The Adult Attachment Projective

David Joubert

This chapter reviews the development, validation and use of the Adult Attachment

Projective (AAP; George, West, & Pettem, 1997). The AAP is a projective instrument

designed to provide a valid assessment of adult attachment states of mind, as

conceptualized by Bowlby (1969). The AAP consists of a set of eight pictures depicting

situations that activate the attachment behavioral system, such as loss of an attachment

figure or physical threat to the self. The individual is asked to make up a story based on

the pictures, and specific probes are used to insure that essential elements are included in

the story. Research evidence to date provides strong support for the reliability and

validity of the AAP. Results of studies looking at inter-rater reliability and construct

validity have been especially impressive. The use of the AAP in psychological

assessment is illustrated using an actual clinical case.

 2

Description of the System

Conceptual framework

Attachment theory (Bowlby, 1969, 1973, 1980) is concerned about the ways that

early relationships with primary caregivers contribute to the development of one’s

personality, adaptation, and emotional well-being. Attachment is conceptualized as a bio-

behavioral system designed to promote proximity with the caregivers in times of stress, in

such a way as to optimize emotional regulation (Sroufe, 1988). Research over the past

three decades has provided strong support for Bowlby’s original hypotheses. Attachment

has been linked to a wide variety of outcomes and appears to influence functioning in

many areas, including cognitive, social and emotional development, in both adults and

children (Main, 1995; Thompson, 1999). The link between internal representations of

attachment figures and relationships and various forms of psychopathology in adults is

well-established (see review by Dozier, Stovall, & Albus, 1999). In addition, several

papers have been published in recent years which describe clinical applications of

attachment theory with a range of populations and problems (Eagle, 1997; Fonagy et al.,

1996).

One area of interest for attachment-oriented researchers and clinicians concerns

the assessment of individual differences relative to Bowlby’s (1969,1973) concept of

“internal working models” of attachment relationships. Because caregivers respond

differently to infant’s expressed needs for proximity, attachment researchers have

theorized that the caregiver’s response patterns become integrated into the child’s

developing psyche as a representational “template” for predicting and interpreting

parental behavior in response to this child’s attachment needs. Bowlby (1973, 1988)
 3

proposed that this template would serve as an integral part of the individual’s personality,

and determine the person’s typical response to stressful situations and relationships.

Ainsworth’s (Ainsworth, Blehar, Waters, & Wall, 1978) seminal work with

children using the Strange Situation procedure documents the presence of specific

patterns of child attachment behavior and caregiver’s response in mother-infant dyads.

Secure children demonstrate a capacity to express needs for proximity with caregivers in

a clear and modulated way. Upon reunion following separation with the attachment

figure, the child is soothed and can return to exploration and play within a short period of

time. Cross-sectional and longitudinal research has shown that secure children generally

show higher levels of functioning or development compared with insecure children

(Grossmann, Grossmann, & Waters, 2005; Sroufe, Egeland, Carlson, & Collins, 2005;

Thompson, 1999). Their behavior in the Strange Situation is thought to reflect an

internalization of caregivers as available and responsive (Sroufe & Fleeson, 1986).

Insecure children, on the other hand, exhibit behaviors that suggest defensive

interference in attachment-related situations (Bretherton, 2005; George & West, 1999;

Solomon, George & De Jong, 1995). Insecure-avoidant children attempt to limit their

interactions with the caregiver as a way to avoid experiencing distress related to the

attachment figure’s unwillingness or inability to respond to their signals for proximity.

Insecure-ambivalent children, on the other hand, both desire and resist proximity and

soothing from the attachment figure. Paradoxically, insecure-ambivalent children can

express much distress during separation with the parent. It is the incapacity to be soothed

or comforted by the caregiver that distinguishes this group.

 4

Another attachment configuration in children, historically identified later by Main

& Solomon (1990), is composed of children for whom “organized” attachment patterns

(e.g., secure, avoidant, ambivalent) are undermined throughout the Strange Situation by

behavioral disorganization, dysregulation and conflict, such as freezing,

approach/avoidant conflict behavior, withdrawal, failure to recognize the attachment

figure, and trance-like states. These signs of attachment disorganization often appear

especially during the moments of reunion following separation from the attachment

figure. This attachment group is termed disorganized. The disorganized group has been

found to be significantly at risk for negative outcomes in a variety of developmental

areas, including cognitive, interpersonal, emotional, and physical. Not surprisingly,

disorganized infants and children are over-represented in clinical populations (see review

by Lyons-Ruth & Jacobvitz, 1999).

While research looking at attachment in children has been progressing steadily for

several decades, interest for attachment-related processes in adults is relatively new by

comparison. Bowlby (1969) and Ainsworth (Ainsworth, Blehar, Waters, & Wall, 1978)

believed that an individual’s attachment history is an important building block for

personality development, and therefore remains relevant “from the cradle to the grave”

(Bowlby, 1969). According to attachment theorists, early interactions with caregivers

become internalized as representations or models of relationships (“internalized working

models”), which can then be used to predict and understand social interactions, especially

those involving close emotional bonds or intimacy. Such representational models tend to

remain stable over the person’s life, although significant evidence exists that such models

can and do change over time (Fonagy, 1999; Waters, Merrick, Treboux, Crowell, &
 5

Abersheim, 2000). A thorough assessment of attachment can be of considerable research

or clinical interests because it provides the professional an empirically established

conceptual lens to better understand the meaning of interpersonal behavior in terms of

attachment processes (e.g., proximity-seeking) related to the individual’s past experiences

with caregivers. One implication of attachment theory, consistent with other theoretical

formulations (e.g., object-relations theory), is that such models will determine the

individual’s behavior in current relationships with significant others, thus accounting for

existence of phenomena such as transference in therapeutic work or partner selection and

processes in adult relationships.

Part of the challenge of examining internal models of attachment in adults for

clinical or research purposes lies in selecting a form of assessment that is appropriate for

the construct. A construct-appropriate assessment of attachment in adult and adolescent

populations must consider the following points : 1) attachment in adults consists of a set

of internal representations, which exist in a more or less integrated form in the

individual’s psyche, and 2) attachment-related internal processes (i.e., thoughts, fantasies,

affects, defensive exclusion) are more likely to be accessible when the individual’s

attachment behavioral system is activated. Such considerations clearly make certain

forms of assessment more relevant than others.

The search for a reliable and valid method for assessing attachment in adults has

been the object of considerable attention in the literature for the past two decades. Work

has focused primarily in two areas: 1) the assessment of attachment-related cognitions

and self-reported perceptions using paper-pencil questionnaires, and 2) the assessment of

 6

representational models of attachment relationships using unstructured or semi-structured

methods.

The use of self-report questionnaires to assess attachment has stemmed from the

work of social psychologists such as Hazan & Shaver (1987) and Bartholomew

(Bartholomew & Horowitz, 1991) looking at the construct of “romantic attachment”. A

large number of such questionnaires now exist in the literature, and recent evidence

suggests that these self-report measures tap into similar romantic attachment constructs

(see Crowell, Fraley, & Shaver, 1999). That is, scores obtained on those instruments

appear to reflect conscious perceptions of self and others, relative experiences of anxiety

in interpersonal relationships, and management of distress through behavioral avoidance.

Unfortunately, self-report questionnaires tend to correlate better with other self-report

data than with other outcomes (e.g., observations, peer or parent report, physiological

markers), which makes their validity questionable for the purpose of assessing a construct

such as adult attachment, whose foundations lie in the individual’s internal

representations of early relationships (Waters, Crowell, Elliott, Corcoran, & Treboux,

2002). Further, these measures were designed following Ainsworth’s original infant

attachment model, poorly assess disorganized attachment, and do not adequately integrate

theory or research in attachment on adults (George & West, 1999; West & George,

1999).

A more fruitful approach to the assessment of adult attachment, as conceptualized

by Bowlby and others, is represented through the use of narrative and semi-structured

techniques employed to access the individual’s representational world. Among these

techniques, the Adult Attachment Interview (AAI; George, Kaplan, & Main, 1985) is
 7

considered the “gold standard” for the assessment of attachment in adults. The AAI is a

semi-structured interview covering various aspects of the individual’s relationships with

his or her primary caregivers. The interview is designed to assess current representational

structures and states of mind relative to attachment, by eliciting autobiographical

memories concerning attachment-related experiences. Information obtained during the

interview is not considered “factual”; rather, the main focus of interest lies in the

individual’s capacity to use autobiographical memory to “inform” current representations

of attachment relationships. The classification system focuses on coherence of discourse

when the individual discusses attachment with attachment figures.

The AAI uses a classification system that is designed to assess conceptually

categories as developed by Ainsworth (1978) and Main (Main & Solomon, 1990) in their

research with infant-mother dyads. Secure-Autonomous individuals are characterized by a

coherent discourse and complex, well-integrated representations of attachment figures.

This is illustrated, for example, by their capacity to present balanced, reality-based

descriptions of primary caregivers, and by the presence of metacognitive activity

throughout their discourse. Individuals classified as Dismissing tend to present with

difficulties providing descriptions of attachment figures that are informed by

autobiographical memory. They frequently report an incapacity to recall specific

attachment-related events, tend to minimize the importance of early attachment

relationships, or provide unrealistic, idealized descriptions of caregivers. Preoccupied

individuals appear to be highly ambivalent about attachment relationships. Their

descriptions of relationships with primary caregivers lack coherence and are poorly

integrated. They may experience confusion or appear enmeshed with past relationships
 8

with caregivers. Last, individuals classified as Unresolved with respect to loss or trauma

show manifestations of behavioral, affective or cognitive dysregulation when discussing

relationships with early attachment figures. Such manifestations might include, for

example, disorientation or signs of dissociative process, discourse manifestations that are

interpreted similarly to the interferences in attachment behavior observed in disorganized

infants.

There has been abundant research using the AAI over the past 25 years, a full

description of which goes beyond the scope of this chapter (see overview by Hesse,

1999). What is evident from using the AAI, however, is that it is costly to use in terms of

time of administration and coding. This contributed to a search for alternative ways of

measuring attachment, leading to the development of numerous self-report

questionnaires, as described earlier. It is also evident that there are other important

attachment-related constructs that the AAI does not measure. Evaluation of idealization

or the inability to recall childhood experience, for example, does not provide specific

information about the defensive exclusion processes that lead to idealized images of

attachment figures or memory interferences. Carol George and Malcolm West have

developed a method that identifies the same basic AAI attachment groups, preserves

attachment theory’s emphasis on representational structures and process-oriented analysis

of discourse, examines attachment-derived operationally defined defensive processes, and

is more user friendly than the AAI.

Projective methods have traditionally played an important role in the assessment

of attachment in children. In 1972, Hansburg described the Separation Anxiety Test

(SAT), which used pictures similar to the TAT to evaluate responses to separation in
 9

older children and adolescents. Subsequently, Klagsbrun & Bowlby (1976) modified the

SAT for use with younger children. This tradition is still very much alive as researchers

have continued to develop projective assessments, including further modifications of the

SAT (e.g., Slough, Goyette, & Greenberg, 1988) and doll-play story-stem methods (e.g.,

Bretherton, Ridgeway, & Cassidy, 1990; Solomon, George & De Jong, 1995), to evaluate

individual differences in attachment representations in children. However, development

of similar methods to evaluate attachment in adults or older adolescents has not been

undertaken until recently.

Stimuli

George, West, & Pettem (1997) used their extensive knowledge of attachment

theory and projective testing to build a coding system that addresses content, process and

linguistic aspects of narratives obtained from exposure to a set of cards visually depicting

attachment-related situations. The Adult Attachment Projective (AAP) includes eight

pictures which are presented to the individual in a pre-determined sequence. The first

picture is a neutral scenario depicting two children playing ball, and is intended as a

“warm-up” for the other cards. The remaining cards describe seven attachment scenes:

Child at Window depicts a child with her back turned to the viewer, looking out of a

window; Departure presents two characters, man and woman, facing each other with

suitcases next to them; Bench depicts a person sitting alone on a bench; Bed shows a

child lying in bed with his arms forward and an adult character sitting at the end of the

bend; Ambulance depicts two characters, a child and an older woman, looking out of

window at an ambulance outside; Cemetery shows a man standing in front of a grave; and
 10

Corner illustrates a child standing in a corner, his head turned with an arm extended

forward.

The AAP cards were specifically designed to elicit attachment-related states of

mind, and for this purpose the level of detail in the background (i.e., the environment

surrounding the characters) was kept to a minimum. In addition, facial features were

omitted from the drawings, in order to heighten the level of ambiguity and facilitate the

expression of “projective” material. The administration sequence activated the attachment

behavioral system in a way that is analogous to the “separation from the caregiver” and

“introduction of a stranger” episodes during the Strange Situation. The visual themes that

compose the AAP set relate to events that are known to elicit attachment, such as

separation, threat, illness, or death.

Administration

The administration is similar to that of the “apperceptive” class of projective

techniques in general, although there is a more “structured” component based on the

presence of specific probes and instructions. Individuals are presented with the cards one

at a time, and are told to create a story based on their perception of what is happening in

the picture, what led to the scene, what might happen next, and what the characters are

thinking or feeling. The interviewer uses probes corresponding to these elements

(present, past, future, character’s internal states) to insure that they are present in the

narrative.

Scoring Categories

The evaluation of attachment states of mind using the AAP focuses on three

primary dimensions: 1) content, 2) discourse, and 3) defensive processes. The content

 11

dimension refers to specific features of the narrative and the characters included in it. The

discourse dimension has to do with the way that the individual constructs the narrative

and attempts to build a coherent plot. The defensive processes dimension refers to the

manner in which attachment-related information may be excluded from the person’s

consciousness. An overview of scoring categories is included in Table 1.

Insert Table 1 about here

________________________________________________________________________

Content. The way that content features are considered varies according to the

nature of the situation depicted on the card. Pictures presenting a character alone are

coded according to the level of Agency attributed to the character. The concept of

“agency” refers to the individual’s capacity to negotiate the activation of the attachment

behavioral system, through summoning internal cognitive resources, receiving help or

comfort from an attachment figure, or taking action. Three levels of agency are thus

possible: 1) Internalized Secure Base (use of internal resources as illustrated through the

use of expressions such as “think about”) or Haven of Safety (use of an attachment

figure), 2) Capacity to Act (use of action), or 3) Absence of Agency.

In addition, pictures depicting characters alone are assessed for the degree of

Connectedness shown in the narrative. This term refers to the degree to which

relationships are desired and established by the character. Stories are examined to

determine if Connectedness is present, blocked, or simply not described (i.e., not an

element in the story).

 12

By contrast, pictures describing characters dyads are scored according to the level

of Synchrony included in the narrative. The concept of synchrony relates to the extent to

which dyadic relationships are reciprocal, goal-corrected attachment-caregiving

relationships, following Bowlby (1969). Three levels are possible: 1) Synchrony

(evidence of a reciprocal, goal-corrected partnership between the characters), 2) Failed

Synchrony (evidence of a more “functional” relationship, not characterized by genuine

reciprocity), and 3) Absence of Synchrony (lack of relationship described between

characters).

Discourse. The evaluation of the characteristics of the individual’s discourse is

based on two variables. The first, Coherency, involves an assessment of the degree to

which each individual narrative is constructed in a logical, consistent, articulated, and

intelligible way, following Main’s (1995) reformulation of Grice’s (1975) rules of

conversational discourse to attachment interview material. This approach evaluates four

criteria, including Quality (how “good”, cohesive or creative the plot is), Quantity

(sufficient level of detail provided without being excessive), Relation (whether the

elements included are relevant to the plot or not), and Manner (the way that the narrative

is told). The evaluations of these specific four variables are used to generate an overall

relative coherency code that designates a story as highly, moderate, or not coherent. In

addition, each story is rated according to whether Personal Experience is present or not.

This variable refers to the use of any element pertaining to one’s own life or experience

in the narrative construction process. Recent work on attachment has shown that the

presence of such material can indicate problems in maintaining appropriate self-other

boundaries (Solomon & George, 1999).

 13

Defensive processes. The assessment of defensive functioning has traditionally

been a central element of projective personality testing (e.g., Schafer, 1954). The concept

of “defense” in attachment theory refers to how people process attachment-related

material while at the same time maintaining psychological equilibrium as well as

preserving existing representational structures (Bowlby, 1980). The purpose of defenses

is, therefore, to terminate or modulate activation of the attachment behavioral system in

order to prevent potentially disequilibrating subjective experiences of distress or

discomfort. Projective testing is an excellent way to observe defensive processes at work

because the ambiguous nature of the stimuli constitutes an implicit threat to the self

(Lerner, 1991). The AAP is designed to assess the ways that individuals process and

exclude attachment-related thoughts, emotions, and needs.

Bowlby (1980) identified three primary attachment-related defensive processes:

1) Deactivation, 2) Cognitive Disconnection, and 3) Segregated Systems. Solomon et

al.’s (1995), and later George & West (2001), have clarified and operationally defined

these defensive exclusion processes based on attachment research in children using

projective doll play stories and the AAP, respectively. Deactivation effectively takes the

individual’s attention away from all attachment-related information, focusing instead on

exploration and mastery. In other words, deactivation implies a form of attachment “shut

down” that excludes certain aspects of distressing attachment information, events, and

emotions before reaching a conscious level. On the AAP, an individual using deactivation

as a defense will often provide stories devoid of themes involving distressing attachment-

related material (e.g., avoid themes of needing help, sadness, anger). Deactivation is

evidenced by themes of personal strength or power, achievement, peer relationships, or

 14

social status. Deactivation is also observed through attempts at minimizing or

normalizing situations, stereotypical statements, or negative evaluations of characters.

Cognitive Disconnection involves splitting attachment-related information from

its source in order to attempt to terminate the activation of the attachment behavioral

system and the discomfort associated with it. Such defensive process, in contrast with

deactivation, usually fail to terminate the attachment behavioral system, as the actual

source of the activation is maintained out of consciousness and, therefore, never really

addressed. Cognitive disconnection manifests as deep-seated uncertainty and

ambivalence about attachment figures and their capacity or willingness to meet

individual’s needs (Solomon et al., 1995; George & Solomon, 1999).One example of this

process is the evocation of anger and aggression as a way to retaliate against perceived

insult or frustration by the attachment figure. Although angry behavior is likely to be

experienced by the person as tension-relieving, a potential solution to psychological

discomfort, and a way to “distance” oneself from the source of the frustration (i.e., the

other person), its function in attachment-related situations is to elicit caregiving behavior

and create proximity with the attachment figure (Bowlby, 1973). Bowlby used the term

“functional” anger to describe appropriate, clearly directed angry signals to attachment

figures in achieving proximity. He differentiates this from “dysfunctional” anger, the

pervasive use of diffuse or augmented anger signals. Dysfunctional anger shifts the

relationship from successful proximity to threatening to push the attachment figures

away, implying a loss of autonomy vis-à-vis the attachment figure and, to some extent, a

loss of self-other boundaries. This representational status does not allow for a resolution

of the underlying issues (e.g., unavailability of the Object). On the AAP, cognitive
 15

disconnection is manifested in narratives involving, for example, uncertainty, anger,

ambivalence, radically opposite interpretations, or glossing over attachment distress.

Segregated Systems most often refer to themes of danger or threat to the self.

Issues of vulnerability, abandonment, physical threat or danger, abuse, disasters or

catastrophic events, or a need to escape a situation (flee, become invisible, hide, run

away) are considered segregated systems markers. Themes of emptiness or isolation also

constitute markers for this type of defense. Any indication of odd, bizarre or peculiar

ideas are coded as segregated systems, along with manifestations of dissociative process,

intrusive material, or extreme constriction. Segregated systems coded in the AAP to

determine (1) their presence and (2) whether or not they are contained or resolved, a

representational quality that suggests at least a minimal degree of integration within the

individual’s psyche.

Forms of resolution of segregated systems typically include the use of internal

resources to manage attachment-related distress, assistance or care from others, use of

action as a problem-solving strategy, or psychologically leaving the location of the

dysregulating attachment material, as indicated by statements such as “…moving on with

his life”. Segregated systems markers that are not contained by these story features

indicate that some traumatic material involving attachment relationships is left

unintegrated, and blocked from consciousness, that is, segregated from the rest of the

psyche. Bowlby (1980) stressed that such material may emerge from the unconscious if

attachment is activated and contribute to behavioral, emotional or cognitive

disorganization, the expression of which may be subtle or readily observable in the

individual’s behavior.
 16

Classification process

Assessment of attachment states of mind using the AAP culminates in a

classification process that evaluates the coding variable patterns of the individual’s

protocol according to four groups conceptually similar to the framework developed by

Ainsworth (1978) and Main (Main & Solomon, 1990) for child and adult attachment,

respectively.

The Secure-autonomous (F) group is characterized by the capacity for internal

representations of attachment to portray integrated Agency; that is, the portrayal of

attachment figures as providing an “internalized secure base” or external comfort is the

primary mechanism for managing stress and conflict. Relationships with attachment

figures are important and satisfactory, as portrayed by successful Connectedness when

attachment distress occurs in contexts when the individual is alone. Secure-Autonomous

people portray attachment relationships in dyadic relationships as Synchronous, that is, as

functioning in a goal-corrected partnership based on sharing and reciprocity. They are

aware of interpersonal boundaries (no Personal Experience). Their AAP protocols are

diverse in terms of patterns of defensive processes; there is a wide range of individual

differences in how Deactivation, Cognitive disconnection, and Segregated Systems are

evident in the narratives. Overall, secure individuals tend to show less evidence of

defensive process than insecure individuals. Their stories tend to be more coherent than

those of insecure individuals, although some secure-autonomous individuals are secure

and “incoherent.” Coherency only evaluates story discourse, communicating the story to

another in a “conversational” context. The sin qua non of security in the AAP is the
 17

individual’s ability to portray flexibly integrated representational attachment

mechanisms, as evidenced in Agency, Connectedness, and Synchrony.

Dismissing (Ds) individuals tend to emphasize Agency as a form of taking action

that does not involve attachment figures as a primary way to manage internal conflict.

This action also includes functional activity (Synchrony), such as interactions or activities

performed together (e.g., go on a vacation, have a meal) that circumvent emotions or

events related to attachment distress. Although reciprocal in nature, functional Synchrony

does not portray comfort or mutual enjoyment as important to the attachment-caregiving

relationship. Dismissing individuals may describe Connectedness to others, but these

story elements often involve peer relationships rather than attachment relationships as the

primary way of meeting needs for social interactions and security. Their narratives also

involve a blend of Deactivation, Cognitive Disconnection and Segregated Systems

markers that are contained. Dismissing individuals are typically moderately coherent or

incoherent. The defining feature of Dismissing individuals is the effective use of

deactivation to filter distress from attachment events and thoughts, a form of defensive

exclusion that appears to be used as a substitute for flexible integration and internalized

security.

Preoccupied (E) individuals generally show an absence of Agency, functional

Synchrony with attachment figures, and lack of Connectedness with others because they

are alone or blocked from relationships. Their protocols are often restricted and include

much uncertainty or contradictory statements. While a range of defensive processes may

be observed in their story narratives, the defining feature of individuals preoccupied with

attachment experiences is the dependence on Cognitive Disconnection to manage

 18

attachment. Their narratives demonstrate the ability to evoke deactivation; however, these

attempts are widely unsuccessful due to the continuous activation of the attachment

system and their sensitivity to attachment distress. Preoccupied individuals tend to

provide stories containing segregated systems markers. Minimal Agency, inconsistent

Connectedness, functional Synchrony, and representations of non-related individuals

(e.g., strangers, neighbors) as comforting prevent attachment representations and the self

from hyperarousal. Cognitive disconnection is the underlying representational mechanism

related to their incapacity to create a coherent narrative, which in many ways parallels

their failed attempts at terminating the attachment behavioral system.

The Unresolved (U) group is defined by a state of vulnerability to dysregulation

due to attachment-related trauma which has not been sufficiently integrated. All

Unresolved individuals also maintain “organized” representations of attachment and

behavioral strategies (Main & Solomon, 1990). The expression of such dysregulation in

the individual’s functioning may depend in part on the primary organized attachment

classification (F, Ds or E). Consequently, there is no prototypical AAP protocol for U

individuals, as they constitute a heterogeneous group. The defining feature of this group

involves the presence of at least one narrative in which a segregated systems marker is

not resolved. Other characteristics of the protocols are a function of their primary

attachment configuration. For example, U individuals with F as a primary attachment

configuration will be more likely to show some evidence of integrated Agency,

Synchrony, or Connectedness, albeit not in every story. For the time being, researchers

and clinicians using the AAP refer to U individuals as a specific attachment profile (i.e., a
 19

“group”) in itself, although future research should be able to tell us whether other ways of

considering this variable may be more helpful.

Research Evidence

Although the AAP is still a relatively new method, recent research has provided

some interesting data supporting its use with a variety of populations. The majority of

studies to date have investigated the instrument’s construct validity or test-retest

reliability.

Validity

The original validation study performed by the original authors (George, West, &

Pettem, 1997) intended to provide some basic data to support the validity and reliability

of the AAP. Because the AAP was developed as a valid, cost-effective alternative to the

AAI, the authors compared attachment classification on both instruments across various

samples, including 25 at-risk mothers, 23 women participating in a large-scale study on

depression, and 48 (42 women, 6 men) individuals recruited in community, university

and clinical settings. In this study, the authors were able to show strong convergence

between AAP and blind-rated AAI classifications, as suggested by values of .92 (=.75,

p<.001) and .85 (=.84, p<.001) for secure-insecure and four-group (F, Ds, E, U)

classifications, respectively.

More recent data on construct validity, as well as discriminant validity

(intellectual level and social desirability), is also available for a sample of 144 adult

subjects (100 women, 44 men ranging from 18-65 years or age) recruited from college

and community settings in California and Alberta, Canada. In this study, the authors

(George & West, in press) essentially replicated their previous findings by obtaining a
 20

97% (=.88, p<.000) agreement for secure-insecure and 90 % (=.84, p<.000) agreement

between four-group AAP and AAI classifications. Discriminant validity was documented

by a lack of association between attachment status and verbal intelligence (as measured

by the Vocabulary and Similarities of the WAIS) and social desirability (Bindra

Inventory of Desirable Responding).

Because attachment theory involves understanding the importance of qualities of

parent-child relationships for the development of the child, establishing an empirical link

between attachment states of mind in the caregiver and salient features of child

functioning may provide support for the validity of the method used to assess adult

attachment. In a series of studies, Moss and colleagues (Béliveau, Cyr, & Moss, 2002;

Cyr, Béliveau, & Moss, 2003; Moss, Cyr, & Dubois-Comtois, 2004) investigated such a

link using the AAP as a measure of adult attachment. More specifically, they examined

whether mothers’ attachment states of mind were valid predictors of child development,

as documented by Strange Situation separation-reunion behavior and behavior problems,

in a normative sample recruited from various daycare centers in a large urban setting. The

authors found that secure-autonomous mothers were more likely to raise children

assessed as secure in the Strange Situation, whereas insecure (Ds, E, or U) mothers were

more likely to have insecure children. They reported an overall match of 73 % between

mother and child attachment status. Moreover, children of dismissing and unresolved

mothers appeared to be more at risk for presenting behavior problems, as compared with

children of secure-autonomous mothers. In an interesting look at potential variables that

may explain this association, Moss and her colleagues found that some factors related to

the mother herself (e.g., stressful life events, feelings of helplessness), as well as factors
 21

related to her actual behavior in her relationships with her child (e.g., discourse,

conversational patterns), contributed to the development of psychosocial vulnerabilities

in the child. Data was also obtained pertinent to the discriminant validity of the AAP, as

attachment classification in this sample was independent of socioeconomic level.

Similar results linking attachment states of mind and caregiving representations

and behavior were obtained by Odipo (2002) with a sample of foster mothers. Foster

mothers classified as secure on the AAP displayed more positive affect and capacity to

develop coherent representations of the children under their care than mothers classified

insecure. Insecure mothers reported more negative affect, feelings of burden and

subjective distress, and less ability to adopt the child’s perspective.

Data supporting the construct validity of the AAP also comes from its use with

clinical or high-risk samples. In a study by the original authors (West & George, 2002),

the AAP protocols of 24 women diagnosed with dysthymia were examined in order to

determine the prevalence and understand the attachment mechanisms that characterized

this severe form of depression. The authors found that preoccupied attachment was over-

represented in this sample (58 % for 4-group and 75 % for 3-group classifications),

suggesting that the inability to deactivate attachment distress and the confusion resulting

from the inability to disconnect elements of attachment experiences contributed to this

disorder.

Mezzich and her colleagues (Mezzich, Heliste, George, Kolar, Swaney, Raspet, &

Day, 2006) used the AAP in their follow-up study designed to examine the role of

attachment in Substance Abuse Disorder and Antisocial Personality Disorder in

adolescent girls. Cross-sectional data using the AAP at age 24-28 during their
 22

longitudinal study focused on Substance Abuse Disorder. They report that Substance

Abuse Disorder was significantly related to insecure attachment, and this relationship was

not mediated by Antisocial Personality Disorder. They interpret their findings to suggest

that poor affect regulation as associated with insecure attachment may be a prominent

mechanisms underlying Substance Abuse Disorder.

A series of studies by Buchheim and her colleagues provide concurrent and

predictive validity for using the AAP in non-English speaking and clinical populations.

Buchheim has pioneered the use of the AAP to examine attachment activation in a

neuroimaging environment, using functional magnetic resonance imaging (fMRI)

technology. Buchheim administered the AAP in Germany to clinical patients (described

more fully below) and established convergent validity for the AAP in both the German

and English translation transcripts. Convergent validity for the AAP (coded by a reliable

English-speaking AAP judge from English translations of German transcripts) was

established using AAI classification as the criterion (determined from German transcripts

by a reliable German AAI judge). Correspondence combined for Buchheim’s studies

were as follows: 84% for four classification groups (=.71, p<.000), 91% for secure-

insecure (=.70, p<.000), and 88% for unresolved-resolved groups (=.75, p<.000)

(George, personal communication, 2006).

The first study (Buchheim, Erk, George, Kächele, Ruchsow, Spitzer, Kircher, &

Walter, 2006) was designed to assess the capacity of the AAP to elicit activity in certain

specific visual and semantic brain regions. The authors were interested in learning how

lack of integration of traumatic events, as indicated by the “Unresolved” status, was

linked to brain activity in a normative sample (11 healthy female participants). They
 23

successfully in demonstrated that the AAP task elicits activity in areas of the brain linked

to visual and semantic processes, a theoretically expected finding. More specifically, the

AAP card administration sequence (i.e., increasing activation of the attachment system)

was associated with activation of the right inferior frontal cortex, a region of the limbic

system responsible for affect regulation. The validity of AAP classifications based on

fMRI administration was checked against AAI classifications based on interviews

administered outside the fMRI environment. There was 100% convergence for the

organized-unresolved groups.

An important piece of information uncovered by the authors is that in unresolved

subjects, gradual exposure to more “traumatic” stimuli through the card sequence was

linked with increased activation of medial temporal regions, including the amygdala and

hippocampus. These structures are thought to be involved in the processing of negative

emotions and the retrieval of autobiographical memory. This data is important in that it

constitutes strong evidence supporting the capacity of the AAP to elicit neural activity

patterns that are consistent with activation of the attachment system. Activation of the

attachment behavioral system is thought to be essential to the assessment of attachment in

both adults and children, and may require a form of stimulus that most commonly used

self-report questionnaires cannot provide (George & West, 1999).

Extending this neuroimaging methodology, Buchheim and colleagues (Buchheim,

Erk, George, Kächele, Martius, Pokorny, Ruchsow, & Walter, 2006) investigated the

neural correlates of attachment patterns in Borderline Personality Disordered patients

compared with normal controls. Because of the high prevalence of unresolved attachment

linked to experiences of abuse in childhood for individuals with Borderline Personality

 24

Disorder (BPD), the authors hypothesized that pictures including traumatic segregated

systems marker material in the AAP narratives (e.g., abandonment, abuse) would be

associated with different neural activation patterns in BPD subjects when compared with

controls. BPD patients showed significantly more traumatic segregated systems language

indicators in response to alone pictures than nonpatient controls. Alone picture fMRI

response scans showed significantly greater anterior midcingulate cortex activation.

Dyadic picture fMRI response scans showed greater right superior temporal sulcus

activation and less right parahippocampal gyrus activation.

In addition to providing further evidence of the validity of the AAP for the

purpose of eliciting attachment-related representations, Buchheim’s work clearly

illustrates difficulties experienced by individuals with BPD in inhibiting their internal

states when exposed to conflictual attachment-related situations. Such a finding is quite

consistent with the existing literature on BPD (e.g., Gunderson, 1996; Stone, 1990), and

may contribute to a better understanding of this disorder from the standpoint of

attachment theory.

Buchheim and George (2006) examined attachment status and traumatic AAP

segregated systems markers in a larger adult patient population : 34 BPD (13 inpatients

from the fMRI study and 21 outpatients), 20 Anxiety patients, and 21 healthy controls.

BPD (74%) and Anxiety patients (80%) were largely unresolved; the majority of healthy

controls were secure and dismissing (62%). These classification distribution patterns are

consistent with the clinical attachment literature (Dozier et al., 1999). Borderline and

anxiety patients showed significantly more trauma dysregulation markers in their

narrative responses to alone pictures than controls. Traumatic segregrated systems

 25

appeared predominantly in personal experience material for the anxiety patients. The

intrusion of attachment dysregulation directly related to self material during the AAP

suggests important attachment related processes related to the blurring of self-other

boundaries in anxiety patients. This pattern appears to be related to anxiety patients’

unresolved loss, not abuse (the primary correlate of BPD traumatic dysregulation).

Recent studies have begun to use the AAP with adolescents. These studies are

beginning to establish predictive validity for the AAP during the teen years, although

more studies are needed to examine concurrent and psychometric validity of the AAP for

teens. Theoretically expected associations were found between attachment status on the

AAP and variables related to psychosocial adjustment in a sample of 13-14 year-old

adolescents (Moss, Thibaudeau, Joubert, Beliveau, & Dubois-Comtois, 2006).

Preoccupied adolescents were more likely to use substances and report more

externalizing problems. In a sample of young mothers, Joubert and colleagues (Joubert,

Mazzarello, Moss, Paquette, & Zoccolillo, 2006) found that mothers in the preoccupied

group were more likely to show a significant history of antisocial behavior and their

children were more likely to have problems with aggression. Working with a sample of

adolescents referred to an outpatient mental health clinic, Webster and Joubert (2006)

reported associations between unresolved status and impairments in thought processes,

self-other boundaries, reality testing and emotional constriction, as assessed using the

Rorschach Inkblot Method.

 26

Generalizability

Studies to date have demonstrated that the AAP can be used to measure

attachment status in a diverse range of individuals. It has been validated for adults (age

18-65), predictive validity has been established for adolescents, and it has been used in

normal and clinical populations. Studies described in this review have included samples

originating from the U.S., Canada (including narratives in French), and Germany.

Because of the focus placed on language and narrative process in the AAP system, the

use of this instrument across groups whose primary language is not English might be

problematic. However, results obtained with French- and German-speaking samples

suggest that the AAP assessment system can be used with non-English-speaking

individuals. Studies using the AAP with populations in Italy, Russia, Japan, and Taiwan

are in progress.

The AAP assessment process requires that an individual be able to construct

narratives based on the visual depiction of certain attachment-related situations. Because

preschool-age children typically demonstrate a capacity to construct coherent narratives

based on visual stimuli, educational level requirements for the AAP are fairly low. This is

indirectly supported by research showing no association between AAP classification and

verbal intelligence (George & West, in press).

Reliability

Data aggregated from the original validation samples (George, West, & Pettem,

1997) shows strong evidence of inter-rater reliability, as indicated by values of .93

(=.73, p<.001) for secure-insecure classification and .86 ( =.79, p<.001) for the

standard four-group classification.

 27

George and West’s (in press) replication study obtained similar results. Four-

group inter-rater reliability was excellent, the values reported between 2 pairs of raters

being 90% (=.85, p<.000) and 85% (=.79, p<.000), respectively. Inter-rater agreement

values reported in various other studies using the AAP are generally in the .80 to .90

range.

Temporal stability was also investigated in this study, as attachment is presumed

to be a fairly stable component of one’s personality (Bowlby, 1969). The authors reported

test-retest reliability estimates over a 3-months period of 84% (=.78, p<.000) for the

four-group classification. Stability estimates for specific attachment groups included 82%

for Secure-Autonomous, 96% for Dismissing, 62% for Preoccupied, and 80% for

Unresolved.

It must be pointed out that the values reported above were obtained from research

using reliable, certified coders. Utilization of the AAP requires that coders complete a

training and reliability trial under the supervision of Carol George. It is recommended

that potential coders have a graduate degree in the social sciences and be familiar with

human development, particularly attachment theory as developed by Bowlby, as well as

with the fields of personality, cognition, emotion, and basic psychometrics.

Case Illustration

This section presents an example pertaining to the use of AAP data in a clinical

setting. For this purpose, the AAP protocol of a man referred for individual

psychotherapy is reviewed, along with background information and psychological test

data. The discussion focuses on the way that information obtained from the AAP protocol
 28

can enhance the clinician’s understanding of the individual’s dynamics and personality,

and can by extension inform decision-making by adding attachment information.

Background Information

Mr. L is a 40 year-old, single, Caucasian man who was referred by his treating

psychiatrist to an outpatient mental health clinic for individual psychotherapy. At the

time of the assessment, Mr. L had been clinically depressed for the last 6 months

following the loss of a relative. He was hospitalized for a brief period due to his reporting

suicidal ideation. Mr. L was administered the WAIS-R, MMPI-2, Early Memories

Procedure, Rorschach and Adult Attachment Projective as part of his assessment.

Mr. L is the seventh in a family of 12 children, including six brothers and five

sisters. He grew up on a farm in a rural area. He initially described his childhood as a

“happy” period, without significant problems. Mr. L’s father is described as a soft-spoken

man who used to “hide feelings” and generally kept to himself. His mother is described

as “special”, “nice” and responsible for disciplining the children. Mr. L described himself

as very close to his mother. He reported liking school at the time, and got along well with

teachers and peers. His relationship with his siblings is also described as “good”, without

any major conflicts.

According to Mr. L, adolescence was a difficult time for him. His father partially

lost the use of his hands due to exposure to extreme cold. His father then became severely

depressed and withdrew from family life, isolating himself in his room. This isolation

was punctuated by several suicide attempts. The father finally committed suicide by

hanging when Mr. L was 12. The client reported being traumatized by finding his father

hanging from the ceiling of the barn. He reportedly started skipping school and avoiding
 29

friends following this incident. Mr. L can recall no involvement in romantic relationships

during adolescence.

The client left the family home when he was 17 and moved to the city. He had

several jobs for short periods of time, but felt that he could not stay long in any one place.

He started at that time to frequent gay bars, where he would meet older men who would

promise him financial or material support in order to develop a romantic relationship with

him. Such relationships were typically conflictual and short-lived. Mr. L eventually met a

young woman with whom he started a relationship, which lasted 7 years. However, this

girlfriend was prone to episodes of depression and was often suicidal. Mr. L reported

feeling confused at the time because his girlfriend, who was bisexual, would bring

lesbian lovers to the apartment and requested that the client takes part in sexual activities

with them. Mr. L eventually left the relationship. He resumed frequenting gay bars in

order to find sexual partners.

Mr. L met an older man, a former law enforcement officer, with whom he

established a friendship. The relationship had a sexual component, although Mr. L insists

that sexual activities between them were only on an occasional basis. This man was

married to a woman who was severely ill at the time, and most of his time was spent

caring for her. He allowed Mr. L to stay at their house, where he had his own room. Mr.

L would spend entire weeks at this friend’s house and would only rarely spend time in his

own apartment. Mr. L’s friend experienced problems with depression and made several

suicide attempts during the years of their relationship. Mr. L reported feeling very

anxious at times because of this, even though he claims that this was a “good” period of

his life.
 30

Some time later, the client’s mother was hospitalized due to a serious, terminal

illness. Mr. L spent most of his time taking care of her at the hospital for the 3 months

preceding her death. Following his mother’s death, the client’s younger brother, who had

financial problems at the time, committed suicide by hanging. Mr. L reported feeling

devastated by these losses, especially his younger brother’s suicide. He became

increasingly depressed and was spending all of his time at his friend’s house watching

TV. Approximately a year later, Mr. L’s friend committed suicide by locking himself into

his running car. Mr. L’s depression worsened even further and he was feeling burdened

by the responsibility of taking care of his friend’s wife, whose condition kept

deteriorating. His friend’s wife passed away shortly after, and Mr. L’s depression became

chronic. He was then unable to function and was reporting persistent suicidal ideation. He

was admitted to a psychiatric unit for observation and stabilization. Mr. L reported that

the medication used, as well as the safe, structured environment in the hospital were

helpful in making him feel better.

Upon discharge, Mr. L felt that he was able to provide himself with basic care but

his depression had not entirely remitted. He resumed a romantic relationship with a

former lover who had just opened a small business. This man promised Mr. L material

support as well as partnership in this business as soon as it became financially viable. Mr.

L was hesitant to involve himself in the relationship, but his friend’s promises eventually

convinced him.

Mr. L claims that since these losses, things have changed for the worse and that

he cannot get out of his depression. He reports chronic fatigue, lack of motivation or

interest, persistent sadness, difficulties going to sleep, lack of appetite and social
 31

withdrawal. He also reports drinking large quantities of alcohol on a regular basis. Mr. L

cannot frequent public places such as shopping malls due to severe Agoraphobia. Mr. L is

currently romantically involved with his business partner but reports marked conflicts in

the relationship, partly due to feeling disappointed and frustrated by his partner not

keeping the promises made. This will be Mr. L’s first experience in individual

psychotherapy.

Summary of Psychological Test Data

A brief description and integration of psychological test data will be presented

here, as space limitations prevent us from discussing individual test results in more detail.

Mr. L’s intellectual level is in the low-average range (FSIQ=84), with a non-

significant discrepancy between Verbal (88) and Non-verbal (80) functioning. His lowest

scores were obtained on Object Assembly (3), Similarities (5), Picture Completion (5),

Digit Symbol (5) and Comprehension (6). Significant difficulties were noted in the areas

of visual concentration, non-verbal memory, perceptual organization, as well as abstract

thinking.

Mr. L’s Rorschach protocol, scored using the Comprehensive System (Exner,

1993), is notable for the high number of responses provided (R=35), presence of shading

as a determinant (C’=3, T=4, V=2, Y=3), low frequency of human movement (M=2) or

whole human percepts (H=1), overemphasis on details (W=3, D=25, Dd=7), moderately

high level of distortions in thinking (FABCOM=2, ALOG=1), idiosyncratic or highly

subjective interpretations of reality (Xu%=52), low reactivity to emotional stimulation

(Afr=0.40), absence of cooperative movement or reciprocity (COP=0), presence of food

as content (Fd=2), and chronic, pervasive deficits in coping skills (CDI=5). Mr. L
 32

mentioned the content “fur” in 9 different responses in his protocol. Other responses

alternated between passive (“joined hands”, “small animals”) and more aggressive (“men

pulling, blood, their chest torn apart”, “dragon’s head”) percepts.

Mr. L produced a valid, interpretable MMPI-2 protocol with significant elevations

on scales 2 (Depression) and 4 (Psychopathic Deviate). An examination of supplementary

scales revealed high score on Repression and low scores on Ego Strength, Dominance

and Gender Role – Masculine. Content scales revealed that Mr. L’s depression as

reported on the MMPI-2 is characterized primarily by lack of drive and motivation.

Mr. L’s Early Memories Procedure protocol (Bruhn, 1990) involved many

episodes of “accidents” in early childhood. His first reported memory involved trying to

ride his uncle’s tractor by himself and barely escaping when the tractor tipped over.

Interestingly, his second memory involved picking and eating strawberries with his

mother in a “special”, secret place that no one knew about, where the strawberries were

larger than anywhere else in the area. Subsequent memories involved Mr. L being either

surprised or scared by unexpected events occurring in his life (e.g., his father wounding

himself with an axe).

WAIS and Rorschach data suggests that Mr. L experiences problems accessing

and using mental representations for adaptive purposes. In the absence of external

support, he is likely to become overwhelmed by demands and becomes confused,

frustrated and powerless. This over-reliance on external, rather than internal, adaptive

mechanisms is compounded by his lack of capacity to use concepts, his stereotypical

approach to situations and his passivity. On the other hand, there is no evidence of formal

thought disorder or major impairment in reality testing in his protocols.

 33

Mr. L’s personality dynamics as represented in the test data are characterized by

passive-dependent traits and a tendency to rely on an idealized figure for preserving

psychological integrity. Underlying his submissive and dependent stance lies an intense

fear of rejection or abandonment by others. Mr. L “expects the unexpected”, that terrible

consequences may stem from banal situations. Perceiving himself as unable to face

challenges and problems, his only recourse as he sees it is to rely on others to rescue him.

He is clearly more comfortable in relationships in which he can adopt such a passive and

receptive role. Paradoxically, he does not anticipate positive interactions with others, but

perceives relationships as potential sources of rejection, abandonment. This conflict may

manifest as chronic ambivalence and dissatisfaction with his social life.

Test data clearly indicates marked distress and dysphoria in Mr. L. Predominant

mood states appear to be characterized by subjective distress, anxiety, frustration, apathy

and pessimism. Such affects are likely to predominate in the absence of external support.

In the absence of such support, more regressed or less adaptive ways of managing

emotional dysregulation may appear, such as substance use or compulsive activity.

In summary, test results as well as the client’s history suggest a diagnosis of

Major Depressive Disorder as well as Agoraphobia, with dependent personality traits.

Adult Attachment Projective Protocol

Narratives obtained from the administration of the AAP with Mr. L will now be

discussed, along with their implications for clinical work with this client.

Picture 1 : Neutral

This picture is used as a warm-up for the rest of the task, and is therefore not

coded.
 34

Picture 2 : Child at Window

That’s pretty weird… Little girl, she’s in the front room at her place. I don’t

know, the patio door there… She’s looking outside, she’s waiting maybe her

friend, her little girlfriend or boyfriend, I don’t know… Yeah… I don’t see…/ Or

maybe she’s sad, she’s waiting her parents, I don’t know. (What might happen

next ?) Well, next…/ Uh… If it’s a friend she’s waiting for, maybe they’ll go eat

an ice cream somewhere… hm-hm… Or if she’s still…/ if her parents…/ Maybe

she’s ready, she’s waiting for them to go to the restaurant… yeah…yeah (What is

she thinking or feeling ?) The little girl…she’s anxious, she’s waiting

impatiently, she’s waiting impatiently for either her friend, her girlfriend, or her

parents. She’s anxious, she’s just waiting. Yeah… (Anything else comes to mind )

No, not really…. No.

This narrative is characterized by marked uncertainty about what is going on in the story

and what the outcome will be. One gets the sense that thoughts are often interrupted and

that the subject is hesitant to commit to one story line. In addition, the plot remains vague

and there is very little elaboration. There is too much interference in the development of

the story for this narrative to be considered coherent (Incoherent). No reference to a

personal experience is evoked. There are indications in the story of an attempt to

terminate the attachment system through action (“going to eat an ice cream”) (Agency,

capacity to act). The reader should note the reference to an “oral” theme, consistent with

other test data, although this feature is not part of AAP coding. There are references to a

desire for connectedness with others (peers, parents) throughout the story, although

fulfillment of this need is only evoked in the context of a relationship with peers
 35

(Connected). Attachment figures are representationally absent; there is no mention of

any interaction with parents actually taking place in the story. In terms of defensive

functioning, the narrative is characterized by the presence of marked Cognitive

Disconnection as evidenced by repeated uncertainty (e.g., “I don’t know”, “waiting”) and

entanglement (“anxious”).

Picture 3 : Scene of departure

Oh, oye-oye-oye… Well, I don’t know…Maybe the husband going on a business

trip… and his wife came to drop him at the airport. Here she’s asking him when

he’s gonna be back. Yeah, yeah, yeah… Hum, I don’t…/ (What might happen

next ?) Uh, he could- he could leave- leave and not come back, I don’t

know….yeah….hm (And, what are they thinking or feeling ?) Well if he leaves…/

Maybe he’s got a lover somewhere, he’s anxious to leave… Yeah, anyways, if he

told her that he’s coming back, she’s gonna wait- she’s gonna wait, she’s gonna

learn the bad news later. Hm-mm. Yeah… I…/ That’s what I can see. (Anything

else comes to mind ?) No. No, there’s no…/

This story is much like the first one in that Mr. L experiences difficulties developing a

clear story line. Both the beginning and the conclusion of the narrative are unclear. There

seems to be no logical connection for the transition between the themes of the business

trip and the husband leaving the wife for a lover. The presence of repetitions, fillers and

interrupted thought processes make the narrative difficult to follow (Incoherent). There

is no personal experience in the story. There is no indication of reciprocity in the

relationship, the narrative emphasizing conflict and betrayal (Failed synchrony). As in

the initial story, there is a marked presence of uncertainty (e.g., “I don’t know”, “wait”)
 36

and unfinished sentences in the narrative, thus pointing to the importance of Cognitive

Disconnection as a primary defense mechanism for this person.

Picture 4 : Person sitting on a bench

Oh !... Uh, it might be a little girl, she’s sad, she’s sitting on a park bench. She’s

sitting on a bench, she’s sad, her boyfriend left her. Her boyfriend left her for…/

Hm… So she’s taking it all out, she’s crying, she’s wondering what she’s- what’s

gonna happen. Hm…Yeah. So what’s going to happen next is, she’s gonna quit

crying, she’s gonna go meet one of her friends. Yeah. (And, what is she thinking

or feeling ?) Well, that’s…/ She’s wondering how she’s gonna get through this.

You know, she’s- she’s gonna have the idea to go see her friend. And she’s gonna

talk to her about it, she’s… Hm… Yeah (Anything else comes to mind ?) No, I

think that’s pretty much it for this one.

This narrative is much more coherent than the previous stories. There is a clear plot and

sequential description of a narrative (boyfriend left girl ; she is sad, upset ; she decides to

go see a friend and talk to her). The character’s internal states are clearly identified and

are consistent with the story line. Problems with completing sentences and thought

processes are still noted in the narrative. Overall, this is a coherent narrative although

significant interference with the process is still present (Moderately coherent). There is

no personal experience in the story. The character is portrayed as using action (going to

see a friend) to manage the internal conflict about the loss of the boyfriend (Agency,

capacity to act). We see again the representation of peers for adaptive purposes

(Connectedness). As in the Child at Window, attachment figures are representationally

absent. Defensive functioning is characterized by uncertainty (e.g., “wondering”) and the

 37

inability to complete thoughts, both markers of Cognitive Disconnection. Note the

inclusion of a third person in the story, a reference being made to the character’s

boyfriend potentially leaving her for someone else, but this thought process is quickly

interrupted and is not integrated in the narrative. From this point, object-seeking behavior

(with a friend) becomes the center of the narrative.

Picture 5 : Child and adult on a bed

Hm-mm… No… she’s not coming to tuck him in, she’s coming to wake him up.

That’s a mother who’s coming to wake her- her son. Hm… And there’s- maybe he

had a nightmare. She- she’s coming to wake him up. Yeah, that’s it, he had a

nightmare. Hm-mm. And…..Yeah. I don’t see…/ (And, what might happen next

?) Uh, she can…/ Uh… she can bring him in the kitchen, give him a glass of milk

or something, a cookie, I don’t know. Yeah…that’s what I see (And, what are they

thinking or feeling ?) Uh- Maybe he’s still in shock from his dream, he had a bad

dream. I don’t know what he might be thinking… And her idea is that she’s gonna

comfort him… Hm… (Anything else comes to mind ?) No.

A general plot is easily discerned in this narrative, and the theme evoked is appropriate to

what is described on the card. An illogical connection is observed early in the narrative,

where the mother is coming to wake the son, who just had a nightmare (and is therefore

awake), in order to comfort him. The son’s internal states cannot be elaborated in relation

to the internal conflict (sense of threat, fear). Although some difficulties were present at

the beginning of the narrative, the story can be considered coherent as a whole

(Moderately coherent). No personal experience is noted. The attachment figure is

described as wanting to comfort the son; however, there is no mention of any action on
 38

her part to do so, except to give him food, a substitute for parental care and comfort in

this case. The description of the mother’s actions in the narrative is not consistent with

“attachment-attuned care” and represents a “functional” type of interaction (Functional

synchrony). Defensive functioning on this card is represented primarily by markers for

Cognitive Disconnection, as illustrated by statements relating uncertainty (“I don’t

know”) and by an attempt by the attachment figure to take the child’s attention away

from the distress (giving milk and cookies). The theme of “nightmare” constitutes a

Segregated Systems marker, which in this particular case is resolved by the parent’s

actions and care. Although not “optimal” from an attachment perspective because

functional care never addresses the child’s fear directly, functional care is sufficient to

“contain” the traumatic material and prevent dysregulation.

Picture 6 : Child and adult looking at an ambulance

Eh…. Uh, an ambulance, it’s- it’s coming to pick up the little guy. The

grandmother is gonna go with him, she’s waiting for the ambulance. Yeah, I don’t

know what he did to hurt himself, but the ambulance is coming to pick him up. Ow

! That’s a weird one ! Uh, and he’s gonna go to the hospital in the ambulance.

The ambulance is here ! I don’t know what happened. Maybe he fell down the

stairs. Yeah (laughs). That’s pretty crazy. (And what might happen next ?) Uh, I

don’t know, he’s gonna go to the hospital, he’ll probably come back with his leg

in a cast. I don’t know. Yeah… that’s it. (So what are they thinking or feeling ?)

Uh the grandmother, she’s really upset, her grandson is hurt. They’re only

thinking about getting better. If he hurt his ankle…/ Yeah…that’s it. (Anything

else comes to mind ?) No.

 39

This narrative is characterized by difficulties in elaborating the plot. The discourse is

significantly impeded by comments intended to distance self from the situation (e.g.,

“That’s a weird one”, “that’s pretty crazy”), statements of uncertainty and unfinished

thoughts. There is only minimal elaboration of the circumstances leading to the current

situation, as well as the outcome. There is also minimal description of the character’s

internal states. This narrative therefore does not meet the basic requirements for

coherency (Not coherent). There is no mention of personal experience in the story. There

is no clear description of the attachment figure attending to the child’s distress, an

unusual occurrence on this card (Failed synchrony). As with the previous cards,

defensive functioning is characterized primarily by a strong presence of Cognitive

Disconnection markers, including uncertainty (“I don’t know”, “waiting”) and unfinished

thoughts.

Picture 7 : Man standing in a cemetery

Well, he’s at the cemetery, he’s visiting… a friend or his wife. It might be

someone he lost in a car accident and… he’s…/ (So, what might happen next ?) I

don’t know ! He’ll go back home, sad, just went to the cemetery to visit

someone… He’s sad, he’s gonna go back home and…/ I guess life goes on

(laughs). Yeah… I can’t see…/ it’s weird. (And- what- what is the character

thinking or feeling ?) I don’t know. If it’s someone close who- who died in an

accident or…/ He must be really sad about it. He would like to see this person

again. Hm… yeah, that’s about it.

Mr. L appears to elaborate a narrative only with difficulty on this card, which pulls for

themes of loss or death of attachment figures. The identity of the person who passed
 40

away is never clearly established and there is a sense of disconnection between different

ideas or thought processes as the individual struggles to build a story. No clear outcome

is elaborated and Mr. L is again unsuccessfully trying to distance himself from the testing

situation (“I can’t see…/ it’s weird”; “I don’t know”). The narrative on this card clearly

lacks coherency (Not coherent). No element related to personal experience is noted. The

character is not described as using internal resources, relationships, or action to manage

the activation of the attachment system on this card (No agency). In spite of lacking

elaboration, the story is replete with Cognitive Disconnection markers, including

uncertainty (“I don’t know”), unfinished thoughts, as well as a statement suggesting that

the character is glossing over the issue (“life goes on”). In addition, a Segregated Systems

marker is present in the narrative, as the character is described as “visiting” a dead

relative. The narrative does not provide indications that this marker is contained. The

story must therefore be coded as “Unresolved” with respect to trauma.

Picture 8 : Child in corner

Oh no!… Uh… What’s that?!!… It might be a little guy who end up being

punished and sent to the corner. Maybe a person- a person wants to beat him up,

he’s trying to protect himself. Yeah… And he must be thinking, “I’d like to be

somewhere else”(laughs). So what happened before… He had a fight, he had a

fight with either his little brother, little sister, so his mum or dad sent him to the

corner… Yeah…Hmm… And what he’s thinking, I already told you, he would like

to be somewhere else (laughs) (And what might happen next?) Uh?… (What

might happen next?) Well, he might be asked to apologize to his sister or brother,

if he had a fight with them. Hm-mm… yeah. That’s it.

 41

This narrative shows a clear plot with a specific beginning and end, although some

elements are left vague (e.g., the identity of the sibling). Internal states are minimally

elaborated. Yet, the discourse itself shows less interference than what was observed on

previous cards (Moderate coherency). No personal experience is present. There is a

reference to an attempt by the character at protecting himself from a threat (Agency,

capacity to act), thus resolving the segregated systems marker in the story (“beat him

up”). The narrative shows Cognitive Disconnection markers such as anger (“had a fight”)

and disconnection (“either brother or sister”). A Deactivation marker is also present,

represented by the theme of punishment occurring in the story.

Mr. L’s AAP protocol as a whole shows the following characteristics: evidence of

unresolved attachment-related material; generally low level of coherency; personal

agency as evidenced by capacity to act; absence of evidence suggesting capacity to

mentalize conflicts; desire for and capacity to relate to others, albeit only in the context of

peer relationships; strong representation of Cognitive Disconnection markers in the

narratives. On the basis of the Segregated Systems material that was left unintegrated in

the “Cemetery” story, Mr. L’s attachment status is Unresolved with respect to trauma.

The protocol also suggests that Mr. L is also Preoccupied as his primary attachment

configuration.

The data obtained from the AAP is consistent with the other information obtained

on Mr. L. The presence of unresolved, unintegrated material related to loss and death of

attachment figures certainly seems to match his history of sudden, traumatic losses

throughout his life, and probably accounts for the presence of severe anxiety in certain

contexts (crowds). Suicidal ideation has also been linked to Unresolved/Preoccupied

 42

attachment status (Adam, Sheldon-Keller, & West, 1996). His primary configuration of

Preoccupied is consistent with his history of depression, his passive-dependent

interpersonal orientation, and his chronic ambivalence in the context of attachment

relationships (West & George, 2002). Both Unresolved/Preoccupied attachment status as

revealed through the AAP and the psychological test data (e.g., low Comprehension score

on the WAIS, high scale 4 score on the MMPI-2) suggest that unprocessed anger may

underlie Mr. L’s presentation. Some of his behaviors, such as recurrent episodes of

drinking or compulsive sexual object-seeking, may constitute ways to prevent becoming

overwhelmed by traumatic material or internal conflicts stemming from past relationships

with attachment figures. Quite likely, Mr. L’s level of involvement and preoccupation

with past experiences with attachment figures, as well as his tendency to become

enmeshed in current relationships, do not allow for resolution of the trauma and

integration of conflictual material. Mr. L is thus likely to repeat the same dysfunctional

patterns in his current relationships, in an attempt to resolve the original trauma, while at

the same time making himself vulnerable to re-experiencing loss, rejection and

abandonment by others.

The data obtained from the AAP also includes some positive findings. While Mr.

L’s primary attachment configuration was determined to be Preoccupied, his protocol

showed evidence of agency, in the form of capacity to act, an unusual finding in

individuals with this attachment profile. While Mr. L is prone to experiencing episodes of

helplessness and dysregulation when confronted with situations that trigger unconscious,

segregated traumatic material, it appears that he has the capacity to use action for

adaptive purposes in his everyday life. This finding is in some ways consistent with other
 43

psychological test data, for example the relatively high number of responses provided on

the Rorschach. This statement is not meant to imply that such capacity to act is associated

with “healthy” behaviors in the individual. Indeed, there is evidence that Mr. L uses

compulsive activity as a way to manage internal conflicts and distress. The transition to

healthier, more appropriate ways to use activity to promote well-being could constitute a

therapeutic objective in itself. Given Mr. L’s lack of capacity to mentalize internal

conflicts, more structured, reality-oriented treatment focusing on building appropriate

coping skills might prove to be more beneficial to him than insight-oriented,

reconstructive therapy, at least initially.

A second positive finding lies with Mr. L’s desire for connectedness with others,

as shown on the AAP. This finding is also consistent with other test data (e.g., texture and

food responses on the Rorschach, MMPI-2 Scale 0 within normal limits). The problem,

from an attachment perspective, is that Mr. L’s desire for connectedness is at the level of

affiliative rather than attachment relationships. Affiliative or peer-based relationships

serve different purposes than attachment relationships and cannot assuage attachment-

related internal conflicts and distress. In this case, Mr. L’s ambivalence, anger and deep-

seated insecurity about attachment figure’s availability are not likely to be addressed

through his current relationships. Furthermore, given Mr. L’s level of preoccupation with

relationships and his tendency to assume a dependent stance, individual psychotherapy

should promote the development of appropriate autonomy and boundaries in the context

of a stable relationship with an attachment figure (therapist). The therapist should

probably expect some boundary- or limit-testing behavior, perhaps in a passive-

 44

aggressive way, as the therapy takes its course and an attachment to the therapist as a

nurturing, stable object starts developing.

Future research

While research available to date appears promising, it is obvious that much work

needs to be done before the AAP can be considered a viable alternative to the AAI.

Although the AAP has done well in construct validity studies using appropriate criteria

(e.g., AAI classification) and inter-rater reliability has been excellent across studies, there

is a need for more research using diverse populations in terms of age, ethnicity, language,

socio-economic background, gender and risk/clinical status. For example, it is not known

whether individuals showing limited intellectual or educational level can provide

narratives with sufficient material so that formal coding is likely to be valid.

In short, there is a need to build a research base that addresses a broader range of

adolescent and adult development as well as “problem” populations for the purpose of

theory-building and practical applications. As an important step, new research groups or

individuals need to become involved in research evaluating the psychometric properties

of the AAP, in order to replicate seminal findings and establish the instrument’s

reliability and validity in a more definitive manner. From a professional standpoint,

research is needed to establish the usefulness of AAP data to inform decision-making and

assess clinically-relevant variables. Studies are currently underway in the United States,

Canada, Japan, Italy, Germany, Russia, and Taiwan that should address many of the

points described above in the years to come.

 45

References

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of

attachment: A psychological study of the strange situation. Hillsdale, N.J.: Erlbaum.

Bartholomew, K., & Horowitz, L. (1991). Attachment styles among young adults:

A test of a four-category model. Journal of Personality and Social Psychology, 61, 226-

241.

Beliveau, M.J., Cyr, C., & Moss, E. (2002). Introduction to a projective measure

of attachment. Paper presented at the Annual Attachment Conference, Quebec City,

Canada.

Bowlby, J. (1969). Attachment and loss : Vol. 1. Attachment. New York : Basic

Books.

Bowlby, J. (1973). Attachment and loss : Vol. 2. Separation. New York : Basic

Books.

Bretherton, I., Ridgeway, D., & Cassidy, J. (1990). Assessing internal working

models of attachment relationships: An attachment story completion task for 3 year-olds.

In M.T. Greenberg, D. Cichetti, & E.M. Cummings (Eds.), Attachment in the preschool

years: Theory, research and intervention (pp. 273-308).

Buchheim, A., & George, C. (in press). The representational, neurobiological and

emotional foundation of attachment disorganization in Borderline Personality Disorder

and Anxiety Disorder. In J. Solomon & C. George (Eds.), Disorganized Attachment and

Caregiving. NY: Guilford Publications.

Buchheim, A., Erk, S., George, C., Kächele, H., Ruschow, M., Spitzer, M.,

Kircher, T. & Walter, H. (2006). Measuring adult attachment representation in an fMRI

 46

environment: A pilot study. Psychopathology. 39, 136-143.

Buchheim, A., Erk., S., George, C., Kächele, H., Martius, P., Pokorny, D.,

Ruchsow, M, &Walter, H (2006, submitted). Neural correlates of attachment

dysregulation in borderline personality disorder using functional magnetic resonance

imaging.

Crowell, J.A., Fraley, R.C., & Shaver, P.R. (1999). Measures of individual

differences in adolescent and adult attachment. . In J. Cassidy & P.R. Shaver (Eds.),

Handbook of attachment : Theory, research and clinical applications (pp. 334-365).

New York : Guilford.

Cyr, C., Beliveau, M.J., & Moss, E. (2003). Maternal attachment representations:

Link with child attachment and externalizing behavior problems. Symposium conducted

at the annual meeting of the Society for Personality Assessment, San Francisco, CA.

Dozier, M., Stovall, K.C., & Albus, K.E. (1999). Attachment and

psychopathology in adulthood. In J. Cassidy and P.R. Shaver (Eds.), Handbook of

attachment : Theory, research and clinical applications (pp. 497-519). New York :

Guilford.

Eagle, M. (1997). Attachment and psychoanalysis. British Journal of Medical

Psychology, 70, 217-229.

Exner, J. E., Jr. (1993). The Rorschach : A comprehensive system. Vol. 1 : Basic

foundations (3rd ed.). New York : Wiley.

Fonagy (1999). Psychoanalytic theory from the viewpoint of attachment theory

and research. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment : Theory,

research and clinical applications (pp. 595-624). New York : Guilford.

 47

Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon, G., Target,

M., & Gerber, A. (1996). The relation of attachment status, psychiatric classification, and

response to psychotherapy. Journal of Consulting and Clinical Psychology, 64, 22-31.

George, C., & West, M. (1999). Developmental vs. social personality models of adult

attachment and mental ill health. British Journal of Medical Psychology, 72, 285-303.

George, C., Kaplan, N., & Main, M. (1985). Adult Attachment Interview (2nd ed.).

Unpublished manuscript, University of California at Berkeley.

George, C., West, M., Pettem, O. (1997). Adult Attachment Projective protocol

and classification scoring system. Unpublished manuscript, Mills College.

Grice, P. (1975). Logic and conversation. In P. Cole & J. L. Moran (Eds.),

Syntax and semantics III: Speech acts, (pp. 41-58). New York: Academic Press.

Gunderson, JG. (1996). The borderline patient's intolerance of aloneness:

Insecure attachments and therapist availability. American Journal of Psychiatry. 153,

752-758.

Hansburg, H. (1972). Adolescent separation anxiety: A method for the study of

adolescent separation problems. Springfield, IL: C.C. Thomas.

Hazan, C., & Shaver, P.R. (1987). Romantic love conceptualized as an attachment

process. Journal of Personality and Social Psychology, 52, 511-524.

Hesse, E. (1999). The Adult Attachment Interview: Historical and current

perspectives. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment : Theory,

research and clinical applications (pp. 395-433). New York: Guilford.

 48

Joubert, D., Mazzarello, T., Moss, E., Paquette, D., & Zoccolillo, M. (in

preparation). Attachment and antisocial characteristics in adolescent mothers and their

children.

Klagsbrun, M., & Bowlby, J. (1976). Responses to separation from parents: A

clinical test for young children. British Journal of Projective Psychology, 21, 7-27.

Lerner, P. (1991). Psychoanalytic theory and the Rorschach. Hillsdale, N.J. : The

Analytic Press.

Lyons-Ruth, K., & Jacobvitz, D. (1999). Attachment disorganization : Unresolved

loss, relational violence and lapses in behavioral and attentional strategies. In J. Cassidy

& P.R. Shaver (Eds.), Handbook of attachment : Theory, research and clinical

applications (pp. 520-554). New York : Guilford.

Main, M. (1995). Recent studies in attachment. In S. Goldberg, R. Muir, & J. Kerr

(Eds.), Attachment theory : Social, developmental and clinical perspectives (pp. 467-

474). Hillsdale, NJ : The Analytic Press.

Main, M. & Solomon, J. (1990). Procedures for identifying infants as

disorganized/disoriented during the Ainsworth Strange Situation. In M. Greenberg, D.

Cichetti, & E.M. Cummings (Eds.), Attachment during the preschool years : Theory,

research and intervention (pp. 121-160). Chicago : University of Chicago Press.

Mezzich, A. C., Heliste, J. L., George, C., Kolar, A.,B., Swaney, M., Raspet, D.,

and Day, B-S. (2006, submitted). Adult attachment and substance use disorder in young

women.
 49

Moss, E., Cyr, C., & Dubois-Comtois, K. (2004). Maternal attachment

representations and child attachment at preschool age: The role of mother-child

discourse. In A. Bernier (Chair), A further look into the attachment transmission gap:

New conceptual and methodological approaches. Symposium conducted at biennial

meeting of the International Society for the Study of Behavioral Development, Ghent,

Belgium.

Moss, E., Thibaudeau, P., Joubert, D., Beliveau, M.J., & Dubois-Comtois, K. (in

preparation). Attachment in adolescents and their mothers.

Odipo, C. (2002). Adult attachment classification and foster mother’s perceptions

of their relationship with their at-risk foster children. Unpublished doctoral dissertation.

University of the Pacific, Stockton CA.

Schafer, R. (1954). Psychoanalytic interpretation in Rorschach testing: Theory

and application. New York: Grune & Stratton.

Slough, N.M., Goyette, M., & Greenberg, M.T. (1988). Scoring indices for the

Seattle version of the Separation Anxiety Test. Washington: University of Washington.

Solomon, J., & George, C. (1999). The place of disorganization in attachment

theory: Linking classic observations with contemporary findings. In J. Solomon & C.

George (Eds.), Attachment disorganization (pp. 3-32). New York: Guilford Press.

Sroufe, L. A., & Fleeson, J. (1986). Attachment and the construction of

relationships. In W. Hartup & Z. Rubin (Eds.), The nature and development of

relationships, (pp. 51-71). Hillsdale, NJ: Earlbaum.

 50

Sroufe, L. A. (1988). The role of infant-caregiver attachment in development. In

J. Belsky, & T. Nezworski (Eds.), Clinical implications of attachment (pp. 18-38).

Chicago, IL : University of Chicago Press.

Stone, M. (1990). The fate of borderlines. New York : Guilford.

Thompson, R. A. (1999). Early attachment and later development. In J. Cassidy &

P.R. Shaver (Eds.), Handbook of attachment theory and research (pp. 265-286). New

York : Guilford.

Waters, E., Crowell, J., Elliott, M., Corcoran, D., & Treboux, D. (2002).

Bowlby’s secure base theory and the social/personality psychology of attachment styles:

Work(s) in progress. Attachment and Human Development, 4, 223-229.

Waters, E., Merrick, S/, Trebous, D., Crowell, J., & Abersheim, L. (2000).

Attachment stability in infancy and early childhood. Child Development, 71, 684-689.

Webster, L., & Joubert, D. (2006). Rorschach correlates of unresolved attachment

in adolescents placed in foster care. Poster presented at the annual meeting of the Society

for Personality Assessment, San Diego, CA.

West, M., & George, C. (1999). Violence in intimate adult relationships: An

attachment theory perspective. Attachment and Human Development, 1, 137-156.

West, M., & George, C. (2002). Attachment and dysthymia : The contributions of

preoccupied attachment and agency of self to depression in women. Attachment and

Human Development, 4, 278-293.

 51

Table 1. ADULT ATTACHMENT PROJECTIVE SCORING CATEGORIES

Content Discourse

Agency Coherency

0=No agency 0=Incoherent

1=Capacity to act 1=Moderately coherent

2=Internalized Secure Base or Haven of Safety 2=Highly coherent

Synchrony Personal Experience

0=No relationship described 0=Absent 1=Present

1=Failed synchrony, functional relationship Defensive functioning

2=Synchrony, reciprocity-based relationship Ds=Deactivation

Connectedness E=Cognitive Disconnection

1=Own activity S.S.=Segregated Systems

2=Not connected (Resolved, Unresolved)

3=Connected
 52

David Joubert received his doctoral degree in Clinical-Developmental Psychology in

2006, from the Université du Québec à Montréal, Montréal, Canada. He is currently staff

psychologist at Napa State Hospital in California, working with forensic populations. He

has also been teaching at the School of Education, University of the Pacific, in Stockton,

California. His clinical and research interests include attachment and object relations

theory, forensic psychology, statistical techniques in clinical research, and personality

assessment, with an emphasis on projective techniques.

The author would like to thank Carol George, Ph.D. and Malcolm West, Ph.D., for their

useful comments.

Contact information:

David Joubert, Ph.D.

Staff Psychologist

Napa State Hospital

2100 Napa-Vallejo Highway

Napa CA 94558

U.S.A.

Phone: (707) 254-2675

E-mail: [email protected]