Affect Regulation, Mentalization, and the Development of the Self (Book Review)

Author:  Fonagy, Peter , György Gergely, Elliot L. Jurist, and Mary Targe
Publisher:  New York: Other Press, 2004
Reviewed By:  John S. Auerbach, Winter 2005, pp. 35-38

Rarely, but every so often, a book comes along in psychoanalysis that is widely hailed, from the moment of its publication, as transforming the field. Affect Regulation, Mentalization, and the Development of the Self, by Peter Fonagy, György Gergely, Elliot L. Jurist, and Mary Target, is such a volume, and justly so. All of a sudden at psychoanalytic conferences, everyone is talking about mentalization and reflective functioning, and for this we can thank Fonagy and his colleagues. To be sure, these were already widespread notions in the psychoanalytic literature, thanks again to the prodigious scholarly and research output of Fonagy et al., but now suddenly these concepts are everywhere, and it is almost certain that the publication of their perhaps overlong but nevertheless clinically, theoretically, and empirically rich volume is the main reason that so many of us are talking about these ideas.

In Affect Regulation, Fonagy et al. set for themselves the ambitious task of linking attachment theory, Bowlby’s Darwinian reformulation of psychoanalysis, to complex ideas about intersubjectivity and theory of mind that derive ultimately from Hegelian philosophy. Although a book as long and complex as this one defies easy summary, a brief account of Fonagy et al.’s central arguments is as follows: The authors posit that mentalization, which they define as “the capacity to envision mental states in self and others” (p. 23) and operationalize as reflective functioning, arises as a higher-order transformation of the attachment system in humans and in turn helps to organize human attachments.

Mentalization therefore emerges in the context of the infant-caregiver relationship through early affect mirroring and is essential to the development of intersubjectivity. That is, infants become independent subjects only if they are recognized as such—as beings with minds, wills, and feelings of their own—by their caregivers. Thus, a sensitive caregiver relates to her baby as a subject long before an infant has any conception of other minds and other subjectivities, let alone his or her own. To paraphrase Fonagy et al., an infant develops a mind because the caregivers have the baby’s mind in mind.

Fonagy et al. propose a social biofeedback model of affect mirroring as the mechanism through which infant affect regulation develops and attachment security (or lack thereof) is consolidated. On this theory, sensitive caregivers respond to their babies’ affective displays with contingent marked affective displays of their own, and this contingent marked mirroring of the infant’s emotions is what enables the baby to modulate his or her own affect states. Psychoanalytic developmental research (e.g., Jaffe, Beebe, Feldstein, Crown, & Jasnow, 2002; Stern, 1985) has long held that moderate degrees of caregiver-infant contingency or coordination in affect states are optimal for the infant’s eventual development of adequate affect regulation and attachment security and that a caregiver-infant system with too much or too little contingency—too little or too much mirroring—results in developmental psychopathology. Fonagy et al.’s unique contribution here is the concept of markedness, the exaggerated facial, vocal, and gestural displays that caregivers make when responding to babies. By marking their affective displays, sensitive caregivers differentiate as-if (or pretend) communications from realistic ones.

They also help babies to find their own emotional states in the parent’s face, much as Winnicott (1971/1982) describes, and the baby therefore experiences the parent’s displayed affect as his or her own, rather than as the caregiver’s. Fonagy et al. argue that this process is especially important in the modulation of negative emotions and in the infant’s formation of a constitutional self. If, however, the caregiver fails to mark his or her emotional displays, the infant will see the parental response as reflecting the parent’s actual emotional state and can be traumatized by the uncontained affect if the caregiver’s response is negative. Or, if the caregiver’s response is marked but incongruent with the baby’s affect, the baby will identify with the incorrectly mirrored affect that he or she sees in the parent’s face and will begin to develop a false or, in Fonagy et al.’s terms, an alien self.

Fonagy et al. note that, because every childhood has some parental failures, everyone has aspects of the self that are alien or false, but when parental failures in contingent marked mirroring are pervasive and traumatic, the alien self forms as a structure that is experienced as persecutory and is defended against through mechanisms like dissociation and projective identification. Thus, the authors further argue that the formation of an alien self in the preverbal period disrupts the child’s eventual ability to take a mentalizing stance in understanding the actions both of self and others. Fonagy et al. posit the existence of an interpersonal interpretative mechanism that grows out of the attachment system and that enables people to understand both themselves and others as intentional beings in the philosophical sense of that term—as beings with beliefs and desires, with the capacity for meaning. They further argue that young children, between ages 2 and 5, understand interpersonal meanings through two distinct representational systems: pretend, in which children decouple fantasies from realistic perceptions so that they can play with them, and psychic equivalence, in which fantasy and realistic perception are equated. Pretend functioning clearly descends from states in which parents respond to children with marked affect, and psychic equivalence descends states in which caregivers respond without markedness. In normal development, these two modes of functioning become integrated, around age 5, into a single system that enables a child to understand the difference between appearance and reality—to understand, therefore, that reality, or rather our understandings of it, is representational and intentional. Alternatively, children understand the separateness of their bodies sometime in the second year of life but do not understand the separateness of their minds until sometime in the fifth year (Mayes & Cohen, 1996), and it is only then that they can begin to integrate transitional fantasy with realistic cognition (Auerbach & Blatt, 2001, 2002).

Consistent with these propositions is empirical research cited by Fonagy et al. that securely attached children understand the appearance-reality distinction, belief-desire reasoning, and the nature of false beliefs earlier than do insecurely attached children. Thus, the status of the caregiver-infant relationship is crucial to the child’s cognitive development, and the authors also discuss research findings that not only the child’s but also the parent’s attachment security predicts the child’s mastery of these various theory-of-mind tasks. Insofar as parental attachment security is actually measured through coherence of discourse regarding the mental states of their parents (i.e., through coherent narratives on the Adult Attachment Interview [George, Kaplan, & Main, 1985]), Fonagy et al. are arguing, essentially, that the causal relationship between reflective functioning and attachment security is bidirectional. Conversely, children who are insecurely attached or have developed an alien self show disruptions of mentalization. They continue to function in psychic equivalence mode and therefore show a deficit in the ability, as Fonagy et al. put it, to play with reality. Such children often develop into adults with highly limited capacity to understand mental states and who are frequently diagnosed as borderline. They are often highly concrete in their understandings of human relationships, and their concreteness adds a certain desperateness to their attempts, via splitting, dissociation, and projective identification, to rid themselves of alien aspects of the self, for to such persons, alien aspects of the self feel real and cannot be regarded as fantasies to be played with.

In these last several pages, I have scarcely done justice to the complexity and subtlety of Fonagy et al.’s volume, and in my efforts to identify a single main line of argument in this book, I have unfortunately left out much of its richness. For example, I have not yet mentioned two of this book’s great strengths—that it takes an unapologetically Darwinian approach to its central concepts, affect regulation, mentalization, and the self and that, although of course arguing for the central role of parent-infant relationships in personality development, it takes seriously recent developments in behavioral genetics that indicate how much of personality is apparently inherited, prior to the influence of attachment. That is, Fonagy et al. are doing the hard work of reuniting psychoanalysis with the biological sciences. This is certainly a project with which Freud would have felt much sympathy, although he of course would have had a very different understanding of the link between psychoanalysis and biology than do Fonagy et al. Nevertheless, the attempt by Fonagy et al. to reconnect psychoanalysis to biology is crucial in an age which so many in the sciences think of psychoanalysis as unscientific and irrelevant and in which so many in psychoanalysis remain in denial regarding this problem.

Another strength of this volume is the radicalism of its intersubjectivist logic. Although I will note that Freud’s views on the nature of primary narcissism are far more complex and contradictory than is usually thought (see Auerbach, 1993, for a discussion of this issue), Fonagy et al. are to be commended for their explicit and thoroughgoing rejection of the view that the self arises from an original libidinal cathexis. In particular, they note, “For Freud in infancy and childhood others in the external world are extensions of the self. For us . . . it seems more accurate to see the self as originally an extension of experience of the other” (p. 266). The authors’ position here is entirely consistent with their rejection of a Cartesian view, in which the self is known through direct introspection, in favor of a Hegelian one, in which the self is constituted and known through relationships with others. In essence, Fonagy et al.’s views are part of the relational revolution that has swept psychoanalysis in the last few decades, although this is perhaps not surprising, given the authors’ roots in Bowlby and attachment on the one hand and Hegel and intersubjectivity on the other.

Having said so many complimentary things about this book, I will note here that it is difficult for me to take a proper critical perspective in reviewing this work because I, in my own writings (Auerbach, 1990, 1993, 1998) and together with Sidney Blatt (Auerbach & Blatt, 1996, 1997, 2001, 2002), have proposed ideas quite similar to those of Fonagy et al., indeed so much so that, in an online dialogue on Psy Broadcasting Corporation (PsyBC) in May 2004, it was difficult for us to come up with points of disagreement. Nevertheless, I do have some theoretical disputes with, or at least questions for, the authors, and I have some clinical concerns as well, and I will state them briefly in this review in an attempt to stimulate dialogue on some of the important issues that Fonagy et al. raise.

For me, a main problem in this book is its heavy reliance on the concepts of internalization and externalization, concepts that I suspect virtually everyone in psychoanalysis, regardless of theoretical orientation, accepts. What follows here will be controversial, therefore. In their otherwise evocative clinical examples, the authors write frequently of how aspects of the self, especially the alien self, are externalized and internalized. In my view, internalization and externalization are concepts that are intelligible only within the context of the Cartesian split between subject and object, and as such, they are not actually compatible with arguments otherwise propounded by Fonagy et al. Similar considerations also to apply to Fonagy et al.’s equation of fantasy and pretense with that which is “inner” and of perception and reality to that which is “outer.” Unfortunately, in the space available, I can only assert, not argue, this point. Nevertheless, I think that many passages in Affect Regulation might be clearer if, instead of internalize, the authors used terms like “appropriate,” “reappropriate,” “make one’s own,” or “make part of oneself” and if, instead of externalize, the authors used “attribute to others,” “assign to others,” or “disavow from oneself.” I also think that we can discuss the distinction between fantasy and perception without ever assuming that the former is internal to the mind and that the latter is external to it. The point here is that, even if readers (or Fonagy et al.) do not agree with my argument that internalization and externalization, inner and outer, are Cartesian residues, it is still the case that the authors’ clinical formulations would be clearer and more experiential if they recast their language, and they might be able to develop some new ways of thinking about the phenomenon of projective identification, a concept that is badly in need of rethinking because of its Cartesian heritage.

These considerations apply also to the authors’ use of the term alien self, which they repeatedly describe as being externalized and internalized, projected and introjected and to which they also attribute essentially agentive powers that are properly ascribed instead to a person or to the subject, rather than to a part of the personality. Here the problem is that the term self in contemporary psychoanalytic discourse, not just in the work of Fonagy et al., is used as a kind of homunculus or a subject behind the subject, much in the way that the term ego used to be, back in the heyday of ego psychology, but I think that the only activity that we can properly attribute to the self, or any representational construct for that matter, is that of organizing or guiding mental activity, not of initiating or causing it. Unfortunately, in our language, and in that of Fonagy et al., it is easy to slide from one meaning, the self’s organizing our understandings of the relational world, to the other, the self’s (rather than the person’s) acting in it. Consequently, this book contains many passages in which it sounds as if it is the alien self, rather than the person’s attempt to disavow the alien self, that causes the borderline personality’s disorganized, often controlling and manipulative behavior. As another way of viewing this linguistic problem, consider what would happen if Fonagy et al. substituted the term alien identity for alien self. Although these two concepts are roughly synonymous, we would have a hard time speaking of externalizing an alien identity or of an alien identity’s doing anything to cause or motivate behavior, rather than simply organizing it. To be fair, I would add in this context that the critique I making here of the terms internalization and externalization applies not only to Fonagy et al.’s work but to almost all psychoanalytic theories, but because the terms are so extensively used in this volume, this review is as good a place as any to call them into account.

Additionally, as I have stated, I have some questions about Fonagy et al.’s clinical approach. Although the terms of their developmental theory are clearly intersubjectivist, their technical stance appears to be largely Kleinian, with many interpretations of projective identification and of patients’ concerns with being crazy or mad. Much of the time in reading the authors’ clinical examples, I wondered what would happen if, for example, the clinician in question took a more, for lack of a better term, Kohutian stance, in which the interpretations focused on feelings of shame or of being damaged. More important, however, is that the interpretations provided seem almost always to focus on what the patient thinks is happening in his or her own mind and almost never on what the patient thinks might be happening in the mind of the analyst--that is, on what the patient assumes (or fears or wishes) the analyst is thinking about him or her—and yet this is an area of experience that most intersubjective theorists (e.g., Aron, 1996; Bromberg, 1998; Mitchell, 2000; Orange, Atwood, & Stolorow, 1999) find essential to the psychoanalytic dialogue. In short, if the self is formed by appearing in the mind of the other, initially of the parent, it should follow that it is modified in accordance with the patient’s understanding of how it appears in the mind of the therapist.

On the other hand, another interesting clinical question concerns the authors’ treatment of borderline patients without the extensive limit-setting that the two most well-known and best-researched treatment approaches to borderline psychopathology—Linehan’s (1993) dialectical behavior therapy and Clarkin, Yeomans, and Kernberg’s (1999) transference-focused psychotherapy—hold is essential to the process. The authors describe extensive acting out on the part of their borderline patients without indicating how they addressed such threats to the treatment. Ultimately, the wisdom of framing interpretations in terms of damage or shame instead of madness or of using extensive limit-setting, rather than a less structured approach, in treating borderline patients is an empirical question, and because Fonagy et al. have been champions of empirical research in psychoanalysis, I am uncomfortable in criticizing them here. Nevertheless, their technical approach to treatment requires the same empirical scrutiny that they have given their theories of psychological development.

In any case, having made these criticisms, I can now close this review with words of high praise for the authors of this exemplary volume. Their work is truly a model for how to integrate theoretical and clinical sophistication in psychoanalysis with empirical research, and through their research, these writers have helped to reformulate psychoanalytic theory in ways that will be essential if psychoanalysis is to remain vital to the psychology and psychiatry of the 21st century. In short, I can conclude this review as I began it: Affect Regulation, Mentalization, and the Self is already a celebrated volume within psychoanalysis and justly so. It needs to be read by everyone who is interested in seeing the rebirth of psychoanalysis as a scientific discipline.


Aron, L. (1996). A meetings of minds: Mutuality in psychoanalysis. Hillsdale, NJ: Analytic Press.
Auerbach, J. S. (1990). Narcissism: Reflections on others’ images of an elusive concept [Review essay]. Psychoanalytic Psychology, 7, 545-564.
Auerbach, J. S. (1993). The origins of narcissism and narcissistic personality disorder: A theoretical and empirical reformulation. In J. M. Masling & R. F. Bornstein (Eds.), Empirical studies of psychoanalytic theories: Vol. 4. Psychoanalytic perspectives on psychopathology (pp. 43-110). Washington, DC: American Psychological Association.
Auerbach, J. S. (1998). Dualism, self-reflexivity, and intersubjectivity: Commentary on paper by Sheldon Bach. Psychoanalytic Dialogues, 8, 675-683.
Auerbach, J. S., & Blatt, S. J. (1996). Self-representation in severe psychopathology: The role of reflexive self-awareness. Psychoanalytic Psychology, 13, 297-341.
Auerbach, J. S., & Blatt, S. J. (1997). Impairment of self-representation in schizophrenia: The roles of boundary articulation and self-reflexivity. Bulletin of the Menninger Clinic, 61, 297-316.
Auerbach, J. S., & Blatt, S. J. (2001). Self-reflexivity, intersubjectivity, and therapeutic change. Psychoanalytic Psychology, 18, 427-450.
Auerbach, J. S., & Blatt, S. J. (2002). The concept of mind: A developmental analysis. In R. Lasky (Ed.), Symbolization and desymbolization: Essays in honor of Norbert Freedman (pp. 75-117). New York: Other Press.
Bromberg, P. (1998). Standing in the spaces: Essays on clinical process, trauma, and dissociation. Hillsdale, NJ: Erlbaum.
Clarkin, J. F., Yeomans, F. E., & Kernberg, O. F. (1999). Psychotherapy for borderline personality. New York: Wiley.
George, C., Kaplan, N. & Main, M. (1985). The Berkeley Adult Attachment Interview. Unpublished manuscript, Department of Psychology, University of California, Berkeley.
Jaffe, J., Beebe, B., Feldstein, S., Crown, C. L., & Jasnow, M. D. (2002). Rhythms of dialogues in infancy. Monographs of the Society for Research in Child Development, 66(2, Serial No. 265). Boston: Blackwell.
Linehan, M. (1993). Cognitive-behavioral treatment of borderline personality disorder. New York: Guilford Press.
Mayes, L. C., & Cohen, D. J. (1996). Children’s developing theory of mind. Journal of the American Psychoanalytic Association, 44, 117-142.
Mitchell, S. A. (2000). Relationality: From attachment to intersubjectivity. Hillsdale, NJ: Analytic Press.
Orange, D. M., Atwood, G. E., & Stolorow, R. D. (1997). Working intersubjectively: Contextualism in psychoanalytic practice. Hillsdale, NJ: Analytic Press.
Stern, D. (1985). The interpersonal world of the infant: A view from psychoanalysis and developmental psychology. New York: Basic Books.
Winnicott, D. W. (1982). Playing and reality. New York: Tavistock. (Original work published 1971)


© APA Div. 39 (Psychoanalysis). All rights reserved. Readers therefore must apply the same principles of fair use to the works in this electronic archive that they would to a published, printed archive. These works may be read online, downloaded for personal or educational use, or the URL of a document (from this server) included in another electronic document. No other distribution or mirroring of the texts is allowed. The texts themselves may not be published commercially (in print or electronic form), edited, or otherwise altered without the permission of the Division of Psychoanalysis. All other interest and rights in the works, including but not limited to the right to grant or deny permission for further reproduction of the works, the right to use material from the works in subsequent works, and the right to redistribute the works by electronic means, are retained by the Division of Psychoanalysis. Direct inquiries to Henry Seiden, Publications Committee chair.