Reassurance seeking

Reassurance seeking is asking for an answer you have already been given. Out loud or silently: a question repeated in slightly different words, a result read again, a face checked for its expression, a message reread for tone, a search run for the fourth time on the same symptom.

The thing that makes it worth a page of its own is a single property. Reassurance has a half-life, and the half-life shortens with use. An answer that bought a week in March buys an afternoon by August, and then has to be renewed inside the same conversation. This is not the situation getting worse and it is not a loss of willpower. It is the predictable result of what the behaviour is actually reinforcing.

The reinforcer is the relief, not the information.

When an open question gets closed — by an answer, a search, a glance, a test result — anxiety drops, and the drop is the part the nervous system records. What it learns from the drop is not the content of the answer. It is that this was a category of question which requires closure before it is safe to stop. So the resting level of alarm goes up, the threshold for noticing the next ambiguity goes down, and the next question arrives sooner.

Each round therefore teaches one lesson, and it is the opposite of the intended one: I cannot handle not knowing this without help. Comfort goes up. Tolerance goes down. That is the whole mechanism, and everything else on this page follows from it.

Which also explains why the content of the answer barely matters. Salkovskis and Warwick made the point in 1986 and it has held: reassurance works on the question and leaves the doubt untouched, because the doubt was never a question in the first place. What is being pursued is certainty rather than information, and that distinction is the subject of intolerance of uncertainty, which is the upstream entry for this one. Information can be supplied. Certainty cannot, not by a person and not by a scan — examinations return probabilities, and no examination returns zero.

How fast the drain runs has been measured. In a 1997 follow-up of patients given a clean result after gastroscopy, those who had scored high on health anxiety beforehand were back where they started within twenty-four hours, from the same consultant, with the same words, after the same clean camera. The long version of that study, and of the mechanism, is in why reassurance makes health anxiety worse.

This is why the entry is not filed under health anxiety. The subject changes; the machinery does not.

• Health. Symptoms searched, a patch of skin re-examined, a pulse taken, somebody asked whether this looks like anything. The full pattern is health anxiety.
• Relationships. Do you still want to be here. Are you sure. Are you annoyed with me. Asked at a frequency the relationship cannot absorb, and often asked at night, when no answer will be remembered by morning. The version that has organised itself into a complete loop is relationship OCD.
• Obsessive-compulsive presentations. Confessing a thought to have it declared harmless, asking whether the door was locked, asking the same person twice to be sure the answer was not said carelessly the first time. Rachman's analysis of compulsive checking found a mechanism that applies here exactly: checking degrades the memory of having checked, so the evidence the check produced is itself unreliable, which licenses the next check.
• Generalised anxiety and low mood. Asking whether the decision was right, whether the work was good enough, whether you are a burden. Joiner and Metalsky's work on excessive reassurance seeking found it predicted depressive symptoms over time, and found the interpersonal cost that the next section is about.

Parrish and Radomsky, comparing people with obsessive-compulsive presentations and with low mood, found that reassurance seeking and checking are driven by overlapping factors rather than being separate habits — the behaviour is one thing wearing different subjects. Kobori and Salkovskis mapped how it is actually carried out across anxiety presentations and found the same point from another angle: the forms are many, the function is singular.

Which matters practically, because the indirect forms are the ones that escape notice. Asking nothing out loud and still seeking is the common case. Probing around the edges, inventing an unrelated errand to see how someone reacts, rewriting a message five times, counting the minutes before a reply, going back to a result to read it again. No question was asked, so no rule about asking catches it, and the tolerance falls by the same amount.

The dose-response runs backwards, which is the least intuitive thing here and the most useful.

A fast, confident, total answer is the worst kind. Of course it's nothing, I promise. A hard close produces a large drop; a large drop is a large reinforcement; a large reinforcement means a shorter interval before the next one. So the most generous version of the answer does the most damage, and loving somebody well and answering them quickly are, in this one narrow domain, at odds.

Volume works the same way. Three sources asked separately produce three reassurances and no verification — and the small contradictions between them become new material, because the question was never answerable by information, and now there is a discrepancy to work on. A search engine is the extreme case: free, instant, available at four in the morning, and the only link in the chain that costs nothing, which is what makes it the hardest one to break.

And the interval itself is read as a symptom. When the renewals have to come every few hours, people conclude they are getting worse. They are not getting worse. A shortening interval is the signature of a reinforced loop, not of a deteriorating condition, and it is the single most reliable sign that what is being treated is the doubt rather than the subject.

Mostly undescribed, and the reason this behaviour ends relationships rather than merely straining them.

Start with what is being asked for. Halldorsson and Salkovskis interviewed people about what reassurance seeking is actually for and found that health-anxious participants did not describe seeking support at all. They described seeking resolution. Support means being accompanied while something stays uncertain. Resolution means extracting a verdict. From the doorway the two look identical, and they do opposite things to the loop, because a verdict is a reinforcer and company is not.

Nobody can supply verdicts at that volume indefinitely. The sequence is consistent enough to be worth naming in advance: the answers get shorter, then slower, then flatter, then there is an argument about it, and eventually they stop. The giver's own account of this is usually guilt — they can see that the thing they are being asked for is kindness, and they can also see it is not working, and they have no vocabulary for refusing a kindness.

Then the trap closes. Read from inside the loop, a withheld answer is not neutral. It is the precise conclusion the asker was most afraid of: they have stopped telling me the truth, something is being kept from me, it must be bad. The loop manufactures its own worst evidence out of fatigue that it created itself. Joiner's work puts the outcome in plain terms — the seeking erodes the support it is reaching for, and then the erosion becomes further proof.

This is also why "just stop asking" is useless advice to give and impossible advice to take. It names a quantity and no replacement.

The direction that works is not less contact. It is contact with the closure removed, which is uncomfortable for both people and is the only version that does not feed the loop.

• Ask once, and then decline the renewals. A single clear question is gathering information. The third is buying relief. The line is uncomfortable and it holds, and it is the same line whether the subject is a mole or a relationship.
• Count the indirect forms as asks. Rereading, probing, testing, checking a face. If no question was spoken, the rule still applies.
• Agree it in advance, not in the moment. At the moment of asking, nobody has the composure to renegotiate anything. The useful version, settled in a calm hour, sounds like: I am not going to ask you whether it is fine, because your answer wears off and then I need another one. Sit with me for ten minutes instead. That gives company without closure, which is the combination the loop cannot use.
• Let the gap stay open on purpose, in something small. Send the message without rereading it. Leave one ambiguous remark uninterpreted. Tolerance is built by exposure to the unresolved, which is the finding behind the whole literature, running back to the 1994 work that first identified intolerance of uncertainty as the strongest correlate of chronic worry.
• Expect it to feel like negligence. It will. That feeling is the behaviour's last argument for itself, and it is not evidence.

The scope note. Asking for reassurance is not a disorder and most of it is ordinary life — checking that someone is not upset, asking whether a decision looked sensible, wanting to be told it went well. There is no threshold number of times. The useful question is not how often but what happens afterwards: whether the answer settles things for a reasonable while, or whether it holds for an afternoon and the same question comes back unchanged by evening.

When the pattern has started costing sleep, work, or the patience of the people closest to you, it is also one of the better-understood behaviours in the field and responds to treatment specifically aimed at it, which is a conversation for a clinician rather than for better self-management. The relationship anxiety test reads the version aimed at a partner, the overthinking test reads the loop with no subject attached, and the health anxiety test reads the one aimed at the body.