How Interpersonal Motives Clarify the Meaning of Interpersonal Behavior: A Revised Circumplex Model

http://psr.sagepub.com/Review

Personality and Social Psychology

http://psr.sagepub.com/content/10/1/67The online version of this article can be found at:

DOI: 10.1207/s15327957pspr1001_4

2006 10: 67Pers Soc Psychol RevLeonard M. Horowitz, Kelly R. Wilson, Bulent Turan, Pavel Zolotsev, Michael J. Constantino and Lynne Henderson

How Interpersonal Motives Clarify the Meaning of Interpersonal Behavior: A Revised Circumplex Model

Published by:

http://www.sagepublications.com

On behalf of:

Society for Personality and Social Psychology

can be found at:Personality and Social Psychology ReviewAdditional services and information for

http://psr.sagepub.com/cgi/alertsEmail Alerts:

http://psr.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://psr.sagepub.com/content/10/1/67.refs.htmlCitations:

What is This?

- Feb 1, 2006Version of Record >>

at University of Wisconsin-Madison on October 6, 2012psr.sagepub.comDownloaded from

http://psr.sagepub.com/

http://psr.sagepub.com/content/10/1/67

http://www.sagepublications.com

http://www.spsp.org/

http://psr.sagepub.com/cgi/alerts

http://psr.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://psr.sagepub.com/content/10/1/67.refs.html

http://psr.sagepub.com/content/10/1/67.full.pdf

http://online.sagepub.com/site/sphelp/vorhelp.xhtml


Personality and Social Psychology Review2006, Vol. 10, No. 1, 67-86

Copyright ( 2006 byLawrence Erlbaum Associates, Inc.

How Interpersonal Motives Clarify the Meaning of InterpersonalBehavior: A Revised Circumplex Model

Leonard M. Horowitz, Kelly R. Wilson, Bulent Turan, Pavel Zolotsev,Michael J. Constantino, Lynne Henderson

Department ofPsychologyStanford University

Circumplex models have organized interpersonal behavior along 2 orthogonal di-mensions-communion (which emphasizes connection between people) and agency

(which emphasizes one person's influence over the other). However, many empiri-cal studies have disconfirmed certain predictions from these models. We thereforerevised the model in 4 ways that highlight interpersonal motives. In our revision:(a) the negative pole of communion is indifference, not hostility; (b) a given behav-ior invites (not evokes) a desired reaction from the partner; (c) the complement of abehavior is a reaction that would satisfy the motive behind that behavior; (d)noncomplementary reactions induce negative affect. If the motive is unclear, themeaning of the behavior is ambiguous. This ambiguity helps explain failures in so-

cial support, miscommunications in everyday life, and features of most personalitydisorders. The model emphasizes measurable individual differences: Reactions thatare complementary for one person need not be complementary for another.

Interpersonal theories (e.g., Leary, 1957; Sullivan,1953) began to emerge in the 1940s and 1950s as a wayof explaining phenomena associated with the study ofpersonality and social interaction. These theories weretypically a reaction against prevailing theories of thetime, particularly psychoanalysis and behavioral theo-ries of learning. They were especially appealing be-cause they incorporated new insights about human in-teraction, but still managed to sidestep controversialassumptions of behaviorism and psychoanalysis.

Overthe past 50 years a variety ofinterpersonal mod-els have evolved from these early efforts (see review byKiesler, 1996). In Sullivan's (1953) early theory, for ex-ample, the "theorem ofreciprocal emotion" had empha-sized the reciprocity (or complementarity) that is evi-dent when two partners interact. In later models thistheorem became the principle of complementarity,which Kiesler (1983) formulated this way: "A person'sinterpersonal actions tend (with a probability signifi-cantly greater than chance) to initiate, invite, or evoke

Michael J. Constantino is now at the Department of Psychology,University of Massachusetts.

The authors are grateful to Laura Carstensen, Herbert Clark,Dafna Fuchs, Philippe Goldin, Sam Gosling, James Gross, MartinGrosse Holtforth, Suzanne L. Horowitz, Daniel Leising, KennethLocke, Jan Smedslund, Scott Stuart, Deborah Tatar, JimenaThomann,Larissa Tiedens, Jeanne Tsai, and three anonymous reviewers for theirhelpful comments on an earlier version of this article.

Correspondence should be addressed to: Leonard M. Horowitz,Department of Psychology, Stanford University, Stanford,CA 94305-2130. E-mail: [email protected]

from an interactant complementary responses" (pp.200-20 1). A complementary response was then definedin terms of a two-dimensional interpersonal space. Ac-cording to most interpersonal models, the interpersonalspace is organized around two orthogonal dimensionsthat are often called affiliation (the horizontal axis,which ranges from hostile to friendly behavior) anddominance (the vertical axis, which ranges from sub-missive to dominating behavior). Typically, a behaviorand its complement are said tobe (a) similar with respectto affiliation-hostility pulls for hostility, friendlinesspulls for friendliness-and (b) reciprocal with respect tocontrol-dominance pulls for submission, submissionpulls for dominance (Carson, 1969; Kiesler, 1983,1996). (For reasons explained later we shall use theolder labels communion and agency in this article in-stead of affiliation and dominance.)A prodigious literature has tested the principle of

complementarity, but the results have not been consis-tent (see, for example, Horowitz, 2004; Kiesler, 1996;Markey, Funder, & Ozer, 2003; Orford, 1986; Sadler &Woody, 2003; Strong et al., 1988; Tiedens & Fragale,2003; Tracey, 2004). In general, the principle ofcomplementarity has been confirmed reasonably wellfor behaviors on the friendly side of the interpersonalspace ("friendly dominant" behavior of one personleads to "friendly submissive" behavior ofthe other, andvice versa). But the principle does not seem to hold forbehaviors on the hostile side ofthe interpersonal space.

Orford (1986) critically reviewed the evidence forinterpersonal complementarity as described by the the-

67 at University of Wisconsin-Madison on October 6, 2012psr.sagepub.comDownloaded from


HOROWITZ ET AL.

ories of Carson (1969), Kiesler (1983), Leary (1957),and Wiggins (1982). His review demonstrated thatpeople frequently respond to "hostile-dominant acts ...with ... hostile-dominant behavior" (p. 365), not withthe theoretically expected hostile-submissive behavior.He also showed that people frequently respond to "hos-tile-submissive behavior ... with friendly-dominance"(p. 365), not with the theoretically expected hos-tile-dominance. Even the landmark study by Strong etal. (1988) showed that complementarity is conspicu-ously absent when hostile behavior is examined.Tracey (1994, 2004) reexamined the data of Strong etal. (1988)-and provided new data as well-to showthat friendly behavior has a relatively high base rate,even when the initiating behavior is hostile. This rela-tively high frequency of friendly reactions to hostilebehavior requires some explanation because it flatlycontradicts the principle of complementarity. To date,however, no explanation has been proposed that wouldexplain why complementarity does not seem to holdfor hostile behavior. This revision of the model pro-poses a solution to this theoretical problem.

The revision that we propose will account for othertheoretical difficulties as well. To begin with, all con-temporary interpersonal models take as a starting pointa focus on behavior. The principle of complementarity,as described previously, states that the behavior of oneperson invites or evokes a complementary reaction inthe other person. But the words invites and evokes aresemantically quite different. Invites points to a desireor motive within person A to obtain a particular type ofreaction from person B, whereas evokes points to auto-matic mechanisms triggered within person B that ex-plain B's reaction.We propose the following model, in which we use the

term invites as a way ofemphasizing A's motive or goalin initiating the interpersonal interaction. (For simplic-ity, we sometimes refer to a single motive or goal ener-gizing a particular behavior; but often a given behaviormay be energized and directed by multiple motives orgoals.) By emphasizing A's motive, we also acknowl-edge that B (for B's own reasons) may not accept A's in-vitation. In that case, B would frustrate A's motive, lead-ing to negative affect in A. The frustration of motivesand the resulting negative affect are usually not statedexplicitly in contemporary interpersonal models.'

'Some interpersonal theorists have implicitly acknowledged theimportant role of motives. For example, Kiesler (1996) and his col-leagues (e.g., Kiesler, Schmidt, & Wagner, 1997; Wagner, Kiesler,& Schmidt, 1995) divide the action-reaction sequence into foursteps: (a) Covert processes in person A lead to (b) person A's overtaction, which then leads to (c) covert processes in person B (per-ceptions, emotional reactions), which lead to (d) person B's overtreaction. Covert processes are said to include various subjectivestates, including goals. Thus, these authors do acknowledge therole played by goals and motives, but the theory itself primarilyemphasizes overt behaviors.

Furthermore, when the motive behind a behavior isunclear, the meaning of that behavior is unclear(McAdams, 1985). When A laughs at B's behavior, Amay be laughing with B or laughing at B. The verysame behavior may have different meanings, depend-ing on the motive behind it. This kind of ambiguity canhave numerous consequences, including disagree-ments among observers, miscommunications in every-day life, and failed attempts at social support.We therefore begin with the interpersonal motive as

the theoretical starting point that distinguishes ourmodel from other interpersonal models. In the first sec-tion we examine the organization of motivational con-structs. Then we show how the very same behaviormay arise from alternate motives, causing ambiguity inthe meaning of that behavior. Then we turn to interper-sonal behavior and show how an analysis of behaviorsexposes broad motivational themes or dimensions thatdrive those behaviors. We also reinterpret the principleof complementarity, arguing that one person's behav-ior invites a desired reaction, which constitutes thecomplement of that behavior. (A noncomplementaryreaction is one that frustrates the desire, leading to neg-ative affect.) Then we apply the model to clarifymiscommunications in everyday life, failed attempts atsocial support, and distinctions important to the studyof personality. Finally, we show how the features ofmany personality disorders may be organized around acharacteristic interpersonal motive.

Basic Postulates of the Model

Interpersonal Motives

PI. Interpersonal motives may be organizedhierarchically.

Motivational constructs vary in their breadth orlevel of abstraction. A broad desire, such as a desire forintimacy or a desire for friendships, is of a higher orderthan a narrow desire, such as a desire to spend timewith a romantic partner. That desire, in turn, is of ahigher order than a still narrower desire, such as a de-sire to make a date with a particular person. These lev-els of abstraction may be conceptualized hierarchically(Emmons, 1989). That is, a desire for intimacy consti-tutes a superordinate (more abstract) category, whichsubsumes narrower categories, and those categorieseach subsume still narrower categories. The term mo-tive usually designates a relatively high level of ab-straction (e.g., a desire for intimacy or autonomy),whereas the term goal usually designates a relativelynarrow, more specific category. Intermediate levels ofabstraction are sometimes called personal strivings orcurrent concerns. This way of conceptualizing motiva-tion is common in contemporary psychology (Austin



INTERPERSONAL MOTIVES AND MEANING OF BEHAVIOR

& Vancouver, 1996; Cantor & Kihlstrom, 1987;Cropanzano, James, & Citera, 1992; Emmons, 1989;Klinger, 1987; Little, 1983; McAdams, 1985).

P2. Interpersonal motives fall into two broad,superordinate categories.

When interpersonal motives are conceptualized thisway, we commonly assume that two very broad, ab-stract categories are at the top of the hierarchy, namely,communion and agency (Bakan, 1966; see alsoMcAdams, 1985; Saucier & Goldberg, 1996). A com-munal motive is a motive for a connection with one ormore others; it is a motive to participate in a larger un-ion with other people. An agentic motive, on the otherhand, emphasizes the self as a distinct unit; it focuseson the person's own individual influence, control, ormastery over the self, other people, and the environ-ment. Bakan (1966) expressed the distinction this way:

I have adopted the terms "agency" and "communion"to characterize two fundamental modalities in the ex-istence of living forms, agency for the existence of anorganism as an individual, and communion for theparticipation of the individual in some larger organismof which the individual is a part. Agency manifests it-self in self-protection, self-assertion, and self-expan-sion; communion manifests itself in the sense of beingat one with other organisms. Agency manifests itselfin the formation of separations; communion in thelack of separations. ... Agency manifests itself in theurge to master; communion in noncontractual cooper-ation. (pp. 14-15)

P3. The first expressions of communal andagentic motivation appear early in infancy.

According to attachment theorists, the concept of anattachment system involves inherent motivation(Cassidy, 1999, p. 5). Attachment theorists (seeCassidy & Shaver, 1999) have highlighted the twobroad categories of motives. The infant's attachmentsystem keeps the child motivated to stay close and con-nected to the adult, an early manifestation of a commu-nal motive that increases the child's chances of surviv-ing infancy. Then, as children come to feel sufficientlysecure about the caretaker's availability at times ofneed, they seek to separate from the caretaker and au-tonomously explore the environment (Mahler, Pine, &Bergman, 1975). The motive to separate and explore isthus an early manifestation of an agentic motive. Overtime each motive becomes differentiated into subordi-nate motives. Communion comes to include motivessuch as intimacy, sociability, and belonging to groups.Agency comes to include motives such as autonomy,achievement, and control.

P4. Generally speaking, interpersonal behaviorsare motivated.

When person A initiates an interaction with personB, we assume that A's behavior is purposeful(goal-directed). The person is not necessarily con-scious of the goal (or motive), and for any given per-son, interpersonal goals may range in importancefrom trivial to vital. When an important goal is satis-fied, the person experiences a positive emotion; whenit is frustrated, the person experiences a negativeemotion (Lazarus, 1991). Furthermore, a particulargoal or motive varies in importance from situation tosituation and from time to time within a person. Forexample, a desire to acquire new friends has a higherpriority during some periods of life than others(Carstensen, Isaacowitz, & Charles, 1999). On aver-age across time and situations, some goals and mo-tives are more important to one person than to an-other; they constitute "thematic lines" in the person'slife story (McAdams 1985, p. 62). Being admired, forexample, might be vitally important to one person butrelatively unimportant to another.

P5. A particular behavior may stem from a va-riety of motives, which lend meaning to thatbehavior.

A person who enjoys giving advice may do so formore than one reason-displaying competence andknowledge (agentic), influencing others (agentic),connecting with others (communal). Similarly, a per-son who loves a particular sport may enjoy playing thatsport for various reasons-belonging to the team(communal), displaying a skill (agentic), winningcompetitions (agentic), maintaining a family tradition(communal), and so on. The meaning of a behavior de-pends on all of the motives behind it. Sometimes, how-ever, it is convenient theoretically to identify and ex-amine the one or two most salient motives that energizeand direct the behavior in question.

P6. Coexisting motives may be behaviorallycompatible, or they may conflict.

A person who gives advice to another person maybe trying to satisfy (a) a motive to influence the otherperson, as well as (b) a motive to connect with thatperson. These two motives are behaviorally compati-ble. Sometimes, however, coexisting motives conflictbehaviorally. For example, an agentic motive mayconflict with a communal motive: Suppose a womancompeted with a good friend for an elective officeand won the election. In the process of satisfying herown agentic motive, she may have disappointed andalienated her friend, thereby jeopardizing the friend-ship. Exline and Lobel (1999) discussed this type of



HOROWITZ ET AL.

conflict, showing how strivings for personal masteryand superiority can clash with strivings for commu-nion. For this reason, people sometimes conceal theirsuccess or downplay its significance (Brigham,Kelso, Jackson, & Smith, 1997). Similarly, academi-cally gifted students frequently conceal their superiorabilities from peers through a variety of "camouflag-ing" strategies (Arroyo & Zigler, 1995; Cross, Cole-man, & Terhaar-Yonders, 1991). The conflict seemsespecially salient among people with strong commu-nal needs (Santor & Zuroff, 1997).

When motives conflict in this way, an event that sat-isfies one of the motives may frustrate the other (con-flicting) motive. Psychotherapists are frequently pre-sented with dilemmas of this type. Suppose, forexample, that a greatly overweight adult client invites atherapist to address him by a mocking or teasing nick-name (e.g., "Call me Tiny; everybody does"). Yieldingto a request of this type might satisfy one motive (e.g.,a desire to feel accepted or loved), but it might simulta-neously frustrate a desire to be addressed as a respectedadult. In such cases, it is often possible to sidestep thedilemma altogether by systematically investigating themeaning of each alternative and examining the prosand cons of each. It may be more meaningful (and re-spectful) to the person to have the conflict investigated,discussed, and related to other aspects of the person'slife than to have one motive satisfied and the other mo-tive tacitly frustrated.

Conflicts between a communal motive and anagentic motive seem to be common in everyday life. Itshould be noted, however, that two communal motivesmay also conflict-e.g., a desire to be part of an ad-mired group may conflict with a desire to remainfriends with a nongroup member. Likewise, twoagentic motives may conflict-e.g., a desire to hold amanagerial position of power may conflict with a de-sire to express personal opinions freely. Conflicts suchas these would follow the same principles as those be-tween communal and agentic motives.

P7. Interpersonal goals and motives can bemeasured.

Later we describe two self-report measures con-structed in recent years for assessing interpersonalgoals and demonstrating their stability over time(Grosse Holtforth & Grawe, 2000, 2002; Locke, 2000).They also provide a way of aggregating interpersonalgoals into higher order motives.

Ambiguity of Behavior

A behavior can be ambiguous. That is, the verysame behavior may arise from different motives. Thenext two propositions concern that ambiguity.

P8. When the motive or motives behind an inter-personal behavior are unknown or unclear, thebehavior is ambiguous.

Suppose we know a man's goal, namely, to callMaria for a date this weekend. Can we infer the higherorder motive from which this goal stems? If two menboth plan to call Maria for a date this weekend, are theyboth necessarily trying to satisfy a higher order inti-macy motive? Not necessarily. As shown in Figure 1,one may be seeking intimacy (a communal motive),whereas the other may be seeking the respect, admira-tion, or envy of his friends (an agentic motive). Thus,the goal-directed act itself may be unclear. Only whenwe can locate the behavior in the person's hierarchy ofmotives do we understand its meaning.

If someone sitting next to us on an airplane startedchatting amiably, we might assume a communal mo-tive (to socialize). However, if the person then asked inall seriousness, "Have you heard the Word of the Lordtoday?" a listener might revise his or her interpretationof the person's chattiness and perceive an agentic goal(to proselytize or influence). Many behaviors are am-biguous in this way: When A spills something on B, Bmust judge whether the act was accidental(noninterpersonal) or intentional (interpersonal).When A laughs at B, B must decide whether A's laugh-ter is communal (e.g., laughing with B) ornoncommunal (e.g., laughing at B). When A advises Bto modify B's behavior, B must decide whetherA is be-ing communal (kind, friendly), agentic (controlling,critical), or a combination of both. Inferring anotherperson's intention is an extremely common human ac-tivity that is evident from the first year of life (e.g.,Tomasello, 1995). Inferences about another person'sintentions are also evident in the literature of personperception and impression formation, as evident in theearly writings of Asch (1946) and Heider (1944).

Psychiatric symptoms sometimes provide excellentexamples of ambiguous behavior (Caspar, 1995, 1997;Grawe, 2003). For example, an individual with an-orexia nervosa might aspire to lose weight, but themeaning of the person's behavior (self-starvation)would not be clear until we could locate it in the hierar-chy of motives. For one person, self-starvation might

Communion Motive

i fIntimaT Motive

Desire to fee close to anattractive woman

Goal to call ria for a date

Agency Motive

Motive to affirm self

Desire to displaymasculine superiority

Goal to call Maria for a date

Figure 1. Two possible motive hierarchies for the same goal.




have an agentic meaning: a motive to exercise auton-omy -e desire to display self-control -> desire to re-frain from eating -+ goal to eat nothing but lettuce thisweekend. For another person, however, self-starvationmight have a communal meaning: a motive to be nur-tured by the family -> desire to seem small, thin, andfrail -* desire to lose weight -* goal to eat nothing butlettuce this weekend. (A blend of the two is also possi-ble.) Some theories of anorexia nervosa have empha-sized a communal motive to maintain family harmony(e.g., Minuchin, Rosman, & Baker, 1978), whereasother theories have emphasized an agentic motive todisplay self-control and strength (e.g., Bruch, 1973).Although the personal striving to lose weight is similarin both cases, the meaning of self-starvation is notclear until we can describe the broader motive fromwhich it arose.

P9. Because ambiguous behavior can be inter-preted in different ways, it can lead to amiscommunication between interacting partners.

Examples ofmiscommunications abound in the liter-ature on social support. Difficulties can arise whenever alistener does not understand the speaker's wishes. Aspeaker may want communal support (e.g., empathy),but the listener might offer agentic support (e.g., influ-ence through advice). People report dissatisfactionwhen a listener's reaction does not match their own de-sired form of support (Horowitz et al., 2001, Study 3).Bereaved people and people with chronic illnesses areoften subjected to unwanted advice from well-meaningfriends (e.g., Lehman, Ellard, & Wortman, 1986;Lehman & Hemphill, 1990). Telling a parent of a de-ceased child to "consider yourself lucky that you canstill have other children" dismisses the person's pro-found sense of loss and instead burdens the person withunwanted advice (Lehman & Hemphill, 1990). Ms. Amay believe that she is forging intimacy with Ms. Bwhen she says, "Tell me, my dear, have you always had aweight problem?" but Ms. B may interpret her questionas a hostile criticism. We analyze this problem moreclosely in the following section.

Interpersonal Behavior

The next five propositions describe an empiricalprocedure that reveals the motivational meaning of aninterpersonal behavior. This procedure will help uspinpoint the source of the ambiguity that producesmiscommunications. It will also highlight ways inwhich our revised model differs from earlier circum-plex models.

PIO. Interpersonal behavior bay be representedgraphically within two prominent dimensions of

meaning that correspond to communion andagency.

Interpersonal behavior includes behaviors that fitthe frame "person A [does this to] person B:" "Adominates B," "A blames B," "A ignores B," "Ayields to B," and so on. A variety of data-reductionmethods (such as principal components analysis)have been used to expose the most salient dimensionsof meaning that run through the domain of interper-sonal behaviors. Numerous studies have identifiedtwo particularly salient dimensions such as thoseshown in Figure 2 (see summaries by Horowitz,2004; Kiesler, 1996; Wiggins, 1991, 1996). Most in-vestigators have concluded that these two salient di-mensions provide a good first approximation towardexplaining variation in the meaning of interpersonalbehaviors. The exact amount of variance explaineddepends on the particular scaling method used, theitems selected for study, and the context of the study.Dimensions beyond the first two would certainly addnuance to the meaning of the behaviors, but the firsttwo dimensions seem to provide an adequate first ap-proximation. These two dimensions are thereforeused as a heuristic to help us conceptualize the mean-ing of different interpersonal behaviors.

The first dimension (represented by the x axis) hasbeen called connectedness, affiliation, love, warmth, ornurturance; we use the superordinate term communionfor this axis. The second dimension (represented by theorthogonal y axis) has been called influence, control,dominance, power, or status; we use the superordinateterm agency for that axis. Thus, communion, as thehorizontal dimension, ranges in meaning from "beingdisconnected, indifferent, or distant" to "being con-nected, loving, or close." (Please note that we label thenegative pole of communion disconnected behavior,not hostile behavior. In our view, indifference, ratherthan hostility, is the polar opposite of love, and in alater section we shall explain why we emphasize thispoint.) Agency, as the vertical dimension, ranges inmeaning from "yielding, submitting, or relinquishingcontrol" to "influencing, controlling, or dominating."

YDominaft

S

Scold 0

Ignore

App"Soe

* Pro~t

Love

x * communionY * agency

S Trust

Submit

Figure 2. Interpersonal behaviors placed in a two-dimensionalspace.

71

0 w A

r

at University of Wisconsin-Madison on October 6, 2012psr.sagepub.comDownloaded from


HOROWITZ ET AL.

Each scaling procedure provides a pair of coordi-nates for every behavior to describe that behavior's loca-tion on each dimension. "A protects B" is positive incommunion and positive in agency; "A scolds B" is neg-ative in communion but positive in agency. Behaviorsthat are geometrically close to one another (similar co-ordinates on both dimensions) have similar meanings,so they are positively correlated: That is, people whostrongly exhibit one behavior also tend to exhibit othernearby behaviors. Behaviors that are diametrically op-posite one another have contrasting meanings, so theyare negatively correlated. For example, behaviors thattypically accompany dominating behavior rarely ac-company submissive behavior. Thus, the proximity oftwo behaviors tells us about their degree of correlation(Gurtman, 1994; Gurtman & Bakarishnan, 1998).

Other interpersonal domains have also been scaled.For example, interpersonal traits (e.g., assertive,friendly) are often regarded as labels for an aggregateof co-occurring behaviors and internal experience, par-ticularly motives and goals (e.g., Alston, 1975; Buss &Craik, 1983). Therefore, the domain of traits (aggre-gates of behavior) might also be organized in terms ofthe two underlying classes of interpersonal motives.Wiggins (1979) confirmed this hypothesis. After locat-ing interpersonal traits in a two-dimensional space, hedivided the two-dimensional space into eight regions(or octants), as shown schematically in Figure 3. Oneoctant, for example, contained traits that are high inagency but neutral in communion (e.g., assertive,self-confident, forceful). The composite of traits in thatoctant was called assured-dominant, a higher ordertrait than a one-word trait such as "assertive." Anotheroctant contained traits that are high in communion butneutral in agency (e.g., kind, sympathetic, nurturant).That higher order trait was called warm-agreeable. Athird octant contained traits that are high on both di-mensions (e.g., jovial, enthusiastic, extraverted); it was

(Octant 1)Assured-dominant

'Octant 8)Gregrious-eXtraVeftad

(Octant 2)Arrogant.-calculating

(Octant.3)*Coldhearted-unsympathetic

(Octant 4)Aloof-

introverted

(Octaut 7)Warm-agreeable

(Octant6)Unassuming-ingenuous

(Octant 5)Unassured-submissive

Figure 3. Two-dimensional graph divided into eight octants (tocreate eight personality scales).

called gregarious-extraverted. In this way, Wigginscreated eight separate scales (each located within thetwo-dimensional interpersonal space) to assess each ofthe eight higher order trait octants. Therefore, we canuse Wiggins's eight scales to rate a person (a) on a nar-row trait (such as assertive), (b) on a higher order trait(such as assured-dominant), or (c) on the highest ordertrait, agentic, by appropriately weighting and combin-ing scores on all scales that assess agency. In this way,we can identify different levels of abstraction for traitconstructs, comparable to those portrayed in Figure 1for motivational constructs.A similar procedure has been used to scale and mea-

sure interpersonal goals. Locke (2000) constructed aself-report measure containing 64 items (goals) thatare described in two dimensions corresponding tocommunion and agency. Every item names a particulargoal, and the 64 items (goals) are organized into eightscales that assess higher order categories of motiva-tion. The scales also possess reasonably high test-re-test reliability, so individual differences seem to be rel-atively stable across time.Why are communion and agency so salient as orga-

nizing dimensions in all of these domains (behaviors,traits, and goals)? Our model assumes that communaland agentic motives give rise to all three. That is, com-munion and agency apparently constitute fundamentaldimensions of meaning because they reflect two broadtasks in life that every person encounters from child-hood on (cf. Angyal, 1941; Erikson, 1963). From anevolutionary perspective, Hogan and Roberts (2000)suggested that they reflect the two principal evolution-ary challenges of social adaptation, namely, "gettingalong" (communion) and "getting ahead" (agency).Other writers have described the two tasks as (a) con-necting with other people to form a larger protectivecommunity and (b) achieving a reasonably stable andrealistic sense of one's own competence and control,which helps facilitate instrumental action (e.g., Blatt,1990; Horowitz, 2004).

PI1. Sometimes an interpersonal behavior can-not be located unambiguously in the interper-sonal space.

Earlier interpersonal models have implied that a be-havior occupies a unique position in the interpersonalspace (e.g., Kiesler, 1996). As we have noted, however,behavior can be ambiguous. When a wife says to herhusband, "Let's straighten up before we go out," hergoal may be primarily communal (a relatively high xcoordinate, reflecting a desire for closeness throughteamwork) or primarily agentic (a relatively high y co-ordinate, reflecting a desire to influence her husband'sbehavior). Therefore, the husband and wife may per-ceive the wife's remark very differently. To say that abehavior is ambiguous means that the coordinates of




the behavior on the graph are unknown because the un-derlying motives are unclear. When two people inter-pret the same behavior differently, that difference is apotential basis for misunderstanding.A difference in perspective frequently arises in trou-

bled marriages (Fincham & Beach, 1999): A husbandreturns home late from work, and his wife argues that"he thinks only about himself and his needs." Suchconflict-promoting attributions impair problem solv-ing, increase negative affect, and lower marital satis-faction (Bradbury & Fincham, 1992; Fincham &Bradbury, 1992; Karney & Bradbury, 1997). Finchamand Beach (1999) commented that marital conflict isusually conceptualized in terms of each partner'sgoals, yet "research ... has paid little attention to thegoal construct" (p. 61).

In brief, then, two people may have different inter-pretations of the same interpersonal interaction. WhenA and B interact, the behavior ofeach may be describeddifferently by A, by B, and by an outside observer.

P12. The "complement" of a behavior is the re-action that would satisfy the motive behind it.

When we know the motive behind a behavior, wecan locate the behavior graphically. Then we canspecify the person's desired reaction (see Figure 4).When person A unambiguously dominates person B,A wants B to yield. When A makes an unambiguousbid for intimacy, A wants B to reciprocate closeness.An unambiguous interpersonal behavior invites a de-sired class of reactions (which the partner may ormay not provide), and those desired reactions consti-tute the complement. This interpretation of a behav-ior's complement differs from that of othercircumplex models, which primarily focus on pairs ofbehaviors that uniquely evoke each other. An ambigu-ous behavior, however, has different possible comple-ments, depending on the motive behind it.

What is the formal relation between an unambigu-ous interpersonal behavior and its complement? Likeearlier circumplex models, we say that an unambigu-

AGENCY

A givesfriendly advice

i

CB willinglyaccepts advice

Figure 4. A pair of complementary bel

COMMUNION

haviors.

ous interpersonal behavior and its complement aresimilar with respect to the horizontal axis (connectioninvites connection, detachment invites detachment)and reciprocal with respect to the vertical axis (influ-ence invites deference, and deference invites influ-ence). Therefore, when person A gives friendly advice(warm influence) to person B, A desires warm accep-tance of the advice, as shown in Figure 4. When A tellsB to "leave me alone" (detached influence), A wouldlike B to comply by withdrawing (detached deference).When A tells B that he feels "stuck" over a problem(warm deference), A is inviting B to come to A's rescue(warm influence).Now we can explain why we label the negative

pole of the x dimension disconnected or indifferent,rather than hostile. Whereas disconnectednessseems to invite disconnectedness, hostile behaviordoes not seem to invite hostile behavior. Hostile be-havior, in our view, reflects anger; and anger (nega-tive affect), according to our model, indicates thatan important motive has been frustrated. A personwho wants to be left alone might become angry if apartner kept offering love (Moskowitz & Cote,1995), and a person who wants intimacy might be-come angry if a partner kept being unresponsive.Thus, the assumption of earlier models-that is,that hostile behavior "leads to" hostile behav-ior-would seem to hold only when both partnersfrustrate each other's motives.

An interesting case arises when B reacts to A'sdominance with dominance (or to A's deference withdeference), thereby frustrating A's desire. If two peo-ple keep trying to influence each other (and neitheryields), they may become stuck in a power struggle inwhich neither satisfies the goal of the other. Two peo-ple may also become frustrated (and irritated) if eachkeeps deferring to the other (e.g., "After you, my dearAlphonse." "No, dear sir, after you!" "No, no, I'll fol-low you.").

P13. An interpersonal interaction is affected byeach person's expectations ofthe other's motives.

Holmes (2002) astutely emphasized the impor-tance of each person's expectations of the other whenthey interact. In Holmes' view, expectations are thebuilding blocks of social cognitions. One class of ex-pectations is given special prominence, namely, ex-pectations about a partner's motives. The role of mo-tivational expectations in our revised model becomesevident when we examine the self and self-protectivemotives that arise to protect the person from a vulner-able sense of self.

P14. A partner may interpret a person's motivecorrectly, but choose to respond with anoncomplement.



HOROWITZ ET AL.

Now we can explain why empirical research has sooften failed to support the principle ofcomplementarity.To begin with, recent data have generally confirmed theprinciple of complementarity in cooperative interac-tions (e.g., Sadler & Woody, 2003). Using structuralequation modeling, these authors showed that one per-son's dominating behavior tends to be followed byyielding behavior in the other partner, and vice versa.They also showed that the behavior of one person tendsto match the behavior of the other along the dimensionof friendliness (connectedness).

Similarly, Tiedens and Fragale (2003) providedclear evidence for complementarity in a simple coop-erative interaction. They studied people's nonverbalbehavior and confirmed complementarity along theagentic dimension. In their experiment, each partici-pant worked with a partner-confederate who adoptedone of three physical postures during the task-an"expansive" (domineering) posture, a neutral posture,or a "constricted" (yielding) posture. Expansive con-federates draped one arm over the back of an adjacentchair and rested their right foot on their left thigh,making their right knee protrude. Constricted confed-erates sat slightly slouched, with their legs togetherand their hands in their laps. Each participant's "ex-pansiveness" was then measured (with a ruler) fromthe videotape. Participants working with an expansiveconfederate became increasingly constricted duringthe session, whereas those with a constricted confed-erate became increasingly expansive. In a second ex-periment, participants believed that their skin conduc-tance was being measured, and the apparatus requiredthem to assume a particular posture. Their postureand the confederate's posture were manipulated ex-perimentally to be expansive or constricted. Partici-pants indicated that they liked the confederate betterand felt more comfortable when the confederate'sposture complemented their own.

However, many studies reviewed by Orford (1986)have disconfirmed the principle of complementarity,especially when the initiating behavior was not com-munal. Consider the study by Strong et al. (1988) as anexample. That study confirmed complementarity insome experimental conditions but not in others. The in-vestigators divided the interpersonal space into eightoctants and trained female confederates or actresses toenact behavior in one of the octants. Eighty female stu-dents each interacted with one of the confederates, cre-ating a story together using pictures selected from theThematic Apperception Test (Murray, 1938). Duringthe interaction each confederate enacted a preassignedrole, and every interaction was videotaped and tran-scribed. When the confederate's behavior was friendly(friendly-dominant or friendly-yielding), the partici-pant's behavior was most often complementary. How-ever, when the confederate's behavior was not friendly,the partner's behavior was often not complementary.

To cite one example, when the confederate bragged(detached-controlling behavior), the participant oftenreacted with a connecting behavior-as though theparticipant were trying to transform a cool disengage-ment into a warmer interaction. In brief, people do notreact automatically to disengaged control with disen-gaged compliance.How shall we explain this noncomplementarity? In

our view, partners may have their own salient interper-sonal goals that lead them to refuse a person's invitation.Tracey (1994; also see Tracey, 2004) showed that partic-ipants express connected (warm) reactions consider-ably more often than disconnected reactions-evenwhen the initiating behavior invites a disconnected reac-tion. Our model suggests that partners with strong com-munal goals wouldbe especially disinclined to provide acomplementary reaction. Thus, our revised model doesnot predict inevitable complementarity. In some situa-tions, an invited reaction would conflict with other sa-lient motives, and a noncomplementary reaction mighteven be the modal reaction.

When a reaction is noncomplementary, it shouldfrustrate the person's interpersonal goal. And if thatgoal is important to the person, it should lead to nega-tive affect. Shechtman (2002) tested this hypothesis.She noted that assertive people describe themselves us-ing traits that suggest a strong agentic motive. Self-de-scriptive traits such as dominant, forceful, firm, andcontrolling suggest a strong motive to influence others.Therefore, any partner who tried to dominate an asser-tive person would frustrate the person's agentic motive,and that frustration should induce anger.

In Shechtman's (2002) study, unacquainted partici-pants were introduced and told that they would worktogether on a problem-solving task. They sat in adja-cent rooms, each at a computer, and they were told thatthey would communicate by computer. Their task, theDesert Survival Problem, required them to imaginethemselves to be copilots of an airplane that hadcrash-landed in the desert. They were to rank order 12objects for survival value in the desert (e.g., a flash-light, a quart of water). Each participant was to ex-change initial rankings with the partner and discusseach object. Half of the participants were assertive(they had high scores on a test of assertiveness), andthe others were nonassertive (they had lower scores).

In actual fact, the communications that each partici-pant received came, not from each other, but from acomputerized script. These communications recom-mended changes in the participant's rankings-for ex-ample, that the participant's fourth-ranked object bemoved up to Rank 1. The preprogrammed script alsoprovided reasons for recommending each change. Thelanguage of the message was manipulated experimen-tally. In one condition, the partner's words seemeddominating (e.g., "The flashlight is the only reliablenight-signaling device. Put it higher."). In the other




condition, the partner did not seem dominating (e.g.,"Do you think the flashlight should maybe be ratedhigher? It may be a reliable night-signaling device.").Thus, an assertive or nonassertive participant workedwith an apparently dominating or nondominating part-ner, producing four experimental conditions. From thetranscript of each participant's statements during thesession, every hostile (angry) comment was identified.On average, assertive participants working with a dom-inating partner produced more than 6 times the amountof hostility produced by participants in any other con-dition. Apparently, the assertive person's self-affirm-ing motive to influence the partner was frustrated bythe dominating partner.

Would assertive participants make hostile commentsifthey believed that their "dominating" partnerwas not ahuman being? In the four conditions described previ-ously, the participants believed that they were interact-ing with a person. The same four experimental condi-tions were therefore repeated with one simple change ofdetail. In these conditions the participants were told thatthey were interacting with a computer program that wascontinually updating its internal norms in search of anoptimal solution to the problem. The participants rarelyproduced hostile comments in these conditions, evenwhen an assertive participant was working with a domi-nating computer partner. Apparently, an interpersonalmotive gets aroused by a human being, not by a com-puter. Naturally, computers can frustrate other task-re-lated motives (e.g., completing the task), thereby induc-ing negative affect; but computers do not seem tofrustrate the interpersonal motive that is implied by thetrait "assertive."

P15. Abidforsocial supportis often ambiguous.

The principles described previously also help clar-ify issues of social support. When a speaker describes aproblem to a listener, we assume that the speaker wantssomething from the listener (advice, compassion, helpregulating an emotion). A genuinely supportive reac-tion is one that satisfies that desire. Therefore, a lis-tener has to determine what the problem teller wantsand react in a way that satisfies that want. Stressfulproblematic situations may be classified broadly intotwo categories that correspond to communion andagency (O'Brien & DeLongis, 1996). Some situationsleave people feeling rejected, abandoned, ostracized,or isolated, and in those cases the person may want tofeel securely reconnected, understood, or loved. Othersituations leave people feeling like a failure (inept,powerless, inferior), and in those cases the person maywant to feel more empowered (able to perform,achieve, or do). When a person's sense of competenceis at stake, the person may want tactful advice to helprestore a sense of control or efficacy (P. Brown &Levinson, 1987; Goldsmith, 1994). According to

O'Brien and DeLongis (1996), communal problemsinvolve "strivings for love, intimacy, friendship, affilia-tion, emotional relatedness, belongingness, mutuality,group cohesion, communality, and relation mainte-nance," whereas agentic problems involve "strivingsfor mastery, power, achievement, work performance,and instrumental task completion" (p. 80). Some prob-lems, of course, reflect a combination of both.

Typical reactions to a reported problem also fall intotwo broad categories. Cobb (1976), writing about so-cial support, differentiated between "emotional sup-port," which provides connection, affiliation, orwarmth, and "esteem support," which provides greaterefficacy or agency. Cutrona and Suhr (1992, 1994) alsodistinguished "emotionally supportive" forms of socialsupport from "action facilitating" forms. Trobst (1999;Wiggins & Trobst, 1997) scaled a sample of supportivereactions and empirically derived the familiar two-di-mensional structure organized around communion andagency. Thus, communal problems seem to call forcommunal support, whereas agentic problems seem tocall for agentic support (Horowitz et al., 2001). To helpa person overcome a sense of isolation and feel con-nected, a listener might show compassion or displayunderstanding. To help a person feel more empowered,a listener might suggest, demonstrate, or help the per-son discover an effective solution to the problem. Jef-ferson and Lee (1992) tape-recorded dyadic conversa-tions among coworkers in the workplace and identifiedtwo types of conversations. In one, a speaker describedan agentic problem (e.g., "I can't get this equipment towork"). In reporting a problem of this type, the speakerwould seem to want an agentic reaction (advice). In theother case, a speaker described a communal problem(e.g., "I behaved badly-people must despise me"). Inreporting that kind ofproblem, the speaker would seemto want an empathic response that might undo thesense of rejection or ostracism. According to Jeffersonand Lee (1992), difficulties can arise whenever aspeaker and listener do not understand each other (i.e.,when the speaker's wishes are ambiguous). Thespeaker may want one kind of support, but the listenermay misinterpret the speaker's behavior. In an experi-mental study, Horowitz et al. (2001, Study 3) showedthat people report more dissatisfaction when a lis-tener's reaction does not match their desired reaction.

The Self and Self-ProtectiveInterpersonal Motives

A number of interpersonal motives seem to arise asa way of defending the self from feelings of vulnerabil-ity that are associated with relational schemas(Baldwin, 1992). Early writers (e.g., Adler, 1927;Homey, 1945; Sullivan, 1953) have described thissense of vulnerability and strategies that people use toreassure themselves that they possess desired commu-



HOROWITZ ET AL.

nal (e.g., likeable) and agentic (e.g., competent)self-qualities. More recent investigators (J. D. Brown,1998; Tafarodi & Milne, 2002) have even developedways to separately assess communal and agentic formsof vulnerability.

Attachment theory has been particularly helpful inclarifying the nature of this vulnerability. Attachmentmarks the first appearance of a communal motive in aninfant's life, revealing a powerful motive for the infantto connect with an available adult (Ainsworth, 1982;Bowlby, 1969, 1973). Numerous attachment theoristshave described the vicissitudes of the motive and theirconsequences for a person's subsequent development(see Cassidy & Shaver, 1999). One particularly impor-tant aspect of the theory concerns the development of aperson's images of the self and others (e.g.,Bartholomew & Horowitz, 1991; Crowell, Fraley, &Shaver, 1999; Feeney, 1999; Shaver & Hazan, 1988).For example, a view of oneself as inadequate and ofother people as potentially disapproving, rejecting, orhumiliating can give rise to intense expectations of be-ing hurt and intense motives to protect oneself from po-tential distress.

Our model needs to explain how these self-protec-tive, interpersonal motives acquire strength. The fol-lowing propositions relate interpersonal motives to theself.

P16. Self- and other schemas are acquired, inpart, through interpersonal interactions.

According to Bowlby (1973) and later attachmenttheorists (e.g., Bretherton& Munholland, 1999), infantsbegin to form images ("internal working models") ofother people and the self early in life. Biological predis-positions undoubtedly interact with experience to shapethese schemas. For example, children differ biologi-cally in their susceptibility to anxiety (e.g., Kagan,1994; Kagan, Reznick, & Snidman, 1988), and an anxi-ety-prone child would probably perceive danger in situ-ations that placid children take in stride. If an anxi-ety-prone child were terrified repeatedly byunpredictable interpersonal loss or abandonment, thatchild might acquire vivid schemas and expectationsabout (a) other people as potentially unreliable, (b) theself as potentially abandoned and helpless, and (c) situa-tions that portend danger. The resulting schemas and ex-pectations might then sensitize a child to abandonment,giving rise to a self-protective motive to prevent futureabandonment. In this way, heredity and environmentwould interact to intensify expectations of threat in par-ticular interpersonal situations.We can illustrate the Heredity x Environment inter-

action with experimental data concerning the develop-ment of an insecure attachment. Van den Boom (1989,1994) assessed the temperament of infants at 10 daysof age to identify and study a group of distress-prone

infants. Then, in one experiment, mothers either re-ceived or did not receive special training in caring fortheir distress-prone child. Mothers who received notraining frequently came to ignore the child exceptwhen the child was in distress. When those childrenwere tested in Ainsworth's (Ainsworth & Bell, 1970;Ainsworth, Blehar, Waters, & Wall, 1978) Strange Sit-uation at 1 year of age, the majority were insecurely at-tached to the mother. However, mothers who did re-ceive training learned ways to deal with adistress-prone child and were highly responsive to thechild's needs. When those infants were later tested inthe Strange Situation, they were usually securely at-tached to the mother. Thus, a child's heredity can shapethat child's environment, and the resulting combina-tion of heredity and environment would shape schemasof the self, of other people, and of the environment.Similar mechanisms are discussed by Buss (1987).

P17. Images of the self and others are often de-scribed by fuzzy concepts.

Concepts that are used to describe the self in earlychildhood are usually well defined (see Harter, 1999),but later concepts are not. For example, an earlyself-descriptor such as "I am a boy" has a clear,well-defined referent. However, later concepts such asnerd, stud, sissy, and wimp do not. A concept such assissy is said to have a fuzzy definition because we can-not state the necessary and sufficient criteria. Manypossible criteria come to mind-effeminate,unaggressive, frail, timid, eager to please, cow-ardly-but none is absolutely essential. Some sissieshave one subset of characteristics; others have a differ-ent subset of characteristics. The best we can do in de-fining "sissy" is to identify the most prototypic charac-teristics that people think of when they describe a sissyand judge how well a given person's characteristicsmatch that prototype (Horowitz, Post, French, Wallis,& Siegelman, 1981; Horowitz, Wright, Lowenstein, &Parad, 1981). The better the match, the higher the prob-ability that the person is called a sissy.

Some writers have therefore proposed that theself-image be viewed as a theory about the self, a setof hypotheses that keep getting tested and supportedor refuted (Brim, 1976; Epstein, 1973). If a boy com-pared himself to the prototype of a sissy and observedmany of its characteristics, he might have to classifyhimself as a sissy. Of course, behaviors vary fromtime to time, and by performing a very aggressive orbold act, a boy with marginal characteristics couldshow himself and the world that he does not fit thecategory. Later, however, if the boy were publiclypraised for being "a very good boy who never causestrouble," the sissy hypothesis might again become athreat, requiring behavioral disconfirmation. There-fore, a child who thought he marginally fit the cate-




gory might have to protect his self-esteem by behav-ing in ways that refuted the undesirable hypothesis(e.g., by being aggressive or bold).

P18. Interpersonal tests are performed to affirmor validate a desired self-image.

People sometimes create interpersonal interactionsthat confirm (or disconfirm) a particular hypothesis(Weiss & Sampson, 1986). For example, people takesteps to correct a partner's perception of them when itcontradicts their own self-perception-even when thepartner's perception is more favorable than their own(Swann, 1996). Such steps help "prove" that their ownself-perception is valid. That is, people instigate inter-actions that confirm a desirable hypothesis: Bulliesseem to challenge people who are easy marks as a wayof "proving" that they are tough and strong. Narcissis-tic people seem to solicit admiration as a way of "prov-ing" that they are admirable. Obsessive-compulsivepeople seem to strive for perfection as a way of "prov-ing" that they are beyond reproach. Histrionic peopleseem to draw attention to themselves as a way of"proving" that they are connected to others. In thisway, particular interpersonal motives get strengthenedas a way of affirming desired qualities of the self in re-lation to others.

P19. Feelings of vulnerability arise from char-acteristic expectations in particular situations.

Clarifying an interpersonal situation frequentlyhelps explain a Person x Situation interaction.Holmes (2002) stressed the significance of a person'sidiosyncratic expectations about a given partner'smotives. This point helps clarify the results of Shoda,Mischel, and Wright (1994), who showed that chil-dren at a summer camp differed in the profile of situ-ations in which they displayed verbal aggression. Forexample, some children displayed aggression whenpraised by an adult, whereas other (perhaps equallyaggressive) children did not. As Holmes noted, themeaning of an interpersonal situation (an adult prais-ing the child) may or may not lead to verbal aggres-sion because children differ in their expectationsabout the adult's motives. Some children expectmean-spirited manipulation from a praising adult,frustrating the child's self-protective motive and in-ducing negative affect (anger).

Frustrated Interpersonal Motives(Interpersonal Problems)

When important goals and motives are chronicallyfrustrated, the person suffers negative affect. We now

ask why interpersonal motives get frustrated.

P20. A chronically frustrated goal or motiveconstitutes an interpersonal problem for pne orboth members of a dyad.

Most people seem to be reasonably successful infinding ways to attain desired levels of intimacy,friendship, autonomy, influence, self-efficacy, and soon. Some people, however, are not successful and re-port interpersonal problems. A very shy person, for ex-ample, might yearn for intimacy but avoid social con-tact to protect the self from rejection. By withdrawingfrom others, the person unwittingly invites others towithdraw. In this way, the person's self-protectivestrategy frustrates the motive for intimacy.When important interpersonal motives are chroni-

cally frustrated, the person reports interpersonal prob-lems, such as "It is hard for me to make friends" or "Ifind myself alone too much." Complaints of this kindmay be assessed using the Inventory of InterpersonalProblems (Horowitz, Alden, Wiggins, & Pincus,2000), a self-report measure that contains 64 items(problems) organized in two dimensions that corre-spond to communion and agency. Every item states acommon interpersonal problem, and the 64 items areorganized into eight scales (eight items per octant).

P21. Why do interpersonal motives get frus-trated, causing interpersonal problems?

Suppose a person has a strong desire to affirm theself through assertive behavior (an agentic motive), butthe person generally finds it hard to be assertive (an in-terpersonal problem). How is this interpersonal prob-lem to be understood? Among the possible answers,two are particularly evident from our model.

Motives conflict. Psychodynamic writers (e.g.,Luborsky & Crits-Christoph, 1998; Strupp & Binder,1984; Strupp, Schact, & Henry, 1988) have empha-sized the adverse effect of motivational conflict on aperson's well-being. When two or more motives con-flict within a person, the person has to sacrifice one tosatisfy another. For example, a person with a strong de-sire to behave assertively may forsake that motive tosafeguard a communal motive (preserving harmony inrelationships). Numerous laboratory studies haveshown that people often camouflage or forfeit a desiredagentic goal (e.g., successful competitive behaviorwith friends) to preserve a friendship (e.g., Arroyo &Zigler, 1995; Brigham, Kelso, Jackson, & Smith, 1997;Cross, Coleman, & Terhaar-Yonkers, 1991; Exline &Lobel, 1999; Santor & Zuroff, 1997). In addition,Emmons (1986; Emmons & King, 1988) assessed theamount of conflict people reported among theirtop-rated "personal strivings" (goals). Participantslisted up to 15 goals and rated, for each pair of goals,the extent to which the goals conflicted. The goal "to



HOROWITZ ET AL.

appear more intelligent than I am," for example, wasjudged to conflict with the goal "to present myself in anhonest light." The more conflict participants reported,the greater their level of negative affect, depression,visits to a health center, psychosomatic complaints,and physical illness. Riediger and Freund (2004) alsoshowed that conflicting goals interfere with a person'ssubjective well-being.

Ambiguous behavior is misinterpreted. A sec-ond reason that motives get frustrated is that behavioris misinterpreted. A person who tries to be firm or as-sertive may come across as unreasonable or disagree-able. Other people, misinterpreting the behavior, maythen react in ways that frustrate the person's motive.When Allport (1937) introduced the trait as a unit ofpersonality, he regarded the trait as a summary of fre-quent acts and motives, as well as values, likes, andinterests. Following Allport (pp. 319-324), we wouldsay that a sociable person wants company, an asser-tive person wants to have influence, a theatrical per-son wants attention, a dependent person wants to becared for, a timid person wants safety, and a narcissis-tic person wants admiration. Observers, however,may misinterpret the behavior and frustrate the per-son's motive.

In brief, important interpersonal motives may bechronically frustrated in adulthood, creating interper-sonal problems. When the frustration is severe enoughto produce significant distress or impairment, we speakof a personality disorder.We now apply our model to help clarify the defini-

tion of most personality disorders.

Personality Disorders and FrustratedInterpersonal Motives

This section applies our model to personality disor-ders. In our view, the personality disorders fall on acontinuum with normative personality traits and arebest explained by the constructs and methodology ofpersonality and social psychology. Progress in under-standing these phenomena would seem to require asensible theory that could direct future research.

According to Diagnostic and Statistical Manual ofMental Disorders (DSM-IV-TR; American Psychiat-ric Association [APA], 2000), a personality disorder isdefined in terms of enduring personality traits thatseem to have evolved gradually over the childhoodyears and become well established by early adulthood.The term also implies that the traits have come to bemaladaptive over time, leading to subjective distress orimpaired functioning or both. For example, perfection-ism may have been adaptive at one time in helping theperson protect the self from criticism, but later in lifethat perfectionism may impair the person's function-

ing: A rigid perfectionist may now be unable to com-plete important tasks at work.

Every personality disorder is defined as a fuzzy set:The features (or criteria) of each disorder consist of alist of n characteristics. To receive the diagnosis, theperson must exhibit any m of those n characteristics.No single feature or subset of features constitutes anecessary and sufficient condition for the diagnosis.The features themselves are heterogeneous, describingmotives, current concerns, behaviors, affects, andcognitions. Because every feature is weighted equally,no one feature serves to integrate or organize the oth-ers. In brief, there is no overarching conception that in-tegrates the features, clarifies the person's problem, orexplains why some personality disorders co-occurmore often than others. The study of personality disor-ders is said to contain "a conceptual and theoreticalvoid" (Marinangeli et al., 2000, p. 74).We propose that the features of most (but not all)

personality disorders contain a single salient interper-sonal motive that organizes the other features. This in-terpersonal motive helps us conceptualize the disorderin a way that (a) formulates it in interpersonal terms,(b) emphasizes its continuity with "normal" mecha-nisms, (c) explains why most personality disorders re-quire a fuzzy definition, and (d) clarifies the relation ofthe personality disorders to one another.

The features of most personality disorders may beclassified into one of four types. The first includesthose criteria that explicitly describe a self-protectiveinterpersonal motive or current concern. The clearestexamples appear in the dependent, avoidant, border-line, narcissistic, histrionic, and paranoid personalitydisorders (Horowitz, 2004; Horowitz & Wilson, 2004).The other three types of criteria describe consequencesof that motive-(a) strategies that the person uses tosatisfy the motive, (b) negative affect that occurs whenthe strategies fail and the motive is frustrated, and (c)characteristic ways in which the person tries to reducethe resulting negative affect. We begin by examiningthe interpersonal motives that integrate the other threekinds of criteria.

P22. The features of 6 of the 10 personalitydisorders specifically mention a self-protectiveinterpersonal motive.

Interpersonal motives that are mentioned among thecriteria of a personality disorder describe a motive toattain a desired state ("an approach goal") or a motiveto avoid an aversive state ("an avoidance goal"). A cri-terion of the histrionic personality disorder describes adesire to be the center of attention. A criterion of thenarcissistic personality disorder describes an excessiveneed for respect and admiration. Criteria of the para-noid personality disorder describe a motive to protectthe self from malice, humiliation, and exploitation by




others. Criteria of the avoidant personality disorderemphasize an intense feeling of inadequacy that givesrise to a motive to avoid rejection, disapproval, criti-cism, ridicule, and shame. Criteria of the dependentpersonality disorder emphasize an intense feeling ofinadequacy that gives rise to a motive to have otherstake charge. A criterion of the borderline personalitydisorder describes a motive to avoid being abandoned.

P23. Other features of the disorder may be orga-nized around the motive (or frustrated motive).

Once a self-protective interpersonal motive hasbeen identified, it helps clarify the other features. Asnoted previously, they fall into three categories.

Strategies for Satisfying the Motive

Motives may be satisfied in alternate ways. Ac-cording to DSM-IV-TR (APA, 2000), a person with ahistrionic personality disorder is apt to draw attentionto the selfby using physical appearance or exaggerateddisplays of emotion or both. A person with a narcissis-tic personality disorder is apt to affirm a feeling ofself-worth by exploiting other people or acting likesomeone who is special, important, and entitled, or bydoing both. A person with an avoidant personality dis-order is apt to avoid rejection by minimizing socialcontact, intimacy, and new relationships. A personwith a dependent personality disorder is apt to avoidfeelings of helplessness by pleasing others and gettingthem to take charge. A person with a paranoid person-ality disorder is apt to avoid humiliation by guardingagainst possible malice, disloyalty, and abuse.

Negative Affect When the Motive isFrustrated

When these strategies fail and the motive is frus-trated, negative affect follows. Some features describethis negative affect. According to DSM-IV-TR (APA,2000), a person with a dependent personality disorderis apt to become uncomfortable, anxious, or helplesswhen alone. A person with a borderline personalitydisorder is apt to shift abruptly to a state of depressionor anger or both. A person with a narcissistic personal-ity disorder is apt to become envious. A person with aparanoid personality disorder is apt to become enragedat perceived signs of malice.

Ways of Coping with Negative Affect

The remaining criteria describe how the personcopes with negative affect. People with a borderlinepersonality disorder, trying to feel better, are apt to actout on the self or others through impulsive or suicidalbehavior. Kemperman, Russ, and Shearin (1997)

showed that people who mutilate themselves (e.g., bycutting their wrists) often do so in an effort to feelbetter. Likewise, people with a paranoid personalitydisorder are apt to counterattack. People with a de-pendent personality disorder are apt to seek another re-lationship urgently when a close relationship hasended. People with a narcissistic personality disorderare apt to become arrogant and haughty.

P24. At least two personality disorders inDSM-IV-TR cannot be organized around an in-tegrating motive.

The features of the schizotypal personality disor-der in DSM-IV-TR (APA, 2000) resemble mildsymptoms of schizophrenic disorders-e.g., "oddthinking and speech," "ideas of reference," "odd be-liefs or magical thinking" (p. 701). These criteria donot name (or imply) any integrating motive. Like-wise, criteria of the antisocial personality disorder areorganized around a deficit (Killingmo, 1989) ratherthan a frustrated interpersonal motive: a lack of guiltor remorse plus antisocial behaviors that result fromthe deficit. Therefore, the schizotypal and antisocialpersonality disorders may be categorically differentfrom other personality disorders (McWilliams, 1998).

P25. Personality disorders are said to reflect amaladaptive interpersonal pattern.

Early discussions of personality disorders empha-sized maladaptive interpersonal patterns (e.g., Carson,1969; Horney, 1945; Kiesler, 1983, 1996; Leary, 1957;McLemore & Brokaw, 1987; Pincus & Wiggins, 1990;Strupp & Binder, 1984; Sullivan, 1953). According toone common explanation, the adult person's currentbehavior reflects outdated patterns that once served animportant motive in childhood (Benjamin, 1996). Forexample, excessive compliance may have been adap-tive at one time, leading to praise, affection, and good-will from adults; but now in adulthood, excessive com-pliance may lead to disrespect, impatience, orunwanted advice from others.

Our model broadens the description by focusing ondetails of current behavior. For example, a woman witha histrionic personality disorder, trying repeatedly toestablish communal connections, might seem to othersto be manipulative or self-centered. A man with anavoidant personality disorder, trying repeatedly to pro-tect himself from rejection, might seem uninterested inconnecting. As the person's efforts to satisfy the mo-tive backfire, they frustrate the very motive that theywere meant to satisfy. As a result, the person experi-ences subjective distress, which the person tries to alle-viate in maladaptive ways (e.g., self-isolation). Eachaspect of the formulation contributes to our under-standing of the person-the interpersonal motive, the



HOROWITZ ET AL.

person's strategies, the reason the strategies fail, thenegative affect that results, and the person's (at times)self-defeating ways of coping with negative affect(also see Piper, Joyce, McCallum, Azim, & Ogro-dniczuk, 2001).

People frequently qualify for two or more personal-ity disorders (e.g., Marinangeli et al., 2000). In thatcase, the person's disorder would be formulated interms of two or more organizing motives. For example,the very same person might crave attention (a histri-onic motive) and also strive to avoid abandonment (aborderline motive); both involve communal motives.Indeed, the histrionic and borderline personality disor-ders do frequently co-occur (e.g., Davila, 2001; Wat-son & Sinha, 1998). Likewise, a person might crave at-tention (a histrionic motive) and also crave admiration(a narcissistic motive): One is largely communal, theother primarily agentic; but both reflect a strong desireto influence other people.

P26. A pressing motive may induce (and be re-inforced by) a cognitive bias that sustains themaladaptive pattern.

The paranoid personality disorder illustrates the ef-fect of motives on cognitions: In an effort to protect theself from malice, the person has become wary of otherpeople, and that suspiciousness seems to lower the per-son's objectivity in testing hypotheses about others(Millon & Davis, 2000). That is, the person has appar-ently acquired an extremely salient interpersonalagenda, namely, to avoid humiliation by detectingearly hints of cheating, deception, exploitation, and be-trayal. The person conducts a vigilant (but biased)search for evidence supporting that concern. When aparanoid man applies an interpersonal test and detectssigns of possible malice, he becomes convinced thathis suspicion has been confirmed, and this "discovery"reinforces his original need for vigilance (Millon &Davis, 2000). Evidence to the contrary is simply ig-nored. In this way an intense self-protective motive canaffect cognitions.

To some extent, a cognitive bias may be associatedwith every personality disorder that is organizedaround an intense motive to protect the self (Beck &Freeman, 1990): The borderline person may be biasedtoward perceiving signs of abandonment; the avoidantperson, signs of rejection; the histrionic person, signsof indifference; the narcissistic person, signs of disre-spect. False alarms then provide the "confirming evi-dence" that increases the sense of frustrated motive, in-ducing negative affect.

If interpersonal motives are basic to personality dis-orders, then each personality disorder should occupythe same position in a circumplex of personality disor-ders that the corresponding motive occupies in acircumplex of motives. That is, if most personality dis-orders can be organized around a salient motive, weshould be able to demonstrate empirically that they canbe located graphically in a two-dimensional space de-fined by communion and agency. For example, the his-trionic motive-"to connect with other people by get-ting their attention"-implies a desire to influenceother people to connect. That disorder should thereforeoccupy the upper right-hand quadrant of a two-dimen-sional space (see Figure 5). The dependent mo-tive-"to connect with other people and get them totake charge"-implies that the dependent personalitydisorder should occupy the lower right-hand quadrant.Two disorders that are located very close to one an-other should thus reflect similar organizing motives,and those disorders should co-occur particularly often.

Numerous studies have scaled and graphed the per-sonality disorders. For example, Pincus and Wiggins(1990) administered questionnaires to a large sampleof undergraduate students to identify the interpersonalproblems (frustrated motives) associated with differentpersonality disorders. A principal components analysisshowed that the personality disorders can be organizedwithin two primary dimensions. Other authors haveobtained similar results using samples of students andsamples of psychiatric patients (Blackburn, 1998;DeJong, van den Brink, Jansen, & Schippers, 1989;Matano & Locke, 1995; Morey, 1985; Overholser,1996; Sim & Romney, 1990; Soldz, Budman, Demby,& Merry, 1993; Trull, Useda, Conforti, & Doan, 1997).

Wagner, Riley, Schmidt, McCormick, and Butler(1999) summarized the various results this way: Asshown in Figure 5, people with a narcissistic personal-ity disorder are high in agency and neutral in commu-

Domineering/Controlling

Vindictive/Self-Centered

Cold/Distant

SS

Socially Inhibited

Intrusive/Needy

Self-Sacrificing

Overly Accommodating

Nonassertive

P27. Personality disorders may be organizedgraphically in two dimensions that correspond tocommunion and agency.

Figure 5. Location of personality disorders in the two-dimen-sional interpersonal space. AVD = avoidant; DPN = dependent;HIST = histrionic; NAR = narcissistic; PAR = paranoid; SZD =

schizoid.




nion (in our terms, they want respect and admiration).Those with a paranoid personality disorder are high inagency and low in communion (they want to influenceothers to "leave me alone"). Those with an avoidantpersonality disorder are low in both (they want to pro-tect the self by remaining passive and disconnected).Those with a dependent personality disorder are low inagency and high in communion (they want connectedothers to take charge). Those with a histrionic person-ality disorder are high in both (they want to influenceothers to become connected). The borderline personal-ity disorder, with its many instabilities, does not seemto occupy a consistent graphical location.

Personality disorders frequently co-occur, but somemore often than others (Marinangeli et al., 2000). Thegraphical arrangement in Figure 5 should help us pre-dict which are most apt to co-occur. Those that are nearone another should often co-occur (they reflect similarmotives); Watson and Sinha (1998) found, for exam-ple, that the histrionic and narcissistic personality dis-orders had a comorbidity of 30.4%. However, disor-ders that are diametrically apart should co-occur lessoften; in the same data the histrionic and avoidant per-sonality disorders had a comorbidity of 0%.

Concluding Remarks

We have proposed a revision in the circumplex mod-els that have evolved from Leary's (1957) seminal con-tribution. Our revision has postulated four new assump-tions: (a) The negative pole ofcommunion is taken to beindifference, not hostility. (b) A given behavior invites(rather than evokes) a particular reaction, which thepartner may choose not to satisfy. (c) The complementofabehavior is the reaction that would satisfy the motivebehind it. (e) Noncomplementary reactions induce neg-ative affect. By changing the emphasis to motives, themodel highlights ambiguity about the motive behind abehavior, leading to miscommunications. The modelalso calls attention to individual differences in motives.A reaction, such as self-disclosure, that satisfies one per-son s motive may induce anger in another person. Themodel also helps clarify personality disorders andplaces them on a continuum with "normal" behavior.We conclude the article with comments about future re-search based on the model.

New Measures Now Exist for AssessingInterpersonal Goals

Classical studies of games have compared coopera-tive (communal) and competitive (agentic) motivesduring interpersonal interactions-for example, thePrisoner's Dilemma (e.g., Axelrod, 1980; Colman,2003; Kelley & Stahelski, 1970). In general, however,empirical studies have focused more on the "situation"than the "person" because, until recently, we have

lacked objective measures of individual differences ininterpersonal goals. However, two measures now existthat provide a representative sample of interpersonalgoals.

Locke's (2000) measure contains both approachgoals ("It is important to me that other people under-stand me") and avoidance goals ("It is important to methat I not make a social blunder"), organized along twodimensions corresponding to communion and agency.Each of the eight subscales contains a broad but coher-ent group of goals that lie within an octant (a higher or-der motive). By weighting responses appropriately,one can sum the x and y component of each response toobtain two overall scores that assess the relativestrength of communion and agency.

Grosse Holtforth (2002) also developed a 94-itemmeasure from the statements of people in treatment.Everyday goals, such as avoiding criticism and seekingthe admiration of others, are grouped into avoidanceand approach motives. Grosse Holtforth (2005) dem-onstrated an important relation between the strength ofavoidance motives and distress from interpersonalproblems.

These Measures Can Now Be Used toTest the Model Experimentally

Participants can now be selected to be high or low ina particular motive, and variables can be manipulatedexperimentally to evaluate hypothesized cognitions,affect, or behavior. Here are three examples:

1. Suppose a participant interacts with a confeder-ate in a "getting acquainted" task, and the confeder-ate's behavior fails to satisfy the participant's goals.For example, a participant with strong communal goalsmight interact with a taciturn, unresponsive confeder-ate. Or a participant with weak communal goals mightinteract with an open, self-disclosing confederate. Ac-cording to the model, participants in both of these con-ditions should be relatively dissatisfied. In both cases,participants should also exhibit noncomplementary be-havior that invites the confederate to alter his or her be-havior in a direction that satisfies the participant'sgoals.

2. Interpersonal behavior is often ambiguous, sothe way a participant interprets an actor's behaviorshould depend on the participant's own salient mo-tives. If a videotaped scene depicted an actor teasing apartner, participants might disagree about the amountof friendliness or warmth present in the teasing. If weknew each participant's own interpersonal goals(hence, the person's cognitive biases), we should beable to predict the participant's interpretation of the ac-tor's behavior.

3. We might also construct tasks to evaluate a par-ticipant's ability to provide the kind of social support a



HOROWITZ ET AL.

problem teller would like. Suppose target people wereasked to tell a partner about a recent problem situationin which they (a) "wanted the other person to give ad-vice" or (b) "wanted the other person simply to listenand show compassionate understanding." Each targetperson's description of the problem might be video-taped and shown to other naive participants, who mightthen be asked to respond to the videotaped target per-son. "Accuracy" in social support would then be de-fined by a responder's success in matching each targetperson's desired reaction. A responder's response biasacross videotapes could also be assessed and related tothat person's own most salient interpersonal motives.

A comprehensive measure of interpersonal goalscould also be used to evaluate our view of personalitydisorders. People randomly sampled from the generalpopulation might complete two self-report mea-sures-one assessing interpersonal goals, the otherassessing the degree to which the person endorsestraits resembling those of a particular personality dis-order. People who report the traits corresponding to aparticular personality disorder should also report thecharacteristic motive that is associated with thosetraits. For example, people who report traits resem-bling those of a histrionic personality disorder shouldreport a strong desire to feel connected with others.Locke (2000, Study 3) has already shown a relationbetween characteristic traits and hypothesized inter-personal goals among college students. Once identi-fied, we might then examine, from the person's ownperspective, strategies used to satisfy that motive,emotional reactions that typically occur when thosestrategies fail, and ways of coping with the resultingnegative affect. A measure of interpersonal goalscould also help clarify the meaning of abstract mo-tives such as a motive to be loved. Our model as-sumes that a motive is a superordinate construct thatsubsumes narrower interpersonal goals. Super-ordinate constructs, however, are themselves poorlydefined (fuzzy) concepts. Fehr (1988, 2004) de-scribed a procedure for operationalizing fuzzy inter-personal concepts, such as intimacy and love. In ap-plying Fehr's procedure, individuals first generatefeatures of the construct, and other individuals thenrate the prototypicality of each generated feature. Thetotal set of features constitutes the prototype for love.Examples of highly prototypic criteria for love gener-ated by this method include "a desire to be with theother person," "feelings of caring," "a sense of free-dom to talk about anything," and many others.

If a person has a strong motive to be loved, thenthat person should endorse a number of interpersonalgoals that exist in the prototype-for example, want-ing the other person to "care about me, desire to bewith me, feel free to talk about anything with me,"and so on. However, two individuals may differ as to

which elements of the group-generated prototype oflove they personally consider important. We are neversurprised when a man in a distressed marriage de-clares that he loves his wife, and his wife challengeshis claim by insisting, "If you really loved me, youwould be more open with me." The wife's concept oflove apparently includes "openness" as a feature,whereas the husband's concept apparently does not.Systematic research is needed to better understandmiscommunications of this type that arise from dif-fering interpersonal goals within the same higher or-der construct.

The Interpersonal Approach HasUnique Advantages

Of all the theoretical approaches to personality andsocial psychology, the interpersonal approach is proba-bly the one that is most compatible with the others: (a)Like the biological approach, it assumes that innate tem-peramental differences play an important role in shap-ing an individual's personality and interpersonal inter-actions (Horowitz, 2004). According to theinterpersonal approach, a child's temperament affectsthe caretaker's (and other people's) reactions to a child,thereby shaping the child's environment. (b) Like thecognitive-behavioral approach, the interpersonal ap-proach emphasizes the important role of cognitions(e.g., schemas and cognitive biases) in shaping a per-son's expectancies and interpretations of otherwise am-biguous behavior. As noted previously, these cognitionsthemselves can strengthen a person's interpersonal mo-tives and subsequent interactions. (c) Like the humanis-tic approach to personality, the interpersonal approachemphasizes the self, dyadic relations, communication,and social support-all topics addressed in this article.(d) Like the psychodynamic approach, the interpersonalapproach emphasizes motives, ascribing important psy-chological consequences to frustration that arises fromunsolvable motivational conflict.

Because the interpersonal approach harmonizes sowell with all of these theoretical approaches, it is inte-grative: It draws from the wisdom of all major ap-proaches to systematize our understanding of interper-sonal phenomena (see also Pincus & Ansell, 2003).Although it is integrative, however, it is also unique, pos-ing characteristic questions ofits own. It asks, for exam-ple, what a person is seeking in a particular dyadic inter-action. Does the person want to connect with others forcare, comfort, intimacy, or friendship? Or does the per-son want to be autonomous, demonstrate competence,display superiority, or confirm a desired identity? Whatmotivates a particular person to avoid interactions, telllies without apparent reason, threaten suicide, maintaina program of self-starvation, have temper tantrums, ordisagree "for the sake of disagreeing"? In all of thesecases, the personmay be seeking autonomy or control or




both, or the person may be seeking nurturance or someother form of connection. Finally, if an interpersonalmotive is chronically frustrated, what is the reason forthe chronic frustration?

These, and many other questions that we haveposed, help define the interpersonal approach to per-sonality and social psychology. We have tried to ar-ticulate the major propositions of the model and orga-nize them into a relatively simple theoreticalframework. Using this framework, it should be possi-ble to study social interactions with all of the concep-tual and methodological tools of social psychology,personality, communications, psycholinguistics, andchild development.

References

Adler, A. (1927). The practice and theory of individual psychology.New York: Harcourt.

Ainsworth, M. D. S. (1982). Attachment: Retrospect and prospect.In C. M. Parkes & J. Stevenson-Hinde (Eds.), The place of at-tachment in human behavior (pp. 3-30). New York: BasicBooks.

Ainsworth, M. D. S., & Bell, S. M. (1970). Attachment, exploration,and separation: Illustrated by the behavior of one-year-olds in astrange situation. Child Development, 41, 49-67.

Ainsworth, M. D. S., Blehar, M. D., Waters, E., & Wall, S. (1978).Patterns of attachment: A study of the strange situation.Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Alston, W. P. (1975). Traits, consistency, and conceptual alternativesfor personality theory. Journalfor the Theory ofSocial Behav-iour, 5(1), 17-48.

American Psychiatric Association. (2000). Diagnostic and statisti-cal manual of mental disorders (4th ed., text revision). Wash-ington, DC: Author.

Angyal, A. (1941). Foundations for a science of personality.New York: Commonwealth Fund and Harvard UniversityPress.

Arroyo, C. G., & Zigler, E. (1995). Racial identity, academicachievement, and the psychological well-being of economi-cally disadvantaged adolescents. Journal ofPersonality and So-cial Psychology, 69, 903-914.

Asch, S. E. (1946). Forming impressions of personalities. Journal ofAbnormal and Social Psychology, 41, 258-290.

Austin, J. T., & Vancouver, J. B. (1996). Goal constructs in psychol-ogy: Structure, process, and content. Psychological Bulletin,120, 338-375.

Axelrod, R. (1980). More effective choice in the Prisoner's Di-lemma. Journal of Conflict Resolution, 24, 379-403.

Bakan, D. (1966). The duality of human existence: Isolation andcommunion in Western man. Boston: Beacon.

Baldwin, M. W. (1992). Relational schemas and the processing ofsocial information. Psychological Bulletin, 112, 461-484.

Bartholomew, K., & Horowitz, L. M. (1991). Attachment stylesamong young adults: A test of a four-category model. Journalof Personality and Social Psychology, 61, 226-244.

Beck, A. T., & Freeman, A. F. (1990). Cognitive therapy ofpersonal-ity disorders. New York: Guilford.

Benjamin, L. S. (1996). Interpersonal diagnosis and treatment ofpersonality disorders (2nd ed.). New York: Guilford.

Blackburn, R. (1998). Relationship of personality disorders to ob-server ratings of interpersonal style in forensic psychiatric pa-tients. Journal of Personality Disorders, 12, 77-85.

Bowlby, J. (1969). Attachment and loss: Vol. 1. Attachment. NewYork: Basic Books.

Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation: Anxietyand anger New York: Basic Books.

Bradbury, T. N., & Fincham, F. D. (1992). Attributions and behaviorin marital interaction. Journal of Personality and Social Psy-chology, 63, 613-628.

Bretherton, I., & Munholland, K. A. (1999). Internal working modelsin attachment relationships: A construct revisited. In J. Cassidy& P. R. Shaver (Eds.). Handbook of attachment: Theory, re-search, and clinical applications (pp. 89-111). New York:Guilford.

Brigham, N. L., Kelso, K. A., Jackson, M. A., & Smith, R. H. (1997).The roles of invidious comparisons and deservingness in sym-pathy and schadenfreude. Basic and Applied Social Psychol-ogy, 19, 363-379.

Brim, 0. G., Jr. (1976). Life-span development of the theory of one-self: Implications for child development. In H. W. Reese (Ed.),Advances in child development (Vol. 11, pp. 241-25 1). NewYork: Academic.

Brown, J. D. (1998). The self New York: McGraw-Hill.Brown, P., & Levinson, S. (1987). Politeness: Some universals in

language usage. New York: Cambridge University Press.Bruch, H. (1973). Eating disorders: Obesity, anorexia nervosa, and

the person within. New York: Basic Books.Buss, D. M. (1987). Selection, evocation, and manipulation. Journal

ofPersonality and Social Psychology, 53, 1214-1221.Buss, D. M., & Craik, K. H. (1983). The act frequency approach to

personality. Psychological Review, 90, 105-126.Cantor, N., & Kihlstrom, J. F. (1987). Personality and social intelli-

gence. Englewood Cliffs, NJ: Prentice Hall.Carson, R. C. (1969). Interaction concepts ofpersonality. Chicago:

Aldine.Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking

time seriously: A theory of socioemotional selectivity. Ameri-can Psychologist, 54, 165-18 1.

Caspar, F. (1995). Plan analysis: Toward optimizing psychotherapy.Seattle, WA: Hogrefe & Huber.

Caspar, F. (1997). Plan analysis. In T. Eells (Ed.), Handbook ofpsy-chotherapy case formulation. New York: Guilford.

Cassidy, J. (1999). The nature of the child's ties. In J. Cassidy & P. R.Shaver (Eds.), Handbook ofattachment: Theory, research, andclinical applications (pp. 3-20). New York: Guilford.

Cassidy, J., & Shaver, P. R. (1999). Handbook ofattachment: Theory,research, and clinical applications. New York: Guilford.

Cobb, S. (1976). Social support as a moderator of life stress. Psycho-somatic Medicine, 38, 300-314.

Colman, A. M. (2003). Cooperation, psychological game theory, andlimitations of rationality in social interaction. Behavioral andBrain Sciences, 26, 139-153.

Cropanzano, R., James, K., & Citera, M. (1992). A goal hierarchymodel of personality, motivation, and leadership. In L. L.Cummings & B. M. Staw (Eds.), Research in organizationalbehavior (Vol. 15, pp. 267-322). Greenwich, CT: JAI.

Cross, T. L., Coleman, L. J., & Terhaar-Yonkers, M. (1991). The so-cial cognition of gifted adolescents in schools: Managing thestigma of giftedness. Journal for the Education of the Gifted,15(1), 44-55.

Crowell, J. A., Fraley, C., & Shaver, P. R. (1999). Measurement of in-dividual differences in adolescent and adult attachment. In J.Cassidy & P. R. Shaver (Eds.). Handbook of attachment: The-ory, research, and clinical applications (pp. 434-465). NewYork: Guilford.

Cutrona, C. E., & Suhr, J. A. (1992). Controllability of stressfulevents and satisfaction with spouse support behaviors. Commu-nication Research, 19, 154-174.

Cutrona, C. E., & Suhr, J. A. (1994). Social support communicationin the context of marriage: An analysis of couples' supportive



HOROWITZ ET AL.

interactions. In B. R. Burleson, T. L. Albrecht, & I. G. Sarason(Eds.), Communication ofsocial support (pp. 113-135). Thou-sand Oaks, CA: Sage.

Davila, J. (2001). Paths to unhappiness: The overlapping courses ofdepression and romantic dysfunction. In S. R. H. Beach (Ed.),Marital andfamily processes in depression: A scientific foun-dation for clinical practice (pp. 71-87). Washington, DC:American Psychological Association.

DeJong, C. A. T., van den Brink, W., Jansen, J. A. M., & Schippers,G. M. (1989). Interpersonal aspects of DSM-III Axis II: Theo-retical hypotheses and empirical findings. Journal ofPersonal-ity Disorders, 3, 135-146.

Deutsch, M., & Krauss, R. M. (1960). The effect of threat upon inter-personal bargaining. Journal ofAbnormal and Social Psychol-ogy, 61, 181-189.

Emmons, R. A. (1986). Personal strivings: An approach to personal-ity and subjective well-being. Journal ofPersonality and SocialPsychology, 51, 1058-1068.

Emmons, R. A. (1989). The personal striving approach to personal-ity. In L. A. Pervin (Ed.), Goal concepts in personality and so-cialpsychology (pp. 87-126). Hillsdale, NJ: Lawrence ErlbaumAssociates, Inc.

Emmons, R. A., & King, L. A. (1988). Conflict among personalstrivings: Immediate and long-term implications for psycholog-ical and physical well-being. Journal ofPersonality and SocialPsychology, 54, 1040-1048.

Epstein, S. (1973). The self-concept revisited Or a theory of a the-ory. American Psychologist, 28, 404-416.

Erikson, E. H. (1963). Childhood and society (2nd ed.). New York:Norton.

Exline, J. J., & Lobel, M. (1999). The perils of outperformance: Sen-sitivity about being the target of a threatening upward compari-son. Psychological Bulletin, 125, 307-337.

Feeney, J. A. (1999). Adult romantic attachment and couple relation-ships. In J. Cassidy & P. R. Shaver (Eds.), Handbook ofattach-ment: Theory, research, and clinical applications (pp.355-377). New York: Guilford.

Fehr, B. (1988). Prototype analysis of the concepts of love and com-mitment. Journal of Personality and Social Psychology, 55,557-579.

Fehr, B. (2004). Intimacy expectations in same-sex friendships: Aprototype interaction-pattern model. Journal ofPersonality andSocial Psychology, 86, 265-284.

Fincham, F. D., & Beach, S. R. H. (1999.) Conflict in marriage: Im-plications for working with couples. Annual Review ofPsychol-ogy, 50, 47-77.

Fincham, F. D., & Bradbury, T. N. (1992). Assessing attributions inmarriage: The Relationship Attribution Measure. Journal ofPersonality and Social Psychology, 62, 457-468.

Goldsmith, D. J. (1994). The role of facework in supportive commu-nication. In B. R. Burleson, T. L. Albrecht, & I. G. Sarason(Eds.), Communication of social support (pp. 29-49). Thou-sand Oaks, CA: Sage.

Grawe, K. (2003) Psychological therapy. Seattle, WA: Hogrefe &Huber.

Grosse Holtforth, M. (2005). Fears "behind" interpersonal prob-lems in psychotherapy inpatients. Unpublished manuscript.

Grosse Holtforth, M., & Grawe, K. (2000). Fragebogen zur AnalyseMotivationaler Schemata (FAMOS). Zeitschrift fur KlinischePsychologie, 29, 170-179.

Grosse Holtforth, M., & Grawe, K. (2002). Bern inventory of treat-ment goals: Part 1. Development and first application of a tax-onomy of treatment goal themes. Psychotherapy Research, 12,79-99.

Gurtman, M. B. (1994). The circumplex as a tool for studying nor-mal and abnormal personality: A methodological primer. In S.Strack & M. Lorr (Eds.), Differentiating nor-aland abnormalpersonality (pp. 243-263). New York: Springer.

Gurtman, M. B., & Bakarishnan, J. D. (1998). Circular measurementredux: The analysis and interpretation of interpersonal circleprofiles. Clinical psychology: Science and practice, 5,344-360.

Harter, S. (1999). The construction ofthe self: A developmental per-spective. New York: Guilford.

Heider, F. (1944). Social perception and phenomenal causality. Psy-chological Review, 51, 358-374.

Hogan, R., & Roberts, B. W. (2000). A socioanalytic perspective onperson-environment interaction. In W. B. Walsh, K. H. Craik, &R. H. Price (Eds.). New directions in person-environment psy-chology. (pp. 1-24). Hillsdale, NJ: Lawrence Erlbaum Associ-ates, Inc.

Holmes, J. G. (2002). Interpersonal expectations as the buildingblocks of social cognition: An interdependence theory analysis.Personal Relationships, 9, 1-26.

Homey, K. (1945). Our inner conflicts. New York: Norton.Horowitz, L. M. (2004). Interpersonal foundations of

psychopathology. Washington, DC: American PsychologicalAssociation.

Horowitz, L. M., Alden, L. E., Wiggins, J. S., & Pincus, A. L. (2000).Inventory ofinterpersonalproblems. San Antonio, TX: Psycho-logical Corporation.

Horowitz, L. M., Krasnoperova, E. N., Tatar, D. G., Hansen, M. B.,Person, E. A., Galvin, K. L., et al. (2001). The way to consolemay depend on the goal: Experimental studies of social support.Journal ofExperimental Social Psychology, 37, 49-61.

Horowitz, L. M., Post, D. L., French, R. de S., Wallis, K. D., &Siegelman, E. Y. (1981). The prototype as a construct in abnor-mal psychology: II. Clarifying disagreement in psychiatricjudgments. Journal ofAbnormal Psychology, 90, 575-585.

Horowitz, L. M., & Wilson, K. R. (2004). Interpersonal motives andpersonality disorders. In S. N. Strack (Ed.), Handbook ofpersonology and psychopathology. New York: Wiley.

Horowitz, L. M., Wright, J. C., Lowenstein, E., & Parad, H. W.(1981). The prototype as a construct in abnormal psychology: I.A method for deriving prototypes. Journal of Abnormal Psy-chology, 90, 568-574.

Jefferson, G., & Lee, J. R. E. (1992). The rejection of advice: Man-aging the problematic convergence of a "troubles-telling" and a"service encounter." In P. Drew & J. Heritage (Eds.), Talk atwork (pp. 521-571). Cambridge, England: Cambridge Univer-sity Press.

Kagan, J. (1994). Galen 's prophecy: Temperament in human nature.New York: Basic Books.

Kagan, J., Reznick, J. S., & Snidman, N. (1988, April 8). Biologicalbases of childhood shyness. Science, 240, 167-171.

Karney, B. R., & Bradbury, T. N. (1997). Neuroticism, marital inter-action, and the trajectory of marital satisfaction. Journal ofPer-sonality and Social Psychology, 72, 1075-1092.

Kelley, H. H., & Stahelski, A. J. (1970). Errors in perception of inten-tions in a mixed-motive game. Journal ofExperimental SocialPsychology, 6, 379-400.

Kemperman, I., Russ, M. J., & Shearin, E. (1997). Self-injurious be-havior and mood regulation in borderline patients. Journal ofPersonality Disorders, 11, 146-157.

Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy forcomplementarity in human transactions. Psychological Review,90,185-214.

Kiesler, D. J. (1996). Contemporary interpersonal theory and re-search: Personality, psychopathology and psychotherapy. NewYork: Wiley.

Kiesler, D. J., Schmidt, J. A., & Wagner, C. C. (1997). A circumplexinventory of impact messages: An operational bridge betweenemotion and interpersonal behavior. In R. Plutchik & H. R.Conte (Eds.), Circumplex models ofpersonality and emotions(pp. 221-244). Washington, DC: American Psychological As-sociation.




Killingmo, B. (1989). Conflict and deficit: Implications for tech-nique. International Journal ofPsychoanalysis, 70(1), 65-79.

Klinger, E. (1987). Current concerns and disengagement from incen-tives. In F. Halisch & J. Kuhl. (Eds.), Motivation, intention, andvolition (pp. 337-347). Heidelberg, Germany: Springer-Verlag.

Lazarus, R. S. (1991). Emotion and adaptation. New York: OxfordUniversity Press.

Leary, T. F. (1957). Interpersonal diagnosis of personality. NewYork: Ronald Press.

Lehman, D. R., Ellard, J. H., & Wortman, C. B. (1986). Social sup-port for the bereaved: Recipients' and providers' perspectiveson what is helpful. Journal ofConsulting and Clinical Psychol-ogy, 54, 438-446.

Lehman, D. R., & Hemphill, K. J. (1990). Recipients' perceptions ofsupport attempts and attributions for support attempts that fail.Journal of Social and Personal Relationships, 7, 563-574.

Little, B. R. (1983). Personal projects: A rationale and method for in-vestigation. Environment and Behavior, 15, 273-309.

Locke, K. D. (2000). Circumplex Scales of Interpersonal Values: Re-liability, validity, and applicability to interpersonal problemsand personality disorders. Journal of Personality Assessment,75, 249-267.

Luborsky, L., & Crits-Christoph, P. (1998). Understanding trans-ference: The Core Conflictual Relationship Theme method(2nd ed.). Washington, DC: American Psychological Associa-tion.

Mahler, M. S., Pine, F., & Bergman, A. (1975). The psychologicalbirth of the human infant. New York: Basic Books.

Marinangeli, M. G., Butti, G., Scinto, A., DiCicco, L., Petruzzi, C.,Daneluzzo, E., et al. (2000). Patterns of comorbidity amongDSM-III-R personality disorders. Psychopathology, 33,69-74.

Markey, P. M., Funder, D. C., & Ozer, D. J. (2003). Complementarityof interpersonal behaviors in dyadic interactions. Personalityand Social Psychology Bulletin, 29, 1082-1090.

Matano, R. A., & Locke, K. D. (1995). Personality disorder scales aspredictors of interpersonal problems of alcoholics. Journal ofPersonality Disorders, 9, 62-67.

McAdams, D. P. (1985). Power, intimacy, and the life story:Personological inquiries into identity. New York: Guilford.

McLemore, C. W., & Brokaw, D. W. (1987). Personality disorders asdysfunctional interpersonal behavior. Journal of PersonalityDisorders, 1, 270-285.

McWilliams, N. (1998). Relationship, subjectivity, and inference indiagnosis. In J. W. Barron (Ed.), Making diagnosis meaningful:Enhancing evaluation and treatment ofpsychological disorders(pp. 197-226). Washington, DC: American Psychological As-sociation.

Millon, T., & Davis, R. (2000). Personality disorders in modern life.New York: Wiley.

Minuchin, S., Rosman, B. L., & Baker, L. (1978). Psychosomaticfamilies: Anorexia nervosa in context. Cambridge, MA: Har-vard University Press.

Morey, L. C. (1985). An empirical comparison of interpersonal andDSM-III approaches to classification of personality disorders.Psychiatry, 48, 358-364.

Moskowitz, D. S., & Cot6, S. (1995). Do interpersonal traits predictaffect? A comparison of three models. Journal of Personalityand Social Psychology, 69, 915-924.

Murray, H. A. (1938). Explorations in personality. New York: Ox-ford University Press.

O'Brien, T. B., & DeLongis, A. (1996). The interactional context ofproblem-, emotion-, and relationship-focused coping: The roleof the Big Five personality factors. Journal of Personality, 64,775-813.

Orford, J. (1986). The rules of interpersonal complementarity: Doeshostility beget hostility and dominance, submission? Psycho-logical Review, 93, 365-377.

Overholser, J. C. (1996). The dependent personality and interper-sonal problems. Journal of Nervous and Mental Disease,184(1), 8-16.

Pincus, A. L., & Ansell, E. B. (2003). Interpersonal theory ofperson-ality. In T. Millon & M. Lerner (Eds.), Comprehensive hand-book ofpsychology: Personality and social psychology (Vol. 5,pp. 209-229). New York: Wiley.

Pincus, A. L., & Wiggins, J. S. (1990). Interpersonal problems andconceptions of personality disorders. Journal of PersonalityDisorders, 4, 342-352.

Piper, W. E., Joyce, A. S., McCallum, M., Azim, H. F., &Ogrodniczuk, J. S. (2001). Interpretive and supportivepsychotherapies: Matching therapy and patient personality.Washington, DC: American Psychological Association.

Riediger, M., & Freund, A. M. (2004). Interference and facilitationamong personal goals: Differential associations with subjectivewell-being and persistent goal pursuit. Personality and SocialPsychology Bulletin, 30, 1511-1523.

Sadler, P., & Woody, E. (2003). Is who you are who you're talking to?Interpersonal style and complementarity in mixed-sex interac-tion. Journal ofPersonality and Social Psychology, 84, 80-96.

Santor, D. A., & Zuroff, D. C. (1997). Interpersonal responses tothreats to status and interpersonal relatedness: Effects of de-pendency and self-criticism. British Journal of Clinical Psy-chology, 36, 521-541.

Saucier, G., & Goldberg, L. R. (1996). The language of personality:Lexical perspectives on the five-factor model. In J. S. Wiggins(Ed.). The five-factor model ofpersonality. (pp. 21-50). NewYork: Guilford.

Shaver, P. R., & Hazan, C. (1988). A biased overview of the study oflove. Journal ofSocial and Personal Relationships, 5,473-501.

Shechtman, N. (2002). Talking to people versus talking to comput-ers: Interpersonal goals as a distinguishing factor Unpub-lished doctoral dissertation, Stanford University, Stanford,CA.

Shoda, Y., Mischel, W., & Wright, J. C. (1994). Intraindividual sta-bility in the organization and patterning of behavior: Incorpo-rating psychological situations into the idiographic analysis ofpersonality. Journal ofPersonality and Social Psychology, 67,674-687.

Sim, J. P., & Romney, D. M. (1990). The relationship between acircumplex model of interpersonal behaviors and personalitydisorders. Journal ofPersonality Disorders, 4, 329-341.

Soldz, S., Budman, S., Demby, A., & Merry, J. (1993). Representa-tion of personality disorders in circumplex and five-factorspace: Explorations with a clinical sample. Psychological As-sessment, 5, 41-52.

Strong, S. R., Hills, H. I., Kilmartin, C. T., DeVries, H., Lanier, K.,Nelson, B. N., et al. (1988). The dynamic relations among inter-personal behaviors: A test of complementarity andanticomplementarity. Journal of Personality and Social Psy-chology, 54, 798-8 10.

Strupp, H. H., & Binder, J. (1984). Psychotherapy in a new key:Time-limited dynamic psychotherapy. New York: BasicBooks.

Strupp, H. H., Schacht, T. E., & Henry, W. P. (1988). Problem-treat-ment-outcome congruence: A principle whose time has come.In H. Dahl & H. Kaechele (Eds.), Psychodynamic process re-search strategies. New York: Springer.

Sullivan, H. S. (1953). The interpersonal theory ofpsychiatry. NewYork: Norton.

Swann, W. B., Jr. (1996). Self-traps: The elusive quest for higherself-esteem. New York: Freeman.

Tafarodi, R. W., & Milne, A. B. (2002). Decomposing global self-es-teem. Journal ofPersonality, 70, 443-483.

Tiedens, L. Z., & Fragale, A. R. (2003). Power moves:Complementarity in dominant and submissive nonverbal behav-ior. JournalofPersonalityandSocialPsychology, 84,558-568.



HOROWITZ ET AL.

Tomasello, M. (1995). Joint attention as social cognition. In C.Moore & P. J. Dunham (Eds.), Joint attention: Its origins androle in development (pp. 103-130). Hillsdale, NJ: LawrenceErlbaum Associates, Inc.

Tracey, T. J. G. (1994). An examination of the complementarity ofinterpersonal behavior. Journal of Personality and Social Psy-chology, 67, 864-878.

Tracey, T. J. G. (2004). Levels of interpersonal complementarity: Asimple representation. Personality and Social Psychology Bul-letin, 30, 1211-1225.

Trobst, K. K. (1999). Social support as an interpersonal construct.European Journal ofPsychological Assessment, 15, 246-255.

Trull, T. J., Useda, D., Conforti, K., & Doan, B. T. (1997). Borderline per-sonality disorder features in nonclinical young adults: 2. Two-yearoutcome. Journal ofAbnormal Psychology, 106, 307-314.

Van den Boom, D. C. (1989). Neonatal irritability and the develop-ment of attachment. In G. Kohnstamm, J. Bates, & M. Rothbart(Eds.), Temperament in childhood (pp. 299-318). New York:Wiley.

Van den Boom, D. (1994). The influence of temperament and moth-ering on attachment and exploration: An experimental manipu-lation of sensitive responsiveness among lower-class motherswith irritable infants. Child Development, 65, 1457-1477.

Wagner, C. C., Kiesler, D. J., & Schmidt, J. A. (1995). Assessing theinterpersonal transaction cycle: Convergence of action and re-action interpersonal circumplex measures. Journal ofPersonal-ity and Social Psychology, 69, 938-949.

Wagner, C. C., Riley, W. T., Schmidt, J. A., McCormick, M. G. F., &Butler, S. F. (1999). Personality disorder styles and reciprocal

interpersonal impacts during outpatient intake interviews. Psy-chotherapy Research, 9, 216-23 1.

Watson, D. C., & Sinha, B. K. (1998). Comorbidity of DSM-IV per-sonality disorders in a nonclinical sample. Journal of ClinicalPsychology, 54, 773-780.

Weiss, J., & Sampson, H. (1986). The psychoanalytic process the-ory: Clinical observation and empirical research. New York:Guilford.

Wiggins, J. S. (1979). A psychological taxonomy of trait-descriptiveterms: The interpersonal domain. Journal of Personality andSocial Psychology, 37, 395-412.

Wiggins, J. S. (1982). Circumplex models of interpersonal behaviorin clinical psychology. In P. C. Kendall & J. N. Butcher (Eds.),Handbook of research methods in clinical psychology (pp.183-221). New York: Wiley.

Wiggins, J. S. (1991). Agency and communion as conceptual coordi-nates for the understanding and measurement of interpersonalbehavior. In W. Grove & D. Cicchetti (Eds.), Thinking clearlyaboutpsychology: Essays in honor ofPaul E. Meehl (Vol. 2, pp.89-113). Minneapolis: University of Minnesota Press.

Wiggins, J. S. (1996). An informal history of the interpersonalcircumplex tradition. Journal of Personality Assessment, 66,217-233.

Wiggins, J. S., & Trobst, K. K. (1997). When is a circumplex an "in-terpersonal circumplex"? The case of supportive actions. In R.Plutchik & H. R. Conte (Eds.), Circumplex models ofpersonal-ity and emotions (pp. 57-80). Washington, DC: American Psy-chological Association.'



Date post:	03-Dec-2016
Category:	Documents
Upload:	lynne
View:	215 times
Download:	1 times

How Interpersonal Motives Clarify the Meaning of Interpersonal Behavior: A Revised Circumplex Model

Documents