Reclaiming the Inner Child in Cognitive-Behavioral Therapy: The Complementary Model of the Personality

Many psychotherapeutic approaches invoke the concept of dividing the personality into parts or subsystems. A. T. Beck (1), for instance, suggested the notion of modes, which includes the activation of cognitive, affective, motivational, and physiological systems that together form specific suborganizations within the personality. Compassion-focused therapy (2), internal family systems therapy (3), voice dialogue (4), Gestalt therapy (5), ego-state therapy (6), and schema-focused therapy (7), to name a few, also divide the personality into separate structures or functional entities, also known as subselves or subpersonalities. The aim of therapy, it is assumed, is to identify these parts and integrate them into the conscious part of the personality through the therapeutic process. There are different means to this goal and, apart from classical behaviorism, most schools subscribe to the principle of subconscious forces that influence thoughts and behavior.

A. T. Beck’s (8) discovery of automatic thoughts arguably belongs to this school of thought; subconscious forces (automatic thought processes) exert a strong influence on how people feel and act. Contrary to psychoanalysis, A. T. Beck found that these automated processes can easily be brought into conscious awareness, thus becoming the object of the therapeutic process. He explained,

Some therapists have advocated the notion of an inner child as a primary subconscious force. Quite simply, inner child refers to the child the patient once was and with whom the patient might, to some extent, have lost touch with on the way to adulthood. The objective is to help patients reconnect with this inner child to free themselves from maladaptive emotional and behavioral patterns. Bradshaw’s (9) recovery therapy is one of the better known examples of this approach. In a radio interview, he explained the metaphor of the inner child:

The metaphor of the inner child encompasses only one subself (with many expressions) and stands in contrast to approaches that hold that many subselves coexist. A. T. Beck’s (1) notion of modes and McGinn and Young’s (11) schema mode operate with various subselves, some of which relate directly to a child mode. Contrary to the inner child metaphor (as I use it), they share the idea that many modes, and mixes of these modes, exist and that patients can rapidly switch between them:

There is a straightforward association between particular emotional states and the descriptor. Each mode corresponds to an emotional state, such as shameful or vulnerable child, angry child, or abandoned and lonely child. This thinking, to identify and distinguish between a set number of modes or subpersonalities, is shared in McGinn and Young’s (11), Beck’s (1), and even Bradshaw’s (9) models, which, in Beck’s case, includes links between the psychopathology and the descriptor, such as “anxiety, depressive, hopeless-suicidal, panic, obsessive-compulsive, as well as specific phobic modes, corresponding to each of the disorders or clinical problems” (p. 13).

An alternative to this division of modes into categories according to their clinical features would be simply not to, as in the unitary metaphor of the inner child. With the preceding example in mind, it would be possible to view the various modes as expressions of one and the same. Imagine a small child put into a stressful and utterly unreasonable situation. When children are not coping, it is not uncommon to observe them switching between states, at times even quite dramatically and violently. An example would be a child at the hospital just before a dreaded procedure. The child might switch between crying and sobbing, then becoming aggressive and accusative, then begging and appealing, and then switching back to kicking and screaming before finally giving in and becoming overly obedient or withdrawn. One would not describe the child’s behavior as switching between an angry mode and an obedient mode. Instead, the behavior would be considered perfectly understandable. The behavior of a patient with borderline personality disorder appears much like that of a (distressed) child. In this view, the patient is simply in a state called child mode—not denoted by a particular descriptor. Why call it borderline mode? Calling a child’s behavior borderline in such a situation would make no sense. Even the punitive parent mode can be seen as a child’s attempt at coping when all other strategies have failed. It is an intelligent survival strategy. By internalizing a feared or critical parent, the child hopes to prevent unpleasant trigger situations from arising in the future by watching the situation through the eyes of the critical parent and criticizing him- or herself proactively when inadvertently engaging in some risky behavior that might attract the parent’s disapproval.

Child Part and Automatic Thoughts

Cognitive therapy is based on making patients focus on a particular type of inner conversation known as automatic thoughts. A. T. Beck (8) discovered that a concurrent stream of thought, known as free association, was present when a patient engaged in the usual reporting of thought content. This parallel track would often reflect on aspects of the situation, with comments such as “The therapist thinks I am stupid” or “I’m not making any sense.” The patient would not spontaneously report these types of thoughts or even be aware of them. Contrary to the wisdom of the day, in cognitive therapy it was not always considered necessary or even recommended (8) to make the patient recall childhood memories. Focusing on automatic thought processes and the underlying rules and beliefs they convey was sufficient. However, these rules and belief systems are firmly based in life experiences, particularly those stemming from childhood:

The cognitive model offers an explanation for the origins of automatic thought processes: “Beginning in childhood, people develop certain beliefs about themselves, other people, and their worlds. Their most central or core beliefs are understandings that are so fundamental that they do not articulate them, even to themselves” (12, p. 15).

Automatic thoughts are the gateway to the identification of core belief systems emanating from childhood experiences:

The focus on automatic thought processes as the main source of information guiding the therapeutic process is a fundamental principle of cognitive therapy. The exact same principle holds true when introducing the metaphor of the inner child into cognitive therapy; the inner child is identified via the observation and reporting of automatic thought processes as in cognitive therapy. Now, instead of calling the inner conversation automatic thoughts, it is simply described as a child part, an inner child. The cognitive processes usually identified as automatic thoughts are really just the part of the personality identified as the child part revealing itself. Habitually, people are not aware of this inner voice, as already pointed out, but through training they can learn to pay attention to it. The primary source of information about the inner child does not emanate from the disclosure and analysis of painful historical events from the patient’s past but transpires from the present-moment automatic thought processes. The suggestion is that the child part is identified through the type of automatic thought processes discovered by A. T. Beck, including the schemata and beliefs that underpin them.

Twin Modes

The principle of complementarity in quantum physics solved the fundamental wave-particle problem: that a particle can exist in two fundamentally different states at the same time and, astoundingly, that the state of the particle is contingent on the observer. Danish physicist Niels Bohr proposed that the counterintuitive and illogical appearance of the problem related to the natural limits of human comprehension, especially with regard to duality. The complementary model of the personality suggests, along similar lines, a counterintuitive solution to the understanding of the personality—that it is fundamentally dual in nature, made up of two complementary systems, two ways of processing, not immediately available to people, that can be made the subject of introspection, as with A. T. Beck’s automatic thought processes.

Many patients spend much time, sometimes even the majority of their waking hours, in a negative state of mind, that is, a negative mode. When analyzing their daily experiences, they appear repetitively triggered, even by small events. Fortunately, they also report periods in which they experience a calmer state of mind, feeling relatively okay about themselves and the world around them. I find that patients readily recognize this description. They find a division into two complementary, or twinned, mental states very meaningful, in the same way that they relatively easily absorb the concept of trigger events and automatic thoughts. They recognize that they sometimes (or most of the time) feel they are in a negative mood interrupted by shorter or longer periods of relative peacefulness or calm, that is, the absence of the state of being triggered. Patients who spend the most time in a negative mode generally have more substantial difficulties managing and coping with life.

Let us, for a moment, imagine that there are only two modes of processing: adult mode (the preferred position) and child mode (the unhealthy position). Child mode would largely correspond to the mental state that arises after trigger events in the classical cognitive model, a state of heightened arousal (this is true for depressive states as well). It heralds the activation of dysfunctional core beliefs.

The thoughts and feelings arising from child mode are unpleasant, overwhelming, and confusing. They equal the activation of ineffective cognitive, emotional, motivational, and behavioral schemas, not corresponding well to the demands of the situation. These trigger moments stand in stark contrast to people’s usual sense of self when they are dealing with nontrigger situations, and when their problem-solving skills (and corresponding beliefs) are comparatively much more effective, appropriate, and resourceful.

When the mental state corresponding to adult mode is active, people feel pretty much in control and tend to have a good understanding of the situation and how they feel about it and about themselves. Even if things do not work out the way they want them to, they do not completely lose their self-control. This preferred state of mind is lost once triggered. When child mode processing is activated, people’s information processing style changes dramatically, becoming biased, ineffective, and dysfunctional.

Adult and child mode processing mark the difference between the mental state that arises during trigger events and the mental state that is there before and after. Some would rightly argue that they simply mark the difference between the activation of positive and negative belief systems. So far, I have not described any process that could not be contained within the classical cognitive model. Notice that when one mode is activated, the other is deactivated, and vice versa. My aim is to teach patients to distinguish between the two states of mind. This is essential. I point out that as long as child mode is activated, they will not be able to solve their problems. First, they must learn how to deactivate child mode. I help patients to realize that during trigger events, child mode is desperately trying to fix the situation, forever doomed to failure because it is only a child in a grown-up world.

Experiential Techniques

Schema-focused therapy has taken inspiration from the constructionist movement in cognitive sciences. It signals a rejection of the correspondence theory of truth. People’s beliefs cannot be judged solely on having a direct correspondence to an objective reality. “As a consequence, constructivist therapists tend to target changes in broader systems of personal constructs, rather than disputing circumscribed thought units” (11, p. 185). Originally set out as an adjunct to cognitive therapy to better deal with characterological disorders, schema-focused therapy includes a stronger focus on emotional meaning, exposed by various cognitive and experiential techniques.

When identifying and changing core beliefs or schemas, it is common practice to include early trigger situations arising from childhood experiences. In schema-focused therapy, this process comes under the heading of imagery techniques, which “are among the most dramatic approaches to changing schemas” (11, p. 197). In the section “Restructuring Early Memories,” J. S. Beck (12, pp. 184–192) described the same technique working with a patient called Sally:

• T: Sally, you look pretty down today.

• P: Yes. (Crying.) . . . I got my paper back. I got a C−. I can’t do anything right. . . .

• T: When is the first time you remember feeling this way, as a kid?

• P: (Pause.) When I was 6 or 7. I remember I brought home my report card. . . . (p. 184)

Notice that the therapist brings the patient to remember an historical event or image from childhood that appears to be associated with the activated dysfunctional belief the patient brings up in therapy. It carries the same feeling or meaning. It is this image the therapist aims to restructure by adding new elements to the memory, in this case in the form of a role-play:

• T: Okay, how about if we do a role-play? I’ll play you at age 7; you play your mom. Try to see things from her point of view as much as you can. I’ll start. . . . Mom, here’s my report card.

• P: Sally, I’m ashamed of you. Look at these grades. What am I going to do with you?

• T: Mom, I’m only 7. My grades aren’t great like Robert’s, but they’re okay. (p. 185)

The work is being done using a past event, looking at the situation in a new way, enabling the patient to realize an alternative view of the episode, for example, that her mother was concerned about her poor grades because she herself did not get an education.

Now, I am proposing to use the technique rather differently—the historical events are only considered useful in that they provide information about the particular child part the therapist is working with, and the actual restructuring takes place (by placing the historical child) in the day-to-day trigger events the patient brings to the session. In the case of Sally, it would look something like this:

• P: Yes. (Crying.) . . . I got my paper back. I got a C−. I can’t do anything right.

• T: Let’s assume it was child part that became active when you got that result (it surely sounds like it was because you got so emotionally overwhelmed by the episode). Let us take that little girl who reacted like that out of you, and place her on that chair over there (pointing to the chair), okay?

• P: Yes.

• T: Child part is feeling exactly how you felt in that situation. Like she cannot do anything right. She blames herself. She is sitting right over there (pointing). How old do you think she is?

• P: Maybe . . . 6 or 7?

• T: Yeah, that sounds right, a little girl. Now, we’re the adults sitting over here and looking at her. We know exactly how she feels about herself. How are we going to help her? What does she need?

• P: Tell her that she is OK?

• T: Yes . . . good (pause) . . . How would she respond if you told her that, right now?

• P: She feels a little bit better . . . relieved.

Notice that it is the present-moment child part who is addressed directly in the therapy room. The therapist is not connecting with or restructuring a historical image or child. The image in which the child part is being placed is the adult scenario, in the therapy room. The patient has full access to the thoughts and feelings of the externalized child part, constantly responding to its changing moods with the therapist’s continuous support.

Just as the cognitive model makes a distinction between addressing dysfunctional schemas and the learning of new social competencies, the twin mode model similarly operates with two ways of caring for child part. Nurturing a child (part) takes more than the ability to make it shut up by the force of logical thinking and the weighing up of evidence for or against a certain belief. A child needs both nurturing care and boundaries. In cognitive therapy, the child part usually meets boundaries in the form of the challenging of dysfunctional beliefs. J. S. Beck (13) wrote that “therapists must be creative to devise interventions that demonstrate the invalidity of patients’ core beliefs” (p. 177). This approach has the effect of strengthening the patient’s ability to activate the adult part. J. S. Beck (13) described a role-play with a patient, Kim, who was asked to play her emotional side that held a core belief strongly while her therapist played the intellectual side. As the reader might notice, these two sides of the personality correspond well with the child-adult mode division. It is not that the emotional side is purely emotional; it has its own cognitions that are then thwarted by the intellectual side:

It is interesting that J. S. Beck (13) argued that the patient has an emotional side that harbors emotional reasons, as well as an intellectual side, concepts that are not at all well defined. She portrays a technique in which the patient is encouraged to switch from one to the other vis-à-vis switching from child mode processing to adult mode processing. If this intervention is ineffective, J. S. Beck (13) suggested that “a final intervention for patients who still ‘feel’ their core beliefs to be true even though they intellectually understand them to be invalid involves restructuring the meaning of earlier memories” (p. 179).

Externalization

The aim when working with the twin mode model is to establish and internalize successful adult-child interactions, not by the therapist acting as the adult role model but by activating the patient’s own adult part. The externalization technique is instrumental in this process. Externalization techniques are already well-known; think of techniques such as the externalization of the problem as suggested by White and Epston (14) or the empty chair technique used by gestalt therapists. In twin mode model usage, its most basic form consists of viewing or analyzing a situation as though the child part was actually out there acting all by itself without any adult part present. The therapist is identifying everyday situations in which the child part is active and the adult part has disappeared. The therapist asks the patient to view him- or herself from the outside, being in that situation, as a little child on his or her own, and the therapist tries to imagine how the patient would feel in that situation. What might go through the patient’s mind? How would the patient be feeling? Do those thoughts and feelings correspond to how the patient actually felt in the situation?

Adult part processing (the state of being in adult mode) is represented by the therapist, who recognizes and mirrors the needs of the child part sensitively and empathetically. Just as real children need nurturing and boundaries, so too does the child part. The therapist encourages the patient to redevelop parenting skills by empathetically relating to the child and finding ways to help the child, unveiling the child’s needs. The purpose is to make the child part feel cared for so that it can let go and calm down. Once that has happened, the patient automatically switches back into adult mode processing.

Firman and Russell (15) have described a process whereby healthy adult-child connecting occurs during personality development in childhood. They present an idea of the transpersonal self, founded in psychosynthesis, in which the “I” is only a reflection. During childhood, an “I-self” connection gradually emerges, in part, by internalizing healthy or sound interactions with mirroring others, the other acting as an external unifying center. This mirroring of the I-self connection is then internalized into a unifying center, supporting the further development of a mature I-self relationship. Transposed onto the twin mode model, this process equals the successful mirroring of child mode by adult mode, thereby acting as an external unifying center.

To give a personal example of this process, I often think of how my wife would always go through the day’s events with our children at bedtime. She would ask how their day went and talk about the coming day. She might ask whether they knew the plans for the next day. They usually didn’t have a clue. Being children, it just didn’t enter into their minds to think about the next day; they were simply in the moment just before bedtime. In fact, planning ahead and thinking about important events and interactions with friends, family, or colleagues is very much an adult skill. Children do not learn to do it automatically.

As our children develop into adulthood, they will have internalized these innumerable child-adult (parent) interactions with their mother at bedtime and will themselves automatically prepare for or process events, past and future, without being prompted. They will also teach their own children to do the same. Conversely, if our children were left on their own at bedtime and were not accustomed to going through past and future events like this, would the result not likely have been a reduced ability to parent that part of themselves? They would perhaps end up with a child part (at bedtime) that would be a little bit unsettled and anxious (without knowing why), simply because it had not been prepared for events. It is that child part, without a corresponding adult part, that therapists meet in the therapy room.

I illustrate the implementation of the twin mode model into a cognitive therapy protocol, the twin mode protocol, in more detail in the following case example. When introducing the model to patients, I begin by explaining the cognitive-behavioral model and how it applies to their specific problem. We identify trigger events, automatic negative thinking, and underlying unhelpful beliefs—the usual stuff.

Case Illustration

A female educator, let me call her Anne, came to me with a work-related issue, complaining that she felt she did not perform at work, notably in staff meetings. She felt completely unable to put her views across and described herself as overly agreeable and lacking in confidence. When she forced herself to utter her opinion at meetings, she would often feel very anxious to the point of having a full-blown panic attack, quite unlike the way she felt about herself when she was with the children in the classroom, where she would feel confident and in control. I started out by introducing the child part to her:

• T: Try to picture yourself, your feelings, your behavior, your thoughts when you’re at the staff meeting. . . . Get an image . . . if that was your child part, experiencing what you’re going through, how old do you think it would be?

• P: Maybe, like, seven or eight years old?

• T: Okay, yes, that seems reasonable, doesn’t it? About that age. And that way it would actually make a lot of sense, if she was only seven or eight. She would be anxious among all the grown-ups, believing she ought to be able speak out for herself but finding it quite impossible. Imagine, if we expected that of a real seven-year-old!

• P: Yes, poor girl. I feel sorry for her. She thinks she is not good enough.

• T: Now, let me take that child part and place it over here (pointing to a chair in the corner). Have a look at that little girl sitting over there in the corner. In fact, let’s both sit here right now in this room, as grown-ups, and look at that poor child. How can we help her? She is clearly hurting and desperate. She feels that she’s got this problem she cannot solve—that she doesn’t perform at staff meetings—that there must be something wrong with her. What would you do?

• P: (hesitating) I would give her a hug! I’ll tell her it’s not her fault. . . that she is okay just the way she is!

• T: Okay, do that right now! Say to her what you just said to me (the patient doesn’t have to speak out loud to child part). What happens? How does she feel?

• P: (Hesitating) She feels better. She is relieved that somebody is seeing her, noticing how she feels.

• T: Spot on, yes! But she might also have her doubts; after all, she is not used to getting any help from anybody. . . . You’ll have to prove to her your new conviction: that you will take care of her, be there for her, and not leave her on her own. You might start by apologizing to her for not being there in the past; for instance, say that it has been difficult for you but that you will work hard to become better at being there for her. You could even suggest that you will do your best to be therewith her at the next meeting, allowing her not to speak up, to let her hide behind you.

• P: She’d really like that, I think.

• T: Now, that is what I want you to do with that part of you we have placed over there and described as a child part. Connect with her. Communicate. Dialogue. You told me earlier that you find it difficult to think positively about yourself or be kind to yourself. Now, in this model, you immediately sympathized with that child, you actually wanted to be loving toward her, to comfort her. It would never cross your mind to be dismissive or mean to her, would it?

• P: No, absolutely not!

• T: This is the power of this model. It is still you we’re talking about, a part of you, but still simply you. And now you’re not trying to get rid of that part. Instead, you’re embracing that part.

What if I suggested to Anne that her feelings occur because she is actually in child mode—in effect, entering a mental age of, say, a 7-year-old, thrown into an adult professional situation and therefore desperately trying to participate in a grown-up discussion on equal terms? Would the thoughts and feelings she experiences (“I’m not good enough”) fit the situation? Yes, certainly, if she is in child mode, it all makes sense. The negative mental state she finds herself in when facing this type of situation is perfectly understandable, logical even. Being in child mode, she is completely unable to change the situation. She is dependent on someone else to intervene, a grown-up, helping her. A child is surely not supposed to go to meetings at work. The adult part should attend meetings, not her.

Anne will gradually improve her ability to empathize with that child, even during staff meetings, comforting her, and little by little the child will become less active during meetings; she will feel looked after, freeing up the space for Anne’s grown-up self, her adult part, to take charge of the situation and, while still parenting the child part, being able to act as a fellow colleague in the role of a professional teacher. Because Anne will gradually learn to recognize her child part during staff meetings, she will no longer be caught up in an endless cycle of self-loathing and self-criticism for not speaking up for herself. Suddenly, her negative core belief (i.e., “I am worthless”) will have no foundation.

Twin Mode Formulation

When patients arrive for psychotherapy, they are often going through a personal crisis. Their primary mode of processing is therefore child mode. Patients often feel overwhelmed and hopeless. This is the point at which they seek help and are most motivated to change. The initial part of the intervention (once history taking and formulation-diagnostic considerations have been made) is to teach the patient to distinguish between child mode and adult mode processing. This naturally involves a careful and timed introduction to the concept, taking care not to give patients the impression that they or their problems are childish. Sometimes the twin mode model has to be introduced at a later stage, when the patient-therapist relationship is stronger. If the problem can be dealt with effectively by means of existing cognitive therapy protocols, there might be no need to introduce it at all. Remember, this approach is complementary to cognitive therapy. It is not meant to replace existing cognitive therapy tools. For instance, Socratic questioning is integral, as is the ability to identify trigger events.

When making a twin mode formulation of the patient’s difficulties, I start by noting the differences between the two modes of processing, giving plenty of examples from the patient’s own story. I take care to check with the patient that the explanation is meaningful and empathetic. It is easy for the patient to switch to a judgmental (child) mode if the presentation feels too direct or confrontational. Remember, the patient might be quite unstable and switching in and out of child mode when presenting for therapy.

Therapists look at trigger situations and draw comparisons with the patient’s own history. Does the patient recognize that his or her way of processing as an adult looks like the way the patient reacted, felt, or thought as a child? What if the child the patient once was is still active, in the present moment, and is over and over again being thrown into a grown-up world that it finds extremely frightening and confusing? The therapist and patient try to identify child mode processing. Patients nearly always feel deeply met when the child-alone-in-an-adult-world is identified. There are often aha moments: All of a sudden, the patient’s strong negative feelings and biased thinking style make complete sense when viewing them as a child’s (logical) response to being a child alone in an adult world. The experience of child mode processing has been externalized.

An important element of working with the child part is that the therapist does not directly act as a parental figure, a process that in schema-focused therapy is called limited reparenting (11). Instead, patients are encouraged to activate their own adult part (by externalizing the child part) and form a present-moment relationship with that part of themselves, meeting its needs and noticing how it responds. The difference between the twin modes is emphasized again and again in every image, to teach the patient to distinguish between them, using externalization. It is often a great relief for patients to have an explanation for their odd behavior and tendency to uncontrollable emotion—it is simply the child part taking over in certain situations, causing havoc and disruption in its wake, quite a lot like a real child would be doing if placed in a similar situation. Some patients feel their child part is active nearly all the time; for others, it is clearly more limited to specific areas of their lives.

Amalgamation

Cognitive therapy prides itself on being relatively easy to understand and emphasizes that patients can easily process the basic model and quickly become their own therapists. Successful therapy is measured on patients’ use of the provided tools between sessions and after ending treatment. Initially, however, patients often struggle with the concepts and the pen-and-paper exercises. Some are willing to give it a shot and accept that they have to do their homework and complete the worksheets before they see the benefits from treatment. Others are more skeptical and want proof that the method will really help them before commencing treatment, or they might not feel sufficiently met or mirrored because they expected a more psychodynamic or narrative formulation. Cognitive therapy can certainly appear overly technical and simplistic at first glance. This fact, I suspect, keeps some therapists from competing schools from supplementing their knowledge with formulations originating in the cognitive model.

The integration of the twin mode model with the cognitive model simply acts as an interface for the core cognitive model, in a way that provides immediacy and meaning from the outset, a quality sometimes reserved for psychodynamic and systemic approaches. It is a bit like the Windows operating system was for early computing: It was the layer on top of the command prompt that allowed the user to execute the program straight away, intuitively and effectively, without the need for technical know-how. In the same way, in this model, the negative automatic thoughts that arise in the trigger situation are given a form, a figure, a voice in the shape of the child part.

Patients arguably do not need to know or understand the inner workings of cognitive therapy to acquire the (self-parenting) skills needed in this model. All they have to accept is that they have a child part and that they need to learn how to distinguish it from their adult self and care for it effectively. When the child part is triggered or activated, all they need to ask themselves is the question, “How can I help the child; what does it need?”

The twin mode model and cognitive therapy are perfectly compatible. It is easy to switch between the two ways of working. The aim is to teach patients to deal effectively with the child mode by parenting it appropriately. This can occur through persuasion—that we (the adults) know best and should be in charge, quashing the irrational beliefs of the child mode (as in cognitive therapy), soothing and making ourselves available as caregivers providing awareness, attention, and mindfulness (as in third-wave therapies)—or simply teaching patients how to form healthy child-adult attachments (as in psychodynamic approaches). There is no internal conflict between cognitive therapy and the exploration of the perspectives arising from the inner child metaphor and the complementary model of the personality.