Chapter 8: Attachment and Recurring Conflict
Part II: Interaction and Context
Attachment offers a way to think about what it means to depend on another person. In a recurring disagreement, the question may involve more than whether a task gets done. Someone may also be asking whether a partner will be available, whether closeness is welcome, or whether relying on that partner feels possible. Attachment research gives those questions a place in our understanding of couple conflict. It does not settle what any particular argument means.
The distinction matters when a label seems to explain everything. Calling one partner anxious and the other avoidant can appear to organize a confusing exchange. Yet the label still leaves essential questions unanswered: what happened, what each person expected, what the actual problem was, and whether either person behaved harmfully. This chapter uses attachment to sharpen those questions while keeping the current relationship and the concrete disagreement in view.
What researchers mean by adult attachment
In adult romantic attachment research, two widely used dimensions describe different concerns. Attachment anxiety concerns uncertainty about a partner’s availability and worry about rejection or abandonment. Attachment avoidance concerns discomfort with emotional closeness and dependence. Relatively low anxiety and avoidance are commonly described as greater attachment security (Simpson & Rholes, 2017). These are descriptions of attachment orientations within a research tradition, rather than diagnoses of mental disorders.
A dimension describes variation by degree. Asking how uncomfortable someone feels depending on a particular partner leaves room for more than an either-or answer. It also leaves room to specify which relationship is being described. Throughout this chapter, the terms anxiety and avoidance refer to these attachment dimensions. They should not be read as diagnoses of an anxiety disorder or as explanations for every instance of avoiding a conversation.
Attachment theory also uses the term working models for expectations about oneself and significant others. In plain language, the relevant question is what someone expects when closeness, help, or reassurance is at stake. A theory of those expectations can help organize an inquiry. It cannot establish, from the outside, that a person secretly fears abandonment or intimacy. The distinction between a theoretical explanation and a verified account of an individual’s experience must remain visible (Simpson & Rholes, 2017).
Security should not become a moral ranking. In this guide, a person does not earn authority in an argument by claiming to be secure. Nor does reporting insecurity make someone’s complaint less credible. Whether an agreement was broken still requires attention to that agreement. Whether an accusation is supported still requires attention to the evidence. An attachment description cannot do either job for us.
Why context belongs in the explanation
Reviews connect attachment concerns with responses to stressful relationship situations, while emphasizing that those responses depend on context. Greater anxiety does not mean someone is always demanding; greater avoidance does not mean someone is always withdrawn. What threatens the relationship, what support is being sought, and how a partner responds are part of the account. These are tendencies discussed across studies, not a rule that assigns each person a permanent conflict role (Simpson & Rholes, 2017).
Attachment theory describes two broad ways of managing attachment concerns. Hyperactivation involves intensifying attention to the relationship threat and bids for closeness or reassurance. Deactivation involves distancing or relying on oneself while limiting attention to attachment concerns. These are theoretical descriptions of responses associated with attachment anxiety and avoidance, respectively. They do not establish what any particular request or departure is trying to accomplish (Simpson & Rholes, 2017).
Partial clarity
The distinction can clarify a question without supplying its answer. In the attachment account, repeated requests might be considered in relation to uncertainty about availability. A person’s distance might be considered in relation to discomfort with dependence. Those possibilities still require the context and the person’s account. A missed obligation, an unwanted request, or a need for safety cannot be explained away by naming a possible attachment response.
For understanding a particular disagreement, begin closer to the event. Was someone asking for help, raising a legitimate concern, declining a request, or trying to finish a conversation? What did the person say the moment meant? What did the partner actually do next? Those questions preserve information that would disappear if the entire exchange were summarized as an anxious person pursuing an avoidant person.
An attachment explanation should also remain distinguishable from other explanations. An unanswered message might raise questions about availability. It might also be part of a concrete disagreement about an agreed plan. Identifying a possible attachment concern does not resolve whether the message could reasonably have been answered, whether the expectation was shared, or whether the agreement needs to change. The practical issue remains a separate subject for the couple to address.
Chapter 4 examines demand-withdraw interaction in detail. An attachment orientation and an observed interaction role should not be treated as interchangeable descriptions. Here, the useful question is whether attachment-related expectations help clarify the interaction without replacing what is known about it. We do not need to assign a style to describe a repeated request, an interruption, a refusal to answer, or a stated need for space.
Expectations can differ across relationships
Research on relationship-specific attachment makes an important distinction between expectations about close relationships in general and expectations about particular people. The Experiences in Close Relationships–Relationship Structures questionnaire was developed to assess attachment in relationships with parents, a romantic partner, and a close friend. Its development studies found that attachment scores across those relationships were not interchangeable. A broad orientation does not capture every detail of a particular relationship (Fraley et al., 2011).
That finding changes what an attachment statement needs to specify. “I find it difficult to rely on people” and “I find it difficult to rely on you to keep this agreement” are different statements. The second names a particular relationship and a specific issue. It would be a mistake to translate it immediately into a general personality description. It would be equally mistaken to assume that every relationship-specific concern proves the partner caused it.
The measurement evidence does not decide whose explanation of a disagreement is correct. The studies relied on people’s reports, and the instruments have limitations. They do not independently verify what a partner intended or establish the cause of a person’s expectations. Their contribution here is narrower: they give a reason to specify the relationship being discussed rather than treating one general score as a complete account of the relationship (Fraley et al., 2011).
Consider the difference between asking, “What attachment type are you?” and asking, “What have you come to expect from me when you raise this concern?” The latter is an example of a more specific question, not a validated exercise or a required conversation. It leaves room for the person to describe experience without accepting a label. In an unsafe relationship, there is no requirement to ask it jointly or disclose an answer to a partner.
What attachment labels leave out
The scientific question of categories versus dimensions is more complicated than choosing a familiar label. A study using two large online samples found that self-reported attachment was, overall, more consistent with dimensions than distinct categories. It examined both general and relationship-specific orientations. The authors also identified limits: the study used self-report measures, and it did not directly test whether dimensional scores predicted relationship outcomes better than categories (Fraley et al., 2015).
A later study in Spain explored a hybrid approach that allowed dimensions and profiles to coexist. Among 619 adults, the results depended on the questionnaire and the model-selection decision. For the Experiences in Close Relationships measure, adding a second profile did not meet the study’s significance criterion. The Adult Attachment Questionnaire–Revised supported additional profiles; the authors selected two for simplicity after a three-profile model had provided the statistical optimum. This was not a confirmation of a universal four-type system (Morales-Vives et al., 2025).
Careful with labels
These findings justify care in both directions. They do not support presenting popular labels as settled natural kinds. They also do not justify announcing that every scientific question about attachment categories has been closed. The studies used different instruments and analytical approaches, with selected samples (Fraley et al., 2015; Morales-Vives et al., 2025). For this guide, the appropriate approach is to explain the dimensions, acknowledge the measurement debate, and refrain from using a short label as a verdict about a person.
A label also cannot supply missing observations. If someone leaves a conversation, the fact available is that the person left. An account of what they experienced would add information. Calling the departure avoidance does not, by itself, establish attachment avoidance, the purpose of leaving, or whether leaving was appropriate. The same discipline applies to reassurance: a request for reassurance does not by itself establish an attachment orientation or decide whether the request was reasonable.
This is especially important when a label becomes an argument against a boundary. Declining sex, asking for privacy, refusing insults, and leaving a threatening conversation should not be reclassified as attachment problems simply because another person wants continued access. Consent and safety do not depend on an attachment score. The conduct and circumstances need to be addressed directly.
The current relationship still matters
Attachment theory includes the influence of earlier relationships, but a childhood explanation should not be imposed on a present disagreement. A major review describes experiences across life, including adult romantic relationships, as relevant to attachment expectations and change. Earlier experience is part of the framework; it is not a complete explanation of a particular adult’s response. Nor does the possibility of change establish that a particular partner or educational method will produce it (Overall et al., 2022).
For the reader, this sets a limit on retrospective certainty. A person may connect a present concern with an earlier experience in their own account. That is different from a partner declaring, “This is your childhood talking,” as a way to dismiss current evidence. A present broken agreement does not become irrelevant because it resembles an earlier disappointment. Both the reported history and the current conduct can remain subjects of attention without one cancelling the other.
The same review describes partner buffering: responses that can reduce attachment-related distress or defensive responding under studied conditions. It also discusses important limits, including adversity, the difficulty of providing support, and the need to consider people’s own efforts and capacities. The reviewed couple and family literature primarily involved nonmarginalized families with two married parents, which restricts how broadly its findings can be applied (Overall et al., 2022).
Buffering should therefore be presented as a conditional research topic, not an obligation to manage a partner’s reactions. This chapter does not prescribe a style-specific reassurance script or promise that the right response will create lasting security. Chapter 9 takes up responsiveness, regulation, and recovery more fully. Here, the point is that attachment accounts need room for the partner and the circumstances without assigning either person responsibility for preventing the other’s harmful behavior.
A disagreement described without typing the partners
The following example is fictional. One partner agrees to confirm a weekend visit with relatives by Thursday. Thursday passes without an answer. On Friday morning, the other partner asks again and says, “I cannot make my plans until I know.” The first partner replies, “You have asked me three times. I need a few minutes before we discuss this.”
Several facts are available. There was a Thursday agreement, the confirmation has not happened, a repeated request has been made, and a pause has been requested. The example does not establish either person’s attachment orientation. It also does not establish why the confirmation was missed. Those are missing pieces of information, not invitations to fill in a preferred theory.
Suppose the person waiting says, “When you do not answer, I start wondering whether I can count on our plans.” That adds a reported expectation to the account. Suppose the other says, “I am willing to discuss it, but I want to finish breakfast first.” That adds a stated request. Neither statement requires us to decide that the couple consists of an anxious pursuer and an avoidant withdrawer.
What’s still needed
The missed confirmation still needs an answer. A pause request can be considered without treating it as a diagnosis or automatically excusing the delay. The person who missed the agreement can acknowledge that conduct; the person waiting can describe its impact. Attachment may suggest questions about reliance or closeness if the people themselves find those questions relevant. It does not replace the practical decision about the visit.
Imagine instead that the exchange includes a threat of punishment for making independent plans. The appropriate analysis changes. We should not redescribe a threat as a bid for attachment reassurance and ask the other person to respond more soothingly. Fear, coercion, or retaliation requires attention to safety and accountability. A shared account of recurring conflict is insufficient when it blurs responsibility for threatening behavior.
Where attachment and BTL meet
Break the Loop, the educational framework developed by Aaron Jarrels, uses Old Story to name a familiar meaning or prediction that becomes influential during conflict. It uses Protection Move for the person’s next response as they try to manage what the moment has come to mean. These are BTL terms. An attachment orientation, a moment’s meaning, and an observable response describe different things.
In the fictional example, “I cannot count on our plans” could become relevant to an Old Story if it is a familiar meaning for that person. That does not make Old Story another name for attachment anxiety. Asking for a pause does not establish attachment avoidance, and it should not automatically be labeled a Protection Move. The purpose, context, safety, and the person’s own account remain relevant.
Attachment research can inform questions about expectations and responses without validating the BTL Loop. It also provides no basis for sorting people into the Three Pillars, BTL’s lenses for relational significance. Attachment anxiety should not be treated as the same thing as Importance, or avoidance as the same thing as Autonomy. The complete BTL model belongs in Chapter 11; the fuller discussion of Old Story belongs in Chapter 14.
Chapter 6’s distinction between interpretation and evidence remains useful here. An attachment-based interpretation is still an interpretation when the necessary information is missing. It should not be used to tell someone that a concern is imaginary, that a boundary is pathological, or that a harmful action was inevitable. Understanding a possible influence preserves the need to examine what happened and who is responsible for it.
Using attachment without making it the whole explanation
For a particular recurring disagreement, an appropriately limited account can name the issue, describe the behavior, and include expectations that people actually report. An attachment-based interpretation should be presented as a possibility when the supporting information is limited. Evidence about groups should not be converted into certainty about an individual. Leave the question open when information is missing.
The research reviewed here does not provide a complete account of every couple. Selected questionnaire samples, differing measures, and review-based synthesis place limits on the conclusions. The 2025 profile study did not observe couples arguing, and the relationship-specific questionnaire studies did not establish what caused a particular expectation (Fraley et al., 2011; Morales-Vives et al., 2025). Those distinctions matter whenever we are tempted to move from a useful research concept to a confident explanation of someone’s life.
Safety sets an additional boundary. If there is fear of a partner, threats, coercion, stalking, violence, sexual coercion, or concern about retaliation, prioritize safety. Individual support from a qualified professional or a specialized domestic violence service may be more appropriate than shared conflict work. Joint attachment discussions, vulnerability disclosure, or mutual pattern mapping may be unsafe or inappropriate. Neither insecurity nor a wish for closeness excuses coercion. A reader does not need to determine a partner’s attachment orientation before taking a safety concern seriously.
Attachment can remain part of an accurate understanding of recurring conflict when its explanatory limits stay visible. The next chapter examines responsiveness, regulation, and recovery: processes to consider when asking what may change the course of an interaction, with the same attention to context, evidence, agency, and safety.
Back to the Field Guide table of contents