Conflict escalation is the process through which a disagreement becomes increasingly intense, personally threatening, antagonistic, or harmful. An argument may begin with incompatible preferences or competing interpretations and develop into accusations, retaliation, exclusion, intimidation, or aggression. The defining feature is a change in the interaction: participants increasingly treat one another as adversaries, interpret subsequent behavior through the conflict, and use responses that create further reasons for hostility. Escalation can occur within minutes or unfold across months. It may stop, reverse, plateau, or jump suddenly to dangerous conduct; no universal sequence governs every dispute.
Understanding escalation requires following three interacting levels: the changing meaning attributed to the disagreement, the actions that each participant takes in response to the other, and the surrounding relationship or institution that determines what each person can safely do. The research literature connects these levels to threat appraisal, hostile attribution, anger and rumination, negative reciprocity, relationship dynamics, and power. A recent validation study defines escalation through increasing conflict intensity and qualitative changes in tactics and relationships, while also illustrating the limits of stage-based measurement (Scheppa-Lahyani and Zapf, 2023).
What Is Conflict Escalation in Psychology?
A conflict exists when people perceive incompatibility between their goals, interests, beliefs, roles, rights, or expectations. Escalation describes what happens to the conflict process over time. Its markers include intensification of language, growing emotional hostility, expansion of the issues being contested, increasingly rigid perceptions of the other party, recruitment of allies, and tactics designed to punish or constrain. Frequency and severity are different dimensions. A recurring mild disagreement may remain stable; one threatening act can sharply raise the stakes of an otherwise short encounter.
Disagreement, assertive objection, anger, hostility, aggression, and violence are separate constructs. Disagreement concerns incompatible positions. Anger is an emotion that can accompany perceived obstruction or injustice. Hostility involves antagonistic attitudes or conduct. Aggression, as commonly defined in behavioral research, involves an intention to harm someone who seeks to avoid that harm; violence refers to particularly serious forms of harmful aggression. People can experience anger without acting aggressively, challenge an unfair decision without being hostile, or use apparently calm language to intimidate someone. An integrative review of aggression theory clarifies the relevance of distinguishing motives, interpretations, affect, and behavior (Allen and Anderson, 2017).
The distinction matters because intensity alone does not determine the psychological meaning of an encounter. A raised voice can signal urgency, fear, excitement, or a communication convention. A quiet ultimatum can carry a credible threat. Assessment must consider what was said or done, its surrounding context, whether participants retained freedom to disagree or leave, and how the exchange affected their options. Escalation is an interactional description, not a medical diagnosis and not a personality classification.
The Transition From Disagreement to Hostility
The issue acquires relational meaning
At first, two colleagues may disagree about who should complete a task. Once one interprets the other's objection as a sign of disrespect, irresponsibility, or deliberate sabotage, the conversation changes. The factual question remains, but a second question emerges: What does this treatment say about my standing, trustworthiness, or safety? Subsequent statements are evaluated against that new concern. A practical request may now sound like an accusation; a delay may look like deliberate resistance.
A conflict can therefore broaden from a single decision to the history of the relationship. The participants begin using previous grievances to explain current behavior, and unresolved experiences become evidence for increasingly global judgments. This is one route from issue-centered contention to relationship-centered hostility. It is a plausible process model rather than a claim that every disagreement inevitably unfolds in this order. The Conflict Escalation Questionnaire research found that relationship conflict was more prevalent in its more highly escalated workplace cases, but its cross-sectional design does not establish the direction of causality (Scheppa-Lahyani and Zapf, 2023).
Motives become harder to distinguish from interpretations
During a tense exchange, participants know their own frustrations, previous experiences, and intentions better than those of the other person. They see the other's words and actions, not the full reasoning behind them. An ambiguous response can acquire an intention that has never been established. The person who feels accused then responds to the accusation they believe occurred; the other may hear that defensive response as confirmation of bad faith.
For example, a supervisor asks, “Why was this not finished?” An employee experiences the question as public humiliation and replies sharply. The supervisor interprets the reply as insubordination and increases pressure. Neither step requires a preexisting plan to injure the other. Their interpretations and responses can become mutually reinforcing. At the same time, a supervisor may genuinely intend to humiliate or threaten; the analysis must remain open to evidence of deliberate wrongdoing and differences in authority.
Threat Appraisal and Hostile Attribution
How perceived threats organize attention
People evaluate what an event may mean for their goals, belonging, control, dignity, or physical safety. An appraisal of threat can narrow attention toward possible danger and increase the urgency of responding. In conflict, a perceived challenge to competence, fairness, status, or identity may transform a negotiable question into a personally significant test. Threat appraisal can be accurate, exaggerated, or uncertain. Its psychological effects cannot be used to decide, by themselves, whether the threat is real.
Escalation may be accelerated by feedback between appraisal and behavior. A person who expects hostility attends closely to sarcasm, interruption, or delayed responses. They confront the other person. That confrontation may elicit defensiveness, which supplies fresh evidence for the initial expectation. A stable pattern of reciprocal suspicion can develop even when the participants disagree about what the first hostile act was.
Hostile attribution bias and ambiguity
Hostile attribution bias describes a tendency to interpret ambiguous social actions as intentionally harmful. It is particularly relevant when a comment, facial expression, or delayed reply has several plausible explanations. A systematic review of 25 adult studies found generally small-to-medium associations between hostile attribution and aggression across different samples; the authors emphasized that much of the evidence was cross-sectional and measurement quality varied (Klein Tuente, Bogaerts, and Veling, 2019). These results support the construct's relevance while limiting claims about any one person's future actions.
A 2026-relevant recent literature stream also distinguishes habitual anger from momentary anger. A systematic review and meta-analysis of 35 studies including 10,451 participants reported a moderate association between trait anger and hostile attribution bias, with substantial heterogeneity (Fenerdjian and Gagnon, 2026). This association neither establishes that angry people are necessarily inaccurate nor proves that changing attribution alone prevents aggressive behavior.
An observer cannot infer hostile attribution bias simply because a person perceives a threat. Discrimination, coercion, bullying, and previous mistreatment can provide real reasons for concern. Researchers must distinguish interpretations of objectively ambiguous events from responses to documented threats. Requiring individuals to reinterpret genuine mistreatment as harmless would distort the construct and shift attention away from accountable behavior.
Negative Reciprocity: How One Harmful Response Produces Another
Negative reciprocity occurs when a person responds to an aversive action with a further aversive action. In escalation, the response may be experienced as justified repayment or as a means of discouraging future mistreatment. The other person then experiences the new response as the latest injury and retaliates again. The interaction can become self-reinforcing: each participant treats their own action as a reaction and the other's action as an initiation.
Experimental work directly examines the difference between retaliatory impulse and strategic deterrence. Vandermeer and colleagues used a modified economic-exchange task that generated escalating responses to negative exchanges and tested competing explanations involving reflexive punishment and deliberate prevention of future harm (Vandermeer et al., 2019). The study demonstrates that escalation can be investigated through observable response contingencies. Its controlled exchange paradigm, however, does not reproduce the full history, emotional significance, or unequal power of everyday relationships.
A distinguishing property of this cycle is that participants can disagree about proportionality. One thinks, “I only returned the insult.” The other thinks, “They attacked me without reason.” Memory for who began the conflict and what counted as provocation may diverge. Escalation can continue because the parties repeatedly compare the other's visible behavior against their own internal sense of justification. Identifying the cycle is useful for explanation; it does not distribute blame equally when one person has greater power or has initiated violence.
Anger, Arousal, Rumination, and Self-Regulation
Anger can increase action readiness
Anger often follows an appraisal of obstruction, unfairness, or blame and can motivate approach toward a perceived source of harm. It can also focus attention on changing an unacceptable situation. Berkowitz's cognitive-neoassociation account connects aversive events, automatic associations, anger, and aggressive impulses while recognizing that regulatory processes affect whether an impulse becomes behavior (Berkowitz, 2012). The model explains why a conflict may accelerate even before participants have deliberately evaluated every word they exchange.
Physiological activation may make it harder to slow a conversation or consider alternative explanations, especially when coupled with threat appraisal. Yet arousal is not a direct measurement of hostility or danger. The same bodily response can occur during fear, effort, urgency, or emotional investment. Research conclusions should therefore avoid treating heart rate, visible agitation, or speaking volume as independent proof of aggressive intent.
Rumination can prolong the conflict beyond the conversation
Rumination involves repeated mental rehearsal of an upsetting event or grievance. A disagreement can continue internally after the participants have stopped speaking: one person replays the exchange, imagines better retaliatory responses, and interprets new information through the episode. In an experimental study, participants instructed to ruminate about a provoking person while hitting a punching bag subsequently reported more anger and showed more aggression than relevant comparison groups (Bushman, 2002). The result is evidence against a simple catharsis theory in that laboratory design, not a universal prediction about all reflective discussion, exercise, or memory.
Rumination can also affect the next encounter. Previously neutral remarks may now be connected to a larger story about bad character, contempt, or expected betrayal. A temporary pause between conversations prevents escalation only when conditions permit the conflict to be reconsidered rather than continually amplified.
What contemporary synthesis can and cannot establish
A large systematic review and meta-analysis published in 2026 synthesized 171 studies from 137 articles, representing more than 252,000 participants. It found associations between emotion-regulation difficulties, maladaptive strategies, and aggression; the included literature was heavily cross-sectional and statistically heterogeneous (Smith et al., 2026). The evidence supports emotion regulation as an important correlate and possible intervention target, but does not prove that a single technique will prevent escalation in every environment.
A separate meta-analysis of anger-management activities found that approaches reducing arousal, such as some relaxation and breathing practices, were associated with decreases in anger in the studied settings (Kjærvik and Bushman, 2024). Anger reduction and durable reductions in interpersonal violence are different endpoints. In a dangerous setting, the immediate priority is safety rather than successfully calming the person causing harm.
How Escalating Interactions Become Self-Reinforcing
Criticism and countercriticism
Criticism can be specific or global. “This report contains an error” identifies a changeable issue. “You are always incompetent” converts a task concern into a judgment about the person. A global accusation increases the likelihood that the recipient will defend identity rather than examine the task. Countercriticism then shifts attention to the accuser's faults, broadening the exchange from a solvable discrepancy to a contest over moral standing.
The meaningful distinction is between contesting a proposition and attacking the person who advances it. A strong objection to misconduct or abuse remains legitimate even when expressed forcefully. The scientific question is whether repeated personal attacks, contemptuous framing, and retaliation become part of the reciprocal pattern and displace attention from the actual issue.
Pursuit and withdrawal
Some conflicts follow a cycle in which one person intensifies demands for discussion while the other withdraws or ends the exchange. Pursuit can be experienced as pressure; withdrawal can be experienced as rejection or obstruction; either response can increase the other's distress. A meta-analysis covering 74 studies and 14,255 participants found meaningful associations between demand-withdraw patterns and relational or communication outcomes, while expressly cautioning against assuming that these mostly correlational results demonstrate a single causal direction (Schrodt, Witt, and Shimkowski, 2014).
Diary research with 116 U.S. couples observed demand-withdraw patterns in conflicts occurring at home; the patterns were associated with more negative emotion and less reported resolution, and the person initiating discussion was more likely to take the demanding role (Papp, Kouros, and Cummings, 2009). These data belong to couple interactions and do not automatically describe colleagues, friends, or strangers. Crucially, withdrawal may function as self-protection in the presence of intimidation, violence, or an unfair demand. Labeling all disengagement as maladaptive conceals that safety distinction.
Social amplification and audience effects
A two-person dispute can expand when participants recruit allies, share selective accounts, or seek public validation. Audience involvement can intensify concerns about humiliation, loyalty, and reputation; it may also bring witnesses, accountability, and protection when misconduct has occurred. The meaning of involving others depends on the function of that involvement. Seeking independent support during abuse is fundamentally different from mobilizing a group to harass someone.
Digital interaction can amplify these dynamics through persistence, forwarding, visibility, and missing contextual cues. Yet written communication can also preserve evidence, permit slower reflection, and improve access for people who need more processing time. The same medium can support resolution or aggravate conflict depending on how its affordances are used. Broad claims that online conflict is necessarily worse than face-to-face conflict require stronger comparative evidence than isolated anecdotes provide.
Escalation Models: Stages, Spirals, and Structural Change
Glasl's nine-stage framework
Friedrich Glasl's influential model organizes escalating conflict into nine stages within three broad phases, often described as an early phase in which parties still seek mutually workable outcomes, a middle phase in which adversarial victory grows more important, and a late phase characterized by willingness to inflict increasingly severe losses. Its analytical value lies in drawing attention to changing tactics, narrowed options, public positioning, threats, and the need to match intervention to risk (Glasl, 1982).
The model should be interpreted as a framework for observing qualitative shifts, not as a validated universal timetable. Conflicts can move unevenly, skip manifestations, combine behaviors from different stages, or de-escalate. Scheppa-Lahyani and Zapf developed the Conflict Escalation Questionnaire using workplace conflict samples in Germany and reported initial validity evidence, but the studies were cross-sectional, relied substantially on reported conflict experiences, and did not establish that each conflict passes through every proposed stage (Scheppa-Lahyani and Zapf, 2023).
Spiral and feedback-loop perspectives
An escalation spiral emphasizes action-reaction feedback. A stronger demand invites resistance; resistance is interpreted as obstruction; the next demand becomes stronger. This model describes how an interaction may intensify without treating one side's stated intention as a complete account of the resulting behavior. It is particularly useful when the pattern is visible across several conversational turns or repeated encounters.
A structural-change perspective adds that conflict changes the environment in which future conflict occurs. Once trust collapses, the parties limit information sharing, recruit allies, formalize penalties, or change institutional rules. Each modification affects subsequent incentives and interpretations. A conflict originally about a single decision can generate new sources of disagreement through the very tactics used to manage it. Contemporary interdisciplinary reviews place spiral, stage, and structural accounts alongside one another and emphasize that their metaphors illuminate different temporal and institutional aspects of escalation (August, 2025).
When a model becomes misleading
A stage model can encourage false certainty if an observer assigns a fixed number to a complex relationship without examining evidence. A spiral metaphor can imply moral symmetry where a powerful actor is imposing unilateral punishment. A purely emotional account can obscure disputes over wages, resources, discrimination, or legal rights. A strong explanation integrates behavioral sequences with the distribution of choices, responsibilities, and constraints.
Power Asymmetry, Coercion, and Interpersonal Safety
Ordinary disagreement assumes that participants can communicate their preferences, challenge claims, and leave or pause under reasonably safe conditions. These assumptions may fail when one person controls employment, housing, finances, access to children, immigration status, or the possibility of physical safety. Escalating threats in such contexts involve constraints on autonomy that are not captured by a simple two-sided communication cycle.
Coercive conduct may include intimidation, monitoring, isolation, credible threats, or retaliation for seeking outside help. The affected person's silence, appeasement, withdrawal, or apparently inconsistent response may reflect efforts to reduce danger. These responses should not be treated as evidence of an equal contribution to the aggression. Research about reciprocal negative communication and behavioral training cannot substitute for evaluating coercion, access to protection, and the responsible actor's conduct.
The World Health Organization's clinical guidelines for intimate partner and sexual violence emphasize supportive, survivor-centered responses, careful attention to privacy and safety, and access to practical and specialist assistance (World Health Organization, 2013). The corresponding clinical handbook identifies immediate physical and emotional needs, ongoing safety, and supportive care as core priorities (World Health Organization, 2014). When threats or violence are present, advice to debate more calmly, negotiate privately, or participate in joint mediation may expose the targeted person to additional risk. A safety plan and qualified local support may be more appropriate than a shared discussion.
Safety also matters in institutional conflict. An employee subjected to retaliation for raising concerns may reasonably avoid a direct confrontation and seek a protected reporting mechanism. A child cannot negotiate with an adult under the same conditions as two independent peers. Social status and disability access can affect who is believed, who faces penalties for emotional expression, and whose behavior is labeled difficult. An adequate account of escalation asks who holds authority and what alternatives are realistically available.
How Researchers Observe and Measure Escalation
Self-report and retrospective narratives
Questionnaires and interviews can assess perceived hostility, threatening acts, anger, loss of trust, and severity of tactics. They efficiently capture experiences that cannot always be observed in a laboratory, but depend on memory, willingness to disclose, interpretation of items, and sampling. Conflicting accounts may contain genuinely different information; agreement between witnesses is one source of evidence, not a requirement for every valid report.
The Conflict Escalation Questionnaire provides a concrete attempt to distinguish qualitative levels of conflict and to compare self-assessment with interviewer assessment. Its development and cross-validation in two samples strengthens initial measurement evidence, while the studied settings, retrospective features, sample sizes, and stage-based assumptions constrain generalizability (Scheppa-Lahyani and Zapf, 2023).
Observed conversations and sequential coding
Researchers may record interactions and code behaviors such as interruptions, hostile comments, expressions of validation, threats, withdrawal, or attempts to return to the issue. Sequential analysis tests whether one behavior is more likely to follow another than would be expected from their overall frequency. Such measures can illuminate reciprocal patterns and prevent researchers from relying solely on a participant's global judgment that a discussion was hostile.
Observation also has limitations. People may alter their behavior when recorded. Laboratory discussions often concern predetermined topics and last only minutes. Coding systems can mistake culturally variable behaviors for disrespect if their rules are narrow. Inferring motives from an interruption or facial expression alone exceeds what the observation can establish. The strongest designs combine observable sequences with context and participant perspectives.
Experience sampling and longitudinal designs
Daily diaries and ecological momentary assessment provide information about conflicts closer to their occurrence, including changes in participants' emotions, perceived provocations, and responses. Longitudinal research can investigate whether harmful patterns precede later deterioration in relationships, although even temporal precedence does not automatically eliminate third variables. In relationship science, a meta-analysis of 64 dyadic studies found small associations between communication behaviors and later relationship quality; negative communication also showed an association with later dissolution (Kanter et al., 2022). These results resist simplistic predictions based on a single conversation.
A separate analysis of multiple longitudinal couple studies found that the most robust within-couple associations between negative communication and relationship satisfaction were concurrent rather than consistently predictive of future changes (Johnson et al., 2022). The distinction matters: seeing hostility and distress together does not, on its own, identify which causes the other or whether both arise from external pressures.
Experiments, causal inference, and ethical limits
Experiments can manipulate an ambiguous message, unfair allocation, or opportunity to retaliate and test a specific mechanism under controlled conditions. They offer more direct causal leverage for the manipulated element but usually simplify relationships and restrict permissible harm. Researchers cannot ethically reproduce serious intimidation or violence merely to study how it escalates. Triangulation across experiments, observation, diary methods, and qualitative accounts is therefore essential.
Why Conflict Escalation Differs Across People and Contexts
Development and learning histories
Individuals learn from past experiences how disagreement is likely to unfold. A person who has repeatedly been punished for raising concerns may anticipate retaliation; someone whose disagreements have been resolved fairly may expect negotiation to be worthwhile. These expectations can influence how quickly a threat is perceived, how much information is sought, and what action seems available. Developmental explanations must consider context and opportunity rather than assign fixed conflict traits to an individual.
Culture, identity, and communication conventions
Directness, silence, humor, eye contact, interruption, and the use of hierarchy acquire meaning within particular relationships and communities. A blunt disagreement is not automatically aggressive; indirect communication is not automatically evasive. Cultural differences can produce misunderstanding, but they do not remove universal rights to safety and freedom from coercion. Competent assessment identifies the participants' actual conventions and the real consequences of their actions rather than applying stereotypes based on nationality or neurotype.
Neurodevelopmental and communication differences
Differences in sensory processing, pragmatic language, preferred timing, literal interpretation, or conversational pacing may change how a disagreement is experienced. A delayed response may be needed for processing; an abrupt exit may reflect overload; direct speech may express clarity. These observations cannot determine autism, social anxiety, personality disorder, or any other diagnosis. Conflict escalation is a relational process; diagnostic questions belong to a separate assessment undertaken only when clinically appropriate.
Organizational rules and resource constraints
Workload, accountability, unclear authority, shortages, and inconsistent rules can produce repeated conflict even among people who communicate carefully. A meeting may become hostile because a procedure makes fair participation impossible, not because employees lack listening skills. In these cases, institutional changes may do more to reduce escalation risk than asking individuals to modify their tone while the underlying constraints remain unchanged.
Recognizing Escalation Before the Next Exchange
Several observable developments suggest that disagreement is moving toward hostility. The conversation repeatedly returns to attacks on character; ordinary ambiguity is interpreted as deliberate malice; people seek to punish rather than clarify; past grievances expand the topic; threats or public humiliation appear; and participants increasingly believe that listening or compromise would mean surrender. These are contextual indicators, not a diagnostic checklist. Their importance depends on intensity, frequency, safety, and the freedom of all parties to disengage.
A practical analysis can ask five connected questions. What is the concrete disputed issue? What meanings are participants attaching to the other's conduct? What behavior follows each perceived provocation? Which constraints or rewards sustain that sequence? What would establish that the situation has become safe enough for ordinary discussion? The answers identify possible mechanisms without presuming shared responsibility.
Consider two siblings arranging care for a parent. One believes the other has avoided responsibilities; the other believes their financial contributions have been ignored. A request for a schedule becomes “You never help,” followed by “You always complain.” Older grievances are recruited and the care plan disappears from the conversation. Escalation is visible in issue expansion, global blame, and reciprocal rebuttal. A response that addresses the actual distribution of time and money may help more than simply requiring softer words. The example illustrates a mechanism rather than prescribing a universal solution.
Consider instead a worker whose manager threatens dismissal after a safety complaint. The worker's reluctance to meet privately reflects a realistic power asymmetry. Here the relevant intervention may be documentation, independent advice, or protected reporting. Treating the incident as a symmetric argument between two equally empowered communicators would be inaccurate and potentially unsafe.
What the Evidence Supports About Interrupting Escalation
The strongest general principle is to identify and alter the process that is sustaining harm, while assessing safety before encouraging further contact. In relatively safe disagreements, clarification of the issue, accurate acknowledgment of concerns, and an agreed pause may reduce opportunities for additional misinterpretation. Emotional regulation may help participants slow impulsive retaliation. Structural interventions may be necessary when the conflict is generated by roles, scarce resources, or unfair procedures.
Evidence about specific techniques must be matched to their actual endpoints. Controlled anger-regulation studies measure anger and often short-term aggression-related outcomes, not necessarily long-term relationship repair or prevention of violence. Communication studies may measure observed behavior or satisfaction, not changes in legal rights or institutional fairness. Meta-analytic correlations between emotional regulation and aggression identify promising relationships, while leaving causality, individual differences, and transfer to everyday conflicts incompletely resolved (Smith et al., 2026); (Kjærvik and Bushman, 2024).
A useful distinction is between explaining an escalation and deciding how to intervene. An explanatory model identifies the mechanisms by which danger or hostility increases. Intervention requires additional judgments about proportionality, rights, consent, feasibility, and access to support. Detailed techniques for cooling an exchange or negotiating a solution form their own scientific questions; the present article concentrates on why escalation develops and how its mechanisms can be recognized.
Conflict Escalation, Relationship Outcomes, and the Limits of Prediction
Hostile episodes can degrade trust, make coordination costly, and alter expectations for future contact. Their consequences vary with severity, repetition, repair, power, and available alternatives. A meta-analysis of observed couple conflict classified hostility, distress, withdrawal, problem-solving, and intimacy as different behavioral dimensions and related those categories to relationship satisfaction (Woodin, 2011). The evidence concerns patterns in couple samples, not a universal law predicting the future of any relationship from one argument.
Likewise, repeated negative communication can coexist with dissatisfaction because both reflect difficult material circumstances, earlier injuries, or competing goals. Correlational and longitudinal studies support useful probabilistic statements but have limitations for individual predictions. A person may end a relationship for legitimate reasons after a single serious threat. Another relationship may recover from heated disagreement through genuine accountability and changes in conduct. The ethics of the situation cannot be reduced to a score for anger, conflict stage, or conversational style.
Scientific Confidence and Open Questions
Established: escalation is a meaningful description of increasing conflict severity; interpretations, affect, and behavior interact; reciprocal negative responses and demand-withdraw patterns are observable in defined research contexts. Several established measurement traditions distinguish reported perceptions, observed acts, and longitudinal outcomes (Scheppa-Lahyani and Zapf, 2023); (Schrodt, Witt, and Shimkowski, 2014).
Supported but context-sensitive: hostile attribution is associated with aggression, emotion-regulation difficulties are associated with aggressive behavior, and negative communication has modest predictive relationships with later couple outcomes. Their effect sizes and causal interpretations vary by population, measurement, and setting (Klein Tuente, Bogaerts, and Veling, 2019); (Smith et al., 2026); (Kanter et al., 2022).
Preliminary or limited: the predictive validity of fine-grained stage assignments across cultures and relationship types, the degree to which experimental punishment games generalize to sustained interpersonal disputes, and the long-term impact of particular early-interruption strategies outside controlled settings. Evidence is especially limited where direct research would expose people to serious harm.
Contested or unresolved: how much of escalating behavior reflects deliberate strategy versus automatic responding in a given setting; how emotional, structural, and reputational mechanisms interact over time; and which interventions best protect people facing unequal power. Competing accounts may each explain a portion of the same episode, making carefully designed multimethod research more informative than allegiance to one model.
Frequently Asked Questions About Conflict Escalation
Why do minor disagreements sometimes become major arguments?
An apparently minor issue may acquire meanings about respect, fairness, trust, or authority. When those meanings are treated as threats, reciprocal responses can broaden the dispute. The importance of the original issue does not measure the importance of the relational concerns that emerge during the argument.
Is conflict escalation always caused by anger?
No. Fear, humiliation, distrust, perceived injustice, calculated punishment, institutional incentives, and actual threats can contribute. Some hostile actions are quiet or deliberate. Anger is one possible mechanism among several.
Does escalation always follow Glasl's nine stages?
No. The stages describe recognizable qualitative patterns and have inspired measurement research, but conflicts differ in pace, sequence, severity, and context. A threat can arise early; a prolonged disagreement may remain respectful. Stage assignments should be treated as interpretive tools rather than inevitable trajectories.
Can one person cause escalation while the other is trying to stay safe?
Yes. Intimidation or coercive control may be unilateral, and withdrawal or appeasement may be a protective response. Understanding interaction sequences does not require assigning equal responsibility. Safety and agency take precedence when threats or violence are present.
Is taking a break evidence of conflict avoidance?
A mutually understood pause can create room for reflection. Disengagement may also express a boundary or safety need. Silence used to intimidate, punish, or isolate someone has a different meaning. The function of the behavior and the surrounding power relations matter more than its superficial form.
Does hostile attribution bias mean the threat is imaginary?
No. Hostile attribution bias concerns interpretations of ambiguous events. Some behavior is genuinely threatening, and a history of mistreatment changes what a person can reasonably infer. Evidence about the actual conduct must remain central.
Can a communication technique guarantee that hostility will stop?
No single technique establishes control over another person's choices or makes a dangerous situation safe. In ordinary disputes, changes to timing, framing, emotional arousal, or procedures can sometimes alter a feedback loop. Under coercion or violence, independent protection and support may be required.
Related Articles
Cluster Communication, Conflict and Social Skills — The scientific overview of interpersonal communication, social competence, and conflict processes.
Relationship Repair After Conflict: How Couples Recover Emotional Closeness — Relationship-specific consequences and repair processes following disagreements.
Interpersonal Conflict in Psychology: Causes, Dynamics and Outcomes — Define interpersonal conflict and distinguish needs, roles, values and power.
Conflict De-Escalation: How to Reduce Tension During Disagreements — Cover interruption of escalation, timing, tone, safety and return to discussion.
Conflict Avoidance: Psychology, Motives and Consequences — Examines strategic delay, fear, safety, anxiety and recurring conflict.
Defensive Communication: Triggers, Responses and Repair — Explains perceived threat, feedback reception, blame and interaction repair.
References
Allen, J. J., and Anderson, C. A. (2017). General Aggression Model. In The International Encyclopedia of Media Effects. https://doi.org/10.1002/9781118783764.wbieme0078
August, V. (2025). Escalation and de-escalation. In The Handbook of Social and Political Conflict. https://doi.org/10.1002/9781119895534.ch30
Berkowitz, L. (2012). A different view of anger: The cognitive-neoassociation conception of the relation of anger to aggression. Aggressive Behavior, 38(4), 322–333. https://doi.org/10.1002/ab.21432
Bushman, B. J. (2002). Does venting anger feed or extinguish the flame? Catharsis, rumination, distraction, anger, and aggressive responding. Personality and Social Psychology Bulletin, 28(6), 724–731. https://doi.org/10.1177/0146167202289002
Fenerdjian, C., and Gagnon, J. (2026). The association between trait anger and hostile attribution bias: A systematic review and meta-analysis. Journal of Personality. https://doi.org/10.1111/jopy.70092
Glasl, F. (1982). The process of conflict escalation and roles of third parties. In Conflict Management and Industrial Relations, 119–140. https://doi.org/10.1007/978-94-017-1132-6_6
Johnson, M. D., et al. (2022). Within-couple associations between communication and relationship satisfaction over time. Personality and Social Psychology Bulletin, 48(4), 534–549. https://doi.org/10.1177/01461672211016920
Kanter, J. B., Lavner, J. A., Lannin, D. G., Hilgard, J., and Monk, J. K. (2022). Does couple communication predict later relationship quality and dissolution? A meta-analysis. Journal of Marriage and Family, 84(2), 533–551. https://doi.org/10.1111/jomf.12804
Kjærvik, S. L., and Bushman, B. J. (2024). A meta-analytic review of anger management activities that increase or decrease arousal: What fuels or douses rage? Clinical Psychology Review, 109, 102414. https://doi.org/10.1016/j.cpr.2024.102414
Klein Tuente, S., Bogaerts, S., and Veling, W. (2019). Hostile attribution bias and aggression in adults: A systematic review. Aggression and Violent Behavior, 46, 66–81. https://doi.org/10.1016/j.avb.2019.01.009
Papp, L. M., Kouros, C. D., and Cummings, E. M. (2009). Demand-withdraw patterns in marital conflict in the home. Personal Relationships, 16(2), 285–300. https://doi.org/10.1111/j.1475-6811.2009.01223.x
Scheppa-Lahyani, M. N., and Zapf, D. (2023). Are you threatening me? Development and validation of the Conflict Escalation Questionnaire. Frontiers in Psychology, 14, 1164990. https://doi.org/10.3389/fpsyg.2023.1164990
Schrodt, P., Witt, P. L., and Shimkowski, J. R. (2014). A meta-analytical review of the demand/withdraw pattern of interaction and its associations with individual, relational, and communicative outcomes. Communication Monographs, 81(1), 28–58. https://doi.org/10.1080/03637751.2013.813632
Smith, K., Jones, A., Daly, N., Widdrington, H., Garofalo, C., and Gillespie, S. M. (2026). Emotion regulation and aggression: A systematic review and meta-analysis. Aggressive Behavior, 52(1), e70055. https://doi.org/10.1002/ab.70055
Vandermeer, J., Hosey, C., Epley, N., and Keysar, B. (2019). Escalation of negative social exchange: Reflexive punishment or deliberative deterrence? Journal of Experimental Social Psychology, 84, 103823. https://doi.org/10.1016/j.jesp.2019.103823
Woodin, E. M. (2011). A two-dimensional approach to relationship conflict: Meta-analytic findings. Journal of Family Psychology, 25(3), 325–335. https://doi.org/10.1037/a0023791
World Health Organization. (2013). Responding to intimate partner violence and sexual violence against women: WHO clinical and policy guidelines. https://www.who.int/publications/i/item/9789241548595
World Health Organization. (2014). Health care for women subjected to intimate partner violence or sexual violence: A clinical handbook. https://www.who.int/publications/i/item/WHO-RHR-14.26