ATTACHMENT AND CORE RELATIONSHIP THEMES: WISHES FOR AUTONOMY AND CLOSENESS IN THE NARRATIVES OF SECURELY AND INSECURELY ATTACHED ADULTS

Abstract

This study examines links between attachment states of mind and relationship schemas in a sample of 40 young adults, half of whom were hospitalized as adolescents for psychiatric treatment. Participants were interviewed about their closest relationships, and, using the Core Conflictual Relationship Theme method, their narratives about these relationships were analyzed for the relative frequency with which they expressed wishes for closeness and for autonomy in relation to others. Participants were also administered the Adult Attachment Interview and were classified with respect to security of attachment. Security of attachment was associated with the relative frequency with which participants expressed wishes for autonomy in their narratives about close relationships, even after accounting for current levels of psychological functioning and history of serious psychopathology in adolescence. Security of attachment was not associated with the relative frequency with which participants expressed wishes for closeness. The study suggests that core relational wishes for autonomy are linked specifically with subtypes of insecure attachment. These findings extend what is known about connections between the representation of early attachment relationships and the wishes and needs expressed in current relationships with significant others.

Beginning in infancy, human beings form images of relationships based on their dealings with caregivers, and these images guide expectations of how new relationships will proceed (Stern, 1985). Such mental models, or relationship schemas, help make sense of social interactions and are thought to influence behavior toward others throughout life (Baldwin, 1992; Horowitz, 1994). Two of the most influential theories of how we construct mental representations of human relationships are Freud’s concept of transference and Bowlby’s concept of attachment. This study examines empirical links between these two concepts. It focuses specifically on a dialectic that is central to both attachment theory and psychodynamic concepts of transference: the interplay between strivings for interpersonal closeness and individual autonomy.

Transference

Freud conceived of each person as having a central relationship pattern, “a stereotype plate (or several such), which is constantly repeated—constantly reprinted afresh—in the course of the person’s life” (Freud, 1912, pp. 99–100). According to Freud, we approach new people with personal relationship templates composed in large measure of the frustrated wishes, longings, and expectations left over from earlier unsuccessful attempts at need gratification. We literally transfer these wishes from earlier relationships to later ones. Although the term transference is often used in reference to an individual’s relationship with a psychotherapist, Freud and later psychodynamic theorists concur that relationships of all kinds contain transferred elements that emerge as repetitive themes.

Concerns about closeness and autonomy are key elements of many psychodynamic theories about mental models of relationships, including Freudian conflicts between love and hatred (Freud, 1909/1955), Mahler’s (1972) processes of separation-individuation and rapprochement, and Winnicott’s (1965) ideas about the development of the true self and false self. In addition, empirical evidence shows that closeness and autonomy are among the most prominent themes in individuals’ representations of their close relationships. For example, in the Penn Psychotherapy Project (Luborsky, Crits-Christoph, Mintz, & Auerbach, 1988), a study of 73 patients in psychotherapy, the two most frequently identified wishes in patients’ narratives about relationships were “to be close” and “to assert myself and be independent” (Luborsky, Barber, Schaffler, & Cacciola, 1998).

Assessment of transference themes is complicated by that fact that many elements of relationship schemas are not conscious (Baldwin, 1994). There is both theoretical and empirical support for the idea that some aspects of transference are inaccessible by self-report because they are outside of conscious awareness (Luborsky et al., 1999; Mitchell & Black, 1995). In addition, transference is thought to contain procedural knowledge (i.e., rules and skills for processing social information) that may not be conscious (Kihlstrom, 1987). Mindful of these considerations, in the current study we implemented the Core Conflictual Relationship Theme (CCRT) method (Luborsky & Crits-Christoph, 1990), a widely used method of assessing transference themes that taps into unconscious elements of relationship schemas.

Attachment

Attachment theory (Bowlby, 1969, 1973, 1980) derives from the fundamental premise that the human infant has a biologically based, survival-promoting desire for proximity to caregivers. On the basis of repeated interactions with their caregivers, infants develop internal working models of attachment relationships, which include images of the self as lovable or not and images of the attachment figure as available or unavailable to meet the child’s needs. The young child uses these working models to forecast the caregiver’s availability and responsiveness and to develop behavioral strategies that are designed to maximize proximity to the caregiver. Internal working models of attachment are thought to guide expectations about relationships throughout life. In adult relationships, secure attachment provides the basis for both autonomy and closeness. Autonomous functioning in the world is facilitated by the sense that one has a secure base of support, and intimacy is possible when one feels free to be separate and function independently (Holmes, 1997).

Empirical work supports the theory that adults exhibit a limited number of primary attachment strategies (Bartholomew, 1994; Crowell & Treboux, 1995; Feeney & Noller, 1990; Hazan & Shaver, 1987). Securely attached adults have a sense of the self as lovable and of the other as available in times of need and distress. Valuing both closeness in relationships and autonomy of the self, the securely attached individual is able to use an attachment relationship as a “safe haven” in times of stress and as a base from which to explore the world. By contrast, adults who expect that attachment figures will be unavailable in times of need tend either to be insecure and preoccupied with closeness and worry about others’ availability or to be insecure and dismissing of the importance of and need for others. Cognitively and behaviorally, insecure-preoccupied adults tend to overvalue closeness at the expense of autonomy, and insecure-dismissing adults tend to overvalue autonomy at the expense of closeness.

We cannot measure internal working models of attachment relationships directly, but two major avenues of research aim to provide indirect information about these models:

1. Self-report measures (e.g., Bartholomew & Horowitz, 1991; Hazan & Shaver, 1987) ask adults about their typical modes of forming close relationships. They measure attachment styles that, by definition, reflect only those aspects of internal working models of attachment that are within the individual’s conscious awareness.

2. The Adult Attachment Interview (AAI; George, Kaplan, & Main, 1984; Hesse, 1999) assesses states of mind with respect to attachment by analyzing interviewees’ narratives about relationships with early attachment figures. The AAI (George et al.; Hesse) focuses primarily on the style of an individual’s discourse (i.e., the coherence, fluency, and openness with which an individual describes relationships with early caregivers). The AAI focuses only secondarily on the content of those descriptions (e.g., whether or not parents are described as loving). Indicators of insecure states of mind include awkward pauses, gaps in memory, incoherent discourse, and other signs of defensiveness as interviewees talk about parents and other important attachment figures. Discourse style is thought to reflect a state of mind with respect to attachment (George & Solomon, 1996) that is guided by both conscious and unconscious elements of an individual’s internal working model of attachment. Just as attachment states of mind are thought to shape narratives about early relationships, we expect them to shape descriptions of close adult relationships as well.

Not surprisingly, attachment styles as measured by self-reports and states of mind with respect to attachment as measured by discourse style on the AAI are not strongly related to each other empirically (Bartholomew & Shaver, 1998). For the current study, we use the AAI as a measure of attachment precisely because it was designed to “surprise the unconscious” (Hesse, 1999) and incorporate information about internal working models of attachment relationships that may be in part or wholly outside of the individual’s awareness. Wishes for closeness and autonomy—issues central to both attachment theory and concepts of transference (Holmes, 1997)—may be distorted for defensive reasons (Westen, 1991). For example, Main, Kaplan, and Cassidy (1985) suggested that individuals with a dismissive attachment style may be loathe to recall or reveal experiences in which others disappoint them because such disappointment implies a degree of emotional dependency that is threatening. We expect to tap into the complexity of the dialectic between wishes for autonomy and closeness most effectively using measures that provide information about unconscious as well as conscious representations of relationships.

Linking Transference Schemas and Attachment

Transference and attachment have been operationalized successfully and have spawned rich fields of empirical investigation, but these two avenues of research have remained largely separate, and they are discussed in separate literatures. Many of the earliest studies bridging these two domains have come from clinicians using case reports of individual psychotherapies (Alexander & Anderson, 1994; Holmes, 1997; Szajnberg & Crittenden, 1997). In addition, there is now a small but growing number of empirical studies linking attachment styles with how people represent themselves and their relationships with others.

In several studies, security of attachment has been associated with how people perceive the quality of their working alliances with psychotherapists. Mallinckrodt, Coble, and Gantt (1995) found that, in a sample of women in brief therapy, participants’ memories of attachment to parents were associated with their ratings of the quality of the working alliance with their psychotherapists. Satterfield and Lyddon (1995) found that, among clients seeking counseling at a university clinic, those indicating less trust in the availability and dependability of others on a self-report measure of attachment were more likely to evaluate the relationship with a counselor in negative terms during the early phase of counseling. Eames and Roth (2000) found similar links between fearfulness in attachment and lower ratings of the therapeutic alliance in an outpatient clinic population. In a study of more seriously disturbed clients and their case managers, Tyrrell (1999) found that client and case manager attachment states of mind (as measured by the AAI) interacted in predicting how clients rated the working alliance. Specifically, clients who minimized the importance of their early attachment relationships on the AAI perceived themselves as having better alliances with case managers who were less minimizing of the importance of relationships and vice versa.

In a second area of research, studies have begun to link attachment styles with the quality of internal representations of self and other. Insecure attachment styles have been associated with more malevolent representations of self and other, whereas secure attachment has been associated with more benign images of self and other on a variety of measures of self and object representations, including Blatt’s Object Representation Inventory (Levy, Blatt, & Shaver, 1998), the Rorschach (Rothstein, 1997), and the self-report Structural Analysis of Social Behavior circumplex model of interpersonal relationships (Morrison, Goodlin Jones, & Urquiza, 1997). Mikulincer (1995) found that individuals classified as secure on a self-report measure of attachment had more balanced, complex, and coherent self-images than those classified as insecure. Mikulincer and Horesh (1999) found that attachment style is linked with perceptions of others, specifically that insecurely attached adults tend to perceive in others both the particular negative traits that they see in themselves and those traits that they do not want to possess.

The Current Study

Although attachment theory and many psychodynamic theories of transference postulate concerns about closeness and autonomy as two central elements of internal representations of relationships, we know of no empirical work to date that has attempted to corroborate these links. The study reported here examines the specific interface between attachment styles and core relational wishes for autonomy and closeness. The two interview-based methods used in the study—the CCRT method (Luborsky & Crits-Christoph, 1990) and the AAI—are the most widely used systems for reliably assessing transference themes and states of mind with respect to attachment respectively, but apparently until now they have not been studied together. Both the CCRT and the AAI are idiographic methods of identifying relationship schemas (Baldwin, 1992). Patients’ stories about relationships are the objects of analysis, and, within limits, these stories take whatever form the patient cares to give them. Both methods allow for analysis of relationship schemas that may not be entirely in the patient’s awareness but may be inferred from narratives about specific interpersonal experiences.

The study focuses on relationship schemas in young adulthood, when establishing sustained intimate relationships is a primary developmental task. In theory, how one negotiates the conflicting needs of adolescence sets the stage for this task (Erikson, 1959). Adolescence is a time when both the drive toward autonomy and the wish to maintain relatedness with parents are active and potentially in conflict. Although autonomy and relatedness have sometimes been placed at opposite ends of a relational continuum, research suggests that, for optimal social development, relationships between parents and adolescents should include both qualities (Allen & Hauser, 1996; Allen, Hauser, Bell, & O’Connor, 1994; Collins, 1990; Grotevant & Cooper, 1985; Kobak, Cole, Ferenz-Gillies, Fleming, & Gamble, 1993). The successful integration of autonomy and relatedness in parent–adolescent relationships is thought to pave the way for secure attachment and sustained intimate relationships in young adulthood (Allen & Hauser; Erikson).

In examining these issues, we hypothesized that core relational themes of autonomy and closeness are linked specifically to an individual’s state of mind with respect to attachment in close relationships rather than to some broader index of psychological functioning. Although psychodynamic theory posits links between general psychological distress and the inappropriate imposition of relationship templates onto new relationships (Freud, 1912; Mitchell & Black, 1995), little empirical work has been done in this area, and the few existing studies have not found consistent differences in core relationship themes between those with and without significant psychopathology (Chance, Bakeman, Kaslow, Farber, & Burge-Callaway, 2000; Diguer et al., 2001; Wilczek, Weinryb, Barber, Gustavsson, & Åsberg, 2000). Therefore, we expect states of mind with respect to attachment to be associated with the relative frequencies of wishes for autonomy and closeness independent of participants’ levels of current psychological functioning and of their histories of psychiatric hospitalization in adolescence.

On the basis of the preceding considerations, this study tests two hypotheses:

Hypothesis 1: Young adults’ states of mind with respect to attachment will be associated with the relative frequencies with which they express wishes for autonomy in their narratives about current close relationships. Relative to all expressed wishes, insecure-dismissing individuals will express proportionately more wishes for autonomy than those classified as secure, and insecure-preoccupied individuals will express proportionately fewer wishes for autonomy than those classified as secure.

Hypothesis 2: Young adults’ states of mind with respect to attachment will be associated with the relative frequencies with which they express wishes for closeness in their narratives about current close relationships. Relative to all expressed wishes, insecure-dismissing individuals will express proportionately fewer wishes for closeness than those classified as secure, and insecure-preoccupied individuals will express proportionately more wishes for closeness than those classified as secure.

We expected that the hypothesized relations between attachment classification and wishes for autonomy and closeness would be found even after accounting for participants’ concurrent levels of psychological functioning and histories of psychiatric hospitalization during adolescence.

Method

Participants

The 40 participants in this study constitute a subsample of 146 individuals assessed as part of a longitudinal study of psychological development (Hauser, Powers, & Noam, 1991). On entering this longitudinal study at age 14, study participants were members of primarily Caucasian middle- and upper-middle-class families. Approximately half were recruited from the freshman class of a local high school, and half were nonpsychotic psychiatrically hospitalized adolescents. Conduct and mood disorders were the predominant diagnoses in the clinical cohort at the time of hospitalization at age 14. Children with developmental disabilities were excluded from the study. Of the original adolescent participants, 98% were studied again at age 24 using age-appropriate measures of psychological and interpersonal functioning.

The current study is based on a secondary analysis of data collected during the age 24 wave of assessments using a subsample of 20 male and 20 female individuals, selected in equal numbers from the normal and psychiatrically hospitalized groups. (A subsample was chosen because the labor-intensive methods used in the study necessitated limiting the number of participants.) A central variable in the adolescent phase of the study was ego development as measured on the Loevinger Sentence Completion Test (SCT; Loevinger, 1979). Adolescent ego development, as measured by the SCT, has been shown to predict young adult attachment coherence (Hauser, Gerber, & Allen, 1998), so we expected that choosing participants with a full range of SCT scores would yield a sample in which all three major attachment classifications were well represented. To accomplish this, we used a stratified random sampling technique. From each of eight subgroups, 5 participants were drawn: Median splits were used to divide males and females from the hospital and high school cohorts into those with low and high SCT scores. There were sufficient numbers to allow for random sampling from seven of these groups; however, because of the small cell size, all male participants from the hospital cohort with high SCT scores were included.

At the time of their admission to the study at age 14, the 20 hospital and 20 high school participants did not differ with respect to birth order, number of siblings in the family, and whether parents were living together. At the time of the age 24 assessments, 81% were employed, 46% were married or living with a partner, and 89% had at least some years of formal education after completion of high school. At age 24, participants from the original hospital and high school cohorts did not differ with respect to marital status or employment status. Participants in the high school cohort reported an average of 15.6 years of formal education, and those in the hospital cohort an average of 13.1 years of formal education. Of the 20 participants who had been hospitalized at 14 years of age, 5 reported having been rehospitalized at some point between ages 14 and 24, and 1 of the 20 participants from the original high school cohort was hospitalized for psychiatric reasons during this period. All participants were assessed using the instruments described next.

Measures

Attachment classifications

Attachment classifications were determined using the AAI (George et al., 1984). This structured interview probes individuals’ descriptions of their childhood relationships with parents in abstract terms and with requests for specific memories supporting those terms. Participants are asked to list five words to describe their relationship with each parent and then to describe specific episodes that reflect those words. Questions next focus on early experiences of emotional upset, rejection, loss, separation, and trauma. Participants are then asked to describe how their relationships with parents have changed over time and the current state of each relationship. Analysis of the text of these interviews allows for reliable classification of an individual’s state of mind with respect to attachment.

The AAI coding system (Main & Goldwyn, 1984) was used to classify participants’ states of mind with respect to attachment into one of three categories: (a) secure-autonomous yet freely valuing of attachment; (b) insecure-dismissing of attachment relationships; and (c) insecure-preoccupied with attachment relationships. In addition to one of the these three primary classifications, some participants are given the secondary classification of being unresolved with respect to past loss or trauma. Individuals classified as secure-autonomous maintain a balanced view of early relationships, value attachment relationships, and view attachment-related experiences as influential in their development. Individuals classified as insecure-dismissing use strategies that minimize, dismiss, devalue, or deny the importance of attachment relationships. Those classified as insecure-preoccupied recall memories of nonloving but involving, even role-reversing, parenting in which a child needed to be alert to parental needs in preference to his or her own. Insecure-preoccupied individuals manifest ongoing preoccupation about the availability of important others. Those classified as insecure-unresolved report attachment-related traumas of loss or abuse and manifest confusion and disorganization in the discussion of these particular traumas (Crowell, Fraley, & Shaver, 1999) but not other relationships. Attachment classifications using the AAI have been correlated with numerous aspects of interpersonal functioning and have been shown to be transmitted across generations (Allen, Hauser, & Borman-Spurrell, 1996; Allen, Moore, Kuperminc, & Bell, 1998; Pianta, Egeland, & Adam, 1996; Van Ijzendoorn, 1995).

Relationship narratives

Relationship narratives in young adulthood were collected using the Early Adult Close Peer Relationship Interview (Schultz, Hauser, & Allen, 1990). This semistructured interview was designed to elicit participants’ current networks of social relationships followed by an intensive focus on their two closest peer relationships (one romantic relationship and one nonromantic friendship). Within each of these close relationships, experiences of connection, intimacy, autonomy, and conflict resolution are probed with questions about recent experiences with these peers (e.g., “How much time do the two of you spend together?” and “What is the biggest decision you two have ever made together?”).

Relationship themes

Relationship themes were scored from the relationship narratives in transcripts of the Early Adult Close Peer Relationship Interview using the CCRT method. The CCRT method is an assessment system for extracting relationship schemas from narratives about interpersonal interactions (Luborsky & Crits-Christoph, 1990, 1998). Studies using the CCRT method have found that people have identifiable core relationship themes that are repeated across multiple stories (i.e., themes that are thought to be transferred from earlier relationships to later ones; Luborsky, Crits-Christoph, & Mellon, 1986; Luborsky et al., 1996, 1999). It has been empirically demonstrated that the CCRT method guides the delineation of relationship elements that have considerable stability over time and across relationships (Fried, Crits-Christoph, & Luborsky, 1990; Luborsky & Diguer, 1998).

Relationship episodes are identified in interview material in which the participant speaks of an interaction with another person in sufficient detail to enable scoring of specific components. The CCRT rater identifies wishes, responses from others, and responses from self in each relationship episode describing an interaction with another person. Relationship themes are defined as the most frequently occurring categories of wishes, responses from others, and responses from self across the entire set of relationship episodes. For example, a man may tell several stories about dealings with different people, but in the majority of his narratives the theme may be “I want to be close to the other person, but the other is distant, and I feel angry and withdrawn.” In this study, only wishes were considered because these were hypothesized to be linked specifically with attachment strategies. Each wish was rated as belonging to one of the eight clustered categories listed in Table 1. Because standardized categories of the kind used in the CCRT scoring system can capture the meaning of a particular wish, response from others, or response from self only imprecisely, both the best fitting and next best fitting categories are typically scored and analyzed for each identified component of a relationship episode (Luborsky & Friedman, 1998). This procedure was followed in our data coding and analyses.

TABLE 1.

CCRT Clustered Categories of Wishes: Proportion of Wishes coded as Best Fitting and Next Best Fitting

	Proportion
Wishes	Of all best fitting wishes	Of all next best fitting wishes
To be close and accepting	.32	.29
To assert self and be independent (autonomous)	.13	.12
To oppose, hurt, control others	.10	.10
To be controlled, hurt, not responsible	.10	.09
To be distant and avoid conflicts	.12	.11
To be loved and understood	.09	.11
To feel good and comfortable	.06	.10
To achieve and help others	.08	.10

Good evidence exists for the reliability and validity of the CCRT method (Crits-Christoph, Luborsky, Dahl, & Popp, 1988; Levine & Luborsky, 1981; Luborsky et al., 1986; Luborsky & Diguer, 1998). Clinicians have achieved good levels of agreement when using the CCRT method to formulate dynamic themes in psychotherapy (Crits-Christoph, Cooper, & Luborsky, 1988). The CCRT method has demonstrated convergent validity with several other measures of central relationship patterns (Luborsky, 1988; Perry, 1989). Relationship themes as measured by the CCRT method have been correlated with accuracy of interpretations in psychodynamic psychotherapy (Crits-Christoph, Cooper, et al., 1988). Identification and discussion of CCRT themes in psychotherapy have been associated with general clinical improvement (Crits-Christoph, Cooper, et al., 1988) and, more specifically, with mastery of central conflicts (Grenyer & Luborsky, 1996).

Wishes for closeness constitutes a clustered category in the CCRT system. Inspection of particular coding decisions in this data set revealed that coders routinely used this category to characterize statements that indicated the narrator’s spoken or implied wish for psychological or physical proximity to another person. This seemed consistent with and most analogous to an attachment-related wish for proximity. By contrast, wishes “to be loved” and “to feel good” were used by our coders to characterize more particular desires, such as the desire for concrete assistance with a task, for sexual contact, and for enjoyable activities. Among the eight clustered categories of wishes in the CCRT coding system (see Table 1), the category “to assert self and be independent” most closely represents strivings for autonomy. Subcategories of the clustered wish to be independent include “to be my own person,” “to assert myself,” and “to have self-control.” The wish to be independent was used by coders most often to categorize participants’ statements expressing a desire for independence and autonomy. By contrast, the clustered category “to be distant and avoid conflicts” includes the subcategories “to avoid conflict” and “to not be hurt” and was coded most often in narratives in which the speaker was consciously dissatisfied with or worried about some specific aspect of an interaction with another person. The CCRT wish to be independent was, therefore, taken to represent wishes for autonomy in this study.

In addition to categorizing interpersonal wishes, the CCRT method uses standard categories to code responses of others and responses of the self in narratives about relationships (Barber, Crits-Christoph, & Luborsky, 1998). We did not make predictions about whether attachment style would be related to particular responses of others and responses of the self to others because there was less theoretical support for specific links between these relationship elements and internal working models of attachment.

Psychological distress

Current level of psychological distress was measured using the Symptom Checklist-90-Revised (SCL-90-R; Derogatis, 1977). This self-report measure asks individuals how much discomfort each of 90 psychological and somatic symptoms has caused them during the previous 6 months. The Global Severity Index, a summary score of the SCL-90-R, was used as an index of psychological distress at the time of the young adult assessment. The mean score for the sample was 0.665 (SD = 0.50, range = 0.08–1.83), which is higher than norms reported for a nonclinical sample (M = 0.31, SD = 0.31, N = 974) but lower than those reported for psychiatric outpatients (M = 1.26, SD = 0.68, N = 1,002; Derogatis, 1977).

Procedures

Data collection

All participants came to the laboratory to complete study assessments. Written informed consent was obtained from all participants after the procedures for each interview had been fully explained. The AAI and Early Adult Close Peer Relationship Interview were administered along with other self-report and interview measures as part of the larger longitudinal study. The AAI was conducted early in the protocol to minimize any influence of prior assessments on participants’ AAI responses, and the Early Adult Close Peer Relationship Interview was conducted later in the protocol. Both interviews were audiotaped and transcribed. The AAI takes between 45 and 90 min to complete and interviews last on average about 60 min. The Early Adult Close Peer Relationship Interview takes between 1 and 2 hr to complete, and interviews lasted on average about 90 min. Interviews were conducted by trained interviewers who had no knowledge of participant data collected in adolescence (e.g., psychiatric history, ego development level).

Data coding

The AAI and Early Adult Close Peer Relationship Interviews were coded by raters who worked independently of each other and independently of those who collected the interview data. No raters coded both AAIs and Early Adult Close Peer Relationship Interviews. Raters were unaware of all participant data except information contained in the coded interviews themselves.

Adult Attachment Interview

All AAIs were rated by a coleader of the Adult Attachment Institute Workshops. Interrater reliability of overall classifications was established using two additional uninformed raters, who classified an additional 15 and 21 transcripts, respectively, with a combined agreement with the primary rater of 83% on overall classifications (Cohen’s κ = .79).

AAI classifications of the 40 young adults in this study were as follows: 17 secure, 12 insecure-dismissing, and 11 insecure-preoccupied. There were no gender differences in the distribution of attachment classifications, χ² (2, N = 38) = 0.15, p = .93. Consistent with prior research (Allen et al., 1996), significantly more young adults were classified as secure from the high school cohort (n = 14) than from the hospital cohort (n = 3). Of the 40 participants, 10 (4 secure, 3 insecure-dismissing, and 3 insecure-preoccupied) also received the classification of being unresolved with respect to trauma. However, according to attachment theory, the underlying secure, dismissing, or preoccupied attachment organization is primary and the lack of resolution with respect to trauma is secondary. Participants with unresolved attachment display their lack of resolution only in discussing specific traumatic events, and their underlying attachment classifications are apparent in the remaining AAI discourse (Rosenstein & Horowitz, 1996). Because participants in our study told relationship narratives about adult peers and were not recalling traumatic events from childhood, we expected that their primary attachment strategies would be reflected in their relationship narratives. Thus, we expected that the balance of their wishes for autonomy and closeness would be determined by their underlying secure, dismissing, or preoccupied attachment strategies. In this respect, our use of AAI attachment classifications is consistent with their use in prior empirical studies (Bakermans-Kranenburg & Van Ijzendoorn, 1993; Fonagy et al., 1996; Rosenstein & Horowitz, 1996).

CCRT coding of Early Adult Peer Interviews

Lester Luborsky, the originator of the CCRT method, trained Robert J. Waldinger in the CCRT coding method and reviewed his coding of six Early Adult Close Peer Relationship Interview transcripts. These coded transcripts served as the standard for subsequent assessment of interrater reliability. Raters were two undergraduate students and two recent college graduates, all of whom were psychology majors. They were trained in the CCRT method by Robert J. Waldinger, in consultation with Lester Luborsky. Initial training was conducted using Early Adult Close Peer Relationship Interviews from participants in the larger cohort who were not among the 40 participants included in the study. Raters coded 10 interviews in common, and each interview was reviewed in weekly group meetings before the raters coded the six standard transcripts independently. Raters were trained to reliability in two tasks: (a) identifying relationship episodes within the interview transcript and (b) assigning components within each relationship episode a best fitting and next best fitting CCRT standard category. Raters were considered reliable in their identification of relationship episodes if the beginnings and endings of episodes differed by no more than two lines of the written transcript. On average, pairs of raters demonstrated 96% (range = 89%–99%) agreement in identifying 72 relationship episodes.

The CCRT scoring system uses approximately 30 individual categories to classify wishes. Using cluster analysis, Barber et al. (1998) reduced these individual categories to eight wish clusters. Raters scored narratives using the 30 more finely grained individual categories, and these were then aggregated into the eight wish clusters for purposes of reliability testing and data analysis. To assess reliability in categorizing wishes according to eight clustered categories (Luborsky & Diguer, 1998), weighted Cohen’s kappas were calculated for each pair among the independent raters. According to Cohen (1968), when certain disagreements are less important than others, kappas may be weighted in order to have a more valid estimate of the reliability. This weighting has been used in several CCRT studies (Luborsky & Diguer). In counting matches between judges, the following weights were used: (a) Agreement between judges on the best fitting CCRT clustered category was given a weight of 1.0; (b) a weight of .66 was given when the match was based on agreement between the best fitting CCRT category of Judge 1 and the next best fitting category of Judge 2; (c) when the match was based on agreement between the next best fitting categories of the 2 judges, a weight of .33 was given; and (d) when there was no match between categories, a score of zero was given. Weighted kappas for reliability of pairs of judges ranged from 0.73 to 0.85. These are comparable to those obtained in other studies (Luborsky & Diguer).

The Early Adult Close Peer Relationship Interview elicited a varying number of codable relationship narratives from the 40 participants, ranging from as few as 7 to as many as 44 relationship narratives in one interview. A minimum of 7 relationship narratives is deemed adequate to identify core repetitive relationship themes (L. Luborsky, personal communication, June 1994). A total of 581 relationship episodes were coded for the 40 participants. Given the labor-intensive nature of the coding procedure, coders scored all identified relationship narratives for each participant up to 15. Where more than 15 were available for coding, priority was given to those narratives that were told in response to the most content-neutral questions (e.g., a question about time spent together as opposed to one about the greatest disagreement the participant has had with the other person). This strategy was used to yield relationship episodes that would most closely approximate those elicited by the content-neutral prompts in the Relationship Anecdotes Paradigm Interview (Luborsky, 1998), an interview commonly used to collect narratives that can be coded using the CCRT method. The mean number of relationship episodes coded per participant was 14.6 (SD = 1.7, range = 7–16). Six of the 40 participants told fewer than 15 relationship episodes (mean = 12, range = 7–14), and for 2 participants 16 relationship episodes were coded.

Results

As noted, each wish in a relationship narrative was coded for the best fitting and the next best fitting clustered category from among the eight categories of wishes listed in Table 1, yielding two coded categories for each identified wish. Using both scores for each wish, frequencies were calculated as the proportion of wishes falling into a particular clustered category. Thus, if a participant expressed a total of 20 codable wishes in his or her relationship episodes and 10 of these were categorized by raters as wishes “to be close,” then the proportion of that participant’s wishes corresponding to the wish to be close would be .50. We used the total number of codable wishes in the denominator of these calculated proportions rather than the total number of relationship episodes in which a particular wish appeared. This was done on the premise that repetition within a relationship narrative might indicate the level of an individual’s conscious or unconscious preoccupation with a particular wish and, therefore, would be relevant in these analyses. Table 1 indicates the frequencies with which wishes in this sample of 40 young adults were coded as best fitting or next best fitting.

Our two hypotheses were tested most parsimoniously by constructing a single multivariate analysis of covariance (MANCOVA) model with the proportions of wishes for closeness and autonomy as dependent variables. The arcsine statistic was used to transform both proportions because they represent binomial distributions (e.g., wish for closeness vs. other wish). Attachment classification was included as the independent factor. History of psychiatric hospitalization in adolescence and current psychological functioning as indexed by the SCL-90-R were used as covariates to minimize their potential confounding effects on the relationship between attachment classification and CCRT measures. Interaction terms were not entered into the design because of the complexity of the model and our desire to preserve power in the setting of a relatively small sample.

Before conducting our analyses, we examined possible colinearity among the independent factor and the covariates. There were roughly equal proportions of insecure-preoccupied and insecure-dismissing individuals in both the hospitalized and nonhospitalized subsamples (in the hospital cohort, 8 preoccupied and 9 dismissing individuals; in the high school cohort, 3 preoccupied and 3 dismissing individuals). Attachment classification, gender, and cohort membership were not linked with psychological functioning at age 25. Examination of independent relations between the covariates in the model and the outcome variables revealed no significant links.

Table 2 summarizes the MANCOVA results, using Wilks’s lambda to estimate the multivariate F test. The F test of the overall model was significant; attachment groups differed significantly in the proportions of expressed CCRT wishes. The proportions of expressed CCRT wishes were not linked with historical psychiatric status (hospital vs. high school cohort) or with current psychological functioning.

TABLE 2.

Multivariate Analysis of Covariance Comparing Attachment States of Mind With Respect to Wishes for Autonomy and Closeness

	Attachment states of mind
	Dismissinga		Secureb		Preoccupiedc
Wishes	M	SD	M	SD	M	SD	Fd	Effect size	Tukey testse
Autonomy	.20	.19	.11	.08	.05	.05	5.48	.25	A > C*
Closeness	.26	.23	.33	.22	.37	.20	0.99	.05	No differences

Note. N = 40. Multivariate analysis of covariance results for this analysis, F(4, 68) = 2.75, p < .05. Raw data for autonomy and closeness wishes were proportions, which for these analyses underwent arc-sine transformation. Means were adjusted for two covariates: psychiatric hospitalization (coded as a binary variable) and level of psychiatric symptoms. A = autonomy; C = closeness.

^an = 12.

^bn = 17.

^cn = 11.

^dDegrees of freedom = 2, 35.

^ePost hoc.

^*p < .05.

Hypothesis 1: Attachment Classification and Wishes for Autonomy

Given the statistically significant role of attachment classification in the overall model, our individual hypotheses were investigated using post hoc procedures (see Table 2). Univariate analysis of covariance was used to test the hypothesis that attachment classification would be associated with the proportion of an individual’s coded wishes for autonomy. Attachment groups differed significantly in the frequency with which they expressed wishes for autonomy, F(2, 39) = 5.48, p < .01. Post hoc testing using Tukey’s procedure revealed a significant difference between the insecure-dismissing and insecure-preoccupied groups (mean difference [40] = .15, p < .05). Results of comparisons between the insecure-dismissing and secure groups and between the secure and insecure-preoccupied groups were in the expected directions but were not statistically significant.

Hypothesis 2: Attachment Classification and Wishes for Closeness

Analysis of covariance revealed no statistically significant post hoc relationship between attachment classification and the proportion of an individual’s wishes categorized as the wish to be close (see Table 2). Thus, contrary to our hypothesis, security of attachment was not associated with the frequency of expressed wishes for closeness. Because of the lack of a significant result, further post hoc testing using Tukey’s procedure was not done.

Discussion

This study is one of the first to examine hypothesized links between states of mind with respect to attachment, assessed using the AAI, and interpersonal wishes for closeness and autonomy in young adults’ stories about current relationships, as measured by the CCRT method. As such, it is an empirical investigation of theoretical links between Main’s adaptation to adults of Bowlby’s attachment theory and the concepts of relational templates and core relational themes. The labor-intensive measures used in this study, although limiting sample size, allow for analysis of themes not always identifiable using other methods.

Our first hypothesis—that individuals would differ by attachment subgroup in the frequency with which they express wishes for autonomy—was for the most part confirmed. Those classified as insecure-dismissing on the AAI told stories about relationships that contained wishes for autonomy with significantly greater frequency than insecure-preoccupied individuals. Although the difference between insecure-dismissing and secure individuals did not reach the p = .05 level of significance, there was a trend in the expected direction. In addition, although no statistically significant difference emerged between secure and insecure-preoccupied individuals, the difference between the two means was also in the expected direction.

Analyses failed to confirm our second hypothesis that attachment classification would be related to the frequency with which individuals expressed wishes for closeness in their relationship narratives. This negative finding might be understood as confirming Bowlby’s (1979) ideas about the universality of human beings’ fundamental need for proximity to attachment figures. According to Bowlby, children’s varied attachment-related behaviors (even dismissing behaviors) reflect different strategies for maintaining proximity to attachment figures in the face of these caregivers’ different modes of response (Main, 1981). Extending this model to functioning later in life, adults might be expected to have underlying fundamental wishes for proximity to important others but to cope with expected responses from the environment using strategies based on experiences with early caregivers. Because the CCRT method allows coders to infer the presence of underlying wishes of which the speaker may not be aware, the method may be sensitive to wishes for closeness even in the narratives of individuals whose overt relationship styles are dismissing. However, we must keep in mind the possibility that our failure to find statistically significant differences in the frequency of wishes for closeness among our attachment groups may be the result of insufficient power in our sample to detect existing differences. Another possible explanation is that, because the CCRT wish to be close and proximity seeking in the attachment system are not the same constructs, their overlap is not reflected in the frequency of wishes for closeness in adult relationship narratives.

Case Examples

An inspection of narratives from insecure-preoccupied individuals reveals the presence of many wishes for closeness for every expressed wish for autonomy, reflecting a heightened concern with proximity seeking. By contrast, insecure-dismissing individuals counter almost every wish for closeness with a wish for autonomy in their narratives about their close relationships. Case examples can best illustrate the interplay of wishes for autonomy and wishes for closeness in the narratives of young adults with different attachment classifications. Excerpts from the Early Adult Close Peer Relationship Interviews of 3 study participants are presented in the following case examples. (For purposes of maintaining confidentiality, relevant details and names have been disguised.) These examples typify the narratives of people in the secure, insecure-dismissing, and insecure-preoccupied categories.

Example 1: Secure

J., a male who is classified as secure on the AAI, in discussing a best friend relationship defines closeness as a balance between relatedness and autonomy:

Interviewer. What does closeness mean to you in the relationship?

J. Um, closeness to me means being able to count on somebody but also respecting them as an individual … and not feeling entitled to own them or that they need to be there at the time when I need them, but that somebody who can listen and understand and sort of empathize and, um, someone who really does care.

A sample relationship episode from J.’s Early Adult Close Peer Relationship Interview reflects a balance of concerns about independence and closeness with his girlfriend:

J. She has a desire to be married and have children now … I tell her, I have a desire to be with her and stay with her [closeness], but I know that I’m not ready to be married [autonomy] and I know that I’m not ready to have children now, although I do want children in the future … [closeness], but no, I am not going to marry you now and no, I am not going to be the father of your children now [autonomy], so it really is happening the way I’m wanting it [autonomy].

Example 2: Insecure-Dismissing

B., a male classified as insecure-dismissing, expresses strong wishes for closeness in describing a best friend relationship. However, when asked about closeness directly, he appears to disavow those wishes and to emphasize a fundamental sense of separateness from others.

Interviewer. Do you feel like you understand each other?

B. We don’t get too mushy. We just, you know, we’ve just got a good friendship. We’ve known each other for 4 years, just the kind of thing you just don’t let go. I would miss him more than I think he’d miss me. I’d like to live with the guy …

Interviewer. What does closeness mean to you in this relationship?

B. Closeness? There isn’t—it’s not really close. You know, we’re not very close. We were roommates and, you know, we’re just drinking buddies, and that’s about as close as I get with somebody.

A sample relationship episode from B.’s Early Adult Close Peer Relationship Interview reflects his emphasis on autonomy in relation to his wife:

B. I don’t call when I come home at 4 o’clock in the morning (laughs) [autonomy]. I’m out with some of the boys and I stay at their house for a little while.… She just gets, “Did they have a phone there or something?” And it’s like, “I was at a Chinese restaurant and they had all the numbers in Chinese. I couldn’t figure it out” (laughs). Yah, that’s the one thing she hates about me because she puts the phone right next to the bed in case I ever call and I never call [autonomy].

Example 3: Insecure-Preoccupied

R., a male classified as insecure-preoccupied on the AAI, emphasizes closeness with and similarity to a best friend (K.) while downplaying any differences or mixed feelings in the relationship.

Interviewer. How would you describe your relationship with K.?

R. We’re best friends. We’re very close. He’s the only boy in his family. I’m his brother as far as he’s concerned.

Interviewer. Uh huh. What do you like about him?

R. What do I like about him? Gee, I don’t know, we enjoy doing the same things. You know. I just … I just love him. I don’t really know why …

Interviewer. Do you have any mixed feelings about him?

R. Not really. I love him, you know. I wish he didn’t do some of the things that he does, but he does.

Interviewer. Do you feel like you and he understand each other?

R. Very much so.

Interviewer. In what ways?

R. He just knows my heart and I know his.

A sample relationship episode from R.’s Early Adult Close Peer Relationship Interview reflects his need to feel close to his girlfriend:

R. Like, when things are great between her and me, I feel like a very successful man. … And when we fight and she puts me down and everything, then it affects everything in my life. And that’s the way it’s always been when I’m close to someone [closeness]. And I wish I could separate myself so that if I lost that person I would still have R. [myself] [autonomy].… But it’s just not that way. You know what I mean, I need someone in my life to love in order to make things happen properly [closeness]. I’m not a full person by myself [closeness]. I’m not. I never will be. I need someone else [closeness].

These examples illustrate coping and defending mechanisms used to maintain proximity to important others that are consistent with those described by Bowlby for securely and insecurely attached individuals. In this study, we have not examined the relational contexts in which wishes for autonomy occur. From their study of adolescents, Lamborn and Steinberg (1993) concluded that “because emotional autonomy is a relational construct, it is difficult to assess its significance and sequelae without making reference to the object from whom the adolescent is becoming autonomous” (p. 497). Among young adults, expressed wishes for autonomy may have different meanings and may represent different degrees of defensiveness depending on whether these wishes pertain to relationships perceived as supportive or unsupportive. Further research that puts wishes for autonomy and closeness into the context of different types of relationships may shed light on more subtle individual differences in core relational themes among securely and insecurely attached young adults.

In drawing conclusions from the data presented here, we must keep in mind several important limitations of this study. Our choice of methods is both a strength and a limitation. The CCRT and AAI methods of data collection and scoring are both highly labor intensive, and for this reason only 40 individuals could be included in the study. The preponderance of insecure attachment status among participants from the hospital cohort means that we could not separate the effects of cohort from those of secure versus insecure attachment classification and thus had less than adequate power to detect differences between secure and insecure individuals with respect to wishes for autonomy and closeness. We had greater power, however, to test differences between insecure-dismissing and insecure-preoccupied individuals on the outcome variables because distribution of the two insecure subcategories within each cohort was roughly equal.

The relatively small number of participants carries with it the risk that other meaningful differences exist but could not be detected because of insufficient power in this sample. For this reason, the negative finding for Hypothesis 2 discussed previously (i.e., the absence of links between attachment classification and the frequency of expressed wishes for closeness) must be seen as preliminary and in need of confirmation from further research. In addition, only a limited number of relationship episodes were coded for the majority of participants who told more than 15 relationship narratives in their interviews. It is possible that inclusion of all relationship episodes might have altered our results. In addition, it is important to note that a full analysis of the CCRT involves responses of other and responses of self. This study explored only wishes, so the full CCRT relational template was not available for analysis.

We must also consider a possible methods confound in that the AAI and Early Adult Close Peer Relationship Interviews were administered during the same laboratory session by the same interviewer, and links between narratives from the two interviews may in part be temporal or may be related to the style of the interviewer. Finally, the CCRT method relies on a set of standard categories for classifying relationship themes, a set developed using material from psychotherapy interviews of adults (Barber et al., 1998). A classification system that focuses more specifically on attachment-related themes might yield different results.

The statistical methods used in this study had the advantage of providing a conservative estimate of the relations between attachment and CCRT measures using a single parsimonious model. However, by summarizing each participant’s core relationship theme data in terms of the proportion of wishes coded for closeness or autonomy, selecting only 40 data points of almost 1,000 narrative elements coded, a significant amount of statistical power was lost.

Despite the limitations of this study, theoretically meaningful findings have emerged. First, the study demonstrates expected links between representations of past and current relationships. In our sample, young adults with different states of mind with respect to early attachment relationships speak of their adult relationships using different proportions of wishes for autonomy. By using instruments that tap into less conscious elements of relationship schemas, the study may offer empirical support for the universality of wishes for interpersonal closeness posited by both attachment and psychodynamic theorists. It also lends support to the theory that insecure-dismissing and insecure-preoccupied individuals’ different degrees of emphasis on wishes for autonomy reflect coping strategies in adult relationships based on internal working models of how their caregivers in childhood responded to their needs.

This study is one of the first to link two important domains of research: core relationship themes, as identified by the CCRT method, and attachment states of mind, as measured by the AAI. Expected links between these two measures support the validity of each. It would be useful to compare directly the links between attachment styles and core relationship themes that emerge from self-report measures with those that emerge using idiographic methods such as the AAI and the CCRT. Further research using these and other methods for measuring internal representations of relationships has the potential to elucidate elements of internal working models of relationships that are universal. In particular, the question of whether individuals with differing attachment styles differ in the frequency of their expressed wishes for closeness has important implications for both psychological theory and psychotherapy practice. If confirmed, our finding that insecure-dismissing individuals express wishes for closeness as frequently as their secure and insecure-preoccupied counterparts would guide formulations about the core psychological conflicts of insecure-dismissing individuals who present for psychotherapy. Longitudinal studies may delineate more specifically the processes by which internal working models of attachment influence the development of broader relationship schemas and styles of interpersonal functioning throughout life.

Zusammenfassung

In dieser Studie wird die Verbindung zwischen Bindungsstilen und Beziehungsschemata bei einer Stichproben von 40 jungen Erwachsenen, von denen die Hälfte als Heranwachsende in stationärer psychiatrischer Behandlung war, untersucht. Die Teilnehmer wurden über ihre engsten Beziehungen interviewt und mit Hilfe der Methode des Zentralen Beziehungs-Konflikt-Themas (ZBKT) wurden ihre Narrative über diese Beziehungen analysiert hinsichtlich der relativen Häufigkeit, mit der Wünsche nach Nähe und Autonomie in den Beziehungen zu anderen zum Ausdruck kamen. Die Teilnehmer wurden auch mit Hilfe des Bindungsinterviews für Erwachsene (Adult Attachment Interview) hinsichtlich ihrer Bindungssicherheit eingestuft. Bindungssicherheit war verbunden mit der relativen Häufigkeit, mit der die Teilnehmer Wünsche nach Autonomie in ihren Narrativen über enge Beziehungen ausdrückten, selbst dann, wenn das gegenwärtige psychologische Funktionsniveau und die persönliche Geschichte ernster pathologischer Symptome in der Adoleszenz mitberücksichtigt wurden. Bindungssicherheit zeigte keine Verbindung mit der relativen Häufigkeit, mit der dieTeilnehmer Wünsche nach Nähe ausdrückten. Die Ergebnisse der Untersuchung legen nahe, dass zentrale Beziehungswünsche nach Autonomie in spezifischer Weise mit Untertypen unsicherer Bindung verbunden sind. Sie erweitern damit die Kenntnisse über die Verbindung zwischen der Repräsentation früherer Bindungen und den Wünschen und Bedürfnissen, die bei aktuellen Beziehungsmustern mit signifikanten Anderen ihren Ausdruck finden.

Résumé

Cette étude examine des liens entre des états mentaux d’attachement et des schémas relationnels dans un échantillon de 40 jeunes adultes dont la moitié avaient été hospitalisés pour traitement psychiatrique comme adolescents. Les patients étaient interrogés sur leurs relations les plus proches et à l’aide de la méthode du Thème Relationnel Conflictuel Central, leurs narrations au sujet de ces relations étaient analysées pour la fréquence relative des désirs exprimés de proximité et d’autonomie par rapport aux autres. Les participants ont aussi passé par l’Interview d’Attachement pour Adultes et ont été classés en fonction de la sécurité d’attachement. La sécurité d’attachement était associée à la fréquence relative des désirs d’autonomie qu’exprimaient les participants dans leurs narrations sur les relations proches, même en tenant compte des niveaux actuels de fonctionnement psychologique et de l’histoire d’une psychopathologie sérieuse dans l’adolescence. La sécurité d’attachement n’était pas associée avec la fréquence relative des désirs de proximité exprimés par les participants. L’étude suggère que des désirs relationnels centraux d’autonomie sont liés spécifiquement avec des sous-types d’attachement d’insécurité. Ces résultats élargissent ce qu’on sait sur les connexions entre la représentation de relations d’attachement précoces et les désirs et besoins exprimés dans des relations actuelles avec des autres significatifs.

Resumen

Este estudio examina los vínculos que existen entre diferentes tipos de apego (attachment) y esquemas de relación en una muestra de cuarente adultos jóvenes, la mitad de quienes fueron hospitalizados para tratamiento psiquiátrico cuando eran adolescentes. Se preguntó a los participantes acerca de sus relaciones más cercanas y, usando el método del Tema Nuclera de Relaciones Conflictivas (Core Conflictual Relationship Theme, CCRT), se analizó la frecuencia relativa con la que expresaban su deseo de proximidad y autonomía en relación a otras personas.También se administró a los participantes la Entrevista de apego adulto (Adult Attachment Interview) y se los clasificó respecto a apego seguro. El apego seguro se asoció con la frecuencia relativa con la que los participantes expresaban, en sus narrativas, deseos de autonomía en sus relaciones más cercanas, aun después de tener en cuenta los niveles actuales de funcionamiento psíquico y la historia de psicopatología seria en su adolescencia. El apego seguro no estuvo asociado con la freucencia relativa con la que los participantes expresaban deseos de proximidad. El estudio sugiere que los deseos de autonomía en el núcleo relacional están relacionados con subtipos de apego inseguro. Estos hallazgos amplían lo que se conoce sobre las conexiones entre la representación de las relaciones de apego temprano y los deseos y necesidades expresados en las relaciones corrientes (current) con otros significativos.

Resumo

Este estudo analisa as relações entre as representações mentais da vinculação e os esquemas de relacionamento numa amostra de 40 jovens adultos, metade dos quais estiveram hospitalizados enquanto adolescentes para tratamento psiquiátrico. Os participantes foram entrevistados sobre os seus relacionamentos mais próximos e, usando o método “Core Conflictual Relationship Theme”, as suas narrativas sobre os relacionamentos foram analisadas em relação à frequência relativa com que expressaram desejos de proximidade e de autonomia em relação aos outros. Aos participantes foi também administrada a Entrevista de Vinculação do Adulto (Adult Attachment Interview) e foram classificados com respeito à segurança da vinculação. A segurança da vinculação estava associada com a frequência relativa com a qual os participantes expressam desejos para a autonomia nas suas narrativas sobre relacionamentos próximos, mesmo considerando os níveis de funcionamento psicológico actuais e a história de psicopatologia grave na adolescência. A segurança da vinculação não estava associada com a frequência relativa com a qual os participantes expressaram desejos de proximidade. O estudo sugere que os desejos relacionados centrais para a autonomia estão especificamente relacionados com os subtipos de vinculação insegura. Estes resultados expandem o que se conhece acerca das ligações entre a representação das relações de vinculação precoces e os desejos e necessidades expressas nos relacionamentos actuais com outros significativos.

Sommario

Il presente studio esamina la relazione tra gli stati mentali dell’attaccamento e gli schemi relazionali su un campione di 40 giovani adulti, metà dei quali adolescenti ricoverati per un trattamento psichiatrico. I partecipanti sono stati intervistati circa le loro relazioni più intime e, attraverso il metodo dei CCRT è stata misurata la frequenza con quale i soggetti nelle interviste esprimevano desideri di maggiore vicinanza o autonomia nei confronti degli altri. Ai partecipanti è stata somministrata anche l’AAI e sono stati classificati secondo la sicurezza del loro attaccamento. L’attaccamento sicuro è risultato essere associato alla frequenza con la quale erano espressi desideri circa una maggiore autonomia nei confronti delle relazioni più strette. L’attaccamento sicuro non è risultato essere associato con la frequenza con la quale i soggetti avevano espresso desideri di maggiore vicinanza nelle loro relazioni. I risultati della ricerca suggeriscono che i desideri relazionali centrali d’autonomia sono collegabili con i diversi sottotipi d’attaccamento insicuro. Questi risultati vanno ad accrescere le nostre conoscenze circa il rapporto tra le rappresentazioni dell’attaccamento e i desideri e i bisogni espressi nelle relazioni con altri significativi.