Critical thinking

73

Chapter 3

Development

We articulated the multifaceted nature of mentalizing in the previ- ous chapter, and we concluded by proposing that, of all the ordinary language and clinical concepts, mentalizing is uniquely suited to capture the essential unity of understanding oneself and others. As we elabo- rate in this chapter, the basis of that unity is the developmental process of coming to be a self in relation to others.

Of course, as a clinician, you need not know how mentalizing devel- ops in order to engage your patients in mentalizing; you have been do- ing that all along. Yet we aim to promote more refined attention to men- talizing in clinical practice, and understanding the developmental conditions that enhance and inhibit mentalizing makes a crucial contri- bution to that refinement. Accordingly, we summarize current research on the development of mentalizing that we consider essential to under- standing how to facilitate mentalizing in clinical practice.

Understanding the development of mentalizing requires counterin- tuitive thinking. Having acquired mentalizing capacities long ago, we adults are accustomed to taking a rich and private inner world for

74 Mentalizing in Clinical Practice

granted. We are naturally sympathetic with Descartes’s dictum: I think, therefore I am. Moreover, we routinely employ our inner world to em- pathize with others through imagination and simulation (Goldman 2006; Stueber 2006). The use of inner experience to understand others creates a sense of the unity of minds and entails reasoning from the in- side out: others are like me. Intuitively, we apply this line of reasoning to development: the child somehow comes to have a mind and then in- fers that others do too.

This natural Cartesian way of thinking obscures what Peter Hobson (2002) claims: “To put it bluntly, if an infant were not involved with other people, then she would not come to think” (p. xiv). As he elabo- rated, “The links that can join one person’s mind with the mind of someone else—especially, to begin with, emotional links—are the very links that draw us into thought” (p. 2). Accordingly, he titled his mar- velous book The Cradle of Thought. Thus Descartes was speaking accu- rately for the adult; speaking for the child in the making, we might put it differently: Mother thinks I am, therefore I am. Accordingly, counter- intuitively, we embrace Vygotsky’s (1978) view that mind develops from the outside in: “An interpersonal process is transformed into an intra- personal one. Every function in the child’s cultural development appears twice: first, on the social level, and later, on the individual level; first, be- tween people (interpsychological), and then inside the child (intrapsycho- logical)” (p. 57, emphasis in original).

In this chapter we first describe the developmental steps through which mentalizing emerges in relationships, and then we elaborate the complex interplay among parental mentalizing, attachment, and the development of mentalizing in the child. We also introduce what will be one of the primary concerns of this book, namely, how attachment trauma undermines the development of mentalizing and thereby con- tributes to developmental psychopathology.

Developmental Progression The capacity to mentalize is a multistage developmental achievement that develops out of a host of capacities evident in early infancy that promote social engagement (Hobson 2002; Klin et al. 2000; Stern 1985). We begin this section by presenting a broad overview of the developing appreciation of the representational nature of mental states, which con- stitutes the ultimate sense of selfhood (Fonagy et al. 2002a). Then we elaborate four developmental processes that play major parts in the sense of self and others as having representational minds: emotion reg-

Development 75

ulation, joint attention, language, and pedagogical interactions. To aid the reader in holding this developmental complexity in mind, we pro- vide an overview of these main contributions in Table 3–1. All facets of development exemplify a central theme of this chapter: contingently responsive mentalizing on the part of the caregiver promotes the develop- ment of mentalizing in the child. We conclude this section with a discus- sion of three prementalizing modes of functioning that can become re- gressively activated in psychopathology: psychic equivalence, pretend mode, and teleological mode.

Mental Representation The progressive development of mentalizing capacities is intertwined with agency, which, in turn, is intertwined with a sense of self and oth- ers. Fonagy and colleagues (2002a) have traced this development in five steps, outlined in Table 3–2. Here we focus on the emergence of mental representation, which we introduced in the context of a sense of repre- sentingness of mental states in Chapter 1 (“Introduction”) and revisited briefly in the context of theory of mind in Chapter 2 (“Mentalizing”).

As sketched in Table 3–2, infants soon learn that they (and others) are physical and social agents, capable of influencing—and being influ- enced by—external objects and other persons. Then, by around nine months of age, adopting a teleological framework, they expect agents’ actions to be rational and goal directed (Csibra and Gergely 1998; Gergely and Csibra 2003). For example, at nine months, infants take it as a matter of course that one computer-animated agent takes a straight path as if to make contact with another, or jumps over an obstacle to do so if need be; yet these infants register surprise when one agent jumps needlessly to reach the other when there is no obstacle in its path. None- theless, notwithstanding their budding appreciation of rationality, nine-month-old infants do not yet comprehend the goal-directed ac- tions of the agent as based on mental states; rather, expectations of agents’ “rational” actions are based solely on observable physical reali- ties (e.g., a straight or obstructed pathway).

During the second year of life, however, infants begin mentalizing the teleological stance (Gergely et al. 1995); they develop a mentalistic under- standing of goal-directed agency. Infants develop a number of pertinent cognitive competencies, albeit rudimentary, during this period (Fonagy 2006): they interpret intentional agents’ actions as stemming from de- sires, wants, and intentions (Wellman and Lagattuta 2000); they have an implicit understanding of true and false beliefs (Onishi and Baillargeon 2005); they engage in shared imaginative play that facilitates coopera-

76 Mentalizing in Clinical Practice

TABLE 3–1. OVERVIEW OF CORE PROCESSES IN THE DEVELOPMENT OF MENTALIZING

Process Description

Understanding of mental representation

A mentalistic understanding of goal-directed actions develops in which multiple mental models or perspectives can be contrasted (e.g., actual versus possible, past from present or future) and, ultimately, meta-representation becomes possible (e.g., reflecting on the accuracy of beliefs and awareness of thinking and feeling as such).

Emotion regulation

Understanding emotions plays a formative role in mentalizing and arises from a mirroring process in which infants’ emotional expressions elicit a high but imperfect level of caregiver responsiveness that is “marked” by ostensive cues that show the infant that the caregiver is representing the infant’s emotional state— not expressing the caregiver’s own emotional state; this process enables the infant to develop self-representations of emotional states that promote emotion regulation and expression.

Joint attention Awareness of attention directed to the self is evident in early infancy, and this awareness is succeeded by awareness of attending conjointly with another to a third object and by purposefully drawing others’ attention to objects. Joint attention includes implicit emotional commentary on objects (as in social referencing) and, in this context, the other’s attention to the self ushers in a special sense of self-awareness (and self-consciousness) as a person among persons.

Language Language acquisition and mentalizing evolve in a bootstrapping process inasmuch as language acquisition requires joint attention and the appreciation of a communicative intention; in turn, linguistic capacities permit explicit mentalizing along with a refined conceptual framework for articulating mental states and reasoning about them (i.e., meta-representation).

Pedagogy The caregiver’s contingent responsiveness to the infant’s emotional states can be construed as a process of implicit teaching, and the development of mentalizing is part and parcel of a human-specific adaptation for pedagogy, that is, an efficient means for teaching and learning cultural information—including knowledge about mental states. Like language, mentalizing and pedagogical learning develop in a bootstrapping process.

Development 77

tive skills (Brown et al. 1996); and they begin acquiring a language to represent internal states (Repacholi and Gopnik 1997). Nonetheless, at this early stage infants are unable fully to separate mental states from external reality; the distinction between internal and external remains blurred (Fonagy 2006).

As we noted in discussing theory-of-mind research in Chapter 2 (“Mentalizing”), between three and four years of age, children develop a full-fledged explicit understanding of representational mental states, as exemplified by passing false-belief tasks that require linguistic artic-

TABLE 3–2. LEVELS IN THE DEVELOPMENT AND REFINEMENT OF MENTALIZING CAPACITY

Level Description

Physical Infants first develop a sense of self and others as physical agents, as distinct from physical objects, on the basis of being self-propelled and animated; physical agency contributes to the sense of self as author of actions (e.g., making one’s limbs move) and as having influence over external objects (e.g., making the ball move), thus facilitating self-other differentiation.

Social In tandem with physical agency, infants develop a sense of self and others as social agents; social agency encompasses awareness that communicative displays exert influence over other social agents (e.g., smiling prompts mother to smile).

Teleological In the second half of the first year of life, infants develop a sense of self and others as teleological agents whose actions are purposeful and goal directed; at this level, infants expect actions to be rational, that is, efficiently directed toward goal attainment within the constraints of external reality (e.g., taking the straightest path to a goal).

Mentalizing During the second year of life, infants begin mentalizing the teleological stance in the sense of interpreting rational, goal-directed actions as governed by intentional mental states, ultimately taking into account the possibility that the mental states governing actions may not correspond to reality (e.g., false beliefs).

Autobiographical By the sixth year of life, children organize memories of their actions and experiences into a causal-temporal framework that permits the development of an autobiographical self and the understanding of self and others through coherent autobiographical narratives.

78 Mentalizing in Clinical Practice

ulation (i.e., anticipating that a child will look for something on the ba- sis of where he wrongly believes it is located rather than on where it is actually located). In effect, in coming to develop an explicit understand- ing of mind as representational, children’s thinking comes into align- ment with cognitive science, which is predicated on the representa- tional theory of mind (Fodor 2003). Perner (1991) has described the monumental developmental significance of this shift:

representation is not just one aspect among others of the mind, but pro- vides the basis for explaining what the mind is. In other words, by con- ceptualizing the mind as a system of representations, the child switches from a mentalistic theory of behavior, in which mental states serve as con- cepts for explaining action, to a representational theory of mind, in which mental states are understood as serving a representational function. One can think of the concept of “representation” as playing a catalytic role in children’s reconceptualization of what the mind is. (p. 11; em- phasis in original)

As Perner (1991) outlines it, the development of representational ca- pacity frees the mind from reality. Going beyond perceptual representa- tions, in the second year of life infants are able to construct and employ multiple models of the same situation, enabling them to formulate means-end relationships (e.g., contrasting an existing state with a de- sired state) and temporal relationships (i.e., contrasting past, present, and future). Then it becomes possible to have in mind something that is not present as well as to experience something that is present in different ways. The capacity for pretend play best exemplifies this decoupling of mental representations from reality (Leslie 1987); an object is represented as if it were something else, thereby acquiring a mental existence. Hob- son (2002) gave the example of his son pretending that a spoon was a car: “The car-thought could be applied to the spoon-object, so that the spoon became a thought-car” (p. 77). Yet the full flowering of mind comes with the capacity for meta-representation (i.e., representing representations, as in “I think, therefore I am”); then the mind becomes aware of itself and its place in the world. The move to this meta-representational level per- mits cognitive and affective self-regulation through metacognition, as described in Chapter 2 (“Mentalizing”). While we might seem to be get- ting abstruse here in delving into meta-representation, this level of un- derstanding is critical to our clinical work: patients in the throes of de- pressive rumination, paranoid projections, and posttraumatic flashbacks have lost a sense of their mind as mental—a sense of representingness. Accordingly, to mentalize, they need our help to keep in mind that their mental states are the work of their mind.

Development 79

In fostering mentalizing in adults for the purpose of improving their functioning, we rest on solid developmental ground. In children, full- fledged mentalizing as marked by meta-representation is associated with a wide range of social-developmental capacities (Fonagy 2006): empathic behavior (Zahn-Waxler et al. 1992) and good relationships with peers (Dunn and Cutting 1999); elaborate pretend play (Taylor and Carlson 1997); more conversational fluency (Slomkowski and Dunn 1996); and greater social competence (Lalonde and Chandler 1995). Notably, there is also a downside to mentalizing: understanding of false beliefs also enables children to deceive and to play tricks and jokes (Sodian et al. 1992). Lastly, meta-representation includes the ca- pacity to understand that behavior is influenced not only by transient mental states but also by enduring personality dispositions, providing the foundation for an emerging self-concept (Flavell 1999). As noted in Table 3–2, self-representations ultimately become organized into an au- tobiographical narrative. Correspondingly, mentalizing includes the ca- pacity to understand others on the basis of their autobiographical selves—as we routinely do in conducting psychotherapy. And, as the practice of psychotherapy attests, mentalizing the self and others at in- creasing levels of refinement—in terms of accuracy, richness, and flexi- bility—is a lifelong endeavor.

Contingent Responsiveness and Emotion Regulation Back to the beginning: Hobson (2002) asserted that “It is through emo- tional connectedness that a baby discovers the kind of thing a person is” (p. 59, emphasis in original). Similarly, Gergely and Watson (1996) pro- posed that, on the grounds that basic emotions are universal and innate, “emotions are among (if not the) earliest mental states that infants at- tribute to minds” (p. 1183). These latter researchers pinpointed a mech- anism that sets the stage for mentalizing emotion, that is, going from the behavioristic stance of using emotional signals to predict behavior to recognizing emotions as mental states in others and oneself. This devel- opmental progression is prompted by a shift in attentional preference. In the first few months of life, infants prefer perfectly contingent re- sponse!stimulus contingencies (Bahrick and Watson 1985), for exam- ple, watching as they move their limbs. By three months of age, infants know that anything which is not invariably and perfectly contingent on their own actions belongs to the external world (e.g., I move my arm and the ball beyond reach stays put).

Gergely (2001) identified a switch in contingent responsiveness around three months of age, at which point infants shift from a preference for

80 Mentalizing in Clinical Practice

perfect contingency to high-but-imperfect contingency, that is, slightly out of sync, novel, and unpredictable. This shift has monumental conse- quences for psychosocial development and for the development of mind in particular: infants switch their preference from attending to their own actions to attending to the emotionally responsive social envi- ronment. The mother’s facial responses to her infant’s emotional states are a perfect example of high-but-imperfectly contingent responses— just what her infant has come to prefer. Whereas infants’ initial focus on perfect contingencies enables them to discover their bodily self in the physical world, their subsequent focus on highly but imperfectly contin- gent social responsiveness enables them to discover their mental self in the social world. This developmental step leads to a corresponding shift in the nature of affect regulation in the mother-infant relationship. Prior to the switch in contingency preference, comforting is provided by the mother’s direct physical ministrations (e.g., comforting touch); subse- quently, comforting can be provided by emotional communications that foster the development of affect representations in the infant.

More specifically, affect representations arise out of the early mirror- ing interactions of caregiver and infant aimed at soothing and downreg- ulating the infant’s arousal. Mirroring—resonating with, reflecting on, and expressing the internal state that the infant displays—is a universal, biologically prepared (instinctual) response on the part of all adults (Meltzoff and Moore 1997). For affect mirroring to support a represen- tational framework, however, the mother must indicate that she is not showing the baby her own feelings, but rather her awareness of the baby’s state. This form of mirroring is characterized by markedness (Fon- agy et al. 2002a; Gergely and Watson 1996), that is, “marked” in the sense of being marked for sale or tagged. By intermingling an accurate reflection with an incompatible affect, or by exaggerating her display of the affect, the caregiver marks the expression as being as-if or pretend. For example, a mother might successfully soothe her infant by inter- mingling a reflection of the infant’s distress with irony in her facial ex- pression, or she might mimic the baby’s frustration while showing con- cern. More generally, typical features of marked emotion include an exaggerated, slowed-down, or partial expression of the emotion; ex- pressions of mixed emotions, combined simultaneously or sequentially; and behavioral cues such as raised eyebrows that frame the expression as intended for the infant’s attention (Gergely 2007).

This process of mirroring emotion can go awry in two ways: either by failing to mark the affect or by lack of contingent responding. That is, expressions that correctly mirror the baby’s state, but lack markedness, may overwhelm him. For example, rather than responding to the in-

Development 81

fant’s frustration with a marked expression, the mother may feel angry and express her anger directly toward the infant. The infant feels such emotional expressions to be the parent’s own real emotion, making the infant’s experience seem contagious, thus experienced as even more dangerous, potentially leading to trauma rather than containment. On the other hand, the mother may respond to the baby with a marked but non-contingent (i.e., inaccurate) reflection. For example, the mother who mirrors the baby’s excited biting of the breast as aggression might say, “Ouch! You are an angry little beastie today!” To the extent that it is internalized, mismatched mirroring generates alien internal experience that contributes to a fragmented sense of self (Fonagy et al. 2002a).

To summarize, the growth in understanding the self as a psycholog- ical (emotional) agent begins with infants’ discovery of their own af- fects through their primary attachment relationships. Infants’ budding capacities to experience their emotions as feelings are based on internal- izing the caregiver’s contingently responsive marked emotional expres- sions. The caregiver’s mental representations of the infant’s emotional states are exemplified in her behavior; intuitively, she presents these ex- pressions to the infant. In turn, the infant begins to develop a mental representation of his or her own emotional state as a feeling—an emerg- ing form of emotional self-awareness. These representations gradually form the basis for mentalizing emotion and thereby for affect regulation and impulse control: feelings become recognizable; they do not have to be acted out; and they can be shared.

Joint Attention Attention not only plays a cardinal role in the ongoing activity of men- talizing but also in the development of mentalizing capacity. Research on joint attention in infancy, like that on emotional mirroring, illustrates the Vygotskian perspective that the sense of self develops from the out- side in (Vygotsky 1978). Developmental research (Eilan 2005; Franco 2005; Reddy 2005) has elucidated the progression of joint-attention be- haviors as enumerated in Table 3–3.

Reddy (2005) contends that “the self is the first target of others’ at- tention that the infant experiences” (p. 106) and that “attention to the self may be the most direct and powerful experience of attention that is possible” (p. 86). Awareness of others’ attention to the self is evident in the first few months of infancy. First, two-to-four-month-old infants re- spond to others’ gaze at the self in varied ways. Illustrating how closely emotions are tied to attention, infants show responses ranging from pleasure and interest to ambivalence to distress and avoidance. Second,

82 Mentalizing in Clinical Practice

infants also direct others’ attention to the self, for example, by making utterances to bring about face-to-face engagement. At this time, infants also seek to reengage attention when it has been disrupted. In later months, infants’ direction of others’ attention becomes more refined; at seven to eight months, for example, infants go beyond soliciting atten- tion to the self as a whole when they draw attention to specific aspects of the self (e.g., an exposed tummy) or specific actions (e.g., showing off and clowning). Later, infants direct others’ attention to objects in hand (9–11 months) and then to distal objects (e.g., at 10–14 months, by point- ing). Reddy maintains that the objects and topics of attention, not the par- ticipation in attentional engagements, become more complex over the course of development.

Engagement in mutual attention in early infancy provides the infant with the experience of being the object of another’s attention. Of course, the early infant’s engagement in attention contact does not require a concept of attention as an internal, unobservable state of mind (Gómez 2005); much less does the infant have the capacity to represent the self as being represented in the mind of the other. As Reddy (2005) noted, the ob- jects of mutual attention become increasingly complex over the course of development. Whereas mutual attention in early infancy is dyadic, joint attention in the second half of the first year becomes triadic, in- volving the self, another person, and a third object. This triadic self- other-world relationship is fundamental; knowledge of any one re- quires knowledge of all three (Davidson 2001). Accordingly, Tomasello (1999) characterized the development of joint attention in relation to objects as the “nine-month social-cognitive revolution” (p. 11). Hobson articulated beautifully the significance of this triangulation for develop- ing a sense of multiple mental perspectives, the essence of mentalizing:

TABLE 3–3. MILESTONES IN THE DEVELOPMENT OF JOINT ATTENTION

• Responding emotionally to others’ attention to the self

• Directing others’ attention to the self as a whole and later to specific aspects and actions

• Following others’ gaze toward objects

• Directing others’ attention to objects in hand

• Directing others’ attention to distant objects by pointing

• Ensuring that others’ attention is engaged before pointing

• Checking others’ emotional responses to objects or situations (social referencing)

Development 83

What triangulation gives is a fixed point, a kind of pivot, around which different things are brought together. The pivot is the world. The two at- titudes bear upon the same thing in the world. A single thing is experi- enced as having two meanings. It is this that prompts the infant to sep- arate out her own attitude from that of the other. It is not just that the mother reacts and the infant feels new things—it is also that the mother and the infant are reacting to the very same object. Through this experi- ence of having both her own and her mother’s attitudes to the same things, the infant learns something about things on the one hand and at- titudes on the other. In reading her mother’s reaction to a toy, the infant learns something about the toy; but at the same time, the toy tells her something about her mother. What it tells her is that her mother is dif- ferent from herself, in a particular way. It tells her that her mother has an attitude to the toy that is separate from her own attitude to the same toy. (Hobson 2002, p. 109)

Triangulation in joint attention reflects a complex developmental process (Eilan et al. 2005). Joint attention entails first having the infant’s attention directed toward an object by someone else (e.g., looking where mother looks) and subsequently the infant’s directing another’s attention to an object (e.g., prompting mother to look where the infant looks). Joint attention is involved not just in following gaze but also in pointing and gesturing for others, holding up objects to show others, bringing others to specific locations to see an object, and offering objects to others by holding them out. Thus, as Hobson (2002) described, through joint attention, the infant develops an inchoate sense of sharing and comparing experience with another person.

A distinction between two forms of pointing illustrates the infant’s incipient transition from a behavioristic to a more mentalistic stance. In- fants first learn that pointing has the instrumental effect of inducing an- other person to do something (e.g., by pointing to the cup, the infant can induce mother to bring it). Subsequently, they learn to point to direct the other’s attention to something (e.g., wanting mother to see the kitty)— that is, for the sake of sharing attention in the sense of generating some joint emotional engagement in relation to the object (Liszkowski et al. 2004). An important milestone in the development of joint attention, ev- ident at about 18 months, is checking to ensure that the other’s attention is engaged before pointing—what Franco (2005) called “the seed of mentalizing” (p. 142). We will be discussing the pedagogical nature of such interactions shortly; here we note that pointing in joint attention also can have an interrogative function, that is, as a request for relevant information about the object of attention (Fonagy et al. 2007).

Joint attention serves a crucial epistemic function in enabling the in- fant to learn about the world. Closely related to joint attention is social

84 Mentalizing in Clinical Practice

referencing, although the relative developmental timing of the two phe- nomena is unclear (Reddy 2005). In social referencing, the infant checks the mother’s emotional reaction to a novel object, such as a noisy me- chanical toy, to determine how the infant should regard it; for example, as interesting or dangerous (Moses et al. 2001). More generally, joint at- tention entails implicit emotional commentary on objects of mutual inter- est, and Eilan (2005) maintains that such expressions of emotion can be construed as primitive predications: “Reciprocated smiles are recipro- cated comments on the world” (p. 28). We discussed the adaptive as- pects of mentalizing emotion in Chapter 2 (“Mentalizing”); to a consid- erable degree, we rely on emotional knowledge of the world, and we begin acquiring that emotional knowledge through social referencing:

there must be some sense in which this use of affective responses is im- bued with some sense of getting it right in some respect. Things are scary or not, funny or not. The suggestion is that this primitive sense of right and wrong begins to be manifested as social referencing sets in, where the appropriateness or not of the response is precisely what the child is seeking reassurance about. One way of putting this is that affect regula- tion becomes, at this stage, regulation of the appropriateness of the re- sponse to the world. (Eilan 2005, p. 28, emphasis added)

Thus far, we have described joint attention in its triadic form as in- volving the infant’s and mother’s attention to a third object. Yet Toma- sello (1999) also has described how, in this triadic context, attention to the self takes on new significance; self-awareness at a new level of complex- ity emerges:

As infants begin to follow into and direct the attention of others to out- side entities at nine to twelve months of age, it happens on occasion that the other person whose attention an infant is monitoring focuses on the infant herself. The infant then monitors that person’s attention to her in a way that was not possible previously, that is, previous to the nine- month social-cognitive revolution. From this point on the infant’s face- to-face interactions with others—which appear on the surface to be con- tinuous with her face-to-face interactions from early infancy—are radi- cally transformed. She now knows she is interacting with another inten- tional agent who perceives her and intends things toward her. When the infant did not understand that others perceive and intend things toward an outside world, there could be no question of how they perceived and intended things toward me…. infants at this age also become able to monitor adults’ emotional attitudes toward them as well—a kind of so- cial referencing of others’ attitudes to the self. This new understanding of how others feel about me opens up the possibility for the develop- ment of shyness, self-consciousness, and a sense of self-esteem. (pp. 89– 90; emphasis in original)

Development 85

In this passage, Tomasello has captured the dawning of the sense of self as a special sort of object: a person among persons. There is unity here, a sense of being like the other in being mentalized by the other. To reiterate, at this stage, self-awareness (as awareness of others’ aware- ness) is far from being mentalized explicitly on the part of the infant; yet, along with emotional mirroring, it is part of the foundation on which mentalizing develops.

Language Mentalizing and language develop in a bootstrapping fashion: incipient mentalizing capacities evident in joint attention are required for the ac- quisition of language; yet increasingly refined linguistic capacities are required for the development of a full-fledged representational theory of mind:

young children’s entry into the world of possibility is the crucial route into other minds. Moreover, it is a route that opens up only as their lan- guage abilities develop…. language mediates our experience of the world—how it can be used to re-present events so that what is communicated need not match what is (or was) in the world. As chil- dren acquire language, they acquire the ability to think about possible, re-presented scenarios, and they become able to imagine what other people might think, want, or feel. A new world opens up for them as they realize that people’s minds are private from one another and sepa- rate from the real world. (Astington and Filippova 2005, p. 210, empha- sis in original)

Here we summarize this bootstrapping developmental process and then note how compromised mentalizing capacities, as are glaringly evi- dent in autism-spectrum disorders, result in compromised language use.

Tomasello (1999) argues that linguistic reference can be understood only in the context of joint attentional scenes, which he defines as “social interactions in which the child and the adult are jointly attending to some third thing, and to one another’s attention to that third thing, for some reasonably extended length of time” (p. 97). Furthermore, these joint attentional scenes are defined intentionally: “they gain their iden- tity and coherence from the child’s and the adult’s understandings of ‘what we are doing’ in terms of the goal-directed activities in which we are engaged” (p. 98). In addition, children must be able to engage in role reversal within these joint attentional scenes: they are spoken to, and they speak. Consistent with the dependency of language acquisition on joint attention, children develop a large vocabulary if their mother is adept at following into their attention (i.e., talking about objects they

86 Mentalizing in Clinical Practice

are already paying attention to), and children who spend more time en- gaged in joint attentional activities with their mother subsequently de- velop a larger vocabulary.

Not only are other minds the route to language but also, as Asting- ton and Filippova (2005) contend, language is “a route into other minds,” because it “serves two purposes: communication and represen- tation” (p. 211). More specifically, “human language is used both as an intraindividual representational system and as an interindividual com- munication system” (p. 212). The link between communication and rep- resentation is evident empirically in relations between language and mentalizing: children begin developing the capacity to mentalize ex- plicitly when they learn to employ words for mental states, and individ- ual differences among children in their inclination to refer to mental states (e.g., in conversations with siblings and friends) are predictive of their subsequent performance in theory-of-mind tasks (Brown et al. 1996; Hughes and Dunn 1998).

Both semantic and syntactical aspects of linguistic competence have been implicated in explicitly comprehending false beliefs (Astington and Filippova 2005; de Villers 2000). Children must learn to distinguish between “the way things are in the world and the way things are repre- sented in the mind” (Tager-Flusberg 2000, p. 131); to do so, they must comprehend the attitude of the person whose actions they are interpret- ing. Mental-state verbs, such as “think,” “know,” “guess,” and “forget,” serve to qualify the attitude of the person being interpreted (e.g., knows versus forgets) with respect to a given proposition (e.g., the cookies are in the cupboard). Thus, the semantics and syntax of language facilitate such reasoning about mental states—and the ability to juggle knowl- edge of mental states and reality when the two are in conflict, as they are in false-belief tasks (Astington and Filippova 2005).

Fully competent use of language requires mentalizing capacity, and persons with autism who have developed some capacity for language illustrate this connection (Hobson 2002). Tager-Flusberg (2000) has noted impairments in pragmatic use of language across the entire au- tism spectrum. Examples of characteristic impairments include pro- noun reversals (e.g., “I” versus “you”), failure to attend to providing new information in conversation, difficulty maintaining an ongoing topic of discourse, difficulty interpreting figurative speech, and failure to appreciate jokes or sarcasm. Persons with autism also have difficulty with narrative discourse, being unable to organize a sequence of events that coheres with respect to the motives, beliefs, and emotions of the characters. As Tager-Flusberg (2000) summarized:

Development 87

pragmatic impairments in autism are found across different discourse contexts. These impairments include: a narrower range of functions served by language; problems understanding that communication is about intended rather than surface meaning; failure to view conversa- tions as a means of modifying and extending the cognitive environment of a conversational partner; and failure to view narratives as a means for communicating about both events and psychological states. What is striking about these impairments in communication is that they occur to some degree across the entire spectrum of autistic disorder. Across all ages, ability levels, and language levels, deficits are found in some or all of these aspects of pragmatics and communication. They are even con- sidered to be one component of a broader autism phenotype, found among some proportion of first degree relatives with autism. (p. 127)

Yet mentalizing-related deficiencies in linguistic communication are by no means restricted to autism. On whatever basis, a failure to men- talize—for example, to hold the other person’s mind in mind so as to provide needed contextual information—is likely to be associated with such pragmatic impairments of language (Sperber and Wilson 2002). Clinically, the thorough entanglement of mentalizing with language is evident in alexithymia (Krystal 1988; Taylor and Bagby 2004)—literally, not having words for feelings:

It is hard to deny…that a shallow or impoverished vocabulary for emo- tional self-description makes for a shallow emotional life; and, con- versely, that richer conceptual resources make for correspondingly en- riched possibilities of emotional response. A person whose conceptual universe of the emotions is limited to the two possibilities of feeling good and feeling not-so-good will certainly fail to be subject to (and not just fail to notice) the range of responses possible for some other person with the emotional vocabulary of Henry James. (Moran 2001, pp. 40–41)

Henry James, we therapists are not; yet some of our best work will be helping our patients find the right words for their feelings and en- riching their feelings in the process.

Pedagogical Interactions From the point of view of the infant’s development, we might construe contingently responsive mentalizing on the part of the caregiver as im- plicit teaching regarding the infant’s state of mind—and emotional states in particular. This teaching and learning process is crucial for the devel- opment of the subjective sense of self and emotion regulation (Fonagy et al. 2007). More specifically, Gergely and colleagues (Csibra and Gergely 2005; Gergely 2007; Gergely and Csibra 2005) have explicated

88 Mentalizing in Clinical Practice

the development of mentalizing as one facet of a broader “human- specific adaptation for ‘pedagogy,’ a communicative system of mutual design specialized for the fast and efficient transfer of new and relevant cultural knowledge from knowledgeable to ignorant conspecifics” (Gergely and Csibra 2005, p. 463). Knowledge about minds is a major component of cultural knowledge and, central to our concerns, peda- gogical interactions include teaching about minds, including the mind of the learner.

This uniquely human capacity to teach and to learn cultural infor- mation is extraordinarily efficient in comparison to trial-and-error or observational learning. Observational learning is relatively efficient for naturalistic purposes; for example, the infant can learn to get an object in a box by means of taking off the cover through first observing some- one else do it (where the other person is not intending to provide in- struction). Yet pedagogical learning is essential for acquiring arbitrary cultural knowledge (e.g., learning how a gadget like a spoon is to be used and, of critical importance, learning a language of arbitrary sym- bols). Paralleling the point we just made regarding the relation of joint attention and mentalizing to language acquisition, the capacity to relate on the basis of the pedagogical stance develops earlier than language and full-fledged mentalizing and, indeed, facilitates their development. Yet, in another bootstrapping process, once acquired, language and mental- izing enhance the capacity for pedagogical learning.

The infant’s capacity to take the pedagogical stance does not emerge de novo but rather depends on a set of innate proclivities, including a preference for eye contact and the characteristic intonation of infant- directed speech (“motherese”). These innate behavioral tendencies pre- pare the infant to be receptive to ostensive communication cues that signal the teacher’s intent to convey information to the infant; such cues in- clude establishing eye contact, raising the eyebrows, widening the eyes, tilting the head forward, and addressing the infant by name. The osten- sive cues trigger in the infant a receptive attentional and interpretive at- titude, namely, the pedagogical stance (Gergely 2007)—as if the infant were to say, “I get it; you want to teach me something. OK, I’m ready, go ahead!”

Once the pedagogical stance is established, the communication must specify the referent of the knowledge to be transmitted (e.g., an object that is novel to the infant) as well as the relevance of the information to the infant (i.e., as comprising new knowledge for the infant). Thus, for the teacher, pedagogy entails mentalizing (knowing what the infant knows) and metacognition in particular (the capacity to represent her own knowledge in a form that is accessible to the infant). Concomitantly,

Development 89

the pedagogical stance presupposes benevolent intent on the part of the teacher, that is, a helpful intent to convey accurate and useful informa- tion. As Gergely observes, this pedagogical effort is costly to the teacher, who must mobilize the needed attentional resources, cognitive effort, and attachment motivation. In short, pedagogy requires a benevolent, infant-minded caregiver (Gergely 2007)—attributes that characterize a se- cure attachment relationship. In the course of secure attachment, infants develop a receptivity to teaching that Gergely and colleagues character- ize as epistemic trust, namely, confidence in the caregiver as a reliable source of information. Insecure attachment, by contrast, carries the risk of inadequate or misleading pedagogical interactions along with faulty learning—including faulty learning about one’s own state of mind (Fon- agy et al. 2007). Then epistemic mistrust is liable to develop.

In Gergely’s (2007) view, all the interactive phenomena we have been discussing as foundational for mentalizing exemplify pedagogical communication, including joint attention, pointing for the purpose of drawing the other’s attention to an object, imitative learning, and—not least—language. Crucially for present purposes, emotional mirroring exemplifies all the features of pedagogy: emotionally attuned caregiv- ers intuitively teach infants about the infants’ subjective emotional states. This pedagogical process of teaching infants about their own emotional states, albeit intuitive, is extremely complex. Consider, for example, a mother interacting with her infant son. The mother must employ ostensive cues to gain her infant’s attention, while making it clear to her infant that he is the referent of her communication. The “markedness” of the mother’s expression, as we discussed it earlier in this chapter, is an essential cue that her emotional expression refers to him and not to her. The infant, then, employs her communication to de- rive information about himself—a mentalized representation of his emotional state that he can employ in the service of self-representation and can integrate with his own internal (e.g., somatic) cues. Thus he mentalizes his emotional state by connecting his bodily and muscular sensations with a mental representation derived from his mother ’s marked emotional response. This developmental process of beginning to mentalize emotion extends mentalizing to include internal states of the self; in effect, the infant is being socialized to pay attention to him- self and to his internal world. Concomitantly, the infant develops an in- creasingly solid subjective sense of self along with the capacity for emo- tion regulation (Fonagy et al. 2007).

Gergely and colleagues (2007) also applied their pedagogical ac- count to the phenomenon of social referencing, which we discussed in the context of joint attention. Infants employ social referencing, for

90 Mentalizing in Clinical Practice

example, to adopt attitudes toward unfamiliar objects (e.g., checking their mother’s emotional expression to see if a novel toy is safe to play with). Gergely and colleagues presented experimental evidence that a communicator’s (e.g., mother’s) ostensive cues prompt the infant to in- terpret the interaction as a teaching event such that they are informed about the quality of the object, not the mental state of the teacher. This is a minimalist account of mentalizing activity in social referencing in the sense that the infant need not fathom the mother’s mental state as such but rather must gather information about the nature of the object. Yet such pedagogical learning entails a default assumption that teachers are conveying accurate and universally shared knowledge about the world. Mentalizing comes into play when children come to learn that not all others are bearers of accurate or trustworthy information; this lesson requires understanding that all others are not entirely like- minded but rather have separate minds with different mental contents (e.g., knowledge). Thus, as Fonagy and colleagues (2007) put it, in com- ing to mentalize, children must learn not only that others have minds but also that they have separate minds.

Prementalizing Modes Shedding further light on the development of mentalizing, Fonagy and Target (Fonagy 1995, 2006; Target and Fonagy 1996) distinguished two modes of experience that ultimately become integrated in the mentaliz- ing mode, namely, psychic equivalence and pretend. When mentalizing collapses, individuals are liable to revert to these prementalizing modes, which also include the teleological mode. We will be referring to these four modes of experience throughout this book, and we sum- marize them in Table 3–4.

In the psychic-equivalence mode, the very young child equates the in- ternal world with the external world: the world is how the mind repre- sents it; alternative perspectives cannot be generated. Equating appear- ance and reality, children under four years of age do not know that a sponge shaped like a rock is a sponge despite looking like a rock (Flavell et al. 1987). Accordingly, very young children are potentially beset by fears associated with their imagination: if they think there is a monster in the closet, a monster is in the closet (world=mind). They cannot un- derstand that mental phenomena such as thinking, knowing, imagin- ing, and dreaming are generated by the brain and mind (Estes et al. 1989). Our second author gives an example of his son’s distress in the psychic equivalence mode:

Development 91

Some years ago, at about age four, the second author’s son asked him to bring a Batman costume from one of his trips abroad. Wanting to please his son, and with considerable effort, the author found a Batman cos- tume in a shop that sold theatrical costumes. On his return, his son tried it on, looked at himself in the mirror, cried, and demanded that it should be taken off and put away immediately. He proceeded to put on an old skirt of his mother’s around his shoulders and ran around happily pre- tending to be Batman. In the expensive costume he appeared to be Bat- man and therefore was Batman. This is psychic equivalence.

The pretend mode liberates children from psychic equivalence (Fon- agy 2006): in play, with the decoupling of internal and external reality (Leslie 1987), children retain awareness that their experience does not mirror the outer world (Dias and Harris 1990). Yet inner and outer must be kept separate:

The second author’s son, aged two and a half years, was playing that an upside down chair was a tank and that the legs were shooting ammuni- tion. He was asked: ‘Is this a chair or a tank?’ He stopped playing imme- diately, put the chair the right way up, and walked away. He knew that the object was a chair and not a tank. Yet, in the pretend mode, bringing external reality into contact with the play undercut imagination.

TABLE 3–4. MODES OF EXPERIENCE

Mode Description

Psychic equivalence World=mind; mental representations are not distinguished from the external reality that they represent, such that mental states are experienced as real, as in dreams, flashbacks, and paranoid delusions.

Pretend Mental states are separated from reality but maintain a sense of unreality inasmuch as they are not linked to or anchored in reality.

Teleological Mental states such as needs and emotions are expressed in action; only actions and their tangible effects—not words—count.

Mentalized Actions are understood in conjunction with mental states (as contrasted to the teleological mode), and mental states have neither an exaggerated sense of reality nor unreality but rather are appreciated as representing multiple perspectives on reality (as contrasted with the psychic equivalence and pretend modes).

92 Mentalizing in Clinical Practice

Neither the pretend mode nor the psychic-equivalence mode can cre- ate the optimal relation of the mind to external reality, albeit for opposite reasons: psychic equivalence is too real, whereas pretend is too unreal. In normal development the child integrates these two modes in coming to mentalize: mental states represent reality (unlike pretense) but are not equated with reality (unlike psychic equivalence). The essence of the representational mind is to be capable of adopting multiple perspectives on the same interpersonal situation, thus linking—but not binding—the mind to reality. As Perner (1991) described, the mind represents reality as being a certain way; father is seen as driving recklessly or as being in a hurry or as being frustrated. Moreover, full development of mentaliz- ing skills includes an awareness of this representational relation of mind to reality—other perspectives are possible. Linking representational mind and world, mentalizing enables the recognition—implicitly or ex- plicitly—that others’ actions are understandable given their mental states, that is, their particular take on a situation. Concurrently, mental- izing entails awareness—implicit or explicit—that others interpret one- self in this way. This concurrent recognition affirms one’s existence as a mental agent, representing the full flowering of the intersubjective expe- rience of the self as the object in joint attention that Tomasello (1999) identified in the nine-month social-cognitive revolution.

As we will discuss later in this chapter in connection with attach- ment trauma, integrating psychic equivalence and pretend in the men- talizing mode does not render them obsolete; older children and adults are capable of regressing into these two modes. Moreover, we all regress into the psychic-equivalence mode—world=mind—on a daily basis when we dream. Furthermore, the teleological mode always remains a third prementalizing regressive possibility in which mental states like desires and affects must be expressed in action (Fonagy 2006): unable to mentalize his emotion, the patient with borderline personality disorder cuts himself and displays his bleeding arms to express in action what he cannot express in words.

Clinical Implications To explicate the implicit: psychotherapy, a decidedly pedagogical en- deavor, potentially replicates the developmental processes we have out- lined in this section to help patients function more consistently in the mentalizing mode. Psychotherapists’ empathy will entail highly but im- perfectly contingent marked emotional responding. The psychotherapy process can go awry to the extent that the psychotherapist expresses in- tense and unmarked (real) emotion, abetting contagion, or expresses

Development 93

emotion incongruent with the patient’s emotion, abetting internaliza- tion of alien experience. These emotional interactions take place in the context of joint attention, wherein the patient is the object of the thera- pist’s attention (and vice versa); the implicit and explicit emotional commentary ideally promotes emotional knowing of the self in relation- ship. To be effective as an agent of change, however, this interchange must take place in the mentalizing mode: flexibly linked to reality with- out becoming too intensely real (psychic equivalence) or too detached (pretend, as in intellectualizing or using psychological jargon).

As we will continue to reiterate, psychotherapy is best construed as providing developmental help (Hurry 1998). Furthermore, as our em- phasis on high-but-imperfect contingency implies, the psychothera- pist’s interventions should be only slightly discordant with the patient’s experience. In the context of child education, Vygotsky (1978) provided a model for experience-near interventions in his concept of the zone of proximal development, namely, “the distance between the actual developmen- tal level as determined by independent problem solving and the level of poten- tial development as determined through problem solving under adult guidance or in collaboration with more capable peers” (p. 86; emphasis in original). In moving patients forward, we must be just slightly ahead of them, pro- viding the necessary scaffolding through our own mentalizing.

Attachment and Mentalizing If mentalizing capacity develops through relationships, the quality of those relationships will be pivotal. This section proceeds as follows: First, we describe the complex interplay between parental mentalizing, secure attachment in the child, and the child’s mentalizing capacities. Next, we examine the relations between mentalizing and attachment in the context of intergenerational transmission processes. Then we con- sider the impact of attachment trauma on mentalizing capacity, and we also include some qualifications regarding our general thesis that the benevolent relationship climate of secure attachment is optimal for mentalizing. We conclude the section with comments regarding the clinical implications of all these findings.

Secure Attachment Research has demonstrated a complex developmental interplay among parental mentalizing of the infant and child, child mentalizing capacity, and attachment security in infancy and childhood. Extensive research suggests that secure attachment is conducive to children’s mentalizing ca-

94 Mentalizing in Clinical Practice

pacities, but contingently responsive parental mentalizing within the attach- ment relationship plays the key role in facilitating the child’s mentalizing.

Meins and colleagues’ longitudinal studies have contributed sub- stantially to understanding the developmental trajectories of mentaliz- ing in the context of attachment relationships, and her research pro- gram will serve to anchor our review of the broader literature. Meins (1997) proposed the concept of maternal mind-mindedness to refer to the mother’s “recognition of her child as a mental agent, and her proclivity to employ mental state terms in her speech” (p. 127). Meins and col- leagues (2001) measured mind-mindedness in mothers’ interactions with their six-month-old infants in a play situation, employing an index directly reflecting explicit mentalizing: mind-related comments on the in- fant’s activities that were appropriate to the infant’s behavior at that moment (i.e., contingently responsive); examples include, “Do you rec- ognize that?” “Are you thinking?” “You’re just teasing me!” Such mind- related comments reflect “the mother’s proclivity to use language to frame the interaction in a mentalistic context” (p. 641) and thus indicate the mother’s inclination to relate to the infant on the basis of her own mental representations of the infant’s mental states. In this ground- breaking study, the assessment of maternal mind-mindedness in the interaction at six months predicted the infant’s security of attachment measured subsequently at 12 months. This finding makes intuitive sense: secure attachment is measured by the extent to which the infant confidently reaches out to the mother for comforting after a period of separation in a laboratory separation-reunion paradigm (Ainsworth et al. 1978); when distressed, an infant is more likely to reach out for com- fort to a mind-minded caregiver.

Lundy (2003) extended Meins and colleagues’ findings by assessing mind-minded comments in mothers and fathers in relation to interac- tional synchrony as well as subsequent security of attachment. Synchro- nous parent-infant interactions are conceptualized as reciprocal and mutually rewarding, and exchanges were considered synchronous in this study if they involved at least three contingent steps between par- ents and infants. Replicating Meins and colleagues’ results, Lundy found that both maternal and paternal mind-mindedness predicted attachment security; moreover, she found that mind-mindedness was associated with interactional synchrony for both parents and that syn- chrony mediated the relationship between mind-mindedness and subsequent attachment.

In earlier research, Meins and colleagues (Meins 1997; Meins et al. 1998) had found that security of attachment at 12 months, in turn, pre- dicted subsequent theory-of-mind performance on a range of tasks. For

Development 95

example, whereas 83% of children with previous secure attachment passed a false-belief test at age four, only 33% of those with insecure at- tachment did so. A subsequent longitudinal study (Meins et al. 2002), however, failed to replicate this earlier finding, showing instead that maternal mind-mindedness at six months of age but not attachment se- curity per se at one year predicted theory-of-mind performance at four years. The authors concluded that “the relation between attachment and ToM [theory of mind] can be explained best in terms of individual differences in mothers’ mind-mindedness” (p. 1723).

An extension of this research program (Meins et al. 2003) showed that maternal mind-mindedness at six months also predicted children’s performance on a stream-of-consciousness task at 55 months, a task that assessed their inclination to attribute active thought processes to an ex- perimenter in various states of activity (i.e., sitting quietly and waiting, looking at posters on a wall, and solving a puzzle). Children were asked: “What about her mind right now? Is she having some thoughts or ideas, or is her mind empty of thoughts or ideas?” Thus children who had mind-minded mothers at six months were more inclined to attribute thinking to another person at 55 months. Notably, these researchers also assessed mothers’ mentalizing at the child’s age of 48 months by asking mothers to describe their child and determining the extent to which their descriptions referred to aspects of the child’s mental life (i.e., as op- posed to behavioral, physical, or very general descriptions). Mothers’ six-month mind-mindedness regarding the interactions related strongly to their mentalistic descriptions at 48 months, showing continuity in maternal mentalizing. Yet only the six-month measure and not the more concurrent 48-month measure related to children’s performance on the theory-of-mind and stream-of-consciousness tasks, suggesting that gen- eral descriptions of the child are less predictive of the child’s develop- ment than contingently responsive mentalizing interactions.

Meins and colleagues (2003) concluded that “mothers’ appropriate comments on their infants’ mental states might provide a linguistic and conceptual scaffold within which infants can begin to understand how mental states determine behavior” (p. 1208). Plainly, these interactions are occurring before the infant has developed the capacity to use lan- guage, much less the capacity to mentalize; hence the findings suggest that the mothers’ proclivity to mentalize and to draw the infant into contingent interactions provides an interactive foundation for the later development of mentalizing:

Appropriate mind-related comments at 6 months index the beginnings of mothers’ and infants’ joint attention to mind, in which mothers’ ap-

96 Mentalizing in Clinical Practice

propriate labeling of their infants’ current mental states helps draw in- fants’ attention to the existence (and perhaps functional significance) of mental states and processes. In due course, and perhaps at a slightly later stage in development, this kind of interaction presents infants with opportunities to integrate subjective information on their mental states with an external linguistic comment on the behavior that results from these states. (Meins et al. 2003, p. 1208)

As Meins and colleagues’ research attests, there is nothing magical about the relation between secure attachment and mentalizing: secure attachment constitutes a relationship climate conducive to mentalizing interactions; conversely, as described earlier in this chapter, contingent mentalizing responsiveness is conductive to affect regulation that ce- ments the secure emotional bond, as Meins and colleagues’ (2001) find- ings also suggest.

Additional research is noteworthy in part for illustrating the range of assessments of parental mentalizing conducive to secure attachment and child mentalizing. Oppenheim and colleagues (Koren-Karie et al. 2002; Oppenheim and Koren-Karie 2002), for example, videotaped par- ents and children interacting in various contexts and then interviewed the parents regarding their perceptions of their children’s thoughts and feelings as well as their own. This assessment can be done for infants (e.g., in free play) as well as for preschool children (e.g., when building a house together). These researchers focused on maternal insightfulness regarding the motives underlying their child’s behavior in the interac- tion. Insightfulness encompasses understanding and acceptance of the child’s motives; an emotionally complex and undistorted view of the child, including positive and negative features; and openness to unex- pected behavior that leads to continual updating of her view of the child. This research showed that mothers who were classified as posi- tively insightful regarding their interactions with their child were most likely to have securely attached children. Notably, insightfulness accounted for variability in attachment above and beyond maternal sensitivity, that is, appropriateness of responsiveness to the child in the interaction (Koren-Karie et al. 2002).

As we have indicated, there is a complex interplay between parental mentalizing of the child, secure attachment, and the child’s subsequent development of mentalizing capacity. Although Meins and colleagues’ research showed inconsistent relationships between security of attach- ment per se and the child’s mentalizing capacities, Fonagy and col- leagues (1997a, 1997b) found both concurrent and predictive relation- ships between attachment security and mentalizing capacities. Their research program assessed security of attachment in three- to six-year-

Development 97

old children with projective test stimuli depicting various separation scenarios to which the children told stories in response to specific prompts. The children also were administered a belief-desire reasoning task that required appreciation of false beliefs and disappointment. Not only did concurrent attachment security as measured by the projective test correlate with successful performance on the belief-desire reason- ing task but also this successful child mentalizing was predicted by mother-infant and father-infant attachment security.

Sharp, Fonagy, and colleagues have carried the study of parent- child correspondence in mentalizing forward into later childhood (Sharp et al. 2006). As they point out, it is advantageous to study older children inasmuch as they can talk about what is on their mind. Em- ploying 7- to 11-year-old children, the study used cartoon scenarios in- volving distressing peer-related scenarios (e.g., a child sitting alone on a playground). Children were asked to mentalize by imagining what the cartoon subject’s peers were thinking, and their mothers were asked to mentalize by guessing their child’s responses—mentalizing their child’s mentalizing. Inaccurate maternal mentalizing was associated with inaccurate mentalizing in children, namely, children’s unrealisti- cally positive views of what other children might be thinking, an at- tributional bias that has been shown to relate to disturbances in peer relationships (Sharp et al. 2007). Moreover, inaccurate maternal mental- izing was associated with high levels of psychopathology in children, as measured by self-report and parent report.

Gottman and colleagues (1996) have studied what we would call op- timal mentalizing of emotion in parents of older children. Specifically, they studied what they call parental “meta-emotion philosophy” and its relation to child functioning. By meta-emotion, they refer to attitudes toward emotion—what we would call a mentalizing stance toward emotions. Specifically, meta-emotion philosophy bears on “parents’ feelings and thoughts about their own and their children’s emotions, their responses to their child’s emotions, and their reasoning about these responses” (p. 245). In researching parental meta-emotion, they observed that some parents adopted an emotion-coaching strategy with five components:

parents (a) said that they were aware of low intensity emotions in them- selves and in their children; (b) viewed the child’s negative emotion as an opportunity for intimacy or teaching; (c) validated their child’s emo- tion; (d) assisted the child in verbally labeling the child’s emotions; and (e) problem solved with the child, setting behavioral limits, and discuss- ing goals and strategies for dealing with the situation that led to the neg- ative emotion. (p. 244)

98 Mentalizing in Clinical Practice

As we would construe it, parents with an emotion-coaching philoso- phy are mentalizing emotion in themselves and their children as well as actively promoting mentalizing emotion in their children, although the researchers did not specifically assess children’s mentalizing. In a longi- tudinal study of children beginning at age five and ending at age eight, Gottman and colleagues (1996) found that parental meta-emotion and emotion coaching were associated with children’s lowered physiological arousal and greater emotion-regulation capacities as well as with better functional outcomes as measured by academic achievement and peer re- lations. Moreover, a subsequent study (Katz and Gottman 1997) showed that parents’ meta-emotion philosophy contributed to resilience in their children, specifically, being one factor that buffered children against the adverse effects of marital discord, separation, and divorce.

We have reviewed research on parent-child interactions in detail be- cause this research provides a straightforward model for clinical inter- ventions. In closing this section, however, we mention some additional studies to underscore the sheer breadth of empirical support for the re- lation between attachment and mentalizing. Precocious mentalizing in childhood as reflected in theory-of-mind task performance has been as- sociated with more reflective parenting practices (Ruffman et al. 1999), the quality of parental control (Cutting and Dunn 1999; Vinden 2001), parental discourse about emotions (Denham et al. 1994; Meins et al. 2002), the depth of parental discussion involving affect (Dunn et al. 1991), and parents’ beliefs about parenting (Baumrind 1991; Vinden 2001). We have argued that such effective parenting is likely to promote the child’s acquisition of a coherent conceptual apparatus for under- standing behavior in mentalistic terms. It is not surprising, for example, that parents whose disciplinary strategies focus on mental states (e.g., a victim’s feelings or the non-deliberate nature of transgressions) have children who learn quickly to appreciate the importance of mental states (Sabbagh and Callanan 1998). Furthermore, we would expect tol- erance for negative affect to be characteristic of secure attachment and, concomitantly, a family environment that facilitates mentalizing. Ac- cordingly, family-wide talk about negative emotions, often precipitated by the child’s own emotions, has been shown to predict later success on tests of emotion understanding (Dunn and Brown 2001); this finding is consistent with the fact that reflecting on intense emotion without being overwhelmed is a marker of secure attachment (Sroufe 1996). Similarly, the number of references to thoughts and beliefs and the relationship specificity of children’s real-life accounts of negative emotions relates to precocious mentalizing (Hughes and Dunn 2002).

Development 99

Intergenerational Transmission of Attachment To take a step back: What are the origins of parental mind-mindedness? Fonagy and colleagues’ research program on the means by which at- tachment security is mediated by mentalizing (Fonagy et al. 1991a, 1991b, 1994; Steele et al. 1996) extended a productive line of investiga- tion of the intergenerational transmission of attachment (Fraiberg et al. 1975; Main 1991; Main et al. 1985). While they were pregnant with their first child, mothers were interviewed about their own attachment his- tory with the Adult Attachment Interview (Main and Goldwyn 1994). Fonagy and colleagues’ initial study of 96 mother-infant pairs revealed that the mother’s security of attachment in relation to her parents as- sessed prior to the birth of her infant predicted her infant’s security of at- tachment with her at 12 months (Fonagy et al. 1991a). Fonagy and col- leagues subsequently found the same pattern for father-infant pairs (Steele et al. 1996), and these findings have been replicated consistently by other research teams (Benoit and Parker 1994; van IJzendoorn 1995; Ward and Carlson 1995).

Fonagy and colleagues proposed that attachment security is trans- mitted in part by the securely attached parent’s mentalizing capacity:

we assume that the quality of the infant’s attachment to the parent is in- trinsically linked to two factors both present and measurable before the birth of the child: (1) the parent’s internal working model; and (2) the parent’s capacity to reflect on the current mental state of the child and to reflect on, and exert control over, his or her own expectations of rela- tionships as these influence his or her behavior toward the child…. The child is likely to be securely attached if either the parent’s internal model of relationships is benign, dominated by favorable experiences, or…the parent’s reflective function [mentalizing] is of sufficient quality to fore- stall the activation of working models based on adverse experiences in- appropriate to the current state of the relationship of child and care- giver. (Fonagy et al. 1995, p. 258)

Slade and colleagues (Slade 2005; Slade et al. 2005) further validated Fonagy and colleagues’ (1995) model of intergenerational transmission by assessing not only mothers’ security of attachment and mentalizing capacity with respect to their own attachment history but also the moth- ers’ ability to mentalize in relation to their infant (at 10 months); they too assessed the infants’ subsequent security of attachment (at 14 months). These researchers assessed mothers’ mentalizing of their infant with a 90-minute Parental Development Interview. The interview covers such matters as the mother’s perception of her infant and the infant’s feel- ings; her thoughts and feelings about separations from her infant; and

100 Mentalizing in Clinical Practice

her view of herself as a parent as well as her view of the way her expe- riences with her own parents influenced her parenting. Mentalizing was assessed on the basis of the mother’s awareness of mental states, her rec- ognition of developmental aspects of mental states, and her awareness of mental states in relation to the interviewer.

As previous research had done, Slade and colleagues’ study con- firmed a strong relationship between the mother’s security of attach- ment and her mentalizing capacity in relation to that attachment. In ad- dition, the mother’s security of attachment in relation to her parents predicted her ability to mentalize in relation to her child. Moreover, ma- ternal mentalizing capacity in relation to the child was predictive of the child’s attachment security. Finally, the researchers found suggestive evidence that the link between the security of the mother’s adult attach- ment measured in pregnancy and her infant’s security of attachment at 14 months was mediated by the mother’s mentalizing capacity. As Slade (2005) asserted, “the centrality of the parent as mediator, reflector, interpreter, and moderator of the child’s mind cannot be overempha- sized” (p. 273). More specifically:

A mother’s capacity to hold in her own mind a representation of her child as having feelings, desires, and intentions allows the child to dis- cover his own internal experience via his mother’s representation of it; this re-presentation takes place in different ways at different stages of the child’s development and of the mother-child interaction. It is the mother’s observations of the moment to moment changes in the child’s mental state, and her representation of these first in gesture and action, and later in words and play, that is at the heart of sensitive caregiving, and is crucial to the child’s ultimately developing mentalizing capacities of his own. (p. 271)

Importantly, Slade and colleagues’ research demonstrated that the mother’s predominantly mentalizing stance toward her infant, consis- tent with her ability to mentalize in relation to her own attachment history, promoted attachment security in her infant. Yet attachment de- velops in relation to behavioral interactions with the infant. Thus Arnott and Meins (2007) took this line of investigation yet another step. As others had done, they assessed (prenatally) the mother’s and fa- ther’s security of attachment in relation to their own parents as well as their capacity to mentalize about their early attachment experience. In addition, as Meins and colleagues had done in their previous research, they independently assessed each parent’s mind-minded comments to their six-month-old infant in a free play situation, and then they did a follow-up assessment of their 12-month-old infant’s security of attach-

Development 101

ment to each of them. Thus they were able to put all the components of the developmental pathway together.

Although there were some differences between mothers and fathers, the following general pattern held for both parents: security of parental attachment (to their parents) was positively related to mentalizing ca- pacity (regarding those attachments); both security of attachment and mentalizing capacity were related to mind-minded comments to the in- fant at six months; and mind-minded comments were predictive of the infant’s security of attachment at 12 months. Two general developmen- tal pathways for both parents were identified: parental secure attach- ment was linked to infant secure attachment through high parental mind-mindedness, and parental insecure attachment was linked to in- fant insecure attachment through low parental mind-mindedness. Thus the combination of parental secure attachment and mind-mindedness was a powerful predictor of infant attachment for both parents.

To summarize this line of research we have reviewed, the capacity to mentalize in the attachment relationship is part and parcel of secure attach- ment. As Slade and colleagues’ work demonstrates, parents who demon- strate mentalizing capacity in relation to their own childhood attachments are thereby more likely to develop rich, accurate, and flexible mental rep- resentations of their infant; as Meins and colleagues’ work shows, this mentalizing capacity is manifested in mind-minded, contingently respon- sive interactions with their infants. In turn, paralleling their parents, the in- fants develop secure attachment as well as mentalizing capacities. This broad intergenerational trajectory is summarized in Figure 3–1. Fonagy (2006) has articulated the import of this developmental process:

As we currently formulate it, the mother’s secure attachment history permits and enhances her capacity to explore her own mind and pro- motes a similar enquiring stance towards the mental state of the new hu- man being who has just joined her social world. This stance of open, re- spectful exploration makes use of her awareness of her own mental state to understand her infant—but not to a point where her understanding would obscure a genuine awareness of her child as an autonomous being. The awareness of the infant, in turn, reduces the frequency of behaviours that would undermine the infant’s natural progression towards evolving its own sense of mental self through the dialectic of interactions with the mother. (p. 70)

Attachment Trauma Just as secure attachment relationships are the crucible for optimal men- talizing, trauma in attachment relationships is inimical to mentalizing (Fonagy et al. 2007). Attachment trauma carries a double meaning, refer-

102 Mentalizing in Clinical Practice

ring to trauma inflicted in an attachment relationship as well as to the concomitant undermining of the capacity for secure attachment (Allen 2001). Abusive and emotionally neglectful behavior plainly reflects an absence of empathizing in the caregiver; inflicting trauma exemplifies mindblindness (Allen 2007). Consequently, whereas parental mentaliz- ing is conducive to secure attachment, childhood maltreatment engen- ders the most profound form of attachment insecurity, namely, disorga- nized attachment (Main and Solomon 1990; van IJzendoorn et al. 1999). The disorganized infant is not able to develop any consistent strategy for relating to the attachment figure; accordingly, Main (1999) character- ized disorganized attachment as fright without solution.

As Fonagy and Target (1997a, 2002) have proposed, childhood mal- treatment undermines the development of mentalizing capacity and thus poses a dual liability: the maltreatment evokes distress and simul- taneously undermines the development of capacities for affect reg- ulation. More specifically, attachment trauma promotes a defensive withdrawal from the mental world—at worst, a phobic avoidance of mentalizing (Fonagy et al. 2007). Awareness of the mind of the abuser is terrifying to the child, because “he will be confronted with attitudes toward himself which are extremely painful to recognize: hatred, cruelty, indifference” (Fonagy and Target 1997a, p. 693). This self-protective re- treat from the mental world undermines the child’s reflective capacities, for example, as shown by poor performance on theory-of-mind tasks, diminished capacity for pretend play, and the relative absence of lan- guage referring to internal states. Impaired mentalizing, in turn, under- mines coping with an abusive relationship in a vicious circle:

FIGURE 3–1. INTERGENERATIONAL TRANSMISSION OF SECURE ATTACHMENT AND MENTALIZING.

parental security of attachment parental mentalizing capacity

mind-minded interactions with infant

infant secure attachment (comfort seeking)

enhanced mentalizing capacity in childhood

Development 103

Poor comprehension of mental states associated with maltreatment am- plifies distress, activating the attachment system. The need for proxim- ity persists and perhaps even increases as a consequence of the distress caused by abuse. Mental proximity becomes unbearably painful, and the need for closeness is expressed at a physical level. Thus, the child may paradoxically be driven physically closer to the abuser. The child’s ability to adapt to, modify, or avoid the perpetrator’s behavior is likely to be constrained by limited mentalizing skills, and exposure to further abuse is likely to occur. The paradox of proximity seeking at the physical level concurrent with psychological avoidance lies at the root of the dis- organized attachment consistently seen in abused children. (Fonagy et al. 2000, p. 111)

Main and Hesse (1990) proposed that frightening or frightened behav- ior on the part of the parent is the key contributor to disorganized at- tachment; such behavior might include looming or invading the child’s space, playing frightening games, timidity, or parental dissociative states. Lyons-Ruth and colleague (Lyons-Ruth and Jacobvitz 1999) identified two patterns of maternal behavior associated with attach- ment disorganization: hostile intrusiveness and helpless withdrawal, both of which reflect misattunement and aversion to the child’s attachment behavior along with an inability to empathize with the child—a failure to mentalize in the context of attachment-related interactions.

Building on Lyons-Ruth’s approach, Slade and colleagues’ (Grien- enberger et al. 2005; Slade 2005; Slade et al. 2005) research program also considered the adverse side of attachment. These researchers were es- pecially interested in “a mother’s capacity to regulate her child’s affect at times of heightened arousal,” and they hypothesized that “her be- havior at times of distress and negative affect (rather than global sensi- tivity, per se) will be most crucial in determining the child’s attachment security” (Grienenberger et al. 2005, p. 300). Non-mentalizing maternal behavior included affective communication errors (e.g., laughing when the infant is crying), role or boundary confusions (e.g., demanding a show of affection), fearful behavior (e.g., as evident in a squeaky voice), intrusive behavior (e.g., pulling the infant by the wrist), and with- drawal (e.g., not acknowledging the infant after a separation). The re- search showed that these non-mentalizing interactions were associated with high levels of infant insecurity, especially as manifested in disor- ganized attachment. Notably, paralleling secure attachment, an inter- generational transmission process also has been demonstrated for dis- organized attachment. Parental disorganized attachment is associated with atypical maternal behavior and with disorganized infant attach- ment (Goldberg et al. 2003). Moreover, preliminary evidence that the ca- pacity for change in attachment organization decreases over develop-

104 Mentalizing in Clinical Practice

ment underscores the danger that persistent trauma will lead to long- term disorganization of attachment, with attendant poor development of social cognition and substantially raised risks of psychopathology (Kobak et al. 2006).

Slade and colleagues’ research buttresses the conceptual links be- tween trauma and impaired mentalizing (Fonagy and Target 1997a). On the one hand, as these researchers proposed, “mothers who were able to openly reflect on their children’s affect and intentions would be better equipped to handle infant vulnerability without becoming over- whelmed by their own unintegrated fear or hostility.” On the other hand, “Parents who are lacking in RF [reflective functioning] may become eas- ily dysregulated or disorganized by their infant’s distress as they fail to distinguish between their own feelings and those of their children” (Grienenberger et al. 2005, p. 308). If we take egocentrism—attributing one’s own mental states to others—to be the default non-mentalizing mode (Decety 2005), we can see how attachment trauma exemplifies its worst instantiations. Owing to their own trauma history, parents are lia- ble to relate to their distressed child, in effect, in the psychic-equivalence mode. As Slade (2005) summarized:

Disturbed and abusive parents obliterate their children’s experience with their own rage, hatred, fear, and malevolence. The child (and his mental states) is not seen for who he is, but in light of the parents’ pro- jections and distortions. The infant then takes on the parent’s hatred and aggression, a primitive form of identification with the aggressor. (p. 273)

Consistent with the foregoing review, a wealth of research links mal- treatment to developmental deficits that we would anticipate to be as- sociated with impaired mentalizing. Maltreated children’s narratives show disturbances in parental representations as well as self-represen- tations (Macfie et al. 2001). Specifically, abused children’s capacity to generate complex representations of the parent in conflict-imbued set- tings decreases with development; similarly, their self-representations become increasingly simplified and exaggerated (Toth et al. 2000). Mal- treated children are impaired not only in cognitive aspects of mentaliz- ing, such as theory-of-mind tasks (Cicchetti et al. 2003; Pears and Fishler 2005) but also in emotion-focused mentalizing (Pears and Fishler 2005; Rogosch et al. 1995; Smith and Walden 1999). For example, as Fonagy and colleagues’ (2007) review of research indicates, maltreated children are less likely to respond empathically to other children’s distress; they show more emotionally dysregulated behavior; they talk about internal and emotional states less often; and they have difficulty understanding

Development 105

emotional expressions. Hence, in construing emotionally modulated conversations as the royal road to mentalizing, Fonagy and colleagues draw an analogy from the effects of maltreatment to those of sensory deficits: the learning environment is not only frightening but also im- poverished, lacking the open and reflective communication about men- tal states that most nourishes mentalizing.

In sum, given all we now know about the adverse effects of mal- treatment on the development of mentalizing, we might view the exten- sive empirical evidence linking secure attachment to mentalizing from two perspectives: secure attachment not only entails the presence of mentalizing-promoting behaviors (e.g., mind-mindedness) but also the absence of mentalizing-undermining behaviors (e.g., hostile intrusive- ness and neglect). In short, mentalizing begets mentalizing and, con- versely, non-mentalizing begets non-mentalizing. Hence the intergen- erational transmission process for secure attachment (as depicted in Figure 3–1) is mirrored by an adverse intergenerational transmission process for insecure attachment as depicted in Figure 3–2.

Qualifications In the preceding review, we have documented a well-researched global correspondence between accurate mentalizing and secure attachment, on the one hand, and mentalizing failures and insecure attachment, on the other hand. Yet we also have noted that mentalizing is highly con- text-dependent, fluctuating within a given attachment relationship de- pending on the nature of the interaction at the moment. As we will dis-

FIGURE 3–2. INTERGENERATIONAL TRANSMISSION OF INSECURE ATTACHMENT AND IMPAIRED MENTALIZING.

106 Mentalizing in Clinical Practice

cuss further in Chapter 4 (“Neurobiology”), recent neuroimaging findings have led to some refinements in our understanding of the rela- tionship of attachment to mentalizing. Quite obviously, dynamic fluctu- ations in the emotional quality of the secure attachment relationship will determine the extent to which mentalizing is active at any given point in time (Fonagy 2006; Fonagy et al. 2007). There are three condi- tions under which mentalizing will be inhibited or inactive in the over- all context of a secure attachment relationship. First, activating loving feelings, be they those of a mother for her baby or those in a romantic relationship, sometimes inhibits mentalizing—as will be unsurprising to readers whose passion occasionally has overridden their judgment of romantic partners. Second, threat-related activation of attachment, such as might occur at times of separation, can undermine mentalizing by virtue of evoking overwhelming affect that is uncharacteristic of the re- lationship generally. Third, a stable secure attachment relationship can obviate the need to mentalize, for example, insofar as there is no need to contend with threatening or stressful interactions or to make judg- ments of trustworthiness.

On the other hand, there are three broad reasons for secure attach- ment being conducive to mentalizing. First, inevitable challenges, con- flicts, and fears in the relationship will prompt mentalizing, and the emotional containment the relationship provides will modulate the emotional arousal to a level conducive to mentalizing. Second, while not necessitating mentalizing, a secure attachment relationship pro- motes positive emotional arousal conducive to an interest in mentaliz- ing. Third and most crucial: as we have been emphasizing, in a secure attachment relationship the attachment figure is frequently mentalizing the child, and mentalizing is conducive to further mentalizing. In effect, as the “pedagogical stance” implies, the secure attachment relationship is the ideal practice ground for mentalizing.

While we generally associate mentalizing with benevolence, and the link between secure attachment and mentalizing is consistent with this association, there are important exceptions. As we have just implied, one reason to mentalize is social threat: when relationships are poten- tially exploitative or injurious, mentalizing serves not only to detect the threat but also to facilitate self-protective interactions and negotiations. Thus we learn to mentalize in a competitive-threatened mode as well as in a securely attached, exploratory mode. Hence insecure attachment can promote mentalizing of negative emotions (Harris 1999). Of course, mentalizing in the competitive-threatened mode can be compromised when affect escalates; then we see distortions characteristic of mental- izing in a downright paranoid mode. Thus we have a double-edged

Development 107

sword: social competition and threat, which likely fueled the evolution of mentalizing capacities and associated brain structures (see Chapter 4, “Neurobiology”), can either enhance or undermine mentalizing. On the other hand, partly by alleviating competitive pressures, secure at- tachment can promote the fullest flowering of mentalizing capacities (Fonagy et al. 2007).

The flip side of this competitive-threatened-paranoid relationship paradigm is also noteworthy: mentalizing skills can be employed to threaten, exploit, or harm others (Fonagy 2006; Fonagy et al. 2007; Sharp 2006). Thus, developmentally, mentalizing skills can be associated with malicious teasing, bullying, deception, and relational aggression (e.g., in girls who elicit intimacy and disclosure and then manipulate the re- lationship by threatening to expose secrets). Notoriously, psychopathy can be associated with exceptional mindreading skill employed for ma- nipulative, exploitative, and malevolent purposes.

All these instances of antagonistic social behavior reflect skillful but partial mentalizing; something is missing, namely, empathy (Sharp 2006). In Baron-Cohen’s (2003, 2005) terms, for example, psychopathy entails mindreading without empathy, that is, mentalizing without an appropriate emotional response to the mental states being mentalized accurately (see Chapter 4, “Neurobiology”). We consider mindreading without empathy to be impoverished mentalizing in the sense of a fail- ure of emotional knowing (or knowing through emotion). Thus the lack of appropriate emotional responsiveness reflects a profoundly significant lacuna in otherwise intact mentalizing.

Finally, while we have argued that mentalizing is the developmental wellspring of secure attachment and vice versa, mentalizing is em- ployed, practiced, and refined in the full range of interpersonal relation- ships throughout life. As Dunn’s (1996) research attests, the practice of mentalizing beyond relationships with caregivers begins in childhood in interactions with siblings and peers. Nonetheless, the parallels with secure attachment are noteworthy:

The quality of the child-child relationship was important here: it was during playful cooperative interactions that the children discussed mental processes, and the length of their friendship and frequency of their interaction were positively related to explicit reference to mental processes. (Dunn 1996, p. 512; emphasis added)

Notably, despite mothers’ frequent use of mental state language with their children, four-year-old children use mental state terms more often in conversing with siblings and friends than with their mother, of- ten in the context of pretend play (Brown et al. 1996). Moreover, chil-

108 Mentalizing in Clinical Practice

dren’s performance on theory-of-mind tasks and emotion understand- ing is correlated with the extent of their conversations about mental states in dyadic play (Hughes and Dunn 1998). Thus discourse about emotional states plays a central role in the development of mentalizing capacities (Harris 1999), and such discourse occurs in a wide range of relationships. We hardly strain common sense in extrapolating: what is characteristic of childhood holds throughout life, namely, that positive relationships of all sorts promote mentalizing and vice versa. Of course, antagonistic, competitive, and insecure relationships also can prompt mentalizing when they are short of overwhelmingly threatening.

Clinical Implications As we described in Chapter 1 (“Introduction”), Bowlby (1988) recog- nized that the secure base of attachment not only promotes exploration of the outer world but also enables the exploration of the world of the mind. Research reviewed in this chapter has begun to clarify the means by which psychotherapists provide this much needed secure base.

Ideally, mind-minded therapists keep up a mind-related commentary appropriate to the patient’s current state (Meins et al. 2001), engaging in contingently responsive mentalizing interactions. Ideally, therapists are insightful in the sense of understanding and accepting patients’ motiva- tion, developing an emotionally complex view of patients that sub- sumes positive and negative features, while remaining open to contin- ually updating their views on the basis of new information without introducing distortions based on preconceptions (Oppenheim and Koren-Karie 2002). Similarly, therapists construct autobiographical nar- ratives of their interactions with patients that predominantly reflect their patients’ intentional nature (Slade et al. 2005). Conversely, hostile intrusiveness and helpless withdrawal on the part of therapists will un- dermine their patients’ mentalizing capacities (Lyons-Ruth and Jacob- vitz 1999). Yet we must not convey that a continuously warm and be- nevolent climate will best promote mentalizing in psychotherapy, any more than it would have done in childhood development. Developing greater ability to mentalize in the context of challenge and threats to at- tachment security requires that these emotional states be evoked—and mentalized—in the psychotherapy process.

Two aspects of these clinical implications bear underscoring. First we focus primarily on the patient’s current mental state rather than ex- ploring past mental states. Such a focus on the present might appear to be non-contingent when the patients are considering how they “were” rather than how they “are,” but the mentalizing therapist moves fluidly

Development 109

between these states, considering the current motivation for exploring the past as well as the impact of exploring past states on the current mental state. Second, the “insightful” therapist promoting mentalizing is not seeking causal explanations (i.e., explaining the present on the ba- sis of the past) but rather is continually striving to stimulate a coherent and emotionally meaningful narrative for both patient and therapist. Moreover, as we just noted, this narrative is best developed in the midst of emotional states, positive and negative.

Naturally, a therapist’s ability to hew to these ideals will vary from patient to patient and from time to time with any given patient (Dia- mond et al. 2003)—just as is true of mothering, fathering, or any other relationship. Moreover, we have focused relatively exclusively on the impact of the attachment figure’s mentalizing on the child’s mental- izing, just as we emphasize the impact of the psychotherapist’s mental- izing on the patient’s mentalizing. But we should keep in mind that mentalizing is always a two-way street. Fonagy and colleagues (2007) refer to the often neglected child-to-parent effect; for example, that non- mentalizing behavior in children can evoke controlling behavior in par- ents. The same applies to patient-to-therapist effects. Of course, as the ideal parent would do, we psychotherapists should aspire to mentalize in the face of non-mentalizing behavior.

Recapitulation We introduced this chapter by proposing that, in contrast to concepts such as empathy, mindfulness, metacognition, psychological minded- ness, and the like, mentalizing is uniquely grounded in developmental research. We recognize that all clinicians necessarily mentalize and as- pire to promote mentalizing in their psychotherapeutic work, whether or not they have heard the word and whether or not they are privy to mentalizing’s developmental foundations and trajectory. But we have argued that owing to this developmental grounding, knowledge of mentalizing promises to refine our clinical work: understanding devel- opment clarifies the relationship conditions that promote and under- mine mentalizing and enables us to provide needed developmental help—ideally, sooner rather than later.

Recall the extensive developmental ground we have covered in this chapter. We have characterized the development of the self in terms of increasingly sophisticated agency: infants begin life as physical and social agents, exerting influence over their body, physical objects, and other persons; they develop teleological understanding when they perceive

110 Mentalizing in Clinical Practice

agents as being engaged in rational goal-directed actions. During the second year of life, when infants begin mentalizing the teleological stance, they become able to interpret goal-directed actions as based on mental states such as desires and intentions. By the close of their third year of life, having passed the litmus test of passing false-belief tasks, children fall into line with modern cognitive scientists in employing a representational theory of mind. The representational view of the self continues developing throughout life into a rich inner world associated with an increasingly complex autobiographical narrative, ideally ac- companied by a correspondingly rich understanding of other persons. As we have contended, all this mental development goes from the out- side in: caregivers’ contingently responsive mentalizing in the form of emotional mirroring, joint attention, and pedagogical interactions insti- gate and scaffold this development of the self. But the mentalizing mode of functioning, once achieved, is not consistently maintained; psychopa- thology associated with attachment trauma, for example, entails regres- sion to prementalizing modes: psychic equivalence, pretend, and the action-oriented teleological mode.

Following in Bowlby’s footsteps, we have conveyed one overriding message: to promote mentalizing, psychotherapy must recapitulate op- timal development by creating a climate of secure attachment. Yet this message would be relatively uninstructive without all the developmen- tal research we have described. Decades ago, Bowlby’s collaborator, Mary Ainsworth (Ainsworth et al. 1978), rightly linked secure attach- ment to sensitive responsiveness. But we have learned since that the core of sensitively responsive caregiving is mentalizing—and mentalizing emo- tion in particular. Hobson characterized variations in this process well:

Suppose an infant is in a particular state of mind and is needing an ap- propriate response from a caregiver. She might be excited and joyful, deeply upset, or screaming with rage. The contrast is between an adult who has the mental space to pick up and be sensitive to the infant’s state, so that the infant feels responded-to and somehow encompassed within the adult’s attentive care, and an adult who either cannot perceive or cannot take on board what the baby is feeling. (Hobson 2002, p. 125)

Owing to extensive clinical observations and research, we now know a great deal more about how the development of mentalizing flourishes and, sadly, with research on attachment trauma, we also know how it founders. Thus we are in an increasingly strong position to provide— however belatedly—needed developmental help that will foster a stron- ger sense of self-agency, greater capacities for emotion regulation and impulse control, and an enhanced sense of connection with others.

Development 111

Bowlby (1982) maintained that secure attachment is essential for us all throughout life—“from the cradle to the grave” (p. 208). Many of our patients struggle to attain or maintain secure attachment relation- ships, and an important function of psychotherapy is to assist them in those efforts. As the research reviewed in this chapter attests, we psy- chotherapists assist them in the ways that infant-minded parents do: by mentalizing and promoting mentalizing in a benevolent yet chal- lenging relationship. Employing attention and imagination, we must develop accurate and rich mental representations, always open to novel elaborations while remaining mindful of skewing preconceptions and countertransference distortions. Psychotherapy is indeed a pedagogical relationship insofar as it is characterized by contingently responsive, implicit and explicit mentalizing interactions in which our patients can make use of our representational capacity to enhance their sense of self; inherent in this process is enhanced agency stemming from the process of knowing oneself and making oneself known.

112 Mentalizing in Clinical Practice

Key Clinical Points • Conditions that promote and undermine mentalizing: Attachment

research illuminates the relationship conditions that promote and undermine the development of mentalizing. Contingent, mind- oriented responding promotes mentalizing in psychotherapy as well as in development; the mind-minded mother is a model. Conversely, non-mentalizing interactions (e.g., intrusiveness and withdrawal) are as inimical to psychotherapy as they are to development.

• Role of contingent emotional responsiveness: As it does in development, contingent emotional responsiveness in psycho- therapy provides an external representation of emotional states; this process can go wrong in two ways. First, if the therapist’s emo- tional responses are too intensely real (unmarked), the patient’s emotional state will escalate. Second, if the therapist’s responses are out of sync with the patient’s emotions (non-contingent), a number of things can happen: at best, patients who are unable to correct the misunderstanding will ignore the intervention; at worst, they will internalize the therapist’s misrepresentation. In the face of persistently non-contingent responses, patients may cease to look to the therapist as someone who is able accurately to re-present their experience.

• Prementalizing modes: When mentalizing is lost, three pre- mentalizing modes of subjectivity can come to the fore: psychic equivalence (equating reality with the current mental state, as in concrete thinking), pretend (not relating mental states to reality, as in dissociation), and teleological (expressing mental states through action or somatic reactions). Evidence of prementalizing modes of functioning in therapy is an indication that the therapist’s primary task is not to explore, elaborate, explain, or correct the patient’s pro- ductions but rather to work with the patient to recover mentalizing.