Can a Marriage Survive Without Emotional Intimacy? What Relationship Research Can Predict
Updated: Sep 17
Author: Ukrainian Psychological Hub · Published: September 17, 2026 · Editorial Policy
September 17, 2026 · Editorial Policy
Yes. A marriage can remain legally intact for years with very little emotional intimacy. But that answer becomes misleading unless “survive” is defined. Staying married, functioning as a household, feeling satisfied with the relationship, experiencing personal well-being, and rebuilding emotional connection are different outcomes. Relationship science can identify patterns associated with some of those outcomes, yet it does not provide a validated emotional-intimacy threshold below which a marriage must end, nor a reliable countdown telling an individual couple how long it has left.
That distinction matters because search results often turn this question into a verdict: either emotional intimacy is declared absolutely necessary for marriage, or couples are reassured that any marriage can endure if both people simply try harder. The evidence supports neither shortcut. Long-term relationships are dynamic systems shaped by relationship-specific perceptions, commitment, stress, communication, adaptation, investments, constraints, life transitions, individual differences, and the interaction between partners. Emotional intimacy is important inside that system, but it is not a solitary switch that determines divorce.
In relationship science, emotional intimacy is commonly understood as an interpersonal process in which personally meaningful disclosure is met by a response experienced as understanding, validating, and caring. Laurenceau, Barrett, and Pietromonaco (1998) showed that self-disclosure, partner disclosure, and perceived partner responsiveness contributed to experienced intimacy in everyday interactions. A contemporary review by Arican-Dinc and Gable (2023) describes perceived responsiveness as a central process through which partners feel understood, valued, and cared for. For a broader definition, examples, and mechanisms, see our guide to emotional intimacy in marriage.
Can a Marriage Survive Without Emotional Intimacy? The Short Answer
A marriage can survive in the narrow sense of continuing. Legal marriage is a status, and relationship continuity is an observable outcome. Neither requires that both partners report a particular level of emotional closeness. People remain married for many reasons: affection, commitment, shared history, parenting, practical interdependence, financial arrangements, values, social ties, caregiving, uncertainty about alternatives, hope for change, or the simple fact that leaving a long relationship is consequential. Some of these forces reflect genuine dedication; others function more like constraints. Usually several operate at once.
What the continuation of a marriage does not tell us is how the marriage feels from the inside. Two spouses may stay together while feeling deeply connected. They may stay together while functioning as effective co-parents and household partners but sharing little inner life. They may remain committed while one partner feels lonely. They may also pass through a period of low intimacy during grief, illness, caregiving, a new baby, unemployment, or intense work stress and later reconnect. The same outward fact — “still married” — can therefore represent very different relational realities.
This is why the useful scientific question is not “Can marriage survive without emotional intimacy?” as though the answer were a biological limit. It is: what outcomes are we trying to predict, over what time period, and from which combination of relationship processes? Once those pieces are separated, the evidence becomes much more informative.
What Does “Survive” Mean in Relationship Research?
Staying legally married
The simplest outcome is whether the marriage remains intact. This is the outcome most people have in mind when they fear divorce. Research can identify variables associated with relationship persistence and dissolution, but no single variable determines the outcome. Commitment research is especially useful here because it distinguishes the desire to maintain a relationship from the investments and constraints that can make continuation more likely even when satisfaction is imperfect.
An updated meta-analysis by Tran and colleagues (2019) synthesized 202 independent samples and 50,427 participants examining the Investment Model. Satisfaction showed the strongest aggregate association with commitment, followed by investments and the perceived quality of alternatives. An earlier meta-analysis by Le and Agnew (2003) also found that commitment significantly predicted relationship breakup. These studies are broader than married couples alone, so they should not be converted into a marriage-specific formula. Their central lesson is still relevant: relationship persistence is explained by more than one dimension of closeness or satisfaction.
Relationship quality
A second outcome is the quality of the relationship while it continues. This includes satisfaction, affection, trust, appreciation, responsiveness, conflict, sexual satisfaction, commitment, and other dimensions that couples may value differently. A legally stable marriage can score high, medium, or low on these dimensions. Emotional intimacy belongs primarily to this domain: it concerns the lived quality of psychological connection, not marital status itself.
The largest multi-lab prediction project in this area illustrates the difference. Joel and colleagues (2020) used machine learning across 43 longitudinal dyadic datasets from 29 laboratories. Relationship-specific variables predicted substantially more of relationship quality than broad individual differences. Among the strongest predictors were perceived partner commitment, appreciation, sexual satisfaction, perceived partner satisfaction, and conflict. Yet even with a very large predictor set, future relationship quality was only partly predictable, and change in relationship quality over time was especially difficult to predict. That is a warning against treating any single experience — including low emotional intimacy — as a deterministic forecast.
Personal well-being
A third outcome is the well-being of the people in the marriage. Marital quality and personal well-being are related, but they are not interchangeable. A meta-analysis by Proulx, Helms, and Buehler (2007) synthesized 93 studies and found a moderate association between marital quality and personal well-being, with smaller but still meaningful longitudinal associations. The evidence does not imply that marriage quality is the sole cause of well-being; personal circumstances and psychological functioning can also affect how relationships are experienced. It does show why asking only whether a marriage stays together can miss a major part of the outcome that matters.
Why Emotional Intimacy Matters Without Becoming a Divorce Test
Emotional intimacy matters because close relationships are partly built through repeated experiences of bringing something personally meaningful to another person and discovering what happens next. A disclosure met with curiosity, understanding, validation, and care makes future disclosure easier. A disclosure repeatedly met with indifference, contempt, ridicule, defensiveness, or immediate dismissal can make future openness less attractive. Over time, these micro-interactions can accumulate into a sense that the partner is — or is not — a reliable place for one’s inner life.
Perceived partner responsiveness is useful here because it connects everyday language such as “I feel seen” or “I can tell you what is really going on” to a measurable relationship construct. Arican-Dinc and Gable’s 2023 review describes responsiveness as the perception that a partner understands one’s thoughts, goals, and needs, validates the person, and cares about their well-being. That process is central to intimacy. It is therefore reasonable to treat persistent low responsiveness as relationally important when it conflicts with what one or both partners want from the marriage.
It is not reasonable to turn that importance into a universal diagnostic threshold. Emotional intimacy is not a DSM or ICD diagnosis, and there is no clinical laboratory value for “enough” intimacy. Couples also differ in preferred disclosure, emotional expressiveness, solitude, affection, conversational style, and cultural expectations. Some people experience closeness through long verbal conversations; others experience it through shared activity, quiet companionship, caregiving, touch, humor, ritual, or reliable practical support. The relevant question is whether the couple’s pattern provides the degree and form of connection that the partners themselves experience as meaningful and sustainable.
For readers trying to decide whether what they are experiencing is actually low emotional intimacy, our separate article on lack of emotional intimacy in marriage covers signs, causes, effects, and differential explanations in depth. This article stays with the prognosis question: what follows once a low-intimacy pattern exists?
What Relationship Research Can Actually Predict
Current relationship-specific perceptions tell us more than a single label
The strongest general lesson from modern prediction work is that the way people experience their specific relationship matters greatly. In the 43-dataset project by Joel et al. (2020), perceived partner commitment, appreciation, sexual satisfaction, perceived partner satisfaction, and conflict emerged among the most robust self-report predictors of relationship quality. This does not mean that each variable independently causes the future. It means that a richer picture of the relationship carries more predictive information than a single label such as “emotionally distant.”
That finding also exposes a common mistake in online relationship content. A page may list several behaviors, add them up informally, and then treat the total as a prognosis. Relationship science does not support that kind of home-built scoring system. Reliable prediction requires validated measures, appropriate populations, known follow-up periods, and models tested beyond the sample that produced them. Even sophisticated models leave a large amount of individual variation unexplained.
Communication patterns have predictive value, but the effects are modest
Communication matters, yet the strongest evidence is less dramatic than popular “divorce predictor” claims suggest. Kanter and colleagues (2022) meta-analyzed 64 dyadic studies with 1,784 parameter estimates. Positive and negative communication behaviors showed small associations with later relationship quality. Negative communication also showed a statistically significant small-to-moderate association with later dissolution, while positive communication did not significantly predict dissolution in the pooled estimate. The authors’ conclusion is crucial for prognosis: a single early assessment of communication is only modestly predictive of later relationship functioning.
Specific interaction patterns can still be important. For example, the demand-withdraw pattern — one partner pressures, pursues, or demands while the other avoids or withdraws — has a substantial research literature and has been associated with relational outcomes. But even a recognizable pattern is not a verdict on a marriage. The pattern may vary by topic, stress level, partner roles, timing, and whether the couple can interrupt it. Scientific constructs help describe processes; they do not turn one conversation into destiny.
Commitment, investments, and constraints help explain why marriages can persist
Commitment research is especially helpful for understanding the apparent paradox of a long-lasting marriage with little emotional intimacy. Relationship continuation can be supported by dedication — wanting a future together — and by investments or constraints that increase the cost of leaving. Those forces may coexist. A couple can feel emotionally thin yet remain strongly committed to a shared life; another couple can feel low dedication but remain together because their lives are deeply intertwined.
The Investment Model meta-analyses by Le and Agnew (2003) and Tran and colleagues (2019) show that satisfaction, investments, alternatives, and commitment are meaningfully connected. The 2019 update is especially valuable because of its scale: 50,427 participants across 202 independent samples. The implication is not that constraints make an unhappy marriage desirable. It is that persistence and quality are empirically separable. A marriage can be stable for reasons that do not tell us whether either spouse feels emotionally nourished by it.
Stress can alter how couples experience each other
Low intimacy does not always originate inside the relationship. External stress can change attention, patience, emotion regulation, and the capacity to respond well to a partner. Neff and Karney (2009) found across diary studies that spouses experiencing greater stress were more reactive to daily relationship experiences, illustrating how stress can constrain adaptive relationship functioning. The classic longitudinal review by Karney and Bradbury (1995) likewise argued that marital quality and stability are best understood through the interaction of enduring vulnerabilities, stressful events, and adaptive processes rather than any one isolated factor.
This is one reason trajectory matters. A six-month reduction in intimate conversation during intensive caregiving may have a different prognosis from a decade-long pattern in which attempts at closeness are repeatedly punished or ignored. The surface description — “we barely talk about feelings” — can look similar while the underlying system is different. Prognosis improves when context is added.
Trajectories tell us more than snapshots
Marriage develops through time, so repeated measurement is more informative than a single bad week or good evening. Lavner and Bradbury (2010) followed newlywed spouses and identified distinct trajectories of marital satisfaction rather than one universal decline. Those trajectories differed systematically in later divorce rates. The study does not establish a general-purpose calculator for all marriages, and it focused on newlyweds, but it demonstrates a key principle: level and direction of change can matter together.
For emotional intimacy, this means asking whether connection is recovering, stable, eroding, or fluctuating with circumstances. A couple that has been distant for six months but is becoming more responsive may be on a different path from a couple whose distance has intensified for years. Research is much better suited to thinking in trajectories and interacting processes than to predicting the future from one symptom at one moment.
What Relationship Research Cannot Predict
Research cannot tell an individual couple, “You have 14 months before divorce because your emotional intimacy is low.” No validated model supports that claim. It cannot identify a universal number of conversations, disclosures, hugs, date nights, or emotionally intense exchanges required for a marriage to remain intact. It cannot infer from low intimacy alone that love is gone, that a spouse has an avoidant attachment style, that someone is narcissistic, that infidelity is occurring, or that divorce is inevitable.
It also cannot convert population-level associations into personal certainties. If a behavior is associated with later dissolution across many couples, that still means couples with and without the behavior appear in both outcome groups. Prediction is further limited by measurement choices, the populations studied, relationship stage, follow-up length, cultural context, historical period, and the fact that partners change their behavior partly in response to what is happening in the relationship. Causes and consequences can become intertwined.
The large-scale machine-learning results are particularly instructive. Joel et al. (2020) could predict a meaningful portion of relationship quality from extensive information, but relationship-quality change was largely unpredictable. That is not evidence that relationships are random. It is evidence that forecasting dynamic human systems is difficult even with far more data than a reader has when typing one question into a search engine.
Similarly, the communication meta-analysis by Kanter et al. (2022) found statistically reliable associations while emphasizing small effects. A risk factor is not a fate. This is the difference between evidence-based prognosis and fortune-telling dressed in psychological language.
Staying Married and Having a Good Marriage Are Different Outcomes
People often use duration as evidence of relationship health: if a couple has been together for 20, 30, or 45 years, the relationship must be strong. Duration is meaningful, but by itself it measures endurance. It does not reveal whether spouses feel safe enough to be known, whether affection is mutual, whether one partner feels chronically alone, whether the couple is satisfied with its level of closeness, or whether both people are choosing the relationship freely and actively.
This distinction does not devalue stability. Stability can reflect loyalty, care, successful adaptation, shared purpose, and profound commitment. It simply prevents one outcome from being mistaken for every outcome. The same principle applies in the other direction: a difficult period of low intimacy does not mean the marriage has failed. Many relationships pass through periods when energy is diverted toward survival outside the couple and emotional bandwidth contracts temporarily.
The association between marital quality and personal well-being found in Proulx et al.’s meta-analysis is another reason to keep quality in view. Whether partners stay together is not the only psychologically relevant question. If low emotional intimacy is persistently distressing to one or both partners, that distress deserves attention even if separation is nowhere under consideration.
Three Marriages Can Look Equally Unintimate and Have Different Futures
Consider three ordinary scenes. They are illustrations, not diagnostic templates, but they show why context changes prognosis.
A depleted marriage after a major life transition
One couple has a six-month-old baby. Most conversations are about feeding, sleep, work schedules, appointments, and who can take the next shift. Neither partner feels deeply known at the moment. Yet when one says, “I miss us,” the other answers, “Me too. I don’t know where the energy went, but I want it back.” There is low current intimacy, high external stress, mutual recognition of the problem, and an intact wish for connection. Research on the transition to parenthood shows that relationship satisfaction often changes during this period, but trajectories vary. Our separate guide to marriage after a baby covers that transition in depth.
A stable but emotionally sparse partnership
Another couple has shared a home for 25 years. They cooperate well, rarely fight, manage money smoothly, care for relatives together, and genuinely respect each other. They seldom disclose fears or hopes and would not describe the relationship as emotionally intimate. If both partners prefer this level of closeness and do not experience it as deprivation, applying an outsider’s intimacy ideal may create a problem the couple itself does not have. If one partner is quietly lonely and the other assumes everything is fine, the same outward arrangement has a different psychological meaning.
A relationship in which bids for closeness repeatedly fail
A third couple also talks mostly about logistics, but the history is different. One spouse repeatedly tries to discuss hurt or uncertainty; the other responds with ridicule, stonewalling, anger, or abrupt withdrawal. The first partner increasingly stops trying. Here low intimacy is part of a recurring interaction process rather than a mutually acceptable style. The future still cannot be read from one pattern, but the couple has more than an abstract “intimacy deficit” to examine: it has a repeated sequence that changes whether vulnerability feels worth attempting.
These scenes can produce the same search query. That is why a useful prognosis begins by identifying the mechanism and trajectory rather than treating “no emotional intimacy” as a uniform condition.
Is a Low-Intimacy Marriage Necessarily Unhealthy?
No universal intimacy standard can answer that question for every couple. Relationship science studies constructs such as satisfaction, responsiveness, commitment, conflict, support, affection, and loneliness, but it does not define one culturally neutral amount of emotional disclosure that every marriage must contain. Partners differ in temperament, communication habits, neurotype, family norms, privacy preferences, and what makes them feel close.
The more useful distinction is between a relationship form the partners genuinely prefer and a relationship form one or both experience as painful, lonely, unsafe, or chronically unresponsive. Low verbal disclosure can coexist with affection and mutual understanding. Conversely, frequent conversation can coexist with little intimacy if disclosures are performative, guarded, or routinely invalidated. Quantity and quality are not the same variable.
This also prevents pathologizing marriages that look unconventional from the outside. A couple may have little sexual intimacy but strong emotional intimacy. A couple may have limited emotional language but reliable care, warmth, and companionship. A couple may temporarily prioritize caregiving over couple-focused time. The clinical and relational question is not whether they fit an internet template; it is how the pattern functions for the people living inside it.
Can Emotional Intimacy Return After Years of Distance?
It can. A long period of low intimacy does not prove that connection is permanently unavailable. Emotional closeness is partly produced through interaction, and interaction can change. That said, “it can change” is a statement about possibility, not a promise that every marriage will recover. Outcomes depend on what sustains the distance, whether partners want change, whether they can tolerate vulnerable conversations, whether harmful interaction patterns can shift, and what external stressors remain active.
Intervention research offers evidence that distressed couples can improve, while also reminding us to avoid overselling certainty. A 2025 systematic review and meta-analysis by Joseph and colleagues found a large pooled post-intervention effect across 12 recent studies of couple interventions, but the authors also reported substantial heterogeneity and signs of publication bias. Those limitations matter: a large pooled effect does not guarantee a large benefit for every couple or every intervention.
A 2026 systematic review by Jackson and colleagues applied empirical-support criteria to 40 eligible studies and reached a more cautious conclusion. No couple therapy model met their threshold for strong empirical support under the criteria used; seven models showed modest support, with insufficient sample sizes and lack of independent replication among the major limitations. Taken together, the recent reviews support a balanced statement: couple interventions can help many distressed couples, while the evidence base varies by model and is less definitive than popular descriptions sometimes imply.
For couples who have already decided that they want more connection, our practical guide on how to reconnect with your spouse and rebuild emotional intimacy focuses on the rebuilding process. The prognosis article you are reading answers the prior question: whether low intimacy itself tells you the marriage is doomed. It does not.
A Better Way to Think About Prognosis
Look at the direction of travel
A snapshot answers, “How connected do we feel right now?” A prognosis needs the next question: “What has been happening over time?” Distance that appeared during a definable stressor and is beginning to ease carries different information from distance that has steadily deepened despite repeated efforts to reconnect. Tracking trajectory does not produce certainty, but it adds the time dimension that cross-sectional labels lack.
Look at what happens when one partner reaches
Intimacy cannot be inferred from how many emotional topics a couple discusses. What happens to vulnerability may be more informative. When one partner says something difficult, does the other try to understand? Is the response curious, dismissive, hostile, distracted, frightened, overwhelmed, or caring but clumsy? Perceived responsiveness is central to contemporary intimacy research precisely because the meaning of disclosure depends heavily on the response it receives.
Look at whether the problem is shared
A couple can disagree about how much closeness they want. One person may experience the marriage as peaceful and sufficiently connected while the other experiences the same pattern as loneliness. Prognosis then depends partly on whether the discrepancy can itself become a subject of respectful conversation. Agreement that a problem exists is different from agreement about its cause, and neither requires partners to have identical emotional styles.
Look at the whole relationship, not only intimacy
Commitment, appreciation, conflict, sexual satisfaction, trust, affection, friendship, shared purpose, practical cooperation, external stress, health, caregiving, and life transitions all shape the system around intimacy. The Joel et al. (2020) findings are useful here because they show that robust prediction of relationship quality comes from multiple relationship-specific variables rather than one dominant signal. A prognosis that ignores the rest of the marriage is structurally incomplete.
Separate dedication from constraint
Two couples may both say, “We are staying together,” while meaning different things. One means, “We choose this relationship and want to improve it.” Another means, “Leaving would be financially, socially, or practically overwhelming.” Commitment theory treats these forces differently. Recognizing the difference can clarify why stability does not always equal satisfaction and why a low-intimacy marriage may persist for a long time without answering whether the current arrangement is wanted.
When Professional Support Makes Sense
Professional support can be useful when emotional distance is persistent and distressing, when conversations repeatedly collapse into the same cycle, when one or both partners want change but cannot create it on their own, or when the couple is struggling to understand whether the problem is relational, situational, sexual, medical, or connected to an individual mental health concern. Couple therapy is one option; individual support may be more appropriate when one person needs space to clarify their own experience first.
Therapy should not be framed as proof that a marriage is failing. It is a structured setting for assessment and change. The current evidence base supports several couple-therapy approaches to varying degrees, while the 2026 review by Jackson et al. also shows why responsible writing should avoid claiming that one branded model is universally proven to save marriages.
If emotional distance occurs alongside fear, threats, coercive control, stalking, or physical or sexual violence, “how to rebuild intimacy” is not the first question. Safety and qualified local support take priority. Couple disclosure exercises can be inappropriate when one partner cannot speak freely without consequences. Low intimacy in that context should not be treated as an ordinary communication problem.
Intimacy in Culture: 45 Years
Andrew Haigh’s film 45 Years follows Kate and Geoff Mercer as they prepare for their 45th wedding anniversary and receive news that reopens a piece of Geoff’s romantic past. The film is useful here because almost nothing about the marriage can be reduced to whether it has “survived.” Forty-five years of continuity are undeniable, yet the story asks what one spouse actually knows about the other, what has remained private, and how new meaning can alter the felt history of a relationship without changing a single date on the marriage certificate.
The film is not evidence about marriage prognosis. It is a cultural illustration of the distinction this research makes necessary: duration, intimacy, knowledge, commitment, and satisfaction are related dimensions, not synonyms. A relationship can possess an impressive history and still encounter a moment in which its emotional architecture is reinterpreted.
Marriage in the Artificial Era
The prognosis question is changing in the Artificial Era because the intimate ecology around a marriage now includes systems that can listen, respond, advise, summarize conflict, simulate empathy, and become places of private disclosure. This creates a new possibility: a spouse may still be married and still be socially connected while moving part of their inner life into conversations with generative AI. That shift can feel relieving, threatening, neutral, or useful depending on the couple and the function the system serves.
A 2026 systematic review by Levkovich and Alon examined 21 studies from 2024–2026 on generative AI as an advisory or mediating “third voice” in human couple relationships. Adults used GenAI for advice, emotional support, and communication assistance, but the evidence was preliminary, with limited longitudinal and dyadic research. The strongest intervention evidence did not show a clear advantage over an established writing exercise, and the review identified concerns including reliability, overreliance, authenticity, privacy, and safety.
For prognosis, the conclusion is straightforward: there is currently no evidence-based rule that says confiding in AI instead of a spouse predicts whether a marriage will survive. The relational meaning has to be examined at the level of function and trajectory. Is AI helping someone organize thoughts before a constructive conversation? Is it becoming the only place where vulnerable material is disclosed? Does a partner know and consent to how shared messages are being used? Does the technology reduce or increase human responsiveness? These are research questions for the emerging psychology of marriage in the Artificial Era, not settled diagnostic criteria.
Frequently Asked Questions
How long can a marriage last without emotional intimacy?
There is no scientifically validated time limit. Some marriages remain intact for many years with little emotional closeness; others end while emotional intimacy is still present because of factors unrelated to intimacy. Duration depends on multiple interacting variables, including commitment, investments, constraints, conflict, stress, alternatives, life circumstances, and whether the relationship remains acceptable to both partners. Research can estimate associations in groups, not provide an expiration date for one marriage.
Does no emotional intimacy mean the marriage is over?
No. Low or absent emotional intimacy describes a relationship experience, not an outcome. It may be temporary, situational, mutually acceptable, chronically distressing, or part of a broader pattern. The next questions are how long it has lasted, whether both partners experience it the same way, what happens when closeness is attempted, and whether the pattern is changing.
Can spouses love each other without much emotional intimacy?
Yes. Love, commitment, attachment, affection, loyalty, caregiving, sexual connection, friendship, and emotional intimacy overlap but are not identical constructs. A person may love a spouse while struggling to disclose, listen, respond, or create the form of closeness the other partner wants. The presence of love also does not automatically make a low-intimacy arrangement satisfying for both people.
Is emotional intimacy more important than sexual intimacy?
Research does not support a universal ranking that applies to every couple. Emotional and sexual intimacy are related but distinct, and their importance varies across people and relationships. The question in this article concerns emotional access to each other’s inner world and responsiveness to disclosure. A sexless marriage is therefore not automatically an emotionally unintimate marriage, and an active sex life does not prove strong emotional intimacy.
Can a marriage be happy with little emotional disclosure?
It can if the partners genuinely prefer their pattern and experience adequate connection through the forms that matter to them. Happiness should not be inferred from an outsider’s preferred communication style. The more clinically relevant concern arises when one or both partners experience the relationship as lonely, chronically unresponsive, or unable to hold personally meaningful experience.
What predicts divorce better than emotional intimacy alone?
No single variable gives a reliable individual prediction. Research performs better when multiple relationship-specific variables and contextual factors are considered together. Large-scale work has highlighted perceived partner commitment, appreciation, sexual satisfaction, perceived partner satisfaction, and conflict as robust predictors of relationship quality, while meta-analytic work finds modest prospective associations between communication patterns and later outcomes. Commitment, investments, alternatives, stress, and relationship trajectories also contribute information.
Can couples therapy restore emotional intimacy?
Couple interventions can improve relationship distress for many couples, and a 2025 meta-analysis found a substantial pooled post-treatment effect. But studies vary, publication bias is a concern, and a 2026 systematic review judged the empirical support for specific therapy models more cautiously. Therapy is therefore best understood as a potentially useful intervention rather than a guarantee that intimacy or the marriage will be restored.
What if my marriage feels like a roommate relationship?
“Roommate marriage” is a popular description for a relationship that has become heavily logistical and low in couple-focused intimacy. It is not a diagnosis and does not itself predict divorce. Our separate article on roommate marriage explains the term, why couples may begin to feel like roommates, and how it differs from temporary emotional distance or legal separation.
The Bottom Line
A marriage can survive without much emotional intimacy if “survive” means remain legally intact. That fact alone says little about relationship quality, personal well-being, or whether either partner wants the marriage to stay emotionally sparse. Relationship science does not provide a minimum-intimacy threshold, a universal set of fatal signs, or a countdown to divorce.
What research can do is identify processes that make prognosis more intelligible: perceived partner responsiveness, communication patterns, commitment, investments, alternatives, external stress, adaptive processes, and the trajectory of relationship quality across time. These factors operate together. Their effects are probabilistic, and even large multi-study models leave substantial individual variation unexplained.
The most informative question, then, is not simply whether a low-intimacy marriage can continue. It is what kind of marriage is continuing, whether the current form is acceptable to both partners, what happens when connection is attempted, and in which direction the relationship is moving. That is where prognosis becomes psychologically meaningful.
Related Articles
References
Arican-Dinc, B., & Gable, S. L. (2023). Responsiveness in romantic partners’ interactions. Current Opinion in Psychology, 53, 101652. https://doi.org/10.1016/j.copsyc.2023.101652
Jackson, J. B., Codecá, L., Miller, R. B., Simpson, J. E., O’Leary, A. M., & Renner, S. M. (2026). Empirically Supported Couple Therapy Models for Couple Relationship Distress: A Replication and Update. Journal of Marital and Family Therapy, 52, e70162. https://doi.org/10.1111/jmft.70162
Joel, S., Eastwick, P. W., Allison, C. J., et al. (2020). Machine learning uncovers the most robust self-report predictors of relationship quality across 43 longitudinal couples studies. Proceedings of the National Academy of Sciences, 117(32), 19061–19071. https://doi.org/10.1073/pnas.1917036117
Joseph, B., Joseph, V. K., & Rajan, S. K. (2025). Effectiveness of Couple Interventions in Marital Distress: A Systematic Review and Meta-Analysis. Iranian Journal of Public Health, 54(1), 112–123. https://doi.org/10.18502/ijph.v54i1.17581
Kanter, J. B., Lavner, J. A., Lannin, D. G., Hilgard, J., & 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
Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability: A review of theory, method, and research. Psychological Bulletin, 118(1), 3–34. https://doi.org/10.1037/0033-2909.118.1.3
Laurenceau, J.-P., Barrett, L. F., & Pietromonaco, P. R. (1998). Intimacy as an interpersonal process: The importance of self-disclosure, partner disclosure, and perceived partner responsiveness in interpersonal exchanges. Journal of Personality and Social Psychology, 74(5), 1238–1251. https://doi.org/10.1037/0022-3514.74.5.1238
Lavner, J. A., & Bradbury, T. N. (2010). Patterns of Change in Marital Satisfaction Over the Newlywed Years. Journal of Marriage and Family, 72(5), 1171–1187. https://doi.org/10.1111/j.1741-3737.2010.00757.x
Le, B., & Agnew, C. R. (2003). Commitment and its theorized determinants: A meta-analysis of the Investment Model. Personal Relationships, 10(1), 37–57. https://doi.org/10.1111/1475-6811.00035
Levkovich, I., & Alon, L. (2026). Generative AI as a third voice in human couple relationships: A systematic review. Computers in Human Behavior Reports, 23, 101255. https://doi.org/10.1016/j.chbr.2026.101255
Neff, L. A., & Karney, B. R. (2009). Stress and reactivity to daily relationship experiences: How stress hinders adaptive processes in marriage. Journal of Personality and Social Psychology, 97(3), 435–450. https://doi.org/10.1037/a0015663
Proulx, C. M., Helms, H. M., & Buehler, C. (2007). Marital Quality and Personal Well-Being: A Meta-Analysis. Journal of Marriage and Family, 69(3), 576–593. https://doi.org/10.1111/j.1741-3737.2007.00393.x
Schrodt, P., Witt, P. L., & 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
Tran, P., Judge, M., & Kashima, Y. (2019). Commitment in relationships: An updated meta-analysis of the Investment Model. Personal Relationships, 26(1), 158–180. https://doi.org/10.1111/pere.12268