From the below video: If deceit is fundamental to animal communication, then there must be strong selection to spot deception. And this in turn ought to select for a degree of self-deception—rendering some facts and motives unconscious so as not to betray, through subtle signs of self-knowledge, the deception being practiced.
The conventional view that natural selection favors nervous systems which produce ever more accurate images of the world is a very naive view of mental evolution. What Trivers is saying is that the brain is specifically designed to self-deceive when useful, and that it is foolish to assume evolution would select for accuracy per se.
I think it’s a beautiful, intriguing, and horrifying idea. We see this theme in science fiction. Battlestar Galactica had the Cylons, who didn’t know they were Cylons so they could more effectively blend in with humans. They fell in love, and it was horrifying to learn they were Cylons all along. They were programmed to deceive themselves in order to deceive others.
Total Recall has a similar premise, with Arnold Schwarzenegger tricked into thinking he’s a rebel when he’s actually working for the bad guys. And in Seinfeld, George tells Jerry, who’s worried about passing a polygraph: “It’s not a lie if you believe it.” The point is that we’re better liars when we believe the lie.
Is there evidence for this? Unfortunately, it’s spotty. One prediction of Trivers’ theory is that we should be very good at detecting lies, and less good when the person believes the lie. The evolutionary story is that deceivers and the deceived are in an arms race: as deceivers get better, victims get better at spotting deception—cues like sweat, shifty eyes, subtle signs of self-knowledge. So, because it became so hard to lie without leaking cues, a new strategy emerged: lie to oneself first, then to others.
This predicts that people should be good at spotting conscious lies. But evidence doesn’t support that. Studies on polygraphs and other cues show people are barely above chance at detecting lies. So one of the key predictions of Trivers’ model is falsified.
I’ve argued in my own work that we don’t so much lie and self-deceive as we bullshit. Lying is intentionally saying something false you know is false, for self-interest. Bullshitting is when you don’t really know or care what’s true—you just say what you need to say to persuade, gain status, forgiveness, or advantage. Truth, if present, is incidental.
Bullshitting is safer and more effective than lying because lying carries reputational risks. In small-scale societies, being exposed as a liar could be fatal. So lying tends to be a last resort, used when spin fails or when someone is so powerful they can lie with impunity. For most people, spin, exaggeration, selective evidence, and distortion are the preferred tools.
This reframes self-deception. I think what we’re really doing is self-persuasion, or self-bullshitting. Like lawyers preparing arguments, we rehearse stories in our heads. We practice responses to accusations. We construct narratives that present us as good, noble, and justified, so that when challenged, our answers are smooth, persuasive, and compelling.
The prediction here is different: humans should be good not at spotting lies, but at spotting bullshit. Evidence supports this. Hugo Mercier, in Not Born Yesterday, shows humans are not gullible fools. We’re vigilant, skeptical, constantly alert to hidden agendas. What looks like gullibility—such as conspiracy theories—often stems from hyper-skepticism: an overactive vigilance against deception.
So, conspiracy theories don’t come from gullibility; they come from the opposite. And the better story about self-deception is that it’s not about lying to ourselves so we can lie better. It’s about self-persuasion: rehearsing, spinning, bullshitting—so that we can influence others more effectively.
