Here’s why I called it The Exit Code. In programming, an exit code is what a system returns when it terminates. 0 means clean exit, success. Non-zero means error, crash. The metaphor is literal. A narcissistic system isn’t a person with problems. It’s an operating system with code. Logic that repeats. Programming that executes. The code looks like this: - Extract validation from others (the system needs external input for its own sense of purpose). - Perform constantly (the operating system is the only thing that exists; beneath it is void). - Control through isolation (remove external references so the code is the only logic running). - Repeat indefinitely (the program cannot change itself; it can only execute what it was programmed to do). Here’s what’s observable: the system is relentless. It repeats. It escalates. It persists. But not because it’s powerful. Because you’re still executing it. That’s the key distinction. The person isn’t holding you in the system through force or manipulation or even trauma bonding, though all of those are present. You’re held in the system because you’re still running their code. - You’re still seeking validation (executing their extraction protocol). - You’re still monitoring their mood (executing their control protocol). - You’re still trying to prove yourself (executing their performance protocol). The system runs as long as you’re executing its program. And here’s what’s brilliant about the framework: the exit doesn’t require healing. Doesn’t require the person to change. Doesn’t require a dramatic moment or external validation. The exit requires literacy. Once you can read the code, i.e. once you understand how the system operates, what it extracts, how it controls, why it repeats, you stop executing it. Not because it’s suddenly easy. But because you’re literate to it. You can’t unsee code once you’ve read it. And you can’t keep running a program once you understand what it’s doing. That understanding is the exit. That literacy is the code that terminates the system.

Here's the mistake, and most people make it for years. You assume the other person operates from the same moral ground you do. So when you're wronged, you explain. You present evidence. You describe the pain, expecting that once they see it, they'll regret it. I can't tell you what's inside them. I can tell you what the record shows: whether any explanation ever led to a revised account, an acknowledgement, a change in behaviour. If the answer is no, every time, the ground you're appealing to isn't there, whatever the reason.

What actually works against bait is unremarkable continuation. Not cold, not restrained, not pointedly silent. Just ordinary. 'Mm.' 'Possibly.' 'Anyway.' And on to the next thing. Expect escalation after the first few, because a mechanism that stops paying is tried harder before it's abandoned. That escalation is the sign it's registered. Full piece on Substack. substack.com/@theexitcode/no…

Some remarks exist for one reason: to make you do something, and to be seen doing it. They're small, aimed at something you're known to be tender about, and often delivered in company. They're deniable by construction: 'I was joking,' 'you're reading into it.' And they repeat until they land. I've written about how to read them, why they tend to arrive just before you raise something, and what to do that doesn't deliver the catch. substack.com/@theexitcode/no…

A provocation isn't an attack. It's a fishing expedition, and the catch is your reaction. substack.com/@theexitcode/no…

Coercive control is often described as if it needs locked doors and raised voices. Often it doesn't. It needs something quieter: a reliable consequence every time you act on your own behalf. You speak up, and the reply is hostility. You set a limit, and you're told you're the one being controlling. You raise a hurt, and within minutes you're apologising for having raised it. Each time, the message is the same: standing up for yourself costs more than staying quiet. After enough repetitions, you don't need to be told anymore. You start pre-empting. You soften your requests, drop your plans, keep your opinions to yourself, not because you've changed your mind, but because you've learned the price. That's the mechanism. Your own fear becomes the enforcement. If you recognise this, the most important thing is not to try to fix it alone, inside the relationship, by finding the right words. The right words don't exist when the point is your silence. What helps is contact with people who can see the pattern with you, and a safe plan for what comes next. A domestic abuse service can help with that.

When every attempt to speak up gets reframed as an attack, the point isn't the argument. It's to make you afraid of your own voice.

Accusations can tell you more about the person making them than about the person they're aimed at. That's not a rule, and it doesn't prove anything on its own. But when the same behaviour keeps appearing in the accuser's own life, and is described in the other person's, it's worth recording. Dates, examples, and what happened after.

Sometimes the most useful evidence isn't what someone says about another person. It's what they do themselves. If they accuse someone of lying, do they tell the truth when it's inconvenient? If they describe someone as unreliable, do they keep their own commitments? If they call someone manipulative, do they make decisions openly, or through pressure and secrecy? You don't need to judge their character. You only need to compare the description with the conduct. The gap is where the information is.

People often accuse others of exactly what they're doing. Pay attention to the behaviour, not the accusation.

How do you know if a friend is loyal, or quietly reporting back to the person you left? You don't read their mind. You watch what they do with what you share. Say one small, harmless thing to them, and see whether it resurfaces somewhere else. Say one clear sentence about the boundary you need, and see whether they respect it. Stop sharing the things that matter, and see whether they still ask for them. Loyalty is what happens after you've tested it.

There's a particular kind of person who thinks they're being subtle. They manage their image carefully, choose their words, present a version of themselves that looks reasonable. But behaviour doesn't stay in the script. It shows up in how they treat the waiter when they think it doesn't count, in what they say about someone who left, in the way they respond when they don't get what they wanted. Those moments repeat, and repetition is hard to hide. What makes it useful to you is that you don't have to decode what they're thinking. You only have to notice what keeps happening, across different situations and over time. The gap between the image they present and the behaviour you see is the information. It tells you more about how this person operates than anything they say about themselves.

Look at how they treat people who can't do anything for them: service staff, people who've left, people who disagreed. Then look at how they treat people who are useful to them. If the two columns look different, the difference is information. It tells you what the warmth was for.

People in these patterns are often tested before they realise it. Not in a dramatic way. In a series of small moments that each ask a question. Will you say yes to this slightly unreasonable request? Will you let this joke go? Will you adjust when the plan changes at the last minute? Will you apologise first, even when you weren't wrong? Every answer teaches the pattern something about you: how much you'll give, where your boundary bends, how long it takes you to speak up. The early tests are cheap for the pattern and expensive for you, because they set the terms for everything that follows. You don't have to win these. You just have to notice them, and answer the way you'd answer if nobody were measuring. Say no to the small thing. Hold the boundary at the first push. Let the test find out that the answer is yours.

Watch out for people who make you the joke when others are around. It never looks like cruelty. It looks like humour, teasing, a comment about something you did, delivered with a smile so you'd look oversensitive if you objected. In private they might apologise. In public it happens again. Why? Because an audience makes the correction visible and the apology private. You end up being the one who's 'too serious,' while they stay the funny one. These people don't need to raise their voice to make you smaller. They just need a room full of people laughing. Real friends don't use a crowd to score points. If they tease you, you're laughing too, and they stop the moment it lands wrong. Pay attention to how you feel on the way home from an evening with them. That feeling is accurate, even if nobody said anything you can quote.