Red Flag or Trigger? A Decision Sheet

Gate 0: the safety check

If any of these are true, the sheet stops here.

Fear, control, coercion and violence are not candidates for this exercise. If you are frightened of how somebody will react, if you are managing your own behaviour to avoid a reaction, if money, movement, your phone or who you see is being controlled, if anything physical has happened, or if you are being pressured sexually, then this is not a question about your attachment pattern and no amount of self-examination will settle it. Patterns explain reactions. They never excuse frightening behaviour, by anyone, including you. Endpoint: SAFETY ISSUE, GO NO FURTHER.

The incident, written flat

What happened, with the interpretation taken out.

Write it as it would appear in a transcript. Time, what was said or done, in what order, in whose words. No adjectives about anyone's state of mind and nothing about what it meant. Most people find this unexpectedly hard, and the difficulty is the finding: the flat version usually runs about a third as long as the version you have been running in your head, and the two-thirds that fall away are the part you supplied. Keep both. Gates 1 to 4 use the flat version and gate 5 uses the rest.

Gate 1: observable or inferred?

Could a camera have recorded it?

Split the account into what a camera would have caught and what you concluded. "He replied at eleven the next morning" is observable. "He was punishing me" is inferred. Inference is not automatically wrong and you are allowed to be right about somebody; the point is knowing which pile your case is built from, because a case made entirely of the inferred pile cannot be checked by anyone, yourself included. If the observable pile is empty, that is a strong result on its own and it usually points toward activation.

Gate 2: once, or more than once?

One event, or a pattern with dates against it.

A single instance of almost anything is ambiguous. Three instances with dates is a pattern, and a pattern is much harder for either of you to argue away. So put dates on it. If you cannot produce a second and third example you may still be right, but you are working from one data point and the confidence you feel is arriving from somewhere else. The reverse trap lives here too: eleven examples produced instantly with no dates attached is often the same fear counted eleven times.

Gate 3: the stranger test

Would somebody with no stake see it in the transcript?

Hand the flat version to an imaginary person who knows neither of you and holds no view about the relationship. Not a friend, who has a side, and not the version of a friend in your head, who has your side. Would they see what you are seeing without you pointing at it? If it only becomes visible with the backstory attached, that is worth knowing. Sometimes the history is legitimately what makes it serious, and if so, write down which part of it changes the meaning, because that is the actual case.

Gate 4: only with you, or with everyone?

Does the behaviour appear elsewhere in their life?

Behaviour that turns up with colleagues, family and friends is usually about them, and it is fairly predictable once you know it is there. Behaviour that appears only with you is about the relationship or about the specific pairing, and that difference changes what can be done. This gate catches the reverse case as well: a habit you have been taking personally for a year that turns out to be how they treat everybody, including people they plainly like, which is much less about you than it felt at two in the morning.

Gate 5: how old is the feeling?

The one gate that looks backwards instead of outwards.

Now take the part you supplied. Where have you felt precisely this before, and how old were you? Activation has a signature: the feeling arrives ahead of the thought, it outsizes the event, it is immediately familiar, and it carries a certainty you cannot trace to anything. None of that means the thing did not happen. It means the alarm was calibrated by something older, and an old alarm can still be responding to real smoke. Both is a legitimate answer at this gate and it is the most common one.

Reading your endpoint

Four named outcomes and what each one asks of you next.

THIS IS INFORMATION: observable, repeated, visible to a stranger, not especially familiar. Take it seriously and raise it directly. THIS IS ACTIVATION: mostly inferred, one instance, invisible without the backstory, very old. Settle first and look again in two days rather than acting tonight. THIS IS BOTH: something real happened and it landed on old ground, which is most of the time. Handle the two separately and in that order, soothing the alarm before raising the event, without the six years attached. SAFETY ISSUE: gate 0 already answered it and the sheet stopped there.

Why both is the honest answer most of the time

The framing that forces a choice between a real problem and an old pattern is a large part of why this question stays stuck. People are rarely activated by nothing. Something usually happened, it was usually smaller than the reaction, and the reaction is usually accurate about the past and exaggerated about the present. Holding both at once is harder than picking a side, and it is the only version that lets you raise the real thing without dragging a decade of evidence in behind it.

Landing on activation is not a verdict that you were wrong or an instruction to say nothing. It is a timing instruction. In two days the same event will still be there to be raised, and it will be raised considerably better than it would have been on Thursday night.