Chapter 43: When the Same Harm Keeps Repeating
Part VII: Aftermath and Accountability
A repeated apology can eventually raise a different question.
After one difficult exchange, a partner may recognize what happened, take responsibility, and want to respond differently. A later mistake does not automatically erase that effort. People sometimes need practice, and improvement can be uneven.
But another situation deserves a different kind of attention. The same harmful conduct continues, the same assurance follows, and there is little that the affected person can observe as meaningful change. Each new episode may be followed by a convincing conversation. The next episode nevertheless raises a question that the previous conversation did not settle: What has actually changed?
Not every repetition is the same. A recurring disagreement is not necessarily abuse, and one imperfect attempt at change is not proof that someone is unwilling to change. At the same time, repeated harm should not be treated as an ordinary relationship pattern simply because it is familiar.
This chapter is about recognizing that distinction without pretending there is a universal number of incidents, deadline, or test that settles what another person will do next.
The disagreement and the conduct are different questions
Earlier chapters distinguish the issue a couple is arguing about from the Loop that may occur around it. That distinction remains useful when ordinary, safe disagreements repeat. A couple might continue to disagree about spending or scheduling while gradually changing how they speak and listen.
But the difference between issue and Loop does not erase conduct. When someone repeatedly insults, threatens, deceives, intimidates, or violates an agreement, the question is not only how the conversation escalated. It is also what happened, who did it, what impact followed, and what responsibility remains.
Consider an ordinary disagreement about household spending. Two people may become defensive, interrupt each other, and later identify how each contributed to the exchange. That does not require equal responsibility for every sentence or decision. If one partner repeatedly uses insults, responsibility for those insults remains with the person who used them, regardless of the other partner’s contribution to the disagreement.
Likewise, a conflict sequence does not explain away the underlying disagreement. A person may improve the way they raise an issue while still having a legitimate concern about a recurring broken agreement.
Break the Loop offers an educational way to reflect on recurring conflict. It does not establish that two people share equal responsibility for harmful behavior. It is not a method of determining whether abuse has occurred or whether someone is safe.
The difference between difficulty and repeated harm
An imperfect attempt at change can include a lapse followed by specific acknowledgment, a changed response on other occasions, and continued attention to the original concern. None of these elements guarantees success. They do, however, describe something more concrete than repeating a promise.
Repeated harm can look different. The person who acted harmfully may acknowledge the problem immediately after each incident but return to the same conduct. An agreement may be revised several times while the same violation continues. One partner may consistently decline to discuss an important concern, without offering a safe or workable alternative. A betrayal may recur after earlier assurances.
These situations differ in seriousness and context. An avoided conversation is not the same as violence. A missed agreement is not automatically deception. A difficult withdrawal may be a request for space, a limitation in communication, or a protective choice. Its meaning cannot be established from the other partner’s interpretation alone.
The point is not to sort every relationship into two categories, trying or not trying. It is to resist an easier but less useful shortcut: assuming that the existence of an explanation or an apology answers the question of what keeps happening.
What words can acknowledge, and what they cannot prove
An apology may communicate acknowledgment, responsibility, remorse, or an intention to act differently. Lewis et al.’s (2015) review describes these as recurring elements in how apologies are understood. The review does not identify a single combination of words that reliably prevents later harm.
Trust-repair research also points toward accountability and dependable behavior, but the available literature is varied and heavily concerned with betrayal and infidelity. In a systematic review of 13 articles, Giacobbi and Lalot (2025) found recurring themes rather than a proven, universal procedure for restoring trust.
These findings support a careful distinction: saying what one intends to do and later doing it are different events.
An acknowledgment may be sincere without demonstrating that a behavior has stopped. A promise may be specific without being fulfilled. A detailed explanation may help someone understand the circumstances without cancelling the effect of the conduct.
That does not make apologies meaningless. It means they should not carry more evidentiary weight than they can bear. If the concern is repeated behavior, the next relevant information may come from what happens afterward, not from how persuasive the last conversation felt.
A fictional example: an agreement that keeps disappearing
Imagine two partners, Taylor and Jordan, who agree that either will discuss a substantial shared purchase before making it. Jordan makes a purchase without discussing it. Taylor explains the financial impact and the sense of being excluded from a decision they had agreed to make together.
Jordan apologizes and agrees to talk before the next purchase. Some time later, a comparable purchase happens without discussion. The explanation is different. The agreement is repeated. Eventually Taylor says, “We keep agreeing, but the decision still gets made without me.”
The example does not prove Jordan is acting maliciously. Nor does it prove that Taylor’s preferred financial arrangement is automatically fair. Either person might raise questions about the amount, practicality, or scope of the agreement.
What is observable is narrower: an agreement was made; the behavior did not consistently match it; the underlying decision rule still needs attention. The issue is not resolved merely because the two people understand the emotional meaning of another argument.
In a safe relationship, the partners might reconsider whether the agreement is realistic and mutually acceptable, or decide on a different, explicit arrangement. They might also recognize the continued mismatch between the agreement and the conduct, without pretending to know the reason for it. An educational framework cannot force willingness or supply the missing follow-through.
When the same verbal harm returns
Verbal harm deserves separate attention because it can be mistaken for a predictable part of any difficult conversation.
A sharp remark, criticism of a specific action, repeated humiliation, and a threat are not interchangeable. Disagreement can be direct without being abusive. An angry response may still be subject to accountability. Repeated degrading language is not made harmless by labeling it an Overload response or a Protection Move.
If an insulting exchange repeatedly ends with an apology, a practical question remains: Has the language changed when the next difficult moment arrived? The answer may be incomplete. A person may make genuine changes in some situations and still repeat behavior that harms the other person.
That is not a reason to invent a score or declare an exact amount of progress sufficient. It is a reason to distinguish statements of regret from the continuing impact of observable conduct.
The responsibility to stop using harmful language rests with the person using it. The recipient does not have to become better at asking, listening, staying calm, or finding the perfect words before that responsibility applies.
Avoidance is not always the same behavior
Someone may stop a conversation because they are overwhelmed and need space. Another person may avoid a discussion because it is painful, embarrassing, or difficult to navigate. Yet another situation may involve using silence, threats, withholding of access, or retaliation to control what can be raised.
These cannot be sorted reliably by appearance alone.
In an ordinary, safe disagreement, a requested pause with a voluntary intention to return may allow the substantive question to remain available. Chronic avoidance without a workable return can leave an important decision unresolved. The person affected may have a legitimate concern about the lack of engagement even without knowing why it happens.
When silence or disengagement is coupled with intimidation, fear, isolation, punishment, or threats, treating it as a mutual Loop can be dangerous. The safety question comes first. A person who disengages to protect themselves from coercion is not failing at communication.
The same distinction protects against guessing motives. An observable pattern of missed discussions can be named without announcing that the other person is selfish, incapable of intimacy, or deliberately trying to cause pain. Understanding a possible function is not an excuse for harmful conduct, and the absence of an explanation does not make its impact disappear.
Trust can be affected by repeated betrayal
Broken trust can involve many kinds of conduct, from concealed spending and repeated secrecy to violations involving intimacy or access to personal information. These are not interchangeable in seriousness, and their meanings depend on the agreements and circumstances involved.
Research about amends illustrates why outcomes must be kept separate. In Hannon et al.’s (2010) study of 78 undergraduates in ongoing dating relationships who completed two-week incident diaries, reported amends were associated with greater perceived resolution of betrayal incidents, even after adjustment for reported forgiveness. That association does not demonstrate that making amends prevents future betrayals, that trust returned, or that an intervention caused recovery.
An affected person may understand an explanation and still have questions about reliability. They may accept an apology without feeling ready to restore the same level of trust. They may decide that a particular agreement is no longer workable. None of those responses supplies a verdict about what they must do with the relationship.
Restored contact, forgiveness, renewed trust, and future safety are different questions. No one is obliged to provide one as proof that the other has occurred.
What the research cautions about forgiveness
One longitudinal study is especially relevant to the assumption that forgiving more readily must always improve what happens next.
McNulty (2011) followed 72 first-married couples and examined reported tendencies to express forgiveness alongside partners’ reports of psychological and physical aggression across four years. In that selected sample, greater reported forgiveness tendency was associated with aggression trajectories that remained more stable, while lower reported tendency was associated with declining aggression.
This was an observational study, not an experiment in which couples were assigned to forgive or not forgive. It measured a reported tendency rather than all forms of private forgiveness. The sample was young and predominantly White, and only 37 couples remained at the final analyzed wave. These limits matter.
The result does not establish that expressing forgiveness causes aggression. It certainly does not make the harmed person responsible for another person’s behavior or justify a prescription to withhold forgiveness. It does challenge a simpler claim: that forgiveness can be assumed to prevent continued harm in every relationship.
The more defensible conclusion is also the more limited one. Forgiveness and later conduct must not be treated as the same outcome.
What accountability can and cannot settle
It may help to distinguish several questions instead of folding them into one assessment of whether someone “really means it.”
What happened? Describe observable behavior without exaggerating or minimizing it.
Who is responsible for that conduct? Responsibility is not automatically divided equally because two people were present in an argument.
Was the impact acknowledged? Someone can recognize the effect of an action without requiring the harmed person to adopt an explanation for its motive.
What changed afterward? A stated plan and observable follow-through are different information. It may take further encounters to learn more, but no universal waiting period establishes safety.
What remains unresolved? The concrete agreement, damaged trust, protective boundary, or need for outside help may still be present even after a constructive conversation.
These are educational distinctions, not a validated checklist for predicting whether someone will harm another person again. The answers may be uncertain, contested, or unsafe to seek directly.
They also leave room for an important difference: a person can be genuinely attempting change while another person remains affected by earlier harm. Acknowledging effort does not require that person to declare the harm repaired. Recognizing continuing harm does not require a diagnosis of the actor’s hidden intentions.
When safety takes priority over pattern analysis
Not all repeated harmful conduct belongs in ordinary couple-conflict education.
The Centers for Disease Control and Prevention (2026) identifies intimate partner violence as including physical and sexual violence, stalking, and psychological aggression, including verbal or nonverbal conduct intended to harm or exert control. It can occur as a single episode or across a prolonged pattern. A repeated argument is not automatically intimate partner violence, but the absence of physical injury does not rule it out.
Threats, fear of retaliation, stalking, sexual coercion, intimidation, restricted access to resources, isolation, and other controlling behavior are reasons to stop treating the situation as an ordinary shared Loop. Someone experiencing these conditions does not need to accept reciprocal blame, practice joint interruption, confront the other person, or explain their choices through a Protection Move.
Where there is danger or coercion, individualized assessment and specialized domestic-violence support may be more appropriate than couples exercises. Seeking help may itself carry risk if a partner monitors communications or retaliates, so the safest way to access assistance can depend on the situation. When there is immediate danger, contact local emergency services where it is safe to do so.
This chapter does not determine whether any particular relationship is abusive, provide a risk prediction, or tell a reader to stay or leave. Its responsibility is to make clear that safety and accountability cannot be subordinated to a wish to make the relationship pattern look mutual.
When further support may be appropriate
A repeated concern may also exceed what a self-guided educational exercise can reasonably address without meeting the definition of abuse.
For example, continuing betrayal, repeated violations of consequential agreements, persistent harmful speech, or an inability to discuss important decisions may call for a more individualized assessment of what is happening and what options are available. Qualified professional support can help someone consider their own circumstances. Where coercion, threats, or fear are involved, specialized individual support is the more appropriate starting point than an assumption that joint couple work is safe.
An educational tool cannot substitute for a professional assessment, and professional help is not a guarantee that another person will change. Neither is there a requirement to keep inviting discussion until the other person participates.
A boundary may be about the person’s own participation, contact, resources, or next decisions. It does not need to be framed as a technique for producing another person’s compliance. Where there is a possibility of retaliation, even stating or enforcing a boundary may require individualized safety consideration rather than a generic communication script.
What Break the Loop can still offer, and where it stops
In ordinary, safe disagreement, a person might recognize the familiar Launch, notice Overload, question an Old Story, and choose a different response instead of a familiar Protection Move. Those are BTL teaching categories, not independently validated findings about any particular couple.
The same educational perspective can help a reader distinguish what they themselves did from what a partner did. A person might choose to speak more clearly, refuse an insult, take appropriate space, or address an agreement directly. None of these choices makes another person’s harmful conduct their responsibility. Nor does using NAME IT → SLOW IT → CHECK IT → CHOOSE IT require giving up an important concern or remaining in an unsafe exchange.
The framework does not measure a partner’s sincerity or determine whether repeated conduct will stop. It does not establish why a promise was broken, and it has not been independently shown to stop repeated aggression or betrayal. It cannot replace the need for accountability, observed change, protection, or specialized help.
Sometimes progress is incomplete but real. Sometimes a familiar apology accompanies conduct that continues. Other times the most important question is no longer how both people can interrupt an argument, but whether the harmful behavior is stopping and whether the person affected is safe.
Those questions deserve to remain visible, even when a relationship has become practiced at discussing the same incident afterward.
Back to the Field Guide table of contents