Systems · June 15, 2026

The Pattern, Not the Diagnosis

The Pattern, Not the Diagnosis

Most people arrive at this question already losing. They type some version of "is my partner a narcissist" into a search bar, and the internet hands back a checklist. Grandiose. Lacks empathy. Needs admiration. Tick the boxes, get a verdict. The trouble is the verdict is unanswerable from where you are standing, and reaching for it points you the wrong way.

Consider what the checklist is asking you to do. It is asking you to diagnose the interior of a person whose central skill is appearing one way in public and another way at home. It is asking you to read someone accurately at the moment your own ability to read reality has been worn thin by the relationship you are trying to assess. You will not win that. Nobody wins that. It is the wrong instrument for the conditions.

So put the label down and look at the pattern instead. The pattern has one advantage the diagnosis never will. It shows up in your own week, and you have full access to that.

Two rules make the looking reliable.

Behaviour, not interior. You cannot see what someone feels. You can see what they do, repeatedly, across situations, over time. Read the actions and let them accumulate.

Pattern, not incident. Anyone, on a bad day, can produce a single instance of almost any of these. One eye-roll is not contempt. The two-thousandth is. The thing to watch is consistency, especially on the days when behaving that way cost them something and they did it anyway.

Now the part the checklists leave out. The strongest evidence is not in their traits. It is in what has happened to you.

Notice what you have stopped doing. Notice what your apologies have become. Notice that you pre-edit yourself before you speak, that you draft and delete and redraft, that you have started saying "maybe I overreacted" about events you remember with full clarity. None of that is a fact about them. All of it is a fact about the room, and the room is the thing you are trying to read.

Then apply the one test that settles more than any label. Repair.

You do not need to know what is inside the other person. You need to know what happens when you ask them to own something. Not once, when they were caught and motivated. Across months. Across topics. A person who can hear "this hurt me," sit in the discomfort of having caused it, and change the behaviour, is a person you may be in trouble with for ordinary reasons, but it is not this. A person who, even on their best day, reroutes that sentence until you are the one apologising, is showing you the engine. Repair is the diagnostic, not the personality.

This is also where you guard against getting it wrong. Some people who hurt you are not running this pattern. Someone in acute crisis becomes hard to live with, then comes back when the crisis lifts. Someone in active addiction blame-shifts and breaks promises, then becomes recognisable again when they get clean. Someone carrying untreated trauma can wound you while believing they are the wounded one, and still, when repair is on the table, feel the impact and change. The test is what stays after you strip the bad year out. If what is left is a person who can hear you, the crisis was the cause. If what is left is a person who cannot, even at their best, the crisis was the volume knob.

Hold the label loosely and the pattern matters more than ever. The label shimmers. On their good days, when the early version of them resurfaces, you will feel cruel for noticing the rest, and the name will seem too big. The pattern does not shimmer. The pattern is what your body knows when the car pulls into the drive. It is the thirty kept agreements and the three hundred broken ones, and the way the broken ones land near anything that asked them to be inconvenienced.

There is a quieter reason this reframe matters. The moment you stop asking "are they a narcissist" and start asking "what is this relationship doing to me, and what am I feeding it," you have traded a project you cannot finish for one you can begin this week. You cannot fix another person's interior. You can watch what happens to you over time, and you can stop supplying what is being taken.

You do not have to be certain to begin. The pattern was there before you had a word for it. The word is useful. It is not the evidence. What your body has been telling you for years, while your mind kept overriding it, is the evidence.

The newsletter

One email, when there is something worth sending.

A tool from one of the books, a sample chapter, an early copy before publication, word of a new edition. Every template and excerpt on this site arrives through the same door.

Unsubscribe in one click. Your address is never shared.

Related Articles

Stop making bad work faster

In 1990 Michael Hammer told companies to stop using new technology to make obsolete processes faster. We now have far better technology and appear to have preserved the mistake. Most bad corporate processes were never designed. They accumulated, one control and one spreadsheet at a time, until nobody could explain why any of it existed. The current generation of automation tools can strip enormous amounts of manual work out of arrangements like that, which is not the same as removing the reason the work exists. Drawing on the AICD's 2026 Director Sentiment Index and research from McKinsey, Deloitte and Microsoft, this piece sets out what to ask before approving the next business case promising thousands of hours saved, and why saved hours are a poor measure of anything on their own.

WANTED: TEN YEARS' EXPERIENCE IN SOMETHING INVENTED IN 2022

ChatGPT is three years and nine months old and the ads ask for ten years with it. AI job titles went from 264 to 822 in four years while total postings rose 6 percent. Companies now appoint officers for a thing their own engineers cannot define. Years of experience were a proxy for depth. The proxy failed because the pipes are three years old, and the form still rejects the honest answer.

The money went out. Nobody counted what came back.

Spending on AI is recorded precisely. What it returns, in most organisations, has never been written down. This piece sets out what the published research actually says, including where the widely quoted numbers have been compressed past what the studies support. It covers the Australian position after the mandatory guardrails were shelved, the automated decision-making disclosure obligation that takes effect on 10 December 2026, and the finding that thirty-seven per cent of Australian boards have audited what their organisation is running. It closes on a conflict of interest the category has not addressed and an eighteen-month mistake of the author's own.