Sources and References

Primary textbook — Miller, R. S. (2022). Intimate Relationships (9th ed.). McGraw-Hill.

Supplementary texts — Berscheid, E., & Regan, P. (2016). The Psychology of Interpersonal Relationships. Psychology Press. | Feeney, J. A., & Noller, P. (1996). Adult Attachment. Sage. | Reis, H. T., & Sprecher, S. (Eds.). (2009). Encyclopedia of Human Relationships. Sage. | Fletcher, G. J. O., Simpson, J. A., Campbell, L., & Overall, N. C. (2019). The Science of Intimate Relationships (2nd ed.). Wiley-Blackwell. | Mikulincer, M., & Shaver, P. R. (2016). Attachment in Adulthood: Structure, Dynamics, and Change (2nd ed.). Guilford Press.

Online resources — PsycINFO, Google Scholar, JSTOR, Web of Science; Journal of Personality and Social Psychology, Personal Relationships, Journal of Social and Personal Relationships, Psychological Bulletin, Attachment & Human Development.

Chapter 1: A Conceptual Framework for Relationships

The Nature and Significance of Intimate Relationships

Human beings are fundamentally social creatures whose well-being is inextricably linked to the quality of their close relationships. Decades of research demonstrate that satisfying intimate relationships are among the strongest predictors of physical health, psychological well-being, and longevity. In a landmark meta-analysis, Holt-Lunstad, Smith, and Layton (2010) examined 148 studies encompassing over 308,000 participants and found that individuals with stronger social relationships had a 50% greater likelihood of survival compared to those with weaker ties — an effect size comparable to quitting smoking and exceeding the effects of obesity and physical inactivity. This finding alone underscores why the scientific study of interpersonal relations warrants sustained scholarly attention.

The field of close relationships emerged as a distinct area of psychological inquiry in the late 1970s and early 1980s, coalescing around several key intellectual traditions: social exchange theory from sociology, attachment theory from developmental psychology, and person perception research from social cognition. Prior to this consolidation, relationship research was fragmented across disciplines with little cross-fertilization. The founding of the International Association for Relationship Research (originally ISSPR) in 1984 and the journal Personal Relationships provided institutional homes for this burgeoning field.

Interdependence Theory: The Foundation of Relationship Science

Perhaps the single most influential theoretical framework in the study of close relationships is interdependence theory, originally formulated by Thibaut and Kelley (1959) and substantially elaborated by Kelley and Thibaut (1978) and Kelley et al. (1983, 2003). At its core, interdependence theory proposes that the defining feature of any close relationship is the degree to which partners influence one another’s outcomes — that is, their experiences, emotions, and behaviors.

The theory introduces several foundational concepts. Outcome refers to the net result of rewards minus costs experienced in an interaction. Rewards include any gratifying experience — companionship, sexual pleasure, emotional support, practical assistance — while costs encompass unpleasant experiences such as conflict, boredom, sacrifice of alternative activities, and emotional labor. According to Kelley et al. (1983), the structure of interdependence in a given situation can be described along several dimensions: the degree of mutual dependence, the degree of correspondence (or conflict) of outcomes, and the basis of dependence (whether one partner’s outcomes are primarily determined by the other’s actions or by the joint combination of both partners’ actions).

Comparison Level (CL) represents the standard against which an individual evaluates the attractiveness of a relationship. CL is shaped by past experiences in relationships, observations of others’ relationships, cultural norms, and media portrayals. A relationship in which outcomes exceed CL will be experienced as satisfying; one in which outcomes fall below CL will be experienced as dissatisfying. Critically, CL determines satisfaction but not dependence.

Comparison Level for Alternatives (CLalt) represents the lowest level of outcomes a person will accept given available alternatives. If current relationship outcomes exceed CLalt, the person will remain in the relationship; if outcomes fall below CLalt, the person will leave — regardless of their satisfaction level. This distinction elegantly explains why some people remain in unsatisfying relationships (their alternatives are even worse) and why some people leave satisfying ones (an even more attractive alternative has appeared).

The concept of transformation of motivation is central to understanding how partners in close relationships move beyond narrow self-interest. When faced with a situation where self-interest and partner interest conflict, individuals may psychologically transform the situation — setting aside their immediate personal preferences in favor of broader relational goals such as fairness, long-term harmony, or the partner’s well-being. The degree to which partners engage in transformation of motivation is a key indicator of relationship health and commitment.

Communal Versus Exchange Relationships

Clark and Mills (1979, 1993) drew an influential distinction between two fundamentally different orientations that govern the giving and receiving of benefits in relationships. In exchange relationships, benefits are given with the expectation of receiving comparable benefits in return. The norm is one of reciprocity and equity: people keep mental accounts of inputs and outputs, and imbalances create distress. Exchange norms typically govern interactions with acquaintances, business associates, and strangers.

In communal relationships, by contrast, benefits are given in response to the partner’s needs or to demonstrate care, without expectation of specific repayment. People in communal relationships do not keep score; indeed, Clark (1984) demonstrated experimentally that tracking the other person’s inputs actually decreases satisfaction in communal relationships while increasing satisfaction in exchange relationships. Communal norms characterize close friendships, family bonds, and romantic relationships.

Clark and Mills further showed that the desire to have a communal relationship with someone (communal strength) varies in intensity. Communal strength toward a romantic partner is typically very high — people feel considerable responsibility for their partner’s welfare and experience distress when they cannot meet their partner’s needs. Mills, Clark, Ford, and Johnson (2004) developed the Communal Strength scale to measure individual differences in the motivation to respond to a specific partner’s needs.

The Intimacy Process Model

Reis and Shaver (1988) proposed a process model of intimacy that moved the field beyond trait-based and static conceptualizations toward a dynamic, interaction-level understanding. According to this model, intimacy is not a property of a person or even of a relationship but rather an emergent quality of specific interactions between two people.

The model describes a transactional sequence. Person A engages in self-disclosure — sharing personal feelings, information, or experiences. Person B responds to this disclosure. Whether intimacy is generated depends critically on how Person A perceives Person B’s response. If A perceives that B’s response reflects understanding (B accurately grasps the content and meaning of the disclosure), validation (B communicates that A’s thoughts and feelings are legitimate), and caring (B expresses concern and affection for A), then the interaction will be experienced as intimate. This model emphasizes that intimacy requires both disclosure and perceived responsiveness. One without the other is insufficient.

Reis and Patrick (1996) elaborated the model further, proposing that perceived partner responsiveness — the belief that one’s partner understands, validates, and cares for the self — is the core organizing principle underlying all forms of intimacy and relationship satisfaction. This construct has since become one of the most widely studied variables in the field, with Reis, Clark, and Holmes (2004) arguing that it is the active ingredient common to diverse relationship processes including attachment, social support, and capitalization.

Relationship Types and Their Defining Features

Not all close relationships are alike, and the field has developed several typologies. Intimate relationships can be distinguished from other social bonds by four key features: (1) knowledge — partners have extensive personal and private information about each other; (2) caring — each partner feels genuine concern for the other’s well-being; (3) interdependence — partners have frequent, strong, and diverse effects on each other across many different activities; and (4) mutuality — both partners think of themselves as a couple and incorporate the relationship into their self-concept.

Relationship scholars have also drawn attention to relationship diversity — the fact that intimate relationships vary enormously across cultures, historical periods, sexual orientations, and individual circumstances. Same-sex relationships, cohabiting partnerships, long-distance relationships, arranged marriages, and polyamorous arrangements all represent valid forms of intimate connection that the field increasingly recognizes and studies. Kurdek (2005), for example, demonstrated in a five-year longitudinal study of 95 lesbian, 92 gay male, and 226 heterosexual couples that relationship processes (satisfaction trajectories, predictors of dissolution) operate remarkably similarly across couple types, though some differences emerge in areas like social support from families of origin.

The Influence of Culture and Historical Context

The study of interpersonal relationships cannot be divorced from its cultural context. Relationships are embedded within cultural systems that prescribe norms about love, marriage, gender roles, family obligations, and sexual behavior. Individualistic cultures (common in Western nations) tend to emphasize romantic love as a prerequisite for marriage, personal fulfillment within relationships, and individual autonomy. Collectivistic cultures often emphasize family approval, filial obligations, pragmatic considerations in partner selection, and the couple’s integration into broader family networks.

Historically, the concept of marrying for love is relatively recent in Western history, emerging prominently only in the eighteenth century. Finkel, Hui, Carswell, and Larson (2014) have argued that contemporary Western marriages have ascended what they call “Mount Maslow” — modern partners increasingly expect their marriage to fulfill not just basic needs for safety and belonging but also higher-order needs for self-actualization and personal growth. This “suffocation of marriage” thesis suggests that while the best modern marriages may be more fulfilling than ever before, the average marriage may suffer because expectations have risen faster than the investments of time and energy needed to meet them.

Chapter 2: Research Methods in Relationship Science

The Challenge of Studying Relationships

Relationships present unique methodological challenges that do not arise in most areas of psychology. The fundamental unit of analysis is dyadic — it involves two interdependent people whose data are inherently non-independent. Traditional statistical methods assume independence of observations, making them inappropriate for relationship data without modification. Kenny, Kashy, and Cook (2006) developed the Actor-Partner Interdependence Model (APIM), which has become the standard analytic framework for dyadic data. The APIM simultaneously estimates actor effects (how one’s own characteristics predict one’s own outcomes) and partner effects (how the partner’s characteristics predict one’s outcomes), while accounting for the nonindependence of dyadic data.

Beyond the dyadic nature of the data, relationship researchers face the challenge that many of the phenomena they study — love, trust, conflict, intimacy — unfold over time, are deeply private, and are resistant to experimental manipulation for ethical reasons. No ethics board would approve randomly assigning people to relationships, inducing trust violations, or causing breakups. These constraints have led to creative methodological solutions that span the full range of research designs.

Self-Report Methods and Their Limitations

The most common data collection method in relationship research is self-report questionnaires. Instruments such as the Dyadic Adjustment Scale (Spanier, 1976), the Relationship Assessment Scale (Hendrick, 1988), the Perceived Relationship Quality Components inventory (Fletcher, Simpson, & Thomas, 2000), and the Couples Satisfaction Index (Funk & Rogge, 2007) are widely used. Self-report data are efficient and can capture subjective experiences that are inaccessible through other methods.

However, self-reports are subject to well-documented biases. Social desirability leads respondents to present their relationships in an overly positive light. Sentiment override (Weiss, 1980) means that global feelings about the relationship color reports of specific behaviors — happy partners recall more positive interactions than actually occurred, while unhappy partners recall more negative ones. Retrospective bias distorts memories of past relationship experiences, as current satisfaction heavily influences how partners remember earlier events.

To address these limitations, researchers often use multi-method approaches, combining self-reports with behavioral observations, physiological measures, and partner reports. The convergence (or divergence) between different data sources provides important information about the phenomenon under study.

Diary Methods and Experience Sampling

Diary methods have become increasingly popular in relationship research because they capture experiences close in time to when they actually occur, reducing retrospective bias. The Rochester Interaction Record (RIR), developed by Wheeler and Nezlek (1977), was one of the earliest interaction diary methods. Participants complete a brief standardized form after every social interaction lasting ten minutes or more, recording details such as the interaction partner, duration, intimacy, satisfaction, and who initiated the encounter. Reis and Wheeler (1991) used this method to demonstrate that the quantity of social interaction is less important for well-being than its quality — particularly the quality of the most intimate interactions.

Modern diary studies typically employ daily diary designs, in which participants complete questionnaires at the end of each day for periods ranging from one week to several months, or experience sampling methods (ESM), in which participants respond to multiple random prompts throughout the day. Bolger, Davis, and Rafaeli (2003) provided a comprehensive guide to diary methods, noting their advantages for studying within-person processes (how individuals fluctuate around their own averages) and between-person differences in those processes. For example, Laurenceau, Barrett, and Pietromonaco (1998) used daily diaries to test Reis and Shaver’s intimacy process model, finding that self-disclosure and perceived partner responsiveness each contributed independently to daily feelings of intimacy.

Observational Methods

Behavioral observation provides objective data about how partners actually interact, free from the biases inherent in self-report. The most influential observational paradigm in relationship research is John Gottman’s laboratory procedure, in which couples are videotaped discussing an area of ongoing disagreement in their relationship for a fixed period (typically 15 minutes). Trained coders then rate each speaking turn using coding systems such as the Specific Affect Coding System (SPAFF), which categorizes each moment of interaction into one of roughly 20 affective codes (e.g., contempt, defensiveness, affection, humor, sadness, anger).

Gottman’s (1994) research using this paradigm identified what he called the “Four Horsemen of the Apocalypse” — criticism, contempt, defensiveness, and stonewalling — as behavioral patterns that predict relationship dissolution with remarkable accuracy. In one study, Gottman and Levenson (1992) achieved 93% accuracy in predicting which couples would divorce over a 14-year period based on observational coding of a single 15-minute conflict discussion, though subsequent analyses have suggested somewhat lower predictive accuracy when cross-validated.

Beyond Gottman’s system, other observational approaches include the Interactional Dimensions Coding System (IDCS) used by Julien, Markman, and colleagues, and the Rapid Marital Interaction Coding System (RMICS) developed by Heyman and colleagues, which is designed for efficient coding of large datasets. The still-face paradigm, originally developed by Tronick, Als, Adamson, Wise, and Brazelton (1978) for studying infant-caregiver interaction, has also been adapted for adult relationship research. In this paradigm, one partner suddenly becomes expressionless and unresponsive during a conversation, and the other partner’s reactions are observed. This paradigm reveals the powerful role of partner responsiveness — and its sudden absence — in regulating emotional experience.

Longitudinal and Cross-Sectional Designs

Because relationships are dynamic entities that evolve over time, longitudinal designs are essential. Panel studies follow the same individuals or couples over extended periods, assessing key variables at multiple time points. The Early Years of Marriage (EYM) project, led by Orbuch and colleagues, followed 373 couples from their wedding day over 25+ years. Lavner and Bradbury (2010) followed newlyweds for four years and found that initial satisfaction levels — not the rate of decline — were the strongest predictor of later divorce, challenging the common assumption that relationships decline inevitably.

Cross-sectional designs compare different individuals or couples at a single time point. While useful for establishing associations, they cannot establish temporal precedence or rule out third-variable explanations. A cross-sectional finding that trust correlates with satisfaction cannot tell us whether trust causes satisfaction, satisfaction causes trust, or both are caused by a third variable such as personality.

Experimental and Quasi-Experimental Approaches

True experiments in relationship research are relatively rare but provide uniquely strong evidence for causal claims. Researchers have developed creative paradigms to study relationship processes experimentally. Aron, Melinat, Aron, Vallone, and Bator (1997) demonstrated that pairs of strangers could generate feelings of closeness in a 45-minute laboratory session by engaging in an escalating series of self-disclosure exercises. This “fast friends” paradigm has become widely used in research on closeness, cross-group friendships, and relationship formation.

Quasi-experimental designs exploit naturally occurring groups or events. For example, researchers have compared couples undergoing a specific life transition (e.g., the birth of a first child) with matched couples who have not yet experienced that transition. Intervention studies, in which couples are randomly assigned to receive a relationship education program or a control condition, also provide quasi-experimental evidence about the malleability of relationship processes.

Speed-Dating as a Research Paradigm

The emergence of speed-dating studies in the early 2000s represented a significant methodological advance. Finkel, Eastwick, and Matthews (2007) pioneered this approach, in which participants attend organized speed-dating events and rate each potential partner on multiple dimensions after brief (typically 4-minute) interactions. Because each participant rates and is rated by multiple partners, speed-dating designs disentangle perceiver effects (some people like everyone), target effects (some people are liked by everyone), and relationship effects (unique chemistry between two specific people).

This paradigm has yielded important findings that challenge conventional wisdom. Eastwick and Finkel (2008) found that although participants reported before the event that physical attractiveness was more important to men and earning potential more important to women (consistent with evolutionary predictions), these stated preferences did not predict actual mate choices during the speed-dating event. Both men and women were equally influenced by physical attractiveness and equally uninfluenced by earning potential in their real-time romantic interest.

Chapter 3: Attraction

Physical Attractiveness: More Than Skin Deep

Physical attractiveness exerts a powerful influence on interpersonal attraction, and its effects extend far beyond the domain of romantic relationships. Langlois, Kalakanis, Rubenstein, Larson, Hallam, and Smoot (2000) conducted a comprehensive meta-analysis and found that attractive individuals are judged more positively, treated more positively, and exhibit more positive behaviors and traits than unattractive individuals — the “what is beautiful is good” stereotype identified by Dion, Berscheid, and Walster (1972).

What makes a face attractive? Research has converged on several features. Averageness — the degree to which a face approximates the population mean — is consistently rated as attractive, likely because average features signal genetic diversity and developmental stability (Langlois & Roggman, 1990). Symmetry is also valued, particularly by women evaluating male faces, and is thought to signal good genes and resistance to developmental perturbations. Sexual dimorphism plays a role as well: women generally prefer masculine features in male faces (strong jaw, prominent brow ridges) especially at peak fertility, while men prefer feminine features in female faces (large eyes, full lips, small chin).

The classic matching hypothesis was first tested by Walster, Aronson, Abrahams, and Rottman (1966) in their famous “Computer Dance” study at the University of Minnesota. Researchers randomly paired 376 first-year students for a dance and assessed physical attractiveness, personality, and intelligence. The central finding was striking: the only significant predictor of how much a person liked their date and wanted to see them again was the date’s physical attractiveness. Intelligence, personality, and even similarity had negligible effects. Although the original matching hypothesis predicted that people would prefer partners of similar attractiveness, this first test actually supported an aspirational pattern — everyone preferred the most attractive partner. Subsequent research using designs that allowed for actual choice (rather than random assignment) has generally supported the matching hypothesis: formed couples tend to be similar in attractiveness (Feingold, 1988).

Proximity, Familiarity, and the Mere Exposure Effect

Proximity — physical nearness — is one of the most robust predictors of relationship formation, and its effects were documented in the classic Festinger, Schachter, and Back (1950) study of friendship formation in MIT housing. Residents were most likely to become friends with those living in adjacent apartments or near stairwells and mailboxes. In modern life, proximity operates not only through physical nearness but also through digital proximity — the frequency of encountering someone online.

The mechanism underlying proximity’s effect is largely the mere exposure effect (Zajonc, 1968): repeated exposure to a stimulus increases liking for it, provided initial reactions are not strongly negative. Moreland and Beach (1992) demonstrated this in a classroom setting, where a confederate attended a large lecture class either 0, 5, 10, or 15 times. Although the confederate never interacted with other students, those who saw her more frequently rated her as more attractive, more similar to themselves, and more likeable.

Similarity and Complementarity

The adage “opposites attract” has little empirical support. Instead, decades of research confirm that similarity — in attitudes, values, personality, demographics, and even physical attractiveness — is a strong predictor of initial attraction and relationship satisfaction. Byrne’s (1971) law of attraction proposed that attraction is a linear function of the proportion of similar attitudes shared by two people. While subsequent research has shown this formulation to be oversimplified, the core finding is robust.

Montoya, Horton, and Kirchner (2008) conducted a meta-analysis distinguishing between perceived similarity (how similar one believes a partner to be) and actual similarity (objectively measured overlap). They found that perceived similarity was a much stronger predictor of attraction than actual similarity, suggesting that the belief that “we are alike” may matter more than the objective reality. In existing relationships, perceived similarity predicted satisfaction more strongly than actual similarity did.

The question of whether complementarity — having different but interlocking traits — contributes to attraction has received mixed support. Dryer and Horowitz (1997) found that dominant individuals preferred submissive partners and vice versa in initial interactions, suggesting complementarity on the dominance-submissiveness dimension. However, this effect appears limited to the dominance dimension and does not extend to personality more broadly.

Online Dating and Algorithmic Matching

The rise of online dating has fundamentally altered the landscape of mate selection. By 2020, approximately 40% of heterosexual couples and nearly 70% of same-sex couples in the United States met online (Rosenfeld, Thomas, & Hausen, 2019). Online dating expands the pool of potential partners far beyond one’s immediate social network but also introduces unique dynamics.

Finkel, Eastwick, Karney, Reis, and Sprecher (2012) published an extensive review of online dating research and concluded that while dating sites increase the quantity of potential partners, their matching algorithms show little evidence of improved match quality over random pairing. The proprietary algorithms used by sites like eHarmony (based on similarity matching) and OkCupid (based on user-stated preferences) have not been validated by independent research. Joel, Eastwick, and Finkel (2017) applied machine learning to a large speed-dating dataset and found that individual attributes (actor effects and partner effects) predicted only about 5% of the variance in romantic desire — the vast majority of variance was attributable to the unique dyadic combination of two specific people, which no algorithm can predict in advance.

Cross-Cultural Variation in Mate Preferences

Buss (1989) conducted a landmark study across 37 cultures involving over 10,000 participants, examining mate preferences. Several findings emerged with remarkable cross-cultural consistency: in virtually all cultures, both men and women valued kindness and intelligence highly. However, consistent sex differences also appeared: men across cultures placed greater emphasis on physical attractiveness and youth, while women placed greater emphasis on financial resources and social status.

These findings have been interpreted through multiple theoretical lenses. Evolutionary psychologists argue that sex differences in mate preferences reflect evolved adaptations — women’s preference for resource-holding partners reflects the adaptive problem of securing paternal investment, while men’s preference for youth and beauty reflects the adaptive problem of identifying fertile mates. Social role theorists (Eagly & Wood, 1999) counter that these differences reflect gender inequality: in cultures where women have greater economic power and gender equality is higher, sex differences in mate preferences shrink. Zentner and Mitura (2012) provided cross-national support for this position, showing that gender differences in mate preferences diminished as national gender equity increased.

Chapter 4: The Self in Relationships

Attachment Styles and the Self

One of the most consequential ways in which the self shapes relationships is through attachment style — the systematic pattern of expectations, emotions, and behaviors that individuals bring to close relationships based on their history of interactions with attachment figures. Bartholomew and Horowitz (1991) proposed a four-category model based on two underlying dimensions: model of self (positive vs. negative, reflecting anxiety about abandonment) and model of others (positive vs. negative, reflecting avoidance of intimacy). This yields four attachment styles: secure (positive self, positive other), preoccupied (negative self, positive other), dismissive-avoidant (positive self, negative other), and fearful-avoidant (negative self, negative other).

These internal working models profoundly shape how individuals perceive and behave in relationships. Securely attached individuals approach relationships with confidence, communicate openly about their needs, and respond constructively to partner distress. Anxiously attached individuals are hypervigilant to signs of rejection, experience intense jealousy, and engage in protest behaviors (excessive calling, emotional manipulation) when they perceive threats to the relationship. Avoidantly attached individuals maintain emotional distance, suppress attachment-related thoughts and feelings, and withdraw from partners during times of stress.

Inclusion of Other in the Self

Aron and Aron’s (1986) self-expansion model proposes that a fundamental human motivation is the desire to expand the self — to acquire new resources, perspectives, identities, and capabilities. Close relationships serve this motive because partners incorporate aspects of each other into their own self-concept. Aron, Aron, and Smollan (1992) developed the Inclusion of Other in the Self (IOS) Scale, a single-item pictorial measure consisting of seven pairs of increasingly overlapping circles, where respondents select the pair that best represents their relationship.

Despite its simplicity, the IOS Scale has demonstrated strong psychometric properties and impressive predictive validity. It correlates highly with more complex measures of closeness and commitment, and it predicts relationship stability over time. Aron, Aron, Tudor, and Nelson (1991) demonstrated the cognitive underpinnings of self-expansion: when partners are included in the self, people show confusion between self and partner traits in reaction-time tasks — they are slower to distinguish “my traits” from “my partner’s traits” than from a stranger’s traits, because partner traits have literally become part of the self-concept.

The self-expansion model also predicts that novel and arousing shared activities enhance relationship quality by recapturing the rapid self-expansion that characterizes the early stages of relationships. Aron, Norman, Aron, McKenna, and Heyman (2000) tested this experimentally by assigning couples to engage in either a novel and exciting activity (completing an obstacle course while bound together) or a mundane activity (rolling a ball across the floor). Couples in the novel-exciting condition showed greater increases in relationship satisfaction, supporting the self-expansion prediction.

Self-Verification Versus Self-Enhancement

A fascinating tension exists between two motives that operate in relationships: the desire for self-enhancement (wanting partners to see us more positively than we see ourselves) and the desire for self-verification (wanting partners to see us as we truly see ourselves, even if that view is negative).

Swann, De La Ronde, and Hixon (1994) found that the relative dominance of these motives depends on the stage and type of relationship. In dating relationships and early interactions, self-enhancement motives prevail — people prefer partners who view them positively. However, in committed marriages, self-verification motives become stronger — people prefer partners whose views of them are congruent with their own self-views. This shift has provocative implications: a person with low self-esteem may initially enjoy a partner’s idealization but over time become uncomfortable with it, preferring a partner who sees their flaws accurately. Swann argued that self-verification provides a sense of coherence and predictability that is essential for navigating daily life.

However, this account has been challenged by Murray, Holmes, and Griffin (1996a, 1996b), who found that partner idealization — seeing one’s partner as better than the partner sees themselves — consistently predicted greater satisfaction in both dating and married couples. Murray and Holmes proposed that idealization functions as a buffer against the inevitable disappointments of relationship life, and that over time, targets of idealization actually come to live up to their partners’ positive illusions.

The Michelangelo Phenomenon

Rusbult, Finkel, and Kumashiro (2009) introduced the Michelangelo phenomenon, an elegant model describing how close partners sculpt each other toward their ideal selves. Just as Michelangelo reportedly described sculpture as revealing the form already hidden within the stone, close partners can affirm and elicit behaviors that help each other move closer to their ideal self — the person they aspire to become.

The model distinguishes between partner affirmation (the partner perceives and behaves toward the individual in ways consistent with the individual’s ideal self) and partner disaffirmation (the partner perceives and elicits behaviors that move the individual away from the ideal self). When partners affirm each other’s ideal selves, individuals make greater progress toward their personal goals, experience greater personal well-being, and report higher relationship satisfaction. Drigotas, Rusbult, Wieselquist, and Whitton (1999) provided initial evidence across five studies, demonstrating that partner affirmation predicted movement toward the ideal self, which in turn predicted both personal and relational well-being.

Relational Self-Construal

Beyond individual attachment styles and self-expansion processes, researchers have examined how broadly people define their self-concept in relational terms. Relational self-construal refers to the degree to which individuals include their close relationships as a central component of their identity. Cross, Bacon, and Morris (2000) developed the Relational-Interdependent Self-Construal (RISC) scale, finding that individuals high in relational self-construal are more attentive to relationship information, more responsive to relationship threats, and more willing to accommodate their partners.

This construct is distinct from collectivism. Cross and Madson (1997) proposed that gender differences in self-construal — with women being more relationally interdependent than men — help explain many observed gender differences in relationship behavior, including women’s greater self-disclosure, emotional expressiveness, and relationship maintenance efforts. However, effect sizes for these gender differences are typically small to moderate, and substantial variability exists within each gender.

Chapter 5: Communication

The Centrality of Communication in Relationships

Communication is the primary vehicle through which relationships are created, maintained, and dissolved. The quality of communication between partners is one of the most consistent predictors of relationship satisfaction across diverse populations. Gottman (1994) famously proposed that the ratio of positive to negative interactions during conflict is a key diagnostic indicator: stable, happy couples maintain a ratio of approximately 5:1 (five positive interactions for every negative one), while couples heading for divorce show ratios closer to 0.8:1.

Verbal and Nonverbal Communication

Communication in close relationships occurs through both verbal and nonverbal channels, and the interplay between them is critical. Nonverbal communication — including facial expressions, tone of voice, posture, touch, and eye contact — often carries more emotional weight than the verbal content of messages. Noller (1984) conducted pioneering research on nonverbal communication accuracy in marriage, finding that unhappy couples showed deficits in both encoding (sending clear nonverbal messages) and decoding (accurately interpreting the partner’s nonverbal messages). Critically, these deficits were specific to the marital relationship — unhappy spouses were not less skilled communicators in general, but something about their relationship dynamics impaired their nonverbal accuracy with each other.

Sentiment override (Weiss, 1980) describes the tendency for global feelings about the relationship to override the specific content of communication. In positive sentiment override, partners in happy relationships interpret ambiguous messages charitably and respond positively even to mildly negative messages. In negative sentiment override, partners in unhappy relationships interpret even neutral or positive messages negatively. Hawkins, Carrere, and Gottman (2002) demonstrated that sentiment override mediates the link between relationship satisfaction and behavioral responses during conflict — dissatisfied partners show negative sentiment override, leading them to react negatively even to their partner’s repair attempts.

Gottman’s Sound Relationship House Theory

John Gottman’s research program, spanning over four decades and involving the observation of thousands of couples, has produced one of the most comprehensive models of relationship communication and stability. The Sound Relationship House theory (Gottman, 1999; Gottman & Silver, 1999) proposes that stable relationships rest on seven interlocking components arranged in a metaphorical house.

The foundation consists of Love Maps — detailed cognitive representations of the partner’s inner psychological world, including their worries, dreams, preferences, and life narrative. Above this is Fondness and Admiration — maintaining a fundamentally positive view of the partner’s character. The third level is Turning Toward versus turning away — the extent to which partners respond to each other’s bids for emotional connection. Gottman and DeClaire (2001) define a bid as any attempt to connect emotionally, ranging from a simple comment (“Look at that bird”) to a request for emotional support. In their newlywed apartment laboratory studies, Gottman found that couples who divorced within six years had turned toward each other’s bids only 33% of the time, compared to 86% for couples who remained married.

The remaining levels include Positive Perspective (sentiment override), Manage Conflict (rather than resolve it — Gottman argues that 69% of marital conflicts are perpetual and unresolvable), Make Life Dreams Come True (supporting each other’s aspirations), and Create Shared Meaning (developing a shared culture, rituals, and life narrative).

Emotionally Focused Therapy Communication Patterns

Sue Johnson’s Emotionally Focused Therapy (EFT) model (Johnson, 2004) draws heavily on attachment theory to understand communication patterns in distressed couples. EFT identifies characteristic negative interaction cycles — rigid, self-reinforcing patterns that maintain distress. The most common is the demand-withdraw or pursue-withdraw pattern, in which one partner (typically the more anxiously attached) escalates emotional demands while the other (typically the more avoidantly attached) withdraws emotionally and physically. Each partner’s behavior triggers the other’s, creating a vicious cycle.

Johnson argues that beneath the surface behaviors of pursuing and withdrawing lie unmet attachment needs — the pursuer’s desperate need for reassurance and connection, and the withdrawer’s desperate need to avoid feeling inadequate or overwhelmed. The therapeutic process involves helping couples recognize their negative cycle, access the vulnerable emotions underlying their positions, and develop new interactional patterns that create a secure bond.

Self-Disclosure and Responsiveness

The reciprocal exchange of self-disclosure is a fundamental engine of relationship development. Altman and Taylor’s (1973) social penetration theory proposed that relationships develop through gradually increasing self-disclosure along two dimensions: breadth (the range of topics discussed) and depth (the degree of personal significance and vulnerability of shared information). Early relationships are characterized by broad but shallow disclosure; as trust develops, disclosure becomes both deeper and narrower, focusing on core aspects of the self.

Sprecher, Treger, Wondra, Hilaire, and Wallpe (2013) demonstrated that sustained, reciprocal self-disclosure generates closeness even between strangers, replicating and extending the Aron et al. (1997) fast-friends paradigm. They further found that the listener’s role is as important as the discloser’s — responsive listening drives closeness from the listener’s perspective as well.

Chapter 6: Commitment

The Investment Model

Caryl Rusbult’s Investment Model (Rusbult, 1980, 1983) is the most extensively tested and supported framework for understanding commitment in close relationships. Rooted in interdependence theory, the model proposes that commitment — the subjective experience of dependence on and intention to persist in a relationship — is determined by three independent factors: satisfaction (positive affect derived from the relationship, reflecting the degree to which the relationship exceeds one’s comparison level), quality of alternatives (the perceived desirability of one’s best available alternative to the current relationship), and investment size (the magnitude and importance of resources that are tied to and would be lost upon leaving the relationship).

Investments include both intrinsic investments (resources put directly into the relationship, such as time, emotional energy, and self-disclosure) and extrinsic investments (resources that become attached to the relationship, such as mutual friends, shared possessions, children, and shared memories). According to the model, commitment increases with satisfaction and investment size, and decreases with quality of alternatives.

Le and Agnew (2003) conducted a meta-analysis of 52 studies involving over 11,000 participants and found strong support for the investment model. Satisfaction (r = .68), alternatives (r = -.48), and investments (r = .46) all predicted commitment in the expected directions. Commitment, in turn, was the strongest predictor of relationship persistence (r = .47). The model has been validated across diverse relationship types, including heterosexual dating and married couples, same-sex couples, friendships, and even commitment to organizations and athletic teams.

Johnson’s Tripartite Model of Commitment

Michael Johnson (1991, 1999) proposed a tripartite model distinguishing three qualitatively different types of commitment. Personal commitment reflects the positive desire to continue the relationship, rooted in attraction, love, and relationship identity. Moral commitment reflects a sense of obligation to the relationship based on personal values (e.g., “marriage is forever”) and felt obligations to the partner. Structural commitment reflects external constraints that make leaving difficult, including financial barriers, legal obstacles, social pressure, lack of alternatives, and concern for children.

Johnson argued that these three components are independent and can combine in different patterns. A person may stay in a relationship primarily out of personal desire, primarily out of felt obligation, or primarily because the costs of leaving are prohibitively high. Each combination is associated with different relationship dynamics. Relationships sustained primarily by structural commitment (in the absence of personal commitment) are likely to be experienced as “traps,” while those sustained by personal commitment are experienced as choices.

Adams and Jones (1997) developed measures to assess all three forms and found that personal commitment was the strongest predictor of marital satisfaction, while structural commitment was the strongest predictor of marital stability. This dissociation is important: the factors that keep people satisfied are not identical to the factors that keep them together.

Sacrifice and Transformation of Motivation

A key behavioral expression of commitment is sacrifice — the willingness to forgo self-interest for the benefit of the partner or relationship. Van Lange, Rusbult, Drigotas, Arriaga, Witcher, and Cox (1997) found that more committed individuals were more willing to sacrifice personal interests, and that this willingness was driven by the transformation of motivation described by interdependence theory. Committed individuals psychologically transform situations of conflicting interests by adopting a long-term, partner-oriented perspective.

Importantly, the experience of sacrifice differs depending on the motivation behind it. Impett, Gable, and Peplau (2005) distinguished between approach-motivated sacrifice (sacrificing to promote positive outcomes, such as making the partner happy) and avoidance-motivated sacrifice (sacrificing to prevent negative outcomes, such as avoiding conflict). Approach-motivated sacrifice was associated with greater positive emotions, relationship satisfaction, and personal well-being, while avoidance-motivated sacrifice was associated with negative emotions and lower satisfaction.

The Commitment Calibration Hypothesis

Joel, MacDonald, and Shimotomai (2011) proposed the commitment calibration hypothesis, which suggests that people are not always well-calibrated in their commitment — they sometimes remain too committed to poor relationships and sometimes are insufficiently committed to good ones. This miscalibration can result from investment size bias (people overweight sunk costs), from the status quo bias (preferring to maintain the current state of affairs), and from the psychological difficulty of accurately evaluating complex relationship dynamics. The hypothesis draws attention to the fact that commitment is not always adaptive and that the mechanisms described by the investment model can sometimes trap people in unhealthy relationships.

Chapter 7: Attachment

Bowlby’s Attachment Theory and Its Extension to Adults

John Bowlby (1969/1982, 1973, 1980) developed attachment theory to explain the deep emotional bonds between infants and caregivers. He proposed that the attachment system is an evolved behavioral system that motivates proximity-seeking to protective figures (attachment figures) under conditions of threat or distress. Through repeated interactions with caregivers, children develop internal working models — cognitive-affective representations of the self and others that guide expectations and behavior in close relationships throughout life.

Hazan and Shaver (1987) made the seminal extension of attachment theory to adult romantic relationships, proposing that romantic love is an attachment process and that individual differences in adult attachment mirror the three infant attachment styles identified by Ainsworth, Blehar, Waters, and Wall (1978). They found that the distribution of attachment styles in adults roughly paralleled that in infants: approximately 56% secure, 25% avoidant, and 19% anxious-ambivalent.

Measurement: Categories Versus Dimensions

A major debate in adult attachment research concerns whether attachment is best conceptualized as categorical types or continuous dimensions. Bartholomew and Horowitz (1991) expanded the three-category model to four categories by distinguishing two forms of avoidance (dismissive and fearful). Brennan, Clark, and Shaver (1998) developed the Experiences in Close Relationships (ECR) scale, which measures two continuous dimensions: attachment anxiety (preoccupation with abandonment and rejection) and attachment avoidance (discomfort with closeness and dependence).

Fraley and Waller (1998) applied taxometric analyses to attachment data and found that the data were better fit by a dimensional model than a categorical one — attachment differences exist on a continuum rather than in discrete types. Fraley (2002) developed the ECR-Revised (ECR-R) to improve measurement precision at the dimensional endpoints. Nevertheless, categorical language (secure, anxious, avoidant) remains widely used in the field because it is intuitive and facilitates communication, even though researchers generally acknowledge the dimensional nature of the underlying constructs.

The Adult Attachment Interview

While the ECR and similar self-report measures assess attachment representations in the context of current romantic relationships, the Adult Attachment Interview (AAI), developed by George, Kaplan, and Main (1984/1996), takes a very different approach. The AAI is a semi-structured interview that asks adults to describe their childhood relationships with parents and to evaluate the influence of these early experiences on their current functioning. Crucially, the AAI is coded not for the content of what is said but for the coherence and organization of the narrative.

Main and Hesse (1990) identified four AAI classifications: autonomous/secure (coherent, valuing of attachment), dismissing (normalizing or dismissive of negative experiences, idealizing parents without supporting memories), preoccupied (confused, angry, or passive in discussing attachment experiences), and unresolved/disorganized (lapses in reasoning or discourse when discussing loss or trauma). The AAI has impressive predictive validity: a parent’s AAI classification predicts their infant’s attachment classification with approximately 75% accuracy, even when assessed before the child is born (van IJzendoorn, 1995).

Attachment, Caregiving, and Sexuality

Attachment theory describes not just attachment-seeking behavior but also two closely related behavioral systems: the caregiving system and the sexual/mating system. Collins and Feeney (2000) studied how attachment style influences caregiving in romantic relationships. They found that secure individuals provided more effective support — they were more responsive to their partner’s needs, provided comfort that matched the situation’s demands, and enabled their partners to explore and grow. Avoidant individuals tended to provide less support, while anxious individuals provided support that was often excessive, intrusive, or focused on their own needs rather than the partner’s.

Regarding sexuality, Davis, Shaver, and Vernon (2004) found that attachment anxiety was associated with using sex to feel closer and to reduce insecurity (sex motivated by attachment needs), while attachment avoidance was associated with casual, uncommitted sexual behavior and discomfort with sexual intimacy in ongoing relationships. Birnbaum, Reis, Mikulincer, Gillath, and Orpaz (2006) found that avoidant individuals reported less sexual satisfaction and used sex as a means of maintaining distance rather than closeness.

Earned Security and Attachment Change

While attachment orientations show moderate stability over time, they are not immutable. The concept of earned security (Main, Goldwyn, & Hesse, 2003) refers to individuals who experienced difficult or insecure childhoods but have developed coherent, integrated narratives about their attachment experiences, resulting in secure AAI classifications. Earned-secure individuals function similarly to continuously-secure individuals in their parenting and relationship behavior, suggesting that the working through of early adverse experiences — often through therapy, corrective relationship experiences, or self-reflection — can fundamentally alter attachment representations.

Fraley, Vicary, Brumbaugh, and Roisman (2011) examined attachment stability in a large internet sample retested over several months and found moderate stability (test-retest correlations around .55-.65) with meaningful individual change. Both positive relationship experiences (entering a satisfying relationship) and negative ones (breakup, betrayal) were associated with changes in attachment orientation, supporting a revisionist view of attachment rather than a strictly prototype (unchangeable) view.

Chapter 8: Theories of Love

Taxonomies of Love: Sternberg, Lee, and Fehr

Robert Sternberg’s (1986) Triangular Theory of Love proposes that love consists of three components: intimacy (feelings of closeness, connectedness, and bondedness), passion (drives that lead to romance, physical attraction, and sexual consummation), and commitment (the decision that one loves someone and the commitment to maintain that love). Different combinations of these components yield different types of love: romantic love (intimacy + passion), companionate love (intimacy + commitment), fatuous love (passion + commitment), consummate love (all three), infatuation (passion only), empty love (commitment only), liking (intimacy only), and nonlove (none).

Lee (1973) proposed a typology of love styles drawn from classical Greek concepts: eros (passionate, romantic love), ludus (game-playing love), storge (friendship-based love), pragma (practical, logical love), mania (obsessive, possessive love), and agape (selfless, altruistic love). Hendrick and Hendrick (1986) developed the Love Attitudes Scale to measure these styles and found systematic sex differences: men scored higher on ludus, while women scored higher on pragma and storge.

Fehr (1988, 2006) took a prototype approach to love, asking laypeople to list features of love and then rate their centrality. She found that features most central to the concept of love were trust, caring, honesty, friendship, and respect — notably, these are features of companionate rather than passionate love. This suggests that when ordinary people think of “love,” they think first of a deep, caring bond rather than of sexual passion and romance.

The Neuroscience of Love: Fisher’s Three Systems

Helen Fisher (2004, 2006) proposed that romantic love involves three distinct but interacting brain systems, each associated with different neurochemical substrates and evolutionary functions. The lust system, driven primarily by androgens and estrogens, motivates the search for sexual partners and is relatively indiscriminate. The attraction/romantic love system, associated with elevated dopamine and norepinephrine and suppressed serotonin, produces the focused attention, exhilaration, craving, and obsessive thinking characteristic of early romantic love. The attachment system, associated with oxytocin and vasopressin, supports long-term pair bonding and the deep sense of calm and security that characterizes enduring partnerships.

Fisher, Aron, and Brown (2005) conducted fMRI studies of individuals who were intensely in love, finding activation in the ventral tegmental area (VTA) and caudate nucleus — regions rich in dopamine neurons that are part of the brain’s reward system. This pattern of activation is similar to that seen in addiction, supporting the folk intuition that being in love is a kind of natural “high.” Importantly, Acevedo, Aron, Fisher, and Brown (2012) scanned individuals who reported being intensely in love with their long-term partner (average relationship duration of 21 years) and found VTA activation similar to that in early-stage romantic love, but without the anxiety and obsession-related activation. This suggests that long-term passionate love is possible and neurobiologically real, challenging the widespread assumption that passion inevitably fades into mere companionship.

Compassionate Love

Sprecher and Fehr (2005) developed the Compassionate Love Scale to measure an other-focused form of love characterized by caring, concern, tenderness, and an orientation toward supporting the partner’s well-being, even at a cost to the self. Compassionate love is conceptually related to Clark and Mills’ communal orientation and to the agape love style in Lee’s typology, but it is specifically operationalized as the felt experience of caring for another.

Their research showed that compassionate love for a romantic partner was related to, but distinct from, passionate love and relationship satisfaction. Compassionate love was a stronger predictor of relationship satisfaction than passionate love in longer-term relationships, and it showed greater stability over time. Fehr, Harasymchuk, and Sprecher (2014) conducted a meta-analysis and found that compassionate love was strongly associated with satisfaction (r = .58), commitment (r = .53), and trust (r = .52) across diverse samples.

Cross-Cultural Conceptions of Love

While the experience of romantic love appears to be universal — Jankowiak and Fischer (1992) found evidence of romantic love in 147 of 166 cultures examined — the cultural meaning and significance of love varies considerably. In many East Asian cultures, love is conceptualized as more restrained, considerate, and endurance-oriented compared to Western conceptions. The Chinese concept of yuan (predestined affinity) and the Japanese concept of amae (indulgent dependency) represent culturally specific relational constructs with no direct Western equivalents.

Neto et al. (2000) compared love styles across numerous countries and found that while eros was valued across all cultures studied, its relative importance varied. Cultures with stronger collectivistic values showed higher endorsement of storge and pragma, consistent with the greater emphasis on practical compatibility and family approval in partner selection. Hatfield and Rapson (2005) argued that while the capacity for passionate love is a human universal, cultural display rules and relationship norms powerfully shape how love is expressed, experienced, and evaluated.

Chapter 9: Theories of Emotion in Relationships

Emotion as a Relational Phenomenon

Emotions are not merely private internal states — they are fundamentally interpersonal phenomena that are shaped by, expressed within, and regulated through close relationships. Parkinson (1996) argued that most emotions are inherently relational: their causes, objects, and consequences all involve other people. In close relationships, emotions serve multiple functions: they communicate needs and desires, signal relationship evaluations, motivate approach and avoidance behavior, and regulate the partner’s behavior.

Broaden-and-Build Theory in Relationships

Barbara Fredrickson’s (2001) broaden-and-build theory of positive emotions has important implications for close relationships. The theory proposes that positive emotions — joy, gratitude, interest, love, serenity — broaden individuals’ momentary thought-action repertoires (expanding the range of thoughts and actions that come to mind) and build enduring personal resources (social connections, coping strategies, knowledge, physical health).

Applied to relationships, the broaden-and-build model suggests that positive emotional experiences in relationships create upward spirals of increasing connection and resilience. Waugh and Fredrickson (2006) found that positive emotions at the beginning of college predicted the development of closer friendships, which in turn predicted more positive emotions. Fredrickson (2013) specifically identified love as a “positivity resonance” — a momentary experience of shared positive emotion, mutual care, and biological synchrony between two people. This micro-momentary conceptualization of love complements trait-based approaches by emphasizing that love is something people do in specific interactions rather than something they permanently have.

Capitalization: Sharing Good News

Gable, Reis, Impett, and Asher (2004) introduced the concept of capitalization — the process of sharing positive events with others and deriving additional benefit from the sharing. They found that the way romantic partners respond to the disclosure of positive events is a significant predictor of relationship quality. Responses vary along two dimensions: active-passive and constructive-destructive.

Active-constructive responding (enthusiastic, engaged support — “That’s wonderful! Tell me more about it!”) was the only response type associated with increased intimacy, satisfaction, and trust. Passive-constructive (“That’s nice, dear”), active-destructive (“That will mean more work for you”), and passive-destructive (changing the subject) responses were all associated with poorer relationship outcomes. Remarkably, Gable et al. found that how partners respond to positive events was a stronger predictor of breakup than how they respond to negative events, suggesting that the ability to celebrate together may be even more important than the ability to weather storms together.

Jealousy: Evolutionary and Social-Cognitive Perspectives

Jealousy is a complex emotional response to perceived threats to a valued relationship. The evolutionary perspective, championed by Buss, Larsen, Westen, and Semmelroth (1992), predicts sex differences in jealousy: men should be more distressed by sexual infidelity (because it threatens paternity certainty), while women should be more distressed by emotional infidelity (because it threatens the partner’s resource investment). Buss et al.’s forced-choice studies found the predicted sex difference, and the finding has been replicated across many cultures.

However, the social-cognitive perspective challenges both the methodology and interpretation of these findings. DeSteno and Salovey (1996) proposed the double-shot hypothesis: people are most jealous about whichever type of infidelity they believe implies the other type as well. Because women tend to believe that men who are sexually unfaithful are not necessarily emotionally involved, while men tend to believe that women who are emotionally unfaithful are also likely to be sexually involved, the apparent sex difference in jealousy may reflect different beliefs about the co-occurrence of sexual and emotional infidelity rather than different evolved modules. Harris (2003) provided physiological evidence against the evolutionary account, finding that men did not show the distinctive physiological arousal to sexual jealousy imagery that would be expected if a specialized jealousy module existed.

Emotional Intelligence in Relationships

Emotional intelligence (EI) — the ability to perceive, understand, manage, and use emotions effectively — has been linked to relationship quality. Brackett, Warner, and Bosco (2005) found that couples in which both partners scored high on the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) reported the highest relationship satisfaction, while couples with large discrepancies in EI reported the lowest satisfaction. Malouff, Schutte, and Thorsteinsson (2014) conducted a meta-analysis of 38 studies and found that EI was positively associated with relationship satisfaction (r = .27), with the association being somewhat stronger for ability-based EI measures than for self-report EI measures.

Chapter 10: Trust

The Nature and Importance of Trust

Trust is widely regarded as one of the most essential ingredients in successful intimate relationships. It reflects the confident expectation that one’s partner will be responsive and caring, particularly in situations where one is vulnerable. Without trust, partners cannot be open about their feelings, take emotional risks, or depend on each other in meaningful ways. Trust is not merely a background condition of healthy relationships — it is the active process through which partners navigate the inherent uncertainty of interdependence.

Rempel, Holmes, and Zanna’s Three-Component Model

Rempel, Holmes, and Zanna (1985) — research from Dr. Rempel’s own program — developed one of the most influential models of trust in close relationships. Based on interviews with married couples and factor-analytic studies, they identified three distinct components of trust that develop sequentially as relationships deepen.

Predictability is the most basic component and refers to the expectation that the partner’s behavior will be consistent and reliable. Predictability is based on past experience — when a partner behaves consistently over time, the other comes to expect regularity. This component is grounded in specific behavioral evidence and does not require extensive inference about the partner’s underlying character. Predictability provides a foundation for trust but is insufficient on its own because it does not tell us why the partner is behaving reliably.

Dependability is a deeper component that involves confidence in the partner’s dispositional qualities — specifically, the belief that the partner is honest, benevolent, and genuinely caring. Dependability requires the perceiver to move beyond surface behavior and make inferences about the partner’s character and intentions. It addresses the question: “Can I count on this person when it really matters?” The shift from predictability to dependability represents a qualitative leap because it requires taking a risk — acting as though the partner can be relied upon in situations where the evidence is ambiguous.

Faith is the deepest and most abstract component of trust. It involves a generalized confidence that the partner will be responsive and caring in novel situations that have not yet been encountered and therefore cannot be predicted from past behavior. Faith represents an emotional security — a felt sense that the relationship is fundamentally safe — that transcends any specific evidence. Rempel et al. found that faith was the strongest predictor of love and relationship satisfaction, suggesting that this deepest level of trust is what ultimately binds partners together.

The three components are hierarchically organized: predictability provides the behavioral evidence from which dependability is inferred, and dependability provides the dispositional foundation from which faith emerges. This developmental sequence means that trust cannot be rushed — it must be built through accumulating experiences that allow each level to develop.

Holmes and Rempel’s Trust Development Model

Holmes and Rempel (1989) elaborated a process model of how trust develops (and sometimes fails to develop) over the course of a relationship. They proposed that trust development involves navigating a series of diagnostic situations — situations in which the partner’s self-interest and the perceiver’s needs are in conflict. These situations are diagnostic because they reveal whether the partner will prioritize the relationship over personal self-interest.

In the early stages of a relationship, partners tend to engage in partner-enhancing attributions, interpreting the partner’s behavior in the most favorable light possible. However, as relationships deepen and the stakes increase, partners inevitably encounter situations that create ambiguity about the partner’s motives. The critical question becomes: when my partner does something that could be interpreted negatively, do I give them the benefit of the doubt?

Holmes and Rempel distinguished between trust-testing and trust-building processes. In trust-testing, the perceiver creates or attends to situations that might reveal the partner’s true motives. For example, a person might reveal a vulnerability and observe whether the partner responds with care or indifference. In trust-building, positive experiences accumulate and are integrated into a generalized sense of confidence. Importantly, negative experiences have a disproportionate impact on trust — a single betrayal can undo years of positive experience, a phenomenon consistent with the broader negativity bias in social perception.

The Diagnosticity Problem

A core challenge in trust development is the diagnosticity problem: benevolent behavior by a partner in low-risk situations is ambiguous — it might reflect genuine caring, or it might simply reflect self-interest or social convention. A partner who is kind when it costs nothing provides no diagnostic information about what they would do in situations involving sacrifice or conflicting interests.

Holmes (1991) argued that truly diagnostic situations are those in which the partner must incur a cost or sacrifice an alternative in order to be responsive. This is why critical moments — when a partner stands by you during a crisis, forgoes an attractive opportunity for the relationship’s sake, or responds supportively when they are tired or stressed — carry so much more weight for trust than everyday kindnesses. The diagnostic value of a behavior is proportional to the strength of the competing motive the partner must overcome.

Risk Regulation and Dependence Regulation

Simpson (2007) proposed the risk regulation model, which describes how perceptions of a partner’s regard regulate willingness to become dependent on the partner. When individuals feel confident that their partner values and accepts them, they are willing to take the emotional risks inherent in deepening the relationship — disclosing vulnerabilities, asking for support, expressing needs. When individuals doubt their partner’s regard, they self-protectively withdraw, reducing vulnerability but also reducing the possibility of deeper intimacy.

This self-protective withdrawal creates a tragic irony: the very people who most need reassurance (those low in self-esteem or high in attachment anxiety) are least likely to seek it, because their doubts about the partner’s regard make the risk of rejection feel unbearable. Simpson demonstrated that this self-protective pattern operates automatically and outside of awareness, making it particularly resistant to change.

Murray, Holmes, and Collins (2006) extended this logic with their dependence regulation model, proposing that individuals manage the risks of dependence through a system of “if-then” contingencies: “If I feel valued by my partner, then I can safely risk connecting.” This regulatory system operates dynamically — fluctuations in felt regard produce corresponding fluctuations in willingness to connect, on a day-to-day and even moment-to-moment basis. Murray, Bellavia, Rose, and Griffin (2003) demonstrated that on days when low-self-esteem individuals felt more rejected by their partner (even if these perceptions were inaccurate), they reported more negative views of their partner and less closeness.

Trust Repair After Betrayal

Betrayals — infidelity, deception, broken promises — represent catastrophic failures of trust that fundamentally alter the relationship landscape. Finkel, Rusbult, Kumashiro, and Hannon (2002) examined trust repair and found that the offender’s expression of genuine remorse and the victim’s ability to empathize with the offender were both important predictors of forgiveness and trust restoration. They also found that the victim’s commitment to the relationship moderated the forgiveness process: more committed individuals were more motivated to forgive and rebuild.

Rebuilding trust after betrayal is a long and difficult process. Rempel and colleagues’ work suggests that the deepest component of trust — faith — is the most severely damaged by betrayal because it involves a generalized confidence that is shattered by evidence of the partner’s willingness to harm. Restoring faith requires not only consistent positive behavior over time (rebuilding predictability and dependability) but also a fundamental reappraisal of the partner’s character and intentions. Guerrero and Bachman (2010) found that communication characterized by explicit and sincere apologies, explanations for the betrayal, and demonstrations of changed behavior were associated with greater trust repair.

Chapter 11: Conflict

The Inevitability and Functions of Conflict

Conflict is a natural and inevitable feature of all close relationships. Because intimate partners are highly interdependent, they will inevitably encounter situations in which their preferences, needs, or goals diverge. The critical question is not whether couples experience conflict but how they manage it. Research consistently shows that the manner in which couples handle disagreements is far more predictive of relationship outcomes than the frequency or content of their disagreements.

Gottman (1994) identified several characteristic conflict patterns. Validators discuss disagreements calmly and respectfully, showing mutual support even during conflict. Volatiles engage in intense, passionate arguments but balance negativity with equally intense expressions of affection and humor. Avoiders minimize conflict, agreeing to disagree and sidestepping contentious topics. All three styles can be found in stable, happy marriages, provided the 5:1 positive-to-negative ratio is maintained. It is when couples show hostile patterns — especially the Four Horsemen (criticism, contempt, defensiveness, and stonewalling) — that relationships deteriorate.

Exit-Voice-Loyalty-Neglect Model

Rusbult, Zembrodt, and Gunn (1982) adapted Hirschman’s (1970) economic framework to describe four categories of responses to relationship dissatisfaction, organized along two dimensions: active-passive and constructive-destructive.

Exit (active, destructive) involves behaviors such as threatening to leave, actually separating, or engaging in retaliatory behaviors. Voice (active, constructive) involves discussing problems, seeking compromise, suggesting solutions, or seeking counseling. Loyalty (passive, constructive) involves waiting and hoping that things will improve, giving the partner the benefit of the doubt, and supporting the partner despite difficulties. Neglect (passive, destructive) involves ignoring the partner, spending less time together, avoiding discussion of problems, or engaging in extrarelational activities.

Research using this framework consistently finds that commitment predicts constructive responses (voice and loyalty) and inhibits destructive responses (exit and neglect). Rusbult, Johnson, and Morrow (1986) demonstrated that the tendency to respond constructively to partner provocations — termed accommodation — is one of the key relationship maintenance mechanisms through which commitment translates into relationship-promoting behavior.

Forgiveness

Forgiveness — the process by which a victim sets aside resentment and the desire for retribution toward a transgressor — has emerged as a major topic in relationship research. McCullough, Worthington, and Rachal (1997) defined forgiveness as a motivational change in which the victim experiences decreased motivation for avoidance and revenge and increased motivation for benevolence toward the offender.

McCullough, Fincham, and Tsang (2003) conducted a longitudinal study of forgiveness in romantic relationships and found that forgiveness is not a single event but a process that unfolds over time. Immediately after a transgression, victims experience a surge of unforgiving motivations (avoidance, revenge), which then gradually decline. The speed and extent of this decline are predicted by relationship quality, commitment, the severity of the transgression, the offender’s apology and amends, and the victim’s trait-level tendency toward forgiveness.

Finkel, Rusbult, Kumashiro, and Hannon (2002) found that commitment promotes forgiveness through two mechanisms: committed individuals are more motivated to maintain the relationship (providing an incentive to forgive) and are better able to take the offender’s perspective (facilitating empathy). Worthington (2005) distinguished between decisional forgiveness (a behavioral intention to treat the offender as before) and emotional forgiveness (the replacement of negative emotions with positive ones), arguing that decisional forgiveness is more under voluntary control while emotional forgiveness is a gradual process.

Attachment and Conflict

Attachment theory provides a powerful lens for understanding conflict behavior. Simpson, Rholes, and Phillips (1996) brought dating couples into the laboratory and exposed them to an anxiety-provoking situation (discussion of a major unresolved problem). They found that more anxiously attached women displayed more stress and sought more support during the discussion, while more avoidantly attached men provided less support and displayed more withdrawal. The combination of an anxious woman and an avoidant man produced the most distress and the least effective conflict resolution — the classic pursue-withdraw dynamic described by EFT.

Physiological research has added another dimension to our understanding of conflict. Levenson and Gottman (1983) found that couples whose physiological responses (heart rate, skin conductance, gross motor movement) became highly linked during conflict — a pattern called physiological linkage — were more likely to show declines in satisfaction over time. This linkage appears to reflect a mutual escalation of arousal that overwhelms both partners’ capacity for constructive communication.

Chapter 12: Power

Defining Power in Relationships

Power in intimate relationships refers to the ability of one partner to influence the other’s behavior, thoughts, or emotions. Power is inherently relational — it exists only in the context of interdependence between two people. From an interdependence theory perspective, power derives from the structure of dependence: the less dependent partner holds more power because they are less affected by the other’s actions and have better alternatives.

Principles of Power: Least Interest and Resource Theory

Waller and Hill’s (1951) principle of least interest states that the partner who is less interested in maintaining the relationship holds more power. This principle follows directly from interdependence theory: the less interested partner has lower dependence, faces less to lose from the relationship’s dissolution, and therefore has greater leverage in negotiations over relationship conduct.

Blood and Wolfe’s (1960) resource theory proposed that the partner who contributes more valued resources (income, education, occupational prestige) to the relationship holds more power. While resource theory received initial empirical support, subsequent research revealed that the translation of external resources into relationship power is moderated by cultural norms. In cultures with traditional gender ideologies, men’s resources translate into power more readily than women’s resources, while in egalitarian cultures, the association between resources and power is more symmetric.

Approach-Inhibition Theory of Power

Keltner, Gruenfeld, and Anderson (2003) proposed the approach-inhibition theory of power, which holds that having power activates the behavioral approach system (BAS), leading to increased goal-directed behavior, risk-taking, and attention to rewards, while lacking power activates the behavioral inhibition system (BIS), leading to increased vigilance, inhibited behavior, and attention to threats.

Applied to intimate relationships, this theory suggests that the more powerful partner may become less attentive to the less powerful partner’s needs, less empathic, and more likely to act on their own impulses — effects that can undermine relationship quality. Fiske (1993) proposed a related model arguing that powerless individuals attend more carefully to powerful individuals (because their outcomes depend on the powerful person’s actions), while powerful individuals attend less to powerless ones. In relationships, this asymmetry means that the less powerful partner is often more attuned to the relationship’s dynamics than the more powerful one.

Intimate Partner Violence and Power-Control Dynamics

At the extreme end of power dynamics lies intimate partner violence (IPV). Johnson (2008) distinguished among four types of partner violence based on the dynamics of power and control. Intimate terrorism involves one partner using violence as part of a general pattern of power and control over the other — this is the pattern most commonly seen in clinical samples and shelters, predominantly perpetrated by men. Situational couple violence arises from specific conflicts that escalate to physical aggression — this is the most common form of violence in general population samples and is more gender-symmetric. Violent resistance is violence used by the less powerful partner (typically the victim of intimate terrorism) in self-defense. Mutual violent control involves both partners using violence to attempt to control each other.

This typology has important implications for intervention. Couples therapy may be appropriate for situational couple violence but is contraindicated for intimate terrorism, where the priority must be the victim’s safety. The dynamics of power and control that characterize intimate terrorism — isolation, economic control, emotional abuse, intimidation, and coercion — extend far beyond physical violence and represent a pattern of systematic domination.

Decision-Making and Power Processes

Power in relationships is also expressed through everyday decision-making processes. Couples must make thousands of decisions, from trivial (what to have for dinner) to consequential (where to live, whether to have children). Zvonkovic, Schmiege, and Hall (1994) found that decision-making power was associated with satisfaction, but the pattern depended on the domain: both partners were more satisfied when each had influence in their areas of expertise and when major decisions were made collaboratively.

Chapter 13: Dissolution and Maintenance

The Process of Relationship Dissolution

Relationship dissolution is not a single event but a process that unfolds over time, often following identifiable stages. Duck (1982) proposed a four-phase model of dissolution. In the intrapsychic phase, one partner privately evaluates the costs and inadequacies of the relationship. In the dyadic phase, the dissatisfied partner confronts the other, and the couple negotiates whether and how to repair the relationship. In the social phase, the couple involves their social network, announcing the breakup and seeking support and validation. In the grave-dressing phase, each partner creates a personal narrative of the relationship’s history and demise — constructing a story that protects self-esteem and provides meaning.

Rollie and Duck (2006) updated this model, adding a resurrection phase in which individuals prepare psychologically for future relationships, drawing lessons from the dissolved one. They also emphasized that the process is not strictly linear — couples may cycle back through earlier phases, attempting reconciliation before ultimately dissolving.

Predictors of Dissolution

Decades of research have identified reliable predictors of relationship dissolution. At the individual level, neuroticism is one of the strongest personality predictors — individuals high in neuroticism experience more negative affect, perceive more conflict, and are less satisfied with their relationships (Karney & Bradbury, 1997). Attachment insecurity, poor communication skills, and low commitment also predict dissolution.

At the dyadic level, Gottman’s research identified specific behavioral predictors. The Four Horsemen (criticism, contempt, defensiveness, and stonewalling) predicted divorce with high accuracy. Gottman and Levenson (2000) further distinguished between two types of divorce: early divorces (within the first seven years), which were predicted by high levels of negative affect during conflict, and later divorces (after 14+ years), which were predicted by the absence of positive affect during conflict — suggesting that relationships die not only from the poison of negativity but also from the slow starvation of positivity.

On-Again/Off-Again Relationships

Dailey, Pfiester, Jin, Beck, and Clark (2009) brought attention to on-again/off-again (on-off) relationships, which are far more common than previously recognized. Approximately 60% of adults report having experienced a relationship in which they broke up and got back together at least once. On-off relationships are characterized by greater uncertainty, lower satisfaction, and more conflict than stable relationships, but they persist because of residual attachment, continued contact, investment, and optimistic beliefs that problems can be resolved.

Dailey, Rossetto, Pfiester, and Surra (2009) found that the reasons for renewal often differed from the reasons for dissolution: couples frequently broke up over conflict and incompatibility but renewed because of lingering feelings and perceived partner change. However, the problems that caused the original breakup often persisted, leading to a cyclical pattern of dissolution and renewal.

Post-Dissolution Experiences

Breakups are among the most painful experiences in human life, rivaling bereavement in their emotional intensity. Sbarra and Hazan (2008) applied attachment theory to the breakup experience, proposing that dissolution triggers separation distress similar to the grief response observed in infant-caregiver separations. The intensity of post-breakup distress is predicted by the degree of emotional attachment (not just satisfaction), with those who were more attached experiencing more intrusive thoughts, sadness, and yearning.

However, dissolution can also be a catalyst for post-dissolution growth. Tashiro and Frazier (2003) found that college students reported an average of five positive changes following a breakup, including personal growth, improved confidence, and clarification of desires in future relationships. Growth was more likely when individuals engaged in deliberate reflection and meaning-making about the breakup.

The rise of social media has introduced new challenges for post-breakup adjustment. Marshall (2012) found that maintaining Facebook contact with an ex-partner was associated with greater longing, more negative feelings, and slower recovery. Cyberstalking — the use of digital tools to monitor an ex-partner — is surprisingly common, with Lyndon, Bonds-Raacke, and Cratty (2011) finding that a majority of young adults reported engaging in some form of online monitoring of ex-partners.

Relationship Maintenance and Repair

While dissolution has received extensive attention, researchers have increasingly focused on the processes through which couples maintain satisfying relationships over time. Stafford and Canary (1991) identified five key maintenance strategies: positivity (being cheerful and optimistic), openness (direct discussion of the relationship), assurances (statements of love and commitment), social networks (involving friends and family in the relationship), and sharing tasks (equitable division of responsibilities).

Rusbult and colleagues (Rusbult, Olsen, Davis, & Hannon, 2001) described several relationship maintenance mechanisms that operate to sustain commitment and satisfaction. These include accommodation (responding constructively to partner provocations), willingness to sacrifice (forgoing self-interest for the partner or relationship), forgiveness (letting go of grievances), cognitive interdependence (thinking in terms of “we” rather than “I”), positive illusions (maintaining idealized views of the partner), and derogation of alternatives (perceiving potential alternative partners as less attractive).

Gottman’s research program has also contributed to understanding maintenance. His emphasis on building and maintaining Love Maps (continuing to update knowledge of the partner’s inner world), maintaining the friendship system, and creating shared meaning through rituals, roles, and goals provides a practical framework for long-term relationship maintenance. Gottman and Silver (1999) estimated that it takes approximately five hours per week of focused relational attention to maintain a healthy marriage — a relatively modest investment with potentially enormous returns.

Mate Selection

The process of choosing a long-term partner involves a complex interplay of individual preferences, social constraints, and dyadic dynamics. Filtering models propose that people narrow the field of potential partners through successive stages. Kerckhoff and Davis (1962) proposed that social homogamy (similarity in demographic background) acts as the first filter, followed by value consensus, and finally need complementarity. While the specific filtering stages they proposed have not been strongly supported, the general idea that mate selection involves progressive narrowing based on increasingly intimate criteria remains useful.

Arranged marriages, which remain common in many cultures, represent an alternative pathway to partner selection. Research comparing arranged and love-based marriages has yielded nuanced findings. Gupta and Singh (1982) found that love increased over time in arranged marriages but decreased in love marriages, though this study had methodological limitations. More recent research suggests that satisfaction in arranged marriages depends heavily on factors such as the degree of choice given to the couple, compatibility of values and personality, and the quality of communication that develops after marriage (Madathil & Benshoff, 2008).

The concept of assortative mating — the tendency for partners to be similar — is one of the most robust findings in mate selection research. Partners are similar in physical attractiveness, age, education, religion, political attitudes, intelligence, and personality. Watson et al. (2004) examined 291 newlywed couples and found strong assortment for attitudes and values (r = .50-.60), moderate assortment for religiosity and political orientation (r = .35-.45), and weak but significant assortment for personality traits (r = .10-.20).

Chapter 14: Love and Hate

The Relationship Between Love and Hate

The close relationship between love and hate — the observation that the most intense hatred is often directed at those we once loved most deeply — has long fascinated philosophers, poets, and psychologists. From an interdependence perspective, the capacity for hate is proportional to the degree of interdependence: only a partner who has deep access to our vulnerabilities, whose behavior profoundly affects our outcomes, and in whom we have invested heavily can inspire the intensity of emotion that we call hate.

Fitness and Fletcher (1993) studied emotional experiences in marriage and found that anger and hate in close relationships are qualitatively different from anger and hate directed at strangers or acquaintances. Relationship-based hatred involves a deep sense of betrayal, a perception of injustice, and a desire to harm that is rooted in the violation of expectations born from intimacy and trust. The intensity of these feelings is directly proportional to the intimacy that preceded them.

Unrequited Love

Baumeister, Wotman, and Stillwell (1993) studied unrequited love — the experience of loving someone who does not reciprocate — and found that both the would-be lover and the rejecting target suffer, though in different ways. The would-be lover experiences longing, frustration, and lowered self-esteem. The rejector experiences guilt, anxiety about hurting the other person, and frustration at being unable to end the other’s suffering. Both parties’ accounts reveal systematic self-serving biases: would-be lovers emphasize their persistence and the injustice of rejection, while rejectors emphasize their kindness and the other person’s failure to accept reality.

Commitment and the Transformation of Love Over Time

The trajectory of love over the course of a long-term relationship is not a simple story of decline. While passionate love typically diminishes from its initial peak, this decline is neither universal nor inevitable. Acevedo and Aron (2009) conducted a meta-analysis and found that romantic love (with the intensity of passion but without the obsession) did not decline with relationship duration, while obsessive love did decline. This suggests that the obsessive, anxious component of early love fades but the core experience of deep romantic connection can endure.

Tucker and Aron (1993) found that self-expansion through novel shared activities can rekindle the excitement of early relationship stages, and Coulter and Malouff (2013) conducted a meta-analysis of 24 randomized controlled trials showing that relationship education and enrichment programs produce meaningful improvements in relationship quality (d = 0.36). These findings underscore that relationship quality is not a fixed outcome of partner compatibility but is substantially shaped by the ongoing choices and investments that partners make.

The study of interpersonal relations reveals that our closest bonds are simultaneously our greatest source of joy and our greatest vulnerability. Understanding the psychological processes that govern these bonds — attachment, trust, communication, commitment, and the complex interplay of emotion and cognition — provides not only scientific knowledge but also practical wisdom for building and sustaining the relationships upon which human flourishing depends.