_____________________________________________________________________________________________________________ Fam Proc 22:69-83, 1983 Circumplex Model of Marital and Family Systems: Vl. Theoretical Update DAVID H. OLSON, PH.D.a CANDYCE S. RUSSELL, PH.D.b DOUGLAS H. SPRENKLE, PH.D.c aProfessor, Family Social Science, 290 McNeal Hall, University of Minnesota, St. Paul, Minnesota 55108. bAssociate Professor, Department of Family and Child Development, Kansas State University, Manhatten, Kansas 66502. cAssociate Professor, Department of Child Development and Family Studies, Purdue University, Lafayette, Indiana 47907. This paper updates the theoretical work on the Circumplex Model and provides revised and new hypotheses. Similarities and contrasts to the Beavers Systems Model are made along with comments regarding Beavers and Voeller's critique. FACES II, a newly revised assessment tool, provides both "perceived" and "ideal" family assessment that is useful empirically and clinically. The theory should be simple enough for the average therapist to understand. When important issues are clearly understood, the therapist is not distracted by clients who are experts in complexity and obfuscation. Jay Haley, Leaving Home The Purpose of this paper is to provide a theoretical update and respond to the constructive comments made by Beavers and Voeller (3). We are grateful for critics who take our work seriously enough to digest it, dissect it, and point the way for further development. We concur with Beavers (2) when he discussed "the excitement of two people who respect each other working hard on a problem.... Learning results from mutual exploration rather than indoctrination" (p. 17). The challenge burden of a developing theoretical model like the Circumplex is the continual process of clarifying, refining, and modifying it. We welcome this opportunity to defend and sharpen our thinking about various aspects of the model. Our goal is to improve, expand, and operationalize the model so that it can become more conceptually integrative, empirically valid, and clinically useful. We will begin this paper with a brief overview and theoretical update of the Circumplex Model. Then we will address the issues raised by Beavers and Voeller (3). We will conclude with some comments regarding the Beavers Systems Model and discuss some newly developed assessment tools for the Circumplex Model. Family Cohesion, Adaptability (Change), and Communication Family cohesion, adaptability, and communication are three dimensions of family behavior that emerged from a conceptual clustering of over fifty concepts developed to describe marital and family dynamics. Although some of these concepts have been used for decades (power and roles, for instance), many of the concepts have been developed recently by individuals observing problem families from a general systems perspective (pseudomutuality, double binds). After reviewing the conceptual definitions of many of these concepts, it became apparent that, despite the creative terminology, the terms were conceptually similar and dealt with highly related family processes. One family process had to do with the degree to which an individual was separated from or connected to his or her family system and was called family cohesion. The second dimension was family adaptability, which focused on the extent to which the family system was flexible and able to change. The third dimension focused on family communication between various members. Family cohesion is defined as the emotional bonding that family members have toward one another. Within the Circumplex Model, some of the specific concepts or variables that can be used to diagnose and measure the family cohesion dimensions are: emotional bonding, boundaries, coalitions, time, space, friends, decision-making, and interests and recreation. There are four levels of cohesion, ranging from disengaged (very low) to separated (low to moderate) to connected (moderate to high) to enmeshed (very high) (see Fig. 1). 1 _____________________________________________________________________________________________________________ Figure 1. Circumplex Model: Sixteen Types of Marital and Family Systems It is hypothesized that the central levels of cohesion (separated and connected) make for optimal family functioning. The extremes (disengaged or enmeshed) are generally seen as problematic. Many couples and families that come for treatment often fall into one of these extremes. When cohesion levels are high (enmeshed systems), there is overidentification, so that loyalty to and consensus within the family prevent individuation of family members. At the other extreme (disengaged systems), high levels of autonomy are encouraged, and family members "do their own thing," with limited attachment or commitment to their family. In the model's central area (separated and connected), individuals are able to experience and balance being independent from and connected to their families. Family adaptability is defined as the ability of a marital or family system to change its power structure, role relationships, and relationship rules in response to situational and developmental stress. In order to describe, measure, and diagnose couples on this dimension, a variety of concepts have been taken from several social science disciplines, with heavy reliance on family sociology. These concepts include: family power (assertiveness, control, discipline), negotiation styles, role relationships and relationship rules. The four levels of adaptability range from rigid (very low) to structured (low to moderate) to flexible (moderate to high) to chaotic (very high) (see Fig. 1). As with cohesion, it is hypothesized that central levels of adaptability (structured and flexible) are more conducive to marital and family functioning, with the extremes (rigid and chaotic) being the most problematic for families as they move through the family life cycle. Basically, adaptability focuses on the ability of the marital and family system to change. Much of the early application of 2 _____________________________________________________________________________________________________________ systems theory to families emphasized the rigidity of the family and its tendency to maintain the status quo (Haley, 14, 15, 16). Morphostasis was the system term used to describe the pattern of resistance to change, and morphogenesis was the potential to develop and grow as a system. Until the work of recent theorists (Speer, 39; Wertheim, 43, 44), the importance of potential for change was minimized. They pointed out that systems need both stability and change and it is the ability to change when appropriate that distinguishes functional couples and families from others. Family communication is the third dimension in the Circumplex Model, and it is considered a facilitating dimension. Communication is considered critical for facilitating couples and families to move on the two dimensions. Because it is a facilitating dimension, communication is not graphically included in the model along with cohesion and adaptability. Positive communication skills (i.e. empathy, reflective listening, supportive comments) enable couples and families to share with each other their changing needs and preferences as they relate to cohesion and adaptability. Negative communication skills (i.e. double messages, double binds, criticism) minimize the ability of a couple or family members to share their feelings and, thereby, restrict their movement on these dimensions. Brief Review of Theoretical Models of Family Systems This review will describe theoretical models that have focused independently on variables related to the cohesion, adaptability, and communication dimensions. Most of these models have been developed in the last five years by individuals having a systems perspective of family. Evidence regarding the value and importance of these three dimensions is the fact that these theorists and therapists quite independently concluded that the variables they selected were critical for understanding and treating marital and family systems. Table I summarizes the work of seven theorists who have worked on describing marital and family systems. Most of the recent theorizing about family dynamics and intervention has been strongly influenced by general systems theory as described by Bertalanffy (6) and applied to the family by Jackson (20), Haley (14, 15, 16), Speer (39), Hill (17), and Wertheim (43, 44). The current work has focused on describing both clinical and nonclinical families (Beavers, 2; Kantor and Lehr, 21; Lewis et al., 25) or has been concerned with clinical intervention (Benjamin, 4, 5; Constantine, 8; Epstein et al., 10). Only the early work by the psychologist Timothy Leary (24) and Talcott Parsons (33) was less influenced by the recent systems perspective, although Parson's work had a strong systems orientation in that it related the family to other social systems. Table 1 Theoretical Models Using Cohesion, Adaptability, and Communication COHESION ADAPTABILITY COMMUNICATION OTHERS Benjamin (1974 & 1977) Affiliation Interdependence Epstein, Bishop & Levin (1978) Affective Behavior Control Communication Involvement Problem-Solving Roles Affective Responspiveness French & Guidera (1974) Kantor & Lehr (1974) Affect Dimension Capacity to change Power Anxiety; Role as symptom carrier Power dimension Meaning dimension Dimension Leary (1957) and Constantine (1977) Lewis et al. (1976) and Beavers (1977) Affection Dominance Hostility Submission Closeness Power Negotiation Affect Mythology Autonomy Coalitions Parsons & Bales (1955) Expressive Role Instrumental Role Brief Description of the Circumplex Model Figure 1 illustrates the two dimensions and the four levels of each dimension. The four levels of cohesion were named (from low to high): disengaged, separated, connected, enmeshed. The four levels of adaptability were called (from low to high); rigid, structured, flexible, chaotic. In selecting the adjectives for each level, an attempt was made to avoid the traditional diagnostic labels. Combining the dimensions enables one to identify and describe 16 distinct types of marital and family systems. Although 3 _____________________________________________________________________________________________________________ it is assumed that it is possible to identify conceptually, measure empirically, and observe clinically all 16 types, it is also assumed that some of the types would occur more frequently than others. As with any circumplex model, the more central types are the most common, but it is hypothesized that couples and families having problems would fall more into the extreme types. Once the types were identified and located within the model, it then became clear that there were three basic groups of types. One group had scores at the two central levels on both dimensions (four balanced types); another group was extreme on both dimensions (four extreme types); and the third group was extreme only on one dimension (eight mid-range types). Hypotheses Derived from the Circumplex Model One of the assets of a theoretical model is that hypotheses can be deduced and tested in order to evaluate and further develop the model. The following are hypotheses derived from the model. The first general hypothesis postulates a curvilinear relationship between the dimensions of cohesion and adaptability and effective family functioning. An alternative hypothesis (Hypothesis III) relates to couples and families in which the normative expectations in the family support extreme behavior on these dimensions. I. Couples/families with balanced (two central levels) cohesion and adaptability will generally function more adequately across the family life cycle than those at the extremes of these dimensions. An important issue in the Circumplex Model relates to the definition of balance. The model postulates that a balance of the dimensions is related to more adequate family functioning. Even though a balanced family system is placed at the two central levels of the model, it should not be assumed that these families always operate in a "moderate" manner. Being balanced means that a family system can experience the extremes on the dimension when appropriate but that they do not typically function at these extremes for long periods of time. Families in the balanced area of the cohesion dimension allow family members to experience being both independent from and connected to their family. Both extremes are tolerated and expected, but the family does not continually function at the extreme. Conversely, extreme family types tend to function only at the extremes and are not encouraged to change the way they function as a family. This leads to Hypothesis II. II. Balanced family types have a larger behavioral repertoire and are more able to change compared with extreme family types. Integrating Normative Expectations of Families Although a curvilinear relationship is generally predicted, some important qualifications must be made in terms of the normative expectations and cultural bias of Hypothesis I. Hypothesis III states: III. If the normative expectations of a couple or family support behaviors extreme on one or both of the Circumplex dimensions, they will function well as long as all family members accept these expectations. The normative expectations in our culture provide two conflicting themes that can create problems for couples and families. One theme is that family members are expected to do things together as a family; yet, the second theme encourages individuals to "do their own thing." The theme of independence becomes more prominent as children approach adolescence and has taken on greater importance for increasing numbers of women in our culture. As a result, many American families find that a balance of the two themes has become a difficult issue. Families in our culture still vary greatly in the extent to which they encourage and support individual development in ways that may differ from the family's values. Although parents would prefer their children to develop values and ideas similar to theirs, most parents can enable their children to become somewhat autonomous and differentiated from the family system. A sizable minority of families, however, have normative expectations that strongly emphasize family togetherness, often at the expense of individuation from the family system. Their family norms emphasize emotional and physical togetherness, and they strive for high levels of consensus and loyalty. Some ethnic groups in this country such as Slovak-American (38), Puerto Rican (28), and Italian families (12), and religious groups such as the Amish (45) and Mormons (36) have high expectations regarding family togetherness. These expectations are also common, but less predominant, in many other American families regardless of ethnic or religious orientation. Many of these families could be described as extreme on the cohesion dimension (i.e., enmeshed), and they function well as long as all the family members are willing to go along with those expectations. This hypothesis necessitates that family members be asked about their expectations and satisfaction with their current family system. One way that can be done is to ask them first to describe their family system (perceived) and then describe how they would like it to be (ideal). The "ideal" description provides an assessment of how they would like to be, and the "perceived-ideal" discrepancy can be used as an inverse assessment of satisfaction, i.e., discrepancy relates to higher satisfaction and vice versa. The methodology for testing this hypothesis is now possible with FACES II, which is short 4 _____________________________________________________________________________________________________________ enough to be administered twice: once for "perceived" and once for "ideal." FACES II is described later in this paper. The new methodology with FACES I also facilitates the testing of another hypothesis. Hypothesis IV states: IV. Couples and families will function most adequately if there is a high level of congruence between the perceived and ideal descriptions for all family members. This hypothesis focuses on the satisfaction of family members with their current family system regardless of the type of system they have in the Circumplex Model. This hypothesis also builds on the assumption that it is less important where the family falls on the Circumplex Model (i.e. type of family system) than how they feel about the kind of family they have. In testing this hypothesis and others in the Circumplex, it is important to obtain an assessment from each family member, since family members often disagree about how they see their family (perceived level) and how they would like it to be (ideal level). Family Communication as a Facilitating Dimension Communication is a critical dimension related to the Circumplex Model because it facilitates movement on the two dimensions. This leads to two specific hypotheses linking communication to balanced types and change on cohesion and adaptability. Hypothesis V states: V. Balanced couples/families will tend to have more positive communication skills than extreme families. VI. Positive communication skills will enable balanced couples/families to change their levels of cohesion and adaptability more easily than those at the extremes. In general, positive communication skills are seen as helping marital and family systems facilitate and maintain a balance on the two dimensions. Conversely, negative communication skills impede or prevent marital and family systems from moving into the central areas and, thereby, increase the probability that extreme systems will remain extreme. Positive communication skills include the following: sending clear and congruent messages, empathy, supportive statements, and effective problem-solving skills. Conversely, negative communication skills include the following: sending incongruent and disqualifying messages, lack of empathy, nonsupportive (negative) statements, poor problem-solving skills, and paradoxical and double-binding messages (31). Although many studies have investigated communication and problem-solving skills in couples (37, 41) and families (1, 13, 34), these studies have not specifically tested the relationships of these skills to the hypotheses derived from the Circumplex Model. Family Development and the Circumplex Model VII. To deal with situational stress and developmental changes across the family life cycle, balanced families will change their cohesion and adaptability, whereas extreme families will resist change over time. This hypothesis deals with change in the family system to deal with stress or to accommodate changes in family members, particularly as family members alter their expectations. The Circumplex Model is dynamic in that it assumes that individuals and family systems will change, and it hypothesizes that change can be beneficial to the maintenance and improvement of family functioning. When one family member desires change, the family system must deal with that request. For example, increasing numbers of married women want to develop more autonomy from their husbands (cohesion dimension) and also want more power and equality in their relationships (adaptability dimension). If their husbands are unwilling to understand and change in accordance with these expectations, the marriages will probably experience increased amounts of stress. Another common example of changing expectations occurs when a child reaches adolescence. Like the wife in the previous example, adolescents often want more freedom, independence, and power in the family system. The Circumplex Model allows one to integrate systems theory with family development, a proposal made more than a decade ago by Reuben Hill (17). Building on the family development approach, as described by Hill and Rodgers (18), it is hypothesized that families must change as they deal with normal transitions in the family. It is expected, therefore, that the stage of the family life cycle and composition of the family will have considerable impact on the type of family system. It is first of all hypothesized that at any stage of the family life cycle, there will be a diversity in the types of family system as described in the Circumplex Model. In spite of this diversity, it is predicted that at different stages of the family life cycle many of the families will cluster together in some types more frequently than in others. For example, it would be predicted that premarital couples would tend toward the high range on cohesion (enmeshed and connected) and toward the lower range on adaptability (structured and rigid). In other words, they would fall into the lower right quadrant of the Circumplex Model (see Fig. 1). Table II indicates the levels of family cohesion and adaptability that are hypothesized at critical stages of the family life cycle. It also indicates how distressed and nondistressed couples and families might differ. Unfortunately, as yet there is no 5 _____________________________________________________________________________________________________________ adequate data to test these hypotheses. Table 2 Typical Types of Systems at Critical Stages of Family Life Cycle COHESION ADAPTABILITY QUADRANT OF MODEL Premarital Connected & Enmeshed Structured & Rigid Lower Right Newlyweds Connected Flexible & Chaotic Upper Right Early Marriage Separated & Connected Flexible & Structured Central Area (a) Nondistressed Separated & Connected Flexible & Structured Central Area (b) Distressed One or both spouses disengaged One or both spouses disengaged Extreme Area or enmeshed enmeshed Later Marriage Birth of First Child Connected & Enmeshed Initially Chaotic; later structured Initially Upper Right; Later & Rigid Lower Right (a) Nonproblem families Separated & Connected Flexible & Structured Central Area (b) Problem Families Disengaged & Enmeshed Chaotic & Rigid Extreme Areas (a) Nonproblem Families Primarily Separated; some Connected Flexible & Structured Central Area (b) Problem Families Enmeshed or Disengaged Rigid or Chaotic Extreme Areas (a) Nondistressed Separated & Connected Flexible & Structured Central Area (b) Distressed Disengaged or Enmeshed Rigid Lower Left or Right Child Rearing Adolescent in Family Older Couples w/o children 1. This table summarizes modal patterns for each stage. It is assumed that there will be high variability within these stages, and this summary obscures many of the complexities that need to be more fully described and empirically evaluated. 2. The concepts used to describe the four levels of cohesion and adaptability are presented in the Circumflex Model (Fig. 1). The first author, however, is currently completing a study entitled "A National Survey of One Thousand Healthy Families" that is collecting data from couples and families across the family life cycle from newlyweds to retired couples. This national data set will enable us to describe more adequately the types of family systems at various stages of the family life cycle and will investigate the types of stress and coping styles of these families. The initial data analysis should be completed in early 1984. The model is dynamic in that it assumes that changes can occur in family types over time. Families are free to move in any direction that the situation, stage of the family life cycle, or socialization of family members may require. A retrospective look at a family known to one of the authors illustrates the dynamic nature of the model. Steve and Sally were both raised in traditional homes. Three years after they were married, they became parents for the first time, and Sally resigned from her teaching job. Because of the dependency needs of their son and their own desire for mutual support in this transition period, they developed a moderately high, but not extreme, level of family cohesion. Also, their upbringing led them to be moderately low, but not rigid, on the adaptability dimension. They were comfortable with a rather traditional husband-dominant power structure and segregated role relationship, preferring the relative security of these established patterns to the ambiguities of continually negotiating them. Using the current model, we would classify their family type as structurally connected, an option that seemed to be satisfying to them at the time. When Sally's and Steve's son became a teenager, Sally started pursuing a career, and both parents experienced a good deal of "consciousness raising" about sex roles through the media and through involvement in several growth groups. Because of their son's needs for more autonomy at this age, as well as the parents' separate career interests, they began operating at a lower level of cohesiveness, moving from being connected to being more separated. Furthermore, the family power structure shifted from being husband-dominant to a more shared pattern. Sally exercises much more control in the relationship than previously, and the couple are struggling, almost on a weekly basis, to redefine the rules and role definitions that will govern their relationship. Although they occasionally yearn for the security of their earlier, more structured relationship, both find excitement and challenge in this more flexible relationship style. In short, flexibly separated best describes their current family organizational pattern. 6 _____________________________________________________________________________________________________________ This brief case history illustrates the dynamic nature of the model, which allows for movement within reasonable limits. It also seeks to recognize diverse values and legitimize the diverse organizational ideals of families. None of the four types in the inner circle is designated as "the ideal" at any given stage of the family life cycle, but all are more functional than the extreme types. Comparison of the Circumplex Model and the Beavers Systems Model Family Adaptability: Two Contrasting Models Although both the Circumplex Model and the Beavers Systems Model consider family adaptability a central dimension, they differ in the way in which the dimension is organized and presented. Although it would be tempting to resolve the differences by changes in one model, the difference can be seen as a useful testing ground to determine which approach is more conceptually, empirically, and clinically relevant. Once substantive data are provided for the validity and utility of either approach, the two models can be easily integrated. The integration should not be too difficult, since essentially there is agreement on the second dimension of family cohesion. In both models, family adaptability is built on principles derived from general system theory. The Circumplex Model builds on the concept of systemic change on the continuum from morphogenesis (continual change) to morphostasis (no change). The Beavers Systems Model builds on the concept of systemic growth on the continuum from entropy (death of system) to negentropy (system growth). Both models also utilize similar concepts to define family adaptability: i.e., leadership, style of discipline, and style of negotiation. Although Beavers and Voeller (3) criticize the use of adaptability as a curvilinear dimension in the Circumplex Model, it is actually more consistent with general systems theory than considering it a linear dimension. It is true that numerous family researchers have previously considered adaptability and flexibility to be a linear dimension. Their work, especially in the power area, is conceptually and empirically confusing (Cromwell and Olson, 9). Although early family system theorizing emphasized the morphostatic qualities of families, more recent theorizing by Speer (39), Hill (17), and Wertheim (43, 44) has criticized this morphostatic view of the family as too restrictive and misleading. Wertheim recommended that we consider the more positive and growth-oriented aspects of family systems, i.e., morphogenesis. She also emphasized, however, that too much change (morphogenesis) could lead to chaotic systems, also a problematic consequence. Extreme morphogenesis, tantamount to constant change, would preclude building up even a minimal set of common meanings, values, and expectations, essential for communication and the survival of an intimate, face-to-face group. [43, p. 365] The Circumplex Model hypothesizes, therefore, that both morphogenesis (change) and morphostasis (no change or stability) are necessary for a viable family system. Too much change, which can lead to chaos, or too little change, which can become system rigidity, are seen as disfunctional for family systems. Rather, it is the balance between these two extremes that is hypothesized as most functional to marital and family system development over time. As Wertheim (44) has stated: An ideal adaptive family system can be conceptualized as one characterized by an optimal, socio-culturally appropriate balance between stability-promoting, "self-corrective" processes, or morphostasis and change-pro-moting, "self-directive" processes, or morpho-genesis. [p. 286] The emphasis is on the term dynamic balance, which indicates that the level of change in a family system is both dynamic and seeking a balance between the extremes of chaos and rigidity. This balance does not reflect a "golden mean" or a theory of "moderation," but a sifting and movement back and forth on this dimension. Table III provides an overview of the salient differences between the two models. Although both rely on general systems theory, the Circumplex Model focuses on change, whereas the Beavers Systems Model focuses on growth. More importantly, the Circumplex Model hypothesizes that there is a curvilinear relationship between family adaptability and family functioning. The Beavers Systems Model, on the other hand, postulates a linear relationship between the two dimensions. The latter formulation hypothesizes increasing levels of family functioning, with increasing growth in the systems, both individual and family. Growth reaching to infinity is suggested, but raises the question of whether too much individual and family system growth could eventually lead to the demise of family functioning as individuals grow at different rates and in different directions, i.e., grow apart. Table 3 7 _____________________________________________________________________________________________________________ Family Adaptability: Two Contrasting Models CIRCUMPLEX MODEL BEAVERS SYSTEMS MODEL I. Conceptual Definition of Adaptability I. Relationship to Family Functioning Increasing levels of family system changes (Morphogenesis) Curvilinear Increasing levels of family system growth (Negentropy), with creasing structure and flexibility. Linear (Development and infinitely progressive) III. Relationship to General Systems Theory SYSTEM CONCEPTS TYPE OF FAMILY SYSTEM MEASURING FAMILY LEADERSHIP SYSTEM CONCEPTS TYPE OF FAMILY SYSTEM MEASURING FAMILY LEADERSHIP Morphogenesis is primary Chaotic Changing or lack of leadership Negentropy highest; (high autonomy) Most functional; (negotiation & intimacy high) Equalitarian MODERATE: Balance of Morphogenesis & Morphostasis Flexible and Structured Equalitarian Mid-Range Rigid Authoritarian LOW: Morphostasis is primary Rigid Authoritarian Entropy highest Chaotic Changing or lack ofleadership HIGH: The major distinguishing feature between the two models is the placement of "chaotic" systems. In the Circumplex Model, chaotic systems are placed at the highest level of change whereas in the Beavers Model they are placed at the lowest level of growth. Although Beavers and Voellers maintain that within the Circumplex Model chaotic families are highly adaptive, we maintain they are extreme in regard to change. It is the high level of change (morphogenesis) that is problematic for most families. It would be inappropriate to label either approach wrong, for they are based on different aspects of systems theory. To some extent, it is also a matter of choice. As Beavers has stated: Theorists have a choice in the design of their models. They can use a continuum of functioning with optimal at one extreme and most inadequate at the other or a bipolar curvilinear continuum with optimal functioning in the middle. Further, theorists can choose to use consistently similar continua throughout or to combine the two types. [3, p. 87] Differences between the two models on adaptability become seen as more arbitrary when the dimension is connected as a circle, with the chaotic and rigid ends linked. It then becomes a matter of choice where you cut the circle for descriptive purposes. It also means that findings from either model can be rather easily related to each other. Table III illustrates the similarity in concepts used to describe types of family systems (i.e., chaotic, rigid) and to measure variables related to the dimension (i.e., equalitarian, authoritarian, and changing leadership). Beavers and Voeller (3) maintain that chaotic and rigid systems are "not poles apart but close neighbors" (p. 11). They also maintain that family systems move from being chaotic to rigid and then to equalitarian. Although we would agree that chaotic and rigid systems are dynamically related, there is a lack of empirical evidence indicating that families always change from chaotic to rigid before becoming equalitarian. Both assumptions deserve further investigation. Linking chaotic and rigid systems does, however, cause problems empirically when scales are developed. More specifically, when family leadership patterns are assessed using the Beavers Systems Model, the scale moves from chaos (lack of leadership), to marked dominance (authoritarian) to equalitarian (see Table III). The difficulty with this conceptual and operational leap is indicated by the lack of inter-rater reliability on this scale (r = .45), which is problematic since it only accounts for 20 per cent of the true variance (23). In closing, the order of family adaptability dimension is an open question that needs further investigation to assess the empirical validity, theoretical salience, and clinical utility of the ordering. Family Cohesion: Redefined Without Autonomy Our original definition of family cohesion was "the emotional bonding members have with one another and the degree of 8 _____________________________________________________________________________________________________________ individual autonomy a person experiences in the family system" (28, p. 5). As Beavers and Voellers note, however, this definition confounds the family system concept of cohesion and the individual developmental concept of autonomy. They suggest that with only modest but vital changes, the Circumplex Model can approximate a systems-oriented model, already described, developed from empirical data. If one eliminates autonomy as an aspect of cohesion, a stylistic dimension results, termed centripetal/centrifugal after Erickson and Stierlin. [3, p. 89] Although we consider cohesion to be a curvilinear dimension, autonomy is more adequately described in a linear dimension. Beavers and Voeller concur: "Cohesiveness, in contrast [to autonomy], is curvilinear; that is, too little or too much cohesion is not optimal" (p. 87). Hence, putting them together in the same definition unnecessarily confuses both concepts. In order to maintain the conceptual clarity and family systems orientation, personal autonomy will be dropped from the definition. The revised definition of family cohesion will now be: "the emotional bonding members have with one another." Family cohesion, thereby, becomes more clearly a curvilinear concept that is linked to general systems theory. It is similar to Minuchin's (25, 26, and 27) disengaged-en-meshed dimension and to Stierlin's (39) and Beavers' (2) centripetal-centrifugal dimensions. Although autonomy in individual development is a less integral part of the Circumplex Model than family system dimensions are, we support any efforts to relate individual dynamics to one's current or past family system. In that regard, Newman and Craddock (30) recently completed a study relating family of origin (using the Circumplex Model) to the individual's ego defensiveness and attitude toward feminism. Their hypothesis was supported that members of families that tended to be rigidly enmeshed expressed higher levels of ego defensiveness and less favorable attitudes toward feminism in comparison with members of more balanced (central 4 cells) family structures. This type of study demonstrates one approach in the application of the Circumplex Model and its value in bridging individual and family systems concepts. Another clarifying comment by Beavers and Voeller regarding family cohesion relates to the classification of scapegoating (42) and schism and skew marriage (26). In our original paper, we identified eleven theoretical concepts related to the extremes of the family cohesion dimension. By mistake, we described scapegoating and schism and skew as characteristic of the disengaged family system, i.e., very low cohesion. These should have been described as characteristic of enmeshed family systems, i.e., very high cohesion. Comments Regarding the Beavers Systems Model Before discussing some issues that might further improve the utility of the Beavers Systems Model, it is important to emphasize the considerable conceptual similarity between the two models. We would like to follow the precept: "Seek first those things that unite you before disclosing those things that divide you." Both models are attempts at synthesizing and integrating the diversity of concepts in family therapy. Both models attempt to utilize an empirical base for further development. Both models are concerned with developing a triple threat model, i.e., one that is able to bridge research, theory, and clinical practice. The first issue we would like to raise in regard to the Beavers Systems Model is the placement of "chaos" on the adaptability dimension. As discussed earlier, this is an empirically testable hypothesis that deserves further investigation. A serious difficulty in validating the Beavers Systems Model is the operationalizing and assessment of the major variables. Although some work has been done to develop clinical rating scales, the interrater reliabilities are so low that it raises questions about the validity of their findings. For example, in the study of families described in the book, No Single Thread, by Lewis et al. (25), there was little agreement between four raters on the Global Health/Pathology Scale (r = .42). Although they indicate that this correlation is significant at the .05 level (which it is), the problem statistically is a correlation of .42 and only accounts for 17 per cent of the true variance (i.e. (.422) = .17); the remaining 83 per cent is error variance. Likewise, the interrater reliability on the Beavers-Timberlawn Evaluation Scales was only about .40, which is too low for empirical studies. Even in the most recent study by Kelsey-Smith and Beavers (22), in which they used a "Global Centripetal/Centrifugal Scale" to describe mid-range families from the Tim-berlawn Project (23), the interrater scores were too low to use empirically. The inter-rater reliability from two raters averaged about .52, which accounts for only 27 per cent of the true variance and is of questionable value for empirical validation. Inter-rater reliabilities of at least .90, which account for 81 per cent of the variance, are statistically necessary for empirical studies. This lack of reliability raises questions about the validity of the findings in their studies. That is particularly true, since 9 _____________________________________________________________________________________________________________ the clinical and normal groups demonstrated considerable empirical overlap in their study. Although this is not a problem unique to their work, it is an important issue when attempting empirically to test and validate any model. In closing, it is important to emphasize that challenges and opportunities for development abound in both models. Both are relatively new models that have much in common conceptually. Both, have had limited empirical testing and systematic evaluation to date. Both, however, are deserving of further clinical and empirical research. The challenge for both the developers and the other family professionals is to continue to search and empirically investigate the value of these models theoretically and clinically. Clinical and Research Application of Circumplex Model One of the major goals in developing the Circumplex Model was to provide a conceptual framework that could be used by marital and family therapists. More specifically, the model was designed so that it could be used in diagnosis, in establishing treatment goals, and in assessing the effectiveness of the treatment program. The model does not specify treatment techniques, since it is assumed that most techniques are a theoretical and should be chosen from those that the therapist has in his or her repertoire. Also, the model assumes that the circumplex types are dynamic, can be changed, and are not as static or stable as personality types. Diagnostic Tools Based on the Circumplex Model There are now three assessment tools that have been developed, designed specifically to assess family cohesion and family adaptability. Two are self-report measures, and the third is for clinical assessment. FACES II, acronym for Family Cohesion and Adaptability Evaluation Scales, is a newly developed 30-item, self-report scale that overcomes some of the limitations of the original 111-item FACES. FACES II has high levels of reliability (internal consistency and test retest) and validity (content and construct). It was used in a "National Survey of 1,000 Families," so national norms are available for couples across the family life cycle. Although FACES II was designed for families with children, a "Couple Form" is available for couples without children. A packet of materials containing the FACES II instrument, reliability and validity data, the national norms, and instructions on how to use the instrument are available from the first author. The nine instruments used in the National Survey are also available in a manual entitled Family Inventories. Another advantage of FACES II is that it is short enough so that it can be administered twice. First it can assess how family members currently perceive their system (perceived) and, second, how they would like it to be (ideal). The perceived-ideal discrepancy for each family member provides useful insights into how satisfied family members are with their current family system and provides information on the direction they would like to see their family change. Also, it is valuable to compare how each family member perceives his or her family and how similar family members are in terms of their ideals for their family. A particularly important advantage of assessing both the "perceived" and the "ideal" aspects of family system relates to cultural relativity of the Circumplex Model (see Hypotheses III and IV). It is possible that it is less important where the family falls in the Circumplex Model than how they feel about their levels of cohesion and adaptability. This means that it is important not only to assess how individuals perceive their family but how they would like it to be. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 10 Alexander, J. F. and Barton, C., "Behavioral Systems Therapy for Families," in D. H. Olson (ed.) Treating Relationships, Lake Mills, Iowa, Graphic Publishing, 1976. Beavers, W., Psychotherapy and Growth: A Family Systems Perspective, New York, Brunner/Mazel, 1977. Beavers, W. R. and Voeller, M. N., "Family Models: Comparing and Contrasting the Olson Circumplex Model with the Beavers Systems Model," Fam. Proc., 22, 85-98, 1983. Benjamin, Lorna S., "Structural Analysis of Social Behavior." Psychological Review, 81, 392-425, 1974. Benjamin, Lorna S., "Structural Analysis of a Family in Therapy," J. Counsel. Clin. Psychol., 45, 391-406, 1977. von Bertalanffy, L., General Systems Theory, New York, Braziller, 1968. Buckley, W., Sociology and Modern Systems Theory, Englewood Cliffs, New Jersey, Prentice Hall, 1967. Constantine, L. A., "A Verified System Theory of Human Process," Paper presented at University of Minnesota, Family Social Science, October, 1977. Cromwell, R. E. and Olson, D. H. (Eds.), Power in Families, New York, John Wiley and Halstead Press, 1975. Epstein, N. B., Bishop, D. S. and Levin, S., "The McMaster Model of Family Functioning," J. Marr. Fam. Counsel., 40, 19-31, 1978. _____________________________________________________________________________________________________________ 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 11 Feldman, L. B., "Goals of Family Therapy," J. Marr. Fam. Counsel., 2, 103-113, 1976. Goetzel, V., "Mental Illness and Cultural Beliefs in a Southern Italian Immigrant Family," Can. Psychiat. Assn. J., 18, 219-222, 1973. Guerney, L. F., "Filial Therapy Program," in D. H. Olson (ed.), Treating Relationships, Lake Mills, Iowa, Graphic Publishing, 1976. Haley, J., "Family Experiments: A New Type of Experimentation," Fam. Proc., 1, 165-293, 1962. Haley, J., Strategies of Psychotherapy. New York: Grune and Stratton, 1963. Haley, J., "Research on Family Patterns: An Instrument Measurement," Fam. Proc., 3, 41-65, 1964. Hill, R., "Modern Systems Theory and the Family: A Confrontation," Soc. Sci. Information, 10, 7-26, 1971. Hill, R. and Rodgers, R., "The Developmental Approach, in H. T. Christensen (ed.), Handbook of Marriage and the Family, Chicago, Rand McNally, 1969. Hoffman, L., Foundations of Family Therapy, New York, Basic Books, 1981. Jackson, D. D., "Family Rules," Arch. Gen. Psychiat., 12, 589-594, 1965. Kantor, D. and Lehr, W., Inside the Family, San Francisco, Jossey-Bass, 1975. Kelsey-Smith, M. and Beavers, W. R., "Family Assessment: Centripetal and Centrifugal Family Systems," Am. J. Fam. Ther., 9, 3-12, 1981. Killorin, E. and Olson, D. H., "Clinical Application of the Circumplex Model to Chemically Dependent Families," Unpublished manuscript, 1980. Leary, T., Interpersonal Diagnosis of Personality, New York, Ronald Press, 1957. Lewis, J. M., Beavers, W. R., Gossert, J. T. and Phillips, V., No Single Thread: Psychological Health in Family Systems, New York, Brunner/Mazel, 1976. Lidz, T., Cornelison, A. R., Fleck, S. and Terry, D., "The Interfamilial Environment of Schizophrenic Patients," Am. J. Psychiat., 114, 241-248, 1957. Minuchin, S., Families and Family Therapy, Cambridge, Mass., Harvard University Press, 1974. Minuchin, S., Montalvo, B., Guerney, B. G., Rossman, B. L. and Schumer, F., Families of the Slums, New York, Basic Books, 1967. Minuchin, S., Rossman, B. L. and Baker, L., Psychosomatic Families, Cambridge, Mass., Harvard University Press, 1978. Newman, L. and Craddock, A. E., "Relationships Between Family Structure, Defensiveness and Attitudes Toward Feminism, "Unpublished Manuscript, Department of Psych-ology, University of Sydney, Sydney, N.S.W. Australia, 1982. Olson, D. H., "Empirically Unbinding the Double Bind: Review of Research and Conceptual Reformulations," Fam. Proc., 11, 69-94, 1972. Olson, D. H., "Insiders' and Outsiders' Views of Relationships: Research Studies," In G.Levinger and H. L. Rausch (eds.), Close Relationships, Amherst, Mass., University of Massachusetts Press, 1977. Parsons, T. and Bales, R. F., Family Socialization and Interaction Process, Glencoe, III., Free Press, 1955. Patterson, G. R., "Parents and Teachers as Change Agents: A Social Learning Approach," in D. H. Olson (ed.), Treating Relationships, Lake Mills, Iowa, Graphic Publishing, 1976. Russell, C. S., Olson, D. H., Sprenkle, D. H. and Stilana, R. B., "From Family Symptom to Family Systems: Review of Family Therapy Research," Unpublished manuscript, 1982. Schvandeveldt, J. D., "Mormon Adolescents: Likes and Dislikes Toward Parents and Home," Adolescents, 8, 171-178, 1973. Sprenkle, D. and Olson, D. H., "Circumplex Model of Marital Systems IV: Empirical Study of Clinic and Non-Clinic Couples," J. Marr. Fam. Counsel., 4, 59-74, 1978. Stein, H. F., "The Slovak-American Swaddling Ethos': Homeostat for Family Dynamics and Cultural Continuity," Fam. Proc., 17, 31-45, 1978. Speer, D., "Family Systems: Morphostasis and Morphogenesis, or 'Is Homeostasis Enough?", Fam. Proc., 9, 259-278, 1970. Stierlin, H., Separating Parents and Adolescents, New York, Quadrangle, 1974. Vincent, J. P., Weiss, R. L. and Birchler, G. R., "A Behavioral Analysis of Problem-Solving in Distressed and Nondistressed Married and Stranger Dyads," Behav. Ther., 6, 475-487, 1975. Vogel, E. F. and Bell, N. W., "The Emotionally Disturbed Child as a Family Scapegoat," in Bell and Vogel (eds.), The Family, Glencoe, Illinois, Free Press, 1960. Wertheim, E., "Family Unit Therapy and the Science and Typology of Family Systems," Fam. Proc., 12, 361-376, 1973. Wertheim, E., "The Science and Typology of Family Systems II. Further Theoretical and Practical Considerations," Fam. Proc., 14, 285-308, 1975. Wittmer, J., "Amish Homogeneity of Parental Behavior Characteristics," Hum. Relations, 26, 143-154, 1973.