Why You Knew Better and Did It Anyway

We usually explain regret as knowing better but lacking discipline. This episode argues something more useful: in a hot state, the values on your internal scoreboard change.

I connect Epictetus with hot-cold empathy gaps, self-control replication research and implementation intentions. Then I give you a four-minute handover note: one specific if-then plan, written while you can still see clearly, for the version of you who will not realise their judgement has changed.

Sources and further reading:

- Epictetus, Discourses 1.28
- Seneca, Letters 76.34-35
- George Loewenstein, “Hot-Cold Empathy Gaps and Medical Decision Making” (2005)
- Hagger et al., “A Multilab Preregistered Replication of the Ego-Depletion Effect” (2016)
- Peter Gollwitzer and Paschal Sheeran, “Implementation Intentions and Goal Achievement” (2006)
- Keith Stanovich, Richard West and Maggie Toplak, work on myside bias and rational thinking

Free 7-Day Stoic Challenge: stoicchallenge.co

The Stoic Vault: stoicvault.com

Transcript

SPEAKER_00 0:18

The standard story of regret is that there's a gap between knowing and doing. You know you shouldn't send that message. You know you shouldn't take the drink. Open the tab at one AM, say the thing to your partner you've already said twice this month, and you know it, and yet you do it anyway. So the diagnosis is I have no discipline, I have a weak will, and the prescription is just to become more disciplined. But that story is wrong, not slightly wrong, but wrong in a way that sort of guarantees the fix won't work and never will work. Ovid gave us the line everyone quotes Medea, on the edge of the worst decision in Greek literature says she sees the better course and approves of it, then follows the worse. It's read as a confession of weakness and it's a dead end. But Epictetus reads it differently. He came back to Medea repeatedly in the discourses, and his verdict was cold. She did exactly what she thought was best. In that moment, revenge looked to her like the most valuable thing available. Her error wasn't in her will, it was in her pricing. This is one of the least fashionable stoic claims, but it's also one of the most useful for understanding our decisions and human behavior. Nobody assents or endorses to an impression they judge to be false. Nobody chooses the option that looks worse to them at the moment of choosing, which means every decision you've regretted made complete sense to the person making it. And that person was, of course, you. And you weren't lying to yourself, you were running a different valuation at the time. So the question isn't why you failed to act on what you knew, it's why for roughly ninety seconds you knew something else. George Lowenstein calls this the hot cold empathy gap. In a cool state you can accurately describe what you want and how much it's worth. In a hot state driven by anger or hunger or exhaustion, arousal or loneliness, boredom, the numbers change. Not your access to the numbers, but the actual numbers. George Ainsley's work on discounting shows the same thing but with more precision. We don't devalue future rewards on a smooth curve. We devalue them hyperbolically, which means the curves cross. At a distance the larger later reward is genuinely preferred, but up close the smaller sooner one wins. Both preferences are real and both are yars, and the reversal isn't a lapse in the second self's integrity, it's just basic arithmetic. Notice what this does to this story of discipline. There's no calm, rational you being overpowered by an impulsive you. There are two states of the same nervous system that price the world differently, and only one of them is in the room when the decision gets made. Below about thirty meters, breathing compressed air, divers develop nitrogen narcosis. The old rule of thumb calls it Martini's law. Every ten meters past the threshold is worth roughly one drink. Judgment degrades, reaction time slows, and fine motor control goes. But that isn't what makes it dangerous, actually. What makes it dangerous is euphoria. So narc divers often report feeling excellent, calm, but capable. Some become convinced they should go deeper to see what's down there. A few have been found having offered their regulator to a passing fish. The faculty that would detect the impairment is the impaired faculty. That's the part that transfers. When you're in a state that produces the regretted decision, you don't experience yourself as compromised. You experience yourself as clear headed, in fact, unusually clear headed. The two AM scroll feels like curiosity. The message you'll regret feels like honesty. The seventh drink feels like being present with your friends. From the inside the impairment presents itself as a form of insight. Divers didn't solve this with resolve or more discipline. They solved it with procedure. Planned maximum depths agreed on the surface, gauge checks at fixed intervals whether or not you feel you need one, a buddy whose job it is to notice what you can't. Ascend a few meters and it clears. Nobody expects a narc diver to reason their way out of narcosis because the reasoning is the thing that's bent. And here's something uncomfortable. Intelligence doesn't protect you, and there's a decent case that it might even hurt. David Perkins found that people with higher measured intelligence generated more arguments on informal reasoning problems, but not more arguments against their own position. The extra capacity went entirely into their own side. Keith Stanovich has spent decades on this, coining disrationalia for the specific condition of being smart and still reasoning badly, and showing that my side bias is largely uncorrelated with cognitive ability. Read that alongside Nisbet and Wilson's classic work on introspection where people confidently reported reasons for behavior that demonstrably weren't the causes, and a shape emerges. Your reasoning isn't the judge, it's the defense, barrister, and a good education just means you've hired a better one. Which is why the smartest people you know have the most sophisticated accounts of their worst habits. The account isn't a failure of intelligent, it's actually an achievement of it. The obvious response is to train willpower like a muscle. That model has had a bad decade. The ego depletion literature once close to settled, largely collapsed under the multilab replication. Hager and colleagues ran the effect across two dozen labs in 2016 and found something close to nothing. Whatever is happening when you fail at 11 PM, it probably isn't a depleted reservoir of self-control that push-ups and cold showers will deepen. The Stoics were better engineers than this, though the modern retelling flattens them into people who just tried harder. But if you read the actual practices, and they're almost all preemptive, Epictetus tells you to speak to an impression before you ascend to it, to make it wait while you examine it, which is not a strategy for the moment of temptation so much as a habit you install beforehand. Premeditum allorum, premeditation of future evils, is not gloominess, it's rehearsing the hot state while you're cool, which is the only condition under which you can plan for it with honesty. The move is to stop trying to win the argument at depth and start deciding on the surface. Peter Golwitz's work on implementation intentions is the cleanest modern version. Vague goals like I'll be more careful with my phone in the evening do very little. If then plans that specify the situation and the response produce reliable medium size effects across hundreds of studies. The plan works because it isn't asking your compromise self to deliberate. It's asking it to follow a decision that was already made by someone with better information at hand. So I recommend you do this once properly for one specific decision. There are four steps. One decision, purely factual, no commentary on your character, something like I texted at half eleven, not I always sabotage things. Two, reconstruct the state, not the choice. Bring to mind where you were, what time it was, what you were running on, tired, hungry, angry, bored, alone, maybe with a screen, three drinks in, just off a difficult phone call. This is depth reading. It matters more than the decision itself. Write one if then sentence. If it's after eleven and I'm alone with my phone, then it goes on the kitchen counter and I read something on paper. Specific trigger, specific action, not an intention, a rule. And four, make one change that doesn't require you to be at your best. Move the object. Delete the app. The test is whether it still works on your worst night. Because your worst night is the only night it's for. So we have name the last decision, you regret it, reconstruct the state, not the choice, write one if then sentence, make one change that doesn't require you to be at your best. You are not writing a promise, you're writing a note to someone who will be less capable than you are now, and who will not believe that they are.