Multiple emotions, multiple selves: compassion focused therapy chairwork

Abstract

Compassion focused therapy (CFT) is rooted in an evolutionary view of the human mind as formed of a multitude of contrasting, and often conflicting, motivations, emotions and competencies. A core aim of the therapy is to help clients understand the nature of their mind in a way that is de-pathologizing and de-shaming. The approach is also focused on the cultivation of compassion to work with these difficult aspects of mind. CFT includes the ‘multiple-selves’ intervention which involves the differentiation of threat-based emotion and an exploration of their conflict. Compassion is then applied to the client’s affective world to aid regulation and integration. This paper focuses on clients’ experiences of a chairwork version of multiple-selves, wherein clients personify their emotions in separate chairs. Nine participants with depression were interviewed directly following the intervention and the resulting data were analysed using interpretative phenomenological analysis. Three interconnecting themes were identified: appreciating emotional complexity; the role of chairwork process; and compassionate integration. The results highlight the importance of emotional differentiation in understanding internal multiplicity and conflict in depression, and the role of compassion in creating a sense of personal coherence. The embodied and enactive nature of chairwork was found to be of benefit in identifying and separating emotion, and in developing new forms of self-relating. The paper discusses the clinical implications of such findings for the treatment of depression.

Key learning aims

As a result of reading this paper, readers should:

1. (1) Learn about the ‘multiple-selves’ framework for working with threat emotions.

2. (2) Appreciate the complexity of emotions in depression.

3. (3) Understand how chairwork processes can be used to access, differentiate and address emotional material.

4. (4) Develop insight into how compassion can be used to regulate emotions and integrate aversive experiences.

Introduction

Compassion focused therapy and the multi-mind

Compassion focused therapy (CFT) is an integrative psychotherapy model grounded in affective-neuroscience, attachment theory and evolutionary science (Gilbert, 2020). The broad aim of the therapy is the cultivation of compassion in its various flows: from self-to-others, others-to-self, and self-to-self. The therapy was originally developed to address shame and self-criticism but has developed into an effective treatment for a variety of clinical presentations (Craig et al., 2020) and non-clinical populations (Irons and Heriot-Maitland, 2020). In practice, CFT draws on various methods for developing compassionate competencies and mind-states, including meditational practices, guided imagery, body-focused exercises, and method-acting techniques (Gilbert, 2010). In applying compassion to life difficulties, CFT integrates a range of evidence-based interventions, including cognitive behavioural approaches.

The initial stage of CFT involves generating insights about the complex nature of the evolved human brain and mind. Gilbert (2005) describes the human ‘multi-mind’ as a multiplex of motivations, emotions and cognitive competencies that give rise to conflict and distress. Such conflicts can occur between basic motivations (and their associated mentalities): for example, the tension between cooperating in a group whilst also competing for status and resources (Gilbert, 2017). Conflict can also arise between ancient ‘old-brain’ emotions, drives and action impulses that we share with other animals, and more recently evolved ‘new-brain’ cognitive abilities: e.g. feeling anxious and avoiding a conflict and then criticizing oneself for such feelings and behaviour (Gilbert, 2010). One of the aims of highlighting these inherent conflicts is to depersonalize and decentre from the processes of our biological mind, thereby reducing the individual attribution of blame. Another is to develop ‘mind awareness’ to differentiate the various patterns and potentials that texture the mind and ultimately shape its output (Gilbert, 2020).

CFT also adopts an evolutionary functional approach to emotions, clustering them into three interactive groups dependent on the primary motives they serve: threat, protection and harm avoidance; drive-based resource seeking and acquisition; and rest, digest and soothing (Gilbert, 2020). A course of CFT will include a focus on balancing these emotional systems with an emphasis on enhancing ‘soothing’ parasympathetic feeling-states and emotions which, in mammals, are interlinked with social safeness and care (Gilbert, 2014; Porges, 2017).

Gilbert (2020) suggests that ‘because our experiences are about patterns of functions and processes’ (p. 14), our emotions each have their own body-states, ways of thinking, behavioural impulses, attentional focus, associated memories and linked motivations. Emotions can therefore function like ‘mini-selves’ that shape our internal experience and external interactions in vastly different ways. To add further complexity, the same event/stimuli can trigger multiple emotions that arise in conflict or a mixed, undifferentiated form that can be hard to understand (Gilbert, 2010). Emotions can also be ‘fused’ to one another – becoming dominant or absent in our awareness – via classical conditioning (Ferster, 1973). For example, a client repeatedly shamed for expressions of anger as a child might, as an adult, experience shame or anxiety at their feelings of anger to the point where such anger is difficult to access in conscious awareness (Gilbert, 2007).

One CFT intervention that addresses such emotional complexity is termed ‘multiple-selves’ (Gilbert 2020; Kolts, 2016). This exercise involves the differentiation of the ‘big three’ threat emotions: fear/anxiety, anger/rage and sadness. Clients are typically asked to identify a dispute with someone they care about and the situation is explored via the reactions of each emotional ‘self’ (e.g. what does angry self think, focus on, feel like in the body and want to do?). Clients are also asked about their memories of each emotion (to determine associated learning histories) and what each emotional self ultimately wants (to identify related motives). A recent addition includes asking what the emotional self needs in order to ‘settle’ (highlighting possibilities for regulation). Interactions and relationships between emotional selves are then explored (e.g. ‘what does angry self think of anxious self?’) to reflect on the potential for emotional conflict and the individual’s pattern of emotional responding (e.g. angry self might block a client’s access to sad self’s grief). Finally, the compassionate self is accessed with exercises that support ventral vagal activation (such as soothing-rhythm breathing) and connect clients to their compassionate motivation and related qualities of mind (Gilbert and Choden, 2013). The compassionate self is then guided to reflect on the situation (with an enhanced capacity to mentalize both parties) and to relate compassionately to each emotion in turn.

The ‘multiple-selves’ framework can be introduced via a white board exercise but is also undertaken via chairwork, utilizing the method’s processes and procedures to enhance emotional differentiation and new forms of internal relating.

Compassion focused therapy chairwork

Chairwork is defined as a ‘collection of experiential interventions which utilize chairs, their positioning, movement, and dialogue to bring about change, principally through the facilitation of here-and-now interactions with parts of the self’ (Pugh et al., 2020; p. 2). Its origins are in the psychodramatic approach of Jacob Moreno (1948) with notable developments in Gestalt (Perls, 1973), experiential (Greenberg et al., 1993) and integrative therapies (Young et al., 2003). In form, chairwork can involve empty chair dialogues (e.g. speaking to imagined/remembered ‘others’ symbolically held in an empty chair), multi-chair techniques (whereby the client moves between such parts or representations, enacting both roles), and role-plays (whereby the client and therapist explore and rehearse interpersonal or intrapersonal interactions) (Pugh, 2020). In function, chairwork is used for various therapeutic ends including to enhance motivation and resolve ambivalence (Pugh and Salter, 2018), address internalizations of critical or abusive ‘others’ (Kellogg, 2015), or to elaborate alternative, positive, self-schemata (Chadwick, 2003). Chairwork has been incorporated into various evidence-based treatments (e.g. Paivio and Nieuwenhuis, 2001) and is effective as a stand-alone intervention (Shahar et al., 2012).

Given the multifarious applications of chairwork, across a diverse range of therapeutic modalities, Pugh and Bell (2020) have suggested underpinning principles and processes that unify the method. They conceptualize chairwork as fundamentally grounded in the principle of self-multiplicity, supported by the process of separation: i.e. the externalization and differentiation of parts of the self in set positions. Once separated, parts of the self can then exchange information via the process of animation. This might involve personifying aspects of the self (e.g. imagining the self-critic’s form and character in an opposite chair) or embodying them (e.g. moving to the critic’s chair and taking on its posture, gestures and voice-tone). Finally, transformation is achieved via dialogue: for example, cultivating or integrating aspects of the self via verbal and sensory communication.

Chairwork can be used throughout a course of CFT treatment but is typically associated with ‘set-pieces’: two-chair self-compassion, chairwork to address self-criticism, and multiple-selves (Kolts, 2016; Bell, in press). CFT chairwork is unique in focusing on the compassionate self/chair, which is used to tolerate, integrate and transform other aspects of the self. Previous research into CFT’s self-critic chairwork (Bell et al., 2020a; Bell et al., 2020b) has highlighted how the unique qualities of chairwork support the application of compassion. This includes personification of internal experience to facilitate self-directed empathy and mentalization; embodiment to identify patterns of the mind and the functions they serve; and separation of self-parts to witness and modify relationships between them. It is hypothesized that the multiple-selves framework delivered via chairwork (outlined in Table 2) will similarly benefit from the methods’ procedures, processes and properties, in the service of mind awareness and emotional differentiation.

Aims and emphasis on depression

This is the first study of the multiple-selves chairwork intervention. As such, the research is exploratory in nature and seeks to understand clients’ experience of the approach with the aim of improving its clinical application. The qualitative methodology of Interpretative Phenomenological Analysis (IPA) was chosen for its emphasis on lived experience and related meaning-making (Smith et al., 2009), and for the precedence of its use in client accounts of CFT (e.g. Bell et al., 2017).

The study focuses on clients with depression due to the complicated nature of emotions in this presentation. Whilst depression has been termed a ‘malignant sadness’ (Wolpert, 1999), it is an emotionally complex, emergent state of mind and brain, formed of multi-layered interactions between an individual’s biology, psychology, and social world (Gilbert, 2007). Depression can therefore have multiple pathways and be formed of various combinations of different experiences that give rise to the same symptoms and diagnosis (Irons, 2014). In emphasizing the evolutionary basis of depression, Gilbert (1992) suggested that depression is linked to blocked/arrested/ineffective defences, which can create problems in different emotions. For example, depression is associated with increased anger and irritability, but also with problems in being assertive in interactions with others and an internalization of anger-based self-criticism and shame (Gilbert and Miles, 2000). Depression is also associated with feelings of entrapment and blocked flight, as well as arrested help-seeking and a freezing of expressions of grief and sadness (Gilbert, 1992; Gilbert and Gilbert, 2003). There is also the toning down of positive affect systems linked to safeness (Gilbert et al., 2008). Depression can therefore be conceptualized as an avoidance or shutting down of core affects, which are linked to fundamental human defences, needs and motivations (Greenberg and Watson, 2006). The multiple-selves chairwork is therefore an opportunity to help clients access, tolerate, and express their threat-based emotions and to work compassionately to find new adaptive solutions.

Depression is also associated with difficulties in differentiating negatively valanced emotions, particularly in the context of high stress exposure (Starr et al., 2020a). Depression is typically experienced as generalized, global distress with clients ‘unaware that their malaise is exacerbated by their inability to discriminate and process their core emotions’ (Greenberg and Watson, 2006; p. 44). Lower negative emotion differentiation (NED) is also associated with greater parental depression, greater authoritarian parenting style, and lower parental attachment security (Starr et al., 2020b). Developing a compassion-based intervention that supports NED may therefore hold particular benefit for clients with depression.

Method

Recruitment and eligibility

Ten participants were recruited from two ‘Improving Access to Psychological Therapies’ (IAPT) teams within the National Health Service (NHS). Due to COVID-19 restrictions to in-person treatment, only nine participants were ultimately included in the study (a large number within IPA studies, Smith et al., 2009). The eligibility criteria required clients to have a ‘provisional diagnosis’ of Major Depressive Disorder and a score of 10 or above on the Patient Health Questionnaire (PHQ-9) (Kroenke et al., 2001) at the start of treatment. The score of 10 or above is used as a ‘cut off’ for clinical depression within IAPT services, and the identification of a provisional diagnosis is standard practice for IAPT therapists (National IAPT Programme Team, 2011). All participants were receiving CFT and the multiple-selves intervention as part of their routine treatment and were identified by their therapists for the research. See Table 1 for a summary of participant characteristics.

Table 1. Participant characteristics

To describe the sample, the following measures were administered at interview: Other as Shamer Scale (OAS) (Goss et al., 1994); Forms of Self-Criticizing/Attacking and Self-Reassuring Scale (FSCRS) (Gilbert et al., 2004); and the Beck Depression Inventory (BDI-II) (Beck et al., 1996). Whilst not based on statistical analysis, OAS results (mean = 37.33; SD = 17.21) were higher than non-clinical average (mean = 20.0, SD = 10.1) (Goss et al., 1994). FSCRS results were as follows: inadequate-self (mean = 28; SD = 6.65), hated-self (mean = 8; SD = 4.47) and reassured-self (mean = 12.22; SD = 6.62). Compared with a normative study of the FSCRS (Baiao et al., 2015), the present participants scored similarly to a clinical sample on inadequate-self (mean = 27.47; SD = 7.51); above non-clinical, but below clinical, samples on the hated-self; and below non-clinical, but above clinical, samples for the reassured-self. The mean results of the BDI-II (mean = 26.88; SD = 14.53) indicated ‘moderate depression’ (Beck et al., 1996).

Therapist eligibility and characteristics

Therapists were required to have undertaken a minimum of an introductory training in CFT (typically three days) in addition to specific training in the intervention (as part of an advanced CFT skills workshop, typically three days, or via a two-hour workshop delivered by the first author for participating services). The six therapists involved in the study included: four females and two males; four White-British, one British-Pakistani and one White-other; and an age range of 30–63 years (mean = 40.83). All participants practised CBT, in addition to CFT, as their main modality of therapy. Therapists were offered an hour supervision on the intervention by the first author.

Intervention

The multiple-selves intervention occurred during a single, face-to-face session. Due to the process-based nature of CFT, the intervention took place at various points in treatment ranging from session 7 to session 17 (mean = 12.77). Prior to the intervention, all participants had completed CFT psychoeducation and formulation, in addition to ‘compassionate self’ practices whereby clients are guided to embody and enact their compassionate motivation (via body-based exercises, guided imagery, and acting techniques) (Gilbert, 2010). The multiple-selves session was audio recorded and reviewed by the first author to ensure adherence to the steps outlined in Table 2.

Table 2. Outline of the multiple-selves intervention

Data collection

A separate, face-to-face interview was undertaken with each participant immediately following the session (see Table 3 for the interview schedule). The interview was semi-structured (based on the interview schedule), audio recorded, and transcribed verbatim by the first author.

Table 3. Interview schedule

Data analysis

The six-stage process outlined by Smith et al. (2009) was used to analyse the data. This included initial case-by-case notation and thematic labelling. Such themes were refined into super-ordinate themes via processes such as abstraction, subsumption and polarization. This process was repeated for each participant, followed by the development of cross-case super-ordinate themes and the production of a final narrative account supported by illustrative examples from the raw data. The first author maintained a reflective diary to identify and manage personal beliefs and presumptions. Such reflections were summarized and shared with the other authors who independently audited the transcripts, initial analysis, and thematic development. Any divergence of analysis was resolved by consensus.

Results

The analysis generated three interconnected super-ordinate themes (see Table 4).

Table 4. Summary of themes

Theme 1: Appreciating emotional complexity

Multiplicity and differentiation

All participants were able to access and express experiences from each emotional ‘self’ in a way that highlighted distinct mind–body patterns for each emotion. The identification of such individual patterns gave rise to an impression of the self as multifarious: formed of agentic ‘parts’, separate ‘mind-sets’, ‘mind-frames’, or ‘modes’ of being. These emotional selves were found to offer differing perspectives and ways of reasoning, alternate behavioural impulses and motivations, in addition to separate feeling states. Such selves were frequently referred to as autonomous ‘characters’, with Alice making repeated reference to the personified emotions in the film ‘Inside Out’ (Docter and Del Carmen, 2015). In relation to the problem chosen for the exercise, these selves offered new perspectives and potentials, and a focus on alternate aspects of character (Amy: it ‘really opens your mind to different facets of your personality’). Participants also identified how moving between emotions and mind-sets helped to break down ‘bad’ global experiences that were previously indistinct, undifferentiated, or singular. Participants identified the structured separation of emotion as key to clarifying their initial ‘jumble’ of reactions:

‘It is hard for me to decipher, I feel all of these things at the same time, hard to pick it all out I just need to filter it a bit more … because I feel everything is like a whirl wind, my brain is like a hurricane, going round at once, maybe slowing down or like you are in the middle so you can pinpoint the bits’ [Charlie]

Similarly, Anya outlined how the process of slowing down and separating her ‘mish-mash’ of experience helped her to ‘unravel’ her emotions in a way that brought clarity and options. Universally, participants identified how the ordered nature of the exercise allowed for a deliberate analysis of their reactions to ‘hone’ in on the particular:

‘Break it down into its building blocks, instead of just thinking I’ve just got anxiety, well at least break it down, to actually address it at a more basic level’ [Chris]

Such analysis also allowed participants to distinguish between emotions that had previously been confused (James: ‘because of the crossover between them’). In learning to ‘appreciate’ the nuance and detail of greater emotional complexity, participants reported a sense of playfulness, curiosity and discovery, that they themselves found surprising:

‘Literally within the matter of a few minutes, it is quite fun to see how much your brain can just flip like that really’ [Emma]

Dominance, absence and interaction

Having separated emotions, participants identified that certain emotions were dominant and instantly available whilst others were absent, difficult to access, or avoided. Anxiety was the most dominant (n = 5), followed by anger (n = 3), whilst sadness was the most absent or avoided (n = 5), followed by anger (n = 4). Participants presented with idiosyncratic emotional profiles, revealed by their experiences in each chair: e.g. the chair/emotion that was easier to sit with or leave:

‘Those are my default modes so they are the easiest to slip into … but for the sad self it takes a lot of analysis’ [Anya]

As in Anya’s reflection, such in-session experiences were deemed indicative of broader personal patterns and tendencies. For some participants this provided new insights and ways of describing their daily functioning; Alice, for example, used the metaphor of drivers and passengers to describe how her life was directed by anxiety ‘in the driving seat the whole time’. For others, emotions were acknowledged as completely absent from their lives and awareness or, as Tim explained, were actively ‘controlled’ in ways that had interpersonal consequences (e.g. difficulty being assertive when anger was restricted). Participants also identified that their emotions could be internally or externally directed, blocked or absent at an inter- or intra-personal level. Chris, for example, identified that his anger was blocked externally but expressed inwardly in the form of self-criticism:

‘No, always anger at myself. I very rarely get angry at other people’.

Emotions were also experienced as interactive and dynamic, with one emotion ‘overriding’, ‘overwhelming’, or reacting to another. This was again experienced as an in-session process as selves morphed into others and ‘voices’ battled to be heard. As Amy explained, this process made the exercise feel like an ‘explanatory’ means to ‘assess’ and witness emotional conflicts in real time. Participants experienced various emotional sequences that highlighted the cyclical nature of their reactions, with each emotion acting as ‘fuel’ to the other. Such cycles included self-focused anger at the vulnerability of anxious self, or, in turn, anxiety at the destructive urges of the angry self. For Charlie, sadness was freighted with anxiety due to the fear of becoming subsumed in the emotion:

‘Sadness, because I find myself going more into that emotion and it frightens me how I can become so sad … it feels like swimming and you are just sinking and you can’t swim to the top’ [Charlie]

Tim, similarly, identified an avoidance of sadness due to the depth of associated grief from multiple bereavements. For other participants, certain emotions could only be accessed after experiencing and expressing other emotions, with the exercise offering a means to access the emotional ‘root’ of the problem (Emma: ‘where is it actually coming from’). One of the benefits of the exercise was felt to be the active balancing of emotion, down-regulating some emotions whilst holding space for others:

‘It has been a long time with one side being dominant, so I really appreciated that, giving the sides of myself that don’t have a dog in the fight because they are never around ’ [Amy]

This helped participants to look below the predominant ‘surface’ emotion and give voice to aspects of the self that had been inaccessible or disowned.

Theme 2: The role of chairwork process

Embodiment to identify and access selves

Participants particularly valued chairwork’s emphasis on embodiment and enactment. When prompted to ‘become’ each emotional self, participants universally reported an ability to ‘tap into’, identify, and explore associated bodily experiences. Such sensations were experienced intensely, in an immediate ‘amplified’ manner: as if ‘you could flip a switch’ (Charlie). Participants also noted how being each emotional self impacted the whole of the body and shaped its expression:

‘Everything from body language to the tone of my voice changed because I was accessing a different emotion, it was quite marked … you can very much see and hear the difference in my voice and the difference in how I was sat and my movements were very noticeable’ [Kerry]

As in Kerry’s description, such tangible expressions offered additional insights into the nature of each emotion as they were literally seen and heard by both participant and therapist. The emotional selves became known and differentiated by their somatic makers and the rich granularity of sensations experienced. Participants also reported greater awareness of corresponding changes in mindset, motivation and ‘function’ via changes in physical tensions and urges [e.g. the angry self ‘wanting to smash things’ (Charlie)]. The compassionate self was notably identified by a ‘lighter’, ‘brighter’ feeling tone and sense of ‘peace’.

Participants also found the exercise helpful in moving from an abstract or cerebral discussion of emotion to an exploration grounded in the body and present moment experiencing:

‘You start to feel, really genuinely. In the exercise it wasn’t just hypothetically, it was something that was real, it did make me feel sad, in the exercise I felt all three emotions, in quite a short space of time’ [James]

A similar sentiment was shared by Chris who contrasted the intervention’s focus on in-session affect and internal sensation with the emotional ‘disconnect’ of prior cognitive interventions. Such embodiment and focusing ultimately made each emotion more memorable and easier to recall. Kerry particularly described the exercise as a kind of sensory sampling, with her felt experience resonating in her memory as opposed to lexical and imaginal recall:

‘Remembering how it felt rather that what it looked like, it is almost like if you go into a shop and smell all the different scented candles, it was like a sensory experience rather than picturing myself ’ [Kerry]

The body was also deliberately used to break connection between emotional selves. Participants valued palate-cleansing prompts to ‘get rid of all the remnants’ (Charlie) when moving between chairs, by slowing the breath and ‘shaking it off’ by means of physical gesture. Intentional changes to posture, breathing and facial expression (‘smile’) were also important in accessing the compassionate self.

Standing up, looking back: the benefit of moving chairs

In addition to the role of embodiment, participants emphasized how the physical act of moving between chairs supported shifts in emotion and mentality. When standing up from a chair, participants reported a symbolic sense of stepping out of their emotional selves and leaving them still ‘parked’ behind. Similarly, walking towards a new chair was experienced as an active, intentional means to take on new aspects of the self (James: ‘to step into that shoes, be that person’). This phenomenon was particularly beneficial for participants who felt fused to a specific emotion.

‘I don’t think anyone would be able to do this exercise if they were just sat in one chair, it would be much harder to jump into a new mind-frame or a new self, I think you do need to have the three chairs’ [Anya]

A similar sentiment was shared by Amy, who felt energized by the enactive elements of the session, providing her with a sense of agency and flow in contrast to the ‘stagnant’ nature of her habitual responding:

‘It is really easy to just get stagnant when you are in one position and you are in one mindset … the fact that we moved through the room it didn’t feel static at all, I think that was the key to keeping the blood flowing and making sure your brain is switched on constantly’ [Amy]

The externalization and concretization of emotions onto set chairs, in set positions, not only created the potential to move between emotions at will, but also allowed for new, de-centred, perspectives. Participants were able to look back on their emotional selves with the benefit of physical and psychological distance and separation: the chance to explore the inside ‘from the outside’ (Charlie). For Kerry, such distance provided an overview of her emotions from a new reflective position, at a remove from the usual intensity of her bodily reactions:

‘It was almost like looking into a CCTV screen and they have all the different screens of all the different images … looking out of body, where you look at the scene and pick apart the emotions that are contributing’ [Kerry]

As in Kerry’s example, relating back to empty chairs, where parts of the self had been enacted, generated significant levels of mental imagery. Images of emotional selves were typically visual representations of the self at moments of high affect (James: ‘I could see my younger self ’). Seeing oneself externally represented allowed participants to meet themselves ‘face to face’, as if interacting with another human being (Tim: ‘not you but it is you’). This enabled participants to dialogue with themselves in new ways, applying social-relational skills to their own internal world:

‘This image of what you would look like sitting there, it actually felt like I was giving myself something to talk to. It gave me something to feel real to talk to. It could be three people sat in those three chairs’ [Emma]

Participants ultimately identified an increased ability to be compassionate to themselves and their emotions when externalized or personified in this way.

Theme 3: Compassionate integration

Empathy, acceptance, and integration

In contrast to the emotional selves, the compassionate self was experienced as slower, more reflective, rational, and reasoned. The compassionate self was also capable of empathy for the reactions and experiences of the other selves and was associated with a ‘wiser mind and perspective’ (Kerry). Such empathy was evident as participants reflected on broader contextual factors that explained the presence and intensity of their emotions. This included consideration of situational triggers, an understanding of the other person’s experiences, and an acknowledgement of the impact of childhood relationships. Amy, for example, identified that her anger had become so dominant throughout her life ‘just to be able to survive’. Other participants found relating their emotions to their evolutionary origin and function helped to comprehend and predict their patterns of thinking and feeling:

‘I knew what their role was and I could use that to go off what was going on in my head and it was honestly so helpful’ [Alice]

Accompanying such insights was a global reduction in self-criticism and blame. Participants validated their emotions as ‘normal’ and expected human experiences, rather than as personal faults or problems:

‘Anyone would feel the same way, and I realize if I had watched that scenario unfold with anyone else I would have instantly taken my side. So I was able to leave that feeling of anger with that person and say “you know what, it was never my fault”’ [Kerry]

In their personified form, the emotional selves were treated with acceptance, care and support. The compassionate self was experienced as a ‘parenting’ role, with participants equating their care for themselves to their relationship with family or friends. Genuine affection was also felt as participants expressed a desire to ‘hug’, hold, and sooth their emotional selves. Rather than criticizing, avoiding, or attempting to suppress their emotions, participants suggested they ‘formulate a stronger motivation’ (Chris) in their compassionate self, to accept and work with them in a ‘constructive’ or ‘healthy’ manner. Participants identified the compassionate self as an integrative force that links and connects their various emotions together. Such connection of multiple strands of internal experience created a sense of personal cohesion and integrity:

‘It is just one, it is the whole self, all combined together’ [Anya]

The compassionate self was emphasized as ‘holistic’ in orientation, and able to capture the unifying function and aims of the threat emotions:

‘To acknowledge that it is three sides to the same coin, and how they interlock and relate with each other’ [Amy]

Participants stressed the importance of separating and untangling their emotions for the compassionate self to understand, reorganize, and coordinate them into a coherent whole (Kerry: ‘put myself back together’). This final sense of unity was not achieved by removing complexity or difference between the emotions, but rather by finding a balance and equilibrium to create a ‘richer’ and more intentional life.

‘For me it is about the last compassionate self acknowledging that all of that together makes up one individual … combine it all and make it so you are a fully functioning adult human that is capable of making decisions and living life the way you want to’ [Amy]

Discussion

This is the first study to investigate CFT’s multiple-selves approach to emotions and the first to explore its delivery via chairwork. The results of an interpretative-phenomenological analysis of participant experience generated three interacting superordinate themes: emotional complexity; chairwork process; and compassionate integration. Together, the themes suggest that chairwork offers specific mechanisms that facilitate the core aims of the format: the differentiation and exploration of threat-emotions and their integration with compassion.

Supported by the basic structure of the exercise, participants were very able to experience and express distinct affective patterns in each chair. Echoing previous literature on emotion in depression (Greenberg and Watson, 2006), participants described their initial reaction to their chosen memory as undifferentiated and globally ‘bad’, or as a confusing ‘mish-mash’ and ‘jumble’ of feeling. The separation of emotion helped to break down participants’ reactions into ‘building blocks’ in ways that clarified their constituent parts for further analysis and insight. Whilst providing a degree of organization and filtering, such differentiation also created a sense of expansion and possibility as new perspectives and ways of responding were actualized, tolerated and explored; as Gilbert (2020) has asserted: affective patterns function like ‘mini-selves’ with discrete mind-sets, desires, feelings and histories. This notion of self-multiplicity was well accepted by clients who experienced their emotions as agentic ‘parts’ with different characters and minds. Participants were surprised at their ability to access, inhabit and verbalize the subjective reality of distinct emotional selves and reported experiences of curiosity and playfulness (‘it is quite fun’) at this new capacity.

The above experiences are particularly notable given the bio-psycho-social profile of clinical depression: its submission, immobilization and association with lowered heart rate variability and restricted emotional tonality (Licht et al., 2008). Depression is also linked to fears and suppression of emotion (Beblo et al., 2012), difficulties identifying and differentiating emotions (Demiralp et al., 2012), problems in modulating emotions (Feldman Barrett et al., 2001), and the inhibition of adaptive defence responses, such as healthy assertiveness or help-seeking (Gilbert, 1993). Whilst the current sample is small, the results offer promise in setting out a structured and well-tolerated intervention to target these issues. The exercise appears to generate an appreciation of emotional granularity and a capacity to label affective states, an opportunity to switch emotions in quick succession, and a means by which to approach avoided feelings and inner experience. The exercise also provides a safe method to express and ‘play out’ a range of innate defence responses: the ‘fight’ of anger, the ‘flight’ of anxiety and the adaptive processing of loss associated with sadness. There is the potential that this experiential activation of protective responses may support the broadening of participants’ coping repertoire and role-formation, reversing the sense of entrapment and defeat so associated with depression (Gilbert and Allan, 1998).

The separation of emotions during the exercise also highlighted the complex, idiosyncratic nature of emotion in depression. Participants described specific emotions as dominant or ‘default’, with anxiety being the most prominent, whilst sadness and anger were most frequently described as absent, avoided or ‘controlled’. This supports assertions that depression is not related to a single emotion, but rather ‘is a syndrome, and often avoidance of core emotion is an aspect of this syndrome’ (Greenberg and Watson, 2006; p. 55). Emotions were also shown to be dynamic, interactive and conflictual, with secondary emotions restricting access to the primary ‘root’ of the problem. Examples included anger at the vulnerability associated with anxiety, or a fear and avoidance of sadness related to bereavement. Complex reactions between inter- and intra-personal threat reactions were also captured in the shifting between external to internal foci: i.e. the blocking and redirection of anger from external sources towards internal faults and problems, illustrating a function of self-criticism in protecting from fears of external conflict/rejection (Gilbert, 1992). By highlighting the idiosyncratic patterns underpinning a client’s depression, the multiple-selves exercise might be used to tailor and individualize treatment for a more targeted approach.

Such interactions and conflicts were discovered by participants during the process of the exercise, in difficulties they experienced in entering, leaving, tolerating, or clarifying specific emotions. Rather than being a ‘problem’, such difficulties presented participants with a live means to identify and ‘assess’ their individual profile of emotional responding. Therapists delivering multiple-selves chairwork may therefore benefit from viewing the method as both intervention and assessment, whereby blocks or complications in the task are therapeutically valued for the insights they provide to the client’s broader functioning and learning history. This parallels the way in which CFT utilizes the presence of fears, blocks and resistances to compassion during compassionate mind training, viewing them as a form of ‘intuitive wisdom’ that reveal key areas of personal threat and need (Gilbert, 2020). Whilst process observations might highlight core emotional conflicts in the multiple-selves exercise (e.g. the client becomes tearful when angry), assessment questions for therapists might also include: Which emotions are most and least accessible? When accessing one emotion, do others intrude? Are there discernable sequences of emotional activation (e.g. does sadness tip into collapsed despair)? What is avoided? Do the emotions focus internally or externally? And, given what you know about your client’s life experiences, how do these patterns make sense?

Chairwork processes include separation of ‘parts’ or voices onto chairs, animation of such parts via embodiment or personification, and the facilitation of dialogue between them (Pugh and Bell, 2020). In terms of embodiment (asking the client to change seat and become the part) participants reported intense bodily experiences of each emotional self, which were elaborated and amplified in the adoption of associated body-postures, voice-tones and gestures. As in prior chairwork studies (Bell et al., 2020a; Bell et al., 2020b), embodiment and enactment provided tangible expressions and experiences of each self, making them ‘real’ and memorable whilst also offering ‘action-insights’ into their corresponding mindsets and motivations (Kellerman, 1992). Participants could literally ‘hear’ the character of the emotional self in how their voice sounded, or ‘feel’ the protective function of the angry self in its urge to ‘smash’ and attack. As one client eloquently explained, the selves were sensually sampled as if smelling different scented candles. Embodying parts of the self, and attending to proprioceptive cues, therefore offers novel access to implicit sources of information and highlights ‘bottom-up’ modes of information processing (Kashdan and Farmer, 2014; Michalak et al., 2012). The dynamic interaction between body and mind was also used, intentionally, both to leave and enter self-states: ‘shaking off’ each role between chairs whilst also cultivating (via changes in posture, expression and breathing) the biological and affective conditions conducive to compassion (Gilbert, 2010). When working with the multiple-selves exercise, clinicians may therefore benefit from utilizing the insight that ‘motor expression may intensity congruent emotion, but also dampen other emotions’ (Greenberg and Watson, 2006; p. 88).

The chairwork practice of moving between chairs to separate connection and contact with each self was also found to be helpful, particularly for participants who felt ‘stagnant’ and stuck to a single ‘position’. The physical procedure of standing up and away from the chair was experienced as a symbolic and psychological shift of being, an intentional stepping out of one role and into a next. Phenomenologically, participants still experienced the previous self as firmly ‘parked’ in the prior chair, which led to a reduction in affective intensity. This process of separation and de-centring is proposed as a key mechanism of action for chairwork (Pugh, 2017), and links to the literature on immersion and distancing which suggests that flexibility between first-person (ego-centric) and third-person (observer) perspectives and processing can be beneficial in the treatment of clients with depression (Barbosa et al., 2020). The concrete and embodied change in perspective provided a new reflective overview of the problem (the inner world seen ‘from the outside’), with complex internal relations mapped out in space in a manageable form. Given the importance given to ‘space’, ‘position’ and ‘distance’, an innovation to the exercise might include adaptations of psychodrama’s sociometric methods (Cruz et al., 2018) whereby relationships between emotions can be represented and explored by altering the proximity, arrangement and order of chairs.

The exercise also benefited from personification of emotions, and the spontaneous imagery generated when looking back to an empty chair. As identified in prior chairwork research (Bell et al., 2020b), such imagery augmented the experience of a dialogical encounter, creating a ‘face to face’ focus for the relational exchange. In depicting heightened social signals of distress, the personified imagery of emotional selves was particularly conducive to compassion, allowing for the application of social-relational skills and responses that participants usually reserved for ‘others’ (such as their friends or family members). This involved capacities for empathy and mentalization, taking into account situational factors and causes, as well as feelings of sympathetic connection and an urge to regulate each self with physical touch. Imagery is not an explicit process task in the multiple-selves exercise, but clinicians may benefit from encouraging imaginal representations of each emotion to side-step typical blocks to self-compassion.

As highlighted in the exercise, social threat and feeling threatened can result in segregated and conflicting stereotypic responses (e.g. fight, flight or freeze) and ‘in extreme cases, disintegration and dissociation’ (Gilbert, 2017; p. 35). In contrast, compassion can provide experiences of social safeness and care which ‘enables the integration of inherently disparate and segregated systems’ (Gilbert, 2017; p. 35). As compassion arises from the evolution of interpersonal care and attachment (Gilbert, 2015), it is associated with the effect of care-giving and receiving on regulating a range of neurobiological processes that are conducive to social engagement, psychological flexibility and integration (Keltner et al., 2014; Porges, 2011). Compassion is seen to ‘re-organize the mind’ and create a mentality that focuses on empathic understanding and support (Gilbert and Choden, 2013). The multiple-selves exercise demonstrates how such a care-based mentality might relate externally, drawing a distinction between its pro-social orientation and the disparate patterns of thinking and feeling generated from the emotional selves. Internally, the compassionate self was shown to operate as a ‘relation-forming aspect of mind’ (Gilbert, 2017; p. 62), taking a caring and ‘parenting’ stance towards various distressed aspects of the self. The compassionate self is seen to function in this parental manner and act as internal ‘secure base’ and ‘safe haven’ from which to build courage to re-engage with challenging aspects of one’s life, such as one’s emotions (Gilbert, 2020).

The compassionate self was also associated with a sense of slowing down, and the capacity to be reasoned, reflective and ‘wise’. Such wisdom translated into broader evolutionary and developmental insights integrated from CFT’s psychoeducation and formulation (Gilbert, 2010). These included an appreciation of each emotion’s evolved function and protective role, and the way in which they had been shaped across the individual’s life span. Such insights acted to de-shame and de-personalize the presence and intensity of the participants’ emotion whilst also providing an explanatory frame to understand, accept and predict patterns of thinking, feeling and motivation. In capturing the common protective function of threat-emotions, compassion created a degree of personal coherence and a sense of reorganizing and coordinating the ‘building blocks’ of differentiated experience back into a balanced whole. Kolts (2016) suggests the compassionate self functions like a ‘captain of the ship’, with the ‘superordinate’ role of navigating the storms of life whilst tending to the emotions as if alarmed passengers. This metaphor echoes one used by a participant who experienced anxiety at the ‘driving seat’ of their life. As identified by participants, the benefit of giving the compassionate self this central driving role is not to deny or suppress threat-emotions but to build a ‘stronger motivation’ with which to accept and work with them for a common goal.

Despite the benefits reported by participants, the study is limited in sample size and poses a number of unanswered questions, such as whether the intervention creates any empirical change in symptom reduction or in measures of emotional differentiation or regulation (e.g. Licht and Chabot, 2006). It is also unclear if such benefits are maintained over time, and whether the framework can be utilized contemporaneously by clients in the presence of a stressor. As emotional differentiation is associated with an increased ability to understand and navigate one’s internal world and, in turn, calibrate helpful social responses to interpersonal stressors (Feldman Barrett et al., 2001; Gohm, 2003), it makes sense for CFT to target this capacity to facilitate the multi-directional flows of compassion (between self-to-others, others-to-self, and self-to-self). Future research might therefore focus on evaluating the specific links between emotional differentiation and flows of compassion. Finally, whilst the role of chairwork in facilitating the multiple-selves framework was particularly emphasized by participants, alternative methods (such as white board depictions) remain untested. Given the context of the COVID pandemic, digital adaptations of multiple-selves chairwork could be developed and evaluated to ensure the retention of its unique qualities and processes whilst allowing greater access to the method.