Podcast player

0:00The word is overused — here's the narrow version

Welcome back. Today's episode is about a word that has been stretched until it barely holds anything: gaslighting. It gets used for lying, for arguing, for someone having a different opinion, for someone simply being wrong about a fact in a normal, boring way. I want to shrink the word back down to something usable, because a word that means everything cannot help you identify anything. If everything is gaslighting, then nothing is, and you lose the one tool that would actually have helped you.

Here is the narrow version I use, and I am going to say it slowly because it is doing all the work for the rest of the episode. Gaslighting is a repeated pattern in which your account of reality is treated as a symptom rather than as information. Not a disagreement about what happened — disagreements about memory are ordinary and human, and I will spend real time on that distinction because it is the one listeners write in about most. The distinguishing feature is the reframe: the conversation stops being about the event and becomes about your perception being defective.

Two people can remember the same dinner differently and both be honest. Maybe one of them was more anxious that night and remembers the tone as sharper than it was. Maybe the other was distracted and genuinely missed a comment that landed hard. That is memory doing what memory does — reconstructing, not replaying. What we are talking about today is a different animal entirely. It is when 'I remember it differently' is replaced with 'there is something wrong with the way you remember things,' consistently, over time, in a direction that always benefits the same person.

That last clause matters, so I will say it again: in a direction that always benefits the same person. A single reframe is nothing. Everyone has said 'that's not what I meant' or 'I don't remember it like that' at some point, including me, including you. A hundred reframes that all land on 'you are unreliable and I am not' is a pattern with a shape, and shape is what we are hunting for in this episode, not individual sentences taken out of context.

A listener wrote in a few months ago and said something like, 'I think I might be gaslighting my partner, because I genuinely do remember things differently from him and I get defensive when he brings it up.' I want to answer that directly, right at the top, because it is a generous and honest question and it deserves a real answer rather than a slogan. Getting defensive when someone brings up a hard memory is not gaslighting. Genuinely misremembering things is not gaslighting. What would make it gaslighting is if, every single time, the disagreement conveniently resolves in your favour, and if your standard response to being challenged is not 'let's figure out what happened' but 'the problem is that you're the kind of person who can't be trusted to know what happened.' If you're worried enough to ask the question, that alone is a decent sign you are not the pattern I am describing today — the pattern rarely comes with self-doubt attached.

I also want to head off the other common misuse, which is using this word as a weapon in ordinary arguments. If your partner disagrees with your memory of one conversation and you call it gaslighting, you have just done the thing you are trying to warn against — you have converted a disagreement about an event into a verdict about their character. So the whole point of narrowing the definition is not to hand you a bigger stick. It is to hand you a more accurate one, and accuracy, as you'll hear a lot today, is the entire game.

So here is the shape we are going to build across this episode. First, six specific phrases and exactly what job each one is doing underneath the surface, with the kind of dialogue you'd actually hear in a kitchen at eleven at night. Then why the instinct to argue the facts, which feels like the obvious move, is almost always the wrong one, and what happens to a person's brain over months of trying. Then a practical, low-effort way to keep a record without turning your relationship into a courtroom. And finally, the line — because there is a real line — where this stops being a communication problem worth working on and becomes a safety question worth planning around.

One more thing before we start, because I think it matters for how you listen to the rest of this. None of what follows is a diagnosis, and I'm not going to hand you a checklist that tells you your relationship is definitely this thing. What I can hand you is a set of distinctions sharp enough that you can hold them up against your own week and see what fits and what doesn't. That is the most honest thing an episode like this can do, and it's more useful than a label anyway, because labels don't tell you what to do next — patterns do.

4:34Six phrases and what they're doing

Let's get concrete, because this is the part that actually helps. Six sentences, ordinary enough that you've probably heard at least one of them this year in some form, and underneath each one, what it is actually doing to the conversation.

One: 'You're too sensitive.' This is a category error performed on purpose. You raised a specific event — 'when you said that in front of my sister, it stung' — and the reply answers with a permanent trait about you rather than the event. Notice the structure: you cannot respond to 'you're too sensitive' without talking about yourself, your history, your track record of reacting to things. Which means the event you actually raised has been quietly deleted from the conversation without anyone announcing that it happened. Picture it in real time. You say, 'that comment about my job in front of your friends wasn't kind.' The reply is, 'God, you're so sensitive, I can't say anything around you.' Now you're explaining your sensitivity instead of discussing the comment. What you could say instead, if you want to hold the line without escalating, is something flat and repeatable: 'This isn't about how sensitive I am generally. It's about that one comment. Can we talk about that comment.' You are allowed to just decline the detour.

Two: 'That never happened.' Flat denial, no elaboration, sometimes not even much conviction behind it — that's part of what makes it unsettling, because it's delivered with the same tone as someone stating the time. Its power is not persuasion. You know what you saw, and on some level so do they. Its power is exhaustion. It costs one second to say and can cost you twenty minutes of increasingly desperate detail to contest, describing the room, the exact words, who else was there, and the arithmetic of that eventually teaches you not to raise things at all, because raising things has become expensive and getting your memory validated has stopped being one of the possible outcomes. If you hear this one, a useful internal move is to notice that you are not actually required to win the factual argument in the moment. You can say, 'I remember it differently, and I'm not going to argue about it right now,' and then, later, write it down — we'll get to exactly how in a few minutes.

Three: 'You're remembering it wrong, you were drunk, you were tired, you were upset.' This one is more sophisticated than the flat denial because it attaches the doubt to a plausible-sounding explanation, which makes it much harder to dismiss even when you know you weren't especially drunk, or especially tired, or especially anything. It also builds a precedent that gets reused. Once 'you were upset that night' has been established once, it can be applied retroactively to any inconvenient memory from that entire evening, or that entire month if the person is patient about it. Your memory now carries an asterisk that someone else controls the footnote for. The tell here is usually that the explanation only ever gets deployed against memories that are inconvenient for the other person, never against memories that happen to support their version — if you were 'too tired to remember properly' on Tuesday but somehow have perfect, trusted recall on Thursday when it's convenient, that asymmetry is the actual information, more than either individual claim.

Four: 'I only did that because you...' This is sequence reversal, and it's one of the most common moves I hear about from listeners, maybe because it's so ordinary-sounding that it slides past unnoticed. The response becomes recast as the cause. You raise, 'you slammed the door,' and the answer is 'I only slammed the door because you were nagging me,' and now the conversation is about whether you were nagging, a much more debatable, much more subjective question, while the door-slamming — the actual concrete behaviour you raised — quietly leaves the room unexamined. The two events might both be true. Maybe there was some nagging, maybe there was a slammed door. But notice that only one of them gets investigated, and it's never the one you brought up. A way to keep both events on the table at once is to say, 'we can talk about the lead-up separately — right now I want to stay on the door,' and then actually mean it, and actually return to the lead-up afterward if it's worth discussing, because the goal isn't to dodge accountability either. The goal is sequencing, not erasure.

Five: 'Everyone thinks you're overreacting.' Imported consensus. This one is particularly effective because it cannot be checked in the moment — you are not going to phone the friend group at ten at night to verify — and trying to check it, or push back on it, makes you look exactly like the anxious, overreacting person the sentence describes, which is a neat trap. This is also, worth naming, how what people sometimes call flying monkeys get recruited: the same sentence, or something close to it, is very often being said in the other direction, to those same people, about you, well before you ever hear the phrase yourself. If you ever do get the chance to actually speak to one of these supposedly unanimous witnesses, you often discover the consensus was one person's opinion, reported as a crowd. A reasonable response in the moment is simply, 'I'd need to hear that from them directly, not as a report,' and then not chase it further that night.

Six: 'I never said that. You always do this.' Two moves stacked in one sentence — a denial of the specific plus a generalisation about the whole relationship. The 'always' is doing the heavy lifting here, far more than the denial itself. It converts one contested, specific, checkable event into an already-documented, uncheckable habit, so that even if you somehow proved the specific instance, you'd still be left arguing against a vague permanent character trait that was never actually established anywhere except in that sentence. If you hear 'you always,' a useful, calm question is simply, 'give me three times.' Not to win an argument, but because the answer to that question is genuinely informative — either they can, which is worth knowing, or they can't, which is also worth knowing.

Read together, all six of these do the same underlying job, even though they sound completely different on the surface. They move the subject from what happened to who you are. That's the pattern to listen for underneath the specific words, because the specific words will vary from household to household, from relationship to relationship, but that particular move — event to character, always in one direction — is remarkably consistent once you know to listen for it.

It is worth noticing, too, that these six phrases rarely arrive alone. In a real evening, you might hear a blend — a flat denial followed by a generalisation, followed by imported consensus once a second person gets pulled in over text the next morning. Untangling a real conversation means being able to name each thread separately, even while they are braided together in the moment, because naming them separately is what lets you respond to the actual move being made rather than to the tone, or the volume, or how tired you already are by the third phrase in.

A listener asked whether it counts if the person doing this doesn't seem to realise they're doing it — if it feels more like a reflex than a strategy. Genuinely, for the purposes of this episode, the intent behind it matters less than you'd think, and I say that carefully because I know it's an uncomfortable thing to hear. Whether someone is doing this consciously and deliberately, or has simply learned it as their automatic way of handling conflict because it worked for them before and nobody ever pushed back, the effect on you is close to identical either way. Your job here isn't to diagnose their intentions — you can't, and it's not really any of these six sentences' business whether they were said on purpose. Your job is to notice the pattern and respond to the pattern, which is exactly what the rest of this episode is about.

One more small but useful discrimination before we move on. Sometimes people say one of these six phrases exactly once, in the middle of an otherwise fair fight, because they're tired or embarrassed or cornered, and then later, once things have cooled down, they walk it back unprompted — 'sorry, that wasn't fair, you're not too sensitive, I was just annoyed.' That walking-back matters enormously. It's the difference between a bad sentence and a load-bearing strategy. The pattern we're describing today doesn't walk itself back. It gets repeated, refined, and defended.

12:41Why arguing the facts backfires

The natural instinct, when someone denies a thing you know happened, is to prove it. Better evidence. Better recall. A more airtight, more detailed, more precise sentence that surely this time will be undeniable. I want to explain, gently and without judging anyone who has tried this — because I think almost everyone tries this first, it's the obvious move — why it almost never actually works the way you hope it will.

It does not work because you are not actually in a disagreement about facts, even though it feels exactly like one. You are in a disagreement about who gets to certify reality in this relationship, and that is a completely different kind of dispute, one that evidence is not built to settle. Bring a receipt — a screenshot, a text thread, a photo with a timestamp — and instead of the acknowledgement you're hoping for, you sometimes get something like, 'you kept a receipt? that is exactly what I mean about you, who saves screenshots of their own partner.' Notice what just happened there. The frame didn't just survive the proof, it absorbed the proof and converted it into fresh evidence of your character. You are now defending your habit of keeping records instead of discussing the thing the record proved.

I want to walk through a realistic exchange so this isn't abstract. Say you bring up that your partner promised to pick up your child from football practice and then didn't, and you have the text where they said 'yes, I've got it, don't worry.' You show the text. A healthy response, even a defensive one, looks something like, 'you're right, I forgot, I'm sorry, that's on me.' The pattern we're describing today looks more like, 'okay, I said I'd try, I never said I definitely would, you always twist my words into promises,' followed by, if you push further, 'this is exactly why I don't like texting you anything, you use it as ammunition later.' Watch the sequence: the promise gets downgraded to a maybe, your accurate quotation of their own words gets relabelled as twisting, and finally your habit of having evidence at all gets reframed as a character flaw — being someone who 'uses texts as ammunition.' By the end of that exchange, you are apologising for having a phone.

There is also a cost here that is easy to miss because it's slow and internal rather than dramatic. Every time an argument like this plays out, it trains your brain to check itself before it speaks. At first that's conscious — you deliberately pause before raising something, wondering if it's really worth the fight. Over months, that becomes an automatic habit of pre-emptive self-correction, where you start editing your own perception before it even reaches your mouth, sometimes before you've fully finished forming the thought. You'll notice yourself thinking, mid-sentence, 'well, maybe I am misremembering, maybe I am overreacting,' not because new evidence arrived, but because that thought has become a reflex. That, more than any individual lie, is the real injury of this pattern. Not the specific false claim about the football practice. The habit of doubting yourself first, automatically, before anyone else even has to.

A listener wrote in with a version of this that I think is worth reading almost word for word, because I hear a shape like it a lot: 'I used to be someone who trusted my gut completely. Now I run everything past my sister before I decide if I'm allowed to be upset. I don't know when that started.' That not knowing when it started is itself extremely characteristic. This isn't a change that happens in one dramatic conversation. It happens in about two hundred small ones, each so minor on its own that none of them would be worth an episode by itself, and the cumulative effect only becomes visible in hindsight, usually when someone else — a friend, a sibling, sometimes a podcast — points at the pattern from outside it.

So the goal has to change, and this is really the hinge the whole episode turns on. You do not win the argument about what happened. I'll say that plainly because it's counterintuitive and a little bit deflating the first time you hear it: winning is usually not on the table, not because you're wrong, but because the argument was never actually about the facts to begin with. What you can do instead is keep the record of what happened. Those are two entirely different projects with two entirely different success conditions, and only one of them is realistically achievable inside a relationship where this pattern is active. Winning requires the other person to concede, which is precisely the thing the pattern is built to prevent. Keeping a record requires only you, a pen, and two minutes, which is why we're going to spend the next chapter on exactly how to do that well.

Before we get there, one common objection, because I know it's on some listeners' minds right now: 'but sometimes I really do misremember things, so how do I know I'm not the unreliable one?' It's a fair question and I don't want to wave it away. Everybody misremembers things sometimes — the exact words, the exact order, whether it was Tuesday or Wednesday. The test isn't whether your memory is perfect, because nobody's is. The test is whether being wrong occasionally, in ordinary human ways, is being treated as ordinary and human by the people around you, or whether it's being filed away and used later as proof that you generally can't be trusted. Being fallible is normal. Having your fallibility weaponised is not the same thing, and the difference usually shows up clearly once you start keeping notes, which is exactly what we're about to cover.

It's also worth saying that arguing the facts isn't pointless because facts don't matter — they matter enormously, which is precisely why we're about to talk about protecting them properly, just not by relitigating them out loud in the heat of an argument where the frame itself is being contested. Facts kept quietly and accurately outlast facts fought over loudly and repeatedly. That's the whole strategic shift this chapter is asking you to make.

It also helps to notice what a healthy disagreement about a fact actually sounds like, so the contrast is easier to spot in real time. It sounds like two people comparing notes: 'I thought you said Friday.' 'No, I think I said Saturday, but I can see how it sounded like Friday.' 'Okay, that tracks, sorry.' Notice that both people stay roughly the same size in that exchange, and the fact gets settled without either person's character coming under review. That ordinary, forgettable little exchange is the thing this chapter is contrasting against, and it is worth holding in mind precisely because it is so unremarkable.

19:19Keeping a record without building a case

Here is the version I recommend, and I want to be precise about it, because the constraints around this practice matter every bit as much as the practice itself — get the constraints wrong and you turn a useful habit into something that actively makes things worse for you.

Two lines. That's the target, not a minimum you build up from. Date, and what was said, in their words if you can manage it, because their own words are the most useful data you can capture. That is the whole entry. No analysis, no interpretation, no adjectives describing how upsetting it was, no predictions about what it means for the relationship. Something like: 'Tuesday: said he never agreed to the Friday plan. I have the text where he did.' Done. Close the notebook, put the phone down, move on with your evening. The temptation is always to write more — to explain the context, to vent a little, to make the entry do emotional work as well as factual work. Resist that. The moment it becomes an outlet for feelings rather than a record of events, it starts to blur, and blur is exactly what you're trying to avoid.

Here's a second worked example so the format is completely clear. 'Saturday: said I imagined the argument about money on the 3rd. I remember specific details — we were in the kitchen, it was raining, he mentioned the credit card statement.' Notice that even the second entry stays factual: what was said, what you specifically remember, no adjectives about how it made you feel, no conclusions about what kind of person does this. The entries should be boring to read. Boring is the whole point — boring is what makes them credible, including to yourself, later, when you're tired and start wondering if you're the one who's losing the plot.

Keep it somewhere that is genuinely yours and reasonably private — a notes app with its own lock separate from your phone's general passcode, a small paper notebook that physically lives outside the house, at work, or at a friend's, or an email account only you access, sent to yourself as drafts. If your device is monitored or shared, or if you have any reason to think it might be checked, choose paper, and choose paper you can keep somewhere genuinely outside the home. This isn't paranoia — it's simply matching your method of record-keeping to your actual living situation, the same way you'd choose a different lock depending on the neighbourhood.

Now the constraint, and honestly this might be the single most important sentence in this episode, so I'm going to slow right down for it. This is not a case file. You are not building an argument to present to the other person, to your family, to anyone, because presenting it very rarely produces the acknowledgement you're hoping for, and the attempt to present it usually costs you considerably more than the record itself is worth. I have heard from listeners who assembled months of careful notes, sat their partner down, laid it all out point by point — and the response was some version of 'you've been secretly building a file on me this whole time, that is genuinely disturbing, that says everything about who's really the problem here.' The record got absorbed into the pattern, exactly the way the receipt did in the previous chapter. So the record exists for exactly one reader, and that reader is you, six months or a year from now, at a moment when your own memory has been sanded down by repetition and doubt and you need to check it against something more solid than your current, tired, second-guessing self.

People are consistently, almost startlingly astonished by what re-reading actually does. Individually, each entry looks small, forgettable, arguable — 'maybe I did overreact about that one thing on that one Tuesday.' But twenty of them, read one after another in a single sitting, in a pattern of language that keeps recurring, in a direction that keeps benefiting the same person, do not look small at all. They look like exactly what they are: a pattern. And pattern is the one thing denial genuinely cannot touch, because denial works entry by entry, moment by moment, and pattern only becomes visible when many entries are set down close to the event and then looked at together, later, with some distance.

A practical note on timing that a few listeners have asked about: write the entry as close to the event as you reasonably can, ideally the same day, because memory degrades fast and the whole value of the record depends on it capturing what actually happened rather than what you've since talked yourself into believing happened. If you can't do it the same day, do it the next morning, before the day's other conversations have had a chance to reshape your memory of yesterday's. Waiting a week defeats most of the purpose.

Re-read it once a month — mark a date, put a small recurring reminder somewhere private, whatever works for your life. Not daily; checking it daily tips into rumination, where you're mining the past for grievance rather than orienting yourself in the present, and that serves no one, least of all you. Monthly is orientation rather than obsession — a chance to step back and ask, calmly, 'does this still look like an occasional, ordinary human disagreement, or does it look like a pattern with a direction,' and then to act, or not act, based on an honest answer to that question rather than on whatever mood you happen to be in on any given Tuesday.

One more practical question that comes up often: what do you actually do with the record once a pattern is unmistakably confirmed? The honest answer is that the record itself doesn't obligate you to any particular next step immediately — it's not a summons, it's a tool. Some people use it to finally trust their own instincts enough to raise a boundary clearly and calmly. Some use it to decide it's time to talk to a therapist, ideally one experienced with this kind of relational pattern, and bring the record along as something concrete to work from rather than trying to reconstruct months of fog from memory in the session itself. Some, especially if things have escalated, use it as part of a wider decision about leaving, and if that's a live question for you, please treat this episode as a starting point rather than a complete plan, and lean on people and resources equipped for that particular conversation, which a podcast, however careful, is not.

And a small caution in the other direction too, because balance matters here: a record like this is a mirror, not a verdict. It's entirely possible to keep two months of notes and, on rereading them, decide honestly that most of them describe ordinary friction rather than anything more troubling — that's a genuinely useful outcome as well, and not a failure of the exercise. The record is there to tell you the truth either way, not to confirm a fear you walked in with.

A few listeners have asked whether it is dishonest, in some sense, to keep a private record the other person does not know about. I do not think so, and here is why: a diary of your own experience is not surveillance, and it is not deception, because you are not altering anyone else's reality or hiding anything you would be obligated to disclose. You are simply refusing to let your own memory of your own life be the only copy in existence. People keep journals, medical notes, financial records, and calendars without anyone calling that secretive. This is no different in kind, only in subject.

26:24When it stops being a communication problem

I want to end somewhere more serious than where we started, because I think it would be a disservice to spend this whole episode on phrasing and note-taking without naming clearly the point at which this stops being a communication problem you can work on together and becomes a safety question you need to plan around instead. The line is clearer, and closer, than most people expect going in.

It crosses when naming things gets punished. When raising an issue — even calmly, even gently, even at the best possible moment you could choose — reliably produces a consequence rather than a conversation. A silence that stretches on for days without explanation. A threat delivered with a laugh so it can be denied later as 'just a joke' if you react to it. A sudden financial squeeze, plans cancelled, money moved, timed to arrive shortly after you spoke up. Anything at all that consistently arrives in the aftermath of you telling the truth about your own experience. If accuracy has become dangerous inside your own home, you are no longer looking at a dialogue problem that better phrasing or a calmer tone or a well-kept notebook is going to solve. You are looking at an escalation pattern that needs to be planned around deliberately, with outside help, not gently negotiated with from the inside.

It also crosses a meaningful line when the doubt has gone properly systemic — when you find you cannot decide whether you're even allowed to be upset about something without checking it against a friend first, running the whole scenario past them like a court exhibit before you'll trust your own reaction to it. If that's where you are, please hear this clearly: that is not weakness, and it's not a character flaw of yours to fix through sheer willpower. It is what sustained reframing does to a person, predictably and mechanically, over enough time, to more or less anyone it happens to. And just as importantly, it is reversible. With distance, and with support, the part of you that used to trust its own read on a room does come back. I've heard that from enough listeners, describing it in their own words months or years down the line, to say it with real confidence rather than as a hopeful platitude.

A listener once asked me how to tell the difference between 'I need to leave' and 'I need to have one more honest conversation first,' and I don't think there's a formula that answers that safely for every situation — anyone offering you a clean formula for that particular question is oversimplifying something that genuinely varies. What I can offer is this: if raising things has become something you have to strategise around, time carefully, or brace yourself for, rather than something you can simply do, that shift itself is significant information, regardless of which specific decision you end up making about the relationship as a whole.

The show notes for this episode have a phrase-by-phrase reference guide covering the six patterns we discussed today plus several more we didn't have time for, a longer set of dialogue examples for practising responses out loud before you need them in real time, and regional crisis contacts for anyone who needs them. If any part of this episode described your actual week, this week, rather than a general idea you were turning over intellectually, please open the crisis page rather than the guides, and please reach out to someone in your life you trust, today rather than after this episode finishes.

To close where the middle of this episode pointed us: you do not win the argument about what happened, and that's genuinely all right, because winning was never really the useful goal to begin with. You keep the record. Two lines a day is enough — it takes less time than reading this paragraph took. If you do only one thing from today's episode, let it be that one small, boring, private habit, because boring and private is exactly what makes it durable. See you next episode.

If none of this describes your relationship, that is genuinely good news, and I would rather you finish this episode reassured than newly suspicious of an ordinary partner who occasionally misremembers a Tuesday. This whole hour has been about sharpening a distinction, not about handing you a reason to go looking for a monster where there is only an imperfect, tired, occasionally wrong human being, which describes basically everyone, including whoever is listening to this sentence right now.