What your memory does to the record

The list in your head is not a record. It feels like one — it has items, it has dates, it seems to be sitting there unchanged — but it is being rewritten every time you go through it, and the rewriting all runs in one direction.

This matters because that list is what you will use to decide, and what you will bring to the conversation.

What happens once you suspect

Recall is reconstructive. You do not retrieve a stored copy of an evening in March; you rebuild it from fragments, and what you rebuild it with is your current understanding of the situation. Change the understanding and the rebuild changes with it.

So the same Tuesday that was, at the time, "they were quiet, they'd had a long week" gets rebuilt as "they were quiet, and now I think about it, that was the week it started". Nothing about Tuesday changed. The frame you are rebuilding it inside did.

Three effects, specifically:

Ordinary detail drops out. The parts of that evening that were unremarkable are the parts you have no reason to keep, so they go. What survives is the fragment that fits, and with the context gone it looks much more pointed than it was.

Timing collapses. Things that happened months apart migrate toward each other, because they are stored together now — under the suspicion rather than under their dates. A pattern appears that was never a pattern; it is four unconnected events filed in the same drawer.

Certainty grows without new input. Each time you go over the list it feels more solid, and it is easy to read that as the case getting stronger. It is not. Rehearsal makes a memory feel more confident without making it more accurate, and going over it is the one thing you are guaranteed to be doing a lot of.

Why this is not a reason to dismiss what you noticed

It would be an easy misreading of all this to conclude that your suspicion is a trick of memory and you should drop it. That is the opposite of the point.

Something got your attention, and the thing that got your attention happened before any of this reorganisation began. That original observation is usually the most reliable item you have, precisely because you made it before you were looking for anything.

What is unreliable is everything you have added since — the reinterpreted evenings, the retrospective pattern, the growing sense of how long it has been going on. Those arrived after the filter did.

The distinction worth holding is between what you noticed and what you have since made of it. The first is data. The second is a story built out of the first, and stories get more coherent with retelling whether or not they are true.

Keeping a record you can still trust

If you are going to write anything down — and you should, mostly to stop the rewriting — a few rules make it worth having in a month.

Write the observation, not the interpretation. "Came home at 9, said it was a work thing" is an entry. "Lied about being at work" is a conclusion you will not be able to unpick from the entry later. This is the same split the eighteen signs use, and it is the whole reason each one carries its ordinary explanation on the same line.

Date it when you write it, and note when it happened. Those are two different dates and the gap between them tells you how much reconstruction is in the entry. Something you wrote down the same evening is worth more than something you dated from memory six weeks later.

Write the ordinary explanation next to it, at the time. Not as a discipline in fairness — as a preservation measure. The innocent reading is the part that decays, and if you do not record it while it is still available to you, in a month you will be reading a list where every item has only one possible meaning.

Do not add to old entries. New thought, new entry, new date. Editing an old one is exactly the process this is meant to interrupt.

Keep it somewhere private, and keep it short. A document that becomes a project is a document you are now maintaining, and maintaining it is a form of the watching described in the piece on the waiting interval.

What this changes about the conversation

A written record made this way tends to be shorter than the one in your head, and much less damning. People often find that the list they were carrying had four items and the written one has two, because the other two turned out to be the same evening remembered twice, or an interpretation that never had an event under it.

That is not the record failing. That is the record doing its job.

It also gives you something specific to open with, which is the single thing that most improves how that conversation goes. "You've taken your phone out of the room to answer it four times in the last fortnight" is a sentence you can only say if you wrote it down. "You're always on your phone now" is what the same observation becomes after a month of unwritten rehearsal, and it is unanswerable — which means it starts an argument instead of a conversation.

A structured pass

The eighteen questions are, in effect, this exercise with the categories already drawn. They ask about specific observations rather than impressions, weight them, and hand back a count of what you have actually seen — with the ordinary explanation attached to each, and no verdict at the end, because eighteen self-reports cannot produce one honestly.