The Other, clinical and empirical: A review of Fonagy et al. on Affect regulation, mentalisation, and the development of the self

The following is taken from a presentation delivered late in 2011. Despite the age of the piece, I thought it worth sharing, as it touches on issues from more recent debates such as the nature of psychosis, the meaning of neuroscientific data, and the ethics of treatment. My views on certain matters below, such as phobia, or the nature of signification, have changed since then, but my views on Fonagy are more or less the same.

Anglophone psychology has long objected to the alleged individualism of Freud and
psychoanalysis. Psychoanalytic theory, they say, focuses on the intrapsychic, not the
intersubjective. Adler was one of the first to add a “social‟ element to psychodynamic
theory, positing a lack of “social interest‟ as the cause of every neurotic illness (Adler,
1928/1998). Later, a number of largely US-based psychoanalysts, led by Heinz Kohut,
championed an “intersubjective‟ or “relational‟ psychoanalysis.
The need to come to grips with the social has shown up elsewhere. The leading
neuroscientific theorists – people like Damasio, or Panksepp – invariably pay lip service
to the “social‟ in their formulations, without ever elaborating upon this latter. Cognitive
behaviour therapy has attempted to construct a so-called cognitive analytic therapy, shoehorning Bakhtin’s theory of dialogue into the standard formulations of cognitivism.

Of course, tacking on a social element to an otherwise worldless subject is not to be
confused with a theory of alterity. Likewise, allegations of individualism against Freud
miss the implicitly relational “identity‟ of the subject that Freud constructed in his early
psychoanalytic work. What was implicit in Freud‟s earlier work – a subject woven from
others, part of a circuit of introjected discourses – became explicit by the 1920s.
Civilisation and its Discontents is, in part, a meditation on relations between man and
men. Theorisation of the social can be seen in his paper on Group Psychology (Freud,
1921/2001), where Freud writes that all individual psychology is, at the same time, social

The relations of an individual to his parents and to his brothers and sisters, to the object of his love, and to his physician – in fact all the relations which have hitherto been the chief subject of psychoanalytic research – may claim to be considered as social phenomena. (p. 69)

Only the most profoundly autistic child would fall within the realm of an “individual
psychology‟, since relations to others are established almost immediately upon birth (if
not before). All “natural‟ bodily process – feeding, sleep, the excretory functions – are
rapidly subjected to a regime of socialisation in the child‟s early life. In Freud‟s theory of
psychosexual development, he described this in terms of relations to part-objects, such as the breast. Against this backdrop, the subject passes through stages – the oral, anal,
phallic, and so on. For Lacanians, logical development trumps the chronological, as the
Imaginary and Symbolic come to be knotted with the Real.
Somewhat outside of psychoanalysis, though nonetheless influenced by it, John Bowlby
theorised the attachment of the infant to its caregiver. “Attachment‟ refers to the strength
and quality of the bond between infant and caregiver, usually the mother, and it is one of
the most empirically robust, and consistently verifiable concepts in Anglophone
psychology. Whilst much of empirical psychology can and should be treated with
suspicion by psychoanalysts, I argue here that attachment theory is an exception, and
must be engaged with by anybody interested in the development of human subjects. The
proximity of the other, first grasped by way of touch (and only subsequently by the other
senses and part-drives), is a kind of “transcendental horizon‟ within which all other
developmental moments occur. My discussion today will be an attempt at a Lacanian
reading of one theorist of this transcendental horizon, psychoanalyst Peter Fonagy. I will
examine Fonagy‟s synthesis principally through his magnum opus (co-written with three
others), Affect Regulation, Mentalisation and the Development of the Self.
Fonagy’s project, it seems to me, is an attempt to produce a psychoanalytic account of
childhood relational development, and attendant pathology, that is empirically acceptable
within the framework of mainstream, Anglophone psychology. He and his colleagues are
at pains to tie every theorisation to experimental data. Nonetheless, the work is more
philosophical than the standard psychology text. I shall examine the rudiments of
Fonagy’s theory.
Fonagy begins with his notion of “mentalisation‟, or “reflective function‟, which is a kind
of offshoot of Baron-Cohen‟s “theory of mind‟. In profound autism, the other’s eyes are
not windows to the soul, constituting a gaze, but rather, are mere things. In contrast, the
non-autistic infant seeks eye contact from its first days of life. Eventually, the child
comes to presuppose that others have minds, and, in a somewhat Hegelian fashion, comes to recognise its own mind as an object for reflection. This reflection, according to
Fonagy, needn’t be conscious, but rather, can function as a presupposition that allows the
child to interpret other’s actions and intentions with coherence. This concept of
“mentalisation‟ is at the core of Fonagy’s work: the child needs others “to develop a sense
of self‟, and this, according to Fonagy, is probably the single most important acquisition
of early childhood. (Fonagy does not raise the question of precisely how a “self‟ is to be

How does mentalisation occur? Fonagy’s answer to this has some curious parallels to
Lacan’s mirror stage, with some important exceptions. To quote:

Anxiety is…for the infant a confusing mixture of physiological changes, ideas, and behaviors. When the mother reflects, or mirrors, the child’s anxiety, this perception organizes the child‟s experience, and he now “knows” what he is feeling. The mother’s representation of the infant‟s affect is represented by the child and is mapped onto the representation of its self-state…Within this model, mirroring would be expected to fail if it is either too close to the infant’s experience or too remote from it. If the mirroring is too accurate, the perception itself can become a source of fear, and lose its symbolic potential. If it is absent, not readily forthcoming, or contaminated with the mother’s own preoccupation, the process of self-development is profoundly compromised…
Although this idea is speculative, .it is empirically testable. (p. 35).

For Lacan, the fragmented body finds a unity in its mirror image, shifting the subject
from the specular to the social as he identifies with and assume his mirror image. Images
are always deceptive, and the imaginary register is therefore the domain of
méconnaissance. (Indeed, much of Fonagy’s data lends inadvertent support for Lacan’s
early thesis – see p. 212, for instance. Children as early as 6 months old, and possibly
earlier, are extremely sensitive to the image of a significant other, and to contingency
patterns governing this image). For Fonagy, mirroring is performed by a literal (rather
than imagic) other, or m/other, whose reflections must be “good enough‟. Fonagy
marshals a considerable quantity of empirical data to support this thesis. Should the
mother provide too accurate a reflection of his mental states, the child is overwhelmed. It
is somewhat akin to the lack of a lack. On the other hand, the absence of reflection,
whether through erratic parenting or insecure attachment is just as deleterious. It should
be said, however, that Fonagy’s formulations on this point have important differences to
Lacan’s. First, Fonagy adopts a thoroughly cognitivised discourse, and is subject
to all of the flaws of any such discourse; this entire process is subsumed under something
he calls “representational mapping‟. Secondly, whilst it is rare and admirable to see a
non-Lacanian analyst treat language with any seriousness, Fonagy‟s handling of it here is
questionable. There is no systematic thinking through of a distinction between the
Imaginary and Symbolic. Instead, the symbolic itself comes to be too laden with affect in
the cases of panic or phobia, for instance. “{T]he signifier is not sufficiently
“demotivated” – in other words, it resembles the signified too closely‟. (p. 35).
Elsewhere, Fonagy describes the mother’s “empathic face‟ as the signifier to the infant’s
own “emotional arousal‟ as the signified (p. 127). This seems to me a misunderstanding
of Freud‟s condensation or displacement (or Lacan’s metaphor and metonymy). The
phobic object can only function qua symbolic insofar as it has been displaced as a
signifier from a signified. Freud‟s split between an „affect‟ and an unconscious,
„unacceptable idea‟ is recapitulated, without Freud‟s aporias, in Lacan‟s distinction
between signifier and signified. Where there is a failure to cleave the Saussurian sign into
two, it should, theoretically, at least, follow that there can be no repression, and no
unconscious. The result would be psychosis, not panic.
Nonetheless, the infant is presumed in Fonagy to “read‟ the mental states of others,
primarily its caregivers. Its ability to do so is closely correlated with the security of the
infant-parent attachment, and again, much experimental data is brought to bear on this
point. Secure attachment is associated with favourable outcomes across many tasks,
particularly those calling for symbolic capacity. It is as if attachment supplies the
Bejahung required for Freud‟s subsequent Verneinung, or, to put it in Lacanese, for the
Nom-du-père to become operational, there must first be a primordial Oui between mother
and child.
Fonagy and his colleagues place great weight on the work of neuroscientists, and, like
Damasio, he seems to accept a distinction between a kind of “neural self‟, and a
phenomenological, subjective, and linguistically constructed “I‟ (p. 79). Whilst a little too
much deference is paid to the findings of neuroscience, Fonagy does mount a substantial
critique of genetic explanations of behaviour, the more extreme of which counsel that
virtually no significant trait of a child can be modified by its parents actions.
In contrast, Fonagy and co argue that, in “good enough‟ circumstances, the infant can
attribute emotions to others by the end of its first year (p. 159), and can draw inferences
based on the display of others’ emotions. This evidence leads Fonagy to conclude that the
infant possesses something by way of “representation‟ of emotion. The infant, according
to Fonagy, is counting on the truth of Lacan‟s maxim of Seminar I, namely, that “feelings
are reciprocated‟. The infant needs to feel some degree of “contingent control‟ over the
parent’s emotional expression but must not have (or be duped into believing he has)
complete control. Even when faced with “empathic parenting‟, the child must experience
some degree of disparity, however slight, between his affect and his caregiver’s, lest he be persuaded that his parent is the victim of emotional contagion. Children themselves will
show a preference for this, according to Fonagy’s own experimental data. Children attend
more to imitation-based (highly but imperfectly) contingent image than to the perfectly
contingent one (p. 189). The child, for Fonagy, prefers a high but imperfect degree of
control of the other’s responses – it is as if there needs to be some limit set to the
complete accession of his demands, in order for desire to emerge. Of course, the parent of an infant will often interact with it accompanied by exaggerated prosody and intonational
vocal patterns. This is called “markedness‟ by Fonagy, and “motherese‟ by linguists. In
(later) Lacanian theory, it is called Lalangue, and it is a necessary precondition of the
development of speaking subjects. This markedness of emotional representations forms
part of what Fonagy and co call their “social biofeedback‟ model, which they link to
Winnicott’s model of the mother’s holding function, and Bion’s model of maternal
containment (p. 191).
What has all this to do with the clinic? Well, according to Fonagy, there are a range of
“deviant‟ mirroring styles, which fail in a number of ways. For instance, if the parents
expressed emotion in an unmarked but highly “realistic‟ manner in response to the
infant’s distress, the latter may presume that his own emotion is “out there‟, “belonging to
the other, rather than to himself‟ (p. 193). This leads the child into “psychic equivalence‟,
whereby the mind is not decoupled from the world, and forms the basis for subsequent
projective identification. This has clinical ramifications. For instance, Fonagy sees
Borderline Personality Disorder (BPD) as an extension of psychic equivalence. A
relationship breakdown or rejection is experienced by the borderline as not merely
distressing or mournful, but as literally self-obliterating. It is this “literalness‟ which
constitutes the sine qua non of BPD, and, in children, might be manifest in a difficulty to
pretend whilst playing. In Lacanian terms, this literalness would be placed on the side of
structural psychosis. For the neurotic, the “pretend mode‟ can be a means of working
through painful thoughts and memories without being overwhelmed, since the latter are
recognized as being mere mental entities. In the mode of psychic equivalence, that which
belongs to the subject and world is not so clearly delineated.
An enormous amount of development occurs within a relatively brief period. For
instance, in an experiment in which toddlers were asked to give the experimenter food,
18-month olds could respond according to the experimenter‟s previously exhibited
preferences, such as when they said “yuck‟ or “yum‟ in connection with a food item. In
contrast, a group of 14-month olds could not do the same, and always tended to give
experimenters the food that they liked, showing a failure to distinguish between their own
desires, and the desires of another (p. 238). Around a similar age – 18 to 24 months –
toddlers can clearly distinguish between themselves and their image and, if placed before
a mirror with a rouge mark on their nose, for instance, will seek to remove it from
themselves, and not the mirror image, as would a younger child. Given the enormous
sensitivity of children to their parents’ interactions, abusive or negligent parents, or bizarre
and incoherent expressions of emotion are implicated in Fonagy‟s aetiological models of
What, then, of the Oedipal drama, as we know it in Freud, or in the elaborate re-working
it receives from Lacan in the 1950s? Fonagy’s answer is to use the array of data that he
and his colleagues have amassed to attempt a re-working of his own, wherein
“mentalisation‟ becomes the pivot on which the process turns. At the time of the Oedipal
conflict (that is, around age 4), the child, by “using the parent’s mind, is able to play with
reality‟ (p. 267). Through a secure attachment bond, and with appropriate parental
mirroring, the child is able to internalize triangulation, and find representations for it. (At
this point, mentalisation reads like a cognitivised version of Melanie Klein‟s depressive
position). Whilst “reality‟ in this iteration goes largely unquestioned, the element of the
“pretend‟ here is important, for, as Fonagy notes:

The young child, attempting to make the developmental step between a nonrepresentational and a representational mode of psychic reality, is in a highly vulnerable state…The fantasy of the sexual possession of the parent of the opposite sex is safe as long as it is held in the pretend mode of psychic reality, where concerns with possibility or impossibility, or relation to physical reality, are absent. Pretend desire, even if conscious, arouses no conflict. Only when the representational mode of pretense comes
gradually to be integrated with the experience of psychic reality corresponding to external reality – where the thought suddenly becomes real – do terrifying conflicts arise throughthe increasingly clear image of the feelings of the other, in seeing the child’s wish. Resolution of this dilemma normally arises through the radical restriction of such dangerous fantasies, through the establishment of the repression barrier, first to a pretend mode and later to an unconscious thought. (p. 280).

This looks very much like a passage from proto-psychosis to a fully-fledged, neurotic
subjectivity, of either the hysteric or obsessional variety. Whilst Fonagy clearly indicates
that this movement takes place only through the intermediary of the Other, he does not
have a fully-worked out notion of language or the Symbolic to rely upon. Hence, in the
instance of intergenerational trauma, transmission between generations does not occur by
way of signifiers or identifications, but by a lack of parental mentalisation. If the parents
refuse the trauma entry into “shared reality‟, by which it can be talked about, something
which is not “metabolised‟ is passed from one generation to the next (p. 287-288). It is
the parent or caregiver who lends the child the representations to process “reality‟.
Fonagy comes close without quite articulating the notion of a desire of the Other.
Specifically, he raises the question of the aetiological consequences of an Other whose
desire is abusive, or malevolent. Fonagy argues that, in some instances, the child simply
refuses to conceive altogether of the intentions of the abusive other (p. 346). This, we
might argue, is tantamount to Lacan’s foreclosure. Seen in this light, psychosis as well as
neurosis can be understood as a kind of solution to a problem. In Fonagy speak, abuse
and trauma undermines (but does not necessarily prevent) “mentalisation‟ and reflective
function, because the child “who recognizes the hatred or murderousness implied by the
parent‟s acts of abuse is forced to see himself as worthless or unloveable‟ (p. 353). I was
reminded here of research into “outcomes‟ for child victims of incestuous abuse: the chief
predictor of subsequent psychopathology was neither the type of abuse, nor the role of
counselling, but the reaction of the non-offending parent, who could offer a Symbolic
structure for such difficult events, or, alternatively, foreclose them. For Fonagy as for
Lacan, it is a moment that determines the choice, forced or otherwise, of psychosis or
neurosis. Both analysts are skeptical of the notion of “regression‟, but to the extent that it
can be said to exist within their respective frameworks, it operates on roughly similar
terms. Regression in Lacan can be from the Symbolic to the Imaginary, with the effects
of foreclosure being felt in the Real (terrifying hallucinations, for instance). In Fonagy,
regression is a step back from mentalisation to “nonmentalistic thinking‟ (p. 356), often
accompanied by some disturbing or traumatic event.
Thus, the autistic child or the psychotic in Fonagy is very much like Lacan’s “structural
psychotic‟. (The Borderline is likewise, but Fonagy does not draw out the conclusions of
this. One wonders whether sociopathy also belongs in the same general category). Again,
the link with trauma is crucial:

Trauma may…disrupt the representation of feelings or thoughts by creating a propensity for shifting into the infantile mode of pretend. Some traumatized children grow up with an apparent hypersensitivity to mental states, needing to guess immediately what those around them feel and think in order to preempt further trauma. As part of this, a pseudoknowledge of minds can develop, which is superficial and may be very selective, scanning for particular danger signals and avoiding reflection on meanings or connections. (p. 383)

This sounds very much to me like the “high-functioning‟ end of the psychotic spectrum,
as if it were a description of the paranoiac (or, perhaps, a hint at the paranoic nature of
belief and knowledge). Fonagy characterizes psychotic and borderline states as a failure
of sharing “psychic reality‟ at the level of “symbolic thought‟ (p. 416), the latter being
“blocked‟, to use Fonagy‟s term. This causes thinking to be “felt as words that have been
spoken and cannot be taken back‟. What Fonagy arguably misses, by not having examined linguistics more profoundly, is that the “symbolic thoughts‟ in question are themselves a discourse, and are signs divisible into the signifier/signified distinction. The counterargument is that the psychotic does not mistake thoughts for words by words for things, since the breach in significantion (and subsequent displacement, metonymy, etc) has not been effected through repression.At the descriptive level, Fonagy‟s account of a certain “alienated‟, non-mentalising group of individuals rings true (see p. 471, for instance); the DSM and related systems cannot adequately grasps these phenomena, as Fonagy points out. Nonetheless, the explanation seems frustratingly incomplete, due partly to the familiar lack of Symbolic explanations, but more importantly, I would argue, due to the normative presumptions made throughout his work. The extent to which these
presumptions colour Fonagy’s work remains a problem, as we shall see.

The goals of treatment for Fonagy are very much those critiqued by Lacan throughout his career. There are some exceptions to this. For instance, Fonagy argues that treatment with borderlines (and psychotics) is not necessarily about trying to make the unconscious conscious, and still less to devolve into “symptom management‟. Rather, the “technical priority is the analyst‟s survival‟ (p. 408) or, more specifically, “the preservation of a clear picture of the patient’s mental state in the analyst’s mind‟. In the absence of a “containing or „good-enough‟ parent, the analyst can “provide the necessary frame forplay and reflection‟ (p. 285). The survival of the analyst is one thing; his or her ownImaginary quite another. The confounding of symbolic and imaginary that pervades Fonagy’s work elsewhere reappears here. Indeed, this murkiness is tied to the aims of a “mentalising‟ analysis, in which treatment is supposed to “help an individual to achieve a higher level of intersubjectivity, in terms of deeper experiences with others and ultimately a life experienced as more meaningful‟ (p. 265).

The imaginary is not the only problem. Analysis proper, in a Lacanian sense, at least,
mingles with “adaptation‟ in Fonagy’s work. (I can only assume that this problem is
exacerbated among the non-analytic psychologists who have taken up his work). The
transference is used to recreate “the original attachment context‟, and thereby get the
patient to re-experience parental affect in the “reality-decoupled safe environment of the
analytic situation‟. Fonagy and co provide numerous clinical case studies where this
approach is illustrated, with patients ranging from the overtly psychotic to cases of mere
neurosis. This impulse to create a safe environment within the analytic setting, however,
is paired with the goal of having the analyst provide “corrective emotional and
representational experience for the patient through emotion-regulative marked
externalizations of the patient’s affects in the form of mirroring interpretations or
expressions‟ (p. 316). This, clearly, is a rather different aim to encouraging an analysand
to hold fast to his or her desire, and Fonagy and his followers do not pause to consider
that which in the affective might escape “regulation‟. The idea of an affect which
deceives is not countenanced any more than the notion of an unregulatable jouissance or
anxiety. Mentalisation becomes a kind of cognitive skill, a take-home “strategy‟ that
recreates the worst elements of ego psychology (as critiqued by Lacan) and CBT.
Mentalisation is, in the last resort, about the “correction of maladaptive models‟ of
interaction, and the “strengthening‟ of “mental capacity‟ to interact differently (p. 470).
One might argue, of course, that these “corrective‟ goals can be decoupled from the
theory itself. It may be possible to abstract the cognitive language and unacceptable aims
of Fonagy‟s thesis to leave a kernel of important psychoanalytic thinking, and indeed, I
think this possibility more likely than not. It may be as a consequence of striving so hard to
integrate his psychoanalytic thinking into a university discourse that Fonagy’s work
descends, like so many of his predecessors, into strategy-mongering. Yet this need not be the case. Much of the vast pool of primary research cited by Fonagy could be susceptible to a Lacanian reading, and support a broadly Lacanian thesis on aetiology, and the development of personality structure. The Lacanian aversion to university discourse is not entirely ubiquitous, and the next logical step, it seems to me, is to reexamine Fonagy’s material through a different epistemology. Such an epistemology is supplied by Belgian psychoanalyst Paul Verhaege, but his work will have to be the subject of another discussion, another time.

Adler, A. (1938/1998). Social interest. Oxford: One World.
Fonagy, P, Gergely, G., Jurist, E., & Target, M. (2004). Affect regulation, mentalisation,
and the development of the self. New York, NY: Other Press.
Freud, S. (1921/2001). Group psychology and the analysis of the ego. Vol. 18. The
Standard edition of the works of Sigmund Freud. (trans. J. Strachey). London:


1 thought on “The Other, clinical and empirical: A review of Fonagy et al. on Affect regulation, mentalisation, and the development of the self

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s