Pass on what you have learned. Strength, mastery, hmmm … but weakness, folly, failure also. Yes; failure, most of all. The greatest teacher, failure is. Luke, we are what they grew beyond. That is the true burden of all masters. (Yoda to Luke in Star Wars: The Last Jedi, R. Johnson, 2017)

The purpose of this paper is to introduce a practical application of attachment theory in the context of supervision, with a particular emphasis on the use of the Circle of Security (Powell et al., 2014) as a model for supervision. Given the centrality of the quality of relationship in therapeutic outcomes (Calvert et al., 2016; Hiebler-Ragger et al., 2021), it makes sense to consider centring the relationship in the supervisory context as well. Attachment theory offers a framework for understanding those interpersonal dynamics. The Circle of Security developers offered the field a visual map of attachment theory that clearly and succinctly translates complicated attachment concepts into ideas easily grasped and understood. Glen Cooper, Kent Hoffman, and Bert Powell from the Marycliff Institute in Spokane, Washington, and Robert Marvin from the University of Virginia, Charlottesville, created the Circle of Security diagram and used it as part of a video-based intervention that strengthens parents’ abilities to observe their child’s needs and improve their caregiving responses to those needs (Poulsen et al., in press). The team at Circle of Security International have more recently created an analogous visual for supervisory relationships that emulates the original Circle. The Supervisor’s Circle of Security has not yet been publicly disseminated, and it is part of the purpose of this paper to do that.

A Very Brief History of Attachment Theory

John Bowlby and others concluded that a baby’s attachment to its primary caregivers was a result of a deep bond that begins at birth and lasts throughout the lifespan (Bowlby, 1969/1982; Lorenz, 1935). These researchers also concluded that the bond is established irrespective of the caregiver meeting the child’s physical and emotional needs (Harlow, 1958). From an evolutionary standpoint, separation from the caregivers presents great danger for the infant, and thus proximity must be maintained at all costs. Simultaneously, infants and young children also desire to explore and gain mastery over their environments. Bowlby (1969/1982) hypothesised that when children feel safe and secure, they are able to explore and learn and their brains are wired for growth. Exploration, and the needs associated with exploring, are represented on the top half of the Circle of Security diagram (see Figure 1). However, when children feel threatened, exposed, criticised, or vulnerable, their exploratory system terminates, and their attachment system is activated (Powell et al., 2014). The child’s return to what Bowlby (1969/1982) called the “safe haven” is captured in the bottom half of the Circle (see Figure 1).

Figure 1
Figure 1.Circle of Security: A Visual Map of Caregiver–Child Attachment

Note. From What is the Circle of Security? Developing Specific Relationship Capacities, by Circle of Security International, 2022 ( Copyright 2022 by Circle of Security International. Used with permission of Circle of Security International, Inc.

Ainsworth et al. (1978) through “the strange situation” experiment established three attachment categories: secure, insecure ambivalent (or resistant), and insecure avoidant. These patterns suggest that infants interact with their caregivers based on different cognitive expectations and that these expectations (or internal working models) are based on actual experiences with the caregivers (Kobak, 1999). The state of being securely attached essentially refers to a child’s confidence about the availability and responsiveness of the caregiver (Ainsworth et al., 1978), while insecure attachment refers to solutions that a child has developed to manage their affect in the absence of a mostly available and reliable attachment figure. In the Circle of Security diagram, the attachment figure is represented by the hands, and the job of that figure, using Bowlby’s (1969/1982) own phraseology, is to be the secure base from which the child is free to explore and the safe haven to which the child returns for comfort or protection (see Figure 1).

Of course, caregivers may struggle to meet top-half of the Circle needs or bottom-half needs, and consistent struggles to meet needs is a hallmark of attachment insecurity. In the case of the ambivalently attached child, the struggle is with exploration, and the caregiver pushes for connection. The child’s clinginess and the caregiver’s struggle to soothe are consistent with ambivalent attachment (Ainsworth & Bell, 1970). The avoidantly attached child, on the other hand, responds to the caregiver’s lack of comfort with closeness by turning away, the other primary pattern identified in Bowlby’s (1969/1982) writing and confirmed by Ainsworth’s observational paradigm (Ainsworth & Bell, 1970.)

Later, Main and Solomon (1986) added an additional category to describe a fourth category: disorganised. Disorganised attachment refers to the lack of a cohesive system to respond to the caregiver, which is often believed to be a result of trauma within the primary relationship, whereby the source of comfort is also the source of fear, resulting in a paradoxical situation for the child (Granqvist et al., 2017). The Circle of Security diagram represents disorganisation as a struggle with hands; in this case, the caregiver struggles to balance being bigger, stronger, wiser, and kind, leading to harsh and punitive reactions to the child’s needs and/or abdication and lack of boundaries, either of which creates fear in the young child (Hoffman et al., 2017). These patterns tend to be relationship-specific; hence, a child might have different attachment categories with different caregivers.

The Adult Attachment Interview (AAI) assesses adult patterns and states of mind that parallel the infant attachments previously identified (George et al., 1985): secure (S), dismissive (D), preoccupied (P), and unresolved (U; see Table 1). An additional category of “earned security” was added to the AAI (Pearson et al., 1994). Earned security is a designation based on the observation that the nature of experiences shared in the AAI would ordinarily suggest an insecure or disorganised state of mind; however, individuals with the designation of earned security experienced a different trajectory that led to a secure state of mind (Hesse, 2018).

Table 1.Continuity of Attachment Systems from Childhood into Adulthood
Strange situation (behaviour based) Adult attachment interview (based on state of mind)
Secure Secure
Avoidant Dismissing
Ambivalent/Resistant Preoccupied
Disorganised Unresolved

Milne (2006) refers to reasoning by analogy as a way to reflect on what is known in one area to inform thinking about and application to another area. In this case, applying an attachment framework to a supervisory relationship suggests that when supervisees feel safe and supported, their energies are freed to explore new theories and techniques, to take chances with an untested approach, and to learn about themselves and their clients. Conversely, a supervisory relationship marked by threats of exposure, criticism, and lack of empathy will result in supervisees focusing their energies on defending themselves from attacks and minimising risk.

The Supervisor’s Circle of Security diagram, which is constructed in a similar way to that presented in Figure 1, centres the supervisor as the hands in the relationship—the secure base from which the supervisee explores and the safe haven to which he or she can return (see Figure 2). As developed later in this paper, considering the supervisory relationship from the perspective of attachment can yield insights into relational dynamics that are important to track.

Figure 2
Figure 2.Supervisor’s Circle of Security

Note. From Circle of Security International [Unpublished manuscript], 2023. Copyright 2023 by Circle of Security International, Spokane, Washington. Used with permission of Circle of Security International, Inc.

It feels necessary to include a note here regarding recent developments in the field of attachment concerning the question of universality of attachment given cultural variability in parental practices. Although much work still needs to be done to continue to examine cross-cultural implications for attachment theory, the current research suggests some aspects that are universal and others that are determined by cultural and other contexts (Mesman et al., 2018). These studies emphasise the importance of investigating attachment within the wider social networks in which children grow and develop. Universal aspects of the theory include the universality of attachment, the normativity of secure attachment, the link between sensitive caregiving and attachment security, and the competent child outcomes of secure attachment (Main, 1990; Mesman et al., 2018; van Ijzendoorn & Sagi-Schwartz, 2008).[1]

An Even Briefer History of Supervision

The earliest accounts of supervision in the field of mental health date back to Sigmund Freud (Jacobs et al., 1995). As Freud was inventing the “talking cure”, he realised the importance of the therapist developing a deeper understanding of themselves and their conflicts in order to assist their clients best. The earliest form of supervision was documented in the early 1900s, when Freud and his followers met weekly in Freud’s waiting room in what was known as the Wednesday Psychological Society. Members of the group took turns presenting a case, and drank coffee, ate cake, and smoked while delving into case discussions. The first recorded supervisory case is that of “Little Hans” in which Freud was assisting Hans’s father in treating the boy’s phobia (Jacobs et al., 1995).

As the field of psychotherapy grew and became more formalised, engaging in one’s own treatment while training to become a therapist became a requirement (Frawley-O’Dea & Sarnat, 2001). Whether one could use one’s therapist simultaneously as one’s supervisor became an area of debate. This heralded the teach-or-treat dilemma, which continues to be an issue of debate to this day in some circles (Frawley-O’Dea & Sarnat, 2001). The extent to which a supervisor focuses their discussions and interventions with the supervisee on didactic instructions versus exploring the therapist’s own conflicts, wishes, history, and fantasies remains a supervisory dimension about which there is no clear agreement (Ladany et al., 2005).

A quick search on Wikipedia resulted in 156 different schools of psychotherapies ( Each of these schools uses a different method to train its novices, incorporating various emphases on supervision (Leddick & Bernard, 1980). However, most registration bodies require some form of mandatory supervision. For example, the Psychotherapy and Counselling Federation of Australia (PACFA) requires 10 to 15 hours of formal supervision for certified practising and registered clinical members every 12 months, depending on their client contact hours (PACFA, n.d.), while the Australian Counselling Association (ACA) requires members to complete a minimum of 10 supervision hours per membership year with a recommendation for one hour of supervision for every 20 hours of client contact (ACA, 2019). Similarly, the Psychology Board of Australia requires registered psychologists to complete 10 hours of peer supervision or consultation annually (Psychology Board of Australia, 2019), although no such requirement was found beyond licensing for members of the American Psychological Association (American Psychological Association, 2014).

Why Is Ongoing Supervision Needed?

Rising levels of burnout and stress in mental health care staff are an international concern for health systems (Javier & Vázquez-Cabrera, 2013; J. Johnson et al., 2018; Scanlan & Still, 2013; Yang et al., 2015). Black (2023) reports that in Australia the demand for mental health services is soaring, and that a significant shortage in the health care workforce is leading to rising rates of burnout amongst psychologists in particular and mental health workers in general. She writes, “according to the Australian Psychological Society (APS), before the pandemic one in 100 psychologists were unable to take new patients. It’s now one in three” (para. 7). Furthermore, a workforce study of counsellors and psychotherapists in Australia found that many gained more new clients and were seeing existing clients more often during the pandemic (Bloch-Atefi et al., 2021).

Rosenberg (2018) argues that a confluence of issues in the United States has led practitioners working in the public health sector to feel more stressed and less supported. These factors might resonate internationally. The factors he outlines include the nature of the business (the pressure to see as many clients as possible in as brief a time as possible), clinical and economic problems (practitioners being paid remarkably little for doing very difficult work), the advent of technology (the constant monitoring of staff activities and productivity), and staff burnout and turnover. He writes:

Caring for others without a sense of reciprocation is beyond unrewarding; it is depleting. While one’s [clients] can, at times, be the source of that reciprocation, relying on one’s [clients] as a source of nurturance is neither clinically appropriate nor practically reasonable … I would suggest that there is something within a dynamically oriented supervision or consultation that carries with it a sense of caring, engagement and reciprocity that allows for feeling understood and given to in a way that can potentially offset the forces that lead to burnout. (p. 260)

Rosenberg et al. (2019) conducted a pilot study in which the authors provided weekly psychodynamically oriented consultation with experienced dynamically oriented supervisors over the course of 12 months to mental health practitioners of various levels of experience and across multiple disciplines. Consultation was significantly and positively correlated with level of work satisfaction (vigour, dedication, absorption) and a sense of recent professional development. Additionally, reflective supervision has been identified as an important mitigating factor in the prevention of secondary traumatic stress and reducing the effects of burnout and vicarious traumatisation (Cunningham, 2003; Kassam-Adams, 1995; Kulkarni et al., 2013; Lee & Ashforth, 1993; Lloyd et al., 2002; Schauben & Frazier, 1995).

Models of Supervision

While many models of supervision exist (Bernard, 1979; Greenwald & Young, 1998; Holloway, 1995; Lambers, 2000; Liese & Beck, 1997; Rønnestad & Skovholt, 2003; Stoltenberg et al., 2014; Ward & House, 1998), Frawley-O’Dea and Sarnat (2001) propose that the models fall along a continuum of three dimensions. Dimension 1 concerns the nature of the supervisor in relationship to the supervisee, Dimension 2 involves the supervisor’s focus, and Dimension 3 is the supervisor’s primary mode of participation. In a client-centred model, the supervisor is the expert and derives their authority from their knowledge of theory and technique. Difficulties with the case are formulated in terms of challenges for the client or limitations and deficiencies in the supervisee’s skills or countertransference. This type of supervisory approach works well when there is a specific technique to be taught and the relationship is new and the dyad not yet comfortable with deeper exploration, or alternatively when a supervisee is new to the field and building a knowledge base. However, it also presents challenges because the supervisee’s own reactions and feelings about the case are considered distractions and irrelevant to the supervisory process (Frawley-O’Dea & Sarnat, 2001).

In a supervisee-centred model, the authority of the supervisor still lies in their knowledge and expertise. However, regarding Dimension 2, the data considered relevant diverge greatly from a client-centred approach. Here, in addition to considering the client’s state of mind, the supervisee’s mind is of particular interest. However, the supervisor’s own state continues to be unexamined, while the supervisee’s resistances, anxieties, projections, object representations, and working models take centre stage. In this supervisory model, the intersubjectivity between the supervisor and supervisee is not attended to and challenges with a case or in supervision are attributed entirely to the client or supervisee (Frawley-O’Dea & Sarnat, 2001).

While both of these models (client-centred and supervisee-centred) present a valuable way to understand supervision and attend to the supervisee’s learning, the authors note that these models fail to consider the complexity of the supervisory relationship and the contribution of the supervisor to what occurs in a supervisory session. A third model proposed by these authors is that of a supervisory-matrix-centred model in which the supervisor views themself as a participant in the supervisory process. In this model of supervision, the supervisor derives their authority from their capacity to participate, reflect, process, and interpret themes as they arise in the supervisory relationship (Frawley-O’Dea & Sarnat, 2001).

Over the past 20 years, the term reflective supervision has become popular in the mental health field, and in particular in the field of infant mental health (Shahmoon-Shanok, 2006; Tomlin et al., 2014). The term generally refers to the idea of providing a safe environment for the supervisee to explore their feelings and reactions towards the families they serve. The idea is that of a nested model of supervision (Slade, 2023), in which the child is held by the parent, the parent by the practitioner, and the practitioner by the supervisor. This form of supervision is most closely aligned with the supervisory-matrix-centred model presented above since it requires bi-directionality in the experience of supervision. Heffron and Murch (2012) propose critical skills for the supervisor, namely, mindfulness and attunement, slowing down, containment, sorting and selecting, perspective taking, professional use of self, negative capacity (the capacity to contain one’s impulse to speak before fully understanding or to try to solve the supervisee’s dilemmas), raising concerns, addressing differences of opinion, and spotlighting (asking questions in a way that is not shame-inducing). The authors have recently added critical self-reflection as an important capacity to highlight further the importance of considering one’s social location and issues of diversity, equity, and privilege within the context of supervision (M. C. Heffron, personal communication, September 19, 2023).

In the past few years, the reflective supervision literature has considered issues of privilege, power, race, and equity in what is known as diversity-informed reflective supervision (Clark et al., 2019; Noroña et al., 2023; Shivers et al., 2022). Wilson and Barron (2022) define diversity-informed reflective supervision by offering four guiding principles in relation to supervisors: honest self-examination, embracing and honouring oneself as a supervisor, finding one’s own relationship rhythm, and sitting with tension. The authors argue that these principles promote the development of trusting, safe, and nurturing relationships that allow all participants to engage bravely in shared vulnerability and genuine expression of emotional responses to their work (Wilson & Barron, 2022).

Training of supervisors has also started to focus on the development of specific skills and competencies (Rubin et al., 2007). Within that body of literature, the building of relationships is identified as foundational and essential to the development of any other competency (Calvert et al., 2016). A number of recent studies have highlighted the central importance of the relationship in therapy outcomes (Shedler, 2010). With the therapeutic alliance accounting for 4–26% of the variance in therapy outcomes (Crits-Christoph et al., 2006; Hardy et al., 2007), it follows logically that a focus on enhancing the relational competency of the supervisee becomes a focus of supervision (Calvert et al., 2016). One way to accomplish this is by focusing on the relationship between the supervisor and supervisee as a vehicle for learning.

McWilliams (2021) writes that in a landmark study in 1958, the researchers noted a parallel between the client–practitioner relationship and the practitioner–supervisor relationship. That is to say, the supervisors found themselves feeling and behaving in ways that were mirroring the way the supervisee was feeling about and acting towards their clients (Ekstein & Wallerstein, 1971). This finding was called the “parallel process”, and this phenomenon has since been extensively studied (Baudry, 1993; Caligor, 1981; Sarnat, 2019; Watkins, 2017). Therefore, McWilliams (2021) argues that the supervisory relationship is:

much more interpersonally complex than the transmission of skills from a more competent to a less competent person … [leading to] an emphasis on the supervisory process itself and on the nature of the supervisory relationship, not just the content of the supervisor’s teaching. (p. 34)

Supervision Through an Attachment Framework

Since relational competency is identified as an important focus of supervision, approaching supervision itself through a relational or attachment framework makes sense. Researchers have begun to explore the relationship between attachment theory and supervision (Fredricks, 2018; Hiebler-Ragger et al., 2021; Pistole & Watkins, 1995; Watkins & Riggs, 2012). Watkins and Riggs (2012) offer “that the psychotherapy supervisee–supervisor relationship, while having the potential to develop into an attachment bond, must best be viewed as involving an affective component that leads to the evoking of attachment dynamics” (p. 256). This paper attempts to offer a model for how these attachment dynamics might present within the supervisory dyad.

With the aid of the Supervisor’s Circle of Security diagram, one can conceptualise a supervisory relationship that matches the attachment framework used to describe the interplay between a parent and a child (see Figure 2). The interplay between a parent (represented by the hands) and a child in Figure 1 may be viewed as having a similar structure to that of the interplay between a supervisor and a supervisee, as depicted in Figure 2. Of course, attachment theory posits that the child’s developing working model of the relationship is influenced significantly by the parent’s own history in relationships. The Circle of Security interventions thereby focus on helping the adult shift how they understand and respond to the child by learning to recognise the child’s attachment needs rather than simply “managing the child’s behaviour”. When two adults come together for supervision, each brings their relational history to the encounter in the form of working models developed over many years of interpersonal interactions.

This paper suggests a way to view the Circle using a relational model focusing on the relative contribution of the client, the supervisee, and the supervisor. As depicted in Figure 2, the supervisor–supervisee relationship is bidirectional, and although the supervisor holds the greater responsibility in containment, interpretation, and management of the relationship, both parties must be willing to engage for the relationship to succeed and provide opportunities for growth for both parties. We consider in this article the possibility of struggles around the Circle residing in both the supervisor and the supervisee. The impact of one person on the other may be influenced and distorted by each person’s history, since what is experienced is not always necessarily reality-based but influenced and superimposed by one’s expectations and experiences. For example, Shedler (2006) writes of his experiences as a therapist with his clients:

I begin therapy with all new [clients] in much the same way. I greet the [client], offer him a seat, and invite him to tell me why he has come. But I am not the same person in the eyes of the [clients]. Some see me as a benevolent authority who will advise and comfort them, some see me as an omniscient being who will instantly know their innermost secrets, some see me as a rival or competitor to impress or defeat, some see me as an incompetent bungler, some see me as a dangerous adversary, some see me as a disapproving parent to appease, some see me as sexy and alluring, some as cold and unresponsive, and on and on. These and a thousand other configurations emerge as therapy unfolds. Anyone who has practiced therapy for any length of time cannot help but be struck by the diversity of reactions we elicit from our [clients], and by how far our [clients’] perceptions of us can diverge from our self-perceptions and from the perceptions of others who know us in other contexts. (p. 22)

When applied to the relationship between supervisor and supervisee, one can appreciate that we are considering not only attachment styles but also the way transferences, countertransferences, temperament, defences, anxieties, resistances, and, importantly, race and other forms of diversity may be playing out in the interaction. Referring to Figure 2, how the supervisor and supervisee negotiate trips around the Circle is influenced by many factors.

As noted, adults fall into one of four potential attachment categories, although—except in the more extreme cases—it is argued that one could fall along a continuum of these categories (Hesse, 2018). Applying this methodology results in a matrix in which a supervisor–supervisee dyad can be assigned to one of 16 dyadic categories each presenting a unique constellation of strengths and challenges (see Table 2).

Table 2.Supervisor–Supervisee Dyadic Attachment Combinations
Supervisor Secure Preoccupied (“Top of the Circle” challenges”) Dismissive (“Bottom of the Circle”) challenges Unresolved
Secure Both are comfortable with exploration and seeking/providing support. Supervisee is fearful of autonomy and seeks constant affirmation from the supervisor. Supervisee is uncomfortable seeking support for fear of being perceived as incompetent. Supervisee is likely to be unpredictable; sometimes seeking support and other times withdrawing. The supervisee’s perception of the supervisor is also likely to change unexpectedly.
Preoccupied (“Top of the Circle” challenges) Supervisor is anxious to provide autonomy to their supervisee and is likely to be overly directive. Supervisee may feel stifled or stunted in their ability to grow. Supervisor who is uncomfortable with allowing autonomy is paired with a supervisee who fears autonomy. The supervisee may feel grateful for the support. Supervisor may provide too much support and guidance for the supervisee’s comfort. Supervisee is likely to feel stifled or stunted, similar to one with a secure attachment. Supervisee is likely to oscillate between experiencing the supervisor as supportive and too controlling. When the supervisor sets boundaries, they may experience retaliation from the supervisee.
Dismissive (“Bottom of the Circle” challenges) Supervisor is more comfortable with a supervisee’s bids for independence than their need to be supported or comforted. Supervisee may feel confused by the lack of support or question their competence. Supervisor who is uncomfortable with “neediness” is paired with a supervisee who is afraid of autonomy. The more the supervisor pushes for the supervisee’s independence, the “needier” the supervisee may become. Both supervisor and supervisee are uncomfortable with seeking/providing support. On the surface, things may appear to be working well but the supervisee may not be able to achieve the level of insight required to thrive in their role. Supervisor who is uncomfortable with affect is paired with a supervisee whose behaviours and perception of the supervisor are unpredictable and volatile. The relationship is likely to be tense, with the supervisor providing the minimal amount of support needed to pacify an unpredictable and reactive supervisee.
Unresolved Supervisor longs for intimacy while also fearing and rejecting it. Supervisees may experience the supervisor as reactive and unpredictable. A securely attached supervisee may feel responsible for managing the affect in the relationship. This can lead to resentment or avoidance on the part of the supervisee. A preoccupied supervisee may initially be responsive to this presentation as it reinforces a level of dependence they find appealing. This relationship is likely to deteriorate as the supervisee realises that the supervisor cannot contain or organise the supervisee’s feelings. Supervisee is likely to feel uncomfortable with the expectation of managing their supervisor’s affect and may avoid supervision altogether. The relationship is characterised by misattunement, resentment, hostility, and confusion.

Below is a theoretical exploration of how these combinations of attachment categories might present in the supervisory dyad. Please note where secure attachment is indicated, that also includes the category of earned security.


The easiest and least complicated of these are dyads in which the supervisor and supervisee each possess a secure attachment state of mind. When a securely attached supervisor is matched with a securely attached supervisee, the relationship tends to flow easily. The supervisor is comfortable allowing the supervisee to venture and experiment as they gain new skills, and the supervisee is equally comfortable with testing newly learned skills or applying newly acquired insights. Both supervisor and supervisee are equally comfortable with the supervisee needing support and encouragement in moments of confusion or discouragement. The supervisor is able to operate as an idealised figure, at times allowing the supervisee to borrow their ego strength. Thus, the supervisor is comfortable allowing the supervisee to take the lead whenever possible and only taking charge when necessary. The supervisee also experiences the supervisor as a secure base.

Secure–Dismissive, Secure–Preoccupied, Secure–Unresolved

More complicated and requiring a differential response depending on the presenting issues are the other dyad attachment categories for an insecurely attached supervisee matched with a securely attached supervisor. A securely attached supervisor matched with a dismissive-avoidant supervisee will spend much time dealing with “bottom of the Circle” issues (see Table 2)—that is, a supervisee who tends to be uncomfortable asking for help or seeking support. The supervisee fears that any display of insecurity or uncertainty may communicate a lack of competence and be met with rejection from the supervisor. Conversely, a secure supervisor matched with a preoccupied supervisee will be dealing with “top of the Circle” challenges. That supervisee may be fearful of venturing out on their own, seek constant affirmation from the supervisor, and experience a supervisor’s push towards autonomy as rejection. These interactions will likely manifest themselves in the supervisory relationship and also in the work with the clients. Working with a supervisee who fits an unresolved category will probably be the most challenging in all cases. However, in the best-case scenario, that supervisee is matched with a secure supervisor who is attuned to the supervisee’s needs and is able to respond in a manner that contains their anxieties and projections and is able to move the work forward. Examples of all these permutations are presented in a later section of this paper.

More challenging are the supervisor–supervisee relationships in which the supervisor is insecure. When the supervisor struggles with the top half of the Circle, the bottom half of the Circle or with the role of hands, the supervisee will be affected. Lacking a secure base from which to explore the clinical encounter or a safe haven to return to can lead to challenges in supervision. Using the Supervisor’s Circle of Security diagram (Figure 2) to track those challenges may be helpful. These types of combinations each present a special set of challenges that are discussed briefly below.

Dismissive–Secure, Dismissive–Dismissive, Dismissive–Preoccupied, Dismissive–Unresolved (Supervisors With “Bottom of the Circle” Challenges)

A supervisor with a dismissive/avoidant pattern is often much more comfortable with the supervisee’s bids for independence than with the latter’s needs to be supported, comforted, and helped to organise their feelings. Such a supervisor has a relatively deactivated attachment system, making them more comfortable focusing on thoughts rather than feelings, and thereby keeping themselves emotionally distant. The anxiety experienced by the supervisor when the supervisee is vulnerable or willing to expose weakness may be intolerable. These supervisors tend to prefer to work with supervisees who are more seasoned and less needing of support. In the event that they must work with a less experienced supervisee, they may state that they are supporting the supervisee’s bids for autonomy when in fact they are not providing a level of support commensurate with the supervisee’s professional developmental stage. If this supervisor is paired with a securely attached supervisee, that supervisee may become bewildered at the lack of support they are experiencing, especially concerning “bottom of the Circle” issues. The secure supervisee may begin to question their competency and experience the supervision as one-sided. If a dismissive-avoidant supervisor is matched with a like-minded dismissive-avoidant supervisee, the relationship may seem on the surface to be working well. Both members of the dyad connect in the top of the Circle in the world of ideas rather than venturing into feeling states. Unfortunately, however, the supervisee will experience a failure-to-thrive situation, in which they will not be able to grow professionally or gain new insights about themselves and the way they approach their work. When a supervisor’s dismissiveness is met with a supervisee’s preoccupied attachment style, the situation may become complicated. A supervisor uncomfortable with a supervisee’s perceived “neediness” paired with a supervisee who is fearful of independence and who constantly needs the reassurance and support of their supervisor will result in both feeling dissatisfied. The more the supervisor pushes the supervisee towards greater independence, the more “needy” the supervisee may become.

Finally, a supervisee with an unresolved attachment style is likely to be unpredictable. Sometimes they will seek help from their supervisor to regulate their affect, but at other times they may withdraw. They may feel hopeless, inadequate, helpless, and at risk of losing control. They may also oscillate between feeling that their supervisor is perfect, competent, and capable at times, and controlling, repressive, and difficult to manage at other times. When paired with a dismissive-avoidant supervisor who has little patience for identifying and managing strong affect, it is likely that at best this relationship will be confusing to both. Again, the more the supervisor avoids the affect and what is being played out in the relationship, the more the supervisee will push and demand that their needs be met. If the supervisory relationship is short term, the supervisor may provide the minimal amount of support required to pacify the supervisee and breathe a big sigh of relief once the arrangement is concluded. Unfortunately, neither will have benefited from the encounter.

Preoccupied–Secure, Preoccupied–Dismissive, Preoccupied–Preoccupied, Preoccupied–Unresolved (Supervisors With “Top of the Circle” Challenges)

A preoccupied supervisor is likely to feel significant anxiety providing their supervisee with any level of independence. They have a relatively hyperactivated attachment system rendering them sensitive to rejection while also struggling to regulate emotions. They fear what might happen to the supervisee if the supervisee were given freedom to make their own decisions and learn from the outcomes. Supervision feels best when they can instruct the supervisee on the actions to take in the session; the supervisee goes out to do what they were instructed to do, and upon their return and reporting of the outcome, the supervisor can comment on their success or lack thereof and send them off again with new directives. Often, supervisors fitting this type of presentation speak of wanting to be supportive of their supervisees, of being mindful of liability issues, and of ensuring their supervisees feel comfortable with the work before letting them work independently. Although these are all commendable sentiments, if the supervisor is stunting the growth of the supervisee, the supervisee’s capacity to explore with clients will be constrained. A supervisee with a secure or dismissive attachment style will feel stifled; one with a preoccupied attachment style will feel grateful for the level of support; and one with an unresolved attachment style will oscillate between appreciating their supervisor and complaining that they are too controlling. A preoccupied supervisor may respond by making themselves available at any time, offering additional supervisory hours, and providing much needed regulation, but when their resources become depleted and they begin to set boundaries, they are experienced by the supervisee as abandoning and negligent, and the supervisee may even complain about them internally or to a regulating or licensing body (Shedler, 2006). The supervisee with an unresolved attachment style now has turned the passive into active in an attempt to gain mastery over a likely recurring theme in their relationships.

Unresolved–Secure, Unresolved–Dismissive, Unresolved–Preoccupied, Unresolved–Unresolved

Perhaps most concerning, and likely less frequently encountered, is a supervisor with an unresolved attachment style that has not benefited from reparative work to become earned secure. However, since not entirely out of the question, one should consider the potential consequences of such a situation. An individual whose attachment style is predominantly disorganised and unresolved will likely have problems with boundaries, which may manifest in difficulties with beginning and endings. Intimate relationships are both longed for and feared, setting up a bind for the person that results in them always pushing away the people they most need (Holmes & Slade, 2018). Holmes and Slade (2018) recommend attending to what may seem insignificant or peripheral elements of interactions, such as touch and gaze, with great care because these may carry particular meaning for this group.

The supervisor with an unresolved attachment state of mind may well seek affirmation from their supervisees while at the same time being sensitive to threat. Ultimately, the supervisee may experience the supervisor as reactive and unpredictable. There may be role reversal in the relationship, whereby the supervisee holds the responsibility for managing the affect in the relationship. Consequently, this provides the supervisor with a space to organise themselves, while simultaneously the supervisee feels neglected by the supervisor. A securely attached supervisee may be able to take on that role, albeit reluctantly, but might feel resentful at not being given appropriate support, or alternatively they may view this dynamic as problematic and refuse to engage in it, resulting in significant conflict in the relationship. A dismissive-avoidant supervisee will probably feel uncomfortable with this arrangement and may avoid supervision altogether. A supervisee with a preoccupied attachment style will likely initially be responsive to this presentation because it reinforces a level of dependence in the relationship that may appeal to them. However, as the supervisor will in fact be preoccupied with their own needs and not those of the supervisee, this relationship is likely to deteriorate over time as the supervisee realises that they cannot use the supervisor as a container or organiser of their feelings. Worst of all is the combination in which both the supervisor and supervisee have an unresolved attachment. Misattunement, resentment, hostility, and confusion will presumably abound.


The following are vignettes that may be used as part of a more extensive supervision training. Participants should read the vignettes and discuss how various supervisor–supervisee dyads might respond to the presenting issues based on their dominant attachment system(s). An example is provided for the purpose of illustration. One should note that interactions are typically much more nuanced and of course include more than attachment style as a contributor to the interactions. For example, general temperament, level of stress or exhaustion, dominant defences, and historical contexts all play a role in how events will unfold. Thus, it is understood that these examples are oversimplifications and intended as a way of focusing through a particular lens, in this case that of attachment, and exploring how it may inform the situation.

Vignette 1

A supervisee practitioner is errant with his paperwork. He has sent the supervisor the same formulation for the third time for review, and there continue to be significant problems with it. The supervisor is feeling irritated with him.

Examples of Possible Responses to Vignette 1


The supervisor notices her own irritation and wonders why she is feeling this way. Is her level of irritation commensurate with the offence? Does the supervisee typically require that level of oversight? In this particular situation, upon reflection, the supervisor realises that this behaviour is highly unusual for this supervisee and that her own level of irritation far exceeds the situation. The supervisor then reviews the documentation again and realises that it needs some edits, but they are minor. The supervisor also realises that this particular case is hitting close to home, and therefore her annoyance may be more a reflection of her own defences and projections. In a follow-up supervision, when the supervisee enquires about the status of his formulation, she provides the supervisee with a couple of suggestions for edits, which the supervisee gladly accepts.


The supervisor is uncomfortable examining her own internal state and affect and thus is unable to make the connection between her own history and what this case is engendering in her, nor is she able to assess the validity of her reaction. The supervisor impatiently questions the supervisee about the formulation. The supervisee, who already feels quite insecure about his work, obtains further validation that he is underperforming. The supervisee also fears that the supervisor will refuse to write him a letter of recommendation he is hoping to receive. The supervisee begins to cry, apologises for his poorly written formulation, and promises to do better. Upon returning to his desk, however, the supervisee feels lost and does not know where to start or what to do to please his supervisor.

Here are some additional vignettes for consideration. While reading the vignettes, consider the following:

  • How do you imagine different attachment category combinations in the supervisory dyads might respond to this situation? Pick two possible combinations and imagine how things might play out.

  • Are there particular combinations for whom this situation would be particularly challenging?

  • How might you as a supervisor respond to such a situation? What new insights are emerging for you?

  • The vignettes are intentionally sparse regarding details about the supervisor and supervisee. Are there additional details that would assist in making sense of the interaction? How would these intersect with attachment categories?

Vignette 2

A supervisory dyad has experienced a particularly difficult supervisory hour in which the supervisee challenged the supervisor’s approach to the work. During the following session, the supervisee spends much time discussing a case in which he feels the parents are questioning his work and are being belligerent.

Vignette 3

A counsellor presents a case during group supervision and asks for advice because he feels stuck. This is a case of two adolescent boys who are living with their maternal grandmother. Their mother has been in and out of a drug rehabilitation program and is generally unavailable to them. The counsellor discloses that he dreads meeting with the grandmother, although he typically enjoys his work and has great engagement with the families he serves. He is concerned by the grandmother’s lack of insight into the children’s behaviour and intolerance of it. He describes her as behaving intrusively and inappropriately with the children. Given her own trauma history, the grandmother feels that the children should “get over it”. In group supervision, various members make suggestions about the grandmother’s possible resentment of having to care for her daughter’s children and are disregarded by the counsellor. The supervisor suggests taking the grandmother’s perspective and is dismissed and told that this is not helpful. The supervisor feels ineffective.

Vignette 4

A counsellor who performs home visits as part of her work for a non-government organisation is struggling with a case. The mother of a 2-year-old who is living with her mother (the child’s grandmother) is at risk of being evicted. The supervisee counsellor is frustrated with the mother regarding her hoarding behaviour and the state of disarray of the apartment. The supervisor recalls that this supervisee was similarly frustrated with another family when the condition of their home was unsanitary. The supervisee reports that she has attempted repeatedly to encourage the mother to tidy up the apartment, even offering to clean the apartment with her, but the mother has refused.

Vignette 5

During their second supervisory hour, a supervisee begins to tell her supervisor that she is having a difficult time with a particular case because of her own history of sexual abuse.

Vignette 6

During supervision, a new practitioner provides extremely brief responses to enquiries about cases, such as, “Everything is fine!” and “They are coming along nicely”. The supervisor often feels exhausted after each supervision, having tried hard to engage the practitioner.

Vignette 7

A therapist presents their supervisor with a case of a 3-year-old boy removed from his pre-adoptive home to live with his paternal grandparents. The child has made allegations that his grandfather has “monsters that tickle his penis” and, despite this, child protective services have removed him from his pre-adoptive home and placed him in the grandparents’ house. The supervisor is feeling increasingly agitated and angry as they hear about this case. The therapist seems very sad.


Although it is unlikely that all supervisors and supervisees will engage in an AAI to categorise their attachment status accurately, the concepts presented in this paper can still be usefully employed to create a practical framework that can assist with making meaning of supervisory relationships and also point to possible interventions in the face of challenging dyadic combinations. Ideally, research should be conducted to test the ideas presented in this paper, such as the prediction of certain experiences depending on the attachment styles of the supervisor and supervisee.

As noted earlier, attachment is only one of many variables that influence and inform the supervisory relationship, so care should be taken not to oversimplify complicated human interactions. As Yalom (2020) emphasised, “Labels do violence to people. You can’t treat labels; you have to treat the people behind the label” (p. 17). Therefore, attachment classification is but one such label that can be applied when it provides insight into a situation. Supervisory relationships, much like any other relationship, are best approached from a place of curiosity, wonder, and respect.


The author wishes to express gratitude to Neil Boris, M. D., research liaison and medical director of Circle of Security International, for his thorough review of the manuscript.

  1. See the Handbook of Attachment: Theory, Research, and Clinical Applications (Cassidy & Shaver, 2018) for a thorough review of attachment theory, historical perspectives, biological perspectives, attachment in infancy and childhood, attachment in adolescence and adulthood, and clinical application of attachment theory, as well as emerging topics and perspectives.