Progress through the Early Stages of Assimilation in Play Therapy with a Traumatized Six-Year-Old Girl

ABSTRACT We applied the assimilation of problematic experiences sequence (APES) to a six-year-old girl’s processing of traumatic experiences involving violence and death in play therapy. We analyzed the post-session notes from the first 34 sessions of a much longer treatment, during which the girl repeatedly enacted a drama we called the cottage play, involving characters assumed by the child and characters assigned to the therapist. We distinguished four phases based on changes in play themes. In phase 1, the girl expressed her need for safety in response to an overwhelming internal threat (APES stage 0, warded off/dissociated). In phase 2, she worked to escape and avoid the threat, referenced in the play as monsters and bad memories (APES 1, unwanted thoughts/avoidance). Phase 3 was a brief period of consolidation. In phase 4, she actively faced the trauma, referenced as murderers, soldiers, and death (APES 2, awareness/emergence). Our observations underlined the child–therapist collaboration and dyadic processing. The expressions of the problematic experiences suggested increasing but limited assimilation (stages 0 to 2 out of eight stages). The assimilation model usefully described symbolic processing in play therapy, and the results pointed to tentative elaborations in APES stage descriptions.

In the assimilation framework, personal experiences leave traces that can be reactivated. The metaphor of voices is used as an alternative to the experiences/schemas to underscore the agentic nature of these traces Stiles, 2011). The voices actively strive for expression when they are addressed by circumstances resembling the original experiences. The expression of voices can be seen in the person's verbal expressions, feelings, thoughts, and actions (Osatuke et al., 2011). Normally, experiences are assimilated and the voices become members of a coordinated community in which voices of the person's experiences are smoothly accessible for dealing with life's demands, forming the person's usual self. In contrast to well-assimilated traces, some painful or incongruent experiences remain unassimilated. The voices of such experiences are internally experienced as problematic and distressing, as they were and are unacceptable, traumatic, or somehow in conflict with the dominant voices (Osatuke et al., 2011). Strong negative affects cause people to ward off or avoid those voices (Stiles et al., 2004), but although not smoothly accessible, the problematic (e.g., traumatic) voices may nevertheless respond when addressed by circumstances causing distress and maladaptive behavior.
Theoretically, the process of psychotherapeutic change involves increasing access to, acceptance and integration of the problematic experiences/voices. Progress through the APES can be achieved by building semiotic meaning bridges, which are signs (words, images, gestures) with similar meaning to the author and the addressee. Author and addressee may be different people or different voices within one person. Meaning bridges are words and other signs that express the problem and are mutually understood, perhaps first by the therapist and client and eventually, among the client's internal voices (Stiles, 2011). Building meaning bridges is observable in a client's verbal and nonverbal expressions, making it possible to track the progress of the problematic experience through the APES continuum.
Most assimilation model research has focused on adults and used clients' verbal expressions to track the evolving relation to the problematic experience. However, theoretically, expressions by a problematic voice may include any expressive signs (e.g., digital imagery; Van Rijn et al., 2019), and accordingly, we supposed that a child's play can be seen as expressions by voices, including warded-off voices expressing themselves in ways not recognized by the dominant community. In the Table 1. Assimilation of problematic experiences sequence.

APES 0
Warded-off/ dissociation Content is unformed; client is unaware of the problem. Distress may be minimal, reflecting successful avoidance.

APES 1 Unwanted thoughts/ avoidance
Content is distressing thoughts. Person prefers not to think about it; topics are raised by therapist or circumstances. Affect is more salient than the content and involves strong negative feelings (anxiety, anger, sadness).

APES 2
Awareness/emergence Person acknowledges the problematic experience and describes distressing thoughts, but cannot formulate the problem clearly. Affect includes acute psychological pain associated with the problematic thoughts and experiences.

APES 3
Problem statement/ clarification Content includes a clear statement of a problem -something that can be worked on. Affect is negative, but manageable.

APES 4 Understanding/insight
The problematic experience is placed into a schema, formulated, and understood with clear connective links. Affect may be mixed, like unpleasant recognitions, but also with curiosity, even pleasant surprise.

APES 5
Application/working through The understanding is used to work on a problem; there are specific problem-solving efforts. Person may describe considering alternatives or systematically selecting courses of action. Affective tone is positive and optimistic.

APES 6
Resource/problem solution Person achieves a solution for a specific problem. Affect is positive, satisfied, proud. As the problem recedes, affect becomes more neutral.

APES 7
Integration/mastery Person successfully uses solutions in new situations; this generalizing is largely automatic, not salient. Affect is neutral.
therapeutic collaborative process of play, experiences are expressed and processed through naming, describing, and trying out alternative behavioral and experiential solutions using different characters, understood as symbolic representations of inner voices. As in adult therapy, collaborations would be most likely to succeed if they remain within the therapeutic zone of proximal development (TZPD), the region between the problem's present APES stage and the stage that the child can achieve in collaboration with the therapist (Leiman & Stiles, 2001;Stiles et al., 2016). In successful therapy, the TZPD gradually shifts higher on the APES as the representations of painful experiences gradually become assimilated, that is, the problematic voices become assimilated into the dominant community. This assimilation process may be inferred from observations of the child's play.

Approaching fearful emotions through play in therapy
The assimilation model is consistent with views suggesting that symbolic play offers the child means to express and explore reality, thoughts, feelings, wishes, and relationships (e.g., Winnicott, 1971). In therapeutic settings, especially in the case of traumatic experiences, the child may play out concerns using play to reenact earlier experiences, digest them mentally, and to gain a sense of control over them (Campbell & Knoetze, 2010;Cockle & Allan, 1996). Traumatic experiences are known to influence play behaviors (e.g., Sossin & Cohen, 2011), and play therapy has been found effective in helping children overcome and manage their traumatic experiences (Ryan & Needham, 2001;Slade & Warne, 2016). Posttraumatic play (PTP) has been characterized as intense, grim, devoid of joy and pleasure, somber, compulsively repetitive, ritualistic, lacking an as-if quality, having increased aggressiveness and withdrawal or avoidance, concretization and reduced symbolization (e.g., Cohen et al., 2010;Terr, 1983). The assimilation account concurs with the view that PTP has an unconscious link to the trauma (Terr, 1981), and consequently, the contents and characters of the play can be understood as projections of child's inner reality. At early APES stages, the metaphors of symbolic play can provide a safe distance, allowing the child to explore and gradually approach the threatening feelings and painful experiences. The meanings of these symbolic representations may be discerned when the contents of several play sessions are consolidated into play themes summarizing the main issues the child expresses. Play themes can be viewed as abstractions and inferences made by the therapist. They can represent significant, underlying, emotional issues expressed by the child (Ryan & Edge, 2011). Changes in the themes may then be understood as showing the development, empowerment, and healing of the child in (Cockle & Allan, 1996). The end or shift of a play theme can be interpreted as an indication that the meaning of the traumatic experience for the child has changed and thus that he/she may have moved toward adjustment (Landreth, 2012). In the present study, we pursued this reasoning, investigating change in play themes as possibly reflecting progress in the assimilation of traumatic experiences and analyzing whether development can be tracked and described using the APES.
We also focused on the child-therapist collaboration during the play. We understood the process of theme building as a dynamic co-construction dependent upon both the child's activities and the therapist's actions during play (Ryan & Edge, 2011), a joint effort at building meaning bridges. We focused on the characters assigned to the therapist by the child and on the characters played by the child. In assimilation model terms, roles assigned by the child to the therapist can be understood as enactments of the child voices, an instance of the suggestion that voices may be embodied in other people's activities (Mosher & Stiles, 2009).

Aims of this study
In this qualitative theory-building case study, we examined the process of assimilation in a year-long episode in the play therapy of a traumatized six-year-old girl. We investigated the theoretical suggestion that the play characters, that is, the roles taken or assigned by a child, can be analyzed as expressions of internal voices and that the assimilation of problematic voices would be manifested in play. We used changes in the play themes to infer changes in APES levels. So far as we know, this is the first published case study applying the assimilation model to play therapy, so one aim was to assess the feasibility and appropriateness of this extension of the model.
Qualitative theory-building case studies do not test isolated hypotheses, effectiveness of a treatment, or restrict attention to a few selected variables. Instead, they compare many aspects of a theory to correspondingly many aspects of case observations (Campbell, 1979). Details that might be ignored in statistical hypothesis-testing research, including unexpected features and events, can be treated as evidence, supporting the theory (or not) or demanding a coherent account that is logically consistent with the rest of the theory (Stiles, 2007(Stiles, , 2017. We aimed to compare our case observations with the assimilation model to assess the fit and to identify needs and opportunities for elaborating the theory. Thus, we do not aim at testing the effectiveness of the treatment, although the use of the APES as an outcome measure has been supported in nomothetic studies (e.g., Basto et al., 2018;Detert et al., 2006) and in case studies (e.g., Mendes et al., 2016).

Lisa, her early childhood history, traumatic experiences, and her therapy
Lisa was a six-year-old girl living in a foster home. The early years of her life had been rather unstable with several caretakers (mother, grandparents, shelter home). As social services official had judged Lisa's mother as immature and incapable of taking care of her child, Lisa and her mother had spent several months in a Mother and Child Shelter Home to support mother's parenting. Lisa's father was mainly absent. At the age of three, Lisa was a victim of a violent incident in which her mother was shot and killed by her then-current boyfriend, who also killed himself. One bullet harmed Lisa. After the incident, Lisa spent several days alone with the corpses. We refer to this incident as "the shooting".
The experienced trauma was complicated, as Lisa witnessed a violent act, lost her mother, and was herself in danger and alone. No formal diagnostic assessment was done; however, in the light of her insecure early attachment history, we inferred that Lisa probably experienced a complex trauma (Cook et al., 2017) or Type III trauma (Terr, 1991), defined as caused by traumatic event in the presence of earlier difficulties in the attachment relationship (severe neglect). There were no reports that Lisa had been physically or sexually abused.
After the incident, Lisa was taken into custody and placed in a temporary foster family for a year. At the beginning of the therapy, she had lived for about a year in a permanent foster family. In addition to the foster family, she had occasional contact with her biological father and half-brothers, as well as her biological maternal grandparents' family comprising several still under-aged or young adult uncles and aunts living in the same apartment.
Lisa was referred to therapy by the healthcare district. The therapy was conducted by a private practitioner. It took place once a week, with sessions lasting 60 minutes. The treatment was openended; however, for this study, we chose to analyze an episode comprising the first 34 sessions, as explained later. The therapist (this paper's first author) was a 50-year-old female clinical psychologist originally trained in child CBT and later as an integrative psychotherapist. She used an integrative approach including child-centered play, drawing important episodes (e.g., cartoon strips describing an argument with the foster mother) and themes difficult to share in words (e.g., Lisa's family tree). She discussed and provided psychoeducation about the experience of being in a foster home (e.g., reading and playing board gamed dealing with the topic). However, her attempts to discuss or provide psychoeducation concerning the shooting were mainly rejected by Lisa. At the time of the sessions studied here, the therapist was not familiar with the assimilation model.
When the therapy began, Lisa had problems sleeping alone and prolonged temper tantrums, and she was demanding, oppositional, hyper-vigilant, and sometimes physically aggressive. There were no formal outcome measures or systematic observations of Lisa's behavior at home; however, after the first year of therapy, the foster mother reported a decrease in the frequency and length of temper tantrums and a lessening of sleep problems.

Assimilation of problematic experiences sequence (APES)
The APES is an evolving description of the relation of a problematic experience or voice to the person's dominant community of voices (Stiles & Angus, 2001;Stiles et al., 1991). The sequence is anchored by the eight stages or levels, numbered 0 to 7 (Table 1). The APES is considered as a continuum with possible intermediate stages (Brinegar et al., 2006). Clients may enter treatment with problems at any stage, and any progress across the stages may be considered as improvement. Although the APES can be used as a rating scale, we do not perform formal ratings but used it to conceptualize and convey our interpretations precisely.

Selection of sessions
In this study, we focused on an episode comprising Lisa's first 34 therapy sessions in an open-ended treatment covering the whole first year of her therapy, from fall until the summer break. Lisa spent the predominant part of these sessions in a role-playing activity: a child-directed drama with partially scripted dialog, improvised set-building, and stage directions, that evolved across the treatment. We call it "the cottage play." In this play, Lisa built a cottage from sofa cushions, a folding screen, chairs, and blankets. She and the therapist then enacted roles directed by Lisa. It should be noted, that Lisa engaged also in other activities (e.g., drawing, board games, reading, and dialog), but they were not analyzed for the present study. Lisa played with the dollhouse nine times, but no consistent theme (see definition below) could be identified in these short and rather chaotic play moments as the play never evolved -perhaps due to its strong emotional loading or because it was cognitively too demanding for her. However, it is worth mentioning that in five sessions during the first phase, Lisa named one of the dollhouse dolls as the violator and in three of them quickly changed it to be the foster father. The therapy continued after the summer break with the same therapist, but Lisa did not systematically continue with the cottage play, although she occasionally returned to it during the following years. For this project, we restricted our focus to the cottage play episode occurring in the first 34 sessions.

Qualitative analysis
The therapist wrote a description of Lisa's play after each session. These hand-written post-session notes were transcribed, and for reporting this study's results, relevant words and passages were translated from the original language of the therapy (Finnish) into English by the therapist. The study had ethical approval from the Finnish Social Insurance Institution, and the legal guardian gave an informed consent.
First, we focused on content of the sessions, looking for recurring events or topics in the therapist's descriptions. To distinguish themes in the cottage play, we first identified the following components in each session description: (1) the main events of the play (e.g., building a cottage, appearance of monsters or murderers), (2) Lisa's character in the play (e.g., self-sufficient teen-ager, child), and (3) the therapist's character assigned by Lisa (e.g., teen-ager, lost child). A theme was distinguished if all three components occurred together at least twice. A change of theme was distinguished if there was a change in at least one of the three components. Next, in joint discussions among the authors, the identified themes were named and analyzed to understand their psychological meanings. Based on the themes, the treatment was divided into phases, reflecting major changes in the psychological meanings of the play. Second, to further understand the change process, we analyzed the characters appearing in each phase to distinguish expressions reflecting Lisa's problematic experiences, that is, expressions by her problematic voices. Third, we used the APES to interpret each phase: how the theme meanings evolved across phases and how the changing expressions of problematic voices could be understood as reflecting different levels of assimilation.

Results
We focused on what we considered to be Lisa's central problematic and traumatic experience: The shooting and the associated events. We interpreted the cottage play from the perspective of how it bore on that experience keeping in mind that Lisa's insecure childhood experiences would also have bearing on the play. We distinguished four phases within the 34 sessions. In the following sections, the main play events, characters played by Lisa, and characters played by the therapist in each phase are summarized in italics, translating as closely as possible from the post-session notes. After the summary of each phase, we present our interpretation of the events, the identified voices and problematic experiences, and the predominant APES stage of Lisa's central problematic voices in that phase.

Interpretation of phase 1
The emotional atmosphere during phase 1 was one of fear and loss. L assigned the roles of characters of the same age (teenage) to both herself and T. This allowed L to take an active, dominant role, leaving T in a submissive role. In this phase, T had no agency; she was helpless and directed by L, who gave her instruction on how to act homeless, lost, and fearful until being given a home by L, who acted as a security provider. L's character expressed self-sufficiency. Her dancing was experienced by T as sensual and flirting, not age-appropriate. L seemed to identify with the immature behavior of her deceased mother with no worries, or perhaps with her aunts in their late teens. This reflected her understanding that being mature means not only being agentic, but being attractive, flirting, dominant, and fearless. This was perhaps her only strategy to cope with helplessness and insecurity. This behavior resembled her conduct in real life, where she was often bossy and disobedient. It might reflect her experience of being on her own, the one responsible for her own safety, with a belief that she could not be needy or depend on other. We inferred that her feelings of fear and being lost were dissociated (warded-off) and that she externalized them to the therapist as a means of self-regulation, so she could express them without owning them. Similarly, the play events were distanced from L's daily experiences by placing them in an imaginary cottage (not home), where an imaginary but still familiar and secure person (friend, not mother) offered shelter to fearful T. During the play, T articulated the experiences and emotions related to being alone, homeless, and lost. Some of these expressions were not explicitly mentioned by L, but she observed them carefully. Not commenting was interpreted as "a hit," that is, a good-enough reflection of L's emotional experiences.
We called the first phase "Enacting need for safety and self-regulation" as it seemed that L was working on the issue of being afraid and fearful. She did it symbolically and by proxy, through the emotions expressed by the therapist. We interpreted that L could not feel nor express these emotions verbally or even play herself a fearful child, but needed T to reach and express them. We interpreted L's eagerness and decisiveness in making the therapist feel fearful and her intensive engagement in the same play events over and over again for 17 weeks as an indication of the importance of these emotions and experiences for her.

Voices and problematic experiences identified
L's character was interpreted to express the voice of self-sufficient and fearless teenager who acted as stage manager and director whereas the voices of fear and being lost were expressed through the therapist's character. We understood the latter as strands (elements, derivatives) of her experience of the shooting, along with her early childhood insecurity. It is noteworthy, that T was given a role of a teenager instead of a child; perhaps the voices of being lost and fearful were too scary to be given for a child. We understood that she expressed experiences of "being fearful," "being homeless," and "not having a parent to take care of her." We interpreted this phase as an indication that the problematic voice of the fearful child without a caretaker had been addressed in the therapeutic relationship and was striving to respond, to communicate the importance of the emotions and experiences related to lost, fear, and homelessness. However, these voices were disconnected from the demanding and mature teenager, which we supposed to reflect L's dominant community of voices as self-sufficient relying only on herself. This position was manifested in L's bossy and disobedient behavior at home. It probably pertained to her experiences as child having been left without a safe, caring adult, and to her recent experiences of not having a permanent home, and perhaps derived in part from her reaction to experiencing helplessness alone with the corpses.
Theoretically, we can speculate that the problematic voice related to the shooting was addressed by the circumstances. In the safety of the therapy room, this voice responded, but in a way that the rest of L's community of voices did not recognize. L seemed to have no explicit awareness of the play being linked with her experience. We supposed that the experiences related to the shooting produced unbearable feelings of terror and helplessness, so that even the voices of age-appropriate neediness, could not be accepted by L's dominant community. The repetition of the play can be seen as reflecting lack of a meaning bridge; the problematic voice repeated itself because the dominant community did not understand.
We judged that this first phase suggested that the personal, terrifying experience of the murder and Lisa's time alone with the corpses were mainly unassimilated. She indicated being aware of the actual incident and her history; for example, she included the violator in her play with the dollhouse. However, the problematic voices related to the shooting could not yet be acknowledged or integrated to her community of voices. Although this community did not recognize the source, they felt the threat was apparent enough to trigger the search for safety and the self-sufficiency enacted in the play. The externalization to T and the failure of L's dominant community to appreciate the content indicates APES 0 (warded-off; Table 1).

Phase 2 (sessions 18-25): Locking memories and monsters behind locks (APES 1)
T is a fearful child lost in the forest. L is still 17, but she acts like a secure adult comforting T, e.g., by singing her a lullaby. They go to a "Safe-nest" to eat a snack. In session 19, L is also a child, whose parents are divorced. Later, L says the parents are dead, and claims that she does not want to play anymore. During that and the following sessions (20)(21)(22), L lives in a "Castle-of-joy" (not in the cottage) and has a "Safe-bubble" which has three locks keeping the monsters and bad memories away. She says: "If they come, they can tear a little girl into thousand pieces." Also during session 20, L first says that her parents have divorces, but later she says that the mother is dead, after which she wants to terminate the play. When T asks whether this reminded her of her own mother, she acknowledges that. In session 23, T and L are sisters. L lives in "Castle-of-sadness" and has magic powers that can make bad memories disappear. L moves memories behind the door and 100 locks. She concretely moves a folding screen in front of the therapy room's door and covers the hole in the screen with a blanket. Before doing so, she climbs on a chair and looks from the "window" (hole of the screen) to see if the memories are visible behind the screen.

Interpretation of phase 2
Phase 2 involved changes in both the play events and the characters indicating change of a theme. The appearance of memories and monsters as well as places and ways to gain safety or shelter (e.g., Safe-Nest, Safe-Bubble, and locking memories behind the fence and locks) were new elements. No more doing make-up, drinking beer, or motorbike riding occurred. Changes in the characters concerned mainly T, who was given a new role as a child. The major change in L's character was that she became aware of the emotions related to the painful memories and scary monsters and the character worked hard to find safety from them. Although T's role was still to be fearful, she was more and more in a role of a witness for L's actions as L was finding ways to deal with her memories by herself. L offered T physical shelter, comfort and reassurance by singing a lullaby.
L's searching for power from inside could reflect her early childhood with no stable secure adult and/or her traumatic experience related to the shooting where she had no control over events. These may have engendered her precocious self-sufficiency. During phase 2, she regained control, and at this time, it was not by being a bossy and demanding teenager, but more age-appropriately by creating images of safety by having magic powers. The memories and monsters in the play were not named as corresponding to the real life-events lived through by L, nor were the solutions from real life. However, the fears, needs, and solutions occurring in the play seemed to reiterate the original scary experiences in the form of symbolic play; they were disguised enough that she was not faced with the horrible reality.
In session 20, the word "dead" was mentioned, and L briefly admitted the association of the play with real-life events. This dialog seemed to evoke emotions that were too strong; it triggered avoidance, and L could not continue the play. Her avoidance suggests that explicit associations to the shooting were too overwhelming. However, L could protect herself from these emotions by interrupting the play. L's own and growing ability to voluntarily avoid and control the overwhelming experiences with these safety-images can be seen as her strength: she felt safe enough to begin, tentatively, to face the problem as she felt that she has the control over the amount exposure to the scary experiences. We described the theme of phase 2 as "Striving towards ownership of self-regulation while being fearful".

Voices and problematic experiences identified
During phase, the voice of a fearful child seemed more accessible as the emotions were expressed by the child-character played by T and by L's own character. Giving the role of a child to T before playing a child herself (phase 4) suggests that L was edging closer to her problematic experience, observing and processing the idea of a child being fearful, studying how a child could cope with fear, learning something that could become a resource. She seemed to build an observer position, which allowed her to observe her own behavior, and to observe and internalize the therapist.
L singing a lullaby, acting as a secure adult (session 19) and looking for safety (Safe-bubble, Safenest, Castle-of-Joy; sessions 19-22) suggest that the voice of a needy child had become more tolerable and available, and she could find a self-soothing voice in herself. A new voice of a capable and resourceful child was observed as L resolved the problem of scary memories by herself, and there was a suggestion of a promise that the fearful experiences can be tolerated. L did not look for comfort from T, but found again the magic power from herself indicating that the voices of the fearful and needy child and the capable and resourceful child could now co-exists. The needy child still seemed to be rather self-sufficient; perhaps it needed to emerge together with the competent child. The selfsufficiency of the first phase transformed to more age-appropriate voice expressing capability and resourcefulness reflected in the use of magical powers, and a new voice of a needy child appeared. Thus, it seems that emotion of fear became more tolerable and accessible.
We concluded that L came to acknowledge the existence of her problematic experiences related to the shooting, and we named the experience as "I have seen something threatening" and "I have bad memories". However, the threat and the memories were not faced. Rather, L gathered the strength and psychological means to prevent being overwhelmed by them, to control or turn them off or avoid activities that stimulated them. The assimilation of the problematic experience in phase 2 can be seen as working through APES 1 (Unwanted thoughts/avoidance) on way to APES 2 (Awareness/emergence). The play was filled with imagery indicating looking for safety and avoiding feeling insecurity and insufficiency. In actively avoiding, she implicitly acknowledged the importance of the painful memories. However, she could play the role of someone needing safety and resourcefully finding ways to escape. Thus, both fear and comfort were simultaneously present. The cues that triggered avoidance can be considered as an early version of a meaning bridge -signs that showed the problem was there, albeit used to avoid and escape rather than to face and overcome. Apparently, the fearful and needy child found this soothing, demonstrating that solutions for fearfulness and safety could be found in the metaphors of the play.
L did not verbalize her thoughts, feelings, or understandings; thus, we cannot know for sure what she knew and understood. It is our inference that, unlike the therapist, she did not recognize the association between play and real-life. L's refusal to continue the play after "dead" was mentioned (session 20) suggests that the memory of the shooting was still overwhelming, though some of the emotions related to it and to her earlier unsecure childhood emerged, in attenuated form. This refusal can be interpreted as an active avoidance (APES 1) used appropriately for emotion regulation. L's character recognized the existence of scary memories, especially when she was able momentarily to actually experience it in the play as well as to find ways to avoid the pain and fear related to them. L solved the problem as children commonly regulate their emotions: By putting the scary things away from sight and mind, and going to a place where they can feel safe.

Phase 3 (sessions 26-28): Sisters with their own homes (APES 1 toward APES 2)
In session 26, L and T first have a "sister-home", and L and T are sisters living together. L and T are also dancers. Later, T is looking for a home, and finds it soon. In session 27, L wants to build a cottage "more beautiful than ever." T is no longer lost, as she has her own cottage. L builds the cottages mainly alone without showing bossiness anymore. L talks more than before, and she expresses verbally her sadness and anger about some daily matters. She also expresses fear in the play and seeks comfort from her sister (T) by coming to T's house for the night and having an eyecontact with T. In session 28, board games are played.

Interpretation of phase 3
During these sessions, there was new companionship between L and T as sisters and dancers with their own homes. A beautiful cottage was built, and nobody was fearful or lost. L expressed her ageappropriate feelings both in the play and in the short conversations describing her daily life. The emotional intensity was lower than before, and L seemed to calm down and seek security. In effect, she seemed to take some time off from processing after the intensity of phase 2. We described the theme of this phase as "Dealing with separation -searching for connection".

Voices and problematic experiences identified
The voice of fear was less intensive than earlier and mainly absent while the voice of the needy child was more calmly present than before and it also appeared outside the play. No new problematic experiences were identified during phase 3. The general atmosphere was calm and positive suggesting she was taking an emotional break, maybe consolidating the sense of control achieved in phase 2. L's focus on tolerating the painful emergence of needy child and being able to talk about daily sad matters, suggested to us that her central problematic experience of the shooting was in the later part of APES 1. She could control her attention and avoid the pain if she chose to.

In session 29, L and T are sisters living in a cottage that is the most beautiful ever. Their parents are gone away; later L says they died two years ago. When T expresses her fear, L becomes annoyed and says "Fear goes away by doing something happy". She is annoyed when T asks questions about the parents. In session 30, L wants to play school and be a teacher. T is a naughty pupil. In session 31, L is tearful already when she arrives, and she shows her sadness during the session, and allows T to comfort her. No play occurs in this session.
In session 32, L finds a necklace with heart and key from the toy-cabinet. She says that it is a magical tool that "can transform the evil hearts of all murderers into good". L, T, and L's boyfriend go to "icecastle" and prepare new hearts for everyone. All murdered persons become alive again. Soldiers threaten children in the castle, but they go to "safe-bubbles". During the night, L is murdered, and when T goes to rescue her, she is "only sleeping" and no rescuing is needed. L also asks T: "Do you know that opening the heart and taking the sorrow away hurts?" In session 33, the children are again in the castle, and L has magic powers. During the night she goes out to a "troll-mountain", T is not allowed to follow her. L comes back to the castle and says: "There is blood on the mountains because the trolls killed people". T asks about the time of shooting, and L answers her questions. Later, L plays that one famous singer is her boyfriend, but "he is having heart attack and may die".
During session 34, T asks L about the violent incident and the house where it happened, L answers her questions but claims soon that she does not want to talk anymore. At the end of the session, L wants T to be her child. L leaves the child alone during the night. She observes the fearful reaction of T for some time, and then finally calls her back to the cottage.

Interpretation of phase 4
We identified change of the theme and distinguished phase 4 from phase 3 based on new play events as well as on the change in L's character from teenager to a child experiencing danger. Despite facing horrors and death, L was a competent girl who worked resourcefully using magic powers to gain victory and to deal with the painful experiences named as death or as danger related to death. Later, as she played with the idea of losing a boyfriend because of a heart attack, not violently, she seemed not to be afraid of the death, she was mostly sad. This phase was interpreted as indicating that L symbolically faced the idea of a child witnessing and dealing with death and being threatened by murderer. L's character found a solution to this threat as a girl concurring death, soldiers, and trolls causing death. We interpreted the meaning of this phase as "Facing the death".

Voices and problematic experiences identified
During phase 4, L's character expressed the voice of fearful child and voice of the child needing protection as it was in obvious life-threatening danger. She, however, simultaneously expressed the voice of mastery and competence as she faced the dangers. Perhaps, it was still not possible to be fearful and dependent of someone else's soothing and competences. A new voice of forgiving or denial of ultimate badness was observed as L changed all murderers' hearts into good. Strands of the central problematic experience expressed included "I have seen death", "I have been in danger", and "I was helpless". In this phase, L's symbolic processing seemed to indicate mainly APES 2 (Awareness/ emergence) as she faced the idea that a child can be in mortal danger and can experience the painfulness of violence and horror related to killing and blood. L's character felt self-efficacy and control. It has been claimed that the healing potential of play depends on the child's perception of control over consequences, his/her ability to process cognitively the trauma and to express forbidden feelings (Nader & Pynoos, 1991). Perusing the role of a competent child and identifying with such a character, L could change the passive victim role into an active one, thus gaining a sense of control and self-efficacy as she found solutions to the horrifying play sceneries. She also expressed feelings of fear and even dislike toward the bad soldiers. L was willing to talk a little bit about the actual traumatic incidence as it happened in reality, and this dialog was between real L and T suggesting momentary progress toward APES 3 (problem statement/clarification).

Discussion
Lisa's play during the first 34 therapy sessions had content typical for PTP, including death, instability, concern with safety, and fantasies of rescue Sossin & Cohen, 2011). In overview, the child-directed drama we called the cottage play evolved across sessions through four phases distinguished by changes in the play events, the character assumed by Lisa, and the character assigned to the therapist. We interpreted the phases as reflecting the problematic (traumatic) experience's progress through the first three stages of the APES, from warded off to emergence. Theme-based voices identified in the characters played either by Lisa or the therapist (e.g., self-sufficient and fearless teen-ager, fearful and lost child, needy child, capable and resourceful child) were understood as referring to strands of Lisa's problems (experiences related to her unsecure early childhood and to the traumatic events surrounding the shooting), along with Lisa's internal responses to these problems.
From an assimilation model perspective, Lisa's change from APES 0 in the first phase to APES 2 in the second phase represents clear, albeit slow, therapeutic progress, consistent with her history of severe trauma. Of course, Lisa was far from recovered in APES terms after these 34 sessions. Theoretically, distress may actually increase across the interval from APES 0 to APES 2, as the dissociation and avoidance are overcome and the problematic content emerges in awareness (Basto et al., 2017;Stiles et al., 2004). And, at least superficially, Lisa's themes of murder, blood, and death in phase 4 sound more distressing than the fast motorcycle riding and dancing in phase 1.
Lisa had a good imagination and an ability use symbolic play to share her experiences. These were probably resilience factors and components of her healing. Her ability to act simultaneously as both director and actor in the cottage play was the basis of the dyadic collaboration. Not all children have these assets. Lisa's age and development were optimal for this sort of symbolic elaboration; she was old enough to express herself, but not yet too analytically oriented or self-critical to control her play too much.
The case offered, and we took, opportunities for elaborating and extending the assimilation model to a variety of areas of play therapy. Among the elaborated aspects were studying assimilation in play therapy for the first time, the significance of repetitive play in the APES, the nature of the achievement in the transition to APES 1, the semiotic power of play, and the child-therapist collaboration in symbolic play. Such elaborations (sometimes called abductions) are justified by their ability to explain the new observations while remaining consistent with the rest of the model and previous observations (Rennie, 2012;Stiles, 2007Stiles, , 2017. They are, of course, tentative and subject to research on further cases.

Lisa's progress in assimilating her traumatic experience
Understood within assimilation theory, Lisa's APES progress reflected changes in how well her dominant community of voices was able to hear and acknowledge the voices of her trauma with attendant fear, horror, and loss (Stiles, 2011). We assigned levels from APES 0 to APES 2, to characterize the changing relation of Lisa's central problematic experience (the shooting) to her usual self (dominant community) across the phases of the cottage play. Although clear progress occured, the problematic experience was far from fully assimilated.

Phase 1
The highly repetitive first phase of the cottage play took half of Lisa's first year of treatment (17 out of 34 sessions). In our interpretation, this expressed her need for safety in response to the overwhelming internal threat of the shooting in the context of her early insecure circumstances. We inferred that the repetitive play expressed warded-off content that Lisa's usual self did not recognize (APES 0). Getting too close was cause for stopping the action (e.g., changing the violator to be the foster father in dollhouse).
In our interpretation, the warded-off voices of the shooting were addressed in the session, and they responded. At this stage of assimilation, the experiences were too overwhelming, and Lisa could not have expressed them if she (her dominant community of voices) understood that they concerned the shooting. Thus, this meaning was not understood or acknowledged because it was disguised in the play. But conversely, because the voices of the trauma were not understood or acknowledged (technically, because there was no meaning bridge), they kept saying the same thing. Such repetitive play is common in cases of severe trauma Sossin & Cohen, 2011;Terr, 1983). Lisa's repetitive play was less grim than is often the case, expressing some energy and hope as well as fear (e.g., wild motorcycle rides, dancing), which seemed to reflect her resilience.
As an example of how the material was disguised, the painful content that we called the voice of fear and loss was expressed through characters assigned to the therapist. This joint performance apparently afforded Lisa a safe enough distance from the emotionally charged experiences which significance her usual self did not appreciate. It also allowed her to observe the therapist, who enacted, expressed, and regulated fear. Thus, the externalizing and symbolic disguise were means of emotion regulation. On the other hand, therapists (and investigators) can often recognize symbolic expressions that are not consciously available to their clients, though it may take a good deal of observation and background information to appreciate what the warded-off voice is saying. To stop the repetition, a problematic voice needs a meaning bridge to the dominant community, that is, it needs to be recognized. Conversely, to allow recognizable expressions, the dominant community needs to feel safe from the overwhelming terror. We suggest that by repeating the cottage play, Lisa slowly but incrementally built confidence, ratcheting down the fear (Wilson & Ryan, 2006).

Phase 2
During this phase (sessions 18-25), Lisa's character experienced fear as she worked hard to escape and avoid the monsters and bad memories. Lisa effectively gained some control over the problematic contents them by volitionally shutting down the play (APES 1, avoidance). We interpreted this as significant progress, an indication that her dominant self could acknowledge the existence of the threat and deal with it by actively avoiding it without being overwhelmed. This progress clearly shows the function of APES 1 in the assimilation sequence. If the traumatic material is so terrifying that it cannot be consciously acknowledged (APES 0, dissociated), one cannot move forward. Gaining the capacity to intentionally avoid the traumatic content represents an increment in safety that is essential for further progress. Lisa was repeatedly threatened by fearsome circumstances and successfully avoided or escaped them. The form of this progress must be understood in relation to defense mechanisms typical to her developmental stage. The magical solutions she used in the play during the second phase can be seen as more age-appropriate than the activities during the first phase (living alone, drinking beer, doing make-up).

Phase 3
In contrast to earlier phases, the cottage play during the brief third phase (sessions 26-28) seemed relatively peaceful and benign, with Lisa and the therapist playing sisters with their own homes. Perhaps she was addressing a different problematic experience. We also speculate that what seemed like a break from the central problem may have also served to consolidate the achievement of APES 1, that is, to confirm that her ability to avoid the flood of terror was smooth and secure. Only when that security was achieved could the traumatic content be cautiously admitted to awareness and faced.

Phase 4
During Phase 4 (sessions 29-34), Lisa actively faced and conquered murderers, soldiers, and death in the play. In the symbolism of the play, the pervasive imaginary violence and death could be understood as a way of facing the real violence and death of the shooting. We rated this as APES 2 (emergence of the problematic content). That is, the play events and the emotions experienced were recognizable symbolic expressions of Lisa's central concerns. Importantly, when asked, she acknowledged the link between the play and the events of her life, demonstrating that she could sustain attention to the trauma. During the fourth phase, the cottage play also continued to include many safety features for magically avoiding the dangers. These seemed to reflect a recycling the achievement of successful volitional avoidance of earlier phases. Lisa's play could be seen as reenactment with soothing . It was characterized by repeated reenactments and comprised expression of diverse feelings and a sense of control, which is considered as a healing component of play (Campbell & Knoetze, 2010;Cockle & Allan, 1996).

Spirals, setbacks, and strands
Lisa's assimilation did not occur linearly but in a series of small advances and setbacks. Although four separable phases could be identified in Lisa's play, they were also interleaved or nested. The emerging new play elements often appeared briefly and receded before they became the main content of the play. For instance, the parent's death appeared briefly a few times before Lisa actually actively focused the play, first on divorce and then on death, and finally in Phase 4, verbally shared some explicit thoughts on the actual violent incident. It seemed as if processing of the experiences and change occurred in nested phases, or in a spiral cycle. Cockle and Allan (1996) characterized this as a circumambulation, in which the child symbolically approaches problem in a spiral pattern by coming closer and closer to resolution. In each repetition of the cycle (playing through of a meaningful play scenario), a piece of new information is provided to the conscious mind, which may change the nature, tone, and affective expression of the play. In Lisa's case, conscious, uninterrupted dialogue concerning the traumatic experiences could take place at the end of the fourth phase (a marker of APES 2). The memories gradually became more accessible to her as the treatment proceeded and she symbolically approached the painful content.
Assimilation theory suggests that this spiral pattern of setbacks and advances usually reflects shifting attention among different strands of a problem (Caro Gabalda & Stiles, 2018). Strands of Lisa's central problematic experience include "being fearful," "being homeless," "not having a parent to take care of me," "I have seen something threatening," "I have bad memories," "I have seen death", "I have been in danger," and "I was helpless," as well as the voice of self-sufficient and fearless teenager, the voices of fear and being lost, and the voice of a fearful child. We presumed that these originated partly in her experience of the traumatic events themselves (insecure attachment, witnessing the shooting, thoughts and feelings during her three days with the bodies) and partly in elaborations and responses that subsequently grew up surrounding traces of those experience. Shifting to a lower-APES strand would appear as recycling. In theory, if different strands of a problem are interlinked, they tend to advance together. However, during successful treatment, some strands may advance faster than others, and some may be left behind. Setbacks occur when the client's attention shifts to a lower-APES strand, typically because the therapist directs attention to that strand, or because the therapist tries to guide the client toward highly distressing content before the client is ready, and the client retreats to a different strand at a lower APES level (Caro Gabalda & Stiles, 2013). For example, a few of Lisa's setbacks seemed initiated by the therapist referring to the shooting, thus exceeding her TZPD (cf. Stiles et al., 2016).

The semiotics of play
Judging APES stages in play therapy raises issues about how we understand and interpret symbolic processing in play. At 6 years-of-age, Lisa might have been ready or able to verbally articulate her thoughts and feelings about the trauma or insight into her own processes. Play is semiotic too, however, and the point of play therapy is to offer children a more congenial medium for expression. In therapeutic play, Lisa could express and share a version of her emotional experiences and her search for safety. A further conundrum is how play contributes to assimilation: To what extent can reenactment of traumatic experiences in play have healing potential in the absence of conscious reflection and verbalization? Theoretically, the meaning bridges that underlie therapeutic assimilation need not be verbal. For example, actions, sets, and stage directions have semiotic properties and can be used to communicate interpersonally and intrapersonally. Play can address emotionally powerful content in its own terms. For a child, symbolic play offers a nonverbal, but nevertheless semiotic, path to understanding, along with strategies to build meaning bridges and to handle difficult real-life experiences. For example, repetitive play sequences controlled by the child may allow predictability and development of familiarity with fears and integration of narratives. They offer an experience of control or mastery over feelings and practice in behavior regulation and coping (Wilson & Ryan, 2006). Lisa's symbolic play probably provided means for corrective emotional experience and for expressing warded-off material that were not available in reality, giving her confidence to move to the next APES stage.
For children, play experience can be more effective in alleviating traumatic memories than a purely verbal approach (Nader & Pynoos, 1991). Play may be less precise than words for communicating with others; therapists may not understand exactly what experiences are being expressed. We speculate that children may be better able to understand the meanings of their own play. In particular, they may understand what problematic voices are saying in play to a degree that varies with the APES levels of the content.
A different but related question is, how can a child work productively on a problem that, on an explicit level, he or she is not aware of having, as is often the situation in child psychotherapy, especially when a trauma stems from early experiences? To the extent that the lack of awareness is due to the emotional power of the trauma, the assimilation model suggests that the play may be disguised so that the child's usual self cannot recognize expressions by a warded-off voice. Lisa's progress from Phase 1 to later phases shows how this work can be productive, although theoretically, progress in assimilation requires awareness (Stiles, 2011). For Lisa, assimilation of her trauma would be marked by an awareness of how the play referred to her traumatic history. Although she did not verbalize her problematic experiences in any detail, the extent to which she verbally acknowledged a link with her play helps index her level of assimilation. For example, the contrast between Lisa's shutting down the play when the therapist mentioned death in session 20 (early in Phase 2) with her acknowledgment of the connection in sessions 33 and 34 (Phase 4) is diagnostic of APES 1 (avoidance) versus APES 2 (emergence), respectively.

Therapeutic collaboration and the dyadic subject
As interpersonal meaning bridges develop between therapist and client, they can share experiences and engage in joint activity. In effect, we suggest, the author of the cottage play was a dyadic subject. During the first phase, Lisa was casting and directing the play, whereas emotions were expressed behaviorally and verbally by the therapist. She recognized psychological meanings when Lisa was not aware of them. Throughout the treatment, the roles representing the problematic experience were shifted, divided, and combined. Often, the therapist was assigned roles that seemed to refer to Lisa's most traumatized painful parts, while Lisa played roles that referred to her mother, for example, being a wild teenager in Phase 1 and going to the mountains where the trolls killed people in Phase 4.
This collaborative processing could be perceived as a dyadic assimilation process (HaCohen et al., 2017). The therapist acted as a container for Lisa's painful emotions and provided a sort of auxiliary awareness, as they jointly searched for understanding and expressions. This collaboration seemed to allow Lisa to test her own capabilities and eventually to find trust in them to act alone. In previous play therapy research, the participatory role of the therapist has often been neglected, although the therapist's understanding and supportive, encouraging and empathetic response have been credited with facilitating therapeutic change (Benedict, 2006;Ryan & Edge, 2011). Our findings show a need for further focusing on the therapist and child-therapist collaboration in the emergence of play themes and expression of feelings.
Formulating the collaboration as a dyadic subject underlines the importance of the therapist's involvement and the co-creation of the play, which is in line with earlier notions on the importance of therapist's sensitivity in play therapy (e.g., Crenshaw & Kenney-Noziska, 2014;Ryan & Edge, 2011). Lisa's progress toward being an active subject through building an observer-position emulates the developmental process of evolving emotion regulation: The caretaker first provides wording and soothing and models regulation. Gradually, these regulative functions are internalized by the child and he/she becomes increasingly competent to regulate emotions alone (e.g., Kochanska, 1994). The therapist, working on TZPD (Leiman & Stiles, 2001), modeled ways to express emotions and ways to regulate them when Lisa could not express or regulate them on her own. The product of this work in the first phase of treatment, we surmise, was Lisa's ability to control whether and when she faced the unwanted thoughts, which could be observed in the subsequent phases.

Limitations, generality, clinical implications, and further research
The post-session process notes we used as our primary data were not written for research purposes. We had no video or audio recordings or verbatim transcripts, so we could not do micro-analysis. Aside from the APES, there was no formal assessment of change, although the foster mother reported improvements. We cannot disentangle the interacting effects of therapy, development, and the security offered by the foster home, although probably all worked together. These limitations dictate caution in considering our interpretations.
Case observations, such as our results, are not meant to be generalized as isolated statements. Generality in theory-building research, as in most natural sciences, is a property of the theory, not of a study's findings (Stiles, 2017). A study may justify a small increment of confidence in the theory, and it may stimulate elaborations or extensions, as our does. As indicated earlier, the assimilation model is quite general, meant to describe the process of change in any psychotherapy. Likewise, a single theorybuilding case study by itself seldom justifies any practical recommendation. However, the theory it addresses may have many implications for practice. Among other things, the assimilation model, following elaborations suggested by this study, implies that the APES stages can be used to describe therapeutic progress in play therapy (and hence to suggest the next therapeutic subgoal), that therapeutic work on problematic experiences and memories is possible in symbolic play without explicit verbalization, that repetitive play may reflect the absence of an internal meaning bridge (expressions being repeated because they are not understood), and that gaining the ability to actively avoid a previously dissociated experience (APES 0 to APES 1) may be a prerequisite for facing a feared experience. That is, our study adds confidence to a theory that has these clinical implications, even though it does not, by itself, justify them.
In future research, the assimilation stage descriptions could be elaborated to better apply to child development and the metaphoric nature of symbolic play. Finally, our study underscores the need for longterm study of PTP, especially in cases of traumatic experiences that occurred in early developmental periods when no verbal expressions were available or in the case of Type III or complex trauma.

Disclosure statement
No potential conflict of interest was reported by the author(s).