Companions
Challenge
Blog
←Back to Blog
Relationships

Trust Issues Are a Forecast, Not a Personality: Where They Come From and How to Update Them

ILTY TeamILTY Team•November 19, 2026•15 min read

In crisis? Call or text 988 in the US, 9-8-8 in Canada, or find your country's line at findahelpline.com.

It is a Tuesday night. Your partner's phone lights up face-down on the kitchen counter, and they take it into the other room. You do not ask. You note the time, and you notice you are listening for how long the door stays shut. Later you re-read a friend's "we should catch up soon" for the angle.

If you have typed "why do I have trust issues" into a search bar at 1am, you already know the official answer: something happened, and now you are guarded. What that answer skips is why the guard never comes off with people who have done nothing, and why "just trust them" lands like an instruction to stop seeing.

Trust issues are not a personality but a forecast: a prediction about what people will do, trained on evidence that was real at the time, and kept running by habits that stop new evidence getting in. A personality is something you have. A forecast is something you can update.

What trust issues actually are: a three-layer forecast

In a 1985 framework, Rempel, Holmes and Zanna (1985) described trust in close relationships as three layers: predictability (can you forecast what they will do?), dependability (do they come through when it costs them?) and faith (do you expect them to keep coming through in situations you have not seen yet?). Each layer is a bet about the future. The third is a bet with no data.

Trust issues live at that third layer. You may credit what a person has done so far, but you cannot extend the line, so every new situation is treated as if the record has reset to zero.

Simpson (2007) names the moments where trust is actually tested: trust-diagnostic situations, when your partner's interests and yours pull apart, so what they choose tells you something you could not learn while the interests were aligned. A relaxed Sunday is not diagnostic; a partner choosing between covering for you and looking good in front of their friends is.

People with trust issues are right that these moments matter; the error is treating every moment as diagnostic, including the phone going into the other room.

Where do trust issues come from?

Mostly from accurate learning. A forecast that people will hurt you is usually built from real occasions when people did. The question is not whether the lessons were true but whether you still live in the conditions that taught them.

Betrayal by someone you depended on. Freyd's (1996) definition of betrayal trauma is exact: it "occurs when the people or institutions on which a person depends for survival significantly violate that person's trust or well-being." A parent or partner you could not afford to lose teaches you that trust itself is the exposure.

Gobin and Freyd (2014) connect this kind of betrayal history to how people trust later. In a sample of 124 female students with trauma histories, Platt and Freyd (2015) found that a high-betrayal history predicted more shame and more dissociation after an interpersonal threat. Shame matters here because it keeps the forecast private, and a private forecast is never tested. If the betrayal happened inside an abusive relationship, start with trauma bond.

Inconsistency in childhood, with an honest caveat. The standard story is that unreliable caregiving teaches you closeness cannot be counted on. In the CDC's analysis of 2011 to 2020 survey data, 63.9% of US adults reported at least one adverse childhood experience and 17.3% reported four or more (Swedo and colleagues, 2023). When nearly two in three people share the exposure, it cannot be what separates those who develop trust issues from those who do not.

Nor is attachment style fixed. A meta-analysis of 127 papers covering 21,072 attachment assessments found overall stability of r = .39 from infancy to early adulthood, and no significant stability across intervals longer than 15 years (Pinquart, Feussner and Ahnert, 2013). Childhood may have written the first draft of the forecast; it did not lock the file (attachment styles and anxiety has the fuller picture).

Why don't I trust anyone? Suspicion is ordinary, up to a point

Because some suspicion is standard equipment, and the dial goes a long way before a clinician would name it. In a 2005 study of 1,202 people from a non-clinical sample, Freeman and colleagues found that paranoid thoughts occurred regularly in about a third of respondents, and were linked to avoidant coping. That is one sample, not the population; do not round it to "a third of people are paranoid". Suspicious thinking in people who are not ill is common enough to be unremarkable.

Bebbington and colleagues (2013) mapped the shape of it in 8,580 adults. Endorsement of individual paranoid items ranged from under 2% for the most extreme to nearly 30% for the mildest, and the data fit a continuum with a small, severe class at the far end. One of their four factors was labeled simply "mistrust".

How distrust keeps itself alive

The signs of trust issues in a relationship are mostly quiet, and the quietness is how the forecast protects itself. A model that says everyone leaves should update when someone stays; it usually does not, because low trust runs rules that filter the evidence before it arrives.

Murray, Holmes and Collins (2006) call this risk regulation: an if-then system that trades connection against self-protection, with the dial set by how much you believe your partner values you. If you feel unvalued, a stressful moment reads as a verdict, so you pull back and prepare to be left.

Across 21 days of spouse diaries, Murray and colleagues (2003) found that people who felt less valued read more into stressful events, felt more hurt, and behaved worse toward their partner on the following days. Your partner is now responding to your Wednesday behavior, not your Tuesday fear, and what comes back looks like confirmation.

Over 14 days, Campbell, Simpson, Boldry and Rubin (2010) saw the same signature: lower trust predicted bigger day-to-day swings in how people rated the relationship, and a habit of reading ordinary conflict as more negative. Relationship anxiety describes the reassurance loop that often rides alongside it.

Low trust rarely looks like fights. When Rempel, Ross and Holmes (2001) coded 850 statements from 35 married couples, the low-trust couples did not attack; they made narrow, muted, conflict-avoiding statements. Distrust goes quiet. You stop asking the question whose answer you fear, so you never get the answer that might have changed the forecast.

When it does turn hostile, the evidence is specific. In Downey and colleagues' (1998) study, high rejection sensitivity predicted breakups, and for women high in rejection sensitivity, hostile behavior during conflict was the route by which partners later became more rejecting. That is a finding about that group, not a general law.

How can I learn to trust again?

To get over trust issues, treat the forecast as a model and run experiments on it, at stakes you choose. Deciding to trust is an instruction your nervous system will ignore.

Take graded risks, not leaps

Threat beliefs can update through behavior, and the clearest demonstration is clinical. In the Feeling Safe trial, 130 people with persistent persecutory delusions received a cognitive therapy built on graded behavioral tests of their fears, and conviction in those beliefs fell by a large margin (d = -0.86) relative to befriending (Freeman and colleagues, 2021). That is a treatment for a clinical condition, not for everyday distrust; take only the principle.

Write down a small, specific prediction ("if I say I'm struggling, they'll use it against me later"), pick a situation where being wrong costs little, run it, and record what happened. Then one size bigger. A boundary is what makes a graded risk safe: you decide beforehand what you will do if the prediction comes true. How to set boundaries covers that half.

Run the disclosure loop

In Laurenceau, Barrett and Pietromonaco's (1998) study, intimacy arose from a loop: you disclose, your partner discloses, and you perceive them as responsive to what you said. Emotional disclosure did more than factual disclosure. The muted style from the Rempel study switches this loop off at step one: you cannot gather evidence of responsiveness if you never say anything that needs a response.

Let yourself lean, a little

Many trust issues rest on the belief that depending on someone makes you weaker. Feeney (2007) found the reverse in couples: when a partner accepted the other person's dependence, that person showed less dependence and more autonomous functioning six months later, which the paper calls the dependency paradox. A small, accepted lean is a data point against the model.

The oxytocin shortcut that is not one

Oxytocin has been marketed as the trust hormone. The 2005 Nature paper (Kosfeld and colleagues) reported that oxytocin increased trust in humans. Nave, Camerer and McCullough (2015) reviewed the evidence and concluded it does not robustly show that trust is associated with oxytocin, and a registered replication by Declerck and colleagues (2020) reported "no effect of OT on trusting behaviour". There is no chemical bypass around the experiments.

Repairing trust after a specific breach

If one person's act rebuilt the forecast, the repair has a shape. Kim, Ferrin, Cooper and Dirks (2004) found that the response that works depends on the kind of violation. After a competence violation (they let you down because they could not deliver), an apology repaired trust better. After an integrity violation (they chose to deceive), denial worked better than apology.

ResponseCompetence violation (could not deliver)Integrity violation (chose to deceive)
ApologyMore effectiveLess effective
DenialLess effectiveMore effective

The likely reason denial wins after an integrity violation: an apology concedes that the person was willing to do it. That study was run in a workplace hiring context, not with couples, so read it as a map of how people reason about violations, not a script.

Gottman's (2011) working definition is useful for measuring repair: trust is your partner behaving in ways that raise your rating of the relationship, and betrayal is the zero-sum version, where their gain was your loss. Look for the first over months, not in a single apology. Infidelity is the obvious case, and the formal evidence on recovering from it is thin: in the pilot of Gordon, Baucom and Snyder's (2004) integrative intervention, six couples completed it and most reported less distress. That is a reason to see a couples therapist, not proof that affairs are reliably recoverable.

When is it more than trust issues?

First, when the distrust arrived after a traumatic event and sits alongside other changes. The VA National Center for PTSD lists "You may think the world is completely dangerous, and no one can be trusted" among the ways PTSD shows up, and says that "if symptoms last longer than a month and are causing problems in your life, it could be PTSD." If that sounds like you, see a doctor or mental health professional; our PTSD and trauma page explains where an app fits alongside treatment.

Second, when the suspicion does not track your environment. MedlinePlus describes paranoid personality disorder as "a long-term pattern of distrust and suspicion of others" and notes that people with it "have trouble seeing that their distrust is out of proportion to their environment". Its advice: see a provider "if suspicions are interfering with your relationships or work." Most mistrust is nowhere near this. If people you respect keep saying the forecast is wrong and you cannot find the evidence they see, let a clinician look.

If the distrust comes with hopelessness, or you have started to feel nobody would notice if you were gone, call or text 988 in the US, 9-8-8 in Canada, or find your country's line at findahelpline.com. If the betrayal is happening now, inside a relationship where you are being hurt or controlled, the US National Domestic Violence Hotline is 1-800-799-7233; elsewhere, findahelpline.com lists your country's services.

Frequently asked questions

Can trust issues be a sign of mental illness? Usually not: suspicious thinking is common in the general population and sits on a continuum (Bebbington and colleagues, 2013). It points to a condition when it follows trauma and lasts over a month with other symptoms (PTSD), or when it is long-standing, out of proportion to your environment and damaging your relationships or work (what MedlinePlus describes for paranoid personality disorder). Either way, see a clinician rather than diagnosing yourself.

Can you rebuild trust after betrayal? Yes, in the sense that the forecast can be retrained, though not on a schedule and not by apology alone. It rebuilds through behavior in trust-diagnostic situations, observed over months. After infidelity specifically, the formal evidence is thin enough that a couples therapist is a better bet than any protocol you find online.

Why doesn't my partner trust me? Often because they are running a forecast trained before they met you, and the quiet way low trust shows up hides that it is a forecast at all. You cannot argue them out of it; what helps is being predictable, dependable when it costs you, and responsive when they do risk telling you something.

Can you fix lifelong trust issues? Lifelong overstates how fixed these patterns are: attachment stability across long intervals is weak (Pinquart and colleagues, 2013), so the pattern you learned at seven is not the one you are stuck with at forty. What tends to persist is the habit of avoiding the situations that would update it. Change the habit in small steps and the forecast gets the data it needs to change.

Sources and further reading

ILTY cannot make you trust anyone, and it is not therapy or couples counseling. What it is good for is the experiment: telling a companion the prediction you have not said out loud to anyone, running the small test, and having the prediction still be there when you report back, because the companion remembers what you told it, ends each conversation with two to four concrete steps, and asks how they went the next time you talk. Mr. Relentless is the companion most likely to ask what evidence you have been discarding. Start the 7-day trial.

#trust issues#trust in relationships#betrayal#relationship anxiety#attachment#rebuilding trust

Share this article

ILTY Team

ILTY Team

AI Mental Health Companion

Building an AI companion that actually helps with your mental health.

Get mental health insights in your inbox

No fluff, no toxic positivity — just what actually helps.