Authority and Agency in
Open Dialogue Network Meetings: A Conversation Analysis
Benjamin-Hai Leng Ong
A thesis submitted in fulfilment of the requirements for the degree of Doctor of
Philosophy
Susan Wakil School of Nursing and Midwifery
Faculty of Medicine and Health
University of Sydney
June 2021 Statement of originality
This is to certify that to the best of my knowledge, the content of this thesis is my own work.
This thesis has not been submitted for any degree or other purposes.
I certify that the intellectual content of this thesis is the product of my own work and that all the assistance received in preparing this thesis and sources have been acknowledged.
If my candidature is successful, I will lodge this thesis with the University Librarian to be made available for immediate public use.
Ben Ong, 24th June 2021.
2 Authorship attribution statement
Part of Chapter 1 of this thesis is accepted for publication as:
Ong, B., & Buus, N. (in press). What does it mean to work “dialogically” in Open Dialogue and family therapy? A narrative review. Australian and New Zealand Journal of Family
Therapy.
I designed the study, extracted the data, analysed the data and wrote the drafts of the manuscript with the analytic and editorial support of my supervisor.
Part of Chapter 2 of this thesis is published as:
Ong, B., Barnes, S., & Buus, N. (2020a). Conversation analysis and family therapy: A critical review of methodology. Family Process, 59(2), 460-476. doi:10.1111/famp.12431
I designed the study, extracted the data, analysed the data and wrote the drafts of the manuscript with the analytic and editorial support of my supervisors.
Chapter 3 of this thesis is published as:
Ong, B., Barnes, S., & Buus, N. (2020b). Conversation analysis and family therapy: A narrative review. Journal of Family Therapy, 42(2), 169-203. doi:10.1111/1467-6427.12269
I designed the study, extracted the data, analysed the data and wrote the drafts of the manuscript with the analytic and editorial support of my supervisors.
Chapter 4 of this thesis is published as:
Ong, B., Barnes, S., & Buus, N. (2020c). Downgrading deontic authority in open dialogue reflection proposals: A conversation analysis. Family Process, Advance online publication, 1-
16. doi:https://doi.org/10.1111/famp.12586
3 I designed the study, extracted the data, analysed the data and wrote the drafts of the manuscript with the analytic and editorial support of my supervisors.
Chapter 5 of this thesis is published as:
Ong, B., Barnes, S., & Buus, N. (2021). Eliciting stance and mitigating therapist authority in
Open Dialogue meetings. Journal of Marital and Family Therapy, 47(1), 120-135. doi:10.1111/jmft.12454
I designed the study, extracted the data, analysed the data and wrote the drafts of the manuscript with the analytic and editorial support of my supervisors.
Chapter 7 of this thesis is accepted for publication as:
Ong, B., Barnes, S., & Buus, N. (in press). Developing multiple perspectives by eliding agreement: A conversation analysis of Open Dialogue reflections. Discourse Studies.
I designed the study, extracted the data, analysed the data and wrote the drafts of the manuscript with the analytic and editorial support of my supervisors.
As supervisor for the candidature upon which this thesis is based, I can confirm that the authorship attribution statements above are correct.
Prof Niels Buus, 24th June 2021.
4 As supervisor for the candidature upon which this thesis is based, in my opinion this thesis is sufficiently well presented to be examined and does not exceed the prescribed word limit.
Prof Niels Buus, 24th June 2021.
5 Abstract
Open Dialogue is an approach to working with people and their social networks in times of mental health crisis. The approach centres around principles emphasising the responsiveness, responsibility, and flexibility of the treating team as well as the ability of the therapists to tolerate uncertainty by not rushing decisions on diagnoses or interventions
(Seikkula & Arnkil, 2006). Open Dialogue places particular importance on the promotion of dialogue including an openness to hearing and understanding the experience of the client and network members, a commitment to collaborative understandings through the voices of all those present and following the lead of the network, and a willingness to be public with their own thoughts. Open Dialogue focusses on the moment-to-moment unfolding of a conversation and the therapists’ responsiveness to the changing demands of the interaction.
This thesis uses Conversation Analysis, an approach that focusses on describing the conversational practices and normative expectations that are present in conversations, and how these practices achieve social actions. The study data comes from 14 hours of video recorded therapy sessions by Open Dialogue therapists (n=12) and their clients and social network (n=36) in a child and youth mental health service. This thesis is organised around chapters analysing how therapists construct proposals to transition to a reflecting conversation, how therapists elicit stance positions and introduce new areas of discussion with multiple participants, the functions of therapists repeating the prior talk of the family, and how therapists construct a reflecting conversation. Throughout these studies therapists oriented to issues of deontic authority, or the ability to determine the future actions of others, and epistemic authority, or who has primary rights to knowledge. Therapists assume deontic authority in nominating topics for further development and selecting next speakers, while also deferring to the epistemic authority of the recipients, and promote particular forms of
6 agency by fostering flexibility in responses. These studies demonstrate some of the ways that therapists manifest dialogical principles in practice and provide a more detailed understanding of the work that is done by Open Dialogue therapists.
7 Acknowledgements
Many thanks to my supervisors Prof Niels Buus and Dr Scott Barnes for their patience, careful reading, and thoughtful feedback.
Emma Croker for her limitless encouragement and editing.
And the boys Herman, Doobee, Marius, and Bronson, not the most helpful but always funny.
8 Table of Contents
Statement of originality ...... 2
Authorship attribution statement ...... 3
Abstract ...... 5
Acknowledgements ...... 8
Table of Contents ...... 9
Chapter 1: Introduction ...... 12
Prologue - Why This Now and How to Go On? ...... 12
Origins of Open Dialogue ...... 14
A brief history of family therapy ...... 15
The Dialogical Approaches ...... 18
What Does it Mean to Work “Dialogically” in Open Dialogue and Family Therapy? A
Narrative Review ...... 22
Chapter Summary ...... 43
Chapter 2: Method ...... 44
Conversation Analysis ...... 44
Historical origins of CA ...... 44
Central Concepts in CA ...... 47
Turn-taking ...... 47
Sequences ...... 47
Stance ...... 50
Epistemic authority ...... 51
Deontic authority ...... 53 9 Repetition ...... 55
Disagreement ...... 56
CA Methods ...... 58
Validity checks ...... 59
CA and psychotherapy ...... 59
Other analytic approaches ...... 61
Introduction to Methodological Critique ...... 64
Conversation Analysis and Family Therapy: A Critical Review of Methodology ...... 66
Chapter Summary ...... 81
Chapter 3: Literature Review ...... 82
Conversation Analysis and Family Therapy: A Narrative Review ...... 83
Update on literature reviews ...... 115
Chapter Summary ...... 117
Thesis Aims ...... 117
Chapter 4: Deontic Authority in Reflection Initiations ...... 119
Downgrading Deontic Authority in Open Dialogue Reflection Proposals: A
Conversation Analysis ...... 120
Chapter Summary ...... 135
Chapter 5: Eliciting Stance ...... 137
Eliciting Stance and Mitigating Therapist Authority in Open Dialogue Meetings ... 138
Chapter Summary ...... 154
Chapter 6: Therapist Repeats ...... 156
10 A Conversation Analysis of Therapist Repeats in Open Dialogue Network Meetings
...... 157
Chapter Summary ...... 188
Chapter 7: Eliding Agreement in Therapist Reflections ...... 189
Developing Multiple Perspectives by Eliding Agreement: A Conversation Analysis of
Open Dialogue Reflections ...... 190
Chapter Summary ...... 212
Chapter 8: Summary and Conclusions - Authority and Agency in Open Dialogue
Meetings ...... 213
Power in Open Dialogue ...... 217
Authority in Interaction ...... 218
Agency and Asymmetrical Relations ...... 221
Implications for Open Dialogue and Family Therapy ...... 225
Limitations ...... 230
Future Research ...... 234
Epilogue ...... 235
Interdisciplinary Reflections on Conversation Analysis, Power, and Open Dialogue
...... 236
References ...... 249
11 Chapter 1: Introduction
Prologue - Why This Now and How to Go On?
I will begin by saying a little about myself and how I came to be doing this project.
For you, I hope that it will provide some orientation to the project as well as an insight into my own perspective and potential biases. And for myself, some reflection on where I started and where I have ended up. This story centres around two main recurring professional interests of mine: client-centred approaches, and family systems; and how these interests set me on a seemingly inevitable path towards the Open Dialogue approach.
When starting out as a psychologist, I was interested in the different approaches to psychotherapy and how they theorised mental problems and what could be helpful interventions. I leaned towards more “active” forms of therapy. These approaches, such as cognitive behaviour therapy and acceptance and commitment therapy, felt active, because they had a specified set of techniques or strategies that one can use with clients. However, I found that when one is teaching skills or correcting “dysfunctional” behaviours, at some point people inevitably don’t want to do it or say that it doesn’t work. I found that if I held tightly to the model and the techniques, I could end up in a sort of argument trying to convince people that they needed to try harder or that they would get better over time, while they were telling me that the approach wasn’t working. I gradually developed an aversion to “teaching” people what to do especially when it didn’t fit the way they understood the world.
Consequently, I gravitated towards a client-centred approach (Rogers, 1951, 1967). I was particularly struck by the idea that people could change if you only created the right relationship for them to be able to change. I found this an incredible idea to contemplate: you didn’t need to do anything as such but rather to be a particular way in relationship with others.
12 My first psychologist job was in a Juvenile Justice detention centre. The work was almost exclusively risk assessments and individual therapy. I remember doing what felt like productive work with young people, who often said that it was helpful; they promised not to offend again and that they would never come back. But when they went back into their families, communities, and the same situation, many would regularly re-offend and be back in detention, sometimes after only a few weeks. I came to think that individual psychotherapy was not enough. Rather, it was necessary to look broader and to work with families and broader social networks. Eventually, I did learn more about families through a Masters of
Couple and Family Therapy at the University of New South Wales. Having worked in individual psychology and the justice system, it took a lot for me to think outside of a linear cause-and-effect perspective. The ideas of family therapy, postmodernism, circular causality, and the inextricable positioning of individuals within family and social systems radically changed how I viewed the world. I was also introduced to Open Dialogue. Open Dialogue seemed to embody a humanist way of interacting with people while also emphasising the importance of the social network of family, friends, and professionals around a person. It seemed to incorporate these two themes that I had been trying to connect up until that point.
The core idea of Open Dialogue is the promotion of dialogue between all the voices that are present in the meeting, both internal and external. In response to every statement, the therapist is encouraged to think about “how to go on?” How to respond in a way that promotes dialogue and possibilities rather than closing them down? This question seemed relevant to every part of psychotherapy, especially when the work feels difficult or stuck.
But I found that in reading Open Dialogue articles the question of “how to go on?” was never fully answered. There were common themes about being present, attending to one’s inner voices and creating space for the “not-yet-said”. But how is this actually done? I
13 thought that there must be certain things that the therapist could do that made dialogue more likely and, conversely, things the therapist did that shut down dialogue. But writings on Open
Dialogue tended towards more general descriptions of the clinician’s mindset rather than a focus on specifics. The best answer I had was that if you followed the principles and maintained a dialogical mindset, then you would be appropriately responsive to the situation as it arises. I felt that I needed a better understanding of what happens in Open Dialogue conversations. This led me to this research project with the goal of finding out and describing what actually occurs in Open Dialogue conversations.
Origins of Open Dialogue
In this section, I describe the Open Dialogue approach in more detail. I first outline how the approach has been theorised and described in order to identify what it is this research is actually attempting to investigate. In understanding Open Dialogue it is useful to first place it in its historical context. I will therefore briefly trace the history of family therapy as it applies to the development of Open Dialogue and dialogical ideas. In doing so, it is first necessary to acknowledge the shortcomings of this history. My account of family therapy history and the development of Open Dialogue will necessarily be simplified with broad descriptions of approaches and groupings of family therapy models and frameworks. This history is also geographically limited, as family therapy has developed in different ways in different areas. For example, there are stark differences between the United Kingdom and the
United States with some family therapy approaches almost completely absent in these respective countries (Flaskas, 2010). My perspective is therefore unavoidably influenced by my family therapy training in Australia, and the particular history and development of family therapy in this country, as well as my current placement in time looking back and conceptualising these approaches. Finally, the described differences between models of
14 family therapy seems to have more of a theoretical and educational, rather than a practical function. Practicing clinicians have a much more pragmatic and conditional relationship to models and how they relate to them and utilise them in practice (Flaskas, 2010). Clinical practice necessitates a movement between models or paradigms as required by the family situation (Larner, 1994). With these qualifications aside, I will now briefly trace the development of family therapy in Australia and the conditions for the development of Open
Dialogue and dialogical approaches.
A brief history of family therapy
Cybernetics and modernism
Up until the 1950s, families had been thought of as collections of individual psyches
(Flaskas, 2010). A change of perspective in the 1950s saw the ideas of systems theory or cybernetics, that had previously been developed for the physical sciences, applied to understanding family systems (Jones, 1998; Lorås, Bertrando, & Ness, 2017). These included the idea that the family system is made up of parts and sub-systems that influence each other, so that changes in one part can affect the functioning of the system as a whole. One example of this mechanistic perspective is the concept of homeostasis, where systems work to maintain stability and resist change through a corrective, self-regulating feedback system
(Hayes, 1991; Jones, 1998). Problems in a family are thought to be sustained by these feedback mechanisms (Lorås, Bertrando, & Ness, 2017), and the role of the therapist was to therefore identify and change these maladaptive behaviour patterns. Now referred to as the modernist perspective, these ideas included assumptions that there was an external, normative, objective reality to family systems that was knowable by the therapist. The therapist is thus considered to be an objective, outside observer who can assess and identify a
15 family’s maladaptive patterns and then proceed to intervening and correcting these problems towards a more functional and normative direction.
While these ideas can seem somewhat mechanistic, reductionist and hierarchical they marked significant changes in the way that families were thought about. This “first wave” of family therapy approaches shifted the focus away from more medical and individually focussed descriptions to a more inclusive engagement with people and their place in relationships (Flaskas, 2010).
Postmodern influences
The 1980s was a period of change and transition in family therapy in Australia with the introduction of new and challenging ideas, leading to reactionary and sometimes revolutionary changes to family therapy frameworks (Flaskas, 2010). Flaskas (2010) identifies three sets of influences on family therapy at this time. The first influence involves what is described as second-order cybernetics where the therapist is no longer seen as an external, independent observer. Instead, the presence of the therapist and their relationship with the family influences how the therapist constructs their idea of the family (Cecchin,
1992). In this view, objective reality is unknowable and the therapist therefore cannot objectively observe and describe the family system. The therapist’s knowledge of the family is therefore always provisional, influenced by their own perspective, and needs to be continuously checked against feedback from the family (Lorås, Bertrando, & Ness, 2017).
The second set of influences comes from the feminist critique of family therapy with a strong emphasis on social justice. These criticisms centred around the systems view that tended to de-humanise the family as well as disregard broader contextual issues such as power and gender inequality. The systemic idea of circular causality was problematic as it tended to reduce individual responsibility for violence and abuse. Therapists become increasingly
16 aware of broader social contributors to family problems and the importance of directly addressing issues of abuse in relationships. The third influence comes from a collection of ideas that can be described as the postmodernist perspective including constructivist, social constructionism and narrative ideas. Postmodernism challenged the assumption that there was an objective and knowable reality that could incrementally be more accurately described.
Instead, family therapy came to acknowledge that there were multiple subjective experiences within a family and that it was therefore important to recognise this diversity of perspectives
(Jones, 1998). Postmodern critiques lead to a scepticism about privilege, truth and knowledge and the importance of therapists being aware of the power that was associated with their professional role (Hare-Mustin, 1994; White & Epston, 1990). With these different influences the therapist becomes more of a collaborator (rather than an expert) working with the family to jointly understand their experiences. Family therapy thus moved from an implicit behaviourism to a more cognitive perspective emphasising meaning, narrative and dialogue
(Lorås, Bertrando, & Ness, 2017).
These sets of influences shaped the development of current and emerging family therapy approaches. In Australia, these centred around the Milan systemic approach, solution- focused brief therapy, narrative therapy and dialogical approaches. The Milan systemic approach (Selvini, Boscolo, Cecchin, & Prata, 1980), had already begun a transformation from its origins, by rejecting ideas of hierarchy and objectivity (Cecchin, 1992; Hayes, 1991).
Further developments focussed on the concept of circular (over linear) causality and the importance of questioning as an intervention in itself (Hayes, 1991; Lorås, Bertrando, &
Ness, 2017; Tomm, 1988). The Milan concept of therapist “neutrality”, criticised for its inability to address individual responsibility for violence and abuse, was re-conceptualised as
“curiosity” (Cecchin, 1987). Strategic therapy adapted and developed into other approaches,
17 such as solution focused therapy (Hayes, 1991; Kaslow, 2007). Similarly, early strategic and systemic therapies, with the influence of feminism and Foucauldian ideas of power, have influenced the development of narrative therapy (Flaskas, 2011; Hayes, 1991; White &
Epston, 1990). Despite this apparent diversity, these contemporary frameworks share a number of features including a collaborative approach between therapist and family, an orientation to diversity of culture and relationships, an emphasis on strengths, and an awareness of broader social, contextual and cultural factors (Flaskas, 2011). Also emerging around this period of transition in the 1980s were the dialogical approaches to family therapy.
The Dialogical Approaches
Amongst other contemporary frameworks, dialogical approaches can be considered as reactionary responses to the “therapist-as-expert” role and normative expectations of how families “should” function. Dialogical approaches to therapy have two main intertwined strands: the collaborative language systems approach described by Harlene Anderson and
Harry Goolishian in the United States (H. Anderson & Goolishian, 1988), and the Open
Dialogue approach, coalescing around the work of Jaakko Seikkula and his colleagues at
Keropudas Hospital in Western Lapland, Finland (Haarakangas, Seikkula, Alakare, &
Aaltonen, 2007; Seikkula & Arnkil, 2006). The Open Dialogue approach also draws on a range of other sources from family therapy as well as philosophy (e.g., Andersen, 1987,
1991; Bakhtin, 1981, 1984, 1986; Rober, 1999, 2011; Shotter, 1996, 2006).
The collaborative language systems approach, as it came to be called, is based around the idea that human systems are linguistic systems (H. Anderson & Goolishian, 1988). In this view, human meaning, understanding, and social organisation are constructed through language, rather than through social roles and structures. In essence, problems, and the systems that people arrange around problems, are linguistic systems. Furthermore, if
18 problems are constructed through language, then they can also be “dis-solved” through language. Therapists are advised to take a collaborative “not-knowing” position, meaning that they maintain a curiosity about the clients’ stories (H. Anderson, 2005; H. Anderson &
Goolishian, 1992). This does not mean that therapists do not contribute their own expertise or understandings, but rather they seek to remain curious about the client’s stories and not become too attached to their own theories or understandings about the client. Therapists are perceived as conversational artists who seek to create dialogue and new understandings about the problems that people experience. The collaborative language systems approach places primary importance on language and dialogue in joint, collaborative exploration with clients.
Around the same time that Anderson and Goolishian were developing their collaborative language systems approach, Tom Andersen in Norway developed an innovative use of the reflecting team, making the conversation between the reflecting clinicians transparent and observable to the family (Andersen, 1987). Andersen’s work centred around practices that were respectful, ethical, and helpful rather than being driven by a particular theoretical frame (Flaskas, 2011). His approach rejected the idea of the therapist as an expert, sought to share clinicians’ knowledge with the family, and promoted using everyday language rather than clinical jargon (Lorås, Bertrando, & Ness, 2017). Tom Andersen’s work influenced the originators of Open Dialogue in Finland, who were concurrently developing novel ways of working with people experiencing psychosis. Research and theory on the Open
Dialogue approach is usually described through the seven main principles, which are described below (Haarakangas, Seikkula, Alakare, & Aaltonen, 2007; Seikkula & Arnkil,
2006).
1. Immediate help. The first meeting with the client and their social network is
arranged within 24 hours of their first contact with the service. A crisis is seen as
19 an opportunity for change, as the family and supports are usually actively involved
and there can be a greater openness about previously undiscussed issues.
2. Involving the social network. A client is seen as embedded within their
relationships and any changes will necessarily involve those who are significant
and close to them. With a client’s permission, their personal and professional
network is invited to attend all network meetings. A person’s private network may
include family, friends, neighbours, or co-workers. The professional network may
include people in support services such as individual therapists, social workers, or
hospital nursing staff. Decisions about treatment are made in the meetings with the
network and not in separate meetings with only clinicians present.
3. Flexibility and mobility. The treating team is adaptable to the needs of the client
and network. For example, the treating team may arrange home visits or meetings
in different locations or times to accommodate the needs of the family. Meetings
can also vary in frequency or utilise different psychotherapeutic interventions as
required.
4. Teamwork and responsibility. A team of clinicians are involved in all the meetings
arranged with a family, rather than a series of individual consultations with
different specialists. The composition of the treating team is also designed to suit
the needs of the client and family, and the team jointly accepts responsibility for
the treatment process.
5. Psychological continuity. Clinicians make a long-term commitment to remain
involved with a particular client and family throughout the client’s contact with the
service. This continues if the client is admitted to hospital or other therapies are
introduced.
20 6. Tolerance of uncertainty. Clinicians avoid making quick diagnoses or treatment
decisions. Instead, the treating team is patient, curious, and remains open to
multiple options and choices about the direction of treatment without solely
seeking consensus or agreement.
7. Dialogue. Clinicians seek to create an atmosphere of openness where differing
ideas, meanings and explanations can co-exist and thus allow for multiple
possibilities to emerge. This occurs through generating dialogue and equality
among all participants in the network meetings. The perspectives of the clinicians
ideally have no greater influence or authority than any other perspective.
The central principle of the Open Dialogue approach is arguably the creation of dialogue, even taking precedence over the creation of change (Seikkula & Arnkil, 2006).
However, what constitutes a therapeutic dialogue and how it is promoted in a session is at times difficult to describe. In the following article, accepted for publication in the Australian and New Zealand Journal of Family Therapy (Ong & Buus, in press), I seek to focus specifically on the concept of dialogue, how it has been conceptualised in relation to family therapy, and what specific guidance has been given to therapists.
21 What Does it Mean to Work “Dialogically” in Open Dialogue and Family Therapy? A
Narrative Review
Ben Ong & Niels Buus
Interest in the Open Dialogue approach has expanded internationally over recent years with implementation programs across a wide range of contexts (e.g., J. M. Brown & Mikes-
Liu, 2015; Buus et al., 2017; Gordon, Gidugu, Rogers, DeRonck, & Ziedonis, 2016; Gromer,
2012; Pilling, 2018). A number of recurring challenges in training and implementation has been noted (Buus et al., in press; Ong et al., 2019). One challenging aspect for clinicians is the principle of dialogism or the promotion of dialogue. One source of this difficulty is the different uses of the term dialogue. Cooper, Chak, Cornish, and Gillespie (2013) make a distinction between two distinct but connected usages: the ontological and transformative.
The ontological view presents dialogue as a fundamental aspect of human existence and development. Dialogue is seen as an essential part of how humans grow, develop, come to know themselves, and make meaning of the world. The work of Trevarthen et al. (Bråten &
Trevarthen, 2007; Trevarthen, 1990) is often cited as support for this view. This research shows how the interactions of mothers and their newborn babies are closely coordinated with conversational turn-taking and response structures. The ability to converse and dialogue is therefore thought to be present from birth. This parent-child dialogue is said to later develop into a person’s inner dialogue (Seikkula & Trimble, 2005). Then, when one’s inner dialogue is expressed to others and responded to, one comes to reflect on its meaning and understand more about one’s self (Seikkula & Trimble, 2005).
Another common source of support for the ontological view of dialogue is the work of philosopher Mikhail Bakhtin. For Bakhtin (1981, 1984), the “self” of a person is constantly changing in response to continuing dialogues and people come to know themselves through
22 their relations with others. In every social interaction people bring their own previous experiences, explanations, and understandings, including their own unique speaking practices
(H. Anderson, 2005; Bertrando & Lini, 2019; A. Robinson, 2011; Seikkula & Trimble, 2005).
Meaning is constructed through the interaction between these different experiences and perspectives by exchanging and considering alternative explanations (Andersen, 1987). The
“other” is therefore essential to the creation of meaning, because they inevitably present some different form of response to the speaker’s inner conversation (Rober, 2005b; Seikkula &
Trimble, 2005). H. Anderson and Goolishian (1988) extend this idea, arguing that it is through language that systems, structures and roles are defined and created. Consequently, social structures do not exist outside of “languaged” constructions. From the ontological perspective, dialogue is the means by which people grow, develop, create meaning, and define social relations, but it does not explicitly specify that the dialogue needs to have a particular quality. Instead, it seems that dialogue inevitably occurs whenever we interact with others in the world.
The transformative view is where dialogue can become therapeutic. The transformative view proposes that it is not everyday conversation, but a particular form of dialogue that is helpful and therapeutic (Rober, 2005a). In describing the qualities of a transformative dialogue, authors regularly contrast dialogical and monological conversations
(Olson, Seikkula, & Ziedonis, 2014; Seikkula, 2002; Seikkula, Laitila, & Rober, 2012;
Seikkula & Trimble, 2005). Monological conversations do not adapt to the talk of others because they reject others’ utterances and explanations, contain questions that presuppose one particular answer, and view the other as passive recipient. In contrast, dialogical conversations involve talk that adapts and joins with the previous talk, and is constructed so as to invite further responses and form new, jointly-constructed understandings (Seikkula &
23 Arnkil, 2006). A transformative dialogue thus requires a mutual collaborative participation between all participants, as dialogue is not something that can be created unilaterally.
Consequently, the therapist can only take a dialogical position and hope that this position will be reciprocated by the other participants in conversation. While a transformative dialogue has the potential to create therapeutic change, not all of a therapy session is nor can be dialogical.
Conversation will necessarily involve a variety of responses and interactions that cannot all be dialogical in nature, and monological interactions are necessary in order to, for instance, make agreements or to gather information (Olson, Seikkula, & Ziedonis, 2014).
The transformative view sees dialogue as beneficial due to its ability to create new, multifaceted and collective understandings of a problem (H. Anderson & Goolishian, 1988,
1992; Bertrando & Lini, 2019; Eiterå et al., 2015; Haarakangas, Seikkula, Alakare, &
Aaltonen, 2007; Rober, 2005a). The more voices that are present in a meeting, the greater the number of possibilities for developing new understandings of the problem (Seikkula &
Trimble, 2005) and therefore more possibilities for change. From a dialogical perspective, psychotherapy becomes a process of creating dialogue rather than coming to an agreed description or explanation (Seikkula & Trimble, 2005). Due to a focus on dialogue rather than change, dialogical therapies do not require accounts or theories about the development of dysfunctional interactional patterns (Guilfoyle, 2003) and can thus avoid pathologising explanations.
The ontological and transformative uses of “dialogue” potentially create confusion.
This is because from the ontological perspective the “dialogue” or interaction with others is the mechanism through which people make meaning, but it does not seem to specify that dialogue requires a certain form. This suggests that any form of interaction, even those that are monological, conflictual or abusive, can provide a person with information on their
24 relationship with others, which in turn can inform their views of themselves. In contrast, the transformative view is prescriptive in the sense that only particular types of dialogue are considered helpful or beneficial and they have certain characteristics that stand them apart from monological conversations (Cooper, Chak, Cornish, & Gillespie, 2013). It seems that while all forms or interaction or “dialogue” with others can inform our understandings of ourselves, only particular types of dialogical relations are helpful in promoting what is considered to be positive growth. These two uses of the term “dialogue” are therefore related as the transformative view is built upon and thus arises from the assumptions of the ontological perspective. However, these different uses are also clearly distinct as transformative dialogues can guide people towards particular outcomes that are helpful and therapeutic.
The central principle of the Open Dialogue approach is arguably the creation of dialogue, even to the point that the promotion of dialogue is more important than the creation of change (Seikkula & Arnkil, 2006). However, what constitutes a transformative therapeutic dialogue and how it is promoted is at times difficult to describe. This is partly because the ontological and transformative uses of the term are regularly utilised but not clearly distinguished. There is also a range of writings from a variety of different authors from different theoretical perspectives. To date there has been little to clarify the potential confusion created by the range of writings on the concept of transformative dialogue and how it may be meaningfully applied in therapy. In this article we aim to present and synthesise the writings from the major contributors on dialogical approaches with a specific focus on the practical recommendations for family therapy practice.
Method
25 This article takes the form of a narrative review (Derish & Annesley, 2011; Ferrari,
2015; Green, Johnson, & Adams, 2006) and aims to synthesise and present the main recommendations for therapists on how to practice Open Dialogue and dialogical approaches to family therapy. This review therefore focusses on the transformative view of dialogue and how it has been presented in the literature. The literature for this review began with the writings on the Open Dialogue approach and the primary contributor Jaakko Seikkula. The
Open Dialogue approach incorporates ideas from various other authors. Through citation searches, the literature for inclusion was expanded to include other sources such as the
“reflecting team” work of Tom Andersen; the work on the therapist’s self by Peter Rober; the collaborative language systems approach of Harlene Anderson and Harold Goolishian, and the dialogical-systemic approach of Paolo Bertrando. While the review focusses on and begins with Open Dialogue practices, we expand beyond Open Dialogue to the broader understandings of dialogical practice. We therefore include, where relevant, recommendations from other dialogical approaches, and these are noted when present through the article.
The collected articles were read and the major recommendations for practice were noted. The various recommendations were then compared and grouped by similarity. We present the recommendations grouped into the categories of the therapist’s mindset and the metaphor of polyphony; how to respond to clients during meetings; the involvement of the self of the therapist; and the use of reflections. In each group we describe the presented reasons and benefits of each recommendation before reporting how therapists are advised to bring them into practice. A concise summary of the recommendations is presented in Table 1.
Findings
The Dialogical Mindset
26 Authors often describe the general orientation or dialogical mindset of the therapist.
Similar to the monological/dialogical division, descriptions of the therapist’s mindset contrast two polar opposite approaches: the preferred dialogical position and an undesirable non- dialogical position. The preferred, “dialogical” mindset has been described in a variety of ways, however, they all seem to converge upon a similar type of idea. H. Anderson (2005) and H. Anderson and Goolishian (1992) describe a number of interdependent concepts such as not-knowing, client-as-expert, uncertainty, and conversational partnership. Galbusera and
Kyselo (2018) and Cooper, Chak, Cornish, and Gillespie (2013) describe an openness or a willingness to be changed by the other. Shotter (2006) uses the term “withness” thinking, which involves being relationally-responsive or aware of how the therapist is experiencing the relationship and being responsive to what is required in the moment. Olson, Seikkula, and
Ziedonis (2014) describe this as “responsive listening”, without having an agenda and includes attuning to non-verbal communication. These concepts revolve around the idea that the therapist has expertise in conducting the conversation but not in the details of the clients’ lives. Therefore, through mutual enquiry, the therapist and client attempt to understand the client from the client’s current point of view and not from the clinician’s professional perspective or judgements (H. Anderson, 2002, 2005; H. Anderson & Goolishian, 1992;
Haarakangas, Seikkula, Alakare, & Aaltonen, 2007).
The undesirable alternative to openness is described by Shotter (2006) as “aboutness”: thinking which relates to a more external, theoretical, representational-referential understanding of the other person. H. Anderson (2002) describes this lack of openness as attempting to understand the client through the therapist’s own pre-conceived ideas. This has been described as monological conversation, or trying to convince the recipient of the correctness of the therapist’s position and describe a relationship where the client is not an
27 equal partner in the relationship or conversation (Olson, Seikkula, & Ziedonis, 2014). The dialogical mindset connects with ideas about power (Guilfoyle, 2003) and who has the superior rights and authority to define the client’s situation.
The therapist is advised to adopt a mindset of openness to understanding the client’s point of view and openness to being influenced by the client’s stories. This mindset is summarised by H. Anderson (2012, p. 15) as being a “hospitable host and guest”. This means that therapists communicate the importance of the client and the value of their stories and perspectives, while also remaining aware that they are guests in learning about the clients’ lives. The therapist follows rather than leads the client’s narrative and aims to create an atmosphere of safety that allows the expression of all the present voices as easily as possible
(H. Anderson, 2002; Bertrando & Lini, 2019; Haarakangas, Seikkula, Alakare, & Aaltonen,
2007; Olson, Seikkula, & Ziedonis, 2014; Seikkula, 2008, 2011; Seikkula & Arnkil, 2006;
Seikkula & Trimble, 2005). Finally, therapists are advised to respect and take seriously all contributions and perspectives and avoid coming to one fixed “understanding” of a problem as this can block the development of further dialogue and new meanings (H. Anderson &
Goolishian, 1988).
Polyphony
The dialogical mindset also implicates the metaphor of polyphony, described by
Bakhtin (1984) as the coming together of multiple autonomous co-existing voices. Seikkula
(2008) uses it to refer to the multitude of voices present in a network meeting. This includes what is said by each person in the room (horizontal polyphony) as well as the different
“voices” within each person (vertical polyphony; Seikkula, 2008). These inner “voices” or perspectives in a sense “speak” from the different parts of a person. For example, a person may have different viewpoints from their position as a female, an eldest child, someone who
28 has experienced trauma, as a carer, or as a parent. Polyphony is valued because the presence of multiple voices can guard against a monological perspective. Monologues and ideas of certainty can contribute to feelings of impasse, “stuckness”, and competition over who has the “correct” description of the problem (Rober, 1999, 2005a; Seikkula, 2008; Seikkula &
Arnkil, 2006). These monologues can emanate from the family wanting definitive answers from the therapist about diagnosis or treatment and vice versa when the therapist claims the authority to provide such answers (Seikkula, 2003). In dialogue, a commitment to polyphony views all perspectives as equally valuable, creates more possibilities for change and negates competition over who has the more valid description or diagnosis of the problem. The role and mindset of the therapist is to respond in such a way that promotes further dialogue and creates new understandings that are co-created by all participants (Bertrando, 2007;
Bertrando & Lini, 2019; Seikkula, 2002).
Responding Dialogically in Practice
Open Dialogue places great importance on how the therapist responds to the prior talk of the client, which is informed by the therapist’s dialogical mindset. As a “conversational artist” or an architect of dialogue (H. Anderson & Goolishian, 1992, p. 27) the expertise of the therapist lies not in their ability to identify, diagnose, and intervene in the problems of the family, but rather to respond in a way that promotes polyphonic dialogue. At times a therapist may be invited by the family to respond in a monological way, that is, to provide diagnoses and intervention (Seikkula, 2003). Seikkula (2003) suggests that such monological responses risk the client and family becoming more dependent on the treatment system. A therapist’s responses are therefore designed to explore each person’s inner dialogue so that they are heard by everyone present (Olson, Seikkula, & Ziedonis, 2014; Seikkula, 2008), which increases the possibility for new understandings to be developed by the family (Seikkula,
29 2003). In general terms, the therapist is advised to slow down the conversation to allow silences, time for thinking (H. Anderson, 2002; Haarakangas, Seikkula, Alakare, & Aaltonen,
2007; Seikkula & Arnkil, 2006; Seikkula & Trimble, 2005) and the expression of emotions
(Andersen, 1987; Bertrando & Lini, 2019; Olson, Seikkula, & Ziedonis, 2014; Seikkula,
2011; Seikkula & Trimble, 2005). By allowing time to think and speak, a person is able to put words to past experiences, as well as emotions and current concerns, and thus transform non- conscious embodied experiences into spoken narratives (Seikkula, 2008), which are then open to possible transformation (H. Anderson & Goolishian, 1988).
When responding, therapists are advised to stay close to a person’s meanings and understandings of the problem (H. Anderson & Goolishian, 1988) and these responses should be fitted and follow on from what the client has just said (Seikkula & Arnkil, 2006). The therapists’ own inner dialogues that occur in response to the unfolding conversation may also be a resource that guides their responses as these inner dialogues are potentially meaningful for the family (Rober, 1999, 2011). Therapists may select particular words or issues that stand out to them and then invite the client to explore the meanings and associations of these words (Haarakangas, Seikkula, Alakare, & Aaltonen, 2007). This can involve repeating the exact words of what a client has said in order to encourage elaboration and to preserve their unique ways of viewing the problem (Bertrando & Lini, 2019; Haarakangas, Seikkula,
Alakare, & Aaltonen, 2007; Olson, Seikkula, & Ziedonis, 2014; Seikkula, 2011; Seikkula &
Arnkil, 2006; Seikkula & Trimble, 2005). Other authors suggest that rather than repeating the client’s words exactly the therapist may instead use a similar word of the same meaning or use the therapist’s own words to describe what they think the client is saying (H. Anderson,
2005). This is thought to show that the therapist has understood and is not just repeating the client (H. Anderson, 2005, 2007). In either case, the therapist’s responses are intended to
30 understand and invite further elaboration rather than giving clients new preferred words for understanding their own experience (H. Anderson, 2002) or to interpret or reality check the client’s comments (Seikkula & Arnkil, 2006).
It can happen that when repeating the client’s words the conversation can feel stuck and circling over the same area. In these situations, the therapist can ask questions to invite further elaboration. Preferably these are triadic circular questions (Bertrando, 2007;
Bertrando & Lini, 2019; Olson, Seikkula, & Ziedonis, 2014; Seikkula & Arnkil, 2006) such as asking how a problem has impacted on others or how the relationship between two other people has changed (Selvini, Boscolo, Cecchin, & Prata, 1980; Tomm, 1984, 1988).
Therapists are advised against using questions as tools to gather information, support a hypotheses, or to seed ideas and guide clients to a particular understanding (H. Anderson,
2002). Instead, the purpose of questions is to allow the therapist to be guided by the client and to question the therapist’s own understandings (H. Anderson & Goolishian, 1988, 1992).
In order to promote dialogue, therapists also need to invite each person to talk. This involves managing the speaking and listening time of participants so that everyone has the opportunity speak (Haarakangas, Seikkula, Alakare, & Aaltonen, 2007; Olson, Seikkula, &
Ziedonis, 2014). This can become difficult if one person is speaking more than others and the therapist may have to be active in ending a person’s talk in order to invite other speakers.
There is some variation on how to invite these multiple perspectives. H. Anderson (2012) recommends speaking with each client in turn while the others listen. This involves the therapist engaging with each person’s story individually, being careful not to lead the direction of the conversation, while the others listen without having to respond. In this approach, the therapist actively manages the turn taking in the conversation and is, in a sense, the conduit through which all dialogue flows. Other authors are not so explicit on this point,
31 and it seems possible for family members to speak to each other directly without the intervention of the therapist if this continues to promote dialogue.
Seikkula and Arnkil (2006) suggest beginning a session with an open question and asking everyone to say something early on in the session. The therapist is advised to not preference one idea over another nor seek consensus, but rather to encourage mutual collaboration in understanding the problem (H. Anderson & Goolishian, 1988). In other words, the therapist explores differences rather than negotiating or resolving them (H.
Anderson, 2001; Bertrando & Lini, 2019). The core aim of the therapist’s responding is to further dialogue and not to interrupt it (Seikkula, 2011). However it has not been clearly articulated in the literature how therapists achieve these various goals of mutual collaboration, not seeking consensus and exploring differences and may be an important area of future study.
The Involvement of the Therapist’s Self
Authors regularly stress the importance of bringing the voice of the therapist into the conversation. This has been described in various ways such as self-awareness (Cooper, Chak,
Cornish, & Gillespie, 2013), transparency (Olson, Seikkula, & Ziedonis, 2014), authenticity
(Galbusera & Kyselo, 2018), or being public (H. Anderson, 2001). All share a common perspective that therapists should contribute something from themselves to the dialogue. By contributing to the dialogue, the therapist shows that they are struggling together with the client in their difficulties (Friedman, 1994) and their different ideas can create new meanings
(H. Anderson, 2001; H. Anderson & Goolishian, 1992). Consequently, there have been attempts to create exercises to develop a therapist’s awareness of their inner voices (e.g.,
Mikes-Liu, Goldfinch, MacDonald, & Ong, 2016; Rober, 2010).
32 Again, the primary reason for introducing the therapist’s thoughts into the conversation is to promote further dialogue and should therefore not have an underlying judgement or strategic purpose (H. Anderson, 2002). Therapists are advised to present their ideas not as expert opinions, but as another equal voice contributing to the mutual enquiry into the current problem (H. Anderson, 2005, 2012). The therapist negotiates a fine balance between contributing their voice to the emerging dialogue, and avoiding the creation of monologues and thereby directing or colonising the family with their professional perspectives (Rober & Seltzer, 2010). Bertrando (2007) suggests that the therapist takes a more active role in forwarding their hypotheses and may disagree, challenge or doubt what a client has said (Bertrando, 2007; Bertrando & Lini, 2019).
Rober (1999) makes a distinction between the therapist’s personal self and the therapist’s professional role. Responses from the therapist’s personal self can include their thoughts, memories, images and feelings, while voices centred around the therapist’s role include theoretical concepts, hypotheses, and ideas on how to conduct and manage the structure of the session. When considering introducing the therapist’s personal responses into a session, it is first necessary to become aware of these responses. Through an awareness of their own experiencing, emotional responses, and invitations to act, a therapist can navigate emotionally charged sessions and consider alternative ways of responding that are more constructive (Rober, 2011). Bertrando (2007; Bertrando & Lini, 2019) describes a similar process of finding one’s place. This involves an awareness of the therapist’s own emotions that arise in a session and their attitudes towards the other participants. This also includes an awareness of one’s position in regard to certain domains, ranging from broader social, cultural and political values, to more intimate relationships and individual experiences.
33 Through finding one’s place, therapists are better positioned to help others find and understand their place.
Rober (1999) has described a step-wise process on how the therapist may bring their personal thoughts into the conversation. Firstly, it is important to acknowledge that therapists inevitably experience a range of internal experiences and responses and these are a product of the therapist’s personal history and the current meeting with a family (Rober, 2005a). These inner voices are potential resources (Seikkula, 2008) as they can aid in understanding the therapist’s own position and how they may appropriately respond (Rober, 2011). If personal responses are primarily resulting from the therapist’s personal history then they should not be voiced. The expression of inner thoughts should always promote further openness and dialogue (Bertrando & Lini, 2019). Therefore the therapist needs to carefully consider if their experiences have relevance for the client, and if so, how these experiences can be presented in a useful and respectful way (Rober, 1999, 2005b). If there are any doubts about introducing personal experiences, they should not be voiced and instead be discussed with a colleague or supervisor.
Therapists also have inner dialogues surrounding their professional role. In this regard,
Bertrando and Lini (2019) argue for the usefulness of hypothesising which is generally not part of dialogical practice (Bertrando, 2007) having originated in systemic family therapy
(Selvini, Boscolo, Cecchin, & Prata, 1980). They argue that therapists are always hypothesising about clients and it is better to be aware of these thoughts (Bertrando, 2007;
Bertrando & Lini, 2019). They view “micro-hypothesising” about small events (rather than facts or truths about a family), as a way of creating doubt and uncertainty in the therapist and functions as a way of maintaining neutrality, irreverence, or curiosity in a Milan-systemic sense (Cecchin, 1987, 1992; Cecchin, Keeney, Lane, & Ray, 2018). In this way, micro-
34 hypothesising could be considered a form of inner dialogue. This may help the therapist hold multiple and contradictory perspectives in mind simultaneously without a preference for one idea, thus remaining open to change (H. Anderson & Goolishian, 1988; Bertrando, 2007;
Bertrando & Lini, 2019).
Reflecting Conversations
The public expression of the therapist’s thoughts may occur in the back-and-forth conversation with the family, but can also occur in a more formalised reflecting conversation.
In dialogical approaches, there is some variation in the structure of reflecting conversations, however all share the expectation that the reflection occurs between the clinicians with the family observing. This is stressed to the extent that the therapists should face each other and refrain from eye contact with the family (Olson, Seikkula, & Ziedonis, 2014). The family are expected to not interrupt the conversation but are invited to respond to the therapists’ comments after the reflection is completed. This is said to free the family from having to respond or provide justifications and they can instead focus on listening (Andersen, 1987;
Haarakangas, Seikkula, Alakare, & Aaltonen, 2007; Seikkula & Trimble, 2005). There are a number of potential benefits to reflecting conversations. The therapist is public and transparent in their thinking with the family and this can create a sense of community around the problems (Seikkula & Trimble, 2005). It is also a way of introducing a variety of contrasting perspectives and possibilities to a family that feels stuck around a problem
(Andersen, 1987; Seikkula & Trimble, 2005), and may promote family reflectivity on their situation (Haarakangas, Seikkula, Alakare, & Aaltonen, 2007).
The timing and form of a reflection can occur in a number of ways. Firstly, a reflection can be initiated by one of the therapists at any time during the session (Andersen,
1987). This can occur when the therapist interviewing the family feels it is appropriate or
35 when the reflecting team feels they have something to contribute (Andersen, 1987). A reflection may also be initiated when the therapist is feeling stuck or at a therapeutic impasse
(Rober, 1999). In one form of reflections, a separate reflecting team of around three people observe the session between the family and one or two therapists from behind an observation screen (Andersen, 1987). When it is decided to have a reflection, the participants change places so that the family and the therapists observe the conversation of the reflecting team.
Other variations include the reflecting team operating in the same way but sitting in the same room as the family and the therapists. Another variation has no reflecting team but rather the reflection occurs between the therapists who have been conducting the session. In a further variation, Rober (1999) describes sharing his reflections with the family when he is the sole therapist conducting the session. However, it seems more likely that the reflection occurs between two to three therapists (Andersen, 1987).
The use of reflecting teams is a technique that is utilised by other therapeutic approaches (e.g., White, 1995), but in Open Dialogue, reflections are supposed to have a particular quality. There is a consistent emphasis on therapists presenting their ideas carefully and tentatively with the intention of contributing to the dialogue rather than directing or guiding it (Andersen, 1987; H. Anderson, 2005; Haarakangas, Seikkula, Alakare, &
Aaltonen, 2007; Seikkula & Trimble, 2005). Again, therapists do not strive for agreement but rather a polyphony of co-existing viewpoints (Andersen, 1987). Similar to the advice on how therapists should respond generally, therapists should strive to use the same words that the family has used and highlight ideas that stood out to them. Therapists should also attend to their inner conversation (H. Anderson & Goolishian, 1992) and consider what thoughts, images or experiences and meanings were evoked (Haarakangas, Seikkula, Alakare, &
Aaltonen, 2007; Olson, Seikkula, & Ziedonis, 2014; see also the involvement of the therapist
36 above). If a therapist has an alternative point of view to the family this should be voiced carefully with the right amount of difference so that it can still be heard by the family without offence (Andersen, 1987; Haarakangas, Seikkula, Alakare, & Aaltonen, 2007). Although not writing from a dialogical perspective, Weingarten (2016) describes a good reflection as one that is different enough so that the person has to change and expand to accommodate the reflection. Olson, Seikkula, and Ziedonis (2014) recommend that therapists should discuss treatment plans and recommendations around medication or hospitalisation in reflections. At the end of the reflection the family should always be invited to comment on the discussion of the reflecting team (Seikkula, 2008). This can include questions about their responses, what stood out, or parts that they agreed or disagreed with (Andersen, 1987; Olson, Seikkula, &
Ziedonis, 2014; Seikkula & Trimble, 2005).
Limitations
This review aims to synthesise and present the major practice recommendations on how to work dialogically in Open Dialogue and family therapy. It subsequently includes the writings from the originators and major contributors to Open Dialogue and dialogical approaches. However, there is increasing interest internationally in Open Dialogue and dialogical practices with unique variations of the approach to their local contexts. This means that there is a growing knowledge of unique perspectives and understandings of how dialogue is applied in therapeutic and other settings that is not included in this review. Along similar lines, we do not wish to present these recommendations as immutable directives for practice as one of the strengths of a dialogical approach is the responsiveness to the emerging conversation. Consequently, dialogical practice will be enriched by future contributions from multiple, unique, and varied perspectives. We hope that this review acts as a point of further
37 discussion, revision, correction, and future contributions rather than a prescriptive “how-to” account of dialogical practices.
Another consideration is that our own perspectives as authors will inevitably affect our presentation of these concepts. The first author is a psychologist and family therapist working in a child and youth mental health service and is researching conversational interactions in Open Dialogue sessions. The second author is a nursing academic currently undertaking advanced training in the Open Dialogue approach with a long history of research in Open Dialogue and mental health. We are both involved in providing training and supervision in Open Dialogue. As practicing clinicians, through this review we endeavour to synthesise the various descriptions of Open Dialogue and dialogical practice to inform the work of new and experienced clinicians and to improve training. We acknowledge that our particular interest in examining the conversational practices utilised in Open Dialogue may influence our interpretation of these articles.
Conclusion
In this article, we have attempted to synthesise the main recommendations for promoting dialogue in Open Dialogue family therapy practice. Table 1 presents a concise summary of these recommendations. A particular emphasis is placed on the therapist’s mindset that values the promotion of multiple perspectives and understanding from the client’s point of view. Therapists are also encouraged to be aware of their own inner dialogues and how they can be meaningfully voiced, either in the course of the session or in a reflecting conversation. Across the range of authors reviewed, a consistent theme is that the therapist’s talk should promote further dialogue and all of a therapist’s utterances should emanate upon this core purpose. Despite the different origins of dialogical approaches and the
38 different emphasis of the authors, there is a general consistency in the recommendations for practice.
Authors generally preferred to make general recommendations for practice and were reluctant to be too prescriptive about how therapists should specifically speak. These broad types of recommendations provide for flexibility and adaptability in different situations and seem consistent with the theme of not directing or constraining how clients or therapists should speak. These general types of recommendations also connect with the ontological view of dialogue that takes a more descriptive view of how dialogue operates. However, such descriptions provide limited details on how these processes may work in real life settings, which can be challenging for new clinicians. Ambiguous recommendations also present difficulties when attempting to define what characterises an Open Dialogue approach and how they may differ from other interventions. This makes constructing fidelity criteria and investigating Open Dialogue in traditional forms of quantitative research problematic (Waters et al., 2021). Perhaps with further research and developing practice in different contexts, the details of what characterises a dialogical conversation will become more clearly defined.
As mentioned earlier, there is a clear difference between the ontological and transformative uses of the term “dialogue”. Writings on the clinical practice of Open
Dialogue utilise the transformative view of “dialogue” and provide prescriptive descriptions on what dialogical practice should look like. However, they often offer ontological descriptions as a way of justifying the value of a dialogical approach, which introduces a blurring of the distinction between ontological and transformative uses of dialogue. This creates an inconsistency because in the ontological view dialogue is presented as necessary for meaning making and inevitably occurs when interacting with others, but the transformative view is prescriptive as only particular types of dialogical conversation are
39 therapeutic. While the originators of Open Dialogue present dialogue as something more than an intervention but rather a “way of life” (Seikkula, 2011), this creates a lack of precision and specificity when describing dialogical practices and contributes to ambiguity about what constitutes dialogical approaches. This in turn may create confusion amongst practitioners and potentially undermines the theoretical position of Open Dialogue. Because ontological and transformative dialogues are conceptually distinct, it is conceivable to formulate a dialogical approach to therapy that does not require the ontological perspective.
Open Dialogue presents a promising approach to mental health care that is inclusive and family-centred. However, its departure from treatment as usual and the lack of clear recommendations make it a challenging approach to implement. We hope that this article has simplified the range of practice recommendations from different sources on how therapists may incorporate a dialogical focus in family therapy. We also hope to encourage a more precise use of dialogical language and promote further discussion on what constitutes a dialogical approach that can further refine the conceptualisation and research into Open
Dialogue.
Table 1
Summary of Recommendations for Dialogical Practice
Therapist’s mindset
• Have little, if any, pre-planning before the session about the direction of the meeting
• Start with an open question and follow the lead of the participants
• Attempt to understand the situation from the point of view of each person, not from
your own theoretical perspective or diagnosis
40 Responding
• Slow down and allow silence, time for thinking, and the expression of emotions
• Do not interpret or reality check
• Design responses to explore and understand participants’ meanings and invite
further elaboration
• Encourage elaboration by repeating the same words or parts of the talk used by the
participant, or use your own words with very similar meanings
• Ask questions if the conversation seems to be stuck when using repeats. Aim to
understand rather than guide or direct the participants
• Ask each person to contribute to the conversation
• Assess whether your responses are promoting or hindering dialogue
Involvement of the therapist
• Have an awareness of your inner conversation (both personal and professional)
• Consider if voicing your thoughts will promote or hinder dialogue and if they can be
presented in a helpful way
• Hold in mind multiple perspectives or hypotheses to maintain neutrality
Reflections
• Use reflections as a way for therapists to share their thoughts with each other in
front of the family or to negotiate therapeutic impasse
• Voice reflections tentatively and respectfully with the intention to promote further
dialogue
• Include aspects of your inner dialogue as appropriate
• Ground your reflections through reference to the words of the family
41 • Always invite the family to respond after a reflection has finished
42 Chapter Summary
Family therapy has undergone a number of changes ranging from mechanistic ideas about family systems to approaches that emphasise equality, narrative, and subjectivity. As one of the recent approaches that has emerged from this history, Open Dialogue places particular emphasis on dialogue. However, Open Dialogue is difficult to conceptualise as a form of family therapy, having been positioned as a “way of life” rather than a method or intervention. Furthermore, the central principle of dialogue is somewhat ambiguously conceptualised in current writings, with recommendations on how therapists should conduct themselves in sessions described in terms that emphasise a particular mindset. While Open
Dialogue may not be considered an “intervention”, it still contains a prescribed set of behaviours which have a proposed benefit to clients, even though those recommendations are ambiguously defined. This thesis aims to explore dialogical conversations to identify what practices are employed by therapists and how they may characterise a dialogical approach.
This project necessitates an analytic approach focussing on the micro details of how conversation is produced and constructed in interaction.
43 Chapter 2: Method
Conversation Analysis
Conversational interaction is central to social organisation and human life. Social interaction is used to coordinate social actions; create, affirm, or deny identities; transmit and modify aspects of culture; and to create shared meaning and understanding (Goodwin &
Heritage, 1990). Conversation analysis (CA) is a way of examining conversational interactions that combines language, culture and social organisation (Sidnell & Stivers,
2013). CA seeks to describe the “rules, procedures and conventions” that underlie and make social interaction possible (Goodwin & Heritage, 1990, p. 283). It does this through a focus on talk as it occurs in actual conversations, describing the interaction of the participants through their displayed orientations rather than through inferences about psychological motivations, and through an awareness of how context is both constructed and modified by the interaction. Conversation analysis presents a unique approach to studying interaction and has the benefits of being rigorous and empirical with an established research base. In this chapter, I will briefly outline the historical origins of CA in order to provide a thorough presentation of its theoretical origins and assumptions. I then provide a description of the central concepts of the approach, and how the conversation analytic process is conducted.
This overview will hopefully sensitise the reader to how CA conceptualises and approaches the analysis of interaction. I also discuss other approaches to analysing conversations and the reasons that I have decided to use conversation analysis.
Historical origins of CA
A central area of investigation for social researchers is to describe how social order is accomplished in society (ten Have, 2012). Sociology has sought to investigate social actions via an approach championed by Talcott Parsons (ten Have, 2012). In a Parsonian view,
44 cultural values become internalised as part of a person’s personality, and these drive social interactions (Heritage, 2013b). Meaning and mutual understanding are thought to be possible due to a pre-existing shared set of common knowledge and symbols. In a Parsonian approach, research begins with developing a set of pre-determined theoretical categories or collective dispositions. These are then used to explain motivations and behaviour, with little focus on conversation and language which was instead the focus of linguistics (Heritage, 2013b; ten
Have, 2012). However, the approach of linguistics in the 1960s, influenced by speech-act theory and the work of Noam Chomsky, focussed on studying de-contexualised, isolated, and idealised sentences constructed by linguists (Heritage, 2013b). This is because actual conversations were considered to be disorderly, making analysis difficult as it distracted from the idealised grammatical structure of sentences (Heritage, 2013b).
In contrast to these approaches, Harold Garfinkel developed an approach called ethnomethodology (Garfinkel, 1996). Ethnomethodology starts with social interactions as the topic of analysis rather than assumptions about pre-existing knowledges or idealised grammars (ten Have, 2007). This is because pre-theorised understandings of social interaction interprets ordinary activities through pre-determined concepts and methods
(Garfinkel, 1996). In ethnomethodology, there is an assumption that contained within the design of people’s social actions are elements that make those actions intelligible to others.
Therefore, analysts need only look at the interactions themselves. Ethnomethodology is therefore able to find “what more” exists in the interactions than what is provided for by pre- theorised understandings of interaction (Garfinkel, 1996, p. 6). Ethnomethodology looks at everyday interactions that are routinely produced and understood (but also largely implicit and unnoticed), and how they are made recognisable to others. The goal of
45 ethnomethodology is thus to explicate how social order is intelligibly produced, perceived, and maintained by people in interaction (ten Have, 2004).
Ethnomethodology was a major influence on the development of conversation analysis (ten Have, 2007). CA also studies how people produce meaning and social actions at a local level, but with a particular focus on conversational interaction. Through a focus on conversation as it actually occurs, CA has revealed that social interaction is organised at a minute level of detail including those elements previously considered to be “randomising factors” such as pauses, restarts, “ums”, and “ahs” (Heritage, 2013b, p. 235).
CA incorporates a number of properties that distinguish it from other analytic approaches (ten Have, 2007). First, CA uses recordings and detailed transcriptions rather than summarised data such as coding or statistical analyses, and can therefore be described as operating “closer to the phenomena” under study (ten Have, 2007, p. 9). Second, CA uses recordings from real-life conversations rather than data from experimental or contrived situations. CA studies talk as it is “actually used” by participants rather than what is considered “correct” or “grammatical” and is therefore data-driven rather than researcher- driven. Third, CA views social interaction as organised at a minute level of detail with particular normative expectations that are oriented to by all participants. CA seeks to describe these conversational structures and how they achieve social actions (Sidnell & Stivers, 2013).
CA therefore avoids analyses based on individual motivations or “why” people are acting in a particular way, and instead attempts to describe “how” people achieve social actions. In the next section, I describe some of the central concepts of CA that are relevant to the current thesis.
46 Central Concepts in CA
Turn-taking
CA has revealed that when people talk, they follow certain normative expectations
(Heritage, 2013b). One such aspect of conversation is the system of turn-taking (Goodwin &
Heritage, 1990; Sacks, Schegloff, & Jefferson, 1974). This system consists of turn- constructional units, separated by transition-relevance places, where turns may end and speaker change may occur. Turn-constructional units may include sentences, single words, phrases or clauses. At the end of a turn-constructional unit there are a number of possible options: the current speaker can select the next speaker (for example, through gaze or name), a different speaker may self-select, or the same speaker may continue talking. Listeners monitor the flow of a turn for possible points when a turn will end and where they can begin speaking. At these transition-relevance places there can be overlap with more than one person speaking at once or a short gap between people's turns. Usually this overlap is brief with one person yielding the turn to another. Following on from this system of turn-taking is the concept of sequences.
Sequences
A central idea within CA is that the action an utterance is “doing”, is understood in relation to its position within a sequence of talk (ten Have, 2007). This means that turns are not understood in isolation, but in relation to what has been spoken prior, and how they constrain and influence the interpretation of the following turn. Sequences are a fundamental area of analysis (Heritage, 2013b). A core concept in the organisation of sequences is the adjacency pair, which consists of an initiating, first position utterance or first pair part that makes expectably or conditionally relevant the production of a second position, reciprocally related second pair part, from the recipient (Schegloff & Sacks, 1973). For example, a first
47 pair part of a greeting makes conditionally relevant a greeting in return; a question makes relevant an answer, and an offer makes relevant an acceptance or rejection. The normative nature of adjacency pairs is such that the absence of the expected second pair part is sufficiently unexpected to be worthy of comment and even censure. For example, if an answer is not produced in response to a question, then the speaker can pursue an answer with follow-up prompts or questions (see example 1).
Example 1 (Atkinson & Drew, 1979, p. 52)
1 A Is there something bothering you or not? (1.0) 2 A Yes or no (1.5) 3 A Eh? 4 B No.
Similarly, if a recipient does not know the answer, they may orient to their inability to provide an adequate response through a reason or an account for not knowing. In Example 2,
M is not able to provide an answer to J’s question and provides a reason i.e., “she hasn’t said”.
Example 2 (Heritage, 2013b, p. 250)
J: But the trai:n goes. Does th’train go o:n th’boa:t? M: .h .h Ooh I’ve no idea:. She ha:sn’t sai:d
Adjacency pairs are thus connected with authority (see section below on deontic authority) and are a structural mechanism by which people constrain and hold each other accountable for producing certain responses. They also demonstrate that turns at talk cannot be analysed outside of their sequential context. This is because initiating turns make relevant particular kinds of responses, and responses have particular meanings depending on what types of turns they are responding to. In addition, responsive turns make relevant certain types of responses in a developing causal chain. The concepts of sequence, adjacency pairs,
48 and conditional relevance indicate that conversation follows a principle of enchrony (Enfield,
2013). This means that conversation is progressive and forward-moving with a causal connection between turns. This concept is developed in more detail in Chapter 8 as a way of bringing together the central ideas of this thesis.
The sequential organisation of conversation has a number of implications for the study of conversation, including the creation of context and displays of participant orientations. Context can be understood in a couple of ways. One is that context is something external within which the interaction occurs, for example, a classroom, a courthouse, or a mental health service. These external contextual features provide cues that participants can use to interpret the interaction that occurs. In CA, context is considered at a local level and is demonstrated, created and renewed by sequential interaction itself (Heritage, 2013b). For example, despite the physical external context of the clinic setting and consultation room, not all talk that occurs in the clinic is “institutional”. A therapist may begin a session with talk about the weather or sports, which is contrasted with the therapist’s questions about the family’s personal problems. Personal questions about family problems do not occur in other settings and demonstrate the therapist’s authority and rights to ask about these personal issues, and communicates the beginning of the “therapeutic” part of the session. Context is thus not only external, but also exists within sequences of interaction, and are a way in which
“institutions are ultimately and accountably talked into being” (Heritage, 2013b, p. 290).
Another consequence of the sequential structuring of interaction is the creation of participation frameworks where the roles of participants are constituted and transformed through interaction (Goodwin & Heritage, 1990). On a basic level this involves the complementary positions of speaker and hearer. When a person is speaker, the hearer displays their recipiency through body positioning, gaze, and by forgoing the opportunity to speak,
49 which ratifies the other person’s position as the speaker. The talk and actions of a speaker’s turn also contribute to the participation framework. For example, if the speaker asks a question, this categorises the recipient’s identity as one who has the information and obligation to answer. Participant identity is also constituted in more complex ways. For example, in Chapters 4 and 5, therapists utilise various conversational practices to mitigate their own authority and defer to the authority of the clients, as well as promoting flexibility and agency. This presents a more symmetrical distribution of authority between therapists and the family. A speaker’s identity, the action of their talk and the participation frameworks are therefore reflexively connected and contribute to the interactional context (Goodwin &
Heritage, 1990).
Stance
The concept of stance involves a relationship between a speaker and a stance object.
This relationship can take a variety of forms such as evaluation, assessment, affect, or epistemic. Stance is produced sequentially through linguistic stance acts comprising three subsidiary acts (Du Bois, 2007). These include a stance taker voicing an evaluation of a stance object. This evaluation reciprocally positions the speaker in relation to that stance object. Subsequently, a second speaker can voice their own evaluation and reciprocal positioning towards that same stance object. The degree of concurrence between the two speakers is referred to as alignment and varies along a continuum. In an example from
Rossen et al. (2020), the stance object is a father talking to his son about his suicidal thoughts and the mother (the stance-taker) makes a negative evaluation of the stance object i.e., the father should not speak to their son about the father’s suicidal thoughts. The father has an ambivalent evaluation and stance position towards his actions (the stance object) by agreeing that it was “unhealthy” (to talk to his son about suicidal thoughts) but also justifies his actions
50 by saying that he had no one else to talk to. The mother and the father thus have disaligning stance positions.
Stance involves a public expression of an evaluative position. Through taking an evaluative position, stance is associated with a person’s values and identity and, due to the public nature of stance acts, people become responsible and accountable for the voicing of stance positions. The stance-taker is also an agent in that they are controlling and composing their displayed stance while also being accountable for that stance (Enfield, 2017; also see
Chapter 8). The elicitation and taking of stance positions is a prominent recurring theme in the studies in this thesis. Two types of stance: epistemic and deontic, were particularly prominent in the analyses and are discussed further below.
Epistemic authority
A regular feature of interaction is differences in understandings and knowledges
(Drew, 2018b), referred to as epistemics (Heritage, 2013a; J. D. Robinson, 2013a).
Knowledge is not exclusively internal and subjective, but is displayed to others through interactions that index a speaker's epistemic status (Heritage, 2018; Raymond, 2018).
Furthermore, interactants closely monitor each other’s epistemic positions and design their turns accordingly. As language is a public expression of knowledge claims, CA studies how these claims are “made, manifested, warranted, used and disputed” in interaction (Drew,
2018a, p. 5).
In describing how people make and respond to knowledge claims in interaction, a distinction is made between a person’s epistemic status and stance (Raymond, 2018).
Epistemic status refers to what a person knows or their level of knowledge in a particular domain. Epistemic stance is a person's displayed position in relation to knowledge. For example, if a person asks a question this can suggest a relatively unknowing stance, while a
51 statement implies a knowing or telling stance. Epistemic stance and status interact and often have a complementary relationship where a person with high epistemic status will take an epistemic stance of telling or informing, and a person with low epistemic status often takes a stance of asking. But status and stance are not always aligned, such as when a person with high epistemic status presents a downgraded epistemic stance. For example, a teacher can ask a student a question, not because they have a low epistemic status, but to assess a student’s knowledge (Heritage, 2013b). If there is an incongruence between status and stance, people tend to interpret an utterance by reference to the speaker's status (Raymond, 2018).
Epistemic status and primary rights to knowledge are connected to a sense of self and personal identity (Heritage, 2018) and are displayed in contested epistemics i.e., when there is competition over who has epistemic authority. Heritage and Raymond (2005) found that first position assessments carried superior epistemic rights and authority. This is because first position upgrades were rare and instead commonly downgraded. In contrast, second position assessments were regularly upgraded to present the assessment as independent and not merely responsive to the initiating assessment. This resists the epistemic authority of speakers in first position and contradicts the implication that second assessments were merely responsive to the initiating speaker.
Ekberg and LeCouteur (2012) demonstrated how therapists using cognitive behaviour therapy present proposals and recommendations to clients with a downgraded epistemic authority. In Example 3, the therapist’s (T) recommendations contains low modality verbs
(such as could, may, maybe), I don’t know, minimising terms, delaying devices, and recommendations in the form of questions that the client could accept or decline. The client
(C) resisted the therapist proposal and asserted their own epistemic authority.
Example 3 (Ekberg & LeCouteur, 2012, p. 232)
52 1 T: You could have some sort of way of just (0.8) 2 no:t (0.8) confronting situations I spose there. 3 (1.3) 4 T: where this is gonna come up.=For example 5 you could just not have a cup of tea. 6 C: Mm. 7 (0.7) 8 C: I thought [of] that already hehh
The issue of epistemic rights, how they are presented and negotiated, and their resulting connection to identity are central to interaction. As will be demonstrated in Chapters
5, 6, and 7, Open Dialogue therapists are sensitive to epistemic rights and downgrade their epistemic authority in deference to the authority of the client and family. By downgrading their epistemic authority, therapists also promote flexibility in how clients may respond, thus also promoting client agency (Enfield, 2017).
Deontic authority
CA research has also demonstrated that people regularly attend to issues of power in conversation (Stevanovic & Peräkylä, 2012). As CA is focussed on interactions, power is approached as the displays of the relative rights and responsibilities rather than at a broader level of social discourses. The right to determine actions in a certain domain is referred to as deontics (Stevanovic & Peräkylä, 2012). Deontics can include determining the allocation of speaking turns and selecting next speakers, managing the initiation and closing of sequences of talk, and making requests and decisions (Stevanovic & Peräkylä, 2014). Similar to epistemics, people are said to have differing levels of deontic status, or the capacity to determine future actions in a particular domain, and deontic stance or how deontic authority is presented in interaction (Stevanovic, 2018; Stevanovic & Svennevig, 2015).
Deontic rights are constructed and negotiated through interaction (Stevanovic, 2018).
This is demonstrated through the initiating actions of adjacency pairs as they impose certain
53 obligations on others to respond. For example, an initiating turn can take the form of an assertion, which is a statement about a future fact that requires the receipt rather than the acceptance of the recipient (Stevanovic & Peräkylä, 2012). A similar situation occurs with directives, which involve telling another person what to do without regard for their willingness to comply and do not project non-compliance as a possible response option
(Craven & Potter, 2010). Assertions and directives present a strong deontic stance because they do not require the agreement or non-compliance of the recipient. They also index the speakers’ claims about having the deontic status to make these particular assertions and directives to the recipient.
Alternatively, speakers can utilise a range of practices to present a downgraded deontic stance that is incongruent with their deontic status (Stevanovic, 2018; Stevanovic &
Peräkylä, 2012). These include making proposals that require the acceptance or approval of the recipient such as, positive interrogatives (Asmuß & Oshima, 2012), using conditional verbs such as “could”, providing options, providing accounts or rationale (Svennevig &
Djordjilovic, 2015), the use of “I’m wondering” (Curl & Drew, 2008), or constructing a proposal as a thought that reduces the imposition on the other to respond (Stevanovic, 2013a).
These practices invite the collaboration and joint decision making of the recipient and therefore present and invite the acceptance of a more symmetrical distribution of deontic rights. However, in situations where a speaker with high deontic status presents a downgraded deontic stance, recipients may still orient to the deontic status of the speaker in determining the action of the speaker’s turn (Stevanovic, 2018). Deontic status and stance are thus intertwined and the relative rights of the participants in a conversation are produced and negotiated through sequences of interaction.
54 In psychotherapy, Ekberg and LeCouteur (2014) demonstrate how therapists can exercise deontic authority in directing the conversation, but by implicating the perspective of the client and creating an opportunity for them to disagree, develop a more symmetrical distribution of deontic rights. Clients expressed their resistance to suggestions through examples from their life and thus claimed greater epistemic authority than the therapist
(Ekberg & LeCouteur, 2015). This demonstrates the connection between epistemic and deontic rights as well as the complex ways in which therapists and clients display and negotiate their respective rights.
An analysis of deontic authority is particularly relevant to the study of Open
Dialogue. This is because Open Dialogue places emphasis on collaborative decision making with clients and families, and equality between the polyphony of voices. However, therapists inherit greater deontic authority than clients through their institutional position as “therapists”
(Hare-Mustin, 1994). This creates a situation where dialogical therapists have to continuously counter these existing power relationships (Guilfoyle, 2003). The studies in this thesis demonstrate how Open Dialogue practitioners are sensitive to issues of deontic authority and make attempts to downgrade their authority in various ways. Some downgrading practices do not specify a particular type of response from recipient thus promoting responsive flexibility and client agency.
Repetition
The repetition of prior talk can occur in numerous positions and by different speakers.
Speakers can repeat parts of their own talk for the purpose of self-repair (Schegloff,
Jefferson, & Sacks, 1977), or people can repeat talk of a prior speaker. Repeating the talk of a prior speaker has a wide variety of functions. For example, a partial repeat of the prior turn can mark the location of a problem of hearing, and a full repeat can signal adequate hearing
55 but troubles in understanding the meaning or action of a prior turn (J. D. Robinson & Kevoe-
Feldman, 2010). Repeats can function as response tokens displaying adequate receipt of the prior turn (Maynard, 1997; Schegloff, 1997) or as a form of topic management by nominating a part of the prior talk as a topic for further elaboration from the prior speaker (P. Brown,
1998; Kim, 2002). These repeats implicate a degree of deontic authority as they maintain a particular topical focus in a conversation. Repeats as requests for elaboration have been studied in psychotherapy conversations where they function as therapist requests for elaboration without transforming the client’s talk and thus reducing the introduction of the therapist’s biases (Knol et al., 2020). In Chapter 6, I present a study detailing the various types of repeats utilised in Open Dialogue meetings. One form of therapist repeat specifies a topical area for elaboration, an exercise of deontic authority, but does not specify a particular type of response thus promoting responsive flexibility and agency in the recipient.
Disagreement
Disagreement in conversation necessarily occurs in a responsive turn and may involve the simple voicing of different views (Marra, 2012). However, disagreement may not be explicitly voiced, for example a lack of response, a withholding of agreement, or a change of topic can implicitly signal disagreement. An analysis of disagreement involves not only attention to the content of utterances, but also to the composition or form of disagreements, as well as the relational implications (Angouri & Locher, 2012).
One environment where agreement is made strongly relevant is after an initiating turn containing an assessment. Second assessments offer an evaluation on the same object as the initiating first assessment and are often pursued by the person who made the first assessment
(Thompson, Fox, & Couper-Kuhlen, 2015). Second assessments display not only agreement or disagreement but also the speaker’s stance position towards the assessable object (Du Bois,
56 2007). Assessments display the relative claims of epistemic authority between the two speakers (Thompson, Fox, & Couper-Kuhlen, 2015). By virtue of being presented first, first assessments claim superior knowledge and agency, as second position assessments are presented in response and are interpreted in relation to the first assessment. People can design their turns to either mitigate or upgrade the authority that is attached to the sequential position of their turn, and present a more symmetrical authoritative relationship between the speakers.
For example, first assessments can be downgraded to mitigate their authority while second assessments can be upgraded to claim greater epistemic authority.
In Open Dialogue, the reflecting conversation is an opportunity for therapists to present new ideas to the family. Reflections are an environment where contrasting perspectives and disagreement may be preferred types of responses, similar to political debates, radio talk shows, focus groups, and problem-solving in business situations (Angouri,
2012; Hutchby, 1992; Myers, 1998; Sifianou, 2012). However, reflections are not intended as a confrontational type of talk and therapists have to be tentative in their disagreement so as to not cause offence. Reflections involve the therapists presenting their thoughts about the family, sometimes in the form of assessments. This creates the potential for epistemic asymmetry and negotiation between the therapists. In Chapter 7, I present an analysis of how therapists elide agreement and present a downgraded epistemic stance during reflecting conversations. This allows therapists to present a variety of different ideas without voicing explicit disagreement with each other, which reduces the possibility of breaches to social solidarity. It also cedes epistemic authority to the clients thus downgrading the therapists’ own epistemic authority.
57 In this section, I provided a general description of CA, its assumptions and areas of interest, as well as concepts and research that have been particularly relevant for this thesis.
In the next section I provide an overview of the CA analytic process.
CA Methods
There is not a highly detailed step-by-step structure for doing CA, but there are general steps, questions, and areas to consider in an analysis. The CA analytic process (as described by ten Have, 2007, 2012) first begins with an “unmotivated looking” (Psathas,
1995). This involves an immersion in the data through repeated listening and viewing of recordings and transcripts for phenomena of interest rather than having pre-defined goals (ten
Have, 2007, 2012). After identifying a particular phenomenon of interest, the analysis can proceed in a couple of ways. First, the analyst can seek to understand what actions the participants are doing in their respective turns. These descriptions of actions, which are inevitably based on the researcher’s own experience as a member of a culture, are supported through a comparison with the subsequent talk to understand how the initial talk was understood and responded to by the recipient/s. The analysis may then turn to explicating what conversational practices participants used to jointly accomplish these actions.
Alternatively, an analysis can begin with identifying a particular conversational practice and then investigate the range of actions that this practice may be doing (Schegloff, 1997). An analysis can thus begin with either practices or actions (Kent & Kendrick, 2016). Throughout the process the analyst considers questions such as: how else could the turn have been designed? What other action/s might the turn be doing? How else might the action have been done? (Drew, Walker, & Ogden, 2013). Each example is analysed as an individual case without pre-determined outcomes. The aim is to identify normative structures and patterns of interaction and how these apply to wider social processes (Heritage, 1988). The analysis can
58 then move from a single case study to developing a collection of similar examples that can support and further refine the analysis.
Validity checks
Potter (1996) describes four ways that constructionist approaches such as CA can maintain validity in the analyses. These include understanding utterances through providing data extracts to support the analysis and the next-turn proof. The next-turn proof is the analysis of an initiating turn through how it is responded to by the recipient. This follows from the concept of adjacency pairs described above, where initiating and responding turns are mutually defined. The next-turn proof uses the displayed orientations of the participants rather than their intentions or internal motivations. Analytic validity also comes from findings having congruence with prior research and deviant case analysis. Deviant case analysis involves finding examples that do not fit with the proposed analysis and looks at how the participants in the interaction respond to these deviations from expectations. If participants orient to these unexpected departures this can provide support for the analysis. This is because conversation that flows without difficulties follows normative patterns, but departures disrupt the expectations and expected talk of the participants.
CA and psychotherapy
As an institutional form of talk, psychotherapy differs from everyday conversations due to having certain therapeutic goals. Consequently, CA research on psychotherapy has a unique focus on describing how conversational practices are utilised to achieve institutional tasks, and to understand how referential meanings, emotions, and relationships change through sequences of interaction (Peräkylä, 2019). In order to analyse these types of interactions, Peräkylä (2019) describes a process of identifying sequences of interaction centred around a target action and its surrounding turns. These sequences include 1. Prior
59 action, 2. Target action, 3. Response, 4. Third position turn. For example, through analysing the prior action and the target action Weiste and Peräkylä (2013) demonstrated how therapist formulations can include different types such as highlighting, rephrasing, and exaggerating.
Through analysing the following turns they found that these different types of formulations invited different responses such as confirmation, extended agreement and elaboration, or disagreement, respectively.
The various forms of psychotherapy include particular professional stocks of interactional knowledge (Peräkylä & Vehviläinen, 2003). These include models or theories about interaction that are shared by therapists and inform and shape professional practice.
Depending on the approach, these stocks of interactional knowledge can vary from well- defined prescriptive practices to general descriptions of concepts that are only occasionally utilised. Because these stocks of interactional knowledge are usually generated from theoretical models of psychotherapy, there is a role for CA to examine these stocks from an alternative perspective. CA can be used to falsify or correct, provide more detailed descriptions, add new dimensions, and to expand the range of practices that are described in current stocks of interactional knowledge (Peräkylä & Vehviläinen, 2003).
In this thesis I draw on the general approaches to conducting CA as well as the sequential analysis described by Peräkylä (2019) focussing on various therapist practices as target turns and how they are responded to by other therapists and family members. Chapter 4 focusses on the target action of reflection proposals, Chapter 5 on the conversational practice of “I wonder” and its role in eliciting stances, Chapter 6 targets therapist repeats of the family’s prior talk, and Chapter 7 focusses on therapist reflecting statements. In each chapter,
I discuss how these practices and their interactional functions may connect with or challenge
60 the existing stocks of interactional knowledge (Peräkylä & Vehviläinen, 2003) that are described in the Open Dialogue practice literature.
Other analytic approaches
In this section I describe some of the approaches that have been used for analysing conversations and my reasons for utilising Conversation Analysis.
Dialogical sequence analysis
Dialogical sequence analysis is an approach to analysing talk in psychotherapy based on a particular group of theoretical concepts (Leiman, 2004). Like Open Dialogue, these concepts are inspired by the philosophical works of Bakhtin (1981, 1984, 1986) and
Voloshinov (1972) and centre around the structural composition of talk, namely the addressee, the topic, and the author (Leiman, 2004). Addressees or the recipients of talk influence the design of the talk that is addressed to them. This is because speakers anticipate how the addressees will respond and design their turns accordingly. This is similar to the CA concept of recipient design (Houtkoop-Steenstra, 2000). The addressee is a broad concept and can refer to those who are present, internal states of the speaker (such as a critical aspect of themselves), or more abstract concepts such as social values. The referential object of the turn refers to the content or topic of the talk. By speaking about an object the speaker forms a relationship towards that object as well as the addressee, which is similar to the concept of stance described earlier (Du Bois, 2007). The author of an utterance refers to the speaker, but this can be a complex concept because, like the concept of multiple addressees, the author can speak from multiple positions. For example, a person may talk about themselves from the perspective of another person, such as a parent, a friend, or from society generally.
Authorship then becomes “double-voiced” in that the voice of the speaker is blended with that of an outside observer.
61 Dialogical sequence analysis proceeds by investigating the various interactions, also described as dialogical patterns, between the structural components of addressee, referential object, and author (Leiman & Stiles, 2001). Repeating dialogical patterns are an indication of the self-states of a person and can be used to assess the degree of mental assimilation of the various problematic experiences in their life (Leiman & Stiles, 2001; Stiles et al., 2006).
Dialogical methods for investigations of happening of change
Another approach, this time originating from inside the Open Dialogue field, was developed by Seikkula, Laitila, and Rober (2012) called Dialogical Methods for
Investigations of Happening of Change. The authors contrast this approach with Dialogical
Sequence Analysis as better able to analyse the conversations with more than two participants. This approach combines different concepts such as conversational dominance, symbolic meaning and dialogicity (Seikkula, 2002), as well as other analytic approaches such as dialogical sequence analysis (Leiman, 2004) and the Narrative Process Coding System
(Laitila, Aaltonen, Wahlström, & Angus, 2001). The approach is designed to explore the micro-processes that occur in family therapy interactions with particular attention to the dialogical aspects of conversation.
The analysis involves first dividing the therapeutic conversation into topical episodes.
These topical episodes are then divided into sequences including an initiating utterance and how it is responded to, similar to the CA concept of adjacency pairs as described above.
These responses are categorised according to dominance and dialogicity (Seikkula, 2002).
Dominance is divided into three types: quantitative dominance, or who speaks for the most amount of time; semantic dominance or who introduces new content, words or themes into the conversation; and interactional dominance or who influences the initiation and actions of the other speakers. Responses are also categorised by the extent to which they engage with
62 prior talk and await a response, described as dialogical, or whether they reject other utterances and do not adapt to others, described as monological. Utterances are also categorised as indicative (referring to concrete things) or symbolic (referring to other words or concepts that do not exist tangibly). Utterances are also coded according to whether they are external, describing events; internal, referring to one’s experience of those events; or reflective, referring to one’s emotions and relationship to events and multiple meanings
(Laitila, Aaltonen, Wahlström, & Angus, 2001).
Based on Dialogical Sequence Analysis, the topical episodes are also analysed for voices, addressees and positioning. The authors note that the concept of voices is difficult to define. Voices are described as “traces” within a person that are activated in the current interaction. These voices can refer to past experiences, the perspectives of other people, or a person’s personal or professional roles. Analysis of voices is not only about the content of a person’s utterances, it also includes attention to the “nuance” or the “sense” of a person’s use of those words (Seikkula, Laitila, & Rober, 2012, p. 669). Positioning refers to the point-of- view or perspective from which a person is speaking. This can be described by positions of agreement or disagreement with others. Positioning can thus reflect the different viewpoints of the participants and the relationships that these positions subsequently create. Addressees can refer to the recipient of an utterance. This can be the person who is spoken to directly but can also include the others who are present and observing the interaction. Utterances can also be directed to addressees that are not present such as an absent or deceased parent or broader addressees such as ideologies. Like the idea of voices, addressees can be difficult to identify and involves a degree of interpretation and inference by the analyst.
The approaches of Dialogical Sequence Analysis and the Dialogical Methods for
Investigations of Happening of Change are theoretically consistent with many of the ideas
63 from Open Dialogue as both have a basis in the thinking of Bakhtin (1981, 1984, 1986).
However, this creates the risk of analysing talk from a perspective that is potentially self- affirming rather than providing an outside perspective that can contribute new ideas and potentially critique Open Dialogue practices and its stocks of interactional knowledge. These approaches also cover a wide range of conversational concepts that are difficult to define and require a high degree of subjective interpretation by the analyst.
In this thesis, I wish to approach to the data from an alternative perspective to see what an approach such as CA, which developed separately to the field of psychotherapy, was able to offer. CA is based on the observable details of talk with limited inferences about the internal states of the participants and provides an alternative analytic perspective that is based on conversational practices and the negotiation of social actions. CA also has a broad research base covering various aspects of interaction and multiple sources and articles for comparison. CA thus combines an empirical analytic approach with built in validity measures and an external perspective on Open Dialogue.
Introduction to Methodological Critique
At the time that my literature review was undertaken there was no available research on Open Dialogue using CA. However, Open Dialogue can be viewed as a type of family therapy as it involves sessions with a social network (including families and professionals) and also incorporates a number of family therapy practices. I therefore sought to review the literature on CA and family therapy as a way of understanding the current findings and areas of interest in the field and to better understand how CA had been conducted and utilised. In the following article published in Family Process (Ong, Barnes, & Buus, 2020a), I undertake a review into past research on CA and family therapy in the form of a methodological critique. There has been one previous methodological critique by Tseliou (2013), however
64 this incorporated both CA and discourse analysis, as well as reviewing studies including both couple and family therapy. There are significant methodological differences between CA and discourse analysis as well as differences in the approaches to couple and family therapy. In this review, I focus specifically on CA and family therapy and outline the main methodological strengths and shortcomings of CA research into family therapy, culminating in a number of recommendations for future research. As there is no previous agreed upon quality criteria for CA research, I also conducted a review of recommendations for CA research and propose a set of quality criteria specifically for CA research. This review makes an argument for more rigorous application of CA analytic processes and quality criteria that I hope translates into my own research.
65
Conversation Analysis and Family Therapy: A Critical Review of Methodology
,† BEN ONG* SCOTT BARNES‡ NIELS BUUS*,¶,**
This article critiques the use of conversation analysis (CA) as applied to the study of family therapy. Searches of relevant databases and journals as well as citation searches were conducted in April 2018 for relevant articles. Inclusion criteria included the explicit use of CA either solely or in combination with discourse analysis and discursive psychol- ogy. This resulted in the inclusion of 25 articles that were reviewed against a guideline for the evaluation of qualitative research to which five items specific to CA were added to ensure a specific and balanced evaluation of the studies. Articles generally had a good application of quality criteria although there was a variation in detail of transcription, application of sequence analysis, and a limited use of validity testing. CA has the potential to complement existing research on family therapy but requires a rigorous application of process and quality criteria. The article provides recommendations for future CA research into family therapy.
Keywords: Conversation Analysis; Family Therapy; Review
Fam Proc 59:460–476, 2020
INTRODUCTION onversation analysis (CA) seeks to describe the structural components of everyday, spoken Cinteraction and the particular normative expectations for how it is conducted (Have, 2007; Sidnell & Stivers, 2013). For example, CA has described how talk is divided into turn constructional units followed by transition relevance places where speaker change may expectedly occur (Sacks, Schegloff, & Jefferson, 1974). Speakers also utilize particular resources for selecting the next speaker, extending turns, or suspending talk to repair problems of hearing or understanding. CA has shown how systems of talk are flexible and are imbued with normative expectations that are utilized and modified by participants in their particular context. From early in its development, CA has been concerned with both everyday conversation and institutional talk. The latter involves particular institutional-specific goals,
*Susan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, University of Sydney, Camperdown, NSW, Australia. †Child and Youth Mental Health Service, Nepean Blue Mountains Local Health District, Penrith, NSW, Australia. ‡Department of Linguistics, Macquarie University, North Ryde, NSW, Australia. ¶St. Vincent’s Private Hospital Sydney, Darlinghurst, NSW, Australia. **Institute of Regional Health Research, University of Southern Denmark, Odense, Denmark. Correspondence concerning this article should be addressed to Ben Ong, Susan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, Sydney Nursing School, University of Sydney, 88 Mallett St., Camperdown, NSW 2050, Australia. E-mail: [email protected].
Family Process, Vol. 59, No. 2, 2020 © 2019 Family Process Institute doi: 10.1111/famp.12431
66
objectives, and trajectories, such as in psychiatry and GP consultations (Drew & Heritage, 1992). In psychotherapy (Peräkylä 2012; Peräkylä, Antaki, Vehvilainen, & Leudar, 2008), CA research has explored the different purposes served by formulations, interpretations, and questions by therapists, and how they are rejected or accepted by clients. CA has also demonstrated how clients can show affiliation or shifts in understanding, and how thera- pists show recognition or empathy to client emotions. Through a detailed description and analysis of talk, CA has offered a unique perspective on psychotherapy and holds substan- tial potential for demonstrating how therapeutic concepts and techniques are actually applied, both successfully and unsuccessfully (Georgaca & Avdi, 2009; Madill, Widdi- combe, & Barkham, 2001). This provides a sound basis for novel recommendations on therapists’ micro-level choices in conversation, including word selection, prosody, and intonation, which can support improvements to the implementation of therapeutic interventions. CA research into psychotherapy has tended to focus on individual therapy, with less attention to family therapy interactions. A recent review by Tseliou (2013) has made a sig- nificant contribution to understanding CA and family therapy (FT) research. Tseliou (2013) took a broad approach that reviewed CA and discourse analysis (DA) research in both couple and family therapy. This review focused specifically on methodology, particularly examining types of research questions, data/sampling processes, types of analyses, epistemological perspectives, knowledge claims, and quality criteria. Tseliou (2013) revealed a number of methodological shortcomings in CA/DA research. For example, there were a limited number of original studies, with authors often publishing multiple articles from the same study. There were also many case study designs, with 17/28 studies analyz- ing only one therapy session. Despite this, 13/28 articles made attempts at generalizations. In relation to quality criteria, articles made most frequent use of reflexivity and exemplars (eight instances of both) and 10 articles made no mention of quality criteria. Deviant case analysis and next-turn proof were less frequent, with three and two examples respectively. Tseliou (2013) concluded that, with careful attention to methodological consistency and rigor, CA and DA have the potential to improve our understanding of therapist and institutional influences on clinical practice and to bridge the gap between family therapy research and practice through improving therapist reflexivity. The Tseliou (2013) review combined both CA and DA research and critiqued them against the same methodological standards. While CA and DA both study discourse and language, there are significant substantive and methodological differences between them (Wooffitt, 2005). Substantively, CA focuses on interaction; particularly how social actions are achieved through moment-to-moment sequences of talk, and the normative expecta- tions underpinning them. DA examines language from a broader perspective, utilizing passages of talk, looking at what functions are served by that talk, and how discourses constrain, oppress, and shape talk and interaction. Methodologically, CA has an analytic process that utilizes collections of a phenomenon, sequential analysis, and the incorpora- tion of normative expectations. CA’s analytic claims are grounded in the observable behav- iors of the participants through use of the next-turn proof and deviant case analysis. DA does not necessarily employ these analytic processes. Analytic claims are, instead, rou- tinely evaluated through the presentation of data for reader evaluation (Wooffitt, 2005). DA’s use of passages of talk makes the use of next-turn proof or deviant case analysis less essential. This distinction between CA and DA is complicated by one strand of DA research called discursive psychology (DP). Like CA, DP studies naturally occurring conversations using Jeffersonian transcription conventions and sequential analyses, often using and citing CA techniques and research (Have, 2007). These similarities can at times make the distinction between CA and DP difficult (Potter, 2012). However, DP retains an interest in how discourse influences psychological as well as social structures (Potter,
67
2012), topics that extend beyond the scope of CA (Have, 2007; Madill, 2015). DP (and DA) research is thus likely to make additional analytic claims about social and psychological influences beyond CA’s focus on the observable conversation. Considering these significant substantive and methodological differences, CA research warrants a review against CA-specific methodological standards. As well as including both CA and DA, the review by Tseliou (2013) encompassed both family and couple therapy. While there appears to be similar therapy skills employed in both family and couple therapy, there has been a differentiation between the approaches through the development of specific therapeutic models (Jacobson & Gurman, 2015; Lebow, 2013). This review will proceed on the basis that there is a difference of focus, tech- nique and therefore conversational sequences between couple and family therapy, and will therefore focus exclusively on FT research. While this inevitably reduces the number of articles appropriate for this review, it also provides a more focused and concentrated examination of family therapy practices, and the application of CA techniques to it. In summary, this review will aim to provide a more directed and focused methodological critique of CA research on family therapy than has previously been completed. By high- lighting areas of stronger analytic practice, the outcomes of this critique will provide researchers with recommendations for conducting and evaluating CA research and hope- fully encourage further use of CA in studying family therapy.
METHODS Evaluation Criteria In determining appropriate evaluation criteria for this review, we will employ general criteria for qualitative research and criteria specific to CA. Although a number of different evaluation criteria have been proposed for qualitative research, there does appear to be some convergence of recommended items (e.g., Blaxter, 1996, 2013; Elliott, Fischer, & Rennie, 1999; Russell & Gregory, 2003). These criteria generally cover appropriateness of the qualitative approach for the research question, connection to the literature, selection of data, analytic processes, clear presentation of findings, and position of the author(s). For the current review, we employed the 20-item criteria for the evaluation of qualitative research as reported by Blaxter (1996, 2013). This decision was based on its convergence with other reported criteria, the analytic detail required by the 20 items, and the authors’ previous experience with the tool. The Blaxter criteria contain items covering aims and connection to the literature, methods, analysis, presentation of findings, and ethics (see Table S1). Each item takes the form of a yes/no question, for example “Is reference made to accepted procedures for analysis?” or “Have measures been taken to test the validity of the findings?” The criteria do not provide a definitive “score” or rating of the quality of articles, but are instead designed as a guide for evaluating qualitative research. The Blaxter (1996, 2013) criteria have a number of items for reviewing analytic prac- tices, but it was necessary to supplement them with criteria specific to CA. This is because research from different epistemological perspectives need to be reviewed against quality criteria appropriate to each perspective (Madill, Jordan, & Shirley, 2000; Willig, 2013). Because there is no agreed set of criteria for the evaluation of CA research (Tseliou, 2013), we created a set of criteria through a consideration of CA’s epistemology and recom- mended CA research practices. We acknowledge that Tseliou (2013) produced a set of reported quality criteria, but these items were generated from the studies themselves, and may not necessarily reflect criteria appropriate for CA-specific research. Conversation analysis is considered a constructionist approach (O’Reilly & Kiyimba, 2015), partly due to the primary role of language in constructing social reality (Potter,
68
1996; Tseliou, 2013). Potter (1996) proposed a set of reliability and validity criteria for con- structionist approaches including deviant case analysis, displays of participants’ own understanding through next-turn proof, coherence or integration with past research find- ings, and readers’ evaluation through presentation of adequate data. These criteria have previously been applied to CA research (Madill et al., 2001). A review of introductory writ- ings on CA research processes (Have, 2007; Pomerantz & Fehr, 2011; Schegloff, 1996; Sid- nell, 2013; Sidnell & Stivers, 2013; Wooffitt, 2005) points toward a generally accepted set of procedures common to CA. These procedures include: data in the form of video or audio recordings of naturally occurring interactions, use of Jeffersonian transcription conven- tions (Hepburn & Bolden, 2013), collecting instances of a phenomenon of interest, and an analysis of the phenomenon with reference to sequence organization, grounded in the ori- entations of the participants via next-turn proof and deviant case analysis. There is thus a high degree of convergence between CA processes and the validity and reliability criteria for constructionist approaches, with both including deviant case analy- sis, the orientation of participants, and presentation of adequate data examples. We will therefore apply the following five quality criteria to the present review:
• Does the study utilize naturalistic data in the form of audio or video recordings? • Does the transcription follow Jeffersonian conventions? • Are analytic claims supported by data extracts? • Are analytic claims grounded in the orientations of participants and demonstrated through sequential analysis, next-turn proof, and deviant case analysis? • Are the results discussed in connection to prior CA research?
It is worth noting that Madill (2015) argues that CA makes realist claims about knowable, objective phenomena that through proper analysis emerge from the conversational data. CA thus seems to inhabit a unique methodological position between realist and construc- tionist epistemologies. CA’s epistemology thus requires further exploration and discussion but, due to the current convergence on a constructionist perspective, the above criteria will be maintained for the current review. Literature Search A literature search was conducted by the first author in April 2018 using the databases PsycINFO, PubMed, and CINAHL due to their coverage of clinical practice and family therapy. If a database did not include the controlled search terms “family therapy” and “conversation analysis”, they were searched as free text search terms. This strategy appeared to provide an acceptable level of precision and recall (Harter, 1986). Screening was initially by reading the article title, abstract, and keywords, and if unclear by review- ing the article’s method section. Inclusion criteria were original empirical research, published in peer-reviewed journals, exploring family therapy and specifically stating the use of CA exclusively or as part of a wider analysis. Only articles in English were included and there was no limit on age of articles. As explained above, articles exclusively on couple therapy were excluded. The search in PsycINFO retrieved a total of 30 references, and 10 articles met the inclu- sion criteria; PubMed retrieved nine references, and four articles met criteria; and CINAHL retrieved two articles and one met the criteria. After duplicates were removed, this resulted in a total of 11 articles. Searches of individual journals in the family therapy and linguistics fields using the same search terms added a further nine articles. Using the citation index in Scopus, each article’s reference list, citing articles, and similar articles were reviewed to find other relevant articles meeting the inclusion criteria. This process
69
was repeated for each of the included articles retrieved and added a further five articles. The literature search identified a grand total of 25 articles for inclusion (see Prisma diagram, Figure 1). Evaluation of Identified Articles As stated above, each article was reviewed against Blaxter’s set of criteria that provides guidelines for evaluating qualitative research (see Table S1). Further, these items were supplemented with five items specific to CA. As each article was read, the authors recorded their justified assessment of the article in light of each of the review criterion. To ensure the rigor of the evaluation process, two authors reviewed each paper. The first author reviewed all articles and the other authors were allocated half each. Authors reviewing the same articles then met in pairs to discuss and compare their evaluations against the quality criteria. Similarities and differences were discussed in the pairs until consensus was reached, referring back to the original articles as necessary. The authors’ combined comments were synthesized by the first author and the key conclusions are presented below. Two articles were also initially reviewed by all authors and discussed as a group to ensure consistency in the application of the review criteria. As previously stated, the aim of this review was to critique the application of CA research on family therapy and to therefore highlight areas of stronger and weaker research practice and to identify areas
FIGURE 1. PRISMA statement. Adapted from “Preferred Reporting Items for Systematic Reviews and Meta- Analyses: The PRISMA Statement,” by D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman and The PRISMA Group, Public Library of Science Medicine, 6(7).
70
that require particular attention for future research, rather than to order individual articles along a continuum of strongest to weakest. The literature search retrieved a number of articles reporting the use of CA in combina- tion with other approaches such as DP, DA, or grounded theory. The breadth of approaches within DA and the realist epistemological perspective of grounded theory would necessitate the inclusion of different quality review criteria to that of CA and DP. This review, however, is focused specifically on the application of CA. Each article included in this review has made some specific statement about the use of CA in the analysis and we argue that they therefore make themselves accountable for the application of CA processes, and are open to critique in relation to specific CA research practices. As a result, we have opted to retain the same evaluation criteria for each article, acknowledging that our critique applies only to each article’s application of CA and not to the quality of other aspects of their analyses. Some details about the first author may also help readers to situate this review and con- sider potential biases. The first author is a research student, psychologist, and family therapist interested in understanding the details of dialogical conversations in family therapy. This review was originally conducted to understand how CA has been applied to family therapy interactions and the major findings of such research. A further aim is to understand what constitutes rigorous CA research and how it can be consistently applied in future studies. The authors have no bias against DA, DP, or other qualitative approaches, but we are particularly interested in CA’s focus on observable, naturally occurring data, and analyses grounded in the orientations of participants.
FINDINGS For ease of presentation, the findings will be grouped according to the sub-categories used by Blaxter (1996, 2013) rather than a description of each of the 20 items. CA-specific criteria included for each category are noted in parentheses. The presented sub-categories are: (1) research aims and connection to literature; (2) methods and data collection (use of video or audio data, Jeffersonian transcription); (3) analysis (next-turn proof and deviant case analysis); (4) presentation (use of data extracts and coherence with previous CA find- ings); and (5) ethics. General findings will be presented under each sub-category. Support- ing examples were chosen due to their ability to clearly illustrate the findings. Table S2 summarizes the findings in relation to each article. First, however, we should note that the use of CA to understand family therapy interac- tions is relatively new, with more than half of the identified articles published in the last 10 years. Similar to Tseliou (2013), we found a small number of published researchers, with two authors involved in a combined total of 16 of the 25 articles. Couture (five articles) appears to have published from a single session, while O’Reilly (11 articles) uses one dataset.
Research Aims and Connection to Literature The research questions for all articles were appropriate for CA through a focus on describing interactions. The majority of articles (n = 21) also had a clear statement about aims, focusing on a particular issue concerning family therapy, such as impasse, blame, or the use of reflecting teams. Examples included: “[W]e focus on the discursive resources through which the therapist is able to repair alliance ruptures occurring between the therapist and family members .. .” (Muntigl & Horvath, 2016, p. 104). And, “We investigate how parents seek to build alignment between themselves and the therapist, simultaneously distancing themselves from their child’s behaviour” (Parker & O’Reilly, 2012,
71
p. 459). Four articles had unclear aims, instead describing a general area of investigation. For example, O’Reilly (2006) described the general concept of interruptions and the less-than-full member status of children which inform the study, but these ideas do not culminate into a clear aim for the study. While a clear statement of aims is recommended for qualitative studies (Blaxter, 1996, 2013), a more exploratory approach tends to be employed in CA. Conversation analysts are encouraged to start with an “unmotivated” look at the data by “bracketing” existing assumptions or goals, and approaching the data open to new discoveries (Have, 2007). It was unclear if researchers started with unmotivated looking, but had to adapt their reporting of aims in terms that are generally accepted for research publications. The literature reviews focused on theoretical concepts related to their area of investigation, rather than specific past CA research. Presumably, this was due to a lack of current CA and FT research. The articles therefore provided a good orientation to their area of investigation and gen- erally provided a clear statement of aims. However past CA research in the area of investi- gation did not generally feature in the article introductions.
Methods and Data Collection Data collection was undertaken using availability or convenience sampling, which is not inconsistent with CA. All articles utilized data from naturally occurring FT conversa- tions and utilized either video or audio recordings, as expected for CA. The sole exception was Friedlander, Heatherington, and Marrs (2000), who in addition to video and audio data also utilized a session that had previously been transcribed and published and they therefore had no access to the original recording. Stated analytic approach and number of studies are listed in Table 1. Seven articles reported using CA exclusively, with the rest using CA in combination with another type of analysis. While the articles tended to provide a good conceptual overview of CA, there was little description about how CA was actually conducted. This is not unusual practice for CA articles despite some clear guidelines on conducting CA (e.g., Have, 2007; Pomerantz & Fehr, 2011). Only Hella et al. (2015) made specific reference to CA concepts in the meth- ods. For example, “... we have scrutinized how the turns are allocated, who is taking the turn after whom, what gaps (or pauses) or overlaps there are in turn-transitions, to whom talk is addressed, which lexical choices are made, especially as regards person pronouns, and whether there are aborted utterances, pauses, or other signs of hesitation” (Hella et al., 2015, pp. 27–28). All articles utilized Jeffersonian transcription procedures, with the exception of Fried- lander et al. (2000), who did not present any data samples. Eight articles did not provide a high level of detail, missing features such as prosody and intonation. The transcription practices were therefore not utilized to their potential, with prosodic elements largely absent from the analysis. TABLE 1 Stated methods in the included studies.
Stated method Number of studies CA 7 CA + DA 5 DP + CA 8 CA + other qualitative approach 2 CA + other (unclear) 3
72
In summary, consistent with CA, the articles overwhelmingly made use of video or audio recordings of naturalistic therapy sessions. They also made use of accepted Jeffersonian transcription conventions, however there was a wide variation in the level of prosodic and intonational detail in these transcriptions.
Analysis The articles in this review employed a range of analytic approaches (see Table 1), thus making direct comparison and evaluation difficult. This review will instead focus specifi- cally on the application of CA and not on an article’s analysis as a whole. Overall the anal- yses were presented systematically with appropriate supporting examples. But, guided by Blaxter’s criteria, this review revealed four main issues. These issues centered around: (1) sequential analysis; (2) validity testing and deviant case analysis; (3) additional analytic claims; and (4) statements beyond presented data. Sequential analysis The sequential organization of conversation is a unique and central idea in CA (Stivers, 2013). Utterances are not analyzed in isolation, but in relation to what has come before and what follows. A related idea is the next-turn proof, where the proposed actions of a speaker are interpreted by reference to how they are responded to by the next speaker. The next-turn proof serves to ground an analysis in the orientations of the participants. Sequences of interaction were a major focus with the majority of articles (n = 20) describing behaviors with some reference to sequential context. The extent to which the sequential context was incorporated into the analysis varied greatly, however. Articles with particular attention to sequential environments included ways that a family member and client spoke in relation to symptoms (Hella et al., 2015), the position and responses to children’s interruptions (O’Reilly, 2006), and how families responded to vulnerability (Pote, Mazon, Clegg, & King, 2011). In these studies, the sequential context was essential to the analyses. For example, Pote et al. (2011, p. 111) describe how a successful topic shift is achieved: “After the father’s initial short bid for a topic change in L5 [the original data extract not presented here], there is a brief pause; as nobody challenged his bid, the father held onto the turn and continued to introduce his own agenda. This is a successful bid for topic switch because everybody else shifts onto the content of the father’s turn and abandons the previous topic .. .”. This analysis does not just look at the father’s turn (i.e., the initial bid for a topic switch) in isolation but also incorporates the responses of the other participants (i.e., a pause and shift to new content). The analysis thus involves the sequential position of the utterance and uses the following turns as supporting evidence for the analysis. Another example is from O’Reilly (2006, p. 560), “Lee makes several attempts to draw attention from the therapist and take the conversational floor. He interrupts by addressing the therapist by name on two occasions [the original data extract not presented here]. ... The family members treat Lee’s potentially interruptive turns in a negative way with Steve telling him to ‘Shut up’ ... and his father telling him to ‘shut up’... Lee persists in his attempt and Lee’s turns are treated as interruptive as Mr Niles acknowledges that the therapist’s turn is incomplete: ‘He’ll talk to you in a minute when he’s finished’”. Lee’s talk is analyzed as interruptive because of the responses of the other participants, not by ana- lyzing Lee’s talk in isolation. Again, the utterance is not analyzed in isolation, but is given meaning through the responses by the other participants. Although most articles made reference to sequences, the level of detail of sequential analysis varied. Rather than employing a detailed analysis of sequences, many articles
73
(n = 13) focused on identifying, describing, and grouping behaviors into categories. For example, Friedlander et al. (2000) coded and categorized various ways that therapists responded to blame in FT sessions. They provide detailed descriptions of these categories but focus solely on the responses of the therapist without reference to the preceding or fol- lowing talk. The therapists’ talk is therefore analyzed in isolation from its sequential con- text. Another example comes from O’Reilly and Parker (2014) who describe how talk is constructed as inappropriate for children. In this example the authors are presenting evi- dence to support the analysis that talk can be cast as inappropriate due to temporal factors: For example, “Dad: = they do it
74
Additional analytic claims One application of CA to FT interactions is scrutinizing and critiquing accepted institu- tional practices (Tseliou, 2013). A series of articles achieve this through highlighting how children are responded to regarding interruptions (Hutchby & O’Reilly, 2010; O’Reilly, 2006, 2008). In these studies, the analyses indicate that children can be ignored or inter- rupted without comment but when therapists interrupt adults there is some recognition or apology. These findings draw attention to interactions that may conflict with FT ideals of giving equal voice to all participants. But these studies also add additional analytic claims; namely that parents answer for children because they are responding to the moral implications of the therapist’s questions (Hutchby & O’Reilly, 2010). These interactional patterns are interpreted with reference to an overarching concept of moral implications that, while plausible, are difficult to demonstrate within CA’s focus on structures and sequences of actions. Other similar interpretations concern protection and vulnerability talk (Pote et al., 2011), membership status of children (O’Reilly, 2006, 2008), and blame and responsibility (Parker & O’Reilly, 2012). A strict application of CA would avoid additional analytic claims beyond the observable orientations in the conversational data. However, as discussed previously, DA and DP approaches maintain an interest in the intersection between conversational practices and psychological and social processes. Therefore, interpretations through psychological and social mechanisms were consistent with such approaches. Issues around the combination of CA with other analytic approaches will be considered further in the discussion. Statements beyond presented data Tseliou (2013) noted that, despite a postpositivist epistemology and single case studies, many articles made generalizations (13/28) or realist claims (17/28). A similar finding arose from our review. In 13 articles, we found the presence of claims that were outside what the data could provide. These included statements about frequency using general descriptors without supporting data, interpretations and conclusions that were not supported by the presented findings, and overgeneralized recommendations for practice without sufficient sequential context to evidence their effectiveness. For example, O’Reilly (2007, p. 238) states that “I have demonstrated two central points: that (1) parents can and do work to construct the child as naughty; and (2) there are various ideas about the cause of naughtiness”. In the presented examples, parents describe behaviors but do not directly construct the child as “naughty” (this is done by siblings and the therapist). As for the second point, the presented data do not demonstrate a variety of ideas about naughtiness, as only one example of a proposed cause is presented (i.e., a “nervous problem”). The stated conclusions are thus not supported by a strict reading of the presented data. Other overstatements involved recommendations that were extrapolated from, but not directly consistent with, the presented data. For example, Parker and O’Reilly (2013, p. 504) state that their data “illustrates that validation ... has potential to circumvent disen- gagement or facilitate re-engagement”. They presented three supporting examples; two are of only the therapist speaking (which did not demonstrate how these validation attempts were received), and in the other example the child showed no signs of re-engagement. On the basis of the presented data it remained difficult to recommend validation as a tool for engagement due to the limited demonstrated evidence of its effectiveness. The remaining 12 articles made conclusions that were consistent with the scope of the presented data. For example when describing the use of vocalized noises, O’Reilly (2005a, p. 760) concluded that “Children use an active noise that is in some way derived from or representative of the conversation that is taking place. This serves a simple short way into the interaction. When the adults use active noising, it serves to graphically illustrate a
75
narrative, analogy, or argument, strengthening and authenticating a sense of experien- tially grounded accurate recall and reporting”. The conclusions were a close fit with the presented data and did not include interpretations beyond the data. About half the articles in this review contained some form of statements beyond a strict interpretation of the presented data. This varied from overstatements about frequency of behaviors to recommendations for practice without directly supporting evidence.
Presentation of Data This section includes reader evaluation, coherence, or integration with previous research, and reflexivity or statement of the author’s position. The quality criterion of reader evaluation involves the presentation of data extracts for verification of findings. The articles, overall, provided good supporting evidence for their analytic claims with appropriate data extracts. The two exceptions were Friedlander et al. (2000), who presented no data, and Williams and Auburn (2015), in which the majority of their data was presented in a separate document in supporting information. The criterion of coherence involves the integration between a study’s results and past research and provides a means of checking the validity of claims (Potter, 1996). This crite- rion was evaluated by reviewing each article’s discussion section for the author connecting their current findings to past research. Evidence of integration with existing research is difficult to achieve considering the small number of current CA and FT studies. The authors connected their studies to past research in two main ways; firstly, through connection to existing CA research (n = 8), including more general CA findings on interaction or specific CA and FT research. For example, Parker and O’Reilly (2013, p. 177) connected their research on therapists exiting a session to consult with a reflecting team to previous CA research: “Linguistically and nonverbally effective terminal sequences can be established by attention to closing down the topic and using closed rather that open utterances (Schegloff & Sacks, 1973). White, Levinson, and Roter (1994) found that in the vast major- ity of visits to the doctor, it was the doctor who initiated the closing of the session with patients generally displaying agreement with that closure. This also seems to be the case of temporary closing within sessions, for example, nonverbally the therapist may stand at an ERP [exit relevant place] and indicate leaving the therapy, or verbally could ensure that sentences are completed with final intonation and no invitation for further contribution.” In this example, the authors are connecting their results, i.e., that therapists can temporarily close a session by verbal and nonverbal means, to earlier CA research on closing topics via terminal sequences, and evidence that in doctor consultations, it is the doctor who tends to initiate the closing of the consultation. Secondly, authors connected their findings to similar topic areas from outside CA. This generally occurred through reference to psychotherapy concepts or other discursive research. For example, in his discussion, Stamp (1991) connected his interpretation of the family’s talk to Minuchin’s assertion that a family is a system that regulates and shapes the responses of its members. Stamp also connects the analysis to Liang’s ideas about social selves and the formation of identity through stories. The remaining nine articles presented their results independently without a discussion of their connection to previous research. The role of reflexivity in CA research is unclear. It is a recommended feature of qualita- tive research but is not typically discussed in relation to CA processes. In nine articles the authors made mention of either their professional background or theoretical position in relation to family therapy but included no discussion of the potential impacts on their research findings. Two studies (O’Reilly, 2005b; O’Reilly & Lester, 2016) included a brief
76
discussion on reflexivity but these took a confirmatory rather than critical perspective on how authors’ perspectives may have impacted on the analysis.
Ethics The articles report a range of responses to ethics. Only two articles make explicit men- tion of receiving ethical approval. Ten articles reported following ethical codes or princi- ples, seven reported obtaining consent or anonymizing data. Five articles made no statement about ethics.
DISCUSSION This review aimed to provide a detailed methodological critique of CA and FT research and to highlight potentially problematic research practices in order to guide the evaluation of CA articles and to support sound research practices in future studies. The articles under review were critiqued against a guideline for evaluating qualitative research and five CA-specific quality criteria. The articles under review had a number of strengths. The aims and research questions were appropriate for the use of CA and the articles presented their arguments clearly and systematically with appropriate supporting extracts from the data. Appropriately for CA, the articles made use of naturalistic video or audio data and directed attention to the anal- yses of sequences as well as utilizing Jeffersonian transcription conventions. However, this review also raised a number of concerns, including varying levels of detail in the tran- scriptions, differing attention to the analysis of sequences, and a lack of validity testing. Some articles also included additional analytic claims such as reference to internal states or moral order, or focused on the categorization of behaviors without further analysis of how these behaviors achieved certain actions. Conversation analysis’s focus on interaction and avoiding analysis or accounts of inter- nal processes can be both a strength and criticism of the approach (Madill, 2015). CA’s agnosticism toward the psychological reality of certain phenomena and its position outside FT theory may allow CA researchers to make observations not previously countenanced by accepted FT theory or research. Conversely, CA may not provide the type of analysis of interest to researchers and clinicians concerned with conversational themes and internal processes. This may partially explain the number of studies combining CA with other ana- lytic approaches that can encompass these broader issues of psychological and social pro- cesses. CA thus risks being subsumed by DP in areas such as psychotherapy and FT where internal processes can be of central importance. CA researchers could, and perhaps need to, utilize current psychotherapy research to direct investigations into areas of interest to therapists in order to make it mutually beneficial to themselves and psychotherapists (Madill, 2015; Stiles, 2008). A return to CA’s roots in ethnomethodology and the principle of unique adequacy or acquired immersion (Have, 2007) may help align CA with the interests of family therapists, and highlight the importance of therapist involvement in research. The tendency in the articles to group behaviors into different categories may reflect assumptions about social actions. Enfield and Sidnell (2017) argue that by categorizing an action the categorizer makes relevant and accountable a particular interpretation. The analytic ascription of action may also serve a purpose for researchers. For example, describing the less-than-full status of children, identifying family members using protec- tion talk, or the therapist constructing a relational explanation may reflect the theoretical positions of the authors. Categorizing of behaviors or the ascription of actions are not nec- essarily problematic, but due to their potential implications their use is best acknowledged
77
and discussed. This speaks to a broader point that while CA may be seen to make realist claims (Madill, 2015), an analysis is still an interpretive construction and therefore neces- sitates some form of reflexivity by the authors (Tseliou, 2013). A solution suggested by Enfield and Sidnell (2017) is that rather than beginning with the notion of definitive actions, analysts should instead consider the details and tokens of talk that are attended and responded to in their sequential context. For FT research this may involve attention to which parts of an utterance therapists respond to, how participants show their orienta- tion and understanding of previous turns and how sequences develop. The application of quality criteria is also an important outcome of this review. While most studies used appropriate types of data and transcription methods, there was a lack of validity testing, concurring with the review by Tseliou (2013). This is a striking omission considering the range of literature on criteria for good qualitative research. While there are a number of recommendations for conducting CA (e.g., Have, 2007; Pomerantz & Fehr, 2011; Sidnell & Stivers, 2013), they have, as yet, not been incorporated into a single, widely accepted set of criteria, thus making reviews of CA studies potentially less systematic. Ambiguities about CA’s ontological and epistemological positions are also likely a factor. As discussed, there is some disagreement about whether CA has a realist or constructionist perspective (as well as whether these classifications are relevant for CA at all), potentially creating some confusion about appropriate validity criteria. As a discipline, CA remains quiet regarding these issues, remaining resistant to polar categorization (Madill, 2015; Stiles, 2008). Future CA and FT research could benefit from paying closer attention to adhering and reporting quality criteria to justify and support its findings. This review highlighted that in FT research, CA was often combined with other analytic approaches, with only seven out of the 25 articles included in this review stating the use of CA exclusively. This use of multiple analytic approaches combined with a lack of details about how the analyses were conducted creates confusion around the analytic process. For example, this review noted the presence of additional analytic claims that were outside the analytic capabilities of CA, but still consistent with a DA or DP approach. Without a clear description of the analytic process, and the respective scopes of their claims, it becomes unclear if authors are applying CA inappropriately or utilizing other analytic approaches. It is therefore incumbent upon researchers to be more specific about the ana- lytic processes they are applying rather than stating the use of a general approach that is open to interpretation. Similarly, when using a combination of analyses, a clearer descrip- tion of how they were variously applied and combined seems necessary in order to both assure the reader of the application of analytic rigor and to avoid the perception of a hap- hazard application of qualitative methods. This review also points to the complexities of using CA with FT. FT involves multiple participants with different perspectives in an institutional setting with institutional tasks, in a heightened emotional environment, in lengthy conversation, with the possibility of incipient blame, accountability, suspicion, and defensiveness. The conversation analyst is faced with a difficult task of describing and analyzing these long and complex interactions. The studies in this review highlight the potential applications of this approach but also the difficulties associated with developing a detailed and rigorous analysis. CA also makes use of prosody and intonation because these are important resources for interaction utilized by participants. As noted above, there has not been much attention placed on these elements of talk in FT, making them a still unexplored area of interaction. Our interest in conducting this review has been to set out how CA has been applied to FT and also to see what may constitute rigorous CA research in the FT context. We find CA’s focus on observable behavior, using only the talk that is available to the interactants (without needing to posit internal states), and the importance of grounding an analysis in
78
the orientations of the participants, a refreshing and respectful way of researching inter- action. Our experience of this review has been mixed. We see the potential of CA to inform FT practice in new ways but also clearly recognize the lack of a consistent application of CA’s methods. This review may therefore be affected by our interest in maintaining partic- ular standards for CA.
Recommendations Based on the findings of this review we propose the following recommendations for future CA research into FT as a guide for both researchers and readers of this research. Rather than a definitive list we hope to elicit further discussion and refinement in order to maintain high standards in CA and FT research:
• Include previous CA research in the introduction of the study to better situate it in con- versation analytic traditions. • Be clear on what analytic approaches are being utilized and provide details on how they are conducted. • If multiple analytic approaches are being utilized, provide details on how they interact (or not) and contribute to the research findings. • Provide appropriate data extracts to support the analysis with appropriate level of tran- scription detail for the analysis undertaken. • Support the analysis through reference to surrounding talk via the next-turn proof, and work to account for possible alternative interpretations. • Utilize validity checks and deviant case analysis. • Avoid collecting categories of talk without further analysis of how that talk achieves its proposed action/s. • Confine the analysis and conclusions to within observable data and avoid interpretation though social or psychological motivations. • Discuss coherence and disagreements of results with previous CA research. • Include some reflective discussion of the author’s position and possible impacts on the analysis. • Include information on ethical approval or procedures.
Limitations The search methods for this review only focused on published peer-reviewed papers in academic journals, thus excluding grey literature, dissertations, and book chapters. Arti- cles were only included if they made specific reference to using CA, either alone or as part of a larger analysis. This review could therefore be critiqued for only including a subset of the work on CA and family therapy. Two independent authors were involved in a majority of the articles (n = 16) including collaborations with other authors. This review could therefore be critiqued as a review of the work of two particular authors rather than the research approach as a whole. It is still important to highlight the restricted number of authors as it reflects the developing state of CA and FT research and the need for greater diversity in the field. Despite this limitation, the current review still highlights the potential pitfalls when using CA that future researchers could avoid. CONCLUSION Conversation analysis promises to make a unique contribution to the understanding and practice of FT. This includes describing therapy as it actually happens, thus
79
potentially providing clinicians with greater insights into applying therapeutic principles in real-life situations, alerting clinicians to micro aspects of conversations that may improve or impede the therapeutic project, and critiquing accepted practices and understandings about family therapy, thereby promoting therapist self-reflection. CA thus has the potential to provide clinicians with specific practice recommendations on the micro-level of conversation including word selection, intonation, and prosody. The application of CA to FT is a relatively new endeavor, with a limited number of authors and research articles. This body of work has explored a number of issues of importance to family therapy but CA’s focus on observable behaviors may not fit with clinical interest in psychological processes. The application of CA to FT is also a complicated task considering the complexities of multiparty interactions in FT, and CA’s standards for analytic claims. The findings of our review suggest that future research has a solid base on which to build, but should proceed with careful attention to rigorous analytic processes, and a critical, reflective awareness on the part of the researchers if CA is to fulfill its potential usefulness for FT.
SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Table S1. Blaxter evaluation criteria. Table S2. Summary of articles and review criteria
80
Chapter Summary
The work of conversation analysts has revealed how talk is ordered and structured at a minute level of detail. Participants in interaction display an orientation to these normative aspects of conversation and utilise various conversational practices to achieve various social actions. These practices are observably communicated to recipients and are therefore also available to the conversation analyst. Like ethnomethodology, CA endeavours to explicate how people participate in social interactions through analysing the interactions themselves rather than through reference to overarching theories of interaction or psychological motivations. CA thus seeks to understand interaction through describing how it is produced and understood by the participants utilising it. CA can contribute a unique perspective to understanding Open Dialogue conversations that is not influenced by Open Dialogue theory and is supported by an extensive research base. CA research into family therapy processes needs to pay particular attention to rigorous analytic procedures and validity criteria as well as to limiting research claims to what can be accounted for by the data and the analysis.
81
Chapter 3: Literature Review
In this chapter, I complement the previous methodological critique with a narrative review containing a summary of research findings. In this article, published in the Journal of
Family Therapy (Ong, Barnes, & Buus, 2020b), I attempt to outline the main topics of CA research into family therapy. This review provides an overview of CA research findings that inform and validate my own research as points of coherence and comparison (Potter, 1996).
82
Journal of Family Therapy (2020) 42: 169–203 doi: 10.1111/1467-6427.12269
Conversation Analysis and Family Therapy: A Narrative Review
a,b,f Ben Ong, Scott Barnesc and Niels Buusd,e
Conversation analysis (CA) can provide insight into interactional micro-processes of family therapy. Past reviews of CA and family therapy have focused on methodology without descriptions of research findings, reducing the likelihood that the findings of CA research are employed to guide practice. This narrative review provides therapists with a description of CA findings that can inform family therapy practice. Systematic searches of databases, individual journals, and citation analyses were completed in April 2018 resulting in twenty-five articles for review. The findings of this narrative review were summarised around three areas: family members’ talk about each other, therapists’ responses to the family, and the use of reflecting teams. The use of reflecting teams and various conversational devices may help engage multiple participants. Future CA research into family therapy could focus on longer sequences, the overall structural organisation of sessions and the interactions of multiple therapists.
a Susan Wakil School of Nursing and Midwifery, University of Sydney b Nepean Blue Mountains Local Health District, Child and Youth Mental Health Service, Penrith, NSW c Department of Linguistics, Macquarie University d Susan Wakil School of Nursing and Midwifery, University of Sydney; St Vincent’s Hospital, Sydney e St Vincent’s Private Hospital Sydney and Institute of Regional Health Research, University of Southern Denmark, Odense fAddress for correspondence: Sydney Nursing School, 88 Mallett St. Camperdown, NSW, Australia, 2050. Email: [email protected]
83
170 Ong, Barnes and Buus
Practitioner points • CA can offer practitioners insight into the interactional micro- processes occurring in family therapy that can inform future practice • Micro-interactions between therapists and family members can indicate their perspectives on sensitive topics, blame and accountability • Blaming talk creates a dilemma for therapists trying to maintain both neutrality and engagement • Family therapists use conversational practices to respond to difficult conversations while maintaining engagement and collaboration • Reflecting teams may help to acknowledge and engage multiple family members
Keywords: conversation analysis; family therapy; narrative review.
Introduction Talk is one way that people negotiate and coordinate their interactions. Conversation analysis (CA) is the study of this talk-in-interaction. It de- scribes and analyses how spoken interaction is structured and how se- quences of talk coordinate actions. CA has a number of features that make it distinct from other approaches that study language and social interaction (Sidnell and Stivers, 2013). Conversation analysts study talk occurring in naturalistic contexts, utilising data from audio and video recordings of real-life interactions. CA then employs a unique method of data transcription to capture the features of conversation, such as pauses, prosody and intonation (Hepburn and Bolden, 2013). The analysis focuses on describing and explicating normative structures, and how particular conversational practices achieve social actions. CA’s standard of proof requires that analytic claims are grounded in the ori- entations of the participants (Wooffitt, 2005). This means that analytic claims are based on the observable details of the talk and how partic- ipants display their orientations or understandings of the talk, rather than by reference to internal states, motivations, or analyst-driven invo- cation of social structures and discourses. CA has a particular interest in conversational structures and actions. The best understood conversational structures are normative expecta- tions. Language has a set of normative expectations that participants
84
Conversation analysis and family therapy 171 orient to, and manage at a local level, in order to maintain conversa- tional order. These structures include a system for turn-taking (including speaker selection and transition to next speaker), repair organisation (how misunderstandings are clarified), and sequence organisation (how paired parts of talk, such as question-answer, greeting-greeting, re- quest-compliance, are developed). Another area of interest is the analysis of actions, i.e. the goal-oriented reasons embodied via people’s observ- able conduct (see Enfield and Sidnell, 2017; Levinson, 2013). Talk is a vehicle for action, and particular conversational practices for speaking (e.g. word and grammar choice) can be utilised in order to achieve par- ticular goals. CA describes how certain practices achieve certain actions through analysing sequences of talk with reference to how participants respond and orient to them. It should be stressed that, for CA, under- standing talk requires attention to the sequential context and how partic- ipants manage their interaction. An analysis therefore requires attention to the local, sequential context in which it occurs and as it occurs for the participants (Maynard and Clayman, 1991), and not by reference to abstract invariant structures (Button and Sharrock, 2016; Lynch, 2009). The research objectives of CA have been described in two broad ways (Have, 2007). One, generally called ‘pure’ CA, focuses on describing the features of everyday conversation as described above. Another has been termed ‘applied’ CA. Applied CA can take a variety of forms, but is gener- ally focused on analysing interactions with a view to specific professional and discipline-specific activities. This is most notably demonstrated in studies of what is termed ‘institutional talk’ such as occurs in the court- room or in news interviews, as well as in health and psychiatric settings (Drew and Heritage, 1992). In psychotherapy, CA has been used to pro- vide greater detail on how psychotherapy techniques are delivered and the interactional functions they serve, and to describe ways that clients respond to interventions. A selection of examples is presented below. CA has examined the various uses of ‘formulations’ by therapists (Antaki, 2008). These formulations are presented as summaries of what a client has said and invite confirmation or disconfirmation, but they also do other interactional work. Formulations also select and exclude various parts of the client’s talk, thereby providing a way for the thera- pists to shape clients’ talk and influence the direction of a session. CA can also describe and identify subtle techniques that are utilised by the therapist that may not be described by current psychotherapy theories. For example, Rae (2008) shows how therapists can use a form of repair called lexical substitution. In these cases, therapists would
85
172 Ong, Barnes and Buus respecify the client’s talk by offering an alternative word. For example, (Rae, 2008, p. 63): Client: It feels a little uncomfortable Therapist: Or a lot uncomfortable Here the therapist respecifies the client’s ‘a little uncomfortable’ to an upgraded ‘a lot uncomfortable’. Rae (2008) describes how this practice can display that the therapist is listening closely and to the client, but also serves to encourage the client to be more explicit about their emotional experience without the use of a direct statement to that effect. CA has also analysed how clients can show disalignment with ther- apist questions or formulations. For example, clients may resist the presuppositions of optimistic questions through downgrading the pos- itivity, refocusing to non-optimistic interpretations, redirecting credit to others, or through joking responses (MacMartin, 2008). CA can therefore reveal greater detail about how clients may reject (or accept) therapeutic interventions and promote therapist self-reflection on their practice (Georgaca and Avdi, 2009; Madill, Widdicombe and Barkham, 2001; Tseliou, 2013). Despite the insights CA can provide on family therapy practice, there does not currently exist a summary or synthesis of the research to date. This means that the results of CA research are generally not readily available for practising clinicians to consider and relate to their professional activities. Recent reviews in the field (Ong, Barnes and Buus, 2019; Tseliou, 2013) have focused on methodological critiques of existing CA and family therapy research without a description or summary of the re- search findings. This narrative review will therefore attempt the fol- lowing aims: (1) to systematically search for articles on CA and family therapy; (2) to provide a narrative review including a summary and syn- thesis of the current findings in the field of CA and family therapy; and (3) to consider how these findings may inform family therapy practice for clinicians.
Methods This study was designed as a narrative review of CA and family therapy. While there is not currently an agreed upon structure for a narrative review (Ferrari, 2015), there does appear to be a general consensus on what a good narrative review should contain (Derish and Annesley, 2011; Ferrari, 2015; Green, Johnson and Adams, 2006). These include
86
Conversation analysis and family therapy 173 having a specific research question for the review, clear inclusion and exclusion criteria, a comprehensive literature search, critical evaluation of the methods and quality of results, a synthesis of the key findings and drawing of conclusions, impact on field, and discussion of the direc- tion that the field needs to take. This review will focus specifically on summarising and discussing the key findings of the articles under re- view and compliment recent critical methodological reviews (Ong et al., 2019; Tseliou, 2013). A systematic search for articles was completed in April 2018. The process included three stages: (1) searches of the databases PsycINFO, PubMed and CINAHL due to their coverage of clinical practice; (2) searches of individual journals in family therapy and linguistics; and (3) reviews of reference lists, citing references, and ‘similar’ references of each article found using the Scopus citation database. We used the free text search terms ‘family therapy’ and ‘conversation analysis’. Inclusion criteria were original empirical research articles published in peer-reviewed journals, exploring family therapy and stating the use of CA exclusively or as part of a wider analysis. This often involved dis- cursive psychology articles that incorporate CA in their analyses. CA and discursive psychology have a number of similarities, with both using naturally occurring conversations, similar transcription conven- tions, and the use of sequence analysis, sometimes making it difficult to distinguish between the two approaches (Have, 2007; Potter, 2012). But unlike CA, discursive psychology may go beyond an analysis based only on conversational structures to explore the influence of broader discourses on psychological and social structures (Madill, 2015; Potter, 2012). All types of family therapy approaches were included, but articles researching couple therapy were excluded in order to focus more spe- cifically on family therapy processes. There was no limit on the age of articles, but only articles in English were included. The search retrieved a total of twenty-five articles (see Figure 1 for PRISMA statement). The articles were reviewed against the twenty-item criteria for quali- tative research (Blaxter, 1996, 2013) as well as criteria relevant for CA. The findings of this critical review of methodology can be found in Ong et al. (2019). To complement the critical review, the current paper sum- marises the research findings and takes the form of a narrative review. All articles were evaluated by at least two reviewers: all articles were re- viewed by the first author, with the two other authors each reviewing half of them. The key reported findings of each article, as reported in its abstract, results and discussion sections, were noted and discussed in
87
174 Ong, Barnes and Buus
Figure 1. PRISMA statement. Adapted from Moher et al. (2009).
pairs by the respective reviewers. These findings were then compared across articles by the first author. Similar articles were grouped together according to their findings and areas of investigation, as well as areas of agreement, disagreement and complementarity to build a summary of current knowledge of CA and family therapy. Consistent with the presentation of a narrative review, only the main findings of the review are presented here (Ferrari, 2015). Other areas of investigation, such as the argument for the usefulness of CA research for examining family therapy are included in the summary table (Table 1) but do not form part of this review. The discussion aims to synthesise the presented findings.
Findings Table 1 lists the articles included in this review with a summary of aims, stated method, participants and setting, and main findings. A review
88
TABLE 1 Summary of reviewed articles and main findings
Stated sample (num- Summarised Stated ber of sessions, partici- Reference aims methods pants, context) Main findings Couture How therapist Critical DA/ Single session Examined ‘forward moving conversations’ (2006) and family CA after impasse. move beyond Therapist, young per- Therapist used process of discovering differ- impasses son, parents ent positions of family members and invita- tion to accept a middle ground. Therapist does this by ‘tentative yet strategic’ (p. 294) statements that are small shifts towards
the middle, stopping clients from making Convers extreme statements, offers extreme formu- lations that are tentative and contestable by family.
ation Post hospital Family responded with agreement or disa- admission greement, offering information, qualified
reformulations, partial uptakes, and listen- analysis ing responses (nods, acknowledgement tokens).
Couture How partici- DA/CA As above How multiple parties are addressed in family (2007) pants address therapy. and
multiple family parties
therapy
Ong, Barnes and Buus TABLE 1 Continued
Stated sample (num- Summarised Stated ber of sessions, partici- Reference aims methods pants, context) Main findings Packaging talk/recipient design: ‘perturbed speech pattern’ (p. 70) to display caution, invite shared curiosity.
Show useful- Collectively soliciting: looking down and ness of DA back and forth at all recipients, veiled ad- dress, addressing others as overhearers. Fishing: ambivalent responses to elicit re- sponses by others. Performative advice-giving: therapist models how to interact through ‘responsive in- volvement’ (p. 74). Couture and How families As above As above Similar findings to articles above. Strong and thera- How therapist introduces topic, how family (2004) pists move negotiate two-way small steps towards a beyond middle ground. entrenched positions Argues that therapists can become more effective by attention to micro-details of How DA can conversation. enhance un- derstanding of therapy processes
TABLE 1 Continued
Summarised Stated Stated sample (num- Reference aims methods ber of sessions, partici- Main findings pants, context) Couture and How the family As above As above Process of advice giving by therapist: (1) Sutherland therapist strategic invitation to a possible middle (2006) gave advice ground, (2) uptake by listener, (3) advice to a family given. Process can be repeated until there is uptake. Therapist uses downgrading practices such Conversation as hesitation, questions, sounding tentative and less directive to balance his authority with family’s autonomy. Friedlander How expert CA/ Data from seven narra- Identified three main categories in respond- et al. construc- Grounded tive/constructionist ing to blame: ignoring/diverting, acknowl-
(2000) tionist and theory therapists edging/challenging, reframing. analysis narrative Sessions sent in by No uniformity in clients’ responses to therapists therapists or from therapist. respond to
commercially and blame available videos or
transcription from family a book
therapy
TABLE 1 Continued
Ong, Barnes and Buus
Summarised Stated Stated sample (num- ber of sessions, partici- Reference aims methods pants, context) Main findings
Hella et al. How par- CA One session Divides session into phases: Relative- (2015) ticipation Psychiatrist, client and prominent Information Phase, Participant- of a relative family member, ne- prominent Information Phase, Evaluation effects the gotiating treatment and Decision Phase. expression When doctor and relative talk, client denies of paranoid symptoms. When client was interviewed he experiences is explicit about symptoms. Psychiatrist uses ‘recycling’ and ‘suspended return’ from earlier in conversation. Relative provides important information but at the risk of moral judgements and creat- ing conflict. Hutchby and Explore CA Multiple sessions Identified examples where parents re- O’Reilly relationship sponded when therapist directs a question (2010) between to the child. Parents could answer before therapist child has a chance to answer or after a questions, pause, or parents asked an alternative speaker question. selection, and family members’ responses
TABLE 1 Continued
Stated sample (num- Summarised Stated ber of sessions, partici- Reference aims methods pants, context) Main findings Explore Two therapists, three relationship families, twenty-two between hours video turn-taking and person deixis
Family therapy clinic Proposed two reasons: parents treat children Conversation as less-than-full-members, parents are responding to moral implications of thera- pist’s questions. Parents speak to therapists about bad be- haviours of child and not themselves. This
can create an accusation-denial sequence analysis between parent and child with parents shifting between addressing therapist and child.
and
family
therapy
TABLE 1 Continued
Ong, Barnes and Buus
Summarised Stated Stated sample (num- ber of sessions, partici-
Reference aims methods pants, context) Main findings Muntigl and How a thera- CA First five minutes Session of a ‘master Horvath pist repairs of demonstration therapist’. Therapist repairs alliance at beginning of (2016) alliance rup- session session by claiming to be nervous like the ture and how client, involves child (i.e. asks her a ques- actions relate tion) and uses humour. to therapy
principles Two therapists, Therapist disagrees with mother’s negative mother, daughter assessment of herself, offers a more positive interpretation or her actions. Early childhood inter- vention and educa- tion centre O’Reilly Investigate the Discursive Multiple sessions Noises used differently by children and (2005a) function of psychology adults. noises for Two therapists, four Children: orienting to topic, to attract atten- children and tion and enter the conversation e.g. animal adults families, twenty-two hours video noises, repetition. Only data from Adults: to authenticate and strengthen a three families used point, provides sound imagery and empha- (reason not stated), sises severity of child’s actions. twenty hours video Family therapy service
TABLE 1 Continued
Summarised Stated Stated sample (num- ber of sessions, partici- Reference aims methods pants, context) Main findings
O’Reilly How families Discursive Multiple sessions Complaints to therapist about a third-party (2005b) present com- psychology service. plaints about Common pattern of complaints: capsule professional Two therapists (also stated as one thera- (summary of the problem, gloss of com- third parties plaint matter), then an expansion of and how pist), three families, twenty-two hours details. Capsule gloss worked up by some- therapists thing as negative, moral fault is formu- respond video Conversation lated, agency is assigned. Family therapy clinic Therapist responses: orientating to the unhelpfulness of complaints (redirect- ing to emotions or agency which is more amenable to therapy), orientating to the
expectation of practical assistance (focus- analysis ing on what will be helpful to discuss in current session), stating what the therapist cannot do.
O’Reilly How children Discursive Multiple sessions Three types: children try to interrupt during and (2006) interrupt and psychol- a delicate issue but are ignored and fall
Two therapists, four how adults ogy/ families, twenty-two silent, children try to interrupt during family respond CA hours video delicate issues and are ignored but persist until a negative acknowledgement is given,
Family therapy clinic children interrupt in a topic relevant way therapy and are attended to.
TABLE 1 Continued
Ong, Barnes and Buus Summarised Stated Stated sample (num- ber of sessions, partici- Reference aims methods pants, context) Main findings
O’Reilly How family Discursive As above Term ‘naughty’ has various levels of mean- (2007) and therapist psychology ing: euphemistic with minimal assessment, construct general term with various meanings, non- descriptions technical and can be used by all, victimless of ‘naughty’ in basic form but can be upgraded with child emphasisers, complex with a variety of meanings depending on the person.
O’Reilly To explore Discursive As above When therapists interrupt adults they orient (2008) ways that psychology to the interruption and usually apologise. therapists in- When therapists interrupt children they terrupt their don’t apologise. clients O’Reilly and How parents Discursive As above Identifies ways that parents say they are good Lester display their psychology Child and adoles- parents: state that they are good parents, (2016) ‘good parent- cent mental health acting in the child’s best interests (with ing’ and service self-sacrifice), coping with the child’s be- how blame is haviour in appropriate ways (and contrast- managed ing with inappropriate ways), appeals to science (biological explanations).
TABLE 1 Continued
Summarised Stated Stated sample (num- ber of sessions, partici-
Reference aims methods pants, context) Main findings O’Reilly and How children’s CA As above How children show disengagement. Parker behaviour White British Passive disengagement (inattention), pas- (2013) indicates dis/ Midlands, lower SES sive resistance (don’t attend to question engagement or attempts to engage), active resistance How therapists (directly refuse to answer or comply). manage Expressing autonomy and evading adult alliance impositions (requesting to cease session Conversation ruptures and not wanting to participate in future sessions). Therapists use acknowledgement and valida- tion to create or reinstate engagement. O’Reilly and How families Discursive As above People use informational (what is said), loca-
Parker talk about psychology tional (where it is said and who is around), analysis (2014) what is ap- temporal (when), personal contextual (who propriate for said it and their role) factors to justify inap- children to propriateness of a behaviour.
hear But some contradictions as parents re-state and inappropriate talk in front of the children,
say derogatory things about children. family Therapist can propose that child’s presence
may be problematic that allows parents to therapy confirm thus maintaining engagement.
TABLE 1 Continued
Ong, Barnes and Buus
Summarised Stated Stated sample (num- ber of sessions, partici-
Reference aims methods pants, context) Main findings Parker and Explore social Discursive As above. Parents align with therapists by: tellings O’Reilly positioning psychology Family therapy centre (general descriptions, specific descrip- (2012) between tions – to build authenticity, derogatory parents and descriptions), dispositional descriptions, children active voicing (to demonstrate parents as reasonable and child’s responses as unrea-
sonable), and evidencing (to substantiate claims). Children can deny accounts about them. Therapist acknowledges that children are talked about and attempts to engage with them directly, also statements that thera- pist believes the parents.
TABLE 1 Continued
Summarised Stated Stated sample (num- ber of sessions, partici-
Reference aims methods pants, context) Main findings Parker and How therapists CA As above Successful exit from session to reflecting
O’Reilly negotiate UK systemic family team involves: preannouncement, minimiz- (2013) exit and re- therapy clinic ing the amount of time taken, identifying/ entry to the creating exit-relevant places (ERPs), and session. How accounting for the departure.
information Unsuccessful attempts to exit: attempting to from reflect- leave at non-ERP and no terminal se- ing team is quence (pre-closing announcement, candi- Conversation imparted date resolution, offered to return to subject and received on return), re-opening conversation.
Successful re-entry when therapist takes first turn, provides feedback from team and positive response from family, non-success-
analysis ful when family takes first turn on a differ- ent topic, therapist doesn’t refer to team. All family members used protection strate- Pote et al. Develop un- CA/ Four video-taped
derstanding sessions gies of each other through topic switch, re- and (2011) Thematic versals of valence (negative to positive talk of vulner- analysis
Three families, one ability and or vice versa), intensification of positive or family protection in member with intel- negative talk, decentering (shift focus away intellectual lectual disability from one person), continued engagement
disability National Health with topic (by therapist to actively address therapy Service intellectual issues of protection). disability family therapy team
TABLE 1 Continued
Ong, Barnes and Buus
Summarised Stated Stated sample (num- ber of sessions, partici-
Reference aims methods pants, context) Main findings Stamp (1991) Demonstrate Microanalysis Single session. Turn-by-turn account of a segment of the value of conversa- 1:08 minute sequence interaction. Analysed with reference to of micro tion interruptions, speaker selection, power analysis for relationships.
family inter- CA Argue for usefulness of microanalytic ap- actions, what Therapist, mother, proach for therapy and theory. family means son, and daughter to each participant, to examine communica- tion patterns in the family
TABLE 1 Continued
Summarised Stated Stated sample (num- Reference aims methods ber of sessions, partici- Main findings pants, context) Stancombe How neutrality Range of ‘Large corpus’ – no Devices therapists use to maintain neutral- and White is produced methods other details ity both in session and in private discus- (2005) and resisted including Child and family sions. Change subject to save face or CA assessment and blame but runs risk of not hearing blamer therapy service in who then repeats blaming statements. north of England Re-formulations to package blame and propose therapist non-blaming versions of events, incorporating elements of opposing Conversation sides. Two paradoxes that the performance of neutrality reinforces blame and therapists’ impartial versions of family members are worked up though a series of practical-
moral judgements. analysis Suoninen How evolving CA/DA/ Six sessions Proposes notion of interactional positions, and interactional Social con- Six therapists provides extracts from sessions and labels
Wahlstrom positions struction- the interactional positions and identities and (2009) frame client ism present.
identities One family, parents, Multiple identities of father were collabora- family Describe son, daughter tively constructed and re-constructed. identity con-
Finnish language with therapy structions of English translation fatherhood
TABLE 1 Continued
Ong, Barnes and Buus
Summarised Stated Stated sample (num- ber of sessions, partici-
Reference aims methods pants, context) Main findings Sutherland How a thera- CA Single session Therapist invites collaboration by tentative and pist used his Therapist, young per- formulations using pauses, restarts, inhala- Couture expertise son, parents tions, selective listening, attention to weak (2007) with family’s agreements, and inviting family shifts in preferences meanings.
Show useful- Post hospital Father starts using similar techniques to ness of CA admission therapist. in investigat- ing therapy interactions How the thera- pist collabo- rated with the family, and how the family mem- bers collabo- rated with each other’s meanings
TABLE 1 Continued
Stated sample (num- Summarised Stated Reference aims methods ber of sessions, partici- Main findings pants, context)
Viaro and To describe CA Ten sessions Describes rights and responsibilities of par- Leonardi ‘rules of ticipants in family therapy sessions. (1983) the game’ Five therapists, Therapist has the exclusive right to: decide in family students and what topic is discussed, who will speak at therapy experienced any moment, interrupt a turn, stop a con- versation, put questions, sum up and make organisational glosses. Conversation
Family therapy centre Family members have duty not to interrupt
in Milan and right not to be interrupted (except by therapist), can propose themes for discus-
sion and act as therapist’s interlocutor (but therapist can accept or reject proposals),
have equal rights between each other. analysis
Rules are not explicitly stated but family members conform to the rules through
reinforcement from therapist. Therapist
can accept or censure rule violations. and
How talk is transformed in main session and
How accessible CA Five families, one ses- family Williams and reflecting team. Auburn poly-vocality sion from each (2015) is deployed
and what therapy functions it performs
TABLE 1 Continued
Ong, Barnes and Buus Stated sample (num- Summarised Stated ber of sessions, partici- Reference aims methods pants, context) Main findings In the main session: the deletion of negative and introspective positions, the selection of aspects that place the individual in relation to others, and transformation to establish the basis of positive connotation.
Three therapists Reflecting session: negative descriptions (one male and two contextualised positively, adult descrip- females) tions of difficulties were omitted, and children’s viewpoints emphasised, relation- ship descriptions privileged over individual explanations, develop and extend hopeful content. Can provide voice to unspoken parts of cli- ent, or parts client is not willing to express Systemic family to family members directly. therapy service
Conversation analysis and family therapy 191 of each study’s findings suggested three main areas of investigation: (1) family members’ talk about each other, particularly in relation to blame and accountability, interruptions, and protection; (2) therapists’ management of the therapeutic relationship, in relation to responding to blame and maintaining neutrality, negotiating impasse and advice giving, and maintaining engagement and inviting collaboration; and (3) the use of reflecting teams.
Family members’ talk about each other Research in this area focused on clients and family members and how they spoke about each other and interacted with the therapist. The articles focused on the particular areas of blame and accountability, interruptions, and protection talk.
Blame and accountability. Research on blame and accountability focused on talk about the identified client, which was overwhelmingly by parents about their children. Parents utilised a range of conversational resources to present themselves as good parents and to avoid any blame or accountability that may be connected to the behaviour of their child. This has been reported as being achieved in two main ways. First, parents can present behavioural examples and dispositional descriptions to emphasise and authenticate the severity of their child’s behaviour (O’Reilly, 2005a; Parker and O’Reilly, 2012). Second, parents can make direct statements about their own good parenting and appropriate responses to their children’s behaviour (O’Reilly and Lester, 2016), and family members can emphasise their reliability as informants (Hella et al., 2015). Such talk serves to locate blame in the child and represent the parents as reasonable actors and therefore not responsible for their child’s behaviours. Attention is thus directed away from the relational and interactional patterns that are the focus of family therapy, to problems within a particular individual.
Managing interruptions. CA research on interruptions in family therapy suggests that interruptions by adults receive different responses to interruptions by children. Adults seem to be able to interrupt or speak for children without sanction, while children’s attempts to interrupt may be ignored or negatively responded to (Hutchby and O’Reilly, 2010; O’Reilly, 2005b). Furthermore, if adults interrupted other adults, there is some recognition of the interruption or even an apology but this did not occur when interrupting children (O’Reilly, 2008). In these
105
192 Ong, Barnes and Buus sessions, there seems to be differential rights to the conversational floor with greater access granted to adults rather than children.
Protection talk. Pote, Mazon, Clegg and King (2011) describe protection talk where family members and therapists protect others’ vulnerability by switching topics, changing the valence of the talk, or shifting focus away from one person. However, therapists can also retain focus on a particular topic in order to further address issues of protection. This research suggests that diverting conversation away from topics may signal avoidance, but it can also serve protective functions for a third party. In summary, this research shows how family members’ interactions can indicate something about the perspectives of the participants. For example, topic switches can indicate sensitive areas that family mem- bers may avoid as a way of protecting each other, talk locating blame in an individual directs the conversation away from relational explana- tions, and children can have lesser access to contribute to conversations. These interactions have implications for therapy, as the therapist makes a decision, for example, to accept the topic switch or retain focus on a sensitive topic, accept an individual explanation for problems or pro- pose an interactional view of a problem, or focus on adults or actively solicit talk from children. Therapists may therefore need to be sensitive to how clients are responding to the conversation so that they can make decisions about the most therapeutically useful direction to take in the interaction. This creates a dilemma for therapists about how to appro- priately respond while still maintaining engagement and collaboration.
Therapists’ responses to the family Much of the CA research on family therapy has focused on the be- haviours of the therapists in relation to particular therapeutic issues. Generally, this seemed to reflect the tension between directing and guiding clients while still maintaining engagement within the thera- peutic relationship. Therapist responses are discussed below in rela- tion to responding to blame and maintaining neutrality, negotiating impasse and advice giving, and maintaining engagement and inviting collaboration.
Therapist responses to blame and maintaining neutrality. Therapists took three main positions in response to blame: alignment, by acknowledging or stating a belief about the blame (Friedlander, Heatherington and Marrs, 2000; Parker and O’Reilly, 2012); disalignment by ignoring and
106
Conversation analysis and family therapy 193 diverting the conversation away from the blame (Friedlander et al., 2000; Pote et al., 2011; Stancombe and White, 2005); or through neutrality by reframing the blaming talk in ways that are consistent with the frame of family therapy. Therapists can reframe talk by focusing on emotions, framing the problem in the context of the relationship, redirecting to talk about resolutions, or using metaphors to expand the theme of the discussion (Friedlander et al., 2000; O’Reilly, 2005b). The responses of therapists may therefore indicate their position regarding the blame, even if not stated explicitly. These studies show ways that therapists work to maintain engage- ment while still avoiding blaming or inappropriate talk. For example, rather than sanctioning parents’ inappropriate talk directly, therapists can propose that the child’s presence may be problematic, allowing parents to confirm and make arrangements to discuss particular topics without children being present (O’Reilly and Parker, 2014). This po- tentially allows the therapist to maintain engagement with the parent by not explicitly sanctioning their inappropriate talk, while also reducing the negative effects on the child. But maintaining neutrality is a complicated interactional task. For example, when a therapist attempts to maintain neutrality by topic switches, this can result in a client re- peating the blaming talk or adding additional evidence, possibly as a result of not feeling heard or acknowledged (Parker and O’Reilly, 2012; Stancombe and White, 2005). Thus, through actions designed to re- duce blaming talk, the therapist may actually increase it. Negotiating disalignment and inviting collaboration. Client displays of disalignment can signal to therapists impending disengagement, which can in turn elicit attempts to re-engage and invite collaboration. Client disengagement could be displayed as direct refusals to participate, requests to terminate sessions, and inattention or not attending to questions (Muntigl and Horvath, 2016; O’Reilly and Parker, 2013). There are also more subtle forms, such as pauses before responding, partial uptakes and qualified responses (Couture, 2006; Muntigl and Horvath, 2016). Therapists’ responses could involve explicitly orienting to the dis- engagement through acknowledgement of client difficulties and vali- dation (Muntigl and Horvath, 2016; O’Reilly and Parker, 2013). Other therapist responses involved techniques designed to further the conver- sation, such as forced-choice questions, selective listening and focusing on weak agreements (Sutherland and Couture, 2007). These responses do not directly acknowledge disengagement but instead select relevant parts of the client’s talk that align with the therapist’s therapeutic
107
194 Ong, Barnes and Buus project and can continue the therapeutic conversation. Alternatively, a therapist can use follow-up responses, such as forced-choice questions, that serve to solicit a response more strongly from an ambivalent client. A series of articles describe a particular model of how a therapist and family move through impasses and the steps involved in therapist advice-giving (Couture, 2006; Couture and Strong, 2004; Couture and Sutherland, 2006). Following an impasse, the therapist elicits and vali- dates the different positions of family members, followed by an invita- tion to accept a middle ground. Advice-giving provided two additional steps. First, the therapist provides small invitations to a middle ground while monitoring the level of uptake by family members. Second, the therapist provides advice, which is accepted when there is a sufficient level of prior uptake by the family. These studies describe how therapy is a dynamic, two-way process. The therapist does not simply provide advice or interventions but monitors and adapts to how the family are responding. Similarly, family members are not passive recipients, but instead display their alignment or disalignment to therapist moves that in turn shape how the interaction develops. When negotiating impasse, giving advice, or inviting collaboration more generally, the therapist can utilise a number of particular resources. These resources could take the form of enacting hesitation or tentative- ness. These included: re-starts; pauses and drawn out words; word choice, such as ‘I imagine’ or ‘probably’; and voicing opinions with rising into- nation, serving to elicit a response (Couture, 2007). These techniques presented the therapist’s suggestions as tentative and contestable, po- tentially inviting a response from the family. Therapists could also invite collaboration from the family in other ways (Couture, 2007; Sutherland and Couture, 2007). This occurred through addressing multiple parties simultaneously through questions and indirect gaze that were directed at the group rather than a single addressee. Talk could also be addressed to others indirectly through veiled address, or by ambivalent or uncer- tain talk with pauses, inhalations, and drawn out vowels to ‘fish’ or invite listeners to respond and jointly contribute or elaborate on a topic. In summary, this research shows how therapists attempt to maintain engagement with multiple family members simultaneously. This can be- come a difficult task when family members have different perspectives, as agreeing with one perspective may isolate other family members, and remaining neutral may increase the presence of blaming talk by participants not being acknowledged. Therapists’ talk can also be re- sponsive to the family, particularly around impasse and disengagement, when therapists can use a variety of techniques to gently introduce new
108
Conversation analysis and family therapy 195 ideas or continue the conversation without directly acknowledging the disengagement.
Use of reflecting teams The two studies on the use of reflecting teams looked at two different areas: the transition points in and out of a reflecting team, and how the talk in the main session was discussed in the reflecting team part of the session. Successful transition in and out of the reflecting team discussion in- volved setting an agenda and active management of the session struc- ture by the therapist (Parker and O’Reilly, 2013). Successful exits from the main session involved the therapist making pre-announcements of upcoming exit and not opening new topics of conversation. Return to the session is more smoothly negotiated by the therapist taking the first turn and providing feedback from the reflecting team. If the family takes the first turn, the conversation may take a different course and not return to the feedback from the reflecting team. So in order to maintain the structure of a reflecting team session, and a smooth transition between sections, the therapist needs to be proactive in setting up, marking and directing transition points within the session. Williams and Auburn (2015) examine how talk that occurred in the main part of the session is incorporated into the talk of a reflecting team. This use of the reflecting team differs from that described by Parker and O’Reilly (2013) because the reflecting team conversation is observed by the client and family. In the reflecting conversation, negative descriptions from the family’s discussion were contextualised positively. For example, in the session a father was described as strict and over-protective, while the reflecting team described him as a fa- ther that loves his daughter (Williams and Auburn, 2015, pp. 543–545). Reflecting team discussions also tended to omit adult descriptions of difficulties and instead emphasised the viewpoints of the children. This is in contrast to other studies where children seemed to be treated as lesser than adults (O’Reilly, 2008). The reflecting team is also used to authoritatively endorse positive interpretations or resolutions, particu- larly emphasising hopeful talk and relational descriptions of problems. In summary, the use of reflecting teams involves a particular session structure requiring an exit from the main part of the session with the family, to a consultation or observation from the reflecting team, and then transition back to the family. In order to maintain this structure, the therapist needs to take an active role by using pre-announcements,
109
196 Ong, Barnes and Buus closing conversations on exit, and actively setting the topic on return from the reflecting team. The reflecting team can also provide more positive, hopeful and relational descriptions of the family’s difficulties and promote the viewpoints of children, thus potentially maintaining engagement with all family members.
Discussion The aim of this narrative review was to identify, summarise and synthe- sise the main findings of research that applies CA to the study of family therapy, and to discuss how this research can inform family therapy practice. The discussion to follow will consider in more detail the main themes raised by this research and particularly the complexities of si- multaneously attending to multiple relationships. It will also address how these issues contribute to understanding family therapy, and pos- sible future applications of CA research in family therapy practice. This review highlights the complexities involved in family therapy, with therapists having to monitor and attend to the engagement of multiple relationships simultaneously. When there is blaming talk, the therapist is placed in a dilemma, where acknowledging the blame risks disengagement with the person being blamed, while redirecting blam- ing talk potentially risks disengagement with the blamer. The articles on blame also highlight issues around the perceived rationale for family therapy. When parents present problems as located within the child this undermines the reason for family therapy, which focuses on relational contributors to problems. Therapists then have to potentially argue for the usefulness of their approach rather than progressing with the work of therapy. Some of these studies describe ways that therapists negotiate engagement issues, through direct acknowledgment and validation, or attending to parts of the client’s talk that further the progression of the conversation (Muntigl and Horvath, 2016; O’Reilly and Parker, 2013; Sutherland and Couture, 2007). However, it is not clear to what extent clients accept these attempts at engagement, or if they are successful in encouraging acceptance of family therapy principles for parents who present individually oriented problem formulations. The work by Williams and Auburn (2015) suggests that the use of a reflecting team may be a way of maintaining engagement with multi- ple family members and acknowledging quieter voices. Williams and Auburn (2015) showed that a reflecting team tended to focus on the perspectives of children over adults, thus bringing forth potentially
110
Conversation analysis and family therapy 197 overlooked voices. The reflecting team also positively reframed nega- tive descriptions, thus potentially aligning with the subjects of negative descriptions. Considering the contributions of multiple participants si- multaneously may be a difficult task for solo clinicians. As observers, the reflecting team may help to ensure that consideration is given to all participants, including those with lesser influence, thus helping to ameliorate any potential alliance breaches. Therefore, therapy with multiple clients may be aided by multiple therapists with multiple per- spectives. But care must also be taken with the use of reflecting teams. First, clinicians need to properly explain and prepare families when using reflecting teams to minimise any difficulties in transition points within the session. Second, while reflecting teams can present more positive interpretations and give a voice to quieter participants, they can also make authoritative endorsements or interpretations that may not fit with the experience of the family. Reflecting teams may therefore need to be cautious about the delivery and presentation of their ideas, and therapists may also need to give sufficient time for discussion with the family about reflecting team feedback. By describing how other clinicians respond to particular situations in therapy, CA can contribute to promoting therapist self-reflection (Tseliou, 2013). One example of this comes from the studies on the role of children during interruptions. A number of articles describe how children can be ignored or spoken for by adults (Hutchby and O’Reilly, 2010; O’Reilly, 2006; Parker and O’Reilly, 2012). This (conceptually at least) conflicts with family therapy approaches that seek the input of all family members and recommend avoiding coalitions or alliances with individuals or sub-groups within a family (e.g. Kerr and Bowen, 1988; Selvini, Boscolo, Cecchin and Prata, 1980). These CA studies show how this can be a difficult task when interruptions occur, and the therapist has to attempt to maintain some fairness and equality between those present. By being aware of these phenomena, clinicians can reflect on their own responses in similar situations and consider if their own practice aligns with their theoretical approach, preferences, and values. Of the articles in this review, only seven reported the use of CA exclu- sively. The majority used CA in combination with other approaches such as discourse analysis or discursive psychology. While there are areas of convergence between these approaches, their different methodological practices and standards can generate analytic claims that are different from those generated via CA, such as accounts for findings that appeal to psychological or social explanations. A discussion of these method- ological issues are outside the scope of this article but can be found in
111
198 Ong, Barnes and Buus Ong et al. (2019). However, a few points are worth mentioning here. In articles reporting a combination of analytic approaches there was not a clear description of how they were respectively applied nor how the findings were collated. It thus becomes unclear if CA is being utilised aberrantly or if authors are drawing on interpretations offered by other approaches. It is therefore important that when using various forms of discursive analyses authors are explicit about how they are each applied and combined in order to maintain transparency and analytic rigour. This review attempts to focus on the findings from a CA perspective, particularly in reference to sequential analysis, and limiting references to analysis though psychological states. We do not wish to undermine the importance of other discursive approaches but to instead describe what can uniquely be offered by CA. A recent special issue of the Journal of Marital and Family Therapy focused on discursive methods in couple and family therapy research (Tseliou and Borcsa, 2018). The issue includes four articles analysing the same couple session, with each utilising a different discursive approach including CA, discursive psychology, discourse analysis and semantic analysis. This issue provides more detail about the relative application, possible findings, and practical value yielded by these different discursive approaches.
Limitations This review is limited to published peer-reviewed research articles and may therefore have omitted other important research not published in this format. CA cannot provide a ‘how to’ of family therapy. CA research can identify interactional problems and ways that they are negotiated, but it is then up to therapists to critically reflect on how these findings are applicable in their own practice and theoretical frameworks. This narrative review presents a synthesis of the main findings, and therefore excludes the analytic subtleties included in the individual re- search papers. This review thus represents some selectivity on the part of the authors to include the main directions and results of this area of research that inevitably excludes certain details.
Future directions Much of the research on CA and family therapy has tended to focus on the behaviours of either clients or clinicians separately, with less
112
Conversation analysis and family therapy 199 focus on sequences of interactions (Ong et al., 2019). CA has the ability to describe sequences of interaction and document the trajectories of different conversational practices, thus highlighting both smooth and problematic ways of navigating interactional projects (Stokoe, 2014). Future research may therefore consider longer sequences of interac- tion with attention to therapeutic or conversational difficulties, and their successful or unsuccessful negotiation. This could provide a more detailed understanding of the progression of therapy, and also provide some further guidance for therapists. Much of the current research focused on sessions with only one therapist, with only a few studies including sessions with multiple ther- apists working as co-facilitators or in reflecting teams. The presence of multiple therapists is likely to have unexplored effects on, for exam- ple, how therapists manage and structure the session between them. These structural differences have potential unexplored implications for how a therapy session progresses and therefore warrant further investigation. CA in other contexts has also been concerned with describing the overall structural organisation of interactions. For example, clear struc- tures have been identified in certain healthcare settings (Robinson, 2013), but this has proved to be more difficult in psychotherapy ses- sions, possibly because psychotherapy and family therapy do not nec- essarily follow a repeated, transparently specific structure. However, there appear to be interactional tasks and projects that are completed in family therapy and this may be a way of describing structural organi- sation. One example of this comes from Couture (2006), who proposes a description of how a therapist and family worked through impasse. It is possible that family therapy involves a series of such tasks, but further research is necessary to describe what these tasks and structures may be, and how they are successfully and unsuccessfully negotiated.
Summary CA research examining interactions in family therapy is a new but growing area of investigation. Family therapy interactions involve mul- tiple participants with differing goals and perspectives, providing a complicated and rich area for CA research. So far, CA researchers have described how families speak about each other, particularly in relation to blame and accountability, how therapists manage the therapeutic relationship when working through blame, impasse, and giving advice, and how reflecting teams have been utilised. These findings can make
113
200 Ong, Barnes and Buus therapists more aware of these processes, and may help therapists re- flect on their own work, and provide some guidance on how interac- tional problems may be negotiated. Further research could broaden the range of family therapy processes under analysis, providing clinicians with greater detail into the micro organisation of their interactions with families.
114
Update on literature reviews
Since these literature reviews were conducted there have been a couple of CA research studies into Open Dialogue. These studies have originated from a research group in Denmark, including my supervisor Niels Buus. I have also been involved to different degrees in the writing and analysis of these articles. Below, I provide a brief description of these studies as they have informed my own research.
Schriver, Buus, and Rossen (2019) investigated therapist reflecting conversations in
Open Dialogue sessions as examples of “my side” tellings (Pomerantz, 1980). A my side telling involves a speaker stating their limited, occasioned knowledge, that may in turn elicit further details from the recipient who is in a more knowledgeable position. The article identifies two types of my side tellings. In reporting my side tellings the therapist claims to restate what a client as actually said or reported to have said. For example, “Maja also says at some point to you get scared... Sara describes that she has had suicidal thoughts” (Schriver,
Buus, & Rossen, 2019, p. 24). In inferring my side tellings, therapists deliver their own formulation or interpretation and are usually marked by the words “I think”. For example, when talking about the fear previously expressed by the family the therapist says, “I also think that Lisbet and Brian carry this with them in each their own way... and react in each their own way” (p. 27), or a therapist describes the client’s situation as “I just think that it’s vulnerable”
(p. 28). These two types of my side tellings could occur together with reporting my side tellings used to provide evidence for and to support the inferring tellings.
This article highlights the importance placed on epistemic authority in Open Dialogue conversations. Clients have authoritative access to their own thoughts and feelings as they form part of their own experiences and are only publicly available in a mitigated form.
However, psychotherapy involves therapists voicing their professionally informed perception
115
of the clients’ experiences. Therapists have a delicate task when they voice their own thoughts while also respecting the epistemic authority of the client. My side tellings represent a way that therapists can manage the relative epistemic access and rights of themselves and the clients. Therapists can recognise the perspective of the family through reporting my side tellings, while also producing and marking their own position through inferring my side tellings. They are also a way that therapists can produce multiple voices in their reflections, and this can be seen as a manifestation of the dialogical concept of polyphony. However, the structure of the reflecting conversation is such that it occurs between the therapists with the family only watching and not participating. Only after the reflection is completed is the family invited to respond. These my side tellings report the perspective of the therapist but they do not invite an immediate response from the family.
Another recent article by the same research team investigates a particular conversational practice produced by therapists in the form of “Y what do you think about what X just said?” and its related variants (Rossen et al., 2020). This practice occurs in sequences that follow a typical form where there is some talk by a network member, then the therapist selects and repeats a particular piece of that talk. The therapist then selects a different network member to comment on that talk through “Y what do you think about what
X just said?” The next speaker then produces a stance position that was either affiliative, ambiguous, or disaffiliative with the initial stance position. Through this practice the therapist selects topics and elicits multiple stances, while also maintaining a degree of neutrality by not directly producing their own stance position. The authors consider how the therapists’ exertion of control over topic and speaker selection is not antithetical to dialogue but rather may allow more voices to be heard and increase the depth of the discussion.
116
Chapter Summary
CA research into family therapy has covered a range of topics including the use of reflecting teams, blame, accountability, impasse, advice-giving, and how therapists respond to these issues. Attention to conversational processes over more abstract descriptions of therapy can provide unique insights not described in theoretical models. More recent studies on Open
Dialogue demonstrate ways that dialogical principles can be enacted in therapy sessions as well as the complicated work of therapists that has not been previously described in theoretical writings on Open Dialogue. These conversations reveal how therapists are attending to issues of epistemic authority and neutrality in their work, eliciting the stance positions of other participants, and exercising deontic authority in guiding a session.
Thesis Aims
Following the perspective of ethnomethodology and CA, the position that I take in this thesis is that human communication necessarily involves some form of expression including either verbal or embodied communication. Even though Open Dialogue at times emphasises the therapists’ embodied experiencing of the interaction with clients, there still remains something communicated by clients and in turn something communicated by the therapist.
Furthermore, because these interactions are all that is available to therapists and clients, there is a practical importance for therapists to better understand how these conversations are structured. The internal experience of the therapist and client may be relevant and important in Open Dialogue theorising, but these experiences are only available through what is observably communicated between them, and this communication forms the focus of this thesis. This thesis therefore has the following aims:
• To explore, using conversation analysis, some of the conversational practices present
in Open Dialogue sessions.
117
• To describe different ways that dialogical principles may be represented in interaction.
• To provide more detailed accounts of the work that therapists do, and to reveal
conversational practices and ways of interacting that may not have been described in
existing writings on Open Dialogue.
• To examine and contribute to the Open Dialogue professional stocks of interactional
knowledge.
118
Chapter 4: Deontic Authority in Reflection Initiations
This chapter focusses on the deontic authority of the therapist to direct the course of a therapeutic conversation. As described in the introduction, there has been a gradual movement in family therapy from the idea of the therapist as an expert on diagnosing and intervening in the problems of a family, towards greater collaboration and joint decision-making throughout the treatment process. This is a perspective embodied by the Open Dialogue approach (Olson,
Seikkula, & Ziedonis, 2014). In the following study, published in Family Process (Ong,
Barnes, & Buus, 2020c), I focus on a specific conversational environment where therapists make proposals to transition to a reflecting conversation. In these situations, therapists are proposing a change in the structure of the session and are exercising their authority to direct the course of the session, however they attempt to do this in a collaborative way.
119
Downgrading Deontic Authority in Open Dialogue Reflection Proposals: A Conversation Analysis
BEN ONG*,†
SCOTT BARNES‡
NIELS BUUS*,§,¶,**
The Open Dialogue approach promotes collaboration with clients and families in decisions about the direction of therapy. This creates potential problems for Open Dialogue therapists who seek collaboration but also have responsibility for managing the session. Using conversation analysis, we examined 14 hours of video recordings of Open Dialogue sessions, and specifically how therapists proposed the transition to a reflecting conversation. We found that, when making proposals to reflect, therapists routinely downgrade their deontic authority (i.e., adopt a less powerful, more collaborative position). They did this through framing proposals as interrogatives, providing accounts, and by prefacing their proposals with “I’m wondering”. More heavily downgraded proposals made acceptance less salient, potentially risking transition to the reflection. These findings provide more detail on how theoretical concepts such as “collaboration” and “power” are actually displayed and negotiated in practice and can contribute to a more nuanced understanding of what constitutes Open Dialogue.
Keywords: Open dialogue; Power; Authority; Reflection; Conversation analysis; Deontics; Proposals
The Open Dialogue approach to working with families emerged during a period of transition in family therapy. This transition, influenced by postmodernist and feminist critiques, resulted in an increased awareness of the power attached to a therapist’s institutional position, and a greater emphasis on promoting collaboration and joint decision-making (Flaskas, 2010, 2011; Hare-Mustin, 1994). Open Dialogue originated in Finland with people experiencing psychosis and has since been adapted to numerous other countries and settings (Brown & Mikes-Liu, 2015; Buus et al., 2019; Gordon, Gidugu, Rogers, DeRonck, & Ziedonis, 2016; Tribe, Freeman, Livingstone, Stott, & Pilling, 2019). The approach is characterized by a number of principles (Seikkula & Arnkil, 2006), which
*Faculty of Medicine and Health, Susan Wakil School of Nursing and Midwifery, University of Sydney, Sydney, NSW, Australia. †Nepean Blue Mountains Local Health District, Child and Youth Mental Health Service, Sydney, NSW, Australia. ‡Department of Linguistics, Faculty of Medicine, Health and Human Sciences, Macquarie University, Sydney, NSW, Australia. §St. Vincent’s Hospital Sydney, Sydney, NSW, Australia. ¶St. Vincent’s Private Hospital Sydney, Sydney, NSW, Australia. **Institute of Regional Health Research, University of Southern Denmark, Odense, Denmark. Correspondence concerning this article should be addressed to Ben Ong, Sydney Nursing School, 88 Mallett Street, Camperdown, NSW 2050, Australia. E-mail: [email protected]. Family Process, Vol. x, No. x, 2020 © 2020 Family Process Institute doi: 10.1111/famp.12586
120
include providing immediate help by seeing the client within 24 hours of referral, including the client’s social network in treatment meetings (such as family, friends, work or school colleagues, and other involved professionals), and continuity of the treating team by maintaining the same therapists for the duration of the client’s contact with the service. Other principles emphasize collaboration between therapists and the family, the creation of dialogue, and the use of reflections (Olson, Seikkula, & Ziedonis, 2014). In Open Dialogue, the reflection is a conversation between the therapists about the thoughts and feelings that arise in them during the session and is observed by the family who are present in the same room (Schriver, Buus, & Rossen, 2019). Furthermore, the usual turn-taking structure of the session is altered because the reflection occurs only between the therapists. The family members are encouraged to listen until after the reflection, when they are invited to comment on the therapists’ reflections. In order to maintain the structure of the reflection, with the therapists speaking and the family listening, the therapists need to communicate to the family that they are transitioning to a reflection so that the reflection can progress as planned. The initiation of a reflection is potentially problematic because the therapist is directing the session and this conflicts with a philosophical preference for a collaborative orientation and following the lead of the client (Seikkula, 2011; Seikkula & Arnkil, 2006). There is currently little empirical research on how collaboration in Open Dialogue conversations actually occurs. The research methodology of conversation analysis (CA) provides a way of exploring conversations empirically. CA studies talk as it actually occurs in real-life interactions and describes the conversational practices that people utilize to achieve social actions (ten Have, 2007; Sidnell & Stivers, 2013). This involves detailed analysis of verbal and nonverbal communication with particular attention to how turns are designed (including word choice, intonation, prosody) and how conversations incrementally develop. Conversation necessarily proceeds as a series of turns at talk, meaning that someone’s turn at talk occurs in response to previous talk. To understand the meaning and function of a turn, it is necessary to analyze that turn in relation to the sequence of turns surrounding it. CA has been extensively used to study everyday talk, as well as “institutional” forms of conversation including psychotherapy (Georgaca & Avdi, 2009; Peräkylä, 2012, 2019; Peräkylä, Antaki, Vehvilainen, & Leudar, 2008). This research has described therapeutic actions as they occur in real-world sessions with a particular focus on sequences and formulations (Antaki, 2008; Peräkylä, 2019). CA’s granular focus on therapist and client conduct holds substantial potential for promoting therapist self-reflection (Tseliou, 2013) and expanding our understanding of practice beyond that described in psychotherapy models and theories (Peräkylä & Vehviläinen, 2003). The use of CA to describe family therapy interactions is a less developed, but growing area of investigation. Recent reviews demonstrate how CA has been utilized to describe how families and therapists discuss blame, accountability, impasse, advice giving, and how talk occurs in reflecting teams (Ong, Barnes, & Buus, 2019, 2020; Tseliou, 2013). For example, parents can emphasize their own good parenting and make dispositional descriptions of their children’s behavior to avoid blame themselves (Hella et al., 2015; O’Reilly, 2005a; O’Reilly & Lester, 2016; Parker & O’Reilly, 2012). Other research has focussed on how therapists work to address disengagement through acknowledgment and validation (Muntigl & Horvath, 2016; O’Reilly & Parker, 2013). Therapists also respond by selecting
121
and focussing on parts of the clients’ talk that aligned with the therapist’s intended therapeutic direction (Couture & Sutherland, 2006). This allowed the therapist to progress the therapeutic conversation without directly addressing the disagreement. Other research has examined the relative rights of adults and children to speak and participate in the session (Hutchby & O’Reilly, 2010; O’Reilly, 2005b, 2008). For example, children could be interrupted by adults without sanction, while children’s interruptions were more often ignored. However, when therapists interrupted parents they apologized or explicitly oriented to their interruption. CA research on family therapy has thus described how therapists and families negotiate different rights and responsibilities in the therapeutic relationship. In CA, the authority and responsibility to decide and determine courses of action within a particular domain are referred to as deontic authority (Stevanovic & Peräkylä, 2012; Stevanovic & Svennevig, 2015). A distinction is made between a person’s deontic status and stance (Stevanovic & Peräkylä, 2014). Deontic status refers to a person’s level of authority relative to other participants in a certain domain of action. This can be accounted for via their personal history, expertize, or position in institutional structures. For example, the institutional position of being a “doctor” inherits a higher level of deontic status regarding diagnosis than being a “patient”. Deontic stance refers to the level of authority that a person displays in interaction, which is accomplished via their moment-by-moment interactional conduct. Deontic stance varies along a gradient (Stevanovic, 2013b). A person may present themselves as having a heightened deontic stance, indicating entitlement to direct others’ actions, or a downgraded deontic stance, indicating limited entitlement to direct others’ actions. This may or may not accord with the deontic authority imbued by their deontic status. In relation to family therapy, therapists are institutionally accorded the deontic authority to direct the course of the session by their status, but they have the option to present a downgraded deontic stance in order to enact a more symmetrical distribution of deontic rights. In workplace conversations between church staff, Stevanovic and Peräkylä, (2012) described differences in turn design and how these reflected the speaker’s deontic stance. For example, an assertion about a future plan displays and claims a higher deontic stance because it states the next course of action without requiring acceptance from the recipient. Conversely, a proposal seeks the acceptance of the recipient before the conversation can progress (Sacks, Schegloff, & Jefferson, 1974; Schegloff & Sacks, 1973), and consequently downgrades the speaker’s deontic stance. Speakers can further mitigate the imposition to respond to a proposal by prefacing it with “I was thinking”, and thereby constructing the proposal as a “thought” (Stevanovic, 2013a). Similarly, Curl and Drew (2008) identified that when making requests, speakers often orient to deontic rights. If people made requests with the form “can/could you...” this framed their request as being unproblematic and easily fulfilled by the recipient. But if the request incorporated the words “I wonder if...” the speaker displayed an orientation to their low entitlement to make that request or anticipated difficulties in their request being granted by the recipient. Grammatically, the addition of “I wonder...” transforms a statement from a request or a directive to a reporting of the speaker’s own thoughts, which mitigates the imposition on the recipient to respond and downgrades the speaker’s deontic authority (Craven & Potter, 2010; Pomerantz, 1980; Stevanovic, 2013a). In a CA study of cognitive behavior therapy, Ekberg and LeCouteur (2014, 2015) demonstrate how therapists and clients negotiate deontic rights in proposals 122
for behavioral change. They show how therapists can first elicit a client’s ideas for behavioral changes, and then modify and add their own ideas to these suggestions. This shows how therapists are active in guiding the conversation while also collaborating and coimplicating the client’s responses into suggestions for future actions. If therapists made more direct suggestions without first eliciting and coimplicating the clients’ suggestions then clients could invoke reasons, and the resulting deontic authority, to reject the therapist’s proposals. This highlights the dilemma faced by therapists when balancing the achievement of therapeutic goals with a collaborative perspective. CA and the concept of deontics can therefore provide a method for accessing power as a set of practices and normative orientations employed and negotiated in conversation, rather than an abstract concept. In this study, we utilize CA to explore how therapists in Open Dialogue orient to deontic authority when proposing a transition to reflections.
METHOD Data and Ethics Data were drawn from a corpus of ten video recordings, totalling fourteen hours of Open Dialogue sessions conducted in a child and adolescent mental health service in suburban New South Wales, Australia in 2018. Participants included therapists utilizing the Open Dialogue approach (n = 12), as well as identified clients of the service, their family, and other professionals involved in their care (n = 36). Ten therapists had completed a 5-day foundation training and had a minimum of 2-year experience and supervision in Open Dialogue. Seven of these therapists were also undertaking advanced training in Open Dialogue run by clinicians from Finland, in either a 3-year therapist course, or a 2-year course to become trainers. Two therapists had informal orientation to Open Dialogue through their more experienced colleagues. Therapist participants were informed about the study by the first author (who is a colleague in the same service). Therapists then approached the first author if they wished to participate. Participating therapists then approached clients and their families if they wished to participate. With all participants it was stressed that participation was voluntary. All participants provided written, informed consent before participation in the study. The therapy sessions were video and audio recorded with three cameras positioned around the room in order to capture details of each participant’s embodied (e.g., direction of gaze, hand gestures) and verbal communication. The study and recruitment procedures received ethics approval from the Nepean Blue Mountains Local Health District ethics committee (reference number: HREC/17/NEPEAN/135) before data collection commenced. Names and identifying information have been changed in the extracts below to ensure participant anonymity.
Analytic Process Video recordings for all sessions were reviewed and transcribed verbatim. Following the procedural recommendations for CA (ten Have, 2007) we began with a period of “unmotivated looking”. This involved reviewing the video recordings and transcriptions without prior expectations or hypotheses for interactional phenomena of interest (Psathas, 1995; Schegloff, 1996). During this process, we noticed the reoccurrence of a conversational practice regularly utilized by therapists: the words “I’m wondering” (and its morphological variants). We
123
then focussed our analysis on this conversational practice (Kent & Kendrick, 2016). On closer examination, it became evident that the use of “I’m wondering” occurred in different turn positions with apparently different actions. One prominent location for “I’m wondering” was when therapists were proposing a transition to a reflection. This leads us to examine proposals to commence reflections in particular. We identified fifteen examples of reflection proposals, and these were then transcribed in greater detail according to CA conventions (Hepburn & Bolden, 2017, see Table S1). Utilizing the transcript and the original video recordings, each example was analyzed with attention to conversational features including: the sequential position of the proposal through reviewing the previous turns; the addressed recipient(s) through gaze and verbal resources; the design of the proposal, particularly word choice, intonation, and prosody; and the responses of the recipients of the proposals (ten Have, 2007). The analyses of the different examples were then compared to identify if there were systematic normative patterns realized via the different forms of proposals (Heritage, 1988). We also identified examples of deviant cases to support our analysis (Potter, 1996; Sidnell & Stivers, 2013). Deviant case analysis is a validation tool that identifies examples that do not initially fit with the proposed normative patterns. On closer examination, if the responses of the participants show an orientation to the contrastive behavior that is consistent with the proposed normative reasoning, this lends further support for the analysis.
FINDINGS Overview of Findings Reflection proposals were only made by the therapists conducting the session. Furthermore, eleven of the fifteen proposals were addressed to other clinicians, with four examples addressed to the family. The transition to reflections were therefore entirely therapist-initiated and predominantly therapist-negotiated activities. In addition, therapists commonly presented a downgraded deontic stance when proposing transition to a reflection through either asking permission or by downgrading practices such as turns prefaced with “I’m wondering”. In only one example did the therapist not seek some form of assent or confirmation. We now present examples to illustrate these findings.
Downgrading Authority Through Asking and Accounts Therapists downgraded their deontic authority though a polar (yes/no) question and providing accounts for their proposals. By asking if they can transition to a reflection the therapist downgrades their deontic stance as the transition becomes contingent upon the acceptance of the recipient. Similarly, an account is offered when a person’s behavior departs from expectations and the speaker orients to a need to justify their actions, thus downgrading their deontic authority (Heritage, 1988; Houtkoop-Steenstra, 1990).
124
In Extract 1, a client and their mother are disagreeing about seeing a sister who is not present (data not shown). After closure of this sequence and a lapse in the conversation (line 1), T2 addresses T1 with an interrogatively formatted “can we have another reflection,” (line 2). The use of “can” suggests that the speaker has high entitlement and/or anticipates minimal difficulties with acceptance of their proposal (Curl & Drew, 2008). After a nodding acceptance from T1, T2 then addresses the proposal to the family (line 4). Mo’s “sure-” (line 5) reflects her positive deontic stance and orients to T2’s proposal as seeking her deontic authority to accept (Seuren, 2018). Although the design of the therapist’s proposal suggests minimal difficulties with acceptance, it also makes transition to the reflection contingent upon the acceptance of the recipients. Consistent with the form of the therapist’s question (Stivers & Enfield, 2010), the recipients provide acceptances via nodding and a clipped “sure-”, respectively. In addition, at line 7 the therapist provides an account though “we’ll try an keep it sho:rt¿” This further contributes to the therapist’s downgraded deontic stance by presenting the reflection as being short and consequently of little imposition on the family. Extract 2 provides a more direct demonstration of the therapist’s deontic authority despite their deontic downgrades.
T1 initially starts with “↑would” (line 2), suggestive of an upcoming interrogative turn format. This is abandoned in favor of a declarative turn format with rising intonation (lines 2 and 4) “maybe this is a good (0.3) time ta tu::rn ta (.) the team?” (referring to the reflecting team). The therapist downgrades their deontic authority through beginning their turn with “maybe”, and the justification that “this is a good (0.3) time:” for a reflection. The therapist also seeks acceptance for the proposal though rising intonation and eye gaze (Stivers & 125
Rossano, 2010) toward one of the reflecting clinicians (R1, line 4). Despite these indications of a downgraded deontic stance, there is evidence the therapist maintains a high level of authority to direct the session. The design of T1’s talk and other conduct from lines 2 and 4 indicates that the proposal has multiple addressees. Firstly, the wording of the turn suggests that T1 is addressing the client and the family because the reflecting team is referred to in the third person. But during this turn, T1 also gazes toward one of the reflecting therapists (line 4). T1’s proposal is addressed both to the family and the reflecting team via verbal and embodied modalities respectively. The sister orients to the expectation that the proposal requires acceptance though her “okay” (line 7). However, there is no response from R1. T1 continues to pursue a response with an extension of their turn “and (0.5) see what they have to sa:y?” This extension is still worded as addressing the family, and while there is no verbal response from the reflecting clinicians, they show their acceptance of T1’s proposal through embodied action, by turning their chairs to face each other to start the reflection (line 9). The reflecting team thus performed the proposed action by starting the reflection, but without first vocalizing a commitment to performing it (Stevanovic & Peräkylä, 2012). The response of the reflecting team suggests that the therapist’s turn is received not as a proposal requiring acceptance, but as a directive requiring compliance. So, despite the therapist’s efforts to downgrade their authority, recipients still orient to the high deontic status that is associated with the therapist’s role as the leader of the session. Despite the high deontic status of the therapist, they refrain from a directive to transition to a reflection (Craven & Potter, 2010). Instead, therapists proposed reflections through polar interrogatives and offering accounts. These practices downgrade a therapist’s deontic authority as they make acceptance of the proposal necessary before the reflection can proceed. However, the polar interrogative form of proposals suggests that the therapists view their proposals as being relatively unproblematic (Curl & Drew, 2008). So while polar interrogatives and accounts downgrade the therapist’s proposal they do so in relatively minor ways and maintain a focus on the (likely unproblematic) acceptance or rejection of their proposal. Recipients provide minimal forms of acceptance either verbally or through embodied responses, thus displaying an orientation to and acceptance of the therapist’s deontic authority to make the proposal. In the next section, we describe examples where therapists’ reflection proposals fall further along a downgraded deontic gradient.
Downgrading Authority Using “I’m wondering” The phrase “I’m wondering” and its morphological variants were commonly utilized, having occurred in ten of the fifteen examples of reflection proposals. In the following examples, we demonstrate how “I’m wondering” is used to further downgrade the speaker’s deontic stance. Just prior to Extract 3, the client has been talking about his mother cleaning when she is stressed.
126
At lines 1–3, the client recalls his mother picking something up after he has said he would do it. This sequence closes with laughs and silences. T2 then proposes transitioning to a reflection (lines 9–11). Beginning with “so (0.5) I’m wondering” the therapist marks the following talk as a report on their current thoughts (line 9). As mentioned, “I wonder” and presenting a proposal as a thought mitigates the imposition on the recipient and downgrades one’s entitlement to make that proposal (Curl & Drew, 2008; Stevanovic, 2013a). T2 also provides an account for having a reflection: “it’s a good time” (lines 9). At line 11, T2 ends their turn with rising intonation thus seeking a response. When no response is forthcoming from the family, T2 continues their turn with an expansion specifying what the reflection will involve. T2’s turn again ends with rising intonation that seeks a response from the family (line 13). An accepting response comes from all family members through nodding and a “yep” from Mo, and a later “yeah” from Cl. Mo’s “that’s cool” at line 16, is also a positive deontic stance orienting to the therapist’s talk as a proposal and her deontic rights to accept (Seuren, 2018). After this acceptance T2 seeks further confirmation with “yeah? (0.4) okay¿” at line 18, in overlap with client’s acceptance followed by confirming nods from Mo and Fa. T2 presents their proposal as something that they are “wondering” about, as being in need of justification, and requiring the acceptance of the family. However, the family’s minimal responses of acceptance do not provide an indication of the potential barriers foreshadowed via the therapist’s turn design. It seems that the recipient orients to the deontic status of the therapist rather than their downgraded deontic stance. In Extract 4, the therapist’s downgraded authority and invitation for joint decision- making is even more clearly marked.
127
At lines 2 and 4, T2 invites T1 to speak having noticed some previous preparation to speak by T1. T1 presents a deontically downgraded reflection proposal beginning with “I wondered if” (line 9). At lines 12–13, T1 provides a specific reason for the downgrade: “but I don’t know (.)>where you’re at<”. T1 explicitly expresses the potential problem of proposing a reflection at this point as uncertainty about whether T2 is agreeable to a reflection, thus orienting to high contingency as well as low entitlement in making the proposal. This is further highlighted by T1 seeking confirmation at line 18, even after three instances of agreement by T2, including “yeah” and nodding at lines 14, 15, and 16. Through these responses, T2 provides weak and minimal acceptance of the proposal. T1 may therefore be orienting to these minimal and weak acceptances as not adequately embracing their heavily downgraded proposal. Furthermore, unlike proposals that utilize a polar interrogative, and more clearly set up the relevance of acceptance or rejection, proposals with “I’m wondering” are more ambiguous and equivocal, providing a broader range of possible responses by recipients.
Contrastive cases In this section, we provide further support for our analysis though presenting a contrastive example (Extract 5) and a deviant case analysis (Extract 6). In Extract 5, a therapist proposes a reflection to another therapist, but, instead of the usual minimal acceptance from the recipient and transition to the reflection, there is a period in which deontic rights and the proposal are explicitly negotiated. Extract 5 starts with Cl stopping Mo from speaking about Cl’s difficulties getting ready in the mornings, followed by some joking and laughing. T2’s proposal begins at line 18 and includes the downgrading features of “I’m wondering” and an account that they “should” have a reflection. T2 also proposes an alternative action, that is, if they should ask the third therapist, Greta, “what she’s thinking” (line 23). T2’s proposal thus makes relevant not acceptance but a greater degree of collaboration from the recipient who has to make a selection between two alternatives. By only providing two choices T2 constricts the number of possible responses that T1 can expectedly provide and presupposes that there will be some form of change in the session, ostensibly an upgraded deontic position (Antaki & Kent, 2015).
128
After T1 selects the option of speaking with Greta (line 25), T2 does not explicitly acknowledge this response and instead invites further collaboration. This turn begins with an account “I’m not sure how we should do this” before again presenting two possible options i.e., should one or both of them talk to Greta. At line 29, T1 proposes “all three” of them speak together in overlap with T2. T2 again does not respond to T1’s proposal but instead finishes this turn, and then after a gap, asks the family if they have a preference (line 32). T2 thus seeks collaboration from T1 on multiple occasions but when it is forthcoming it is not acknowledged. T2 seeks collaboration in decision-making from both T1 and Mo, but this results in overlooking T1’s decision. From line 37 there is a back and forth wrestling of deontic authority between T1 and T2. T2 begins formulating a decision on how to progress the session “maybe we’ll (0.2) maybe”. Although downgraded with “maybe”, T2 is in the process of deciding the direction for the session, thus claiming deontic rights to decide. In overlap, T1 proposes “all three?” repeating their earlier unacknowledged suggestion, consequently reclaiming some deontic authority in deciding the next part of the session. T2 then repeats “all three of us will (.) just (0.5) ask greta what she’s hea:rd¿” At line 40, T1 confirms this with “↑o↓kay”. With falling intonation, this works to confirm T2’s acceptance of T1’s earlier suggestion. This sequence thus displays a back and forth claim and counter-claim of deontic authority between T1 and T2.
129
This example deviates from the usual trajectory of reflection proposals where the therapist proposes a reflection with various downgrades of their deontic stance and seeking the acceptance from the recipients. The proposal is then usually followed by minimal acceptance by the recipient with transition to the reflection. But, in this example, the therapist making the proposal seeks the input of the other therapist, but then does not explicitly acknowledge the other therapist’s response. The therapist thus makes a downgraded deontic proposal but their following actions are not consistent with this downgraded position. This results in competition for deontic authority between the therapists and risks
130
destabilizing the session and its methodical development. Extract 6 is an example of a deviant case where the therapist does not follow the usual form of reflection proposal, i.e., downgrading their authority and seeking acceptance. This is met with a disaligning (i.e., unsupportive) response from a recipient, followed by the therapist amending their turn and seeking acceptance (although cursorily). Extract 6 begins with C3 protesting against her sister (Julia) previously saying that C3 does not like birds (data not shown).
131
At line 11, T1 initiates reflection by stating that they want to check with their col- league, Charlie, if he has other things on his mind. T1 then physically turns toward T2 marking the beginning of the reflection. T1’s initiation of reflection departs from the previous examples as the account for the reflection is provided first, and the reflection is stated as a unilateral decision without further downgrading elements nor seeking acceptance. This reflection initiation also occurs toward the end of the session and this movement toward ending the session may explain T1’s transition to a reflection without first seeking acceptance. At line 16, C3 disaligns with the transition with a loudly spoken “NOT again”. At line 17 and 19, T1 adds an increment to their earlier turn (now addressed to T2) if there is “anything else that you’ve heard?” that he may want to add. T1 then momentarily turns back to C3 seeking agreement “if that’s okay” before returning to T2, this appears to be a delayed response to C3’s “NOT again” at line 16. C3 provides acceptance at line 22 with “that’s fine”. This example differs markedly from the forms described above. When T2 departs from the usual form of reflection proposal it is met with a disaligning exclamation from one of the recipients. In response, T2 suspends the transition to the reflection and orients to the need to gain acceptance before the reflection can proceed. The participants thus orient to the implicit preference that the therapist first gain acceptance from clients before a reflection can proceed.
DISCUSSION When proposing a transition to a reflecting conversation, therapists using Open Dialogue routinely work to downgrade their deontic authority. These downgrades fall along a deontic gradient from relatively minor downgrades, utilizing polar interrogatives that present the proposal as relatively unproblematic, to more heavily downgraded proposals incorporating multiple downgrading features such as the use of “I’m wondering”, providing accounts, and specifically seeking acceptance and confirmation from recipients. The polar interrogative forms make directly relevant the acceptance or rejection of the proposal by the recipient. Consequently, they maintain a focus on the progressivity (i.e., unproblematic development) of the conversation toward the reflection, while also requiring the collaborative acceptance of the recipient/s. Proposals utilizing “I’m wondering” are more equivocal and do not make acceptance or rejection as strongly relevant. These proposals downgrade the therapist’s authority and invite collaboration, but also potentially obscure the action of transitioning to a reflection. As illustrated in Extract 5, this invitation to collaborate can result in the therapist repeatedly exerting their deontic authority as they seek and respond to the suggestions of other participants. In general, recipients accepted reflection proposals quickly and without negotiation, apparently orienting to the deontic status, rather than the stance of the therapists. The persistence of therapists’ deontic downgrades may therefore serve functions other than the seeking of collaborative decision-making. Various approaches to family therapy have emphasized therapist expertize with mechanistic perspectives on family problems and the therapist as an external agent intervening in the family system (Flaskas, 2010). However, clinicians working in Open Dialogue construct their identities through distancing themselves from the traditional roles as “experts” or “fixers” of mental health problems and promote a willingness to be vulnerable in front of clients (Schubert, Rhodes, & Buus, 2020). Writings on Open Dialogue and dialogical approaches similarly recommend that
132
therapists take a position of “not-knowing” (Anderson & Goolishian, 1992), following the lead of the client (Seikkula, 2011), and tolerating doubt and uncertainty (Seikkula & Arnkil, 2006). Doubt, uncertainty and the pursuit of collaboration are thus central theoretical aspects of Open Dialogue. However, these concepts are ambiguous and ill-defined providing therapists with little specific guidance on how these concepts may be brought into practice. Peräkylä and Vehviläinen (2003) have described how CA’s detailed analysis of actual conversations can expand our understanding of therapeutic practice by providing more details on how ambiguous therapeutic concepts, such as collaboration, are actually achieved in practice. This study thus demonstrates one way that the general concept of “collaboration” may be manifested by therapists through conversational downgrading of their deontic authority. Further research in this area could greatly improve our understanding of what constitutes dialogical therapy and thus improve teaching and clinical practice. A limitation of this study is that although the participating clinicians have a number of years of experience in Open Dialogue they were also undertaking advanced training at the time of data collection. Therefore, their work may reflect their attempts to incorporate new ideas and techniques into their practice and may not represent “good” Open Dialogue practice. However, there does not currently exist any accepted measures of fidelity in Open Dialogue research. Furthermore, the principles of Open Dialogue are sufficiently ambiguous to allow for a range of possible manifestations in different contexts. In fact, it has been proposed that Open Dialogue will necessarily need adapt and develop to the needs of the local contexts where it is applied (Buus et al., 2017). So rather than providing definitive directives for practice, we instead hope that this study continues a “dialogue” about what constitutes dialogical practice. This study also empirically highlights the pervasiveness of deontics in clinical practice with therapists regularly downgrading, and thus orienting to, differences in authority. Differences in power and authority are unavoidable elements of a therapeutic relationship (Hare-Mustin, 1994; White & Epston, 1990). But the role of power in Open Dialogue has been under-developed, with power seen as simply impeding the creation of dialogue (Guilfoyle, 2003). Our findings suggest that deontic authority and directing the course of a session are not necessarily contrary to the values of Open Dialogue, but rather something that is relevant, downgraded, and negotiated by participants to varying degrees. Enfield (2013) asserts that relationships are defined by the nature of the interactions between people and their relative enactment of rights and duties. In this view, an awareness and negotiation of authority are potentially meaningful in developing a certain forms of relationships between therapists and clients. However, further research is necessary in order to understand the complex negotiations of authority in different parts of an Open Dialogue session. We can conclude that experienced therapists undergoing advanced training in Open Dialogue orient to deontic authority in their interactions with clients and work to present a downgraded deontic stance that is weaker than their institutional status would implicate. However, further research systematically exploring the deontic relations across everyday conversation, Open Dialogue, and other therapeutic approaches is necessary in order to confirm if these forms of deontic downgrading are unique to Open Dialogue. Previous findings on requests and proposals in institutional and everyday contexts (Curl & Drew, 2008; Stevanovic, 2018) suggest that deontic issues are, to some extent, generically relevant for human communication, even if there is some particularization of
133
them in the Open Dialogue context.
SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Table S1. Transcription Symbols.
134
Chapter Summary
This study highlights how deontically downgraded stance proposals are one way that therapists promote collaboration and joint decision making in practice. But, although therapists make attempts to downgrade their deontic authority when proposing reflections, recipient responses orient to the deontic status of the speaker rather than to their presented downgraded deontic stance. This suggests that although therapists make attempts to downgrade their authority, this authority remains relevant and is never completely removed.
The presentation of a downgraded deontic stance serves functions in addition to seeking collaboration and joint decision making, such as promoting alternative relationships and identities. Different relationships are defined by different constellations of rights and responsibilities (Enfield, 2013). By downgrading their authority when proposing reflections, therapists present a disavowal of rights to make unilateral decisions about the structure of the session, which invokes a particular identity for the therapist that differs from psychotherapy approaches that are more directive (Schubert, Rhodes, & Buus, 2020). Furthermore, heavily downgraded forms of proposals allow recipients the freedom and flexibility to respond in a variety of different ways as per the concept of conditional relevance (Chapter 2). For example, a polar interrogative, such as “can we have a reflection?” makes relevant a “yes/no” response. While more open forms of turn design, such as, “I wondered if there was space for reflection” make relevant a wider variety of responses. More downgraded forms of reflection proposals seek collaboration from the recipient while also placing fewer restrictions on types of responses. This promotes agency because the recipient has greater flexibility in how they may expectedly respond (Enfield, 2013).
In summary, deontically downgraded reflection proposals have a range of possible implications. These include seeking collaboration and joint decision making on the progression of the session, promoting alternative therapist-client relationships and identities, 135
and promoting agency through flexibility by reducing the obligation of the client to respond in narrowly defined ways. Downgraded deontic stances can contribute to the existing stocks of interactional knowledge in Open Dialogue as they reveal one way that the concept of
“collaboration” is actuated in interaction. This theme of balancing authority while also giving the recipient a wider variety of ways to respond is present in the next chapter on how therapists select next speakers while also allowing multiple ways for them to respond and voice their own stance positions.
136
Chapter 5: Eliciting Stance
In this chapter, I present a study accepted and published in the Journal of Marital and
Family Therapy (Ong, Barnes, & Buus, 2021). This article explores how therapists utilised a particular conversational practice “I’m wondering” for the purposes of eliciting stance positions from the session participants. A central idea in Open Dialogue is the concept of polyphony, referring to the presence of multiple voices. However, articles describing the
Open Dialogue approach provide little detail on how the concept of polyphony is enacted in conversation. This study describes the elicitation of various stance positions as one way that polyphony can be promoted in Open Dialogue conversations.
137
Journal of Marital and Family Therapy 47(1): 120–135 doi: 10.1111/jmft.12454
© 2020 American Association for Marriage and Family Therapy
Eliciting Stance and Mitigating Therapist Authority in Open Dialogue Meetings
Ben Ong University of Sydney Child and Youth Mental Health Service
Scott Barnes Macquarie University
Niels Buus University of Sydney St. Vincent’s Hospital Sydney St. Vincent’s Private Hospital Sydney University of Southern Denmark Open Dialogue is a collaborative systemic approach to working with families in crisis. A core feature is the creation of dialogue through the elicitation of a multiplicity of voices. Using conversation analysis, we studied 14 hr of Open Dialogue sessions. We found that therapists recurrently produced utterances containing “I’m wondering.” These utterances topicalized particular issues and invited stance positions from other participants while also allowing the therapist to mitigate their deontic authority and present potentially disaligning stances. Therapists thus exercised authority in eliciting stances, but provided recipients with multiple avenues for responding. These findings illustrate that therapist authority is not necessarily antithetical to dialogue and, in well-crafted forms, may even be necessary for the creation of polyphony through the elicitation of multiple stances.
Open Dialogue is a collaborative and systemic approach to working with people and their families, who are experiencing psychological crises (Gromer, 2012; Haarakangas, Seikkula, Ala- kare, & Aaltonen, 2007; Seikkula, 2003; Seikkula & Arnkil, 2006). Originating in Finland, it has growing interest in Scandinavia and internationally (Buus et al., 2017; Gordon, Gidugu, Rogers, DeRonck, & Ziedonis, 2016; Ong et al., 2019; Razzaque & Wood, 2015; Rosen & Stoklosa, 2016). Open Dialogue is characterized by a number of principles including immediate help; the inclusion of a person’s social network; and the flexibility, continuity, and responsiveness of the therapy team. The central principle of the approach, however, is the promotion of dialogue through the elicitation of multiple perspectives (Olson, Seikkula, & Ziedonis, 2014). These multiple perspectives, also known as polyphony, include the different voices of all the people present in the therapy session (including the therapist), as well as the various thoughts, emotions, and responses within each person (Seikkula, 2008). The role of the therapist in Open Dialogue—and dialogical approaches
Ben Ong, BPsychHons, MCFT, Susan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, University of Sydney, Australia. Nepean Blue Mountains Local Health District, Child and Youth Mental Health Service, Penrith, NSW, Australia; Scott Barnes, BSpPath (Hons), PhD, Department of Linguistics, Faculty of Medicine and Health Sciences, Macquarie University, Australia; Niels Buus, BN, MScN, PhD, Susan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, University of Sydney. St. Vincent’s Hospital Sydney, Australia; St. Vincent’s Private Hospital Sydney, Australia and Institute of Regional Health Research, Univer- sity of Southern Denmark, Odense, Denmark. Address correspondence to Ben Ong, Susan Wakil School of Nursing and Midwifery, Faculty of Medicine and Health, University of Sydney, Australia; Nepean Blue Mountains Local Health District, Child and Youth Mental Health Service, Penrith, Sydney Nursing School, 88 Mallett St. Camperdown, NSW 2050, Australia; E-mail: [email protected]
138
generally—is to create the conditions and opportunities for these multiple voices or perspectives to be expressed in the therapeutic conversation (Anderson & Goolishian, 1992; Olson et al., 2014; Seikkula, 2002, 2008). Therapists must balance being responsive and respectful of each person’s contributions while also managing the participation and speaking time of each family member (Haarakangas et al., 2007; Olson et al., 2014). Dialogical approaches also promote equality and collaboration between the therapist and the family. Therapists are advised to understand the problem from the perspective of the family, follow the lead of the family, and to base their responses on what the client and family have said (Ander- son, 2002; Anderson & Goolishian, 1988; Seikkula, 2011; Seikkula & Arnkil, 2006). This creates a potential dilemma for therapists as they are advised to be collaborative and to follow the lead of the family while also having to be directive to manage the session. For example, therapists may need to manage speaking times so that multiple perspectives can be heard, or to make decisions on when to transition to a reflection, or to decide what topics are important to explore in greater detail. There has been scant empirical investigation of how therapists actually manage these poten- tial conflicts in real-life therapy sessions. One research technique that can reveal how such family conversations are actually conducted is conversation analysis (Ong, Barnes, & Buus, 2019, 2020; Schriver, Buus, & Rossen, 2019; Tseliou, 2013). Conversation analysis (CA) is an approach to analyzing verbal and non-verbal interaction as it occurs in real-life situations (Goodwin & Heritage, 1990; ten Have, 2007; Sidnell & Stivers, 2013). CA focusses on the observable details of interaction to describe the conversational practices and normative expectations that people utilize and orient to when coordinating their social activi- ties. Put another way, CA seeks to describe how people achieve various social actions through con- versation and the methods they employ to achieve them. CA has been applied to different aspects of family therapy (Ong et al., 2019, 2020; Tseliou, 2013). For the current study, two features of conversation are particularly relevant: stance (Du Bois, 2007) and deontics (Stevanovic, 2018; Ste- vanovic & Peräkylä, 2012, 2014). The expression of stance is a conversational, linguistic act that invokes some form of eval- uation (Du Bois, 2007). Stance can take a number of possible forms, and may invoke attitudes, emotions, epistemics (knowledge), and/or deontics (power) (Stevanovic & Peräkylä, 2012; Stivers, 2008). Rather than focusing singularly on types of stance, Du Bois (2007) instead recommends describing how the stance act occurs in interaction. According to Du Bois (2007), a stance act is an interactional achievement constructed through a sequence of turns at talk. In a stance act, a speaker evaluates a particular stance object. Consequently, this same person is reciprocally positioned in relation to the stance object. Other people may also describe their stance position towards that same stance object. The relative level of agreement between these two stance positions varies along a continuum from alignment to disalignment (Du Bois, 2007). The analysis of stance therefore requires attention to interlocking sequences of talk rather than to the internal states of individuals (Kärkkäinen, 2006). An understanding of stance, on these terms, has the potential to add depth to family therapy practice and research by describing family interactions and the positions people adopt as they actually occur in conversation. For example, Everri, Fruggeri, and Molinari (2014) demonstrated how an analysis of stance can be used to understand the micro-transitions involved in the development of identities in families. Stance can therefore aid clinicians in developing empirically informed hypotheses about family therapy and describing how family members construct identities (Everri & Fruggeri, 2014). Logren, Ruusuvuori, and Laitinen (2019) have used analyses of stance to examine responses to self-disclosures in group diabetes health counseling sessions. They found that, after a self-dis- closure by a group participant, other participants responded in a variety of ways. These included an aligning stance through sharing the same experience, or
139
by disaligning through offering a contrasting perspective. Their focus on group processes meant that they excluded responses by the clinician. The clinician, however, has a unique position in relation to stance acts because of the deontic authority associated with their institutional position. Deontic authority refers to the rights of a person to determine future actions in a particular domain (Stevanovic, 2018; Stevanovic & Peräkylä, 2012). While power can be considered as the ability to unilaterally impose consequences, authority is seen as legitimate and obeyed by the sub- ject of authority with free will (Stevanovic & Peräkylä, 2012). Deontics specifically refers to how power is expressed and negotiated by participants in interaction. Due to our focus on conversa- tional interaction, we focus on deontic authority rather a broader representation of power as some- thing external to interaction. A person with a high deontic status, in a particular domain, has greater authority relative to others. Depending on the context, this authority can come from their personal history, level of expertise, or their institutional position. Deontic rights are not fixed or static but instead are regularly displayed and re-negotiated in conversation (Asmuß & Oshima, 2012). Deontic stance describes how a person presents their authority in interaction. A person may have a high deontic status but, for various interactional reasons, present a downgraded deontic stance (Stevanovic, 2013). In family therapy, therapists have the authority or deontic status, inher- ited through their institutional role, to direct and guide a session (Guilfoyle, 2003; Hare-Mustin, 1994). For example, Ong, Barnes, and Buus (in press) have examined how Open Dialogue thera- pists propose a transition to a reflecting conversation. This study found that although the thera- pists had a position of high deontic authority, when they proposed transitioning to a reflection they made efforts to present a downgraded deontic stance. They did this through prefacing their proposals with “I’m wondering,” asking permission to have the reflection, and by providing rea- sons for their proposals. The therapist’s deontic status may therefore create potential difficulties for a dialogical therapist who wishes to promote dialogue and particular forms of participation from clients and family members. Ong et al. (in press) focused on deontic authority specifically in relation to therapists’ proposals for reflections and did not examine how deontic authority inter- acts other therapist activities such as the elicitation of stance. Using CA, the aim of this study is to explore how therapists following an Open Dialogue approach elicit multiple stances and manage their own deontic status and stance in family sessions.
METHODS
Data The data for this study come from a corpus of 10 Open Dialogue sessions. A total of 14 hr of video recordings were collected in 2018 in a child and youth mental health service in Western Syd- ney, New South Wales, Australia. Each session was video- and audio-recorded with three cameras placed in different points around the room to capture the verbal and embodied communication from each participant. Participants included 12 clinicians training in the Open Dialogue approach, and 36 community members including clients and their personal and professional networks, such as family and support workers. Ten of the therapists had completed a 5-day training program in Open Dia- logue and had at least 2 years of experience in both Open Dialogue practice and supervision. At the time of data collection, seven of these therapists were participating in advanced training in systemic practice through either a 3-year Open Dialogue therapist course, or a 2-year Open Dialogue trainer program. This training was conducted by visiting clinicians from Finland. The remaining two thera- pists had an informal orientation to Open Dialogue practice by their more experienced colleagues. The recruitment process involved information sessions about the study run by the first author,
140
who works in the same service as the therapist participants and has not been involved in any ser- vice provision to the participating clients and families. If therapists wished to participate, then they approached the first author to provide consent. The participating therapists then approached cli- ents and families that they worked with to provide basic information about the study. If clients and their families were interested, then the first author provided them with more information about the study. If clients and families still wished to participate, then they provided written con- sent. There was no prior selection of sessions to be recorded. Instead, recorded sessions were col- lected based on mutual availability of the therapists, families, and the first author to set up the recording equipment. The study was approved by the Nepean Blue Mountains Local Health Dis- trict Ethics Committee (reference number: HREC/17/NEPEAN/135) before data collection began. All identifying features have been changed in the presented extracts to maintain anonymity and confidentiality.
Analytic Process The analytic process followed recommended procedures for CA (e.g., ten Have, 2007; Pomer- antz & Fehr, 2011). This included a basic transcription of the recordings and then reviewing the recordings and transcripts together, adopting an “unmotivated looking” approach. This involved inspecting the data without prior expectations, hypotheses, or assumptions for phenomena of interest (Psathas, 1995; Schegloff, 1996). During this process, we noted the recurrence of a conver- sational practice used overwhelmingly by therapists: the phrase “I’m wondering” and its morpho- logical variants. Taking this conversational practice as our starting point, we sought to describe how it was utilized, and what actions and interactional functions it served (Kent & Kendrick, 2016). We identified all instances of this practice (n = 119) which were then transcribed in greater detail according to CA conventions (Hepburn & Bolden, 2017, see Appendix 1). This included all instances produced by all participants, although most were produced by therapists (n = 112, 94%). We found that “I’m wondering” can occur in different positions in speaking turns, but were mainly employed at the beginning of utterances and in the course of an utterance. “I’m wonder- ing” was utilized for different actions in these different positions. Ong et al. (in press) examined how “I’m wondering” was utilized at the beginnings of turns and specifically in therapists’ propos- als to have a reflecting conversation. In this analysis, we report on uses of “I’m wondering” that are employed not at the beginning of turns, but in the middle of turns as part of an ongoing utter- ance. We identified 25 examples of “I’m wondering” as part of an ongoing utterance, of these, only four examples were spoken by family members. We then analyzed each of these examples with par- ticular attention to turn-taking, sequence and sequential organization, and recipient responses. We then compared each example for similarities between turn design and what actions each example was doing to identify common patterns (Heritage, 1988). We also looked for examples of complex and deviant cases that did not seem to fit the common patterns (Potter, 1996). Ultimately, this iter- ative analytic process was directed toward ensuring that analytic findings were robustly grounded in the moment-by-moment sense- making of the parties to the interaction.
FINDINGS
Overview of Findings The use of “I’m wondering” by therapists as part of an ongoing utterance arises in a particu- lar location and serves a number of functions. It is preceded by a set-up that is backwards look- ing and acknowledges or topicalizes some prior talk by the family. “I’m wondering” then introduces some new configuration for the interaction. This is through selecting a new speaker
141
and eliciting a new stance position, and/or through introducing a new matter for development. The talk following “I’m wondering” is recipient-focused as it seeks comment on matters within the recipient’s epistemic domain. “I’m wondering” implicates the therapist’s own stance position- ing, which makes them responsible and accountable for what they say next. However, it also mit- igates the epistemic authority of the therapist, making the stance position they adopt defeasible. This allows the therapist to introduce a controversial idea or question, which can later be dis- avowed if required. Consequently, “I’m wondering” is a concession, by the therapist, to the epis- temic authority of the recipient regarding the matter being raised. In addition, therapists work to design their stance-eliciting questions as open and thus allowing for greater flexibility in the responses of the recipients. This means that, even though the therapist is exercising their deontic authority by selecting the next speaker and nominating a topic for discussion, the design of these turns affords the recipient some responsive flexibility. As we will demonstrate, in some cases, this flexibility causes problems for turn uptake. Below, we present specific examples to illustrate these findings.
Eliciting Stance and Mitigating Deontic Authority In Extract 1, we present an example of how the therapist topicalizes a part of the family’s prior talk. Using “I wonder,” the therapist transitions to the stance-eliciting component while also miti- gating the therapist’s deontic authority to direct the recipient’s response. At the beginning of this extract, the family have been talking about difficulties caused by the client (C3), (Isabella) and her distress. This includes a stance position that describes the negative effect that the distress has been having on the mother (data not shown) and on everyone else having to look for things that the cli- ent has lost (lines 1–4).
Extract 1 (S4.E2, 16:12) T1: therapist 1; Mo: mother; Fa: father; C1: child 1; C2: child 2; C3: child 3
1 C1 like (0.2) [go:d forbid ] one of=
2 C2 [wasn't it one of her]
3 C1 =her:: like toys goes missing (0.2) .h (.)
4 coz we=all be looking for it
5 (0.8)
6 C2 mm'na:
7 (0.5)
8 C1 HH um: h what else [((laughs))]
9 C2 [((laughs))]
10 .hh
11 C1 [u:m ]
12 T1 [i feel] like i'm (.) i (0.2) >haven't (.)
142
13 quite heard (0.3) as much from: (0.5) natalie
14 and isabella (0.3) i know you started the
15 conversation natalie by saying:=
16 Mo =a'£wa (0.2) [ha]
17 T1 [.h] (0.7)
18 the effec[t: ]=
19 Mo [((cl. throat))]
20 T1 =(0.3) [an then] there's been a lotta=
21 Mo [hh ]
22 T1 =conversation about the effect (.)
23 -> an i wonder what (0.5) what's going on:
24 (0.5) listening to that or
25 (1.3)
26 C3 >what do you think [about]<
27 T1 [what ] (0.5)
28 wha[t's come u]p¿
29 Mo [tch .hh ]
30 (0.3)
31 T1 an i'll [give you a] cha:n[ce oin a seco]nd=
32 Mo [yeah: ]
33 [wa: it's: ]
34 T1 =[issao]
35 Mo [it's:] true but (0.3) what (0.6) you know:=
36 C3 [mm hm]
37 Mo =(.) that's:: h (1.2) >how=it=i:s (.)
38 it's not (0.3)
39 Fa no=
40 Mo =ex[actly] how i can just >say oh guys i'm=
41 T1 [mm. ]
143
42 Mo =going to hawaii for a week (.)
43 um: manage like=
At line 12, therapist 1 (T1) commences a new turn about not having heard as much from Nata- lie and Isabella, indicating them as recipients and potential next speakers. Through the set-up, T1 then acknowledges and summarizes the talk so far as being about “the effect” (of Isabella’s dis- tress). This topicalizes talk about “the effect” as the background for the following question. The beginning of the therapist’s turn thus indicates the addressee and potential next speaker, as well as delimiting the topical space for the subsequent parts of their turn. At line 23, T1 uses “and I won- der” to transition to the stance-eliciting, recipient-focused component; the question “what’s going on listening to that or... what what’s come up.” By prefacing the question with “and I wonder,” the therapist implicates their own epistemic stance, while the following question is epistemically tilted toward the recipient. This means that the therapist is both reporting on their own thoughts (or wonderings), and also eliciting the stance of the recipient. At line 25, there is a long silence of 1.3 s signaling a lack of uptake by the recipients. At line 26, C3 progresses the conversation (Stivers & Robinson, 2006) with a question that possibly com- pletes T1’s prior utterance: “what do you think about.” However, T1 does not adopt C3’s pro- posed question design and instead selects a more open and less restrictive question to the mother of “what’s come up?” T1 then addresses C3 directly stating that C3, referred to as “Issa,” will get the chance to speak in a second (lines 31 and 34). The therapist thus explicitly demonstrates their deontic authority to select the next speaker. So, although the therapist is directing the session through speaker, topic and question selection, “I wonder” combined with the non-specificity of the stance-eliciting question removes a certain degree of the directness associated with a more specific question. The mother responds first with an aligning stance about “the effect” with “it’s true” at line 35. She then responds to earlier comments about how she is managing as a criticism, as evi- denced by a justification “that’s how it is” (line 37) and that that she can’t just to go to Hawaii and leave everyone to manage on their own (lines 40– 43). The mother thus initially responds to the therapist’s topicalizing of “the effect,” but then returns to an earlier part of the conversation for further comment. The therapist’s question including “I wonder” and the non-specificity of asking about “what’s come up” provides the mother with the freedom to respond to an earlier part of the conversation that was not specifically selected by the therapist. The therapist thus exerts the deon- tic authority to select the next speaker while mitigating their authority to dictate the content of that speaker’s turn.
Before Potential Disalignment In Extract 2, we present an example of “I’m wondering” when the therapist is about to intro- duce a potentially disaligning stance position. At the beginning of the extract, the mother (Mo) is describing a previous interaction where her child’s school asked if the child (Tayla), feels she needs to stay home to care for her mother.
144
Extract 2 (S7.E11, 1:18:04) T1: therapist 1; T2: therapist 2; Mo: mother; Cl: client
1 Mo =the schoo:l arksed (.) tayla:: (0.6)
2 er (0.5) couple a weeks ago (.) .h in a
3 meeting (.) .hh (0.3) if (.) tayla feels (.)
4 like (0.8) she can't go to school because
5 she needs to care for me? (0.5) .hh (0.2) an
6 um (0.2)
7 T2 hm=
8 Cl =that's not (0.4)
9 Mo no:(h):
10 Cl [it's [not ]
11 T2 [that's [( )]
12 Mo [w- ] acs- actually [was .h]=
13 T2 [orighto]
14 Mo =(0.2) rea:lly offensive (.) to say that=
15 T2 =hm
16 Mo .h because if they know me a bit why would
17 they say that, (.) .hh (0.2) [you know]=
18 T2 [hm: ]
19 Mo =(0.2) [know] us
20 T2 [hm ]
21 (0.3)
22 T1 hmm.
23 T2 .h (.) but look i i don't kno:w i'm a
24 clearly that was offensive .h (.)
25 -> >but i'm wondering if mum's not
26 well would you wanna< sta:y by mum?
27 (0.8)
145
28 T2 is part of you: wants to be with mum
29 because she's sad?
30 (1.5)
31 Cl ye:s: (.) b[ut]
32 T2 [a ] bit
33 (1.0)
34 Cl [ye]a:h (0.7) >but when a< (.) [but i have=
35 Mo [e- ]
36 T2 [yeah
37 Cl =to go to school
38 T2 of course
39 Mo oha ha hao
40 Cl huhm
41 Mo >oha ha ha [ha hao<]
42 T2 [heh ] huh huh
43 (0.2)
44 Mo .hH (.) sorry about that [darlin]
45 T2 [ oyea:]:ho=
46 Mo =huh [huh huh
47 Cl [mm
48 T2 yea::h
At line 14, Mo puts forward her stance position that the school’s comments were “offensive.” The therapist’s response begins at line 23 by connecting with the Mo’s previous talk. The therapist begins with an account: “I don’t know I’m a” signaling potential disalignment with the Mo’s stance. But the therapist then self-repairs their turn to add a more aligned: “clearly that was offen- sive.” T2 appears to orient to the delicacy of the topic at this point and before presenting their next question, makes an attempt to first establish an aligned stance with the Mo. T2 continues at line 25 with, “but I’m wondering,” foreshadowing an upcoming contrast or transition of some kind. This contrast manifests in two ways. First, there is a change in addressee, with the therapist inviting a stance position from the client. Second, the therapist presents the question “if mum’s not well would you wanna stay by mum?” which contrasts with their previously established, although mod- ulated, alignment. That is, the therapist introduces a question that is disaligning and potentially “offensive” to the mother, and risky for the therapeutic relationship. It also places the Cl in a diffi- cult position because if she agrees then she
146
is contradicting her mother’s prior stance. “I’m wonder- ing” implicates the therapist’s ownership of the stance position conveyed via the question. However, “I’m wondering” also makes the stance defeasible and something that the therapist can disavow or distance themselves from as necessary. The disaligning nature of the question is evi- denced at line 27, where there is silence and no uptake from Cl. T2 redesigns the question to ask if a “part” of Cl wants to be with Mum. This question is more acceptable in that Cl can agree to a part of her wanting to be with Mum and not necessarily contradicting her mother’s prior stance. This redesigned question is again met with little uptake with a long silence at line 30, then a drawn out “ye:s:”. T2 responds by further underscoring that Cl’s agreement is only partial with “a bit”. Cl then qualifies her agreement with an explanation that she still needs to attend school. Cl thus works to align with both T2 (by agreeing that she would like to stay home with her Mum) while also aligning with her mother, that she still needs to go to school. Mo responds with laughing (line 41) and an apology to the client for her feeling like she needs to stay home with the mother (line 44). The mother then goes on to talk about how she feels that she needs a carer at the moment (data not shown). The therapist’s introduction and elicitation of a disaligning stance results in the mother having a new appreciation of how her own mental health problems are affecting her daugh- ter. In this example, the therapist works to first build some form of alignment with the mother before introducing a disaligning and potentially problematic stance position. Using “I’m wonder- ing” and some reframing of the stance-eliciting question, the therapist introduces and elicits a con- trasting and disaligning stance position from a different speaker. Despite the sensitivity of the topic, these disaligning stances are produced with no apparent offense displayed by the original speaker, who actually develops a new appreciation for how her daughter is feeling.
Problems With Turn Uptake Extract 3 presents an example of “I’m wondering” occurring with the introduction of a dis- aligning stance position that results in problems of alignment with the recipient. Just prior to the beginning of Extract 3, the family have been asked if the conversation so far has made anything clearer for them.
Extract 3 (S11.E9, 1:24:34) T1: therapist 1; Mo: mother; C1: child 1
1 C1 .h (.) it's hard because
2 £amanda hasn't sa(h)id [anythi(h)ng]£
3 Mo [oohuhuhuhoo ]
4 C1 ohuho
5 (0.3)
6 C1 .hh (0.2) huh huh
7 Mo but-
8 (0.2)
147
9 C1 [.h]
10 T1 [i ] guess i'm: (0.4) #okindao curious about
11 (0.7) o.ho (0.7) ohhho (0.7) mtk .hh (0.7)
12 coz two people can co:me to a session:
13 (0.6) waiting for the other operson:o (0.7)
14 ta- (0.6) do something odifferento
15 -> (0.9) >but i guess i'm wondering: (.)
16 .h (0.4) if y- (.) you've under (.) if (.)
17 where:: your thinking is (.) in terms of
18 (1.2) you:r (0.4) part in your relationship
19 with your osistero
20 (0.5)
21 C1 mtk whaddayu mea:n
22 (0.7)
23 T1 so: (.) ah- (1.1) oouhoo (.) i i se- i really
24 hea:r you want (0.5) you would love for
25 amanda to be different? (0.6) i don't know
26 if she ca:n (.) wi:ll is ready: (0.4) .h
27 -> (0.7) so: (.) i guess i'm curious about
28 (2.0) have you f:- (0.2) thought of (.)
29 >er'er ha'have you got any thoughts about
30 (2.3) you: (0.3) as (.) part of this sister:
31 (0.3) orelationship.o
32 (1.5)
33 C1 well i feel like (0.8) .hh (0.3) >i dunno
34 it's-< (0.2)
35 know her- (.) what amanda’s feeling
At lines 1 and 2, child 1 (C1) adopts the stance that it has been hard because her sister Amanda has not said anything. The therapist’s turn starting at line 10 does not vocally receipt C1’s assess- ment, suggesting some disalignment with C1’s stance. T1 begins the turn with “i guess i’m: (0.4) #okindao curious about,” and then allows a series of silences to develop, interspersed with audible
148
breaths. T1 then changes tack, altering the grammatical shape of the turn, and re-beginning with “coz” (i.e., because). This commences the set-up component, with the “two people” who “come to a session ... waiting for the other person... to do something different”. The therapist disaligns with C1’s stance that her sister not saying anything has been problematic, and instead focuses on them both expecting the other to change. T1 seems to orient to this disalignment and designs their turn as a parable that refers to C1 and her sister in the third person and thus, to some extent, lessens the focus on them personally. At line 15, the therapist pauses with 0.9 seconds of silence and transitions to a contrasting question beginning “but I guess I’m wondering”. This signals a change of sorts, through “but,” and a mitigation of the therapist’s stance with “I guess I’m wondering”. This recipient-focused section, however, does not go smoothly, with T1 repeatedly repairing their emerging talk. T1’s repairs center around the word “if” which projects a question with a yes or no answer (line 16), and settles on a question containing “where your thinking is” that projects a more open range of responses. T1 seems to again orient to the delicacy of their question and the implied disaligning stance and works to present their question in a carefully worded way. C1’s reply involves a pause and the repair initiator “whaddayu mean” that marks some trouble with the therapist’s turn (line 21). This may be due to the therapist’s disaligning stance and possible rebuke, or due to the ambiguity of the therapist’s question. T1’s response effectively repeats their prior turn but in a more direct fashion. T1 first orients to problems in alignment by acknowledging C1’s stance (lines 23-25). T1 then makes a more explicit stance statement that they do not know if Amanda can, is willing, or ready to change (lines 25 and 26). The second part of T1’s turn again contains a number of silences and repairs as they work to design their turn in a particular way. T1 first marks the transition to the question part of their turn with “I guess I’m curious about”. This is followed by a long silence of 2 seconds at line 28. T1 then establishes their desired area of C1’s response, particularly her thoughts. This is again followed by another long silence of 2.3 seconds at line 30, before T1 establishes what they want C1 to comment upon, that is, C1 as part of her sister relationship (line 30). The complexity and ambiguity of T1’s question from lines 28 to 31 allows for a range of possible responses from C1. There is again some difficultly in C1’s response beginning with a long silence at line 32 and a slow uptake at line 33 with “well I feel like... I dunno.” C1 then returns to her original stance position about the difficulties that arise when she does not know what her sister is thinking. T1 and C1 both seem to maintain their disaligning stance positions. Similar to Extract 1, the therapist exercises their deontic authority to select the next speaker. However, the therapist also mitigates their authority to determine the content of the recipient’s answer. This occurs through multiple efforts to design their turn in particular way, orienting to the delicacy of the topic, and to place fewer restrictions on the possible responses by the recipient. This example also demonstrates that this form of stance elicitation does not always progress smoothly. The therapist introduces a disaligning stance position that combined with an open form of ques- tion results in difficulties with uptake from the recipient. The therapist consequently has to adjust their turn in response to the evolving talk with the family. To further support our analysis, we present a deviant case. This involves identifying an exam- ple where the practice under investigation, “I’m wondering” as part of an ongoing utterance, is present in an interactional moment that does not appear to fit the proposed normative expecta- tions for its use. By examining how participants deal with these departures from the expected use, we can provide a complementary source of evidence for its typical functions. In Extract 4, the ther- apist’s use of “I wonder” occurs in a similar sequential position but with a different type of action, and thus receives a different response to the prior examples.
Extract 4 149
(S11.E14, 1:39:00) T1: therapist 1; T2: therapist 2; C1: child 1; C2: child 2
1 T2 it's kinda like a wa:ve isn't ¯it
2 C1 mm hm
3 (0.5)
4 C1 literally
5 (0.2)
6 C1 [mm hm]
7 T2 [.hhh ] (0.3) an:d um (0.6) you know (0.4)
8 >g- ge-it gets up there it gets a bit<
9 hea:te:d and then you come dow::n and then
10 (0.2) there's=a (0.7) you know you ya
11 connected and a bit of laughi::ng an:d .hh
12 (.) an then it kinda goes up again:
13 (0.2) particularly around .hh (.) those
14 triggers arou:nd (.) my house your [house]
15 C1 [mm hm]
16 C2 huh::
17 T2 you know:: (.) um (1.0)
18 mtk (.) an:d (0.3) uh (.)
19 -> but yeah i just wonder about
20 that (.) r- riding the wa:ve (0.2) when
21 you're together
22 (0.2)
23 T1 mm
24 C1 mtk the (0.2) literally that's what our
25 relationship is like (.) .h like a wave
In Extract 4, the therapist takes a stance position in the form of an image for the clients’ rela- tionship at line 1, that is, that it is like a wave. This turn ends with a tag-question seeking the agree- ment of C1, which is quickly forthcoming with “mm hm” at line 2. This is followed by a pause before further confirmation by C1 with “literally” at line 4, and, after a silence, another “mm
150
hm” at line 6. Despite this agreement, the therapist expands on the wave metaphor, by providing an explanation for how it fits the clients’ relationship (lines 7–14). This may be because the interpreta- tion of the image had not been explicitly voiced and/or that there was no uptake from the other family members. The therapist’s description is followed by agreement from C1 in overlap (line 15), and a laugh by C2 (line 16). T2 continues their turn at line 17 with a number of silences and restarts before ending their turn with “but yeah... I just wonder about that... riding that wave when you’re together.” C1 responds soon afterwards, restating her previous agreement that “literally” that is what the relationship with her sister is like. This example differs from those above as the therapist and C1 has already established an align- ment of stance over the wave image before the therapist’s utterance containing “I’m wondering.” The first part of the therapist’s turn also includes a description and explanation of the therapist’s own wave image rather than parts of the family’s prior talk as in the previous examples. Further- more, the second part of the therapist’s turn does not implicate an alternative stance position, it does not select a different speaker for an alternative stance position, nor does it end with an inter- rogatively formatted component. Consequently, C1 does not put forward a novel stance. This deviant case thus illustrates that the therapist can utilize a similar turn structure incorporating a backwards looking set-up, an “I’m wondering” transition, and a recipient focused ending. But, in this example, the therapist does not use “I’m wondering” to introduce or elicit a new stance posi- tion, as it has already been previously accepted. This results in a confirmation of the therapist’s presented stance rather than other alternative stance positions.
DISCUSSION
The therapist’s use of “I’m wondering” within an ongoing utterance functions to implicate the therapist’s own stance position while also making it defeasible and deferential to the epistemic authority of the recipient. It is also used to transition to some new matter within the recipient’s domain of knowledge or responsibility. This new material includes topic variation or selecting a new speaker and effectively elicits further stance positions from the recipient. “I’m wondering” and the design of the question component of the therapist’s turn allows for a range of possible responses from the recipient. The therapist thus exerts some control over the speaker and topic selection while mitigating their deontic authority on how a recipient may respond. In general conversational terms, dialogue can be described as how prior talk is incorporated and selectively reproduced by a following speaker (Du Bois, 2014). Dialogical conversations are characterized by a parallelism between utterances. This can include repeating certain words or phrases or using a similar sentence structure. Stance acts can therefore be considered dialogic in the sense that they derive from and engage with previous turns at talk (Du Bois, 2007, 2014). As mentioned previously, a stance act involves multiple stance takers voicing their stance position toward a common stance object (Du Bois, 2007). In the Open Dialogue interactions explored in the present study, the therapist acted as an intermediary or an elicitor of stance. The therapist does this by thematizing a particular prior stance utterance for development, and selecting a recipient, who then becomes accountable for responding to that stance. However, this requires the therapist to exercise their deontic authority to manage the participation of speakers, and to determine the topical focus of their responses. If these findings are interpreted through the theoretical lens of Open Dialogue principles, then these findings could be seen as one way that the dialogical concept of polyphony may be actualized, that is, through the therapist eliciting and rendering stances, and asking other members of the meeting to comment on those stances. If so, then some form of struc- ture or direction is necessary in order for polyphony and dialogue to
151
emerge. Dialogue requires the participation of multiple speakers across sequences of turns-at-talk, and can thus be considered as an act of distributed agency (Enfield, 2013, 2017). Agency, the ability to create action, is made up of a number of elements including controlling (or determining that a behavior will occur), composing (the selection and execution of a behavior), and the anticipation of how that behavior will be responded to. When applied to the interactional moments explored in this study, it is clear that the therapist exerts some control over others’ behavior by selecting the speaker and the stance position for comment, while the client or family member, within these con- straints, composes a responsive action. The activity of proposing, eliciting, and responding to stances occurs sequentially across a number of turns, with different speakers and in this way is an activity jointly distributed and co-constructed by therapists and families. Another aspect of agency is accountability (Enfield, 2013, 2017). This is because stance-taking involves a public expression and, consequently, ownership and accountability for adopting a particular value position (Du Bois, 2007). Stance acts, like actions generally, make the stance-taker subject to evaluation, and potential sanction from others if their actions do not align with various expectations (Enfield, 2013). As discussed, there is a distribution of agency in the enactment of a stance act shared by the therapist and the stance-taker. However, accountability is not equally distributed. For example, if the therapist elicits a disaligning stance from a client, it is the client who is held responsible and accountable for articulating that stance, even though it was the therapist who was responsible for instigating it. Therapists thus have the deontic authority to elicit stances from clients; however, they have limited accountability for any sanction that may result from that same stance act. Thera- pists seem to be implicitly aware of the possible accountability associated with their stance eliciting questions. This is reflected in their use of “I’m wondering,” which makes their stance position defeasible. Open Dialogue and dialogical practices have generally under-theorized the role of power in the therapeutic relationship despite some findings that power is regularly oriented to by therapists (Guilfoyle, 2003). It seems that the accountability associated with the elicitation of stance is something that needs to be considered in Open Dialogue. A limitation of this study relates to the fidelity to the Open Dialogue approach. While the ther- apists in this study have a number of years of experience working in mental health and Open Dia- logue, at the time of the study they were still undertaking advanced training in Open Dialogue. This means that they may still have been developing their skills in the approach, and may conduct their sessions differently after further training and experience. While there are some indicators about what elements signify fidelity to Open Dialogue (Olson et al., 2014), there is not currently an accepted way of defining when therapeutic forms of dialogue occur in conversations. Despite this limitation, this study does highlight a particular conversational practice that has relevance for the Open Dialogue approach and, as such, makes a potentially useful contribution to understanding how Open Dialogue may be conducted. We wish to stress that these practices do not provide a “how-to” of Open Dialogue. In our examples of problems in turn uptake, in extracts 3 and 4, we demonstrate that these interactional forms are not a “strategy” or “technique” that can be separated from the local interactional con- text. Rather, we demonstrate how these forms have been utilized but also adapted and modified through the therapists’ responsiveness to the sequential unfolding of the interaction. This study also examined “I’m wondering” in a narrowly defined turn position (i.e., in the middle of turns). As mentioned earlier, a similar study by Ong et al. (in press) examined “I’m won- dering” at the beginnings of turns and specifically when therapists made proposals to transition to a reflection. While in both of these studies “I’m wondering” serves a similar function of downgrad- ing the therapist’s deontic authority, they also demonstrate the different actions that
152
this practice can serve. Future research could therefore explore “I’m wondering” further in similar and other turn positions to provide a more complete account of this practice. Another direction of future research is how common this practice is to other approaches to psychotherapy. The use of “I’m wondering” and the downgrading of therapists’ deontic authority may be not be unique to Open Dialogue but rather common to other approaches that emphasize a collaborative orientation. Future research could examine whether downgraded deontic authority is a point of difference between collaborative and more directive approaches, or whether these downgraded positions serve other functions such as maintaining therapeutic engagement or achieving a mandate for furthering the therapist’s agenda. Writings on Open Dialogue have tended to describe the principles and elements of the approach in general conceptual terms, such as dialogue and collaboration. But how these conver- sations actually occur in real-life settings has received limited empirical investigation. This study demonstrates that an analysis of the micro-details in therapy conversations can provide insight into dialogical practices that cannot be recovered from the theoretical principles of Open Dialogue alone. CA thus provides a means to study dialogical and family therapy conversations in a way that is systematic and empirical. Further analysis of conversations in Open Dialogue holds poten- tial for revealing further actions and practices utilized by therapists that are not yet described, which can offer a greater range of techniques to therapists employing Open Dialogue.
153
Chapter Summary
Chapter 4 analysed how therapists attend to and manage issues of deontic authority when proposing reflections, with therapists having high deontic status and presenting a downgraded deontic stance. Chapter 5 builds on that work, by demonstrating how therapists exercise authority by selecting particular parts of the clients’ talk for further discussion and selecting the next speaker to comment on the stance positions that they have selected. In addition, therapists defer to the epistemic authority of the family. They do this through designing their turns in such a way that stance positions are presented as defeasible and seek the response of the recipient, and by not nominating particular types of responses. This reduces the recipients’ responsive obligations, which promotes responsive flexibility and agency over the control and composition of their turns. Therapists thus exercise authority over topic and speaker selection, while mitigating their authority over how a client may respond, and deferring to the recipient’s authoritative epistemic position.
The conversational practices in this study i.e., nominating topics for development, the use of “I’m wondering”, and selecting the next speaker, make specific contributions to the stocks of interactional knowledge on how the principles of polyphony, dialogue, and a “not- knowing” position may be promoted in conversation. As demonstrated in Chapter 4, therapists work to promote flexibility and agency in client responses, and present their own position and identity as one who exercises authority to direct the session but does not have the authority to describe the experience of the family.
This chapter also introduces the idea of dialogical syntax (Du Bois, 2014). Dialogical syntax involves the reproduction of syntactical forms of the previous speaker and is present in the initial part of therapists’ turns with a backward looking component. As dialogue involves a responsiveness to the prior talk (Olson, Seikkula, & Ziedonis, 2014), syntactical
154
reproduction may represent the conversational manifestation of dialogue. This is considered further in the next chapter where I examine therapist repeats of the talk of the prior speaker.
155
Chapter 6: Therapist Repeats
In this chapter, I present an unpublished manuscript, reporting a study examining therapist repeats in Open Dialogue sessions. In this study, I review how Open Dialogue writings recommend the use of repeats in order to promote thinking and reflection by clients.
I designed the study, extracted the data, analysed the data and wrote the drafts of the manuscript with the analytic and editorial support of my supervisors. Consequently the article includes the use of the pronoun “we”.
156
A Conversation Analysis of Therapist Repeats in Open Dialogue Network Meetings
The various approaches to family therapy each have their own professional stocks of interactional knowledge (Peräkylä & Vehviläinen, 2003). These stocks of interactional knowledge refer to the various models and theories that inform and shape clinical practice and can vary in their level of specificity. In Open Dialogue, the stocks of interactional knowledge place particular emphasis on therapists following the lead of clients in a session.
Therapists are advised to have little pre-planning of a session and to instead begin with an open question and then adapt their responses to what has been raised by the clients (Seikkula,
2011; Seikkula & Trimble, 2005). Therapists’ responses are designed to elicit further dialogue around the unique understandings of each client; to do this, therapists are advised to incorporate the language of the clients into their own responses and questions. This is said to display a “responsive listening”; listening without an agenda that allows clients to express important stories that they have not previously shared (Olson, Seikkula, & Ziedonis, 2014, p.
14). For example:
“The therapist aligns his or her words to the patient’s and family’s, and respects the
definitions and language voiced by each. The therapist can do this by using their
words and expressions. The therapist may ask a person to “say more” about a general
topic or story and may also ask more specific detail-oriented questions.”
(Haarakangas, Seikkula, Alakare, & Aaltonen, 2007, pp. 227-228).
Therapists are thus advised to “align” their words with those of the client and family.
A similar idea is expressed elsewhere as “incorporating the familiar language of the network members” (Seikkula & Trimble, 2005, p. 462). This follows other recommendations in Open
Dialogue not to impose the therapist’s “jargon, interpretation, and hasty conclusions” (Olson,
Seikkula, & Ziedonis, 2014, p. 31) or to “colonize” the family’s understandings (Rober &
Seltzer, 2010, p. 123). 157
In addition to incorporating the words of the clients, therapists are advised to actually repeat the clients’ prior talk word-for-word:
“...through careful attunement to each speaker, the leader generates each next question
from the previous answer (e.g., by repeating the answer word for word before asking
the question or by incorporating into the language of the next question the language of
the previous answer)” (Seikkula & Trimble, 2005, p. 462), and
“Most often, the teams’ answer takes the form of a further question, which is based on
and has taken into account what the client and family members have said. Often this
means repeating word by word some part of the utterance and encouraging further
speaking on the subject” (Seikkula, 2011, p. 182).
Therapists therefore not only align their language with clients by incorporating it into their responses, they are advised to repeat the clients words verbatim. Olson, Seikkula, and
Ziedonis (2014, p. 13) provide an example of the therapist issuing an exact repeat of the client’s prior turn:
“David: I didn’t expect a direct answer.
Therapist: Yeah. You didn’t expect a direct answer.
David: I didn’t expect an answer. I knew there was a training program in Open
Dialogue. I didn’t think it would be a direct possibility. I thought maybe, you’d say
at a future point. I was surprised. It made me think that I was on the right track.”
In this example, David is talking about making initial contact with the Open Dialogue team and didn’t expect a “direct answer” from them. The therapist’s response, with the addition of a “yeah” and changes to deixis, is an exact repeat of David’s prior talk that takes up all of the therapist’s turn. This is said to invite David to elaborate on his prior talk (Olson, Seikkula, &
Ziedonis, 2014). David’s next turn includes a repeat as well as an elaboration saying that he 158
didn’t think that it was a possibility that the team would work with him and that they would instead say that it was a possibility at a future point in time.
The next extract from Seikkula (2011, p. 190), includes an additional way that therapists can incorporate repeats of the clients’ talk:
“M: I have not been recognised.
Tl : You have not been recognised?
M: Throughout my life I've been excluded from the family. At last I want to get rid
of this symbiotic mess.
Tl : You said that ‘Throughout my life I've been excluded from the family’. Then you
said that ‘At last I want to get rid of this symbiotic mess’. It sounds like you are
saying two things at the same time?
M: ... yes ... that's what I said ... But so far I cannot say anything more about it.
Tl : ... yeah.”
This extract provides two examples of how repeats can be utilised in Open Dialogue. There are verbatim repeats that, with minor additions, take up the whole of the therapist’s turn. This was displayed in the first example and the first lines of this current example. In addition, therapist repeats can incorporate the clients’ words into an extended response. This is displayed in the therapist’s turn: “You said that ‘Throughout my life I've been excluded from the family’. Then you said that ‘At last I want to get rid of this symbiotic mess’. It sounds like you are saying two things at the same time?” The therapist repeats the words of the client but they take the form of a “quotative expression” (Knol et al., 2020) that is incorporated into a longer response.
In summary, writings on Open Dialogue practice recommend that therapists follow the lead of the clients within sessions. Therapist responses are designed to align, incorporate, and repeat aspects of the clients’ talk. These responses are thought to allow clients the 159
opportunity to hear their words repeated back to them, to reflect on them, and to encourage elaboration on issues that they may not have discussed previously. As noted in the examples presented above, these repeats can take a couple of different forms. They can include repeats that occupy the whole of the therapist’s turn or can function as quotes of the client’s talk that is incorporated into a longer statement or question. In this study, I focus specifically on therapist repeats that occupy the whole of the therapists’ turns. These repeats are a clearly delineated practice and represent a more manageable sample for analysis.
Prior CA research has shown that repeats can serve a variety of functions depending on their sequential positions. When repeats of prior talk by another party occur immediately following the first saying and with rising final intonation, they have a “questioning” character with a number of possible functions (Kim, 2002). These repeats can indicate a problem of hearing or understanding by the recipient, and project a repairing response from the speaker of the prior turn (J. D. Robinson & Kevoe-Feldman, 2010; Schegloff, 1996, 1997; Schegloff,
Jefferson, & Sacks, 1977). Second position repeats may also seek confirmation from the prior speaker, or display a speaker’s stance towards the prior utterance, such as disagreement
(Kim, 2002). Second position repeats with falling final intonation have different functions.
They can provide confirmation and register receipt of the prior turn (Puchta, Potter, & Wolff,
2004; Schegloff, 1996, 1997), display agreement (Kim, 2002), or elicit expansion and additional information from the prior speaker (P. Brown, 1998; Kim, 2002; Knol et al.,
2020). Repeats after a responsive turn (e.g., in third position) have different actions and can display a speaker’s receipt or acknowledgement of a previous response (Schegloff, 1996,
1997). Third position, receipting repeats can result in no response and close a sequence, or be treated by recipients as initiating repair resulting in responses of confirmation (Schegloff,
1997). In summary, a repeat cannot be understood by reference to its form or sequential
160
position alone, and a more accurate analysis needs to consider what actions a repeat implicates and how the repeat is responded to by the recipient.
Other research has examined the use of repeats in institutional talk. Puchta, Potter, and Wolff (2004) examined repeats as markers of receipt in market research focus groups.
They found that repeats by focus group moderators were produced with a “list-like” intonation, that register receipt, do not seek elaboration, and orient to a response from another group participant. This models and shapes the desired responses from the participants as well as making the relevant responses highly visible to those who have commissioned and are observing the focus group research. In psychotherapy, Knol et al. (2020) described therapist repeats as mirroring the client’s talk back to them. Similar to the recommendations of the
Open Dialogue approach, they argue that unlike formulations (Antaki, 2008), repeats did not transform the talk of the client, thus avoiding therapist interpretations and reducing possible bias. Instead, mirroring repeats selected a salient phrase or quoted the prior talk of the client, which nominated prior topics for elaboration and functioned as more explicit requests for expansion. These mirroring repeats facilitated further explorations while also presenting the therapist as non-directive and impartial.
This study focuses on repeats that occupy the whole turn of Open Dialogue therapists in response to the prior talk of the family. Using conversation analysis, we aim to analyse the different functions of therapist repeats and to consider how these repeats impact on sequences of interactions.
Method
Data and Ethics
Data for this study included fourteen hours of video recorded Open Dialogue sessions from a youth mental health service in Western Sydney, NSW, Australia. Participants included therapists (n=12), clients and the clients’ networks which included family members and other 161
professional staff (n=36). The majority of therapists (n=10) had a minimum of 2-years of experience in Open Dialogue including foundation training and regular supervision. Seven of these therapists were participating in advanced training as Open Dialogue therapists or trainers. The remaining two therapists had an informal orientation to Open Dialogue from their more experienced peers. Open Dialogue sessions do not have a particular form of
“intervention” per se but rather focus on the concerns presented by the family with the therapists aiming to maintain a mindset of openness to understanding the experiences of the family (Olson, Seikkula, & Ziedonis, 2014; Ong & Buus, in press). At some point in the session the therapists have a reflection where they speak to each other, voicing their thoughts and responses to the talk of the family. During the reflection, the family listen and are only invited to respond when the reflection ends (Andersen, 1987; Ong, Barnes, & Buus, 2020c).
The aim of reflections and the session generally is to create opportunities for multiple perspectives to be voiced without necessarily seeking consensus.
Potential therapist participants were approached by the first author with details about the study. Interested therapists then completed a consent to participate and identified potential clients and their networks for participation. The first author provided more detailed information to interested clients and their families. Participation in the study was voluntary and data collection only commenced after all participants provided written informed consent.
The sessions were video recorded with three cameras placed at different points around the room to capture verbal and embodied aspects of their interactions. The study was approved by the Nepean Blue Mountains Local Health District ethics committee (reference number:
HREC/17/NEPEAN/135) before data collection commenced. Identifying information in the presented extracts has been altered with the use of pseudonyms to ensure anonymity.
Following family therapy conventions participants are named according to their family role
162
e.g., Grandmother, Father etc., and children are listed in order of age e.g., Child1, Child 2...
In the analysis all non-therapist participants are referred to as “clients”.
Analytic Process
All video recorded data was transcribed using conversation analysis conventions for talk (Hepburn & Bolden, 2017). We then focused on identifying all occurrences of therapist repeats in response to the talk of the clients. We focussed on examples where the repeats occupied the full turn of the therapist (with minor additions). This was to capture the types of repeats discussed in Open Dialogue writings. We included repeats that occurred within the next few turns following the first saying. We excluded self-repeats by the therapists as they were not repeats of the talk of the clients. We focused on full and partial lexical repeats of prior turns and included repeats with changes to reference, deixis, intonation or prosody.
While we began with a focus on a pre-defined conversational practice of repeats, we did not approach the data with a pre-determined view of the function of these repeats.
We identified 160 instances of therapists repeating the words of the client or their social network. Each example was analysed in relation to the sequential position of the repeat and the response it received. These examples were compared and grouped according to the similar functions they served. Below we present typical examples of the different types of repeats identified in the data; all extracts were analysed by all the authors and the presented analysis reflects our combined analysis of the data.
Results
We found that therapist repeats served a variety of interactional functions. We first present repeats that initiated repair and repeats that functioned as receipts. We then present repeats that nominated topics for elaboration, and explore how the equivocality of some repeats provide for a number of different responses. These types of repeats were congruent wit h Open Dialogue recommendations to repeat the words of the family in order to encourage 163
elaboration. Finally, we present repeats that occurred outside of transition relevance places, which did not disrupt the turn in progress. Consequently, these types of repeats functioned to display receipt. In the extracts that follow, the therapists’ repeats are marked with bold italics and the first saying is marked in bold.
Repeats as Other-Initiations of Repair and Third Position Receipts
In CA, repair refers to a set of practices for interrupting the progression of a conversation in order to deal with problems in speaking, hearing, or understanding (Hayashi,
Raymond, & Sidnell, 2013; Schegloff, Jefferson, & Sacks, 1977). The phenomenon of repair includes three main parts: the trouble source (the part of speech that is later repaired), the initiation of repair (the signalling of a problem) and the repair solution. We found that one function of therapist repeats is the other-initiation of repair, or repair initiated by a party other than the speaker of the trouble source. (Hayashi, Raymond, & Sidnell, 2013; Schegloff,
Jefferson, & Sacks, 1977). As we focus on therapist repeats, these examples involve the production of a trouble source turn by a family member, the initiation of repair by the therapist (through a repeat), and the provision of a repair solution (either by the initial speaker of the trouble source or by another family member). In Extract 1, the therapist uses a repeat that results in a repair solution provided by the prior speaker, the Grandmother.
Extract 1
S6.rpt3
GMO: grandmother; CH1: child 1 (Gregory); CH4: child 4; TH1: therapist 1; TH2: therapist
2
1 GMO: >so would you< (.) [guys: (.) like (.)= 2 CH1: [( ) 3 GMO: =sam (.) gregory to organise another night? 4 (0.2) 5 CH4: .h yep 6 GMO: yeah¿ 7 → TH1: another:_ ((gaze to GMO)) 164
8 (0.3) 9 GMO: night 10 TH1: >another [night]< (.) yeah 11 TH2: [mm? ]
At line 1, the Grandmother asks the children if they would like their brother Gregory
(who lives with a different relative) to visit them an extra night per week. At line 3,
Grandmother initially says “Sam” (who is a different brother) and self-repairs her turn to
“Gregory”. This is answered by Child 4 with “yep” (line 5) followed by confirmation from
Grandmother with “yeah”. At line 7, Therapist 1 initiates repair with a repeat of “another” mirroring Grandmother’s turn at line 3. This repeat includes a sound stretch on the last syllable of “another” with flat intonation and gaze to Grandmother thereby selecting her as the recipient. This repeat displays that Therapist 1 adequately received the first part of the turn and frames the trouble source of the repeat as occurring after “another” (Walker &
Benjamin, 2017). Grandmother responds by completing the repeat of her original turn, began by Therapist 1, with “night”. Grandmother orients to Therapist 1’s repeat as initiating a repair of hearing by providing a repair solution of “night”. The Grandmother did not orient to the repeat as initiating a repair of understanding otherwise she would provide an explanation of the meaning or an account of her prior talk. The repair solution is receipted by Therapist 1 with a repeat receipt of “another night” (discussed further below) and a confirming “yeah”.
The initiation of repair does not occur immediately after the trouble source as there is an intervening sequence of answer and receipt between the Grandmother and Child 4 where the therapist is not directly addressed.
As demonstrated in this example, therapist repeats that initiate a repair of hearing are followed my minimal responses by clients who only provide the part of their turn that was identified as unheard. Consequently, they function to maintain intersubjectivity, as the
165
therapist and clients maintain shared understanding (Albert & de Ruiter, 2018), however they do not elicit expansion or elaboration in the way proposed by Open Dialogue writings.
In Extract 2, the therapist’s repeat again other-initiates repair but, on this occasion, the repair indicates a problem of understanding rather than hearing. In these examples, there is some form of elaboration by the client in contrast to the minimal responses by clients in repairs of hearing. At the beginning of Extract 2, the Mother has described a fight between the Client and her sister (Samantha) that resulted in the client attending court.
Extract 2
S5.rpt1
MOT: mother; CLT: client; TH1: therapist 1; TH2: therapist 2
1 MOT: mtk .h (.) so we've been to court an: um:: 2 (0.4) the ay vee oh was gra:nted an: (.) 3 now she's not allowed to go: anywhere 4 near dad's place::? 5 (1.3) 6 MOT: [or her sister or grandmother,] 7 CLT: [or my local mac (.) donald]s¿ 8 (1.0) ((T1 & T2 gaze to Cl)) 9 TH1: mtk .h 10 (0.3) 11 CLT: hhuh 12 (0.4) 13 → TH2: or your local mac[donald’s,] 14 MOT: [.h ] samantha works 15 at the local mcdonalds 16 TH2: ooohoo 17 TH1: ri:ght 18 (2.3)
At lines 1-4, Mother describes going to court and the outcome being an AVO
(Apprehended Violence Order), stating that the Client is not allowed near her Dad’s place
(where the incident occurred). At line 6, Mother adds that another condition of the AVO is that the Client is not allowed to be near her sister or grandmother. In overlap, the Client adds that she is also not allowed to go near the “local McDonald’s” (line 7). This is followed by a
166
silence at line 8 with Therapist 2 directing gaze towards the client. There is no forthcoming clarification from the Client and at line 13, Therapist 2 repeats “or your local McDonald’s” with slightly rising intonation and gaze to the client. This is a full repeat of the Client’s turn that rules out the therapist having a problem of hearing and instead indicates a problem understanding the whole of the turn (J. D. Robinson & Kevoe-Feldman, 2010). At line 14,
Mother orients to this repeat as a repair initiation that requests an explanation by answering that the Client’s sister Samantha works at the local McDonald’s (as mentioned at line 6, the client is not allowed to be near her sister). The therapist was not previously aware that
Samantha works at the local McDonald’s and this epistemic asymmetry between the family and the therapist (as well as the full repeat ruling out a problem of hearing) is relevant in the
Mother ascribing the therapist’s repeat as an initiation of a repair of understanding (J. D.
Robinson, 2013a). At line 16, Therapist 2 responds with a change of state token “oh”, indicating that the Mother’s answer was new information (Heritage, 1984) and that the repeat at line 13 was indeed seeking a repair of understanding. The repeat of the client’s words and the epistemic asymmetry is sufficient to successfully initiate a sequence of repair, although the trouble source was uttered by the Client, while the repair solution was provided by the
Mother. In contrast to the minimal response from the client in Extract 1, this form of repair elicits elaboration and new information that the therapist was previously unaware of. While the mother’s response was new information for the therapist, it was factual information that did not elicit further discussion or dialogue.
In Extract 3, the therapist uses a partial repeat that functions to indicate a problem of understanding and yields additional information from the client. This extract also demonstrates how the functions of other-initiated repair to correct problems of hearing and understanding can also elicit elaboration and cross over with other conversational actions.
Just prior to Extract 3, Therapist 1 has asked the Client about not liking the rules during a 167
recent hospital admission (data not shown). At lines 1-2, the Client speaks about not necessarily having a problem with the rules in the hospital but how the staff acted and how they treated him.
Extract 3
S10.rpt5
CLT: client; MOT: mother; TH1: therapist 1
1 CLT: >it=wasn't so much< (.) ru:les it was just ha- (.) 2 the way they a:cted an the way they treated you_ 3 (0.5) 4 MOT: oan justo spoke to you [( )] 5 CLT: [yea:]:h 6 (0.6) 7 → TH1: the way they a:cted, 8 (0.5) 9 CLT: like they jus=treat you like you're 10 younger than you were, 11 (0.3) 12 TH1: orighto
At line 7, Therapist 1 repeats “the way they acted” (from Client’s talk at line 2) with slightly rising intonation, indicating that Therapist 1 correctly heard the Client while also making relevant a response from the recipient (Stivers & Rossano, 2010). Similar to Extract
2, there is a marked epistemic asymmetry with Therapist 1 enquiring about something they are not informed about. After silence at line 8, the Client provides more information, i.e., the staff treated the residents as younger than they were. This is receipted by Therapist 1 with
“right” (line 12). The therapist’s repeat is thus treated as indicating a problem of understanding, and elicits an extended response from the Client rather than, for example, confirmation. In contrast to the previous extract, the Client responds with information that was not only factual in nature but included something that was potentially emotionally-laden for the Client, which is aligned with recommendations on the use of repeats in Open
168
Dialogue. This extract is also similar to Extracts 5-8 below, where the therapist repeat selects and topicalises a part of the Client’s talk and elicits elaboration. We have presented this repeat as a repair of understanding as the repeat is responded to with a clarifying response and then confirmation from the therapist, treating the Client’s response as adequate. Repeats therefore do not necessarily meet the proposed categorical descriptions due to ambiguity in the function of the initiating turn and the response. Despite these ambiguities, these extracts demonstrate how the therapist’s use of repeats to initiate repair can elicit minimal responses providing factual information to responses that are potentially emotionally-laden, which could lead to further discussion and dialogue.
In summary, therapist repeats can function as other-initiations of repair, and target a variety of trouble source types. The repeats in this study took different forms, including a sound stretch and partial repeat (Extract 1 and 3) and a full (but modified) repeat with slightly rising intonation (Extract 2). In these cases, the repeat quickly elicited a repair solution from the speaker of the trouble source or another party with epistemic authority over the matter.
Therapists also use repeats to display their acknowledgement and receipt of the family’s talk and to close sequences of talk (Schegloff, 2007). These examples usually occur in a sequence closing third position in a sequence consisting of a therapist question, family member response, and therapist repeat. In Extract 1 (above), Therapist 1’s first repeat of
“another” (line 7) initiates a repair sequence on a problem of hearing. After the repair solution is provided by the Grandmother at line 9, Therapist 1 repeats the whole turn “another night” (line 10) displaying their receipt of the repair solution.
In Extract 4, the Therapist asks the Client if she is in year nine of school (line 2). The
Client answers with a disconfirming response, indicating that she is in year eight. This is receipted by Therapist with a repeat of “year eight”. There is then a silence before Therapist asks about a new topic (data not shown). 169
Extract 4
S11.rpt1
THP: therapist; CLT: client
1 (0.2) 2 THP: >you're in yea::r< ni:ne n[ow? ] 3 CLT: [year] eight 4 → THP: year=eight, 5 (1.3)
This example demonstrates the use of the therapist’s repeat in third position to signal receipt or acknowledgement of the family member’s answer. These repeats display the therapists’ adequate hearing of the family members’ talk and do not lead to repair sequences, even though the repeat ends in slightly rising intonation (like in Extract 2). This may be explained by the sequential position of the repeat and the associated symmetrical distribution of epistemic knowledge. The therapist’s original question implies an epistemic asymmetry with the recipient having a stronger epistemic position relative to the therapist. After the recipient provides an answer there is a more symmetrical distribution of knowledge. The therapist repeat that follows occurs in a sequential position after they have been “informed” displaying their receipt of this new information. Heritage (1984) has demonstrated how recipients of new information mark their internal change of state through the use of “oh”.
Through the use of repeats to initiate repair and to display receipt, therapists demonstrate that they are closely attending to the talk of clients and incorporating that talk into their responses. The initiation of repair also maintains intersubjectivity or shared understanding that is part of a dialogical approach. But the use of repeats to initiate repair and to display receipt are utilised in everyday conversation and thus do not form a conversational practice that is distinctly institutional nor unique to Open Dialogue (even though they may be employed to accomplish situated institutional objectives). They also do not fulfil the
170
intentions of allowing the client to hear their words repeated back to them in order to encourage reflection and elaboration. In the next section, we describe the use of repeats to elicit elaborated responses from the family. These again are not distinctly institutional practices however they do align with how repeats are generally described in Open Dialogue writings.
Repeats Nominating Topics and Eliciting Elaboration
Therapist repeats could function to nominate topics for further elaboration. In these environments, the therapist’s repeat followed some form of assertion or telling by a family member, nominated a topic for further development, and initiated a sequence involving further elaboration from the prior speaker. In Extract 5, the therapist’s repeat elicits elaborating responses from multiple members of the family. At the beginning of the extract,
Child 4 has been describing differences in her sister (the client) since her sister has become mentally distressed.
Extract 5
S4.rpt12
CH1: child 1; CH2: child 2; CH4: child 4; MOT: mother; TH1: therapist 1
1 CH4: i dunno she=used to laugh more 2 (0.5) 3 CH4: [to=me] 4 CH1: [yeah ] 5 → TH1: ah she used to laugh more, 6 MOT: [mm:_ ] 7 CH1: [er her] [hugs (.) are different as ¯we:ll, ] 8 CH4: [she laughs (.) sti:ll too an (.) b]ut 9 (0.3) it's [just ] 10 CH1: [>it's really] wei:rd< 11 (0.4) 12 CH4: [it's weird laugh] 13 CH2: [it's different ]
171
At line 1, Child 4 states that her sister (who is identified as the client) used to laugh more. After a silence, Child 4 adds an increment “to me”, emphasising that her prior talk was her own observation. This occurs in overlap with her older sister (CH1; who is not the client) who agrees with “yeah”. At line 5, Therapist 1 begins with “ah” displaying receipt of new information (Heritage, 1984), and continues with a repeat of Child 4’s “she used to laugh more”. This turn has a newsmarking form of [“oh” + partial repeat] that displays receipt of new information, and also nominates the topic of the client laughing for further discussion
(Maynard, 1997). This is met with responses from three people in rapid succession. Mother responds with stretched “mm” and flat intonation, but she does not say anything further. In overlap (line 7), Child 1 adds that the client’s hugs are different as well. Child 4 responds in overlap with Child 1 adding that her sister still laughs (line 8). This clarifies her prior stance by suggesting that she used to laugh more in the past. After some silences in Child 4’s turn
(lines 8-9), Child 1 adds in overlap that the client’s laugh is “really weird”. Child 4 agrees at line 12 with “it’s a weird laugh”.
Therapist 1’s newsmarking repeat has a projective aspect that selects and topicalises a part of Child 4’s prior talk for further elaboration. The repeat is treated by the recipients as requiring a response, with responses from multiple family members including possible confirmation from the Mother, additional information from Child 1, and a clarifying statement from Child 4. The repeat thus nominated a topic for further discussion and prompted elaboration but did not specify a particular type or format of response promoting responsive flexibility (Enfield, 2017).
In Extract 6, the therapist’s repeat similarly results in elaboration and clarification of a family member’s prior talk. Prior to the extract, the family have been discussing how one of the children (the identified client) would want to do the same activities as the other family
172
members. The extract begins with the Mother describing how they used to “manipulate” the client into doing what they all wanted to do.
Extract 6
S4.rpt13
MOT: mother; TH1: therapist 1; TH2: therapist 2; CH1: child 1; CH2: child 2
1 (0.4) 2 MOT: we=learnt ho:w to manipulate he:r into what 3 (0.5) the family was doing? (0.3) 4 an that worked (0.5) perfectly. 5 (0.6) 6 TH1: (alr[ight)] 7 TH2: [ mm]: ] 8 MOT: [we w]ant]ed her to go somewhere .h (.) 9 or didn't want her to go somewhere 10 we'd put it in a wa:y (0.3) 11 that was the [best] for the family= 12 TH1: [yeah] 13 MOT: =and best for her .h (.) 14 so she thought she was getting the best (.) 15 deal where we'd made the deal (.) 16 .h that was (0.5) our wishes 17 (0.3) 18 CH2: huhuhuh .h (0.2) >it [sounds really] bad (.)= 19 MOT: [you know ] 20 CH2: =but it was like a a'huhuhuh 21 CH1: huhuh .h h 22 (0.4) 23 MOT: [yeah well that's] 24 → TH1: [it sounds really] ba:d, 25 (0.2) 26 CH2: .hh 27 MOT: well= 28 CH2: =YEAH like it's like man:ipulate 29 >like manipulate is such< a: (.) 30 you know >but i get what you mean (0.3) ah 31 (0.2)
From lines 2 to 16 the Mother describes how the family “manipulated” the client by presenting what they wanted to do in such a way that the client would chose the option that they wanted her to. The client would consequently feel that she got the best deal when the
173
family had actually controlled how the options were presented. At line 18, Child 2 begins a turn with laughter and saying that “it sounds really bad but it was like”. Child 2 orients to a moral accountability for the family’s “manipulation” of one of the children. This turn is left unfinished and ends with more laughter. At line 23, the Mother begins a turn that projects a contrasting position to Child 2 that justifies her description or actions with “yeah well that’s”.
However, the Mother’s turn occurs in overlap with Therapist 1 and the Mother abandons her turn. At line 24, Therapist 1 repeats “it sounds really bad”, nominating the prior talk of Child
2 as a topic for further discussion. After a short silence, the beginning of Child 2’s response is not grammatically completed as she suggests, but does not explicitly state, that “manipulate” is a word that sounds bad. Instead, she turns towards her Mother and ends the turn with “but I get what you mean” suggesting an attempt to affiliate with the Mother’s position. The repeat by Therapist 1 thus nominates a particular part of the prior conversation for further discussion and is treated as requiring some clarification of the repeated talk. The client also reflects on the implications of her response by not fully completing her turn and attempting to align with her Mother. The repeat makes relevant elaboration and in doing so makes the recipient’s response public and accountable.
In Extract 7, the therapist’s repeat occurs in third position with downward intonation, which is similar to the sequential position and intonation of third-position repeats displaying receipts. However, after a silence the prior speaker provides further details thus treating the repeat as a relevant environment for elaboration. Just prior to this extract, Child 2 has mentioned that he has a big thing to say, but at line 1, he declines to elaborate with “never mind”.
Extract 7
S9.rpt4
CH2: child 2; SFA: stepfather; TH1: therapist 1 174
1 CH2: onever mindo 2 (0.5) 3 SFA: >no what's your big< thing 4 you said there was a big thi:ng¿ 5 CH2: [( )] 6 TH1: [er is] (0.2) m’¯yea:h (0.2) ah- (0.2) 7 is there a big thing? 8 CH2: communication. 9 (0.4) 10 → TH1: comm[uni ]cation. 11 SFA: [o(right)o] 12 (0.8) ((SFA and TH1 maintain gaze to CH2)) 13 CH2: ( ) (0.3) we're not ta::lking¿ (0.4) 14 TH1: oyea:ho 15 (0.2) 16 CH2: a::n: (0.3) it (0.2) >doesn't< (0.3) 17 go well (0.4) [like] 18 TH1: [ yea]:h 19 CH2: everything goes ta shit 20 (0.7)
At line 3, the Stepfather pursues a response from Child 2. At lines 5 and 6, there is overlap between Child 2 and Therapist 1. Child 2’s turn is inaudible and he abandons the turn. Therapist 1 hesitates briefly after the overlap and re-authors the Stepfather’s prior question to “is there a big thing?” Similar to issues of epistemic primacy (Heritage &
Raymond, 2005; Thompson, Fox, & Couper-Kuhlen, 2015), this re-authoring implicates the therapist’s independent arrival at the question and deontic primacy and authority to direct the session. At line 8, Child 2 responds with “communication”. There is a silence before the therapist repeats “communication” with falling intonation at line 10. The Stepfather also provides a receipt of “right” in overlap with the talk of Therapist 1. Therapist 1’s repeat of
“communication” has a similar placement and intonation as third-position repeat receipts (as in Extract 4). In this example, Child 2 orients to the repeat as a relevant environment for elaboration by responding with further details that when the family are not talking things do not go well (lines 16-19).
175
As displayed earlier, repeats in third position can display receipt of the prior information and close down further discussion on the topic (Schegloff, 2007). These can include news receipts (Maynard, 1997) or acknowledgement tokens (Gardner, 2001) that acknowledge prior talk but do not project elaboration. Other response tokens can close down sequences of talk such as “oh” or assessments (Schegloff, 2007), or change of activity tokens, such as “okay” and “alright”, which display varying degrees of readiness to transition to a new topic or activity (Gardner, 2001). However, as demonstrated in Extract 7, and argued by
Schegloff (2007), third position repeats are potentially equivocal, and may close or extend a sequence of talk. The next example demonstrates how the equivocality of repeats may not lead to the expected response, with the therapist then following up with another statement that elicits elaboration.
In Extract 8, the client responds to the therapist’s repeat with confirmation. The therapist does not receipt the Client’s confirmation and instead issues a revised assertion.
Prior to Extract 8, the participants are talking about the Client’s worries about eating and the client’s Father has said that the Client has had similar worries in the past (data not shown).
Extract 8
S1.R1
CLT: client; FAT: father; THP: therapist
1 CLT: befo:re >i didn’t feel like 2 i was gonna [die,]< (.) .hh 3 FAT: [mm ] 4 (0.2) 5 THP: mm= 6 CLT: =an ah: >i would ea:t and i would< be: 7 (0.4) i would be alright 8 (0.6) + (0.8) + + ((Th nods)) + 9 → THP: you’d get (0.2) you’d be alri:ght, 10 CLT: Yeah 11 (0.2) 12 THP: you’d recover
176
13 (0.3) 14 THP: Like 15 CLT: no i’d jus- (.) i’d eat (.) 16 an then i wouldn’t worry about 17 (0.6) what i’m eating or something 18 an i’d feel like i’m eating correctly 19 (0.5) 20 THP: ommo
At line 1, the Client disagrees with his Father by stating that although he has previously been concerned about his eating, in the past he did not feel like he was going to die when not eating correctly. There is no response at line 4 and a minimal receipt by the
Therapist at line 5. The Client then self-selects to speak again and adds further details about what his experience was like previously around eating, i.e., he would eat and then he would be alright (line 6-7). There is again a long silence of 1.4 seconds (line 8), which includes nods from the Therapist. At line 9, the Therapist begins a turn with “you’d get” that projects a summary or formulation of the Client’s prior turn. Instead, the Therapist self-repairs with
“you’d be alright”, a repeat of the Client’s words with slightly rising intonation. Combined with the evident epistemic asymmetry, this repeat is designed to seek a response from the
Client. At line 10, the Client treats the Therapist’s repeat as seeking confirmation, responding with “yeah” and no further elaboration. After silence at line 11, the Therapist adds, “you’d recover”, which conveys a different description of the client as being “alright”. Initially, there is no uptake by the Client and the Therapist starts another turn with “like” at line 14 projecting self-repair. The Therapist is interrupted by the Client at line 15, rejecting the
Therapist’s interpretation “you’d recover” by saying that he would eat and then wouldn’t worry about what he was eating. In Extract 7, the Therapist’s repeat receives a “yeah” from the Client, suggesting that the client analysed the Therapist’s repeat as seeking confirmation.
The Therapist, however, does not receipt the Client’s confirmation and instead responds by revising their turn, thereby treating the Client’s confirmation as inadequate. Here, then, we 177
can see how the indeterminacy of some repeats may result in multiple attempts from the therapist to solicit elaborative talk from clients and others.
In summary, in addition to initiating repair or marking receipt, repeats were used to create sequential environments that encourage elaborative responses from clients without employing an action or practice that implements stronger constraints on responsive talk (e.g., an interrogative turn). This affords the recipient a wider variety of possible responses within a topical field determined by the therapist. These uses of repeats by therapists is closely aligned with the Open Dialogue literature where therapists are advised to incorporate and repeat the words of clients in order to encourage elaboration. However, it is not clear if these repeats also lead clients to reflect on their prior talk or only encourage elaboration within the conversation.
Repeats as Receipts Outside of Transition Relevance Places
178
In the following extracts, we present instances of therapist repeats outside of transition relevance places, and in overlap with the talk of a family member. These repeats did not disrupt the progression of the talk and did not elicit a response and occurred in two different environments. In one environment, the therapist repeats a word in overlap with the current speaker. In the other environment, the therapist repeats a word from the turn of the prior speaker, in overlap with a new speaker. Extract 9 is an example of the therapist’s repeat occurring in overlap with the talk of the current speaker outside a transition relevance place.
This is followed by a silence with no response, and then the Mother self-selects with a response to the Client’s prior turn. Just prior to the extract, the conversation has been about whether the Client has found her individual therapy sessions helpful. The Mother has proposed that the Client finds it helpful speaking to her therapist and making “plans”. From line 1, the Client responds that nothing seems to come out of the plans.
Extract 9
S5.rpt9
CLT: client; THP: therapist 2; MOT: mother
1 CLT: mtk .hh (1.3) but like- (0.2) the thing is (.) 2 like the action plan never like (.) 3 comes into place or: (.) anythin:g (0.4) 4 just- (0.2) gets said (.) an then (0.3) 5 that's it (.) like it gets said (0.2) 6 an (.) then [nothing else happen]s 7 → THP: [+gets sa:id ] +((gaze to CLT with slow nod)) 8 (1.3) 9 MOT: mtk .h (.) well usually that's because our 10 family comes up with som:e (.) amazing dra:ma 11 (0.6) to throw a spanner in the wo(h)rks:
At lines 4-6 the Client states that the action plan just “gets said” and that nothing happens as a result. At line 7, the Therapist repeats “get said” in overlap with the end of the
179
Client’s turn “nothing else happens” at a place when the Client’s turn is not yet complete.
The repeat is spoken at a clearly audible volume, with emphasis on “said”, gaze to the Client, and slow nods. This is followed by a silence of 1.3 seconds at line 8, before the Mother says that the family usually creates some sort of drama, as a reason for why the plans are not enacted. The Therapist’s repeat occurs at a non-transition relevance place and is not responded to by any of the participants, nor does the Therapist reissue their repeat or revise their turn in order to pursue a response during the long silence at line 8. Instead, the Mother responds to the prior talk of the Client rather than the Therapist’s repeat. The repeat is not treated as an initiation of repair nor a request for elaboration but, along with the slow nods and lack of follow up from the Therapist, seems to function as a receipt of the Client’s talk that does not normatively project a vocal response.
In Extract 10, the Therapist repeats the words of the prior speaker in overlap with the talk of the next speaker. This example begins with the Father speaking about being unable to
“correct” the behaviour of his daughter (the client).
Extract 10
S4.rpt8
FAT: father; THP: therapist; CH1: child 1
1 FAT: >obut youo< CAn't correct her (.) 2 normally [a kid] you can correct an say okay= 3 THP: [no: ] 4 FAT: =(.) i: do it fo:r you¿ (1.3) impossible 5 (0.5) 6 CH1: >or she won't< let you clean it up at a:ll 7 FAT: [mm] 8 THP: [mm]= 9 CH1: =>>or go<< i don't make a mess: 10 or she won't clean it up at all 11 .h (0.2) >>(which means) sh-<< (.) 12 like makes (.) mo::re (.) angry cos 13 she's not appreciating at all what you're doing 14 (0.6)
180
15 FAT: that's an ex[ample so she] (.) ca:n't= 16 → THP: [+ooappreciatin:g_oo] +((CH1 gaze to FAT)) +((THP maintains gaze to CH1)) 17 FAT: = correct her so if:: (.) i- it's (.) it’s: 18 (0.3) 19 ((continuing talk about client))
At lines 1-2 and 4, the Father says that he is not able to “correct” the behaviour of his daughter, and that with other children one is usually able to do things for them but that this is
“impossible” in their current situation. In overlap at line 3, the Therapist’s “no” is consistent with the Father’s stance position (Du Bois, 2007). From lines 6-13, Child 1 talks about some of the difficulties experienced with the client, who is her younger sister. At lines 12-13, Child
1 states that she can feel angry because her sister doesn’t appreciate what she is doing for her.
There is a silence at line 14 after which the Father self-selects as next speaker summarising the prior talk as an example of not being able to correct the behaviour of the client. At line
16, in overlap with the Father, the Therapist repeats “appreciating”, a word used by Child 1 in the prior turn. This repeat occurs in overlap outside of a transition relevance place and is spoken very quietly with flat intonation. During the repeat, the Therapist maintains gaze towards Child 1, selecting her as the addressee of the repeat, but Child 1 turns to look towards her Father who has begun a turn before the Therapist. The Therapist’s repeat is not oriented to by the participants nor the Therapist as requiring a response, and the Father continues his turn with no apparent disruption. The repeat thus functions as a belated receipt of Child 1’s prior turn, which forgoes a substantial turn without interrupting the progression of the conversation.
In Extract 11, the Therapist’s repeat outside a transition relevance place occurs in overlap with the turn of the next speaker. But rather than receipt, as in the previous turn, the repeat serves the function of returning to the prior topic and avoiding a possible argument.
181
The extract starts with the Mother saying that the Client didn’t participant in a school program when he was admitted to hospital, because “he didn’t like the teachers there”.
Extract 11
S10.rpt4
MOT: mother; CLT: client; FAT: father; TH1: therapist 1; TH2: therapist 2
1 MOT: (0.4) he didn't like the teachers there? 2 (1.1) 3 CLT: an their ru:les 4 (0.6) 5 MOT: an their ru::les 6 (0.7) ((mutual gaze between CLT and MOT)) 7 MOT: you can tell them becaus:[:e] 8 FAT: [do]n't look at 9 [mum you need to look at the other] people= 10 MOT: [it's:: (0.2) you: ] 11 FAT: =that are here mum's (.) not 12 [gonna be your saviour]+[so ] ya’know you talk= 13 → TH1: [+their ru::les: ] +((gaze to CLT)) +((CLT gaze to TH1)) 14 TH2: [hm.] 15 FAT: =to the people [(an they'll) help ya] 16 CLT: [yea:::h li]ke- 17 (0.9) if you got sca:rs an stuff you gotta 18 cover up twenty four seven,
At line 3, the Client adds an increment to the Mother’s turn that he did not like the school because of “their rules”. After a silence, the Mother directs the Client to speak (line 7).
At line 8, the Father directs the Client to look at the therapists rather than his Mother and that his Mother is not going to be his saviour, a reference to the Client not speaking for himself.
The Therapist’s repeat at line 13 of “their rules” with emphasis and sound stretch occurs in overlap with Father’s talk and their gaze is directed to the Client. This repeat topicalises the earlier talk about “their rules” and competes with the Father’s turn (Schegloff, 2000). The
Client is being addressed simultaneously by the Father telling him to speak to the therapists, and the Therapist inviting him to speak about the earlier topic of “their rules”. The
182
Therapist’s repeat also directs the Client away from responding to the Father’s sanctioning and away from a possible argument. The repeat thus serves two functions, a request for elaboration, and avoidance of a conversational trajectory that may result in overt conflict.
In summary, when repeats outside of a transition relevance place occur during the talk of the same speaker as the first saying, they generally do not receive a response or impact on the sequential trajectory of the turn. They therefore function as receipts of part of the ongoing turn that do not require a response. When repeats outside of a transition relevance place occurred during the talk of the next speaker, there was not the same consistency in function.
In Extract 11, the repeat occurred in competition with another speaker and served dual functions of requesting elaboration on a therapeutically relevant topic and diverting the conversation away from a potential argument. Repeats outside of a transition relevance place generally do not interrupt the progression of the conversation and display the therapist’s receipt, acknowledgement, and attention to the developing talk of the clients. In this sense, they are consistent with an Open Dialogue approach, however they generally do not lead to the elaboration that is proposed in writings on Open Dialogue repeats.
Discussion
In this study, therapists’ repeats were found to serve a wide variety of interactional functions. They can initiate repair, mark receipt, nominate a topic for further discussion, and direct the conversation away from an undesired area. We found that the initiation of repair and third position receipts functioned in a way similar to those described in previous CA research on everyday conversation. Repeats in other positions encouraged elaboration but could also be treated as requests for confirmation or initiation of repair. In these situations, the design and sequential placement of the repeat nominated a particular part of the prior talk for development while providing a variety of fitted ways to respond.
183
Repeats that elicited elaboration are strongly aligned with the stocks of interactional knowledge in Open Dialogue recommending the use of repeats. This practice enables therapists to nominate a topic for further discussion without, for example, the use of interrogatively formed questions, which may potentially constrain the types of responses a recipient provides. Instead, these repeats display receipt of the prior talk, and select a part of that talk for further development, and project a range of potentially fitted responses.
Promoting such flexibility in responding aligns with Open Dialogue principles in that therapists can nominate topical areas for further comment while lessening the restrictions on how the family may respond. Previous CA research into Open Dialogue has found that therapists use different practices that generate similar outcomes. For example, Ong, Barnes, and Buus (2021) describe how therapists utilise the phrase “I’m wondering” to elicit stance positions from family members. These forms of turn design allow for a variety of responses from recipients and, to some extent, mitigate the authority of the therapist. These findings also align with the work of Knol et al. (2020) who describe how psychotherapists’ mirroring repeats serve as requests for elaboration that avoid adding the therapist’s interpretation.
Other functions of repeats such as the initiation of repair and displays of receipt (both in sequence closing positions and outside of transition relevance places) did not have the same coherence with Open Dialogue writings, but still displayed some consistency with the approach. By repeating the prior talk of the clients, the therapist displays their attention to the unfolding conversation and the particular words used by clients. So, even if repeats did not elicit elaboration and further talk, they at least displayed a therapist’s active engagement with the unfolding conversation.
Our results suggest that the stocks of interactional knowledge on the use of repeats in
Open Dialogue practice provide an oversimplified view of how repeats are utilised in actual therapeutic conversations, because repeats serve a variety of functions that are not accounted 184
for by current descriptions of dialogical practice. For example, therapists’ repeats are not a neutral mirroring back of the family’s words. Instead, repeats select and nominate particular parts of the prior talk for further development and thus reflect some professional decision- making by the therapist, which in turn guides the conversation in particular directions. This is similar to other therapist responses such as formulations, which provide a summary or
“upshot” of the client’s talk (Antaki, 2008; Antaki, Barnes, & Leudar, 2005). Previous CA research has similarly demonstrated that Open Dialogue therapists select parts of prior talk and exclude others in order to elicit stance positions from different participants (Ong, Barnes,
& Buus, 2021; Rossen et al., 2020). So, while the use of repeats limits the therapists’ responses to the actual words of the client, they still involve an exercising of authority through the selectivity and guiding of the conversation by the therapist.
This research suggests that the collaborative ideals of Open Dialogue are not consistently applied in a session. For example, even within apparently desirable practices like repeats, some selectivity, directing of the session, and exercising of authority by the therapist is unavoidable. In addition, economic and institutional demands direct therapists to work in ways that are not always consistent with a dialogical approach (Dawson, River, McCloughen,
& Buus, 2019, 2020; Ong et al., 2019). Current descriptions of Open Dialogue do not adequately address the authority of the therapist and the potential benefits of selectivity or control of the session to enable the promotion of dialogue (Ong, Barnes, & Buus, 2020c,
2021). The current theoretical descriptions of Open Dialogue mainly consist of aspirational ideals, while more practical and pragmatic accounts are relatively underdeveloped. The previously theorised functions of repeats are a clear example of how theoretical descriptions do not capture the complexity of therapeutic interactions.
An examination of the interactional details of Open Dialogue raises the question about what constitutes dialogical practice, and if it is something that is present in all actions 185
of the therapist or something that is applied at certain points in the interaction. Seikkula
(2011), for instance, raises whether dialogical practice is merely another form of psychotherapy or more a “way of life”. The implication is that a dialogical perspective pervades all aspects of a therapist’s life both in interactions with families and personally. In this view, Open Dialogue is not considered an “intervention” as the therapist does not set out to actively shape the behaviour of the client in a particular way (Seikkula, 2011). But, it has been regularly demonstrated by CA research that conversation, including psychotherapy conversation, is an interactional achievement that is co-created and influenced by those present in the interaction (e.g., Kupetz, 2014; Roca-Cuberes, 2016; Schegloff, 1982) regardless of their prior intentions. At other times, Open Dialogue has been treated as an intervention with attempts to develop a fidelity criteria including proposing a system of categorising utterances as “dialogical” or “monological” (Olson, Seikkula, & Ziedonis,
2014). An Open Dialogue session is said to be achieved when at least 2/3 of the conversation is deemed “dialogical” (Olson, Seikkula, & Ziedonis, 2014). Open Dialogue has thus been paradoxically positioned by leading authors in the field as both a “way of life” that guides all interactions and does not seek to intervene or change others, as well as a prescriptive type of interaction with the development of fidelity criteria evocative of other forms of psychotherapy.
A limitation of this study is that, while the participating therapists were experienced in the Open Dialogue approach, they were undertaking advanced training at the time that data was collected. These results may therefore represent a step in their process of applying a dialogical approach and a focus on following current prescriptive advice. This connects with a larger issue about what constitutes the essential practices of a dialogical approach and the current lack of an agreed and operationalised fidelity criteria. We therefore present these results as an indication of the real-world practices of therapists committed to a dialogical way 186
of working. Further research will be able to support whether these practices are also utilised by more experienced clinicians and if other functions of therapists’ repeats are evident.
Open Dialogue originators have encouraged the use of repeats, because they are thought to allow the client to hear and reflect on their own words (Haarakangas, Seikkula,
Alakare, & Aaltonen, 2007; Olson, Seikkula, & Ziedonis, 2014; Seikkula, 2011; Seikkula &
Arnkil, 2006; Seikkula & Trimble, 2005). CA research focuses on the observable details of conversation and any inferences about the internal thought processes of conversational participants must be grounded in their displayed interactions. This study was not able to assess the extent to which repeats encouraged reflection by clients. We were able to demonstrate that certain repeats were able to elicit client elaboration and further details on their prior talk, but we were not able to show that this elaboration was connected to internal reflections or new insights. This question may be addressed in future CA research, possibly with the addition of other methods to investigate the internal thought processes of participants.
This study is also limited in its ability to make recommendations for practice. While we describe a number of ways that repeats have been utilised in Open Dialogue sessions, we are not able to provide recommendations on how they should be used. Instead, we would suggest that close descriptions of the normative implications of conversational practices can provide therapists with an understanding of the repertoire of practices relevant for Open
Dialogue, which they then can implement in context-sensitive ways. In doing this, we hope to contribute to a more nuanced understanding of the conversational work that is done in therapy, and to highlight the limitations of presumptions about the functions of different conversational practices without reference to empirical findings about their organisation.
187
Chapter Summary
In this chapter, the use of repeats has similar characteristics to stance elicitations
(Chapter 5). This is because the repetition of the prior talk of the client nominates this talk as a topic for response, whether it be to initiate repair, mark receipt or to encourage elaboration.
In this sense, the therapist is again exercising their deontic authority by selecting topics for further development. But, as demonstrated in this study, the repeat does not clearly specify a particular type of responsive turn. For example, repeats suggestive of receipts, can be responded to with elaborated responses or confirmation. By using repeats rather than more explicit interrogative forms, the therapist presents a downgraded deontic stance that provides for responsive flexibility and agency as the client can respond in a number of fitted ways.
This study suggests that the Open Dialogue stocks of interactional knowledge can be complemented with a more nuanced conceptualisation of authority in light of how therapists exercise authority by selecting and nominating topics for further development, while also mitigating that authority by promoting flexibility in the clients’ responses. The therapist’s role and identity is again presented as one who mediates the conversation but does not determine the epistemic content of clients’ responses.
188
Chapter 7: Eliding Agreement in Therapist Reflections
In this chapter, I focus on reflecting conversations. Reflecting conversations or reflections have a unique organisation because they occur only between the therapists while the network observe quietly without interacting with the therapists. Only at the end of the reflection are the network members invited to comment upon the reflection. Reflections can serve multiple purposes in different types of psychotherapy. In Open Dialogue, reflections are a way that the therapists’ thoughts, feelings, emotions, images, and associations are explicitly introduced into the conversation. Reflections thus represent a very different interactional environment to the previous chapters where the conversation primarily occurs between the therapists and the client and family network. In this chapter, I present a study, accepted for publication in Discourse Studies (Ong, Barnes, & Buus, in press), that examines a particular practice in reflections where therapists do not explicitly agree but instead elide agreement with each other’s prior assertions.
189
Developing Multiple Perspectives by Eliding Agreement: A Conversation Analysis of
Open Dialogue Reflections
Collaborative approaches to psychotherapy are client-centred. This means that therapists focus on eliciting the perspectives of the client or family and position the client as an “expert” in their own lives (H. Anderson & Goolishian, 1992; Rogers, 1951). The therapist also has expertise, in the form of professional training and experience that is potentially helpful to the family. There can be a tension around how a therapist’s expertise is introduced into the conversation without minimising or invalidating the perspectives of the family. A collaborative approach such as Open Dialogue (Haarakangas, Seikkula, Alakare, &
Aaltonen, 2007; Seikkula & Arnkil, 2006) introduces the expertise of the therapist in two main ways. One way is to position the therapist as an “architect of dialogue” or a
“conversational artist” whose skill lies in asking certain questions and promoting dialogue between the family members (H. Anderson & Goolishian, 1988, p. 371). Another way introduces the therapists’ perspectives more explicitly through a technique known as reflections (Andersen, 1987).
In Open Dialogue reflections, the therapists pause the session with the family, and turn to each other to speak about their thoughts, images, feelings, or experiences that arose during the conversation with the family (Andersen, 1987; Olson, Seikkula, & Ziedonis, 2014;
Ong, Barnes, & Buus, 2020c). The family is encouraged to listen to this conversation without speaking until the reflection is finished, at which point the family is invited to respond to the therapists’ reflections (Seikkula, 2008). In reflections, therapists are encouraged to be tentative, use the words of the family, and to present their thoughts in a way that avoids offence (Andersen, 1987; Haarakangas, Seikkula, Alakare, & Aaltonen, 2007). Therapists are also encouraged to avoid agreement with each other so as to present a range of different ideas
(Andersen, 1987). Reflections are thus an explicit way that the therapists’ perspectives are 190
introduced into the session, and are thought to introduce further possibilities for dialogue (H.
Anderson & Goolishian, 1988; Seikkula & Trimble, 2005). Considering the importance of reflections in Open Dialogue, it is necessary to have a more explicit and thorough understanding of how these conversations actually occur, rather than only theoretical descriptions. Studying real-world Open Dialogue sessions provides an opportunity to describe dialogical processes, and how therapists construct dialogical reflections. One way of investigating such interactions is through conversation analysis, which examines the micro- processes of interaction and how people utilise various conversational practices in order to achieve social actions (Goodwin & Heritage, 1990; Heritage, 2013b).
Conversation analysis (CA) uses video and audio recordings of naturally occurring interactions to describe and analyse spoken and embodied interaction. CA has two main lines of interest, describing both the normative structural elements of conversations, and how people use these structural aspects to achieve and negotiate situated social actions. In CA, the analysis of social actions is not through reference to the internal processes of the participants but rather through the orientations displayed by the participants via their interactional conduct. CA has been used to study a wide range of interactions in different approaches to family therapy (for more detailed reviews see Ong, Barnes, & Buus, 2020a; Ong, Barnes, &
Buus, 2020b; Tseliou, 2013).
One area of interest in these studies is the use of reflections in family therapy which are briefly described below. Although not specifically described as such, Muntigl and
Horvath (2016) analyse a brief reflection in a Structural Family Therapy session. In this example, one of the therapists voices a perspective that disagrees with the participating mother while preserving face and not addressing the mother directly. The reflection also provides the mother with the opportunity to observe and consider the therapists’ talk without the opportunity or obligation to respond. In another CA study, Williams and Auburn (2015) 191
described how therapists in reflecting teams adopt stance positions that are orientated towards relational descriptions, and promoting hope and positive connotations. For example, a participating stepfather that was labelled by the family as strict and over-protective, was described by the reflecting team as having “love” towards his step-daughter. In Open
Dialogue reflections, Schriver, Buus, and Rossen (2019) used CA to describe two types of reflections. These included reporting my-side tellings, where therapists described their observations of what the family said and did, and inferring my-side tellings, where therapists provided their interpretations of what the family said. These studies show how therapists voiced different assertions and adopted stance positions towards the family that were generally positive and relationally-focussed. These assertions vary along a continuum from observations of what a family said or did, to more inferential or interpretive statements about the family. While these studies have focused on how the family is discussed by the reflecting therapists, they did not explore how therapists collaboratively constructed their reflections.
Because reflections require therapists to make various assertions about a family, they would seem to implicate responses of agreement or disagreement from the receiving therapist.
Therapists are thus tasked with presenting their ideas about the family while also negotiating and managing the normative demands of conversation and their own potentially contrastive positions in their reflections. It is not clear how therapists present different ideas in reflections nor how they manage issues of agreement and disagreement.
Past CA research suggests that there is a preference for maximising agreements and avoiding disagreements in conversation (Pomerantz & Heritage, 2013). For instance, agreements with assessments are usually provided quickly with little delay and can contain upgraded responsive stances, e.g., if a person says that it is a “beautiful” day then an upgraded second assessment from the recipient can be that the day is “gorgeous” (Pomerantz,
1984, p. 59). Upgraded second assessments also display a speaker’s independent assessment 192
and thus regain a degree of epistemic authority when speaking from second position (Enfield,
2011; Heritage & Raymond, 2005; Thompson, Fox, & Couper-Kuhlen, 2015). In contrast, disagreements are delayed until later in a turn though silences or adding a preface to the beginning of the turn (Pomerantz, 1984). Disagreements are a dispreferred response and they are produced less often in favour of responses that promote interactional progressivity and, consequently, social solidarity (Heritage, 2013b; Sifianou, 2012). Because disagreements in reflections potentially risk social solidarity, their introduction requires some interactional effort.
In psychotherapy, disagreement can occur when therapists introduce alternative ways of thinking and acting that contrast with a client’s perspective. As a catalyst of change, disagreements can be considered a helpful and necessary part of psychotherapy. However, disagreement needs to be responded to and resolved in some way so that the interaction can resume a therapeutic trajectory (Muntigl & Horvath, 2014). CA studies have identified that therapists can respond to client disagreements by retreating from their prior formulation to affiliate and strengthen the client’s alternative position (Muntigl, Knight, Watkins, Horvath,
& Angus, 2013), or the therapist can maintain their position in either a convergent and supportive way or a divergent and unsupportive way, implying that either the therapist’s or the client’s understanding was in need of correction (Muntigl et al., 2013; Viklund,
Holmqvist, & Zetterqvist Nelson, 2010; Weiste, 2015). More recent studies have shown how therapists work to downgrade both their deontic and epistemic authority in Structural Family
Therapy and Open Dialogue sessions (Muntigl & Horvath, 2020; Ong, Barnes, & Buus,
2020c, 2021). By downgrading their epistemic authority through phrases such as “what I hear...” (Muntigl & Horvath, 2020) or “I’m wondering...” (Ong, Barnes, & Buus, 2021), therapists defer to the epistemic authority of their clients and promote reflection and elaboration. As a whole, these studies show how therapists balance introducing new ideas 193
that may be beneficial to the client with having to maintain therapeutic engagement, social solidarity, and respect for the epistemic authority of the client and family. At the same time, such practices also have the effect of maintaining normative structures of sequences, promoting aligning responses from their recipients.
Reflecting conversations between therapists constitute a different interactional environment. This is because therapists do not have the same obligations to maintain the therapeutic relationship between one another as they do with clients. However, therapists likely still have to display some form of solidarity because overt disagreement may potentially confuse the family and derail the therapeutic process. In Open Dialogue reflections, therapists have to manage the institutional task of introducing multiple ideas to the family while also mitigating the potential negative effects of overt disagreement. At the same time, they must also manage generic conversational contingencies such as preference and normative expectations of aligning responses. In this study, we aim to identify how Open
Dialogue therapists respond to each other’s reflections, manage institutional tasks and generic conversational contingencies, and explore the implications of the elided agreement and potential disagreements for reflecting conversations.
Method
Participants and Data
Study participants consisted of Open Dialogue therapists (n=12), their clients and the clients’ social networks including family and other professionals (n=36). Ten therapists had completed a 5-day foundation training in Open Dialogue with a minimum of 2-years of experience working in the approach. Seven therapists were undertaking advanced training as
Open Dialogue therapists or to be Open Dialogue trainers. The remaining two therapists had informal orientation training in Open Dialogue through their more experienced colleagues.
Open Dialogue is centred around a number of principles focussing on providing immediate 194
help, including families and social networks, maintaining therapist continuity through the therapeutic process, and promoting dialogue (Haarakangas, Seikkula, Alakare, & Aaltonen,
2007; Seikkula & Arnkil, 2006). Decisions about treatment are made collaboratively between therapists, clients, and their families and can include other forms of psychotherapy if it is jointly decided on. Open Dialogue can thus be considered a form of family therapy centred around promoting dialogue as well as a collaborative decision-making process (Ong et al.,
2019). In this study we focus only on one aspect of Open Dialogue, the reflecting conversation.
The first author provided information on the study protocol and aims to potential therapist participants. If therapists were interested in participating, they then approached the first author and provided written consent. The participating therapists then identified families who may be interested in participating. If families were interested in participating, the first author provided them with detailed information on the study. If they still wished to participate, they provided written consent. Data collection consisted of video and audio recordings of 10 Open Dialogue sessions, totalling 14 hours of video. Three cameras were placed around the room to document verbal and non-verbal conduct from all participants.
The study was approved by the Nepean Blue Mountains Local Health District ethics committee (reference number: HREC/17/NEPEAN/135). All data presented in this article has been anonymised to ensure participant confidentiality, with personal names replaced with pseudonyms.
Analytic Process
The analytic process followed the procedures recommended by ten Have (2007). This included first viewing and transcribing the recordings verbatim and an “unmotivated looking” for interactions of interest (Psathas, 1995; Schegloff, 1996). These initial viewings yielded a number of conversational practices of interest that have been published previously (Ong, 195
Barnes, & Buus, 2020c, 2021). For this study, we examined therapist assertions during reflecting conversations. We identified 17 reflecting conversations across the 10 sessions, each averaging 5 minutes in length. We noticed that, at times, therapists did not explicitly agree or disagree with each other. We focussed our analysis on identifying examples of absences of explicit agreements, specifically focussing on environments where an absence of agreement occurred after an inferential form of reflection and were thus noticeably absent
(Schriver, Buus, & Rossen, 2019). Examples were transcribed in detail according to CA conventions capturing details such as intonation, prosody, overlap and silences (Hepburn &
Bolden, 2013, 2017).
Findings
We found that therapists’ initiating reflections consisted of assertions about the family that were not responded to with overt agreement/disagreement, thus departing from normative expectations. Instead, responding therapists tended to elide agreement or failed to endorse the other therapist’s reflection through silences, prefaces, and re-starts. These responding turns then proceeded to defer to the epistemic authority of the client, assert epistemic rights from second position, emphasise a positive perspective, or voice multiple perspectives. Through eliding explicit agreement, therapists introduced multiple perspectives rather than one agreed upon singular perspective from the therapists while also attending to contingency.
Deferring Epistemic Authority to Client
In Extract 1, the absence of explicit agreement is displayed through silences and prefaces, and a proposal to direct the conversation towards the client and their primary epistemic rights and authority. In this extract, the therapists are discussing a prior conversation with the family where the client was feeling uncomfortable with direct eye
196
contact and the therapists tried looking in a different direction while speaking to the client. At the beginning of Extract 1, T2 is reflecting on whether looking away was easier for the client.
Extract 1
S8.E11R – 48:35
T1: therapist 1 (Julie); T2: therapist 2; Tracey: client
1 (1.6) 2 T2 >i- so i guess julie< that's one of the- 3 (0.2) also the things i (.) wonder about huh 4 .hh (0.3) is (0.2) was it easier to 5 have the conversation about the i q testin:g 6 (0.2) 7 T1 mm 8 (0.3) 9 T2 without the lookoing?o 10 (0.2) 11 T1 → yea:h (0.3) i wonder (0.3) oif=sortao (.) 12 tracey might (0.3) be able to tell us? 13 T2 hm 14 (0.4)
At line 2, T2 addresses the co-therapist, T1 (Julie). There is a long beginning to the turn with multiple components as T2 focuses on achieving a particular wording. The turn includes
“I guess”, suggestive of an upcoming evaluation and response seeking (Kärkkäinen, 2007) and “I wonder” which is associated with low entitlement and high contingency (Curl &
Drew, 2008) and a downgrading of epistemic and deontic authority (Ong, Barnes, & Buus,
2020c, 2021). T2 thus foreshadows an upcoming evaluative position as well as displaying their low rights to make that evaluation or potential difficulties in the recipient’s ability to answer. T2’s turn ends with a yes/no interrogative about whether the prior conversation
(about IQ testing) was easier for the client (Tracey) when they weren’t looking at her directly. This refers to an earlier part of the session where the client was feeling anxious and the therapists tried not looking at the client directly when speaking to her (data not shown).
197
This interrogative is preceded by “also the things I wonder about”, transforming the interrogative into a speculative statement or inferring my-side telling (Schriver, Buus, &
Rossen, 2019). T2’s turn contains multiple components that project a number of possible responses i.e., an interrogative projecting a confirming or disconfirming polar response
(Raymond, 2003), or a complex assertion projecting agreement/disagreement. In any case, T2 topicalises the client’s response to the conversation about the IQ testing as the focus of their turn.
T1’s response does not include explicit agreement (beginning at line 11) and contains a number of mitigating features. At line 11, T1 begins with “yeah”. This “yeah” has the quality of an acknowledgement token, which does not precisely conform to the expectations set up by the previous interrogative, and instead conveys weak and/or neutral receipt of the turn more broadly. This is followed by a number of within-turn silences without non-verbal agreement markers that also avoid providing straightforward agreement with relevant parts of the prior turn. T1 continues with “I wonder” and “if sorta”, foreshadowing an upcoming proposal with low entitlement (Curl & Drew, 2008). “I wonder” was also present in T2’s initiating turn (line 3) thus communicating a similar downgraded epistemic position and maintaining a degree of syntactic continuity (Du Bois, 2014). T1 then proposes directing the question towards the client, Tracey (line 12) who has epistemic authority and rights to comment on their own internal experience. The conversation then moves out of the reflection to speak with Tracey directly (data not shown).
In summary, T2’s initiating turn presents a complex speculative action with a number of mitigating features that downgrade their authority and certainty. T1’s response displays an absence of explicit agreement (through acknowledgement and lack of non-verbal agreement markers), a similar downgraded epistemic position, and proposes to direct the question to the client. T1 thus avoids participating in further speculation about a topic in that is in the client’s 198
epistemic domain. This allows the client to display and claim epistemic rights on their own experience as well as avoiding a potential breach of social solidarity by not overtly disagreeing and undermining T2’s initiating turn and perspective.
Asserting Epistemic Authority from Second Position
In Extract 2, an absence of agreement, displayed through verbal and non-verbal resources, is later modified to present a re-authored independent stance claiming epistemic authority from a responsive position. This displays how therapists manage and respond to generic conversational expectations during reflections, and how these considerations may interpose in the content of the reflection. Extract 2 begins with the therapists commencing a new sequence of talk about the humour displayed by the family earlier in the session.
Extract 2
S4.D2 – 47:08
C1: child 1; C2: child 2; C3: child 3; C4: child 4; T1: therapist 1; T2: therapist 2
1 (1.6) 2 C3 ((whines)) 3 T2 >there was a lot’ve< [humour thrown]= 4 C4 [shh ] 5 T2 =around=i think- (0.2) [>the family uses it<= 6 T1 [mtk .hh 7 T2 =for resiloience i thi:nko 8 C1 +ooyeah we dooo ((to C2))+= T1 → +upturned mouth and shrug+ 9 T1 → =oi don'to kno:w (0.2) but [they certainly] 10 C1 [ohuh huh heho ] 11 (0.2) 12 T2 mm 13 (1.6) 14 T1 huh 15 (0.7) 16 T1 yea::h. (.) there's a lot've laughter 17 T2 [m¯m, 18 C1 [((whisper)) 19 (1.0)
199
The example begins with a silence of 1.6 seconds (line 1) marking closure of the previous sequence. At line 3, T2 launches a new sequence with an observation there was “a lot’ve humour thrown around”. T2 continues with another component, “I think” latched on to the prior utterance. After a brief silence, T2 completes the turn with an inferential statement that the family uses it (humour) for resilience, in a lower volume and ending with “I think”.
T2’s turn thus consists of a behavioural observation and ends with their own inferential interpretation about the family’s use of that behaviour. T2’s inferential component begins and ends with “I think”, which presents their inferential talk as something that is from their own perspective or opinion while also introducing a downgraded certainty and lack of commitment, and seeking a response such as endorsement or rejection from the co-therapist
(Kaltenböck, 2010; Stevanovic, 2013a). T2 thus presents an inferential statement in the form of a declarative that makes relevant agreement or disagreement from the other therapist. This differs from Extract 1, where the therapist’s initiating turn contained multiple components that could be heard as an interrogative or a complex assertion that in turn makes relevant a greater number of possible fitting responses.
At line 8, C1 whispers to her sister C2, “yeah we do” displaying her agreement that the family uses humour for resilience. It is not clear if it was heard by the therapists. C1’s talk occurs in overlap with T1’s shrug and upturned mouth expression suggesting upcoming disagreement with T2’s prior turn. At line 9, T1 responds with “I don’t know”, which functions as an account for not agreeing with the prior turn and is also associated with avoiding commitment to the current course of action when following evaluations and questions (Weatherall, 2011). T1’s turn has an absence of agreement with all or part of T2’s prior turn through verbal and non-verbal means. T1 continues with an aborted “but they certainly” followed by a short silence. T1’s turn is not yet complete, and the other therapist forgoes taking a substantial turn with a continuer, “mm”, at line 12. After a long pause, T1 200
laughs and, following another silence, provides an agreement token with an emphasised and drawn out “yeah”. T1 then voices their own observation “there’s a lot’ve laughter”, which is an aligning, although re-authored version (Thompson, Fox, & Couper-Kuhlen, 2015) of T2’s prior observation that there was “a lot’ve humour”. In addition to being re-authored, T1’s
“there’s a lot’ve laughter” occurs in an initiating position and through this, T1 states their own independent assertion about the family that is different from T2’s prior position. T1 claims independent access and asserts their epistemic authority (and agency) to produce an independent interpretive assertion about the family (Enfield, 2011; Heritage & Raymond,
2005).
In this extract, the absence of explicit agreement is displayed through non-verbal signs of an upturned mouth and shrugging, and a verbal marker of “I don’t know”. T1 makes an observation about the family that is partially consistent with T2 but, because T1’s observation is an initiating turn and re-authored, it claims an independent assessment of the family.
Through this conduct, T1 does not explicitly agree with T2’s claimed function of the family’s behaviour, even when there is an (albeit possibly unheard) endorsement by a family member during the reflection. This extract also demonstrates how the sequential organisation of conversation and the ensuing implicit assumptions about epistemic primacy and authority influence how therapists produce and respond to reflections.
Emphasising a Positive Perspective
In Extract 3, the absence of explicit agreement is displayed through minimal responses and a return to earlier talk in the conversation that emphasised a more positive focus. Just prior to the extract, the therapists have been discussing that there is a lot of love between the daughter and her mother, and also a lot of negative feelings between the sisters
(data not shown). At line 1, T1 describes the love between the sisters as “obviously it is there
201
but then” before moving on to describe the negative feelings between the sisters as “the hatred which is there”.
Extract 3
S5.D3R – 1:15:52
T1: therapist 1; T2 therapist 2; Mo: mother; Cl: client
1 T1 [>obviously it's< the:re but then,] 2 [((Mo and Cl inaudible talk)) 3 T1 >is=the [hatred which is there.< (0.8)= 4 [((Mo and Cl inaudible talk)) 5 T1 =which i could s:ee becoming [bigger?]= 6 T2 [mm ] 7 T1 =+.hh + T2 +nods+ 8 (0.6) 9 T1 an (.) prevailing over (1.1) 10 thoughts [an fee:li:ngs and wo::rds= 11 T2 [omm.o 12 T1 =[an actions (0.3) .hh 13 T2 [omm_o 14 +(0.6)+ T2 +nods + 15 (1.0) 16 T2 omtk=yeah.o 17 (0.7) 18 T1 omm:o 19 (1.7)+(0.6) + T2 → +head tilt with hand palm up and down+ 20 T2 ( ) >i said i< thin:(.)k (0.8) 21 T1 hmm:. 22 (0.2) 23 T2 → >it=wz important that< (.) they could also 24 (0.2) that they were still able to 25 talk oabout ito 26 (0.3) 27 T2 othat (.) for=me felt importanto 28 (0.3) 29 T1 hmm. 30 (1.7)
T1’s talk at line 1 occurs in overlap with the mother and daughter who are speaking to each other. At line 3, T1 begins a question “is the hatred which is there”. This is followed by
202
an inserted clause “which I could see becoming bigger...” (line 5). The question is not completed and the turn instead ends in an inferential telling in the form of a prediction about the hatred between the sisters becoming bigger. This finishes with rising intonation seeking a response from the other therapist (Stivers & Rossano, 2010). T2 responds with nods but does not provide a verbal response, leading to silence at line 8. T1 then adds an increment to their turn with a further prediction about the hatred then “prevailing” over other thoughts, feelings, words, and actions. During this turn, there is an absence of explicit agreement as T2 provides receipts with a quiet “mm” at lines 11 and 13. T1’s turn, from lines 1-12, has an ambiguous structure as it begins with an interrogative format that is not completed and ends with what could be considered an inferential prediction. There are thus multiple elements of the turn that T2 can respond to; the question and the propositions conveyed via the question or the predictive inferential telling.
T2 provides minimal uptake with silences (lines 14-15), nods (line 14), and a verbal response of a quiet “yeah” (line 16). This is followed by further silences at lines 17 and 19.
The response tokens of nods, “mm” and “yeah” are superficially similar to usual agreement markers. However, in this context and combined with quieter volume, lack of emphasis and silences, these response tokens instead suggest acknowledgement rather than agreement and delay the voicing of disagreement (Pomerantz, 1984). In addition, through these silences, response tokens and delayed responses, T2 is displaying a similar downgraded epistemic stance to T1. The subsequent verbal and non-verbal actions of T2 confirms their lack of explicit agreement with T1’s prior talk as well as a similar downgraded epistemic stance. At line 19, T2 makes a momentary gesture briefly turning their hand palm up and down with a sideways head tilt (intuitively, an “I don’t know” gesture), followed by another silence. At line 20, T2 begins a turn with “I said I think”, foreshadowing something that they said earlier in the reflection. After a silence T2 continues their turn, focusing on the relationship between 203
mother and client and specifically that it was important that they were able to talk about the problems, something that T2 had mentioned earlier in the reflection (data not shown). At line
27, T2 stresses that this is their interpretive evaluation, stating “that for me felt important”.
This specifically locates the importance as occurring for them personally and not as an objective observation while also emphasising the independence of their position (Heritage &
Raymond, 2005; Thompson, Fox, & Couper-Kuhlen, 2015). This is followed by a receipt from T1 and silence that is later followed by a topic change (data not shown).
In this extract, T1’s initiating turn begins with an interrogative form but ends with a strong prediction about the conflict in the family that receives minimal response and no explicit agreement from the other therapist. T1 then adds more detail to their turn that again receives little response from T2. T2 responds by not mentioning T1’s prior talk, instead repeating his own positive interpretation of the family’s behaviour that was mentioned earlier in the reflection. This closes further talk on T1’s previous predictive statement and the conversation moves to a different topic. T2’s turn with an absence of explicit agreement through minimal responses and embodied actions, leads to a change of topic that closes the current conversation. T2 does not explicitly disagree and thus avoids negating T1’s earlier statement but does voice a contrasting positive perspective on the sisters’ relationship.
Eliding Agreement to Voice Multiple Perspectives
In Extract 4, both therapists produce successive speculative statements that are met with an absence of explicit agreement. Extract 3 also demonstrates the non-verbal resources that can be utilised when eliding agreement, and how the absence of agreement can transition to sequence closure through a change of topic. Extract 4 starts with T1 speaking about how the focus of the family has shifted to Harry (previously the concerns were about Harry’s sister). At line 1, T1 begins with a statement speculating on what the conversation has been like for Harry and the family during the session today. 204
Extract 4
S9.D1R – 52:57
T1: therapist 1; T2: therapist 2; Liz: mother; Harry: child; Abe: biological father
1 T1 i'd (.) >i wonder what it's< li:ke, 2 (0.6) for harry an for everybody really da: 3 (0.3) 4 T2 mm_ 5 (0.8) 6 T1 to (.) talk about this toda:y. 7 (0.6) oio don't know if they ca:me 8 prepared ofor+thiso ooor notoo T2 +circular head movement------> 9 (0.2) + ----->+ 10 T2 mtk (.) yea:h. 11 (0.4) 12 T1 [um: ] 13 T2 [>i fel=like<] the focus has
At line 8, T2 responds with an ambiguous circular head movement that is partially vertical (i.e., nodding) and partially lateral (i.e., head shaking). This is accompanied by the verbal token “yeah”, which could be acknowledgement or agreement (line 10). In either case there is not a strong marker of agreement from T2 in response to the speculative part of T1’s turn. Both therapists begin talking at the same time at lines 12 and 13. T2 continues that they 205
feel like the focus has shifted (line 13). T1 agrees via nodding and a prosodically stressed
“mm” (line 15). At line 16, T2 continues that the focus has shifted “from Harry being in the background maybe he’s now”. This suggests an upcoming speculation about Harry that is not immediately completed and is instead followed by a silence and “mm” (lines 17-18).
Following T2’s speculative beginning, T1 offers a collection of uncertainty gestures including a shrug, eyebrow raise and an upturn of both palms, in a gesture of “maybe” or “I don’t know”. These gestures convey an unwillingness to explicitly endorse T2’s assertion while also presenting a similarly downgraded epistemic stance. T2 completes their turn at line 19 and 21, but it is only partially audible. T1 voices a quiet “yeah” at line 22 but, in the context of T1’s other talk and non-verbal gestures, this indicates acknowledgement/receipt rather than agreement. At line 23, T1 begins an acknowledgement of T2’s talk with a quietly spoken modified repeat “it’s something to ta-” which may be “something to talk about”. T1 aborts this turn and exerts their agency and deontic authority by launching a new topic, spoken with increased volume, that the biological father “Abe” is not in the room, thus closing off the prior topic of conversation.
In this example, both therapists produce speculative assertions and both are responded to with verbal and non-verbal gestures including a quietly spoken acknowledgement token, shrugging, eyebrow raises and upturned palms, indicating an absence of agreement with the speculation. The speculations also differ in the degree of directness with the first (lines 1-8) utilising multiple components like in Extract 1, and the second stated more directly as in
Extract 2. At the end of the extract the uncertainty gestures and absence of explicit agreement is quickly followed by a change of topic that closes the current sequence and any further speculation about the effect on the client. The therapists thus produce multiple perspectives without developing sequences involving explicit agreement or disagreement. This allows multiple perspectives to be voiced without preferencing or invalidating the other perspective. 206
Therapists thus produce multiple perspectives while also managing normative conversational expectations.
Discussion
Using CA, we analysed and presented a number of ways that Open Dialogue therapists elided explicit agreement during reflections. Therapists utilised a range of verbal and embodied resources to accomplish this, including the use of delays in responding, within- turn silences, minimal responsive tokens, and bodily movements like shrugs, head tilts, upturned mouth, and hand gestures. This absence of agreement was particularly noticeable when one therapist voiced some form of inferential telling such as an interpretation, speculation, or prediction about the family. In addition, the absence of agreement had an impact on the direction of the reflecting conversation by transitioning into other areas, such as transitioning to a question directed to the client thus deferring to their epistemic authority
(Extract 1), transitioning to a therapist’s independent assessment of the family (Extract 2), and closing down the current topic of discussion with a more positive interpretation (Extract
3), or transitioning to a new topic (Extract 4). Through eliding explicit agreement, therapists regulated the conversation and each other’s actions by receipting/acknowledging prior turns and closing down some areas of conversation and transitioning into others. Therapists thus did not work to resolve their lack of agreement towards either congruent or divergent outcomes before transitioning to new topics. This differs from disagreements between therapists and clients where therapists worked to produce some form of resolution (Muntigl
& Horvath, 2014; Muntigl et al., 2013). Therapist reflections instead included sequences of assertions that produced multiple perspectives, while also attending to normative conversational expectations such as receipt and acknowledgement of prior talk.
The multi-clausal turn design evident in both the initiating reflection and the responses with an absence of agreement suggests that there are multiple factors influencing 207
the construction of therapists’ reflections. These include more proximal and generic conversational factors such as the relative distribution of epistemic access and rights and the epistemic implications of speaking first. There are also institutional influences such as the stocks of interactional knowledge (Peräkylä & Vehviläinen, 2003) on how Open Dialogue is to be conducted. It seems that the epistemic rights that are encoded in conversational structures conflict with the promotion of collaboration and equality recommended by Open
Dialogue. In order to present a collaborative position, Open Dialogue therapists need to design their turns in ways that mitigate the authority that arises not only through their institutional position but also the authority implicated through sequential positioning.
Open Dialogue reflections involve observations as well as inferential statements and assessments about the client and family (Schriver, Buus, & Rossen, 2019). First position assessments claim epistemic primacy by virtue of their sequential position of being first, while further assessments (about the same object) are heard as responding to the first assessment (Heritage & Raymond, 2005). This creates possible tensions and competition between therapists in reflections as it may appear that those who voice reflections first are seen to be introducing new information while subsequent ones may appear to be merely agreeing, even if they had previously thought similar ideas independently. This can result in first position assessments or assertions being downgraded in order to promote a more symmetrical distribution of authority as in Extracts 1, 2, and 4. Similarly, a therapist may upgrade their authority from second position by withholding agreement and making efforts to produce an independent assessment even though it is in alignment with the initial assessment by their co-therapist (see Extract 2). The asymmetries inherent in the structure of conversation do not easily fit with conceptual ideas such as collaboration and equality between voices as promoted in Open Dialogue. These asymmetries and the potential for
208
competition over epistemic authority can potentially distract the family from the content of the reflections and their purpose of introducing multiple perspectives into the conversation.
Therapists also made efforts to explicitly mark their ownership of their presented thoughts during the reflection. By making explicit the ownership of their interpretations, therapists forgo the opportunity to present their interpretations as objective facts and mitigate the implication of superior claims to knowledge about the experience of the family. Issues of epistemic asymmetry and downgrading practices are also important and relevant for everyday conversation, with speakers constructing their talk orienting to the relative epistemic access and authority of their recipient/s (Heritage & Raymond, 2005). The position of a therapist, however, is different from those in everyday conversation as the role of a therapist has certain institutional roles and duties, as well as connotations as an authoritative expert on matters of mental health regardless of the intentions of the individual (Guilfoyle, 2003).
Conceptual writings on Open Dialogue encourage the voicing of different perspectives under the metaphor of polyphony (e.g., Mikes-Liu, Goldfinch, MacDonald, & Ong, 2016; Olson,
Seikkula, & Ziedonis, 2014; Seikkula, 2008). Therapists therefore have a principled imperative to contribute their own thoughts and inferences into the conversation while avoiding making superior claims to understanding the family or generating a single “version” of a family’s experiences. However, therapists’ utterances unavoidably have authority by virtue of their institutional status. In order to promote a collaborative approach with families, dialogical therapists need to constantly work at downgrading the authority connected with their position (Guilfoyle, 2003). So, while epistemic downgrading is present in everyday conversation, it has particular institutional relevance for Open Dialogue therapists who are encouraged (and obligated) to promote collaboration and equality.
CA research has detailed a generic conversational preference to provide agreement with assertions in order to maintain sequence progressivity and social solidarity (Pillet-Shore, 209
2017; Sacks, 1987). This conversational preference again does not easily fit with the institutional imperative to promote multiple polyphonic perspectives; if therapists agree with prior reflections it reduces the possibility for multiple perspectives, while disagreement risks social solidarity. Eliding agreement may be a way of fulfilling both of these opposing demands. Through an absence of explicit agreement, therapists can generate multiple candidate interpretations of the families’ experiences, and avoid explicit disagreement or negation of the perspective of the other therapist, thus minimising disruptions to social solidarity. By avoiding explicit disagreement therapists may also avoid undermining each other’s authority in front of the family. The absence of agreement may therefore be an expected and appropriate response in reflections. This aligns with previous research that has found disagreement to be not necessarily a negative act and, in some situations, is appropriate and even expected. For example, when problem-solving in business settings, disagreement is viewed as useful, necessary, and associated with creativity (Angouri, 2012), and in focus groups or debates, moderators encourage and invite disagreement and different perspectives
(Angouri, 2012; Myers, 1998; Sifianou, 2012). However, disagreement involves some delicacy from the dialogical perspective in order to promote collaboration and avoid making superior claims to the experience of others. Therapists therefore did not produce bald disagreements and instead utilised a number of means of withholding agreement before introducing other perspectives.
A limitation of this study is that most therapists were engaged in advanced training in
Open Dialogue and these practices and actions may reflect their different perceptions of the approach or their attempts at trying new techniques. Further research on this topic involving more experienced therapists may provide an interesting comparison to the current study. This study is also limited in its ability to make firm recommendations for clinical practice. Instead, we present these findings as examples of ways that therapists may produce reflecting 210
conversations and the complex conversational ways that this is achieved. These findings still represent actual conversations and the displayed orientations of therapists. These conversations provide a source of practice-based evidence that can inform current understandings of dialogical approaches by providing details of therapist conduct not previously considered or described in theoretical models (Peräkylä & Vehviläinen, 2003).
The Open Dialogue community may then consider how therapists manage the opposing imperatives described above and if these or other conversational techniques may better achieve the goals of Open Dialogue.
In this study, using conversation analysis, we demonstrated how Open Dialogue therapists elided explicit agreement and some functions this supports in reflecting conversations. The presentation of different perspectives in a reflection is a means by which therapists present diverse ideas to a family that can result in different perspectives and potential disagreement between the therapists. Eliding agreement is a way in which therapists can guide the conversation in particular directions while delivering independent perspectives.
Eliding agreement avoids creating consensus between the therapists, allowing a number of perspectives to be presented in reflections while also avoiding explicit disagreement and potential ruptures to social solidarity. We also hope to have demonstrated how an examination of the details of conversations, as they actually occur, can provide insights that are not considered or proposed by theories about therapy and its conduct. Theoretical principles are necessarily realised in context-sensitive ways and it is important to acknowledge that therapeutic practice must proceed with reference to both theoretical principles and their empirical realisations.
211
Chapter Summary
Through eliding agreement with prior assertions, therapists managed and guided the reflecting conversation by closing down some areas of discussion and transitioning to others.
This was particularly evident in response to reflections containing inferential talk such as interpretations, speculations, and predictions about the family. Therapists also made attempts to explicitly mark their ownership of their assertions and avoided presenting their reflections as objective universal observations about the family. This is coherent with the professional stocks of interactional knowledge where therapists’ perspectives are voiced tentatively.
Eliding agreement thus facilitated the voicing of multiple perspectives while also deferring to the epistemic authority of the observing network. Like in previous chapters, this study again displays how therapists are sensitive to issues of authority and present a downgraded stance to avoid claims of epistemic authority over the understandings of a network’s situation. In addition, this study demonstrates how the structure of conversation and particularly sequential positioning of turns, implicates certain asymmetrical epistemic relations between speakers. These normative assumptions contrast with a collaborative dialogical approach leading therapists to utilise various conversational practices, such as eliding agreement, in order to negate these implicit assumptions.
212
Chapter 8: Summary and Conclusions - Authority and Agency in Open Dialogue
Meetings
The studies in this thesis cover a range of conversational actions and practices in different sequential environments. A recurring theme through these studies, and one that has not been explicitly addressed in prior Open Dialogue writings, is that therapists displayed a consistent orientation to issues of authority. Therapists’ orientation to authority was demonstrated in two recurring and interconnected ways. First, therapists downgraded their deontic and epistemic authority, and second, therapists designed their turns so that recipients had flexibility in composing their ways of responding. These themes were present in all of the presented studies, and are summarised below.
In Chapter 4, therapists downgraded their deontic authority when proposing reflections through using “I’m wondering”, providing accounts, asking permission, and seeking acceptance from families and other therapists. These forms of proposals contrast with more constraining formats such as plain polar interrogatives, which strongly implicate a
“yes/no” response. Instead, therapists’ downgraded proposals sought collaboration and allowed recipients to respond in a variety of ways. But, despite their efforts at mitigation, therapists did not completely eliminate deontic authority. The fact that therapists, and not clients, proposed transitioning to a reflection reveals therapists’ authority to direct the course of a session, which the family was not similarly entitled to do. So, while therapists presented a downgraded deontic stance, their proposals still reflected their authoritative deontic status.
Therapists’ downgraded proposals could potentially serve purposes other than seeking collaboration. By downgrading their deontic authority, therapists imply that they do not have rights to direct the session without seeking some form of permission.
Chapter 5 demonstrated how therapists exercise and mitigate their authority through the elicitation of different stance positions from family members. Therapists exercised their 213
deontic authority by selecting parts of the prior talk or a new matter for further comment and selecting who would be the next speaker to respond. Through selecting and asking about a part of the prior turn, the therapists reveal aspects of their own stance position. But, by using
“I’m wondering”, therapists transformed their turns into declaratives rather than direct interrogatives and made their stance positions defeasible. This enabled therapists to introduce topics for further discussion, while also presenting a relatively “neutral” stance position towards those topics. Therapists’ turns also made relevant a range of different responses from recipients, which deferred to the epistemic authority of the recipient and allowed them some flexibility in how to respond. Therapists thus exercised their deontic authority by selecting the topic and speaker and to some extent controlling the timing of the recipients’ response.
However, the therapist also mitigates their authority through turn design, allowing for flexibility in the composition of the response, and downgraded their epistemic authority by deferring to the epistemic authority of the family. Therapists’ authority is again mitigated but not annulled. Because therapists control the timing and to a lesser extent the composition of the recipients’ response, agency is distributed between the recipient and the therapist although the accountability for the response is perhaps more heavily weighted towards the responding client (Enfield, 2017; Enfield & Kockelman, 2017).
A similar phenomenon was identified in Chapter 6 in the therapists’ use of repeats.
Therapists’ repeats in Open Dialogue sessions serve a variety of functions. One type of repeat with particular relevance to Open Dialogue are repeats that request elaboration. Through these repeats, therapists claimed deontic authority by selecting and nominating parts of the prior talk for further discussion. However, the repeat did not specify a particular type of response thus placing fewer constraints on the recipient and allowing them flexibility in the composition of their response. Therapists again present themselves as those who exercise authority by nominating speakers and topics, but mitigate their authority in relation to the 214
epistemic position of the recipient. Another type of repeat occurred in overlap at non- transition relevance places. These repeats occurred during the talk of a family member, however they were not treated as interruptions. This is because they were generally spoken with low volume and did not strongly implicate a response. Subsequently, they allowed recipients to respond by either treating the repeat as a receipt (that did not require a response), as requiring minimal confirmation, or as a request for elaboration.
In Chapter 7, I explored how therapists negotiated epistemic authority between themselves during reflections. Therapists who present reflections first have epistemic primacy because responsive actions are heard in relation to the first. In order to redress this epistemic normative bias, therapists who made initiating assertions could downgrade their epistemic authority, while therapists speaking in second position, elided agreement and worked to present their reflections as being independently held even when in alignment with the other therapist. Therapists also worked to explicitly mark their ownership of their reflections, for example through “I think”. This marks their reflections as being independently held while also downgrading their epistemic authority, as they did not present their reflections as having any universal authority. By downgrading their epistemic authority, therapists avoided making superior claims about the experience of the family thus deferring to the epistemic authority of the clients and promoting a collaborative relationship and therapist identity.
These findings demonstrate that Open Dialogue therapists attend to issues of deontic and epistemic authority and utilise a number of different conversational practices to downgrade their authority. These conversational practices are not individually unique to
Open Dialogue and can be found in research on everyday conversation. However, the configuration and recurrent use of downgrading practices by Open Dialogue therapists suggests that issues of authority are regularly (and perhaps uniquely) attended to and relevant 215
for Open Dialogue practice. Open Dialogue therapists do not only mitigate their authority, they also exercise their deontic authority by directing the structure of the session, selecting next speakers, and nominating topics for further discussion. Therapist authority is therefore not something that is avoided completely, but is instead mitigated or exercised in specific moments for professional objectives.
The presentation of downgraded epistemic and deontic stances may serve purposes other than the immediate contingencies of the conversation. Different types of relationships, and identities within those relationships, are defined by various constellations of rights and duties (Enfield, 2013). By downgrading their deontic authority to direct a session, therapists present themselves as not having rights to unilaterally make decisions about the session structure. Similarly, through turn design that promotes flexibility in recipient responses, therapists defer to the deontic and epistemic rights of the family. But as demonstrated through this thesis and elsewhere (e.g., Guilfoyle, 2003; Stevanovic, 2018), therapists’ presentations of a downgraded authoritative stance do not negate their authoritative status. Instead, the presentation of a downgraded authoritative stance indexes a particular configuration of deontic and epistemic rights consistent with a dialogical therapist identity that promotes a collaborate approach. Downgraded stances may therefore be connected to moral conceptualisations of therapist identity in addition to conversational contingencies.
Below, I discuss these findings, firstly with reference to how power and authority has previously been theorised in Open Dialogue and how the findings of this thesis can complement and extend this theorising. I then consider how the concept of agency, as conceptualised by Enfield et al. (Enfield, 2011, 2013; Enfield & Kockelman, 2017), can provide a framework for describing the therapist practices of downgrading deontic and epistemic authority and promoting flexibility in family responses. I then discuss the implications of these research findings for Open Dialogue and family therapy practice, the 216
limitations of this thesis, directions for future research, and close with my personal reflections.
Power in Open Dialogue
Power and authority has been given little direct attention in the Open Dialogue literature. In an epilogue on power and empowerment in Open Dialogue, Seikkula and Arnkil
(2006) acknowledge that power is a ubiquitous aspect of human relationships and, from a
Foucauldian perspective, cannot be completely removed from therapeutic interactions. A similar position is presented by Guilfoyle (2003) where power is located in institutions and roles rather than individual intentions. Seikkula and Arnkil (2006) argue that dialogue can transform power relations through a “shared expertise” that lessens the normalising role of therapist expertise on an individual’s conduct. That is, through dialogue, the therapist aligns with and attempts to understand the family rather than intervening to directly shape their behaviour, which suggests a more symmetrical power relationship. However, as discussed in relation to deontic authority, the therapist is only presenting a downgraded deontic stance, while their deontic status remains unchanged (Stevanovic, 2018). Einboden, Dawson,
McCloughen, and Buus (unpublished manuscript) also argue that the descriptions of
“empowerment” by Seikkula and Arnkil (2006) suggest a paternalistic view where therapists in effect bestow power on clients. Finally, “empowerment” amounts to self-regulation, which, from a Foucauldian point of view, is an aspect of subjugation in modern society (Einboden,
Dawson, McCloughen, & Buus, unpublished manuscript; Foucault, 1979). So, while Seikkula and Arnkil (2006) attempt to address the differential power relations in Open Dialogue, their account is somewhat contradictory and incomplete.
Apart from these preliminary ideas by Seikkula and Arnkil (2006) there have been few other discussions of power in Open Dialogue. Instead there has been more focus on
217
conceptual descriptions of the therapist’s mindset (Ong & Buus, in press). These recommendations tend towards broader descriptions of power by advising therapists to avoid hierarchy, to promote equality between different perspectives, to structure the meeting according to the input of the clients, and to maintain a “be with” rather than a “do to” perspective (Olson, Seikkula, & Ziedonis, 2014; Ong & Buus, under review; Shotter, 2006).
These recommendations involve the display of a flatter, egalitarian orientation where the therapists’ ideas are considered as having no greater or lesser importance than those of the family. But, as previously noted, the therapist’s authority is never completely removed
(Einboden, Dawson, McCloughen, & Buus, unpublished manuscript; Guilfoyle, 2003; Hare-
Mustin, 1994; Seikkula & Arnkil, 2006). Furthermore, the positive outcomes of Open
Dialogue can be seen as a product of therapists actually exercising power through following a systematic program of inviting network members to meetings and promoting cooperation
(Seikkula & Arnkil, 2006). So, while Open Dialogue recommendations advise therapists to promote equality and collaboration so that dialogue may occur, they do not directly acknowledge nor address the inherent and inescapable differences in authority between clients, families, and therapists, nor how power can be meaningfully employed to achieve desirable outcomes. In addition, these recommendations tend to remain at a more conceptual level of description and do not make any explicit recommendations on how these ideals are enacted in actual practice. An interactional account of power can address these concerns by analysing how power is produced and negotiated in practice.
Authority in Interaction
Due to their focus on conversational practices and actions rather than broader social discourses, CA researchers tend to refer to authority - rather than power - and specifically deontic authority, or the ability to determine the future actions of others in a particular
218
domain (Stevanovic & Peräkylä, 2012). At an interactional level of description, differences in power and authority do not exist outside of language, but are encoded, recycled and reaffirmed though conversation (H. Anderson & Goolishian, 1988; Stevanovic, 2013b). In essence, power relations and authority are talked into existence, and they have a persisting reality via this medium of reproduction. CA research addresses authority (and power) through examining how it is displayed and responded to in interaction. This means that CA research does not presuppose the overarching concepts of power in social discourses and institutions as described in more conceptual critiques of family therapy. CA instead explores how power structures are grounded in the orientations of participants through their interactions.
Asymmetrical power relations are not only present in the ways that people design their turns, they are also embedded in the normative structures of conversation. The normative structures of conversation guide the systematic organisation of talk (for example through turn-taking structures, adjacency pairs, organisation of repair) as well as a moral-inferential framework, through which situated identities are realised in ways relevant for the culture. These asymmetries activate different participation frameworks and identities (Goodwin & Heritage,
1990) and are present due to the sequential or enchronic nature of conversation.
Enchrony (Enfield, 2011, 2013) is a meta-concept that subsumes the already mentioned CA concepts of sequence, conditional relevance, and progressivity in conversation
(see Chapter 3). Enchrony is the driving force that progresses conversation through sequentially organised turns that are causally connected. Initiating turns can be graded in their effectiveness in bringing about a relevant response, and responsive turns vary in the appropriateness of how they fit with the prior initiating turn. Following the principle of enchrony, the responsive turn then provides a context through which subsequent turns are understood, forming a chain of interlocking turns that are understood in relation to each other and their sequential placement. Initiating and responsive turns, also known as adjacency pairs 219
(Chapter 2), are therefore indivisible as they are mutually defined. As initiating turns set up an expectation that the recipient will produce an appropriately fitted responsive turn, they also implicate deontic rights (Stevanovic & Peräkylä, 2012). Departures from these expectations gain attention and the person not producing the expected turn is vulnerable to inferences about failure to comply, including censure. So, while initiating and responsive turns are mutually defined, they are deontically asymmetrical because the initiating turn makes relevant a particular type of response from the recipient who can be held accountable for not producing that response. As demonstrated in Chapters 4, 5 and 6, Open Dialogue therapists are, at least implicitly, sensitive to this deontic asymmetry given that they downgrade their deontic authority when proposing transitions to reflections, when eliciting stance positions, and when repeating the words of the client, respectively.
The sequential enchronic structures of conversation also creates asymmetric relations in epistemic authority. Heritage and Raymond (2005) found that unmarked assessments in initiating positions assume primary rights to knowledge. This is because saying something first implies an independent assessment (and rights to do so), while the responding turn is unavoidably heard as responsive to the first saying. If people find grounds to resist these implicit associations with sequential positioning, then they may include additional elements in their turn by either downgrading the authority of their initiating assessment, or upgrading their authority when speaking in second position. The importance of epistemic rights lies in their connection to a sense of self, value, and personal identity (Heritage, 2018). This is also true of deontic authority, in both everyday and institutional conversations, and the various ways that people defend their deontic rights (Stevanovic, 2013b; also see Chapter 3 on
Methods). The sequential structure of conversation thus creates an asymmetry in knowledge because saying something first is connected to having greater epistemic authority. Through various means people can design their talk in ways to modify these default implications 220
(Heritage & Raymond, 2005). These can include downgraded first assessments, and various forms of upgraded second assessments including ways of reclaiming a first position slot
(Heritage & Raymond, 2005). In the study on therapist reflections (Chapter 7), I describe a similar practice where therapists’ responses deviated from normative expectations by eliding agreement to first position assertions. Therapists thus claimed independent access and authority to comment on the family, and also enabled therapists to present a wider variety of perspectives for families to consider.
In this section, I have presented how the enchronic nature of conversation promotes an asymmetrical distribution of deontic and epistemic rights and authority based on who is speaking in initiating positions (Enfield, 2011, 2013). This thesis demonstrates that despite the implicit asymmetries present in both conversational structures as well as institutional roles, therapists can design their talk in ways that downgrade their epistemic and deontic authority in deference to the epistemic and deontic rights of the family, and design their initiating turns to promote a wider range of possible responses from families. In the next section, I present the concept of agency and how it can serve as a unifying concept to account for the various forms of epistemic and deontic downgrading present in the therapist conversations in this thesis.
Agency and Asymmetrical Relations
The therapists’ downgrading of their authority and promotion of flexibility in recipient responses can be understood under the broader concept of agency. Mackrill (2009) describes different ways that client agency has been constructed in psychotherapy research i.e., in relation to processes or stages of change, as an aspect of diagnosis or personality type, as external variables, as how clients participate in therapy sessions, as narratives through the life course, and across various situational contexts. The conceptualisation of agency can take
221
a wide variety of forms ranging from theoretically derived concepts to more interactional descriptions without a single unifying definition nor a detailed account of how it operates on a micro level.
The work of Enfield (Enfield, 2011, 2013; Enfield & Kockelman, 2017) takes a different approach and describes agency from the perspective of conversation analysis and semiotics. This conceptualisation of agency focuses on interaction and how people’s actions are represented moment to moment. It is therefore sensitive to fluctuations in a person’s displayed actions and how agency is responded to by others. An agent is an entity that enables or constrains the conduct of others (Enfield & Kockelman, 2017). Agents can therefore take different forms such as a person, process, event, institution, or even theoretical models of therapy, such as Open Dialogue. In psychotherapy, therapists are enabled or constrained by their models and work contexts, and in turn act as agents that enable or constrain the responses of their clients.
Agency consists of two main dimensions: flexibility and accountability (Enfield,
2013; Enfield & Kockelman, 2017). Flexibility involves a person’s production of an action and includes three elements: controlling, which is determining where and when a particular behaviour will occur; composing, the selection of the type and form of behaviour including verbal and embodied actions; and subprehending, or the anticipation and commitment to how that behaviour will be interpreted and responded to by others, or why one does something and the anticipated effects that may arise from it. The element of controlling connects with the position of a turn in a sequence of talk, while the element of composing relates to turn design and whether it is unmarked or contains downgrading or upgrading components. Through flexibility, agents compose their actions in ways that can either downgrade or upgrade their epistemic and deontic authority, which in turn makes claims about their relative epistemic and deontic entitlements. Flexibility thus connects with the enchronic nature of talk. 222
The second dimension of agency is accountability and includes the agent’s behaviour being subject to evaluation by others; the entitlement or rights of the agent to produce that behaviour, and the obligation, duty or responsibility of the agent to produce that behaviour.
Agency is a way that accountability is distributed, because agents who have greater control over their actions also have greater accountability. This again connects with enchrony and conditional relevance because particular initiating actions claim certain deontic or epistemic entitlements, and recipients have the obligation to respond to these initiating turns in certain ways. In our usual ways of thinking, the elements of flexibility i.e., the control, composition, and subprehension/commitment all reside within one agent. However, these elements can be split and distributed amongst different individuals (Enfield, 2017), such as when someone delivers a speech that has been written by someone else. The speech is composed by a speech writer, while the speaker is responsible for the control or production of the speech as well as the accountability for the speech.
To summarise, the enchronic nature of human communication promotes asymmetry and accountability between interactants. Talk is sequentially ordered so that initiating actions claim epistemic and deontic authority and make relevant certain types of responses.
Recipients are then obligated to produce a fitted response, and are therefore accountable and vulnerable to possible censure if the expected response is not present. Initiating turns, including assessments, involve one person “speaking first” thus implicating an independence of thought and epistemic authority. The concept of agency combines these processes. Agency involves flexibility through controlling, composing and subprehending the effects of a behaviour. An instance of control and composition of a behaviour is a turn produced in conversation (i.e., in an initiating or responsive position), and the design of that turn (e.g., the presence of unmarked or downgraded authority). Control and composition thus relate to issues of asymmetry in authority and accountability in conversation. Agency also involves 223
the obligation and entitlement to produce behaviours as well as the subsequent evaluation from others. This again accounts for the asymmetrical differences in authority and accountability.
The components of agency i.e., flexibility and accountability, are a useful lens through which to conceptualise participation in Open Dialogue interactions. In Chapter 4, therapists utilised various means to downgrade the deontic authority in their proposals to transition to a reflection. These downgraded proposals provide a responsive framework where recipients have flexibility over the control and composition of their responses.
Subsequently, recipients are not obligated to produce a particular type of response. In Chapter
5, therapists utilised similar conversational practices to select next speakers and to elicit stance positions. Therapists exercised some control over recipient responses by selecting next speakers and nominating topics, however their turn design promoted flexibility in the composing of the recipients’ responses. In Chapter 6, therapists utilised repeats, which served multiple possible functions including initiating repair, marking receipt or requesting elaboration. Therefore, when responding to a repeat of a family member’s prior talk, the recipient can treat the repeat in a variety of ways while still composing a fitted response. In
Chapter 7, therapists elided agreement in reflections, contrary to normative expectations, in order to provide more flexible responses on a wider range of topics. Therefore, the concept of agency and particularly the promotion of flexibility in recipient responses appears to be of primary importance to Open Dialogue clinicians. By producing initiating turns that do not make strongly relevant a narrowly prescribed response type, recipients have greater flexibility in composing a range of fitted responses and controlling whether a response is even produced. In addition, because recipients may produce or elect not to produce a range of conditionally relevant responses, they can design their turns to mitigate their accountability.
That is, they are not obligated to produce certain types of responses and can thus design their 224
talk to mitigate anticipated evaluations. The concept of agency as presented by Enfield provides a way of describing the work of Open Dialogue therapists that is empirical and grounded in observable interactions and can provide a means of operationalising the theoretical concepts described by the originators of Open Dialogue. Agency also provides a way of describing the research in this thesis at a level removed from the fine detail of CA, making these findings potentially more accessible to practicing therapists. In the next section,
I focus on how these concepts can be applied to Open Dialogue and family therapy practice.
Implications for Open Dialogue and Family Therapy
The concept of agency has particular implications for Open Dialogue and collaborative approaches to family therapy generally. When therapists present a downgraded stance position and do not constrict response types, families have greater flexibility over the control and composition of their turns at talk. With initiating turns that promote such flexibility in responding, recipients are not obligated to produce certain limited types of responses. Flexibility also reduces accountability in relation to the distribution of agency between multiple individuals (Enfield, 2013). For example, if a therapist asks a question about a sensitive topic, they subprehend possible fitted responses that the recipient is accountable, through obligation, to produce. But if the answer to such a sensitive topic may cause some hurt or division in the family, then the therapist will have to some extent controlled the occurrence of the talk. Although the therapist was responsible for controlling the talk, it is the speaker or composer of the turn who is more likely to be held accountable for that talk. Agency is thus distributed between the therapist and the speaker, however the accountability is not equally distributed. The recipient may resist the constraints of the question however they then risk being held accountable for not producing a fitted response.
By designing initiating turns that promote flexibility, therapists can remove the obligation
225
and accountability for recipients to respond in prescribed ways. The ways that therapists design and present their initiating turns therefore have consequences for recipients in family therapy settings, particularly if their initiating turns are direct questions on sensitive topics and the recipient is strongly obligated to answer.
Promoting flexible responses in Open Dialogue and family therapy amounts to a commitment to certain configurations of agency where therapists provide family members with the opportunity to speak, as well as the flexibility to control and compose how they speak to mitigate their subprehended accountability. This means that the therapist may not receive the types of responses or information that is necessary for other aspects of family therapy such as hypothesis testing or an analysis of sequences. Instead, the therapist is re- positioned or de-centred from their role as an expert to that of a moderator where they work on selecting speakers and eliciting flexible responses. This moves authority and responsibility away from the therapist and towards the family, where the therapist does not provide advice for the family nor elicits certain types of information for strategic purposes, but re-focusses this work back towards the family.
Despite this downgrading, therapists’ authority is not completely annulled. In fact, therapists utilised deontic authority to direct the session (Chapter 4), and selected topics and speakers (Chapters 5 & 6). By directing the conversation, the therapist and the family can explore topics that may be important and not otherwise discussed and allow space for quieter family members to be heard. Similar to the position of Guilfoyle (2003), authority is not antithetical to dialogue but may in fact be necessary for dialogue to occur. Therapists can thus work towards a balance in their authority by guiding the conversation towards areas that they think are therapeutically indicated and by selecting speakers, while also being mindful to not direct the family to respond in a narrowly determined way.
226
These ways of speaking by therapists align with the professional stocks of interactional knowledge (Peräkylä & Vehviläinen, 2003) of Open Dialogue, formally located in the many recommendations for conducting Open Dialogue sessions. These stocks of interactional knowledge are to varying degrees normative descriptions of how therapeutic practice should be conducted. Theoretical writings on Open Dialogue regularly advise therapists to follow the narrative of the family members rather than leading the conversation
(H. Anderson, 2002; Bertrando & Lini, 2019; Haarakangas, Seikkula, Alakare, & Aaltonen,
2007; Olson, Seikkula, & Ziedonis, 2014; Seikkula, 2008, 2011; Seikkula & Arnkil, 2006;
Seikkula & Trimble, 2005). Therapists are also advised to be open to understanding from the perspective of the client and not to “monologue” them with their own perspective (H.
Anderson, 2002, 2005; H. Anderson & Goolishian, 1992; Galbusera & Kyselo, 2018;
Haarakangas, Seikkula, Alakare, & Aaltonen, 2007; Olson, Seikkula, & Ziedonis, 2014;
Seikkula, 2002; Seikkula, Laitila, & Rober, 2012; Seikkula & Trimble, 2005). The design of therapists’ turns to mitigate the obligation and accountability of client responses and the promotion of flexible composing of responses from families can be seen as practice manifestations of the recommendations to avoid monologue and follow the lead of the client.
These studies demonstrate how promoting such flexibility requires particular forms of turn design.
Contrasting these conversational findings with theoretical descriptions of Open
Dialogue highlights how therapeutic practice can be conceptualised at different levels of granularity. The theoretical perspective places emphasis on the therapist’s mindset and general recommendations for promoting idealised aspirational descriptions of collaboration and dialogue (Ong & Buus, in press). Other descriptions of Open Dialogue present the approach even more broadly as a “way of life” (Seikkula, 2011). In this view, Open Dialogue is connected to a broader sense of individual identity and a way that therapists may interact 227
with others in all aspects of life, not only in the therapeutic work setting. These views describe Open Dialogue at a coarser level of granularity. From a conversational perspective, the studies in this thesis focus on the more granular micro-details of conversation where descriptions of therapeutic practices remain at a fine level of specificity. These can be broadened out to the concept of agency but remain focused on forms of therapist talk that promote different agentic relations. This raises questions on how dialogical practices should be described and what level of detail is adequate for useful therapeutic interactions to occur.
There are of course attempts to clearly describe dialogical processes with more practical applications. For example, Rober (1999, 2005b) describes specific steps in determining whether to bring the therapist’s inner conversation into the session with clients.
Olson, Seikkula, and Ziedonis (2014) have attempted to describe twelve key elements of
Open Dialogue with more specificity than described in the seven principles. But, these key elements stem from a theoretical perspective that have not been operationalised or empirically validated (Waters et al., 2021). Furthermore, theoretical writings on Open
Dialogue also tend to conflate the transformative therapeutic uses of the term with the ontological perspective (Cooper, Chak, Cornish, & Gillespie, 2013; Ong & Buus, in press), which again promotes idealised versions of dialogue that are not necessarily consistent with therapeutic practice.
While these idealised overarching and aspirational theoretical descriptions from the originators of Open Dialogue (e.g., Haarakangas, Seikkula, Alakare, & Aaltonen, 2007;
Olson, Seikkula, & Ziedonis, 2014; Seikkula & Arnkil, 2006) may promote a wider range of applications, they are also ambiguous and imprecise which can be a challenge for new practitioners to apply. They also make the Open Dialogue approach difficult to define because a wide range of practices may fit with these more general sorts of descriptions. In current attempts to measure fidelity in Open Dialogue, there appears to be a pervasive pattern 228
whereby fidelity is considered to occur when therapists are appropriately oriented to the principles of Open Dialogue in their practice (Waters et al., 2021). Such an approach makes the assessment of fidelity subject to confirmation bias where fidelity is achieved when therapists are adequately trained and orient to the principles. This differs from standard approaches to measuring fidelity where independent assessors compare practices against standardised protocols (Waters et al., 2021). A conversational perspective may be able to address these issues. From a conversational perspective, Open Dialogue can be reduced to more easily described phenomena such as seeking stances from multiple parties, presenting downgraded deontic and epistemic stances, and question design promoting flexibility in responding. These more concrete elements of talk are perhaps more easily teachable to new practitioners and can readily be identified and measured by researchers. However, Open
Dialogue values an adaptability of the therapist to the developing conversation and a balance is likely necessary to avoid the risk of therapists delivering Open Dialogue in a mechanised fashion that only repeats certain forms of talk without due attention to the current demands of the conversation.
CA can also potentially operationalise aspects of Open Dialogue for research purposes. With a few exceptions (e.g., Seikkula et al., 2006; Seikkula, Alakare, & Aaltonen,
2011), the core principles of Open Dialogue have proved be difficult to research in traditional quantitative ways (Waters et al., 2021). Instead, there have been multiple qualitative studies mainly focussed on implementation and consumer and clinician experiences (Buus et al., in press). CA provides an alternative method that can potentially bridge these two types of research approaches. CA can empirically analyse real world conversations that are observable and identifiable and has an established research base on which to draw (see Chapter 3). It can also provide an outside perspective that is less susceptible to self-perpetuating ideas about
Open Dialogue. CA thus has the benefits of being complementary to Open Dialogue, through 229
a focus on interaction, while also having an established empirical basis. In addition, the research findings from this thesis can potentially inform Open Dialogue research by providing more detailed descriptions of dialogical practices that can be easier to identify and measure for researchers.
CA research has developed a unique way of analysing interaction based on how people display and negotiate social actions. Having developed from outside of psychotherapy, CA utilises different terms based on a different perspective on interaction.
Peräkylä and Vehviläinen (2003) propose that CA can play both a critical and complementary role that can promote a dialogue between CA research and professional stocks of interactional knowledge. CA can critique current practices by identifying, falsifying and correcting theoretical assumptions, and can complement existing theories by providing more detailed accounts of what therapists actually do and identify new dimensions of practice that have not previously been theorised. For example, due to a focus on displayed conversational practices, CA does not analyse or interpret behaviours through psychological motivations. In contrast, psychotherapy tends to involve a level of abstraction and interpretation of behaviours through psychological phenomena. CA can provide an outsider perspective that does not utilise the same types of descriptions as Open Dialogue theory, and as such can provide a unique perspective on the work of therapists that can challenge the core self- perpetuating beliefs from within the field. This is especially important in Open Dialogue, which is a relatively new approach that has so far centred around the work of a relatively small number of originators.
Limitations
As described in Chapter 3 and Ong, Barnes, and Buus (2020a), we developed a set of validity criteria for CA studies by combining the prior work of Peräkylä (2019); Pomerantz
230
and Fehr (2011); Potter (1996); Schegloff (1996); Sidnell (2010); Sidnell and Stivers (2013); ten Have (2007) and Wooffitt (2005). These criteria include: the use of video recordings of naturally occurring interactions; Jeffersonian transcriptions; presentation of adequate data extracts to support conclusions; analysis grounded in sequences and displayed orientations of the participants; and coherence of findings with past CA research. In the presented studies, I have generally met these goals by presenting multiple extracts following Jeffersonian transcription conventions and analysing sequences rather than individual turns at talk. Each article also contains reference to the connections and points of difference to prior research. In addition, I also sought to explicitly identify, acknowledge and exclude other possible interpretations of the data (Ong, Barnes, & Buus, 2020a). However, it has not always been possible to incorporate deviant case analysis as a validity measure. Deviant case analysis relies on having a relatively well defined conversational practice and a sufficiently large collection for comparing and contrasting these examples (ten Have, 2007). There are explicit uses of deviant case analysis in Chapter 4 (Ong, Barnes, & Buus, 2020c) and Chapter 5 (Ong,
Barnes, & Buus, 2021). But in Chapter 6 on therapist disagreement and Chapter 7 on the use of repeats, I was focussing on identifying and analysing the actions of particular practices rather than establishing normative orientations of the participants. In these chapters validity checks are provided through sequential analysis and the orientation of the participants, and coherence with past research.
Another potential limitation of this research comes not from the application of CA but the use of CA itself. CA’s focus on analysing observable conduct, grounded in the sense- making displayed by participants, is potentially both a positive and a negative attribute.
While it provides CA with greater analytic power by reducing interpretative accounts of behaviour, it may make CA findings somewhat difficult for therapists to accept. Firstly, a focus on behaviour can potentially be confounded with behaviourism. Behaviourism 231
(Watson, 1994) maintains a strict focus on behaviour with therapeutic interventions designed to change behavioural contingencies while disregarding internal cognitions. While many aspects of behavioural theory have been accepted and incorporated into approaches such as
Cognitive Behaviour Therapy, behaviourism was widely criticised as too narrow to adequately describe the range of human motivations (Schnaitter, 1999). While CA focuses on behaviour it does not disregard the internal orientations of people. The concepts of deontic authority and epistemic authority are examples where people are internally monitoring each other’s relative status in regard to these domains and designing their talk accordingly. CA therefore does not disregard the internal world of interactants, but these internal states must be displayed and observable through interactions and not based only on preconceived theoretical notations. This distinction may not be obvious to clinicians unfamiliar with CA resulting in the potential risk that CA is dismissed as Behaviourism.
A second problem in the acceptability of CA research is that family therapy and psychotherapy generally involve a degree of inference about behaviour. While psychotherapy models may describe behaviour, they interpret these behaviours as signs or symptoms of broader concepts and theoretical models. For example, a child’s lack of response to a question may be interpreted in various ways depending on the therapist’s model and understanding of the child’s past behaviours. The child may, for instance, be seen as anxious, depressed, oppositional, inattentive, or traumatised. The point is not which is the correct description but rather that there is a commonplace meta-level ascription of inferences abstracted from the observed behaviour in models of psychotherapy. While this practice may be necessary for psychotherapy models, this predilection may make CA studies seem incomplete to practicing clinicians due to a lack of second-order interpretation of internal states. In addition, CA research can at times be very detailed and technical. For example, in
Chapter 7, we described therapist repeats. This involved a technical discussion on how 232
repeats can be categorised and what actions they are doing from a CA perspective. Such points can be more meaningful for a researcher and have seemingly less relevance for practitioners. CA research may therefore fail to engage with practitioners as it may be too technical, or does not fit with the current ways of thinking in psychotherapy practice. As discussed in Chapter 3, another analytic approach such as Dialogical Methods of
Investigations of Happening of Change is closely connected to the terms used by Open
Dialogue. While there are potential drawbacks to using such an approach (see Chapter 3), the results may be more acceptable and recognisable to Open Dialogue clinicians. Similarly, other qualitative approaches such as thematic analysis can produce themes at a level of abstraction that is more readily acceptable to therapists.
Another limitation of this thesis is the ability to make firm recommendations for practice. CA research acknowledges that interaction is highly contextualised and the actions of various practices are dependent on the emerging sequence of interaction. Consequently, it is difficult to make recommendations for practice that are applicable to the wide range of situations experienced by therapists in family therapy sessions. However, there have been CA studies where findings can be directly applied by clinicians. For example, when general practitioners asked “are there some other concerns you’d like to address during this visit?” versus “are there any other concerns you’d like to address during this visit?” they reduced the chance of the patient leaving the appointment with an unmet concern (Heritage & Robinson,
2011). The change of one word using “some” instead of “any” had a significant impact on patient responses. Similarly, asking if someone was “willing” to attend mediation tended to result in greater acceptance than other forms (Stokoe, 2013). The findings of this thesis are not so conclusive; however, they do provide an insight into conversational practices that may aid in better understanding and researching Open Dialogue. However, future research can
233
cover a wider range of conversational practices and potentially make clearer recommendations for therapists.
Future Research
This thesis represents some of the first studies utilising CA to investigate Open
Dialogue. This developing area of research therefore has multiple opportunities for further investigations. Further CA research may reveal other therapist practices not currently described that can further the understanding of Open Dialogue practice. Further CA research could clarify and describe the essential conversational practices of most relevance to Open
Dialogue that can then be applied to the development of fidelity measures for future research.
Some potential areas of research include the use of silences by therapists, therapist practices that close down and open up topics, the overall structural organisation of Open Dialogue sessions, differences in the conversational practices of the family before and after therapist reflections, and relative speaking times and types of contributions of different parties. Open
Dialogue has also been introduced in a number of different areas around the world and future research could explore the various manifestations of the approach in these different settings.
234
Epilogue
This thesis has primarily been undertaken for personal reasons. Learning about family therapy and Open Dialogue has taught me to look at the world in different ways and has expanded my own personal perspective. Open Dialogue has inspired me to take on a research project that I had not previously seriously considered and has lead me to Conversation
Analysis which has in turn completely changed how I view human interaction. The most transformative part of this project for me has been looking at my profession from the perspective of CA. This outside perspective has been particularly helpful for me in understanding more generally the work that therapists do and how psychotherapy is arranged around certain theoretical perspectives. In closing, I thought it useful to again include my reflections on how this project has influenced and changed me personally. The follow article is currently under review with the Australian and New Zealand Journal of Family Therapy.
235
Interdisciplinary Reflections on Conversation Analysis, Power, and Open Dialogue
Throughout my career as a psychologist, there have been a number of times where what I was learning professionally significantly changed how I viewed the world personally.
One time was during my undergraduate psychology degree and early career training. I was learning about models of psychotherapy and various ways of conceptualising mental health problems and this changed the way I understood how people thought, felt and responded to difficulties in their lives. It provided me with concepts with which to categorise and order what I saw happening around me. Another, more revolutionary change for me was learning about family therapy. Family therapy took me from an internal, individualistic view to something broader and more relational. It is difficult for me to adequately describe the extent that these ideas changed how I thought about the world. Ideas such as circular causality, postmodernism, and dialogue were challenging, partly because they departed from the more individualistic and modernist ideas that pervade a lot of psychology training, and because they undermined certainty and made therapy and relationships much more messy and tentative. I became particularly interested in Open Dialogue, which seemed to move away from more authoritative descriptions of family interactions to one based more on the present moment and the various contrasting perspectives of everyone present. Personally and professionally I found it increasingly difficult to undertake prescriptive interventions or to see problems as located only in one person.
I am now in another stage of change. I am coming to the end of a 4-year PhD project researching Open Dialogue. Open Dialogue writings describe a particular way of communicating and interacting with families (Ong et al., 2019; Ong & Buus, in press). I wanted to know more about what made Open Dialogue conversations unique and what therapists actually did to make dialogue more likely to occur. I recorded 14 hours of Open
Dialogue sessions and used an approach called conversation analysis to explore the 236
conversational features of these sessions. Although it is often associated with linguistics, conversation analysis developed out of sociology and an approach called ethnomethodology
(Heritage, 2013b). I did not think it at the time, but viewing family therapy via a different discipline with a different set of assumptions and concepts, revealed a new perspective that has made me reflect differently on family therapy and the construction of psychotherapy generally. In this article, I will describe how this new perspective has shaped my thinking, practice, and approach to family therapy. I will therefore write in the first person to emphasise that these ideas reflect my own personal development and how they have changed the way that I view family therapy rather than focusing on the empirical findings of my research, which readers can find elsewhere (Ong, Barnes, & Buus, 2020a, 2020b, 2020c,
2021, in press). I will briefly say a little about conversation analysis before moving on to some of the more specific ideas that have been interesting to me.
Conversation analysis looks at how conversational interaction is structured. For example, there is a normative expectation that one person speaks at a time, there are mechanisms to select who is the recipient of talk and the projected next speaker (Sacks,
Schegloff, & Jefferson, 1974), and there are ways that people manage and negotiate problems in hearing and understanding (Hayashi, Raymond, & Sidnell, 2013). Conversation follows normative structures that we all implicitly orient to, although we may not explicitly know what these “rules” are. Conversation analysis focusses on the micro-details of what people say and how they say it (i.e., content, prosody and intonation) and how these practices achieve social actions. Conversation analysis differs from other forms of analysis because it maintains a focus on these minute details of interaction and how they are utilised, rather than studying conversational interaction to identify broader themes, the presence and effects of social discourses, or the connection to psychological phenomena and motivations like other analytic approaches such as thematic analysis, discourse analysis, and discursive psychology. 237
Conversation analysis has been applied to everyday conversations as well as “institutional” forms of talk like family therapy (for reviews see Ong, Barnes, & Buus, 2020a; Ong, Barnes,
& Buus, 2020b; Tseliou, 2013). Such research is still relatively new and has focused on issues such as how blame and accountability is discussed by parents (Hutchby & O’Reilly,
2010; Kiyimba & O'Reilly, 2015; O’Reilly, 2007, 2014; O’Reilly & Lester, 2016), how therapists negotiate impasse and engagement breaches (Couture, 2006; Muntigl & Horvath,
2016; Sutherland & Couture, 2007), and the use of reflections (Parker & O’Reilly, 2013;
Schriver, Buus, & Rossen, 2019; Williams & Auburn, 2015).
A revelatory idea for me from conversation analysis is that talk is a vehicle for action.
That is, we utilise various conversational practices in order to achieve social actions. For example, we have ritualised reciprocated greetings such as “hello, how are you?” that serve the action of opening conversations. We can do the action of requesting information through the conversational practice of interrogatives such as “where were you this morning?” or through a less direct my side telling such as “I was trying you all day and the line was busy for like hours” (Pomerantz, 1980, p. 186). Both of these conversational practices make relevant a response of providing information or an account of what the recipient was doing, but they go about achieving this action in different ways.
Conversation analysis has a unique way of looking at interactions. For me it was a difficult learning process as I had to look at interactions in a way that was very different from my usual perspective as a family therapist. But through this process of looking at therapy interactions from an external perspective, I ended up with some very different ideas on family therapy practice than what I was originally intending. Below I present three of these ideas that I summarise as: Interactions are sequential, Psychotherapy involves abstraction, and
Power is unavoidable and not inherently “bad”. I present these points as reflections on my own current understandings and how they have challenged and changed my own thinking. 238
While my research has focussed on Open Dialogue sessions, these ideas apply to my thinking about family therapy and psychotherapy more generally.
Interactions are Sequential
One finding from the conversation analytic research literature that has stood out to me is that conversation is organised around sequences of multiple turns. The core of the sequential organisation of conversation are “adjacency pairs” (Schegloff & Sacks, 1973).
These involve an initiating turn or first pair part (such as a question) that makes relevant the production of a responsive second pair part (such as an answer). First pair parts set up expectations on what will be produced next, and second pair parts reveal how the first pair part was received. They are thus indivisible pairs as they provide information on how each is interpreted. For example, if someone says, “What are you doing tonight?” this can be treated as an action that is seeking information through a response such as “Staying home and doing an assignment”, or it can be responded to as an invitation through a response of “Sorry, I’m busy”. First pair parts may project certain actions, but second pair parts confirm whether that action was received as such. The implication of this is that the intentions of the original speaker are not immediately relevant to the conversation as the response dictates how the initiating turn was received and how the conversation progresses. The original speaker may correct the second speaker, through a repair sequence (Hayashi, Raymond, & Sidnell, 2013), which would then display their original intentions. But these intentions are not necessarily present in the initiating turn nor relevant to the analysis. This was a very strange idea to me, I think because as a therapist I was so used to inferring people’s internal motivations for why they said or did something. But, if you consider that conversation analysis studies conversations and how they are constructed and not psychology it makes more sense.
The sequential organisation of conversation has changed the way that I think about family therapy. Firstly, like all interaction, therapy is a conversational achievement. This 239
means that interaction is not only a transfer of information from one person to another, that is, people do not construct fully formed ideas and then deliver them, via speech or embodied interaction, to another person, who in turn decodes and comprehends the message. Instead, conversation is produced, modified, and adapted though close attention to how the other person is responding. Goodwin (1979) has shown how people can modify their talk mid- utterance in response to differences in gaze from their recipients! Stivers (2005) has shown how doctors can change their prescribing due to certain responses of a child’s parent. The prescription of medication is not only a product of an objective examination of symptoms and diagnosis by a doctor, but is jointly constructed by the conversational participants, the doctor, patient, and family members.
If talk is a conversational achievement, jointly constructed, then therapist actions are only those actions they are intended to be if they are responded to and (in a sense) consummated by clients in a fitted way. For example, I may be attempting to “validate”,
“normalise” or “empathise” with a client through various conversational practices, but these are not unidirectional actions only on my part. My initiating action is only received as such if the client considers or responds to it as “validating”, “normalising”, or “empathic”. In other words, therapists do not simply “give” validation, or any other sort of action, it is only validation if it is received as validating. Actions are jointly constructed between therapist and client. Similarly, the words of the client should be considered in their sequential context and not in isolation; this means that what a client tells you is connected to how and what you said immediately prior. For example, a client may say that they are not suicidal, but if the preceding question was “You’re not going to do anything silly are you?” then this is not likely an accurate representation of their mental state. The client may instead be responding to the contingencies and expectations set up by the therapist’s question.
240
I found that the sequential nature of conversation changed how I was approaching my own research on Open Dialogue. Originally, I was interested in identifying what I was calling
“dialogical moments”. I wanted to identify these moments and see how they were created in order to give some advice to therapists on how to make these moments more likely to occur.
But because interactions are sequential and constructed turn by turn, I am now not even sure if such a thing can exist. I think that there may be “therapeutic moments” where clients feel a sense of being heard and understood but these occur only within sequences of interaction and are not restricted only to one moment in time. Furthermore, conversation analysis only works with observable data, making the study of these “moments” extremely difficult unless they are observed or communicated in some way (unless you believe in telepathic communication). Dialogue by its nature, cannot therefore be one moment in time, but rather something that is incrementally developed and jointly constructed, over time, through sequences of turns at talk.
Psychotherapy Involves Abstraction
When analysing therapy conversations, I was dismayed at how often I would unconsciously interpret the interactions through my pre-existing theoretical ideas about family therapy, even though I thought I was using a conversation analytic approach! It took consistent effort (and feedback from my supervisors) for me to look past theoretical models and concepts from Open Dialogue and family therapy. It was surprising how effortlessly I reverted back to my usual way of viewing and interpreting interactions with clients and how difficult it was to consider these interactions from a conversational perspective. It made me think that psychotherapy is rarely conceptualised or considered at a conversational level.
Instead, the interactions that occur within therapy are interpreted through theoretical models.
Models of psychotherapy regularly contain some explanation of how mental health or family problems develop and then propose a set of procedures or techniques on how to intervene in 241
those problems (T. Anderson, Lunnen, & Ogles, 2010). These models include a set of concepts or constructs, which aid clinicians in having a defined set of phenomena to look out for. Like Piaget’s (1999) concept of assimilation, behaviours of the client are understood and interpreted through reference to the therapist’s preferred theoretical model. Because interaction is complex and rapidly produced, sequences of talk are difficult to analyse in detail especially when conversation is produced in real-time. Theoretical models are helpful because they provide a way of parsing these complex behaviours and interactions into manageable categories. This selection, inference, and abstraction aids us in quickly making sense of complicated information. For me, it was surprising that this process was so pervasive and effortless, that I hardly noticed I was doing it! While there are benefits to models in psychotherapy, I think there is a potential problem if we accept our interpretations as facts rather than seeing them as inferences and interpretations that are abstracted from the actual behaviours of our clients.
A similar process was evident in the development of conversation analysis and linguistics. Prior work in linguistics studied idealised forms of grammatical sentences because everyday conversation, with its pauses, repeats, re-starts, “ums”, and half-finished sentences, was considered too disorderly for proper analysis (Heritage, 2013b). The actual ways that people spoke was disregarded because it did not fit with the predetermined theories about how people should talk! Later conversation analytic research revealed that these sorts of “randomizing factors” are in fact meaningful and ordered at a minute level of detail
(Heritage, 2013b, p. 235), but it was not until the development of ethnomethodology that researchers began to look at how people actually interacted. The revolutionary idea from ethnomethodology (Garfinkel, 1996), a precursor to conversation analysis, was that researchers did not need to develop predetermined theories about behaviour (ten Have, 2007).
Instead, how people made sense of the world was already present in their interactions. 242
Therefore, how people interacted and the methods through which they understood each other’s interactions, became an object of study in itself. Rather than interpreting behaviour through pre-defined categories, conversation analysis looks at how people display their understandings and orientations through their interactions.
When I reflect on psychotherapy, therapists seem to observe verbal and embodied behaviours and then make inferences about a person’s intentions and motivations according to their theoretical models of psychotherapy. These interpretations are potentially highly variable. For example, a client may not respond to a question in a therapy session and depending on the perspective of the therapist, this behaviour can be interpreted as
“depression”, “resistance”, a “personality disorder”, “stonewalling”, a “trauma response”, or
“under-functioning”. Psychotherapy thus involves abstraction, because our understandings about a client are removed from their actual displayed behaviours. Even social constructionist, narrative, and dialogical approaches, which denounce a modernist objective view of reality, still utilise their own sets of concepts for describing client problems and prescribe particular ways of speaking and interacting with clients. As argued by Larner
(1994, 2011), it is not desirable, nor even possible, to completely disregard a modernist paradigm in favour of a postmodern one. Like ethnomethodology and conversation analysis,
Open Dialogue aims to understand interactions through the sense making of the participants themselves, rather than through pre-determined theories of ideas. But, like all approaches to psychotherapy, Open Dialogue still utilises sets of concepts to describe and understand human difficulties. It is not that these ways of understanding are problematic in themselves, but I think that it is important to be aware that we are making these inferences and that they are theoretical constructs, so that unlike the linguists of the past, we don’t potentially miss the details that are present because we are too focussed on our pre-determined theories. Our theories are only useful if we can adapt our view and accommodate new information (Piaget, 243
1999), if we can hold them lightly and irreverently (Cecchin, Keeney, Lane, & Ray, 2018), and integrate and move between scientific-realist and social constructionist approaches through a paramodern stance (Larner, 1994, 2011).
Power is Unavoidable and Not Inherently “Bad”
Critiques around power and inequality have shaped and guided existing family therapy models and prompted the development of new approaches focussed on greater collaboration with clients and families (Flaskas, 2010). Power is often conflated with hierarchy, control and domination (Guilfoyle, 2003), making it a problem for collaborative- dialogical approaches which aim to promote equality (Olson, Seikkula, & Ziedonis, 2014) and avoid “colonizing” the narrative of the family (Rober & Seltzer, 2010). This is because power resides in institutions and social discourses, not in the intentions of individuals
(Guilfoyle, 2003; Hare-Mustin, 1994; White & Epston, 1990). So, even if therapists have good intentions or collaborative preferences, they cannot escape the power that is ascribed to their institutional role. This can sometimes make Open Dialogue therapists feel like they need to be passive and are not allowed to say anything in sessions (Ong et al., 2019).
Due to a focus on interaction, conversation analysis addresses power through the concept of deontic authority i.e., the ability to determine the future actions of others
(Stevanovic, 2013b, 2018; Stevanovic & Peräkylä, 2012, 2014). This type of authority is attached to the structure of conversation because initiating turns make relevant certain responses and therefore constrict the types of responses that a person can expectedly produce without sanction. Initiating actions therefore claim deontic authority. For example, by asking particular questions about, say, family relationships or past traumatic experiences, a therapist asserts and claims the right to ask those questions while the client and family do not have the rights to ask those same questions of the therapist. The conversational structures of initiating
244
and responsive turns thus encode an asymmetrical power relationship. So, while power is inextricably connected to institutions and roles, it is through conversational interactions that power is claimed, reproduced, re-affirmed, and manifested. As noted by Heritage (2013b, p.
290), “it is within these local sequences of talk, and only there, that these institutions are ultimately and accountably talked into being.” People may of course, resist these default positions. For example, parents may exert their authority to describe the relevant symptoms for their child and disagree with doctors’ treatment recommendations (Stivers, 2007) and therapists may downgrade their deontic authority (Ong, Barnes, & Buus, 2020c, 2021). But these examples are notable because they are departures from expectations (Potter, 1996).
This way of viewing authority has led me to question my role as a therapist. I have previously leaned towards a collaborative, social constructionist, dialogical perspective where power and authority seemed antithetical. I’ve come to think that an absolute equality of perspectives as is sometimes proposed by dialogical approaches, may constitute an ideal rather than an achievable reality. As described by Rober and De Haene (2017), a position of
“openness” and “hospitality” of the therapist also invokes conditionality and appropriation, with certain unavoidable expectations, obligations, inequality, and authority. I am now inclined to think that because asymmetries in power are engrained in social institutions, roles, and conversational structures, then instead of attempting to remove therapists’ authority, it is necessary to consider more deeply how that authority is utilised. My current thinking is that therapists should downgrade their authority in situations where it is important to elicit the multiple perspectives of the family. Such downgrades can include hesitation in talk such as pauses, re-starts, and words such as “I wonder” and “I think” (Couture, 2006; Ong, Barnes, &
Buus, 2021). In these situations, the family’s knowledge is prioritised. Therapists may introduce their perspectives but these should be downgraded and explicitly marked as coming from the therapist’s perspective and open to being challenged by the family. But, in situations 245
where a family member has not contributed to the conversation, where only one person is speaking, or there is an important issue that needs to be discussed, then the therapist can exercise some authority in inviting different people to speak or to ask questions that maintain the focus on a particular topic. This seems to suggest a form of therapy centred around the emerging conversation rather than more abstracted explanatory concepts. I acknowledge that this is not perfect and the therapist has discretion to decide if a person has spoken “enough” or what issues are “important” to focus on, and these decisions will likely be based on some form of theory or personal preference. I don’t think that this undermines the values of Open
Dialogue or other relational and collaborative approaches. But rather, these models can be expanded to acknowledge and incorporate how therapists flexibly adopt different positions of authority in therapy in response to the changing demands of the therapeutic environment to promote an ethical approach (Larner, 2011; Rober & De Haene, 2017).
Implications for Practice
I have been wondering if this different perspective has changed the way that I actually practice, and I think that there have been a few differences. I feel that I am less cautious about directing the session. I still feel that it is important to follow the lead of the client and the family and what they think is important to talk about. However, I am more comfortable directing the session, asking direct questions, and inviting quieter family members to speak. I have also become less interested and involved in discussions with colleagues about the family without the family present, and speculating about diagnoses or the motivations behind different interactions. I am interested in discussing with other therapists how we conducted ourselves in the session and if it could have been done differently, and at the same time I have less patience for discussions about people outside of the session due to the risk of reinforcing our own ideas rather than those presented by the family. I think that I am also more careful about how I ask questions in sessions. Not that my questions are more finely 246
structured, but rather I tend to stop, hesitate, repeat, apologise, hedge, explain the reason for questions, and explicitly give people the option not to answer. I think this is due to a cautiousness about introducing too much of my own ideas and potentially restricting how the family may respond.
Conclusion
In my PhD project, I made a conscious effort to strictly follow a conversation analytic process and see what it could reveal about Open Dialogue. This was very difficult. I found myself continually describing the conversation in terms derived from family therapy or Open
Dialogue principles rather than looking at the conversational structures and how the participants themselves were responding and orienting to the developing interaction. Like transitioning from an individual to a relational view, I found it difficult to transition from a therapeutic to an interactional perspective.
I was originally interested in knowing more about conversational processes and hopeful of finding ways that therapists can make “dialogical moments” more likely to occur.
I was hopeful of coming up with ideas on what therapists should and should not do in order to promote dialogue. And while I have a better understanding of some of the different things that therapists do, I have not been very successful in generating a clear list of “dos” and
“don’ts”. Instead, I think that the most significant outcome of this research has been my own personal development rather than an important contribution to the field of family therapy.
Considering Open Dialogue and family therapy from an interdisciplinary perspective has changed the way that I view the construct of therapy. I tend to look at therapist sessions as sequences of actions with people proposing, agreeing, resisting, challenging, downgrading, claiming and struggling for authority. This has been difficult to some extent as it leads me to question the process of therapy and the concepts that we use to describe it. Despite this discomfort, I feel that it is ultimately useful to look at family therapy beyond the discourses 247
of the family therapy community and to question my own assumptions. Ultimately, I think this questioning ensures that our practice remains relevant and adaptable to changes in society. Research on the minutiae of actual therapeutic practice, such as through conversation analysis, is one way of investigating this intersection between theory and practice. Such investigations are mutually beneficial as both theory and practice can be analysed, critiqued, modified, and extended in response to a greater awareness of the intricacies and complexities of the work that therapists actually do that is not captured in only theoretical descriptions
(Peräkylä & Vehviläinen, 2003).
Being able to work on a PhD is a privilege that few people have the opportunity to pursue. For me, it has been a rewarding experience but in ways that were unexpected and not directly related to the outcomes of the research. These insights have been unsettling and have made me question the therapeutic process, and how I view family therapy interactions. But, I hope that they have made me more attentive to what is happening for families rather than seeing them through my pre-conceived, and often theoretically-driven ideas about family functioning.
248
References
Albert, S., & de Ruiter, J. P. (2018). Repair: The interface between interaction and cognition.
Topics in Cognitive Science, 10(2), 279-313. doi:10.1111/tops.12339
Andersen, T. (1987). The reflecting team: Dialogue and meta-dialogue in clinical work.
Family Process, 26(4), 415-428. doi:10.1111/j.1545-5300.1987.00415.x
Andersen, T. (1991). The reflecting team: Dialogues and dialogues about the dialogues. New
York: Norton.
Anderson, H. (2001). Postmodern collaborative and person-centred therapies: What would
Carl Rogers say? Journal of Family Therapy, 23, 339-360.
Anderson, H. (2002). In the space between people: Seikkula’s open dialogue approach.
Journal of Marital and Family Therapy, 28(3), 279-281.
Anderson, H. (2005). Myths about “not-knowing”. Family Process, 44(4), 497-504.
Anderson, H. (2007). Dialogue: People creating meaning with each other and finding ways to
go on. In H. Anderson & D. R. Gehart (Eds.), Collaborative therapy: Relationships
and conversations that make a difference. New York: Routledge.
Anderson, H. (2012). Collaborative relationships and dialogic conversations: Ideas for a
relationally responsive practice. Family Process, 51(1), 8-24.
Anderson, H., & Goolishian, H. A. (1988). Human systems as linguistic systems: Preliminary
and evolving ideas about the implications for clinical theory. Family Process, 27(4),
371-393. doi:10.1111/j.1545-5300.1988.00371.x
Anderson, H., & Goolishian, H. A. (1992). The client is the expert: A not-knowing approach
to therapy. In S. McNamee & K. J. Gergan (Eds.), Therapy as social construction (pp.
25-39). London: Sage Publications.
Anderson, T., Lunnen, K. M., & Ogles, B. M. (2010). Putting models and techniques in
context. In B. L. Duncan, S. D. Miller, B. E. Wampold, & M. A. Hubble (Eds.), The 249
heart and soul of change: Delivering what works in therapy (pp. 143-166).
Washington DC: American Psychological Association.
Angouri, J. (2012). Managing disagreement in problem solving meeting talk. Journal of
Pragmatics, 44, 1656-1579.
Angouri, J., & Locher, M. A. (2012). Theorising disagreement. Journal of Pragmatics, 44,
1549-1553.
Antaki, C. (2008). Formulations in psychotherapy. In A. Peräkylä, C. Antaki, S. Vehviläinen,
& I. Leudar (Eds.), Conversation analysis and psychotherapy (pp. 26-42). New York:
Cambridge University Press.
Antaki, C., Barnes, R., & Leudar, I. (2005). Diagnostic formulations in psychotherapy.
Discourse Studies, 7(6), 627-647. doi:10.1177/1461445605055420
Antaki, C., & Kent, A. (2015). Offering alternatives as a way of issuing directives to
children: Putting the worse option last. Journal of Pragmatics, 78, 25.
Asmuß, B., & Oshima, S. (2012). Negotiation of entitlement in proposal sequences.
Discourse Studies, 14(1), 67-86. doi:10.1177/1461445611427215
Atkinson, J. M., & Drew, P. (1979). Order in court (1st ed. 1979. ed.). London: Palgrave
Macmillan UK.
Bakhtin, M. M. (1981). The dialogic imagination four essays. Austin: University of Texas
Press.
Bakhtin, M. M. (1984). Problems of Dostoevsky's poetics. Minneapolis: University of
Minnesota Press.
Bakhtin, M. M. (1986). Speech genres and other late essays. Austin: University of Texas
Press.
Bertrando, P. (2007). The dialogical therapist. London: Karnac.
250
Bertrando, P., & Lini, C. (2019). Theory and practice of systemic-dialogical therapy.
Australian and New Zealand Journal of Family Therapy, 40(2), 176-189.
doi:10.1002/anzf.1365
Blaxter, M. (1996). Criteria for the evaluation of qualitative research papers. Medical
Sociology News, 22(1), 68-71.
Blaxter, M. (2013). Criteria for the evaluation of qualitative research papers. Medical
Sociology online, 7(1), 4-7.
Bråten, S., & Trevarthen, C. (2007). Prologue: From infant intersubjectivity and participant
movements to simulation and conversation in cultural common sense. In S. Bråten
(Ed.), On being moved: From mirror neurons to empathy (pp. 21-34). Amsterdam:
John Benjamins Publishing Company.
Brown, J. M., & Mikes-Liu, K. (2015). Special issue: Dialogical practices. Australian and
New Zealand Journal of Family Therapy, 36(1), 1-5. doi:doi:10.1002/anzf.1101
Brown, P. (1998). Conversational structure and language acquisition: The role of repetition in
Tzeltal. Journal of Linguistic Anthropology, 8(2), 197-221.
doi:10.1525/jlin.1998.8.2.197
Button, G., & Sharrock, W. (2016). In support of conversation analysis’ radical agenda.
Discourse Studies, 18(5), 610-620. doi:10.1177/1461445616657955
Buus, N., Bikic, A., Jacobsen, E. K., Müller-Nielsen, K., Aagaard, J., & Rossen, C. B.
(2017). Adapting and implementing open dialogue in the Scandinavian countries: A
scoping review. Issues in Mental Health Nursing, 38(5), 1-11.
doi:10.1080/01612840.2016.1269377
Buus, N., Kragh Jacobsen, E., Bojesen, A. B., Bikic, A., Müller-Nielsen, K., Aagaard, J., &
Erlangsen, A. (2019). The association between Open Dialogue to young Danes in
acute psychiatric crisis and their use of health care and social services: A retrospective 251
register-based cohort study. International journal of nursing studies, 91, 119-127.
doi:10.1016/j.ijnurstu.2018.12.015
Buus, N., Ong, B., Einboden, R., Lennon, E., Mikes-Liu, K., Mayers, S., & McCloughen, A.
(in press). Implementing Open Dialogue approaches: A scoping review. Family
Process.
Cecchin, G. (1987). Hypothesizing, circularity, and neutrality revisited: An invitation to
curiosity. Family Process, 26(4), 405-413. doi:10.1111/j.1545-5300.1987.00405.x
Cecchin, G. (1992). Constructing therapeutic possibilities. In S. McNamee & K. J. Gergan
(Eds.), Therapy as social construction (pp. 86-95). Los Angeles: Sage.
Cecchin, G., Keeney, B. P., Lane, G., & Ray, W. A. (2018). Irreverence : A strategy for
therapists' survival. London: Taylor and Francis.
Cooper, M., Chak, A., Cornish, F., & Gillespie, A. (2013). Dialogue: Bridging personal,
community, and social transformation. Journal of Humanistic Psychology, 53(1), 70-
93.
Couture, S. J. (2005). Moving forward: Therapy with an adolescent and his family.
Dissertation Abstracts International: Section B: The Sciences and Engineering, 66(6-
B), 3401.
Couture, S. J. (2006). Transcending a differend: Studying therapeutic processes
conversationally. Contemporary Family Therapy: An International Journal, 28(3),
285-302. doi:http://dx.doi.org/10.1007/s10591-006-9011-1
Couture, S. J. (2007). Multiparty talk in family therapy: Complexity breeds opportunity.
Journal of Systemic Therapies, 26(1), 63-80. doi:10.1521/jsyt.2007.26.1.63
Couture, S. J., & Strong, T. (2004). Turning differences into possibilities: Using discourse
analysis to investigate change in therapy with adolescents and their families.
252
Counselling and Psychotherapy Research, 4(1), 90-101.
doi:10.1080/14733140412331384098
Couture, S. J., & Sutherland, O. (2006). Giving advice on advice-giving: A conversation
analysis of Karl Tomm's practice. Journal of Marital and Family Therapy, 32(3),
329-344. doi:http://dx.doi.org/10.1111/j.1752-0606.2006.tb01610.x
Craven, A., & Potter, J. (2010). Directives: Entitlement and contingency in action. Discourse
Studies, 12(4), 419-442. doi:10.1177/1461445610370126
Curl, T. S., & Drew, P. (2008). Contingency and action: A comparison of two forms of
requesting. Research on Language and Social Interaction, 41(2), 129-153.
doi:10.1080/08351810802028613
Dawson, L., River, J., McCloughen, A., & Buus, N. (2019). ‘Should it fit? Yes. Does it fit?
No’: Exploring the organisational processes of introducing a recovery-oriented
approach to mental health in Australian private health care. Health, Advance online
publication. doi:https://doi.org/10.1177/1363459319889107
Dawson, L., River, J., McCloughen, A., & Buus, N. (2020). ‘Every single minute and hour is
scrutinised’: Neoliberalism and Australian private mental health care. Sociology of
Health & Illness, 42(2), 277-292. doi:https://doi.org/10.1111/1467-9566.13009
Derish, P. A., & Annesley, T. M. (2011). How to write a rave review. Clinical Chemistry,
57(3), 388-391.
Drew, P. (2018a). Epistemics - The rebuttal special issue: An introduction. Discourse Studies,
20(1), 3-13.
Drew, P. (2018b). Epistemics in social interaction. Discourse Studies, 20(1), 163-187.
Drew, P., Walker, T., & Ogden, R. (2013). Self-repair and action construction. In G.
Raymond, M. Hayashi, & J. Sidnell (Eds.), Conversational repair and human
understanding (pp. 71-94). New York: Cambridge University Press. 253
Du Bois, J. W. (2007). The stance triangle. In R. Englebretson (Ed.), Stancetaking in
discourse: Subjectivity, evaluation, interaction (pp. 139-182). Amsterdam: John
Benjamins Publishing Company.
Du Bois, J. W. (2014). Towards a dialogic syntax. Cognitive Linguistics, 25(3), 359-410.
Einboden, R., Dawson, L., McCloughen, A., & Buus, N. (unpublished manuscript). Power,
position and social relations: Is the espoused absence of hierarchy in Open Dialogue
naïve?
Eiterå, A., Hansen, L., Vind, B., Hansen, A., Castella, J., & Sørensen, E. (2015). Open
dialogue - key markers and their context.
Ekberg, K., & LeCouteur, A. (2012). Negotiating behavioural change: therapists' proposal
turns in Cognitive Behavioural Therapy. Communication & Medicine, 9(3), 229.
Ekberg, K., & LeCouteur, A. (2014). Co-implicating and re-shaping clients' suggestions for
behavioural change in cognitive behavioural therapy practice. Qualitative Research in
Psychology, 11(1), 60-77. doi:10.1080/14780887.2012.709917
Ekberg, K., & LeCouteur, A. (2015). Clients' resistance to therapists' proposals: Managing
epistemic and deontic status. Journal of Pragmatics, 90, 12-25.
doi:http://dx.doi.org/10.1016/j.pragma.2015.10.004
Elliott, R., Fischer, C. T., & Rennie, D. L. (1999). Evolving guidelines for publication of
qualitative research studies in psychology and related fields. The British journal of
clinical psychology, 38(3), 215-229. doi:10.1348/014466599162782
Enfield, N. J. (2011). Sources of asymmetry in human interaction: Enchrony, status,
knowledge and agency. In J. Steensig, L. Mondada, & T. Stivers (Eds.), The Morality
of Knowledge in Conversation (pp. 285-312). Cambridge: Cambridge University
Press.
254
Enfield, N. J. (2013). Relationship thinking: Agency, enchrony, and human sociality. New
York: Oxford University Press.
Enfield, N. J. (2017). Distribution of agency. In N. J. Enfield & P. Kockelman (Eds.),
Distributed agency (pp. 9-14). New York, NY: Oxford University Press.
Enfield, N. J., & Kockelman, P. (2017). Distributed agency. New York, NY: Oxford
University Press.
Enfield, N. J., & Sidnell, J. (2017a). The concept of action. Cambridge: Cambridge
University Press.
Enfield, N. J., & Sidnell, J. (2017b). On the concept of action in the study of interaction.
Discourse Studies, 19(5), 515-535.
Everri, M., & Fruggeri, L. (2014). The individual-system relationship: methodological cues
from the stance-taking process analysis. Journal of Family Therapy, 36(3), 219-236.
doi:10.1111/1467-6427.12025
Everri, M., Fruggeri, L., & Molinari, L. (2014). Microtransitions and the dynamics of family
functioning. Integrative Psychological and Behavioral Science, 48(1), 61-78.
doi:10.1007/s12124-013-9248-9
Ferrari, R. (2015). Writing narrative style literature reviews. Medical Writing, 24(4), 230-
235. doi:DOI: 10.1179/2047480615Z.000000000329
Flaskas, C. (2010). Frameworks for practice in the systemic field: Part 1 - Continuities and
transitions in family therapy knowledge. Australian and New Zealand Journal of
Family Therapy, 31(3), 232-247. doi:10.1375/anft.31.3.232
Flaskas, C. (2011). Frameworks for practice in the systemic field: Part 2 - Contemporary
frameworks in family therapy. Australian and New Zealand Journal of Family
Therapy, 32(2), 87-108.
255
Foucault, M. (1979). Discipline and punish : The birth of the prison. New York: Vintage
Books.
Friedlander, M. L., Heatherington, L., & Marrs, A. L. (2000). Responding to blame in family
therapy: A constructionist/narrative perspective. The American Journal of Family
Therapy, 28(2), 133-146. doi:10.1080/019261800261716
Friedman, M. (1994). Reflections on the Buber-Rogers dialogue. Journal of Humanistic
Psychology, 34(1), 46-65.
Galbusera, L., & Kyselo, M. (2018). The difference that makes the difference: A conceptual
analysis of the open dialogue approach. Psychosis, 10(1), 47-54.
doi:10.1080/17522439.2017.1397734
Gardner, R. (2001). When listeners talk response tokens and listener stance.
Amsterdam/Philadelphia: John Benjamins Publishing Company.
Garfinkel, H. (1996). Ethnomethodology's program. Social Psychology Quarterly, 59(1), 5-
21.
Georgaca, E., & Avdi, E. (2009). Evaluating the talking cure: The contribution of narrative,
discourse, and conversation analysis to psychotherapy assessment. Qualitative
Research in Psychology, 6(3), 233-247.
doi:http://dx.doi.org/10.1080/14780880802146896
Goodwin, C. (1979). The interactive construction of a sentence in natural conversation. In G.
Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp. 97-121). New
York: Irvington Publishers.
Goodwin, C., & Heritage, J. (1990). Conversation analysis. Annual Review of Anthropology,
19(1), 283-307. doi:10.1146/annurev.an.19.100190.001435
256
Gordon, C., Gidugu, V., Rogers, E. S., DeRonck, J., & Ziedonis, D. (2016). Adapting open
dialogue for early-onset psychosis into the U.S. health care environment: A feasibility
study. Psychiatric Services, 67(11), 1166-1168. doi:10.1176/appi.ps.201600271
Green, B. N., Johnson, C. D., & Adams, A. (2006). Writing narrative literature reviews for
peer-reviewed journals: Secrets of the trade. Journal of Chiropractic Medicine, 5(3),
101-117. doi:10.1016/S0899-3467(07)60142-6
Gromer, J. (2012). Need-adapted and open-dialogue treatments: Empirically supported
psychosocial interventions for schizophrenia and other psychotic disorders. Ethical
Human Psychology and Psychiatry, 14(3), 163-177s.
doi:https://doi.org/10.1891/1559-4343.14.3.162
Guilfoyle, M. (2003). Dialogue and power: A critical analysis of power in dialogical therapy.
Family Process, 42(3), 331-343. doi:10.1111/j.1545-5300.2003.00331.x
Haarakangas, K., Seikkula, J., Alakare, B., & Aaltonen, J. (2007). Open dialogue: An
approach to psychotherapeutic treatment of psychosis in Northern Finland. In H.
Anderson & D. R. Gehart (Eds.), Collaborative therapy: Relationships and
conversations that make a difference (pp. 221-233). New York: Routledge.
Hare-Mustin, R. T. (1994). Discourses in the mirrored room: A postmodern analysis of
therapy. Family Process, 33(1), 19-35. doi:10.1111/j.1545-5300.1994.00019.x
Harter, S. P. (1986). Online information retrieval: concepts, principles, and techniques.
Orlando: Academic Press.
Hayashi, M., Raymond, G., & Sidnell, J. (2013). Conversational repair and human
understanding: An introduction. In G. Raymond, M. Hayashi, & J. Sidnell (Eds.),
Conversational repair and human understanding (pp. 1-40). New York: Cambridge
University Press.
257
Hayes, H. (1991). A re-introduction to family therapy clarification of three school. Australian
and New Zealand Journal of Family Therapy, 21(1), 27-43.
Hella, P., Niemi, J., Tirkkonen, J.-M., Hintikka, J., Koponen, H., & Koivumaa-Honkanen, H.
(2015). Talking about paranoid experiences: Interactional dynamics of accomplishing
disclosure of psychotic symptoms in triadic conversation. Communication &
Medicine, 12(1), 25-40. doi:http://dx.doi.org/10.1558/cam.v12i1.20214
Hepburn, A., & Bolden, G. B. (2013). The conversation analytic approach to transcription. In
J. Sidnell & T. Stivers (Eds.), The handbook of conversation analysis (pp. 57-76).
Chichester, UK: Wiley-Blackwell.
Hepburn, A., & Bolden, G. B. (2017). Transcribing for social research. Thousand Oaks, CA:
Sage Publications.
Heritage, J. (1984). A change-of-state token and aspects of its sequential placement. In J. M.
Atkinson & J. Heritage (Eds.), Structures of Social Action: Studies in Conversation
Analysis (pp. 299-345). Cambridge: Cambridge University Press.
Heritage, J. (1988). Explanations as accounts: A conversation analytic perspective. In C.
Antaki (Ed.), Analysing everyday explanation : A casebook of methods (pp. 127-144).
London: Sage Publications.
Heritage, J. (2013a). Action formation and its epistemic (and other) backgrounds. Discourse
Studies, 15(5), 551-578.
Heritage, J. (2013b). Conversation analysis. In Garfinkel and ethnomethodology (pp. 233-
292). Hoboken: Wiley.
Heritage, J. (2018). The ubiquity of epistemics: A rebuttal to the ‘epistemics of epistemics’
group. Discourse Studies, 20(1), 14-56.
258
Heritage, J., & Raymond, G. (2005). The terms of agreement: Indexing epistemic authority
and subordination in talk-in-interaction. Social Psychology Quarterly, 68(1), 15-38.
doi:10.1177/019027250506800103
Heritage, J., & Robinson, J. D. (2011). ‘Some versus ‘any’ medical issues: Encouraging
patients to reveal their unmet concerns. In C. Antaki (Ed.), Applied conversation
analysis. Palgrave Advances in Linguistics. London: Palgrave Macmillan.
Houtkoop-Steenstra, H. (1990). Accounting for proposals. Journal of Pragmatics, 14(1), 111-
124. doi:10.1016/0378-2166(90)90066-M
Houtkoop-Steenstra, H. (2000). Recipient design. In Interaction and the standardized survey
interview (pp. 62-87). ProQuest Ebook Central: Cambridge University Press.
Hutchby, I. (1992). Confrontation talk: Aspects of ‘interruption’ in argument sequences on
talk radio. Text & Talk, 12(3), 343-372. doi:10.1515/text.1.1992.12.3.343
Hutchby, I., & O’Reilly, M. (2010). Children's participation and the familial moral order in
family therapy. Discourse Studies, 12(1), 49-64. doi:10.1177/1461445609357406
Jacobson, N. S., & Gurman, A. S. (2015). Clinical handbook of couple therapy (5th ed.).
New York: Guilford Publications.
Jones, E. (1998). Family systems therapy: Developments in the Milan-systemic therapies.
Chichester: John Wiley & Sons.
Kaltenböck, G. (2010). Pragmatic functions of parenthetical I think. In G. Kaltenböck, W.
Mihatsch, & S. Schneider (Eds.), New approaches to hedging (pp. 237-266): Emerald
Group Publishing.
Kärkkäinen, E. (2006). Stance taking in conversation: From subjectivity to intersubjectivity.
Text & Talk, 26(6), 699-731.
259
Kärkkäinen, E. (2007). The role of I guess in conversational stancetaking. In R. Englebretson
(Ed.), Stancetaking in discourse: Subjectivity, evaluation, interaction (pp. 183-219):
John Benjamins Publishing Company.
Kaslow, F. W. (2007). Appendix D: A brief history of the field of family psychology and
therapy. In F. Shapiro, F. W. Kaslow, & L. Maxfield (Eds.), Handbook of EMDR and
family therapy processes: John Wiley & Sons.
Kent, A., & Kendrick, K. H. (2016). Imperative directives: Orientations to accountability.
Reseach on Language and Social Interaction, 49(3), 272-288.
Kerr, M. E., & Bowen, M. (1988). Family evaluation: An approach based on Bowen theory.
New York: Norton.
Kim, H. (2002). The form and function of next-turn repetition in English conversation.
Language Research, 38(1), 51-81.
Kiyimba, N., & O'Reilly, M. (2015). Parents' Resistance of Anticipated Blame through
Alignment Strategies: A Discursive Argument for Temporary Exclusion of Children
from Family Therapy. In The Palgrave Handbook of Child Mental Health (pp. 559-
577): Palgrave Macmillan.
Knol, A. S. L., Huiskes, M., Koole, T., Meganck, R., Loeys, T., & Desmet, M. (2020).
Reformulating and mirroring in psychotherapy: A conversation analytic perspective.
Frontiers in Psychology, 11. doi:10.3389/fpsyg.2020.00318
Kupetz, M. (2014). Empathy displays as interactional achievements-Multimodal and
sequential aspects. Journal of Pragmatics, 61, 4-34.
doi:10.1016/j.pragma.2013.11.006
Laitila, A., Aaltonen, J., Wahlström, J., & Angus, L. (2001). Narrative Process Coding
System in Marital and Family Therapy: An Intensive Case Analysis of the Formation
260
of a Therapeutic System. Contemporary Family Therapy, 23(3), 309-322.
doi:10.1023/A:1011183016456
Larner, G. (1994). Para-modern family therapy: Deconstructing post-modernism. Australian
and New Zealand Journal of Family Therapy, 15(1), 11-16.
Larner, G. (2011). Deconstructing theory: Towards an ethical therapy. Theory & Psychology,
21(6), 821-839. doi:10.1177/0959354310395061
Lebow, J. (2013). Editorial: Couple therapy and family therapy. Family Process, 52(1), 1-4.
doi:10.1111/famp.12027
Leiman, M. (2004). Dialogical sequence analysis. In H. J. M. Hermans & G. Dimaggio
(Eds.), The dialogical self in psychotherapy (pp. 255-269). East Sussex: Brunner-
Routledge.
Leiman, M., & Stiles, W. B. (2001). Dialogical sequence analysis and the zone of proximal
development as conceptual enhancements to the assimilation model: The case of Jan
revisited. Psychotherapy Research, 11(3), 311-330.
Levinson, S. C. (2013). Action formation and ascription. In J. Sidnell & T. Stivers (Eds.), The
handbook of conversation analysis (pp. 101-130). Chichester, UK: Wiley-Blackwell.
Logren, A., Ruusuvuori, J., & Laitinen, J. (2019). Peer responses to self-disclosures in group
counseling. Text & Talk, 39(5), 613-647. doi:10.1515/text-2019-2042
Lorås, L., Bertrando, P., & Ness, O. (2017). Researching systemic therapy history: In search
of a definition. Journal of Family Psychotherapy, 28(2), 134-149.
doi:10.1080/08975353.2017.1285656
Lynch, M. (2009). Response to Wes Sharrock. Text - Interdisciplinary Journal for the Study
of Discourse, 20(4), 541-544. doi:doi:10.1515/text.1.2000.20.4.541
Mackrill, T. (2009). Constructing client agency in psychotherapy research. The Journal of
humanistic psychology, 49(2), 193-206. doi:10.1177/0022167808319726 261
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). New York: Cambridge University Press.
Madill, A. (2015). Conversation analysis and psychotherapy process research. In O. C. G.
Gelo, A. Pritz, & B. Rieken (Eds.), Psychotherapy research: Foundations, process,
and outcome (pp. 501-515). New York, NY: Springer-Verlag Publishing; US.
Madill, A., Jordan, A., & Shirley, C. (2000). Objectivity and reliability in qualitative
analysis: Realist, contextualist and radical constructionist epistemologies. British
Journal of Psychology, 91(1), 1-20. doi:10.1348/000712600161646
Madill, A., Widdicombe, S., & Barkham, M. (2001). The potential of conversation analysis
for psychotherapy research. The Counseling Psychologist, 29(3), 413-434.
doi:10.1177/0011000001293006
Marra, M. (2012). Disagreeing without being disagreeable: Negotiating workplace
communities as an outsider. Journal of Pragmatics, 44, 1580-1590.
Maynard, D. W. (1997). The news delivery sequence: Bad news and good news in
conversational interaction. Research on Language and Social Interaction, 30(2), 93-
130. doi:10.1207/s15327973rlsi3002_1
Maynard, D. W., & Clayman, S. E. (1991). The diversity of ethnomethodology. Annual
Review of Sociology, 17(1), 385-418. doi:10.1146/annurev.so.17.080191.002125
Mikes-Liu, K., Goldfinch, M., MacDonald, C., & Ong, B. (2016). Reflective practice: An
exercise in exploring inner dialogue and vertical polyphony. Australian and New
Zealand Journal of Family Therapy, 37, 256-272. doi:10.1002/anzf.1166
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & The PRISMA Group. (2009). Preferred
reporting items for systematic reviews and meta-analyses: The PRISMA statement.
Public Library of Science Medicine, 6(7). doi:10.1371/journal.pmed1000097 262
Muntigl, P., & Horvath, A. O. (2014). The therapeutic relationship in action: How therapists
and clients co-manage relational disaffiliation. Psychotherapy Research, 24(3), 327-
345. doi:http://dx.doi.org/10.1080/10503307.2013.807525
Muntigl, P., & Horvath, A. O. (2016). A conversation analytic study of building and repairing
the alliance in family therapy. Journal of Family Therapy, 38(1), 102-119.
doi:http://dx.doi.org/10.1111/1467-6427.12109
Muntigl, P., & Horvath, A. O. (2020). Change in family therapy: Accomplishing authoritative
and moral positions through interaction. Communication & Medicine, 16(2), 169-183.
doi:10.1558/cam.34100
Muntigl, P., Knight, N., Watkins, A., Horvath, A. O., & Angus, L. (2013). Active retreating:
Person-centered practices to repair disaffiliation in therapy. Journal of Pragmatics,
53, 1-20. doi:http://dx.doi.org/10.1016/j.pragma.2013.03.019
Myers, G. (1998). Displaying opinions: Topics and disagreement in focus groups. Language
in Society, 27(1), 85-111. doi:10.1017/S0047404598001043
O'Reilly, M. (2015). ‘We're here to get you sorted’: parental perceptions of the purpose,
progression and outcomes of family therapy. Journal of Family Therapy, 37(3), 322-
342. doi:10.1111/1467-6427.12004
O'Reilly, M., Karim, K., Stafford, V., & Hutchby, I. (2015). Identifying the interactional
processes in the first assessments in child mental health. Child and Adolescent Mental
Health, 20(4), 195-201. doi:10.1111/camh.12077
O'Reilly, M., Kiyimba, N., & Lester, J. N. (2018). Discursive psychology as a method of
analysis for the study of couple and family therapy. Journal of Marital and Family
Therapy, 44(3), 409-425. doi:10.1111/jmft.12288
263
O’Reilly, M. (2005a). ‘Active noising’: The use of noises in talk, the case of onomatopoeia,
abstract sounds, and the functions they serve in therapy. Text, 25(6).
doi:10.1515/text.2005.25.6.745
O’Reilly, M. (2005b). The complaining client and the troubled therapist: A discursive
investigation of family therapy. Journal of Family Therapy, 27(4), 370-391.
doi:10.1111/j.1467-6427.2005.0328.x
O’Reilly, M. (2006). Should children be seen and not heard? An examination of how
children's interruptions are treated in family therapy. Discourse Studies, 8(4), 549-
566. doi:10.1177/1461445606064835
O’Reilly, M. (2007). Who's a naughty boy then? Accountability, family therapy, and the
"naughty" child. Research on Social Work Practice, 17(4), 234-243.
doi:10.1177/1066480707301316
O’Reilly, M. (2008). What value is there in children's talk? Investigating family therapists’
interruptions of parents and children during the therapeutic process. Journal of
Pragmatics, 40(3), 507-524. doi:http://dx.doi.org/10.1016/j.pragma.2007.12.005
O’Reilly, M. (2014). Blame and accountability in family therapy: Making sense of
therapeutic spaces discursively. Qualitative Psychology, 1(2), 163-177.
doi:10.1037/qup0000011
O’Reilly, M., & Lester, J. N. (2016). Building a case for good parenting in a family therapy
systemic environment: Resisting blame and accounting for children's behaviour.
Journal of Family Therapy, 38, 491-511.
O’Reilly, M., & Parker, N. (2013). 'You can take a horse to water but you can't make it
drink': Exploring children's engagement and resistance in family therapy.
Contemporary Family Therapy: An International Journal, 35(3), 491-507.
264
O’Reilly, M., & Parker, N. (2014). ‘She needs a smack in the gob’: negotiating what is
appropriate talk in front of children in family therapy. Journal of Family Therapy,
36(3), 287-307. doi:10.1111/j.1467-6427.2012.00595.x
Olson, M., Seikkula, J., & Ziedonis, D. (2014). The key elements of dialogic practice in open
dialogue. The University of Massachusetts Medical School. Retrieved from
http://umassmed.edu/psychiatry/globalinitiatives/opendialogue/
Ong, B., Barbara-May, R., Brown, J. M., Dawson, L., Gray, C., McCloughen, A., Mikes-Liu,
K., Sidis, A., Singh, R., Thorpe, C. R., & Buus, N. (2019). Open dialogue: Frequently
asked questions. Australian and New Zealand Journal of Family Therapy, 40(4), 416-
428. doi:10.1002/anzf.1387
Ong, B., Barnes, S., & Buus, N. (2020a). Conversation analysis and family therapy: A critical
review of methodology. Family Process, 59(2), 460-476. doi:10.1111/famp.12431
Ong, B., Barnes, S., & Buus, N. (2020b). Conversation analysis and family therapy: A
narrative review. Journal of Family Therapy, 42(2), 169-203. doi:10.1111/1467-
6427.12269
Ong, B., Barnes, S., & Buus, N. (2020c). Downgrading deontic authority in open dialogue
reflection proposals: A conversation analysis. Family Process, Advance online
publication, 1-16. doi:https://doi.org/10.1111/famp.12586
Ong, B., Barnes, S., & Buus, N. (2021). Eliciting stance and mitigating therapist authority in
Open Dialogue meetings. Journal of Marital and Family Therapy, 47(1), 120-135.
doi:10.1111/jmft.12454
Ong, B., Barnes, S., & Buus, N. (in press). Developing multiple perspectives by eliding
agreement: A conversation analysis of Open Dialogue reflections. Discourse Studies.
265
Ong, B., & Buus, N. (in press). What does it mean to work “dialogically” in Open Dialogue
and family therapy? A narrative review. Australian and New Zealand Journal of
Family Therapy.
Ong, B., & Buus, N. (under review). What Constitutes Therapeutic Dialogues? A Narrative
Review.
Parker, N., & O’Reilly, M. (2012). 'Gossiping' as a social action in family therapy: The
pseudo-absence and pseudo-presence of children. Discourse Studies, 14(4), 457-475.
Parker, N., & O’Reilly, M. (2013). Reflections from behind the screen: Avoiding therapeutic
rupture when utilizing reflecting teams. The Family Journal, 21(2), 170-179.
Peräkylä, A. (2012). Conversation analysis in psychotherapy. In J. Sidnell & T. Stivers
(Eds.), The handbook of conversation analysis (pp. 551-574). Chichester: Wiley
Blackwell.
Peräkylä, A. (2019). Conversation analysis and psychotherapy: Identifying transformative
sequences. Reseach on Language and Social Interaction, 52(3), 257-280.
Peräkylä, A., Antaki, C., Vehvilainen, S., & Leudar, I. (2008). Conversation analysis and
psychotherapy. Cambridge: Cambridge University Press.
Peräkylä, A., & Vehviläinen, S. (2003). Conversation analysis and the professional stocks of
interactional knowledge. Discourse & Society, 14(6), 727-750.
doi:10.1177/09579265030146003
Piaget, J. (1999). The child's conception of physical causality. London: Routledge.
Pillet-Shore, D. (2017). Preference organization. In J. Nussbaum (Ed.), Oxford research
encyclopedia of communication. New York: Oxford University Press.
Pilling, S. (2018). Development and evaluation of a social network intervention for severe
mental illness: The ODDESSI research programme. Paper presented at the
International Open Dialogue Research Collaboration (IODRC) Meeting, London. 266
Pomerantz, A. (1980). Telling my side: “Limited access” as a “fishing” device. Sociological
Inquiry, 50(3-4), 186-198.
Pomerantz, A. (1984). Agreeing and disagreeing with assessments: Some features of
preferred/dispreferred turn shapes. In J. M. Atkinson & J. Heritage (Eds.), Structures
of Social Action: Studies in Conversation Analysis (pp. 57-101). Cambridge:
Cambridge University Press.
Pomerantz, A., & Fehr, B. J. (2011). Conversation analysis: An approach to the analysis of
social interaction. In T. A. v. Dijk (Ed.), Discourse studies: A multidisciplinary
introduction (2nd ed., pp. 165-190). London: SAGE.
Pomerantz, A., & Heritage, J. (2013). Preference. In J. Sidnell & T. Stivers (Eds.), The
handbook of conversation analysis (pp. 210-228). Chichester: Wiley Blackwell.
Pote, H., Mazon, T., Clegg, J., & King, S. (2011). Vulnerability and protection talk: Systemic
therapy process with people with intellectual disability. Journal of Intellectual and
Developmental Disability, 36(2), 105-117.
doi:http://dx.doi.org/10.1080/13668250.2011.575771
Potter, J. (1996). Discourse analysis and constructionist approaches: Theoretical background.
In J. T. E. Richardson (Ed.), Handbook of qualitative research methods for
psychology and the social sciences (pp. 125-140). Leicester, England: BPS Books.
Potter, J. (2012). Discourse analysis and discursive psychology. In H. M. Cooper (Ed.), APA
handbook of research methods in psychology: Quantitative, qualitative,
neuropsychological, and biological (pp. 119-138). Washington, D.C: American
Psychological Association.
Preparing and delivering interpretations in psychoanalytic interaction. (2003). In Text -
Interdisciplinary Journal for the Study of Discourse (Vol. 23, pp. 573).
267
Psathas, G. (1995). Conversation analysis: The study of talk-in-interaction. Thousand Oaks,
Calif: SAGE.
Puchta, C., Potter, J., & Wolff, S. (2004). Repeat receipts: a device for generating visible data
in market research focus groups. Qualitative Research, 4(3), 285-309.
doi:10.1177/1468794104047231
Rae, J. (2008). Lexical substitution as a therapeutic resource. In A. Peräkylä, C. Antaki, S.
Vehviläinen, & I. Leudar (Eds.), Conversation analysis and psychotherapy (pp. 62-
79). New York: Cambridge University Press.
Raymond, G. (2003). Grammar and social organization: Yes/no interrogatives and the
structure of responding. American Sociological Review, 68(6), 939-967.
doi:10.2307/1519752
Raymond, G. (2018). Which epistemics? Whose conversation analysis? Discourse Studies,
20(1), 57-89.
Razzaque, R., & Wood, L. (2015). Open dialogue and its relevance to the NHS: Opinions of
NHS staff and service users. Community Mental Health Journal, 51(8), 931-938.
doi:10.1007/s10597-015-9849-5
Rober, P. (1999). The therapist’s inner conversation in family therapy practice: Some ideas
about the self of the therapist, therapeutic impasse, and the process of reflection.
Family Process, 38(2), 209-228.
Rober, P. (2005a). Family therapy as a dialogue of living persons: A perspective inspired by
Bakhtin, Voloshinov, and Shotter. Journal of Marital and Family Therapy, 31(4),
385-397. doi:https://doi.org/10.1111/j.1752-0606.2005.tb01578.x
Rober, P. (2005b). The therapist’s self in dialogical family therapy: Some ideas about not-
knowing and the therapist’s inner conversation. Family Process, 44(4), 477-495.
268
Rober, P. (2010). The interacting-reflecting training exercise: Addressing the therapist’s inner
conversation in family therapy training. Journal of Marital and Family Therapy,
36(2), 158-170. doi:10.1111/j.1752-0606.2010.00192.x
Rober, P. (2011). The therapist’s experiencing in family therapy practice. Journal of Family
Therapy, 33, 233-255. doi:10.1111/j.1467-6427.2010.00502.x
Rober, P., & De Haene, L. (2017). Hospitality in family therapy practice: A further
engagement with Jacques Derrida. Australian and New Zealand Journal of Family
Therapy, 38(3), 378-390. doi:10.1002/anzf.1254
Rober, P., & Seltzer, M. (2010). Avoiding colonizer positions in the therapy room: Some
ideas about the challenges of dealing with the dialectic of misery and resources in
families. Family Process, 49(1), 123-137.
Robinson, A. (2011). In theory Bakhtin: Dialogism, polyphony, and heteroglossia. Ceasefire
Magazine, https://ceasefiremagazine.co.uk/in-theory-bakhtin-1/.
Robinson, J. D. (2013a). Epistemics, action formation, and other-initiation of repair: The case
of partial questioning repeats. In G. Raymond, J. Sidnell, & M. Hayashi (Eds.),
Conversational Repair and Human Understanding (pp. 261-292). Cambridge:
Cambridge University Press.
Robinson, J. D. (2013b). Overall structural organisation. In J. Sidnell & T. Stivers (Eds.), The
handbook of conversation analysis (pp. 257-280). Chicester: Wiley.
Robinson, J. D., & Kevoe-Feldman, H. (2010). Using full repeats to initiate repair on others'
questions. Research on Language and Social Interaction, 43(3), 232-259.
doi:10.1080/08351813.2010.497990
Roca-Cuberes, C. (2016). Diagnosing as an interactional achievement in psychiatric
interviews. In The Palgrave handbook of adult mental health: Discourse and
conversation studies (pp. 191-208). New York, NY: Palgrave Macmillan. 269
Rogers, C. R. (1951). Client-centered therapy. London: Constable.
Rogers, C. R. (1967). On becoming a person : A therapist's view of psychotherapy. London:
Constable.
Rosen, K., & Stoklosa, J. (2016). Finland in Boston? Applying open dialogue ideals on a
psychotic disorders inpatient teaching unit. Psychiatric Services, 67(12), 1283-1285.
Rossen, C. B., Schriver, K. N., Tarber, C., Nordahl, D. V., Thygesen Rasmussen, G., Ong,
B., & Buus, N. (2020). “Y, what do you think about what X just said?” Conversation
analysis of stance-eliciting questions in open dialogue network meetings. Journal of
Marital and Family Therapy, 46(4), 719-731. doi:10.1111/jmft.12435
Russell, C. K., & Gregory, D. M. (2003). Evaluation of qualitative research studies.
Evidence-Based Nursing, 6(2), 36-40.
Sacks, H. (1987). On the preferences for agreement and contiguity in sequences in
conversation. In G. Button & J. R. E. Lee (Eds.), Talk and social organization (pp.
54-69). Clevedon: Multilingual Matters.
Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the
organization of turn-taking for conversation. Language, 50(4), 696-735.
doi:10.1353/lan.1974.0010
Schegloff, E. A. (1982). Discourse as an interactional achievement: Some uses of ‘uh huh’
and other things that come between sentences. In D. Tannen (Ed.), Analyzing
Discourse: Text and Talk (pp. 71-93). Washington D.C.: Georgetown University
Press.
Schegloff, E. A. (1996). Confirming allusions: Toward an empirical account of action.
American Journal of Sociology, 102(1), 161-216. doi:10.1086/230911
270
Schegloff, E. A. (1997). Practices and actions: Boundary cases of other-initiated repair.
Discourse Processes: Discourse Process after Two Decades, 23(3), 499-545.
doi:10.1080/01638539709545001
Schegloff, E. A. (2000). Overlapping talk and the organization of turn-taking for
conversation. Language in Society, 29(1), 1-63. doi:10.1017/S0047404500001019
Schegloff, E. A. (2007). Post-expansion. In Sequence Organization in Interaction: A Primer
in Conversation Analysis (pp. 115-168). Cambridge: Cambridge University Press.
Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction in the
organization of repair in conversation. Language, 53(2), 361. doi:10.2307/413107
Schegloff, E. A., & Sacks, H. (1973). Opening up closings. Semiotica, 8, 289-327.
Schnaitter, R. (1999). Some criticisms of behaviorism. In B. A. Thyer (Ed.), The
Philosophical Legacy of Behaviorism (pp. 209-249). Dordrecht: Springer
Netherlands.
Schriver, K. N., Buus, N., & Rossen, C. B. (2019). Reflective practices in open dialogue
meetings: Reporting and inferential ‘my side tellings’. Journal of Pragmatics, 146,
19-31. doi:10.1016/j.pragma.2019.03.007
Schubert, S., Rhodes, P., & Buus, N. (2020). Transformation of professional identity: An
exploration of psychologists and psychiatrists implementing open dialogue. Journal
of Family Therapy, Advanced online publication. doi:https://doi.org/10.1111/1467-
6427.12289
Seikkula, J. (2002). Open dialogues with good and poor outcomes for psychotic crises:
Examples from families with violence. Journal of Marital and Family Therapy, 28(3),
263-274. doi:10.1111/j.1752-0606.2002.tb01183.x
271
Seikkula, J. (2003). Open dialogue integrates individual and systemic approaches in serious
psychiatric crises. Smith College Studies in Social Work, 73(2), 227-245.
doi:10.1080/00377310309517683
Seikkula, J. (2008). Inner and outer voices in the present moment of family and network
therapy. Journal of Family Therapy, 30, 478-491.
Seikkula, J. (2011). Becoming dialogical: Psychotherapy or a way of life? Australian and
New Zealand Journal of Family Therapy, 32(3), 179-193.
Seikkula, J., Aaltonen, J., Alakare, B., Haarakangas, K., Keränen, J., & Lehtinen, K. (2006).
Five-year experience of first-episode nonaffective psychosis in open-dialogue
approach: Treatment principles, follow-up outcomes, and two case studies.
Psychotherapy Research, 16, 214-228.
Seikkula, J., Alakare, B., & Aaltonen, J. (2011). The comprehensive open-dialogue approach
in Western Lapland: II. Long-term stability of acute psychosis outcomes in advanced
community care. Psychosis, 3(3), 192-204. doi:10.1080/17522439.2011.595819
Seikkula, J., & Arnkil, T. E. (2006). Dialogical meetings in social networks. London: Karnac.
Seikkula, J., Laitila, A., & Rober, P. (2012). Making sense of multi-actor dialogues in family
therapy and network meetings. Journal of Marital and Family Therapy, 38(4), 667-
687. doi:10.1111/j.1752-0606.2011.00238.x
Seikkula, J., & Trimble, D. (2005). Healing elements of therapeutic conversation: Dialogue
as an embodiment of love. Family Process, 44(4), 461-475.
Selvini, M., Boscolo, L., Cecchin, G., & Prata, G. (1980). Hypothesizing-circularity-
neutrality: Three guidelines for the conductor of the session. Family Process, 19(1),
3-12. doi:10.1111/j.1545-5300.1980.00003.x
Seuren, L. M. (2018). Assessing answers: Action ascription in third position. Research on
Language and Social Interaction, 51(1), 33-51. doi:10.1080/08351813.2018.1413890 272
Shotter, J. (1996). ‘Now I can go on:’ Wittgenstein and our embodied embeddedness in the
‘hurly-burly’ of life. Human Studies, 19(4), 385-407. doi:10.1007/BF00188850
Shotter, J. (2006). Understanding process from within: An argument for “withness” thinking.
Organization Studies, 27, 585-604.
Sidnell, J. (2010). Conversation analysis: an introduction. Malden, Mass: Wiley-Blackwell.
Sidnell, J. (2013). Basic conversation analytic methods. In J. Sidnell & T. Stivers (Eds.), The
handbook of conversation analysis (pp. 77-99). Chicester: Wiley.
Sidnell, J., & Stivers, T. (2013). The handbook of conversation analysis. Chicester: Wiley.
Sifianou, M. (2012). Disagreements, face and politeness. Journal of Pragmatics, 44(12),
1554-1564. doi:http://dx.doi.org/10.1016/j.pragma.2012.03.009
Stamp, G. H. (1991). Family conversation: Description and interpretation. Family Process,
30(2), 251-263. doi:http://dx.doi.org/10.1111/j.1545-5300.1991.00251.x
Stancombe, J., & White, S. (2005). Cause and responsibility: towards an interactional
understanding of blaming and ‘neutrality’ in family therapy. Journal of Family
Therapy, 27(4), 330-351. doi:10.1111/j.1467-6427.2005.00326.x
Stevanovic, M. (2013a). Constructing a proposal as a thought: A way to manage problems in
the initiation of joint decision-making in Finnish workplace interaction. Pragmatics,
23(2), 519-544.
Stevanovic, M. (2013b). Deontic rights in interaction. A conversation analytic study on
authority and cooperation. (PhD thesis), University of Helsinki, Finland.
Stevanovic, M. (2018). Social deontics: A nano-level approach to human power play. Journal
for the Theory of Social Behaviour, 48(3), 369-389. doi:10.1111/jtsb.12175
273
Stevanovic, M., & Peräkylä, A. (2012). Deontic authority in interaction: The right to
announce, propose, and decide. Research on Language and Social Interaction, 45(3),
297-321. doi:10.1080/08351813.2012.699260
Stevanovic, M., & Peräkylä, A. (2014). Three orders in the organization of human action: On
the interface between knowledge, power, and emotion in interaction and social
relations. Language in Society, 43(2), 185-207. doi:10.1017/S0047404514000037
Stevanovic, M., & Svennevig, J. (2015). Introduction: Epistemics and deontics in
conversational directives. Journal of Pragmatics, 78, 1-6.
doi:http://dx.doi.org/10.1016/j.pragma.2015.01.008
Stiles, W. B., Leiman, M., Shapiro, D. A., Hardy, G. E., Barkham, M., Detert, N. B., &
Llewelyn, S. P. (2006). What does the first exchange tell? Dialogical sequence
analysis and assimilation in very brief therapy. Psychotherapy Research, 16(4), 408-
421. doi:10.1080/10503300500288829
Stivers, T. (2005). Parent resistance to physicians' treatment recommendations: One resource
for initiating a negotiation of the treatment decision. Health Communication, 18(1),
41-74. doi:10.1207/s15327027hc1801_3
Stivers, T. (2007). Prescribing under pressure parent-physician conversations and
antibiotics. Oxford: Oxford University Press.
Stivers, T. (2008). Stance, alignment, and affiliation during storytelling: When nodding is a
token of affiliation. Research on Language and Social Interaction, 41(1), 31-57.
doi:10.1080/08351810701691123
Stivers, T. (2013). Sequence organization. In J. Sidnell & T. Stivers (Eds.), The handbook of
conversation analysis. Chicester: Wiley.
274
Stivers, T., & Enfield, N. J. (2010). A coding scheme for question–response sequences in
conversation. Journal of Pragmatics, 42(10), 2620-2626.
doi:10.1016/j.pragma.2010.04.002
Stivers, T., & Robinson, J. D. (2006). A preference for progressivity in interaction. Language
in Society, 35(3), 367-392. doi:10.1017/S0047404506060179
Stivers, T., & Rossano, F. (2010). Mobilizing response. Research on Language & Social
Interaction, 43(1), 3-31. doi:10.1080/08351810903471258
Stokoe, E. (2013). Overcoming barriers to mediation in intake calls to services: Research-
based strategies for mediators. Negotiation journal, 29(3), 289-314.
doi:10.1111/nejo.12026
Stokoe, E. (2014). The conversation analytic role-play method (CARM): A method for
training communication skills as an alternative to simulated role-play. Research on
Language and Social Interaction, 47(3), 255-265.
doi:10.1080/08351813.2014.925663
Suoninen, E., & Wahlstrom, J. (2009). Interactional positions and the production of
identities: Negotiating fatherhood in family therapy talk. Communication & Medicine,
6(2), 199-209.
Sutherland, O., & Couture, S. (2007). The discursive performance of the alliance in family
therapy: A conversation analytic perspective. Australian and New Zealand Journal of
Family Therapy, 28(4), 210-217. doi:http://dx.doi.org/10.1375/anft.28.4.210
Svennevig, J., & Djordjilovic, O. (2015). Accounting for the right to assign a task in meeting
interaction. Journal of Pragmatics, 78, 98-111.
doi:https://doi.org/10.1016/j.pragma.2014.12.007
275
ten Have, P. (2004). Ethnomethodology. In S. Clive, G. Giampietro, G. Jaber, & S. David
(Eds.), Qualitative research practice (pp. 139-152). London: SAGE Publications Ltd. ten Have, P. (2007). Doing conversation analysis (2nd ed.). London: SAGE. ten Have, P. (2012). Ethnomethodology and conversation analysis. In H. M. Cooper (Ed.),
APA handbook of research methods in psychology: Quantitative, qualitative,
neuropsychological, and biological (Vol. 2, pp. 103-117). Washington, D.C:
American Psychological Association.
Thompson, S. A., Fox, B., & Couper-Kuhlen, E. (2015). Grammar in everyday talk: Building
responsive actions. Cambridge: Cambridge University Press.
Tomm, K. (1984). One perspective on the Milan systemic approach: Part II. Description of
session format, interviewing style and interventions. Journal of Marital and Family
Therapy, 10(3), 253-271. doi:10.1111/j.1752-0606.1984.tb00016.x
Tomm, K. (1988). Interventive interviewing: Part III. Intending to ask lineal, circular,
strategic, or reflexive questions? Family Process, 27(1), 1-15. doi:10.1111/j.1545-
5300.1988.00001.x
Trevarthen, C. (1990). Signs before speech. In T. A. Sebeok & J. Umiker-Sebeok (Eds.), The
semiotic web 1989 (pp. 689-756). Berlin; Boston: De Gruyter Mouton.
Tribe, R. H., Freeman, A. M., Livingstone, S., Stott, J. C. H., & Pilling, S. (2019). Open
dialogue in the UK: Qualitative study. BJPsych open, 5(4), e49-e49.
doi:10.1192/bjo.2019.38
Tseliou, E. (2013). A critical methodological review of discourse and conversation analysis
studies of family therapy. Family Process, 52(4), 653-672.
doi:http://dx.doi.org/10.1111/famp.12043
276
Tseliou, E., & Borcsa, M. (2018). Discursive methodologies for couple and family therapy
research: Editorial to special section. Journal of Marital and Family Therapy, 44(3),
375-385.
Viaro, M., & Leonardi, P. (1983). Getting and giving information: Analysis of a family-
interview strategy. Family Process, 22(1), 27-42. doi:10.1111/j.1545-
5300.1983.00027.x
Viklund, E., Holmqvist, R., & Zetterqvist Nelson, K. (2010). Client-identified important
events in psychotherapy: Interactional structures and practices. Psychotherapy
Research, 20(2), 151-164. doi:10.1080/10503300903170939
Voloshinov, V. N. (1972). Marxism and the philosophy of language (2nd ed.). The Hague:
Mouton.
Walker, T., & Benjamin, T. (2017). Phonetic and sequential differences of other-repetitions
in repair initiation. Research on Language and Social Interaction, 50(4), 330-347.
doi:10.1080/08351813.2017.1340717
Waters, E., Ong, B., Mikes-Liu, K., McCloughen, A., Rosen, A., Mayers, S., Sidis, A.,
Dawson, L., & Buus, N. (2021). Open Dialogue, need-adapted mental health care, and
implementation fidelity: A discussion paper. International Journal of Mental Health
Nursing, 30(3), 805-813. doi:https://doi.org/10.1111/inm.12866
Watson, J. B. (1994). Psychology as the behaviorist views it. Psychological Review, 101(2),
248-253. doi:10.1037//0033-295X.101.2.248
Weatherall, A. (2011). I don't know as a prepositioned epistemic hedge. Research on
Language and Social Interaction, 44(4), 317-337.
doi:10.1080/08351813.2011.619310
Weingarten, K. (2016). The art of reflection: Turning the strange into the familiar. Family
Process, 55(2), 195-210. doi:10.111/famp.12158 277
Weiste, E. (2015). Describing therapeutic projects across sequences: Balancing between
supportive and disagreeing interventions. Journal of Pragmatics, 80, 22-43.
doi:http://dx.doi.org/10.1016/j.pragma.2015.02.001
Weiste, E., & Peräkylä, A. (2013). A comparative conversation analytic study of
formulations in psychoanalysis and cognitive psychotherapy. Research on Language
and Social Interaction, 46(4), 299-321. doi:10.1080/08351813.2013.839093
White, M. (1995). Reflecting teamwork as definitional ceremony. In Re-authoring lives:
Interviews and essays. Adelaide: Dulwich Centre Publications.
White, M., & Epston, D. (1990). Narrative means to therapeutic ends. New York: W.W.
Norton and Company.
Williams, L., & Auburn, T. (2015). Accessible polyvocality and paired talk: How family
therapists talk positive connotation into being. Journal of Family Therapy, 38(4), 535-
554.
Willig, C. (2013). Introducing Qualitative Research In Psychology (3rd ed.). Maidenhead:
McGraw-Hill Education.
Wooffitt, R. (2005). Conversation analysis and discourse analysis: A comparative and
critical introduction. London: SAGE.
278