which ostensibly is the official business of the meeting (Bolden, 2009, p. 988). While this turn is grammatically and prosodically complete, it is pragmatically unclear what action “open it up” makes relevant for the recipients. This is evident in the silence at line 7, and the Mother’s quiet “hm” at line 8, which forgoes taking the floor with a further silence at line 9. With this lack of uptake, the Therapist continues at line 10 with a more specific statement turning to the official business of the session with “throwing out the question about what will be useful to talk about tonight”. Similar to Extract 1, this turn has dispreferred design features and is marked by a lack of progressivity within its turn-constructional units and the turn overall (Schegloff, 2013).

The Therapist’s turn is possibly complete at line 13, although there is an inbreath which may signal further talk. There is a silence at lines 14 and 16 where the family members forego the opportunity to speak, then the Therapist self-selects and continues with additional turns that specify possible areas that the family may talk about beyond “useful” such as “anything” that is “important” or “worthwhile” but also relevant to them “as a family” (lines 15–21). The Therapist’s nominating of different topical areas to discuss, suggests that they, at least initially, treat the family’s lack of response as a problem of what to speak about rather than who will speak next. There is again also the presence of silences providing multiple opportunities for family members to self-select and start speaking, but there is a lack of uptake from the family and no evidence of preparation to speak next. Toward the end of her turn, the Therapist’s gaze changes from sweeping between the family members to focus on the Sister at line 21 and only moves away after the Sister’s turn beginning at line 25. So, while the content of the Therapist’s initial turn and her gaze does not specify a particular recipient or next speaker, after a lack of uptake the Therapist’s gaze eventually settles upon the Sister, thereby non-verbally selecting her as the recipient. From line 22, the Client also directs her gaze towards her Sister, implicitly nominating her as the next speaker, while the Mother’s gaze is ambiguous at this point. After a delay, the Sister does speak next and, following some negotiation with the family, she launches into a topic for discussion (line 33).

These first two extracts display common features of how a collaborative-dialogic therapist designs their turn that transitions to the official business of the session. After closure of the prior sequence there is a token or turn preface that signals a change of topic, followed by a turn with dispreferred features that asks the family to nominate an “important” or “useful” topic for discussion. This turn does not clearly select a topic nor a next speaker and is consequently followed by a disruption of progressivity in the interaction. This may be followed by additions to the turn specifying general topical areas for comments and, in Extract 2, is aided by therapist gaze that indicates an anticipated next speaker, which is affirmed after some negotiation by the family.

Contingent Variation in Transition to Business

In two sessions from this data set, the therapist does not introduce the business of the session. In these instances, it is once raised by a family member, and in the other it is not introduced at all as the therapists have to instead manage a more immediate issue of the client’s participation. In Extract 3, the Mother raises the business of the session, and it is accepted by the therapist using a similar turn design to that described. Just prior to Extract 3, the Mother has stated that there is a new arrangement where the Client will be living one week with the Mother and Sister, and one week with the father who is not present (data not shown). At line 1, the Mother summarises that there is a lot to process and work out as a result of this new living arrangement.

Extract³

After finishing her prior turn, the Mother brings up the business of the session stating that she has been asking for this meeting to occur for a number of years (line 3). At line 5, there is an acknowledgement from Therapist 1, and the Mother continues by describing the volatility in the sisters’ relationship (data not shown). She then ends her turn stressing the importance of working out how they can get along as they will be sharing housing. The Mother thus introduces the business of the session by nominating a topic, without the invitation of the therapist.

At line 25, Therapist 1 displays receipt with a quiet “okay”, then at line 27 with a question confirming that this is the most important thing to talk about today. Therapist 2’s confirmatory question, uses a similar structure and word selection to how therapists introduce the official business of the session presented in Fig. 1 and Extracts 1 and 2. Therapist 1’s turn begins with a shift marking preface “so”, and then a statement “that’s the most important thing to talk about today”. This turn again is lexically similar, containing what is “important” to talk about accompanied by mildly rising intonation making relevant a response. This turn receives confirmation from Sister at line 29 who elaborates on the importance of the problem but doesn’t elaborate on the problem itself (data not shown). From this point, the conversation progresses similarly to the structure in Fig. 1, with the therapist reasserting the deontic authority to direct the session. After a silence at line 41, Therapist 1 asks “so who would like to start”. By beginning with “so” the therapist indicates the incipiency of progressing to the official business of the session. While in this example, the topic of conversation has been established, the therapist again does not specify a next speaker. This is again deferred to the family to resolve and the progressivity of the sequence is disrupted with a long silence from the family at Line 43. Therapist 1 then nominates possible areas for discussion i.e., the relationship, what they think is going on, and how it can be friendly at times, but does not nominate a speaker. There is a short sequence where the sisters work out who will speak first (lines 52–56) before Sister provides a detailed description of the problem (data not shown).

This example contrasts with the first two because the Mother introduces the business of the session without the prompting of the therapist. In response, the therapist utilises a similar format for introducing the business of the session described in Fig. 1. By asking “who would like to start?”, the therapist re-establishes the deontic initiative in the session, which effectively ratifies or accepts the Mother’s proposal to transition to the business of the session. The form of turn design results in a disruption of progressivity because, although the topic has been established, the therapist did not select a next speaker resulting in the family having to negotiate who will speak next. Here, then, we see how the normative pattern of initiating the business of the session is maintained and validated because despite the Mother’s independent development of the substance of the first topic, the therapist still formally and explicitly establishes the topic of the session and the transition to who will speak first.

In Extract 4 there is no explicit transition to the official business of the session by either the therapists or the family. Instead, the conversation moves to focus on the more immediate issue of the client wanting to leave the meeting, which effectively becomes the business of the session. Prior to the extract, the Client has mentioned that she is feeling nauseous and would like to go to her room to watch television. This has been rejected by the Mother because the therapy session is about to start (data not shown). The extract begins with the end of the opening phase, with Therapist 2 asking the client what it is like for the therapists to be “back here” (i.e., at the client’s home) for the meeting. There is a short post-expansion sequence before the topic closes. After a silence at line 10, Therapist 2 begins what seems to be an orientation to the first question with “I’m wondering” (line 11).

Extract⁴

It is unclear if Therapist 2’s “I’m wondering” at line 11 is the beginning of the business of the session. However, it contains many of the features of the prior examples, i.e., it follows topic closure of the opening phase of the conversation, there is a long silence prior to the turn, the turn is prefaced by “I’m wondering” and possibly marking a transition in the conversation, and there is a pause after this turn beginning consistent with the turn design in the other extracts. However, the therapist does not continue the turn because the Client self-selects, asking the Mother if she can go to her room to watch her TV (television) show as it will make her nausea feel better (lines 12–14). The conversation then continues with talk about the Client feeling nauseous and not wanting to have the session (data not shown). In this extract, the silence within the design of the therapist’s turn provides an opportunity for the Client to self-select and direct the conversation away from the business of the session. Furthermore, the activity of transitioning to the official business of the session becomes irrelevant as the issue of the client feeling nauseous and wanting to leave the session becomes the primary concern that needs to be addressed by the therapists otherwise the session cannot continue. Consequently, the conversation in this session continues around engaging the client in the session and discussion with the Mother about this. The therapist’s turn that begins the business of the session by inviting a speaker and a topic is thus unable to occur unless there is at least an implicit agreement that all parties agree to participate in such a discussion. While the presumption of participation is relevant for all forms of conversation, it is especially heightened when a client is opting out of recipiency as the premise of psychotherapy is to engage with a client’s concerns.

Discussion

Within this data set, we demonstrate an orientation to a normative structure for introducing and transitioning to the official business of an Open Dialogue session. Therapists close the opening phase of the session by a shift marking preface or token and introduce the official business of the session through a questioning turn characterised by silences and restarts that seeks for the family to nominate a topic for discussion. This is followed by a silence from the family while there is a negotiation of speakership. The therapist may then propose broad topics for discussion focusing on what is “useful” or “important” for the family, and/or implicitly select a speaker through gaze. After further negotiation a speaker is selected or self-selects and proceeds to propose a topic, enabling the session to continue. We also found some contingent variation, where a family member introduced the transition to the business of the session. In response, the therapist regained deontic authority by accepting the proposal to transition to the official business of the session and initiating the family task of establishing the next speaker. Further contingent variation was found in an instance where the therapist prepared to transition to the official business but instead had to address the client withdrawing from the session. These contingent variations provide complementary, validating evidence of the normative pattern for introducing the official business of the session. In this discussion, we consider how this observed pattern may connect with theoretical principles of collaborative-dialogic approaches such as the promotion of choice by the family, how the disruption to progressivity affects the conversation, and the implications for collaborative-dialogic practices.

Intersection with Collaborative-Dialogic Practices

Collaborative-dialogic theorising recommends that therapists “follow the lead” of the family by allowing them to determine the topics and the direction of conversation (Ong & Buus, 2021, p. 250), with the therapist’s task being to promote dialogical interactions around those topics as “an architect of the dialogical process” (Anderson & Goolishian, 1992, p. 27). The findings from this study align with this theoretical orientation. Therapists introduced the transition to the official business of the session but deferred the decision on the topic and speaker to the family. Therapists specify the topic area only in broad terms such as topics that are “important”, “useful”, or “worthwhile” to the family and their relationships. This promotes selection of a topic that is institutionally relevant while also promoting flexibility in the family’s choice of topic. This is consistent with prior theorising and interactional research on collaborative therapeutic dialogues, where therapists direct the structure of a session, although in a mitigated way, while deferring the to the family on matters of knowledge topical (Guilfoyle, 2003, 2006; Ong et al., 2023).

The therapists’ apparent goal of promoting family choice through the avoidance of topic and speaker selection results in a period of negotiation between family members about who will respond first. Our analysis suggests that therapists are sensitive to this lack of immediate uptake by clients and families, evidenced by them self-selecting to speak after a silence and proposing possible topical areas for conversation and at times nominating a next speaker non-verbally through gaze. The therapists’ circumspection (at least initially) with nominating a topic or speaker seems to regularly produce such a disruption. This may simply reflect the affordances of multiparty interaction, which mean that next speaker selection requires different forms of coordination than dyadic interaction. However, there are several professionally driven questions that arise around the consequences of beginning the session in this fashion.

Disruption of Session Progressivity

People are incredibly sensitive to the communicative import of silences, and silences between turns are a “marked” occurrence across languages and cultures (Stivers et al., 2009). By producing extended silences as part of the session beginning, therapists’ may risk evoking these normative orientations and create engagement issues; particularly if the family are desiring the therapist to take an active lead in a session. More pragmatically, it may also appear to waste time if the family are involved in lengthy negotiations that delays discussions of important topics when there is limited time and availability of therapists. On the other hand, and in accordance with the theoretical principles of Open Dialogue, the disruption of progressivity may be desirable and set off this contact with healthcare will be different than others. For example, in other healthcare settings, therapists take greater responsibility in meetings through more defined question and answer sequences, the provision of diagnoses, and recommendations for treatment or intervention (Robinson, 2003). By declining to initially specify a topic or speaker and by only providing proposed topics or implicit speaker nomination when there is a disruption to progressivity, the therapist sets up early in the meeting that they will lead the session only in certain respects and thus positions the family has having to actively contribute to the discussion. Buus and McCloughen (2022) have described Open Dialogue as promoting the “responsibilization” (p. 315) of families through increased practical and psychosocial work both within and outside of therapy meetings. The placing of work on families to decide the direction for the meeting activates their participation and could be one way that this “responsibilization” is promoted.

Other conversation analysis research suggests that it is the very disruption of progressivity that is a characteristic of dialogue more generally (van Burgsteden et al. 2022a; Burgsteden et al. 2022b). Within institutional interaction, activities progresses towards particular trajectories and outcomes in line with the goals of the institutional setting (Drew & Heritage, 1992). However, one consequence of this is that alternative perspectives can be lost in the momentum of interactional progressivity. Consequently, in order to promote dialogue (i.e., the interaction between multiple perspectives) the interaction has to be disrupted in some way to allow additional perspectives to be voiced (van Burgsteden et al. 2022a; Burgsteden et al. 2022b). In collaborative-dialogic meetings, the disruption of progressivity (even in the early stages of the meeting) shows the therapist’s willingness to develop a different kind of conversation, including the potentially prolonged negotiation of who will speak next and fostering voicing of multiple perspectives. Therefore, one consequence of this organisation of session openings may be the disruption of endemic patterns of clinical communication to promote more responsibility by introducing forms of discontinuity. Furthermore, the task of creating “collaboration” may entail precisely these forms of disrupted conversations.

Implications for Practice

While this study has specifically investigated a collaborative-dialogic approach, these findings are potentially relevant to any collaborative approach to psychotherapy that preferences the establishment of egalitarian collaborative partnerships between therapists, clients, and families. Even models that advocate a more therapist-directed approach still need to maintain attention to client and family engagement, characteristics, values, context, and preferences consistent with the definition of evidence-based practice (American Psychiatric Association, 2006).

The collaborative-dialogic literature generally provides guidance to therapists through broad aspirational principles for practice (Ong & Buus, 2021). This promotes flexibility in the application of the approach but also a high level of practice variation. CA research has noted the importance of not only considering therapeutic principles, but how these principles are enacted within real conversational settings (Stokoe, 2014, 2020). This study identified that the recommended types of questions for opening a session in a collaborative fashion, creates a disruption in the conversation as families must decide “what” will be spoken about and “who” will speak first. That is, the pursuit of a collaborative agenda can potentially stall a conversation. Therefore, therapists may consider whether to take a more active approach in transitioning to the business of a session. If so, the task of speaker selection arguably comes before topic as a speaker must be established before they can state the topic they want to speak about. Therapists could therefore consider directing who will speak first, before enquiring about the topics for discussion. Relatedly, it has been argued elsewhere that, for collaborative-dialogic approaches, the exercise of deontic authority by therapists, (i.e., directing the session and ensuring that everyone has the opportunity to speak) is very different from claiming epistemic authority over what is privileged knowledge (Ong et al., 2023). Therefore, following Ong et al. (2023), therapists may therefore be proactive in directing the session and ensuring that everyone has the opportunity to speak, while deferring to the family on matters of knowledge and not privileging their own knowledge position. So, rather than asking a family to decide who will speak first, therapists could sequentially select each family member in turn to ask what topics they want to discuss in the meeting, as in other forms of family therapy (e.g., Wallis & Rhodes, 2024). In this way, therapists may take a more active role in directing who will speak next in a session, without compromising the respect for the knowledge or experience of the family. Therapist leadership and control of a session is not necessarily contrary to collaboration, and may in fact be necessary in order to ensure that all participants have a voice (Ong et al., 2023).

Limitations and Future Research

Potential participants were nominated by therapists and this could be seen as introducing some selection bias. In addition, this research took place in a child and youth mental health setting and may therefore not be transferrable or generalisable to other types of clinical encounters. However, CA research analyses how people achieve certain conversational tasks and their implications for the interaction. As such claims about selection bias or generalisability are less relevant than providing empirically-demonstrated phenomena from real-world interactions. Our findings therefore alert therapists to potential ways that their introduction of the business of a session can affect the subsequent conversation. This not only has impacts on the local level of a therapeutic conversation, but also potential implications for theory and if conversational practices align with theoretical values. There will undoubtably be a range of other conversational practices that are utilised by therapists in transitioning to the business of a session that were not identified in this study. Further research could therefore explore the range of other practices utilised at this transition point, as well as transitions to other phases of a session such as giving advice, introducing interventions, or closing the session. In addition, future research could compare how these transitions are introduced by therapists utilising other therapeutic approaches beyond collaborative-dialogic ones.

Conclusion

Within the collaborative-dialogical approach of Open Dialogue, the transition from the opening phase of the meeting to the official business of the meeting is characterised by an elaborated turn design that places the onus on the family to select to will speak first and to select a topic for conversation. While consistent with theoretical principles that promote family choice and agency, this turn design results in a disruption to the progressivity of the conversation as the therapist proposes topics or implicitly selects a next speaker through gaze, while the family negotiate who will be the next speaker. This disruption of progressivity may have potential utility via the therapist delimiting the boundaries of their role and placing more responsibility on family members to work together. We also suggest that therapists may consider taking a more active role in managing the turn taking in a meeting that can potentially minimise disruptions to the conversation without compromising therapeutic principles. Collaborative-dialogic meetings may therefore be characterised by the exercise of particular forms of power, rather than the appearance of its absence (Guilfoyle, 2003; Ong et al., 2023).