Setting Boundaries in a Toxic Relationship

A boundary holds when it describes something you will do rather than something you are asking the other person to do, and when you actually do it the first time it is tested. Most boundaries that collapse were worded fine and enforced never, which is a different problem from not knowing what to say.

This is not a piece about identifying a toxic relationship or handling someone who manipulates you generally. It is about the mechanics of one boundary: how to state it so it is enforceable, and what to do the first time it is pushed against, because that moment decides whether it means anything from then on.

What a boundary actually is

A boundary is a statement about your own future action, not an instruction aimed at someone else’s behaviour. That distinction is the whole subject, because it is also the reason most boundaries are unenforceable as stated.

“Don’t talk to me like that” describes what you want from them. It gives you nothing to do if they do it anyway except say it again, louder. “I’m going to end the call if the shouting continues” describes something entirely within your own control. You do not need their cooperation to end a call, leave a room, or decline to discuss a topic. You need it to make them stop shouting, and that is the part nothing can promise.

A simple rewrite test: if the sentence only works with “or” attached (“stop doing this or I’m leaving”), it is closer to an ultimatum, a demand for a change in someone else with a threat behind it. If it works with “and I will” instead (“if this continues, I will leave”), it is a boundary, because the second half is yours to carry out regardless of what they decide.

How to word one so it survives being tested

Say it once, outside the heat of the moment, as a plain description of your own action. Not mid-argument, and not as a list of everything that has ever bothered you, because both invite a debate about fairness rather than a simple statement of fact.

Three things make the wording durable. It names a specific trigger rather than a mood (“when the conversation turns to my weight” rather than “when you’re being awful”). It names an action you can actually take alone, not one that requires their agreement. And it leaves out the character judgement, because “I need us to stop discussing my weight” survives a bad reaction much better than “you’re obsessed with criticising my body”, even when the second sentence is also true. The first is a statement they can accept without losing face; the second is an accusation, and accusations get defended against rather than heard.

Assertive communication of this kind is not just etiquette. A review of the assertiveness training literature found consistent evidence linking assertiveness training to gains in self-esteem and to broader improvements across the clinical problems it has been used for, describing it as a treatment approach that had fallen out of fashion despite the evidence behind it. [speed-2018-assertiveness] Separately, research examining assertiveness alongside self-esteem and relationship satisfaction found self-esteem to be the stronger predictor of how satisfied people were in their relationships, with assertiveness and self-esteem themselves correlated. [moss-2021-assertiveness] Read together, the practical takeaway is not that saying the sentence fixes the relationship. It is that being able to say it, and mean it, tracks with feeling better about yourself regardless of how the other person responds.

The one thing that actually breaks a boundary

Almost every boundary that fails, fails at the same point: it gets crossed, and nothing happens. Not the wording, not the tone, not even the timing. The gap between the sentence and the action.

The mechanism is ordinary learning rather than anything specific to any one relationship. A rule that is announced but never enforced is, functionally, not a rule, and both people end up knowing it, even if neither says so. The person who crossed it learns the real cost was zero. The person who stated it learns that stating things does not change anything, which is exactly the belief that stops people from trying again.

This is why a boundary enforced calmly and consistently beats a boundary re-explained with more feeling every time it is broken. Volume and repetition are not what make a limit real. The action is.

What to do the first time it gets tested

Carry out the exact consequence you named, with no extra commentary, no re-litigating, and no exception for how good their explanation is. This is the single highest-leverage moment in the entire process, because it is the moment that decides whether every future statement of the boundary means anything.

If you said you would end the call, end the call. If you said you would leave the room, leave it. Do not narrate the decision in the moment (“I’m doing this because you know exactly what you’re doing”), because that reopens the conversation the consequence was supposed to close. Act, and explain later if you want to, once things are calm, if at all.

Expect the first enforcement to produce a reaction, often a bigger one than the original behaviour did. A limit that used to be free to cross is now costing something, and testing whether it is real is a very ordinary response, not proof that the boundary was wrong. What matters is what happens by the third or fourth time: a reasonable person adjusts, and the crossings become less frequent because the cost is now real. If the crossings continue at the same rate while the reaction to the consequence keeps escalating, that is worth paying close attention to in its own right.

When a boundary is being used against you

Some people respond to a stated boundary by finding the exact edge of it and standing on that edge, or by punishing you afterwards in a way that has nothing to do with the original issue. That response is information. A boundary is a tool for a relationship where both people are, underneath the friction, willing to adjust. It is not a tool for changing someone who is committed to crossing whatever line you draw, and treating it as one just gives them a fresh script to work with.

If you recognise that pattern, the more relevant approach is not a better-worded boundary but reducing what you offer someone whose behaviour runs on getting a reaction from you, which is what the grey rock method is for. The two are not the same tool: a boundary is something you hold in a relationship you are staying in and want to work, while grey rock is a way of getting through contact you cannot avoid with someone who is not going to change. Our toxic relationships guide covers how to tell which situation you are actually in.

When to seek help

Speak to a doctor or a therapist if you find yourself unable to state a limit without extensive rehearsal and dread, if boundaries you set are met with a punishing or escalating reaction more than once, or if you have started avoiding entire topics rather than risk the conversation. Any of those is a reasonable, common reason to get support in its own right.

Go sooner if a boundary being tested has ever come with a threat to your safety, your access to money, or your contact with other people. That is a different problem from a communication skill, and it needs a different kind of help. Contact your local emergency services or a crisis helpline if you feel unsafe.

How MyFreud can help

Working out whether a boundary is actually holding is hard to judge from inside a single bad evening. Daily tracking makes the pattern visible over a few weeks rather than one argument, which is usually the difference between “this isn’t working” and “this worked three times out of four and the one failure had a clear cause”. Live coaching sessions are a useful place to rehearse the exact sentence before you need it, since the wording is easier to get right when you are not already upset.

Download MyFreud and start today: App Store or Google Play.