Episode 142: Self Deception and Dogmatism

• Listen to this episode: Apple Podcasts

Unofficial, reviewed transcript. It may still contain mistakes; check the podcast when wording matters.

Speakers are labeled by first name where identified. Unidentified-speaker labels may refer to different people. Bracketed question marks indicate uncertain wording.

Transcript

[00:00:00]  Unidentified speaker: Hello, out there.

[00:00:01]  Peter: This week on The Theory of Anything podcast, Bruce delves into self-deception, moralizing, dogmatism, and related issues. As usual, I found his ideas nuanced and illuminating, and I hope someone else out there gets something out of this too.

[00:00:31]  Unidentified speaker: Welcome to The Theory of Anything podcast.

[00:00:33]  Unidentified speaker: Hey, Peter.

[00:00:34]  Unidentified speaker: Hey, Bruce, how are you doing today?

[00:00:36]  Unidentified speaker: Good.

[00:00:37]  Bruce: This episode is part of my series on concepts versus theories, sort of, but it's really going to be about the concept of self-deception and whether that's a real thing or not. So, but it almost directly relates to past podcasts that we've done on concepts versus theories. So, let me kind of take you through what I have in mind here. A little refresher, maybe? So we can talk about, because it's been a while now. I mean, it's been a while since we talked about words versus concepts and concepts versus theories. So there's a chain of thought about concepts that I've been kind of developing. So starting back at video episode 112, which was words versus concepts, that's quite a while ago now. I talked about concepts for the first time, and I argued that concepts are more important than words. Words being really just shorthand labels for various fuzzy interlocking concepts. Then in episode 126, so quite a while later actually, I covered Popper's negative view of concepts versus his positive view of theories. So his main point was that theories make testable truth assertions while concepts do not. He argued philosophers should care about truth. So concepts were of no real interest other than as instruments by which to explain theories. It was theories, he argued, we should really care about because they were about truth.

[00:02:09]  Unidentified speaker: Okay.

[00:02:10]  Peter: So a concept, just to let me get this straight. So concepts are like these things in the world, maybe like memes in a way, are concepts memes of a sort?

[00:02:24]  Unidentified speaker: They actually are.

[00:02:25]  Peter: So they're memes that are almost like things that are waiting to be turned into theories by human minds.

[00:02:32]  Unidentified speaker: Is that how I can think of it?

[00:02:33]  Bruce: I don't know that they're necessarily are going to be turned into theories. So if you were to take the concept of freedom, this one comes up a lot with anarcho-capitalists. Freedom, by which they mean free trade, and they only mean free trade. So freedom is actually a super broad concept, and trying to turn that into a theory is probably impossible, and just not the right way to go about things because of how broad the concept is. It actually covers all, it's like fairness or coercion. It covers so many different things with slight differences and different connotations, depending on context and things like that. So Popper, I don't have the quote handy, but Popper had the quote where he talked about, I believe in freedom, but I would never try to make a theory out of it.

[00:03:27]  Unidentified speaker: Oh, okay, okay.

[00:03:29]  Bruce: So I think concepts are truly useful, right? Like to be able to talk about, let freedom ring, and I believe in freedom, that's not a meaningless thing for human beings to say, but I would never mistake it for a falsifiable theory, right? And it's more like a very, very general idea. And now various concepts can be turned into theories, and do get turned into theories. So we use the example of sound is a wave like water is a wave. When the first person, it was Chrysippus or something like that, from ancient times, started thinking about the idea that sound might be a wave. The truth is that it wasn't a wave, right? Because a wave just meant a water wave, or waves in the grain or something like that. It was a different kind of wave, and they didn't have any concept of different kinds of wave yet, right? Transverse versus compression, etc. So he was actually, strictly speaking, wrong in terms of what the concept meant at that time. But by stretching it and turning it into kind of a more abstract concept of wave, you eventually wind up with physics implications, right? And you can turn it into something that's super powerful in terms of how you think about things. And so concepts have their place, but the problem is that there is no way to falsify them. So when Chrysippus, I think it was Chrysippus, decided that sound must be a wave and theorized that sound was a wave, if you treated it like a theory based on what the word wave meant at that time, it's a false theory.

[00:05:22]  Bruce: But if you treat it like a concept, then you can stretch it through analogy, and you can make it into something that is more valuable as an abstract entity, okay?

[00:05:33]  Unidentified speaker: And concepts are like that.

[00:05:34]  Bruce: We stretch them and we shrink them as the things go out of use, words and concepts will disappear out of the language when they're not useful anymore. And when we do this, it's super useful. It's always just very fuzzy, right? That's part of why you can't really test them. But the other reason why is because they're meant to be stretched. If a concept is wrong, you can just extend the concept and then it's no longer wrong. So it's very easy. They're easy to vary, if you will, and they need to be. That's not a bad thing if they're being used correctly.

[00:06:11]  Unidentified speaker: Okay.

[00:06:13]  Bruce: So in Episode 127, Popper claimed theories were 100 times more valuable than concepts. But then I critiqued Popper's critique, and I found room for some strong disagreement. Concepts are quite a bit more valuable than Popper thought. If you consider the conjecture side of his epistemology, which I don't think he was thinking about. I also argued that there is no clear demarcation between concepts and theories, and that a concept is sometimes treated as a theory, and a theory is sometimes treated as a concept. Even Popper admits this. For example, the idea that sound is a wave like water is a wave, if treated as a theory, it's false, like I was saying. Because sound waves are compression waves, and water waves are transverse waves.

[00:06:53]  Unidentified speaker: They're actually hybrids, but anyhow.

[00:06:57]  Bruce: But if treated like a concept, it will, treating a wave like a concept, it will start to move towards a very useful but abstract concept of wave that ultimately proves so fruitful that you almost have to have that concept to make certain kinds of progress in physics. So it represents a conceptual Schelling point that leads to fruitful conjectures.

[00:07:18]  Unidentified speaker: Okay.

[00:07:19]  Bruce: So I think my view is, is that Popper is right that concepts can't be falsified, so they don't make truth assertions, but that he's missed what their actual value is. So I agree with Popper that a concept like wave, when being treated as a concept, is not falsifiable. Indeed, concepts, almost by definition, have an amazing level of flexibility that makes them unfalsifiable. You can simply extend the concept as you see fit at any time, basically. I've certainly had people I've talked to on the Internet change what they mean by what the concept is, just almost on the fly, right? To make sure that they can't be refuted. But this unfalsifiable flexibility is exactly what makes them so valuable to the conjecture process. Because of this, I agree with Popper that theories, which are falsifiable by definition, are 100 times more valuable than concepts if we're only looking at the critical side of his epistemology rather than the creative conjecture side of his epistemology. So then, in episode 131, which was called Knowledge as Concept, I argued that CritRats often treat their most cherished theories as concepts rather than as theories because they don't want them to get refuted, because concepts can't be refuted because they're flexible. So I used the CritRat Two Sources Hypothesis, which I've beat up quite a bit on the show, the idea that there are only two sources of knowledge, biological evolution and human ideas. And I used that as an example. So I argued that CritRats actually defined their Theory of Knowledge using those two canonical examples, which is typical of concepts.

[00:09:02]  Bruce: Concepts don't have specific definitions. They're usually, or we've talked about this in past episodes on the show, they're usually actually done through canonical examples. So the concept of a bird is actually rooted in a certain canonical example of a bird. A penguin, though you know it's a bird technically, doesn't fit it well, right? And so it's somehow less a bird, even though you know technically it's a bird. And apparently, as we've mentioned in past episodes, when they actually test this, like if they say, you know, the robin, blah, blah, blah, and then you have a sentence, the bird was walking down the street. And then instead they do the ostrich. It will actually take you longer to realize the word bird means ostrich than it does to realize it means Robin. Because concepts are rooted in canonical examples, not definitions.

[00:09:54]  Unidentified speaker: Okay.

[00:09:57]  Bruce: So what I've argued then is that crit rats actually define their Theory of Knowledge using the two canonical examples, biological evolution and human ideas, which is typical of concepts. They freely adjust their criteria for what counts as knowledge until it matches the two sources hypothesis. Therefore, it's completely impossible for it to not match. And actually, they actually can never make it match all the way. It's just impossible to come up with the definition of knowledge that you can hold consistent that is going to only be the two sources hypothesis. And I've shown this over and over again on this show that it's just impossible to actually come up with a definitive theory and definition that only has those two. So in episode 132, which was called Roughly Testable Theories, I then showed that they, meaning crit rats, did the same thing with their economic theories, particularly their anarcho-capitalism. Specifically, what I argued in that episode was that anarcho-capitalism is outcome based. So you have a canonical idea of what it's supposed to look like, and you actually determine what's right or wrong based on that utopian ideal rather than based on any set of actual theories of good governance. And any time you ask them to specify, well, what's the rule here?

[00:11:16]  Unidentified speaker: They just won't, right?

[00:11:17]  Bruce: They'll just kind of change it on the fly. The example I used was if you, if you point out to them, you know, actually, isn't a government really just an HOA?

[00:11:26]  Unidentified speaker: You're not against HOAs, are you?

[00:11:28]  Unidentified speaker: Isn't the government really just an HOA that grew?

[00:11:30]  Bruce: And they'll say, oh no, it's too large.

[00:11:32]  Unidentified speaker: It's like, okay, is that a criteria?

[00:11:33]  Bruce: Like, are you saying contracts aren't legal if an HOA gets too large? Is that like a limit that you put on contracts? And they don't even know what you're talking about, right?

[00:11:43]  Unidentified speaker: They're like, what?

[00:11:44]  Bruce: No, it's too large.

[00:11:45]  Unidentified speaker: That's bad.

[00:11:46]  Bruce: So what they're really doing is they're comparing in their minds, the argument an HOA is like, is a mini government and in theory, if there was no government over them, it could grow to be the size of a government. Legally, by contract, like it could do it. There's no reason why multiple HOAs couldn't form together till they're the size of a town and then form together till the size of a state, form together until the size of country. It just absolutely could happen, and there's nothing in anarcho-capitalist theory that stops it from happening. But they will simply say, no, that's not what I have in mind because they have a utopian ideal in mind, not a set of actual principles, but that they believe in. Now, in episode 131, Knowledge as Concept, I told the story of a crit-rat who got angry with me when I pointed out that he was on the fly, changing what he meant by the word knowledge to fit the two sources hypothesis. He was easily varying it, as it were. So I'd like to dwell on that exchange a bit, because I feel it has relevance to a topic that I consider important. Just as a quick reminder of this discussion, at issue, back in episode 131, at issue were three criteria that I and the person I was talking to, both agreed were a summary of the Constructor Theory of Knowledge. Those criteria were, knowledge is information that has three properties. It is capable of enabling its own preservation. It can be copied from one embodiment to another without changing its properties. It can enable transformations and retain the ability to cause them again.

[00:13:18]  Bruce: He first signed off that this was a proper summary of what counted as knowledge. I then pointed out that the immune system fits these three criteria perfectly. The knowledge or recipe of an antibody enables its own preservation by being useful in killing a novel pathogen. It gets duplicated, so the different cells hypermutate and try out different antibodies, and the ones that don't work die, and the ones that do work, they receive a signal to keep replicating. So it gets itself duplicated when the bad variants die out. So it's exactly point number one. The recipe for the antibody can be copied from one embodiment to another without changing its properties.

[00:13:59]  Unidentified speaker: That is absolutely true.

[00:14:01]  Bruce: That is how the immune system works. The recipe for the antibody can enable transformations, namely killing a novel pathogen and retain the ability to cause them again.

[00:14:10]  Unidentified speaker: This is absolutely true of the immune system.

[00:14:13]  Unidentified speaker: None of these are in doubt.

[00:14:14]  Bruce: The immune system fits the three criteria of knowledge that comes from Constructor Theory of Knowledge. The knowledge that the immune system uses is even stored in DNA. Non-germline DNA, it's true, and it utilizes a mini evolutionary process of intentional hypermutation. When it receives a signal that there's a pathogen, it sends out a signal, hey, all you cells, try hypermutating, and they know how to hypermutate their DNA until they find the right antibody for that pathogen.

[00:14:49]  Unidentified speaker: Okay.

[00:14:50]  Bruce: So they start this evolutionary process of intentional hypermutation to come up with novel antibodies to defeat a never before seen pathogen. It's a mini form of evolution.

[00:15:01]  Unidentified speaker: Okay.

[00:15:02]  Bruce: There is simply no doubt that the immune system fits these three criteria, which are supposedly what defines something as knowledge, according to the Constructor Theory of Knowledge, and that this person I was talking to had absolutely bought off on and said, yes, this is what defines knowledge. In fact, the immune system is not the only counter-example to the two-sources hypothesis. There are many examples of things that fit these three criteria, including, I might add, Deutsch's walking robot example of supposed non-knowledge in Beginning of Infinity. But that he, Deutsch, didn't consider to be knowledge because he only considered biological evolution and human ideas as knowledge creating. Because of course, he's the one who popularized the two-sources hypothesis. In short, the Constructor Theory of Knowledge and the two-sources hypothesis, they are in contradiction to each other. And that's not really in doubt. As I explained in Episode 131, the CritRat that I explained this to, insisted that there was no contradiction and told me that the immune system didn't count as knowledge because it was short-lived and real knowledge is not short-lived. Though he also insisted that this new clarification only applied to animal DNA, not human knowledge, because human knowledge is often short-lived. Note the obvious gerrymandering of his definition to get back to the two sources hypothesis. Specifically, he says, to be knowledge, there's this fourth criteria, it has to be long-lived, but it only applies to animal knowledge. It doesn't apply to human knowledge.

[00:16:36]  Bruce: And this is very obviously a [gerrymandered?] definition to try to force himself back to the two sources hypothesis. Note also that he's retreating to his canonical examples of knowledge, to find something, really anything, that he can use to declare the immune system, not knowledge. So he thinks in his mind, well, what's the difference between the immune system and animal knowledge? And he thinks about, oh, one of them's long-lived and one of them's not, and he immediately creates this fourth criteria to declare it not knowledge. Now I pointed out to him that long-livedness for animal knowledge was not on the list of three criteria that he had already agreed upon. I asked him if I could add it as a fourth criteria. And he got mad at me and he insisted, no, I couldn't add it. And it was implicitly already part of the three criteria.

[00:17:27]  Unidentified speaker: Now let me read the three criteria again.

[00:17:29]  Bruce: You tell me if short-livedness or long-livedness is part of these criteria, even just implicitly, it's capable of enabling its own preservation. It can be copied from one embodiment to another without changing its properties. It can enable transformations and retain the ability to cause them again. Is that even implicitly saying that it's short-lived or long-lived? I pointed out that if it was implicitly already part of these criteria, what harm was in adding it as an explicit fourth criteria? And he got angrier and he insisted that I was being obtuse. We were talking via text now, mind you. So all I was asking him to do was to let me write down what he had already written down in explaining to me in text, so that it was an official part of the list of criteria of what counts as knowledge. So all I needed to do was copy verbatim what he had written to me, and just take a copy of what he had written and make it a fourth criteria.

[00:18:30]  Unidentified speaker: That's it.

[00:18:30]  Bruce: It was hardly a difficult or tricky ask. So why was he getting so angry over this? This was about as reasonable request as was humanly possible. In my opinion, anyhow, in episode 131, I played this off as him feeling like I was being obtuse as some sort of trick, though what the trick was is not clear.

[00:18:51]  Unidentified speaker: And he didn't want me to get away with it.

[00:18:53]  Bruce: And that's probably fine as far as it goes and probably accurate enough. But I really wasn't playing a trick. So there was no real trick to detect here. And the simple truth is, is that there just really shouldn't have been a problem with me writing down what he himself was telling me in writing was a requirement to be counted as knowledge. Even if I had been playing a trick, this still should have been considered a fair request. It's interesting that his main argument to me was that it was just obvious to him that the three criteria of knowledge under Deutsch's Constructor Theory of Knowledge, including included an implicit criteria, such as the idea that for animals only, not human ideas, DNA must last generations to count as knowledge. But for human ideas, that wasn't a requirement. To him, I think it really was just obvious that that was somehow an implicit criteria of the three criteria that we had agreed upon. I suggested in episode 131 that part of the reason he got upset with me initially was that we were playing different games. I was trying to treat the CritRat Theory of Knowledge as an actual theory where the three criteria and only the three criteria have to universally hold. Whereas he was trying to treat it as more of a concept based on canonical examples. Ultimately, implicitly, whatever criteria was needed was assumed to get back to the two sources hypothesis. I do think this is correct insofar as it goes. I can see why he might have thought, gosh Bruce, you're being so obtuse.

[00:20:26]  Bruce: How do you not see that biological evolution and human ideas create wonders compared to say the immune system? In fact, I think he actually made that argument. Somebody did, I don't know if it was this particular person or not. I've had that argument made to me numerous times. Perhaps feeling angry, he simply didn't want to play whatever game he was imagining me playing, even though my request really was perfectly fair and rational. But if I'm being honest, I have to admit I'm uncomfortable with just trying to label it that way to some degree. So what I want to do is I want to explain why. I'm going to argue that there is something far more interesting we can learn here about human dogmatism or perhaps we can instead call it motivated reasoning. So now in episode 137, we had Ray Scott Percival on the show and talked to him about his book, The Myth of the Closed Mind, Understanding Why and How People Are Rational. In that book, Ray argues that humans have evolved something like robust reasoning module that works automatically, and they can't help but see good reason. It changes their opinion whether they intended it to or not. Now, [having had?] this happen to me a number of times, meaning having my beliefs changed when I did not intend it. I can confess he's correct about this in many cases. But as you can probably tell from my questions to Ray in the interview, I was always a bit uncomfortable with his presentation on this point, because anyone paying attention can tell that humans are nothing remotely close to being idealized rational actors.

[00:21:58]  Bruce: When I got a chance, I asked Ray about—to ask Ray about this in the interview, he admitted that was true and that humans are not idealized rational actors. We do not live on Bertrand Russell Island, as it were, or really anything close to it. His real point there was that even if we did, we'd still have misinformation, groups with irrational beliefs, etc. In his book, he goes so far as to say, he's not asking us to ignore the evidence of our eyes, that people are irrational at times, which is good. His real point is that bad or low content explanations and wrong explanations are at a disadvantage compared to correct or good explanations. That absolutely does seem completely right to me. I'm comfortable with the idea that in the long run, humans are pretty rational, at least if you live inside of a dynamic society.

[00:22:50]  Unidentified speaker: More on that in a second.

[00:22:51]  Bruce: Though I suspect for many, the long run, quote unquote, may be longer than a lifespan. I don't think that's really that abnormal to have someone be dogmatic over something their entire life. Which means that for many people, it is not uncommon for them to die holding an irrational belief that they managed to never challenge too much. That must be, I would think that that's a fairly common occurrence. I suspect that this is common for human beings, even inside of dynamic societies. Obviously, it must be very common in static societies.

[00:23:28]  Unidentified speaker: That was what I was trying to hint at before.

[00:23:30]  Bruce: Because then you have 200,000 years of stasis. There, obviously, it—it must—and that was why I asked Ray questions about how does your theory fit with Deutsch's theory on static societies?

[00:23:42]  Peter: Though to be fair, it might—I think it's also very hard to know what people think in their hearts in truly closed societies. Like, for example, an extreme example in North Korea, it's very, I don't know if anyone really knows what people think.

[00:24:01]  Unidentified speaker: Because their opinions are not out there.

[00:24:03]  Unidentified speaker: So yes, you're right.

[00:24:05]  Bruce: So, you know, I'm going to actually talk about this in future episodes. I've come across some really interesting examples of this that I want to share. I think North Korea, like I would be really surprised if the average North Korean didn't think they lived in a tyrannical society.

[00:24:21]  Unidentified speaker: Maybe I'm wrong.

[00:24:22]  Bruce: Maybe they actually have been brainwashed into not thinking that. But like, I seriously doubt it. I suspect that it's much harder today to create a static society like North Korea. And keep people on board without using violence constantly, right? So I think what we have to look at is like old tribal societies, right? Like, do you think that 200,000 years of tribal societies, that there were lots of people that were thinking, you know what, I don't really agree with the elders on this, and that they were being oppressed.

[00:24:58]  Unidentified speaker: Like, I kind of doubt it, right?

[00:25:00]  Bruce: Like, I'm not even sure how you would do that in a tribal society where everybody knows each other. So, we may have to look at two examples here, like large scale static societies like North Korea and small scale ones to get a feel for how Ray's theory plays with each of these, maybe a little bit differently.

[00:25:19]  Peter: Though, you know, you always have to think, well, what are people going to think about our society in a thousand years when they look back into history? Like, are they going to think, wow, they made all these, they coerced all these children to go to school. And, you know, it might be people really, really in their hearts think that was a good idea or, you know, I mean, they're like, we could go down this road of so many examples. I think it's very hard to predict what they might make of our world.

[00:25:53]  Bruce: Yeah, it seems likely, though, that they'll think we're a bunch of moral [cretins?].

[00:25:57]  Unidentified speaker: Oh, sure, sure.

[00:25:58]  Peter: We put all these people in jail instead of just like convincing them that to act better or something.

[00:26:04]  Unidentified speaker: I don't know.

[00:26:05]  Unidentified speaker: Right, right.

[00:26:07]  Bruce: Do you know, I have to make a point here just for a second. I feel like that's actually a misunderstanding of history. So like, if I were to go back and I were to look at, I don't know, like the era of the Israelites or something, right?

[00:26:23]  Unidentified speaker: Right.

[00:26:23]  Unidentified speaker: Yeah.

[00:26:25]  Unidentified speaker: They had famines all the time.

[00:26:26]  Bruce: I mean, the Bible talks about all their famines or whatever.

[00:26:28]  Unidentified speaker: Yeah.

[00:26:30]  Bruce: So I might think that there are a bunch of moral [cretins?].

[00:26:32]  Unidentified speaker: Why didn't they use tractors?

[00:26:33]  Unidentified speaker: Why didn't they feed their people better?

[00:26:36]  Unidentified speaker: Right.

[00:26:36]  Unidentified speaker: I could make that argument, right?

[00:26:38]  Bruce: And I think when I put it in that way, tractors, then you immediately you think it's a silly argument. But if I were to say, well, why didn't they just have a democracy? You wouldn't see that as a technology in the same way you would see a tractor. So you may not see that actually it is a technology.

[00:26:55]  Unidentified speaker: And they hadn't invented it yet, right?

[00:26:58]  Bruce: So the fact that they're moral [cretins?] because they didn't have a democracy, completely misses the fact that you have to actually first invent the knowledge to not have a king, or whatever it was that the Israelites had back in the era.

[00:27:12]  Peter: So we're basically kind of honing in on certain things when we look at history like that. Like, for example, I think I brought up this before, but the idea that women especially consent to being married, like you think about what a powerful... You know, we take it for granted this, I publicly consenting to marriage, marriage, this I do, you say your vows, right? That's an amazing moral invention in a way that we've developed. And it's so easy to judge people on the past by this, through this lens, but like they didn't have that moral knowledge.

[00:27:57]  Unidentified speaker: It didn't exist.

[00:27:58]  Unidentified speaker: That's exactly right.

[00:27:59]  Bruce: You really have to think of things like governments, laws, norms. You have to think of those as inventions like tractors. And I think that that's the right way to look at history. It seems like when people write history, they tend to do it one of two ways. At least this is my experience. They either turn it into a series of moral parables of, wow, look at how terrible these people were in the past, and what can we learn from how terrible they were? Or they try to take the point of view of the person in the past, and they try to explain what their inputs would have been, how they would have understood things, and then try to help you understand the choices that they made. I think both of those are legitimate in their own way. I have no particular problem with using history as a series of moral parables. We need moral parables. History seems like maybe a good source. But I don't think it's an accurate telling of history, if that makes any sense. I think the other view is the more accurate telling of history.

[00:29:02]  Unidentified speaker: So just a thought there for a second.

[00:29:05]  Unidentified speaker: Yeah, no, that's an intriguing perspective.

[00:29:08]  Bruce: Okay, so how well does this exchange that I had with this CritRat from episode 131, how well does it fit Ray's thesis? Now, I admit, it seems to me that it seems to challenge you at least in a narrow sort of way. Specifically, this CritRat friend was being confronted by extremely good reasoning, reasoning so good that he really was backed into a corner, rationally speaking. Yet it did not automatically change his mind, as Ray's theory would seem to predict. Okay, probably that's not really what Ray intended just to be fair to Ray there. But he does say some things where you might read him that way. But I think in total context, maybe that's not the way you should be interpreting him. To see what I mean, consider the exchange carefully. When this crit rat tried to explain to me that why the three criteria of knowledge, which he argued were correct, and it was the sole three criteria, didn't include the immune system, even though really beyond doubt, the immune system fits the three criteria. He simply treat his theory, quote unquote, as a concept rooted in canonical examples, specifically the two sources themselves, human ideas and biological evolution. He'd consult his canonical examples in his mind. He'd note there's a difference between the immune system and animal DNA in the germline, namely, long-livedness. Okay, DNA gets passed to the next generation, the immune system DNA does not.

[00:30:31]  Bruce: Then he'd amend his definition of knowledge on the fly by adding new criteria, in this case, adding a requirement that to count as knowledge, it must be long-lived, but only for animal knowledge. When I pointed out that human ideas are not always long-lived, but he counts them as knowledge, he found that unproblematic because it still fit his two canonical examples. In the results, he simply declared the requirement that it be long-lived as only applied to animals, not humans. And he was completely comfortable doing that. In short, he'd on the fly adjust his criteria for knowledge to fit the two sources hypothesis because he was treating that as a concept rather than a theory or that's my argument anyhow. But every time he did this, I'd say to him, okay, so you just added something not contained in the original three criteria. May I write it down and make it explicit as a fourth criteria? He'd then get angry with me and insist that I was being obtuse because it was just obviously implied, quote unquote. My argument that he was playing a different game, that he was merely trying to explain the concept of knowledge to me using his canonical examples, doesn't quite work here, it seems to me. I mean, if that was really all that was going on, nothing else was going on, then it seems to me like it'd be pretty rational, pretty easy to say, look, the three criteria are meant to be approximate. You are right that it's not going to be a perfect fit.

[00:31:57]  Unidentified speaker: But this isn't what happened.

[00:31:58]  Bruce: Instead, he got mad and insisted his new criteria were implicit in the original three, even though plainly they were not. Or put another way, when things were positioned such that he had an opportunity to see that there was a difference between the three criteria and his canonical examples, surely this was the very point I was making to him. He couldn't see that I was correct, even though it was very clear I was correct and that he had to insist I not be allowed to write down exactly verbatim what he had already just written to me. Do you see how this is at least to some degree at odds with Ray's thesis? This is kind of the point I'm trying to make here using this example. Despite being presented with an overwhelmingly strong refutation of his theory, his robust reasoning module never did kick in. Now, there wasn't really anything to disagree with me over. For sure, the three criteria do not say, and biological knowledge must last for centuries to count as knowledge, though human knowledge is exempt from this requirement.

[00:32:58]  Unidentified speaker: So the immune system is ruled out.

[00:33:00]  Bruce: It just doesn't say that.

[00:33:01]  Unidentified speaker: It doesn't.

[00:33:02]  Bruce: He then proceeded to make a series of arguments that, to me at least, if I'm being honest, didn't really make any sense. For example, why insist to me that the new criteria is, quote, just obvious when clearly it wasn't obvious to me? Like, why is that even an argument? And even if it was obvious, then why not just write it down? Like I'm asking, what harm could it possibly cause to make it explicit if it is by his own admission already implicit in the original three anyhow? There's just no downside to him if that's the case. So his building anger can't be explained merely as him playing a different game, nor can it be explained via the explanation that he was offering. And I want to emphasize again how easy it would have been for a rational actor to simply say, yes, you're right, that these are just approximate criteria. I am conveying a fuzzy idea about what's special about these two sources of knowledge. He really badly wanted to believe that the three criteria defining the Constructor Theory of Knowledge and the Two Sources Hypothesis were fully aligned, and he was forcibly coming up with wild ways to make them align in his head. Now, we all know or claim we do what was really going on here. He didn't want to see his theory refuted, and he was using his anger to get out of that refutation. That is to say, he was immunizing his theory to avoid refutation by getting emotional. We say stuff like this pretty easily, but we miss what a big deal it is. Because what we're really talking about here seems to be what is popularly called self-deception.

[00:34:34]  Unidentified speaker: What is self-deception?

[00:34:36]  Bruce: Is it even a coherent concept?

[00:34:38]  Unidentified speaker: Could it possibly be real?

[00:34:40]  Unidentified speaker: Could it be wrong?

[00:34:42]  Bruce: What does it mean to deceive oneself? Doesn't deception require someone to be aware of the truth and try to deceive someone else about it? Is it even possible for you to, at the same time, be aware of the truth and also deceive yourself about it? Isn't that just a contradiction? So I had a long conversation about this on Twitter with two people. One was a self-proclaimed crit rat, and the other considered himself more of a classic Popperian, not Deutschian, but still a Popperian. Now, both found self-deception to be a difficult to impossible concept to accept. One of them found it impossible to accept, the other one found it possible to accept, but it should be very rare, was his argument. One argued that it wasn't even a coherent concept. Deception requires a deceiver and a deceived, which implies at least two actors. So you can't, you don't have two actors with one person. So it's just impossible, was his argument. The other pointed out how strange it would be if you could at once know a truth and also be deceived about it. So here's an actual quote from him on Twitter, which I think is a very thoughtful quote, by the way. I'm also not so sure, I'm not sure true self-deception in the sense of a person creating his own blind spots is common. Consider what's involved. A person is aware of some uncomfortable truth and so decides to think, try and think of alternatives.

[00:36:01]  Bruce: To end up self-deceived, one of the alternatives they create will not only have to appear true to them, according to whatever truth-related standards they are applying, it will have to have the property of making them essentially forget why they initially tried to create it. Their initial motivation was a discomfort that was brought on because of their own view that a certain idea was true. This is a difficult thing for someone to just overlook or forget. I think that's a super thoughtful argument. Honestly, I do. It's really pretty solid, isn't it? We act like self-deception is normal. We talk about it all the time. Oh, that person is self-deceived. We act like self-deception is normal enough thing for humans to do. But how can a person at once see the truth of something enough to realize how threatening it is and also somehow manage to not be able to tell how threatening it is? For example, let's pretend my CritRat friend from episode 131 was in fact, immunizing his theory consciously. So we're going to go through a few different ideas here of how we might explain his behavior. Okay, by offering up a theory of mind as to what's going on. So let's start with the idea that he is consciously aware that I'm right. And he's trying to immunize the theory to avoid having to admit that. Okay, so the claim now is that he secretly knows he's adding a fourth criteria on the fly, but he doesn't want to admit to me that's the case. But it's an incredibly obvious that he's mistaken. So he isn't going to fool me, that's for sure. So why bother with such a weak argument?

[00:37:43]  Bruce: It's only going to shove him into an intellectual corner.

[00:37:47]  Unidentified speaker: And to be frank, that's exactly what then happens.

[00:37:51]  Bruce: I ask him to do something completely reasonable, write down an implicit criteria and make it explicit, so we can talk about it. It's really hard to see how this could be anything but a fair request. Moreover, if he was consciously aware, he was secretly adding a fourth criteria, he could not know this was going to be my next move. How could he not know this was going to be my next move?

[00:38:11]  Unidentified speaker: It was just to say, can I write it down?

[00:38:13]  Bruce: It's like moving your king in front of an enemy piece and then getting shocked when he lost the game.

[00:38:17]  Unidentified speaker: So this instance—so his insistence that it's just obvious that his new criteria was already implicitly part of the original three is going to force him into a corner and make his whole argument look bad.

[00:38:30]  Unidentified speaker: And that's exactly what happens.

[00:38:32]  Bruce: I simply point out the obvious, there's no good reason not to write down an implicit criteria and make it explicit. If it was already implied, that's his whole argument after all, then why not write it down?

[00:38:44]  Unidentified speaker: Let's ask a different question.

[00:38:45]  Bruce: Why is he so insistent to the point of getting angry about not writing down a fourth criteria? Why not just admit that the three criteria are more of a rule of thumb? I keep bringing this up, but I really want to drive this point home. This seems like an obvious out for him. The simple truth is, is that if he does let me write it down, it's a tacit admission that the three criteria weren't complete. His whole argument was that the three criteria, which were supposed to be a summary of the Constructor Theory of Knowledge and the Two Sources Hypothesis, are one and the same. Adding a fourth criteria is tantamount to admitting that they aren't equivalent after all.

[00:39:20]  Unidentified speaker: Now, I know that.

[00:39:21]  Bruce: That's why I'm making this point to him.

[00:39:23]  Unidentified speaker: But does he know that too?

[00:39:25]  Bruce: He seems like he must because that's why he's refusing to write down the additional criteria. So he's acting like he has an awareness that writing down this additional criteria that comes straight from him, verbatim, that that is a problem for his argument. He's absolutely acting as if that's the case. So that must be why he ends up doubling down on the, no, really, this fourth criteria was really part of the other three. So what he does is he gets angry to shut down the conversation. The conversation comes to an end because I don't want to push him too far. Now, if we assume all of this is conscious, then the assumption is, is that both of us easily see his claim is false. There is a fourth and probably more criteria that the three criteria don't capture. So we now have, we have now robustly shown that the Constructor Theory of Knowledge is not equivalent to this Two Sources Hypothesis. And this whole conversation gives me, at least, exactly what I came for, because that was my point. I wanted to know if the Two Sources proponents could offer a counter explanation for how to reconcile the Two Sources Hypothesis with the Constructor Theory of Knowledge. I couldn't see a way to do it, and apparently he couldn't either. Here's the thing, what is the point of consciously wielding a weak argument when you can easily see that the person you're arguing with is going to come away feeling like you proved them right, which is what happened? It's very understandable that I felt exactly that way after this conversation.

[00:40:48]  Bruce: Okay, so maybe we can claim that what he's doing here is just saving face in the moment. This is something that we often bring up, right?

[00:40:54]  Unidentified speaker: Oh, they're just trying to save face.

[00:40:56]  Bruce: We often point to people acting irrationally and say that they're saving face, but we ignore what a remarkable thing it is that we're really saying.

[00:41:03]  Unidentified speaker: Let's explore this.

[00:41:04]  Bruce: He is consciously aware he's wrong, but he's going to pretend he isn't to quote unquote save face. If we accept race theory that we can't control our beliefs when confronted with a good explanation, this theory makes a prediction. Okay, this crit rat who we're now imagining knows he's wrong, but he's just trying to save face. So this crit rat who can't possibly be self-deceived because we're theorizing that's impossible, must be consciously attempting and failing to deceive me. But he really knows he's wrong, so he'll change his mind in secret.

[00:41:36]  Unidentified speaker: But here's the thing that's interesting.

[00:41:38]  Bruce: Many months later, this person continues to make arguments on Twitter, that there are only two sources of knowledge, and that Deutsch's theory proves that, and that the entire AI establishment is off-base, and has no hope of making progress until they understand what knowledge actually is.

[00:41:53]  Unidentified speaker: This is a very typical crit rat argument about the two sources hypothesis that I've seen numerous, numerous, numerous times.

[00:41:59]  Bruce: In other words, he can't merely be saving face because he continues to act as if the two sources hypothesis is correct in public months later. If he's really consciously doing this, which is what we're assuming at the moment, this seems like an outright falsification of Ray's whole theory. It's supposed to be impossible to hold on to a belief you know is wrong. Ray's whole argument is that if your life or let's say lots of money was on the line, could you force yourself to believe the moon was made of cheese?

[00:42:33]  Unidentified speaker: I couldn't and probably nobody could.

[00:42:35]  Bruce: We literally can't control our beliefs like that.

[00:42:38]  Unidentified speaker: This is Ray's point.

[00:42:39]  Bruce: I want to emphasize, I know this person pretty well. He's not a, quote, good actor in the sense that he could pretend. I don't think he would be able to behave this way just as an act, is what I'm trying to say.

[00:42:53]  Unidentified speaker: He's also a good guy.

[00:42:54]  Bruce: He's not a bad faith actor either, that would keep pushing the two sources hypothesis in public if he knew it was false already. Does it really make sense to keep doing that two months later, just to quote unquote, save face? I have every reason to believe that he very sincerely thinks I failed to refute Deutsch's two sources hypothesis and that he very sincerely believes he refuted my argument. That is why he goes on to continue to argue that there are clearly—with what are clearly fatally flawed arguments. So the save face explanation, at least if you're assuming it's done consciously, just doesn't make any sense. Is there such a thing as saving face without being aware that's what you are doing? If so, how is that different from simply being self-deceived, i.e. you sincerely though wrongly think your argument is correct despite overwhelming evidence to the contrary. We say people act irrationally to save face, we say that, like we say, oh, they're just trying to save face. But they can't possibly be doing it consciously in a case like this. It must be that saving face is being done subconsciously. What I mean is that the individual is not aware that they are acting irrationally to save face. So why bother calling it save face at all? Why not just accept that they are self-deceived and really do believe that they are making a good argument even while wielding a particularly bad one?

[00:44:19]  Unidentified speaker: Okay.

[00:44:19]  Bruce: So let's assume instead that he's not saving face at all. Instead, let's assume he sincerely isn't aware that he's making a weak argument. Perhaps he really does believe that the fourth criteria, long-livedness for animal DNA, is implicitly part of the other three criteria.

[00:44:33]  Unidentified speaker: Does this explanation make sense?

[00:44:35]  Unidentified speaker: Let's explore it.

[00:44:36]  Unidentified speaker: Here we bump into an immediate problem.

[00:44:38]  Bruce: If he sincerely believes that long-livedness for DNA based knowledge was an implicit criteria, then he should have no objection to me writing it down.

[00:44:48]  Unidentified speaker: Think of it this way.

[00:44:49]  Bruce: We're not assuming he sincerely believes, we're now assuming he sincerely believes it was implicitly part of the other three criteria, and thus unproblematic. So there just isn't any reason under this hypothesis about what he's thinking to not write it down right away.

[00:45:04]  Unidentified speaker: Yet he refuses to do so.

[00:45:06]  Unidentified speaker: Why?

[00:45:07]  Bruce: What possible reason could there be, other than to frustrate an argument that is about to refute him? So he can't merely be mistaken, at some level, he appears to know that my argument potentially refutes his theory, if it's written down. So it seems we've eliminated three possibilities. We've eliminated the idea that he was consciously aware he was making a bad argument. We've eliminated the idea he was consciously just saving face, and we've eliminated the idea that he was blissfully unaware he was making a bad argument. What exactly is left open to us here? It would seem that the only option we haven't explored is something truly strange. It is that he was at once aware he was making a bad argument and about to get refuted, and also unaware that he was. This new theory, let's call it self-deception, posits that he knew all along that if he even bothers to write down the additional criteria, he can foresee it ruins his argument. So he's actively refusing to write it down to avoid losing the argument in his own mind. But if that's true, that means he already knows what is wrong with his argument. That is to say, he already knows he's wrong, and I'm right.

[00:46:17]  Unidentified speaker: But this makes no sense.

[00:46:18]  Bruce: For one thing, he comes across entirely sincere. And as I've already mentioned, this is someone I know pretty well. He's just not a bad faith actor, and he's doing it months later. And we can see him continue to make this flawed argument online to other people months later. It seems unavoidable that he really does think his argument was somehow valid. This seems only consistent with the idea that he sincerely believed he was right. Yet his refusal to write down a falsifying criteria and getting angry, that I even asked him to, both seem only consistent with the assumption that he knew that if he did, his theory would be falsified, implying that he knew he was wrong. It's almost like part of him knew that writing it down would result in a falsification of his theory. So he refused to write it down and got angry to avoid having to. But then somehow the refusal meant to his mind that this theory hadn't been falsified. Again, there seems to be something wrong with Ray's theory here. Here we are with what seems like a fairly transparent act of self-deception, where someone is somehow forcing themselves to believe the rational equivalent of the moon being made of green cheese. Before we accept the admittedly very strange self-deception theory, let's explore one more idea. We might say he sincerely thought I was playing some sort of trick, even though I wasn't, and he was shutting that down. But what exactly are you claiming here?

[00:47:41]  Unidentified speaker: That he sensed the trick but didn't know what it was?

[00:47:43]  Unidentified speaker: Is that the claim?

[00:47:44]  Bruce: You can't claim he knew what the trick was since there wasn't one. Maybe he imagined a trick that wasn't there. Okay, what was the supposed trick that got him angry? Try to come up with something in this context. It's really hard to figure out what the trick might be. And even if there was a trick, why not just write it down and then prove that writing it down doesn't matter? In fact, it's sort of what is required in this circumstance. You got to write it down, say, sure, I'll write that down. It's a requirement. This is why it doesn't matter.

[00:48:12]  Unidentified speaker: And say it's because it's implicit.

[00:48:13]  Unidentified speaker: It's part of this one.

[00:48:14]  Bruce: You would do an explanation from there. What seems like the obvious way to neutralize any supposed trick is to just write it down.

[00:48:22]  Unidentified speaker: This is where things get interesting.

[00:48:24]  Bruce: I did have a trick of sorts, but it was completely transparent. If this crit rat had allowed me to write down any additional criteria they added on the fly, I'd have then used that new criteria to find yet another counter example and force him to write down another criteria until the list numbered in the hundreds. This is, I'd argue, exactly the right thing for me to do. It makes clear exactly what is wrong with the pseudo-Deutsch theory of knowledge by bringing the gerrymandering definitions to the fore so you can see it happening in real time. It also makes clear why this strategy can never work. I'll always be able to find yet another counter-example because there is no clean set of criteria that defines only biological evolution and human ideas. Is this still a trick if it's actually bringing to light a real problem? Isn't that what we would normally call a good argument? Interestingly, he took the one strategy that stopped me from making that point, a completely correct and true point. He refused to allow me to make the point at all by getting angry. So you can't really argue here that he sensed there was a trick, but didn't know what it was, because he picked exactly the only path he could, forced into it by the logic of the situation, by getting angry to avoid letting me make my completely valid point.

[00:49:41]  Bruce: He first claimed all new criteria I forced him to add by offering examples of knowledge that didn't come from the two sources, were all implicit criteria. And then when I called him on that by doing nothing more than saying, can I write it down? And he was the one insisting it was implicit, so why not write it down? He came up with a way to avoid even writing it down, because the mere act of writing it down would force him to see that his theory had a problem. When I started to insist we write it down, using a perfectly reasonable argument, I might add, he simply got angry and that ended the conversation without him having to confront the real problem of his theory.

[00:50:17]  Unidentified speaker: Let me put it this way.

[00:50:19]  Bruce: There is nothing random about his behavior. It was the one and only approach he could possibly use to avoid having to admit to himself that he was getting refuted. If he simply didn't know one way or the other, the only reasonable thing to do was to let me write down the new criteria. If he sincerely believed he was correct, why not? Strange as this may sound, at some level, he must have known that the direction I was going was going to refute his theory. Some part of him clearly saw what I was doing, and why even the mere act of writing down supposedly implicit criteria would ultimately refute his theory. He did the one and only thing he could do to avoid that refutation, he got angry. Yet he then, after the fact, acted as if he sincerely didn't realize my argument had refuted his theory. There is something interesting about human psychology here, namely the fact that angry shutting down of critical discussions like this are not random, but happened precisely when, if the conversation isn't shut down, a refutation will result. Because if the person really and truly didn't see that that was where this was going, why not just allow the conversation to continue? It shows, strange as it may sound, foresight that this discussion needed to be angrily stopped because he sensed it was dangerous to his cherished beliefs. But if he could see it was dangerous to his views, wouldn't that require, as per Ray Scott Percival's argument, that he had already engaged his rational explanation—explanation-making module and could already see the problem with his argument?

[00:51:49]  Bruce: So strange as this may sound, this person is acting as if they can't see the point I'm about to make, even though he must be able to see the point I'm about to make. But if he can see the point I'm about to make, didn't that mean he already understood the point? So why stop me from making it? And if he did not understand the point I was about to make, how did he know to get angry and shut it down? There's a really intriguing mystery here about human nature that I'm really trying to bring out here. So back to the question about whether self-deception is real. I agree there is something strange about the idea of self-deception. It's so commonplace, we're so used to it, that we have forgotten how truly strange it is that human beings can at once be aware that an argument refutes their theory and also not be aware that it does.

[00:52:36]  Unidentified speaker: Yet here we are.

[00:52:36]  Bruce: I have presented a real life case, probably about as clear cut as was humanly possible, where this person was at once aware his theory was about to get refuted. So he stopped the conversation via anger and also went on living his life unaware of what he had done, continued to act as if his theory had not been refuted. I don't see a way around this. Human beings must be able to subconsciously read whether an argument is going to refute their cherished theory, and then send a signal to the conscious mind to do things to avoid the refutation, but without becoming consciously aware that is what they're doing. To put it bluntly, we humans have a capacity to subconsciously know our theory is wrong, while the subconscious mind never tells that to the conscious mind. At the end, the person very sincerely, as did this friend of mine, thinks they avoided a refutation and actively acts as if they did, all while their subconscious mind was able to process the argument and come away with ways to avoid letting the conscious mind know that is what happened.

[00:53:37]  Unidentified speaker: So let me put this somewhat bluntly.

[00:53:39]  Bruce: Ray's theory is only true for the conscious mind. The subconscious mind was well aware I was refuting a cherished theory and came up with ways not to fool me, but to fool their own conscious mind. So now it's just academic. Self-deception is real, period. What about the argument that self-deception is uncommon?

[00:54:00]  Unidentified speaker: Let me make a point here.

[00:54:01]  Bruce: Before listening to this episode, you were well aware of the concept of self-deception, but you probably put little thought into how strange a concept it really is. I would argue that this state of affairs can't possibly have happened, except in a world where you encounter self-deception often enough, that it seems commonplace, so you don't think it's strange anymore. So at least for me, I feel like this settles the question. Self-deception is both real and it's common. I see no way to make sense of what really happened here, unless we're going to fully accept that self-deception is a real thing.

[00:54:34]  Unidentified speaker: And it really happens.

[00:54:35]  Bruce: If this friend of mine's conscious mind managed to actually see the argument, he'd have at least somewhat changed his mind. But his subconscious came up with a way to avoid having to allow him to see the argument.

[00:54:47]  Unidentified speaker: In fact, this is what eventually happened.

[00:54:50]  Bruce: So months later, I got this friend on a Zoom call and actually made the list of criteria physically for him. Now that we were face to face over Zoom, his subconscious could no longer pretend I was being dense. My reactions were on display. He could see how sincere I was. So he had no choice now but to let me make the list of criteria. And each time he avoided refutation by adding a new criteria on the fly, I'd write it down right in front of him on a shared screen.

[00:55:15]  Unidentified speaker: And it worked.

[00:55:15]  Unidentified speaker: He suddenly got it.

[00:55:17]  Bruce: What I had been saying all along and admitted, it made sense and I was raising a valid issue. Apparently, he said, there were sources of knowledge other than the two sources. Then something even more interesting happened. He said, I see your point like that animal memes, for example, can't be knowledge in the genes, pretty much by definition, and your point about the immune system seems fair. But these are rare counter examples, more like exceptions that prove the rule, is what he said. And having now admitted that I was tactically correct, he had found a way to avoid the actual implications of the argument and found his way back to something pretty close to the two sources hypothesis. Though I took the win and we ended the conversation there. And he, to this day, still says to people on Twitter that there are only two real sources of knowledge and that there's these dumb AI people who don't know what real knowledge is, so they'll never solve the problem of AI. So despite now admitting that there are occasional exceptions, nothing really has changed that much, showing a real stickiness to his cherished beliefs, even in the face of refuting counter examples. So let's summarize what we've learned so far. Ray Percival's argument that we have robust reasoning and can't control our beliefs is correct, but really for the conscious mind. But self-deception is a very real human attribute too. Self-deception is when the subconscious can see the truth of an argument is going to refute a cherished belief and sends a signal to the conscious mind to avoid this argument.

[00:56:38]  Bruce: In this particular case, by getting angry to make the critical discussion, to bring the critical discussion to a standstill. There are other techniques the human subconscious has available to stop a dangerous critical discussion. I've argued in past episodes that the single most common way is simply to fail to understand what the other person is actually saying, reimagining it in various ways that the conscious mind knows how to refute. This isn't as hard as it might first seem. Understanding someone is very difficult. There is typically only one correct way to understand someone and an infinity of ways to misunderstand them. Understanding, quote unquote, requires considerable effort and misunderstanding requires no effort. All the subconscious needs to do is send a signal to fail to understand, perhaps even by offering a straw man version that was close enough to fool the conscious mind. It is typically not very hard to keep the conscious mind from ever engaging its robust rationality. All of this requires us to imagine a subconscious that is itself robustly rational enough, though maybe not as much as the conscious mind, to have foresight, to know that the argument, the correct argument if understood, would have been a problem and to avoid that problem. Strange as this may sound, this is what we must imagine happening if we're going to make sense of self-deception as a real phenomenon. Let me just kind of summarize my point of view here, which I'm sure is not going to be a very popular view, but I feel like it's the correct view.

[00:58:04]  Bruce: I believe that the subconscious mind and the conscious mind, I will admit that there's no clear distinction between those two. But I think there's enough of a distinction that we can reasonably talk about the two as separate from each other. I think that the conscious mind is the one that ultimately, if you can get the argument into the conscious mind, forces you to change your opinion. I don't think that's true of the subconscious mind. I think the subconscious mind has a level of awareness of enough robust rationality that it can see, oh, this is a dangerous direction. I can see this argument is dangerous, and it has an ability to not tell the conscious mind so that you can go on believing that you weren't refuted.

[00:58:46]  Unidentified speaker: I think this is just true.

[00:58:47]  Unidentified speaker: I really do.

[00:58:49]  Unidentified speaker: So final thoughts.

[00:58:50]  Bruce: In episode 133, I posed a question. Will AGI's be as dogmatic as humans, or will they be as rational as us? But be more rational than us. So absent a full theory of AGI, we can't definitively answer that question. I think there is at least some reason to believe that some degree of dogmaticness may just be a natural part of being a general intelligence. This is something I want to address in a future podcast. If being dogmatic is required to be general, then an AGI will be dogmatic. And I'm going to argue that AGI will be somewhat dogmatic. But would an AGI need to be self-deceived? This is for me, it's impossible for me to imagine. Human consciousness evolved out of animal intelligence. That was evolution's path to consciousness. So it seems natural that we would have both a conscious and subconscious with a fuzzy boundary between the two. An AGI will not reach consciousness via that route. If we want to, if we want, we can design an AGI to not have a subconscious.

[00:59:50]  Unidentified speaker: Or that's not quite right.

[00:59:51]  Bruce: Rather, maybe it would be, would for practical reasons still have a subconscious, but with the conscious mind having access to inspect its contents, there's no reason why this couldn't be true for an AGI. Imagine if this crit rat friend of mine [had/have?] the ability to inspect his subconscious mind and say, oh, looks like I'm actually getting angry with you because my subconscious mind can see that you are making a fair refutation of one of my cherished beliefs. So it seems you were correct after all. We humans realistically just don't do that. But AGI's, they're, I'm gonna argue that they're going to be able to. I will in future podcasts argue that a degree of dogmatic thinking or motivated reasoning, maybe I should say, is natural to all general intelligences. But here we have a clear cut example of where humans are clearly and I would say even decisively more dogmatic than they need to be to be a general intelligence. So I feel we know that AGI's will be at least somewhat more rational than we humans are. Think Spock here, because they will lack self deception, which really should give them a pretty big boost in terms of their ability to be more self critical than we are with ourselves. And why did we evolve an ability to self deceive if it wasn't strictly necessary? It seems to me there's only one possible answer to this question, shocking though it is. It must be the case that self deception and dogmatic thinking and motivated reasoning have survival value. But what possible survival value could there be for being less rational?

[01:01:24]  Unidentified speaker: Can I put that in a different way?

[01:01:26]  Unidentified speaker: Yes.

[01:01:27]  Unidentified speaker: You mentioned Spock.

[01:01:29]  Peter: Well, bring it back to Captain Kirk.

[01:01:32]  Unidentified speaker: Yeah.

[01:01:32]  Peter: So the world, as I see it, the rational world needs Spock's calculating the odds and Captain Kirk's. It needs C3PO's.

[01:01:47]  Unidentified speaker: Right.

[01:01:48]  Unidentified speaker: And Han Solo's saying, damn the odds.

[01:01:51]  Peter: Let's go through the whatever that was, asteroid stuff or what was that?

[01:01:57]  Unidentified speaker: Whatever.

[01:01:59]  Unidentified speaker: Yeah.

[01:02:00]  Unidentified speaker: The Kessel Run.

[01:02:01]  Unidentified speaker: Yeah.

[01:02:02]  Unidentified speaker: Okay.

[01:02:04]  Bruce: So for now, let's just say, there must be a reason why we evolved self-deception because we didn't have to.

[01:02:12]  Unidentified speaker: It didn't have to be that way.

[01:02:14]  Bruce: So I don't see how you can get around the fact that it must have had survival value.

[01:02:18]  Unidentified speaker: What would be the survival value?

[01:02:19]  Bruce: I'm not going to answer this question today. Maybe in a future podcast, I'll give my opinions on the subject. Obviously, I don't know the answer. I could guess and conjecture about it. But it is interesting the degree to which human beings really do have an ability to self-deceive.

[01:02:35]  Unidentified speaker: It is very common.

[01:02:37]  Bruce: Now, David Deutsch often bristles visibly at the idea of human beings are being irrational. He has argued that humans must be [in the main?], rational or an open society wouldn't work. So it's worth noting that self-deception is not necessarily at odds with his thesis. Consider that the vast majority of decisions you make, say buying something or whether or not to take a medicine or something like that, you have no personal identity at stake in the decision. So we would expect you to not become self-deceived over anything in that case. So put another way, self-deception and irrationality really only take place in a certain narrow circumstance, when you personally identify with the theory being attacked, and so it feels like a personal attack. I think one of the reasons other human beings, never ourselves of course, come across to us as irrational is because we don't spend much time on the internet, having long heated conversations with people about what to purchase, or whether to take aspirin or not, right? Of course, a long heated conversation will only take place if the topic involves people's personal identities being at stake. So it may well be that humans are [in the main?], rational, and also that you mostly encounter them in the rare circumstance where they are irrational, that that's the most common scenario in which you will encounter them. Those two can both be true at the same time. There's just no reason for you to spend time on them when they're being rational. So that's never going to stick in your mind.

[01:04:11]  Bruce: It's never going to be something you're going to spend hours debating with them on the internet. So I don't think David Deutsch is wrong just because self-deception is real, that irrationality is real. This actually fits well with Ray's thesis. Now, I've kind of to some degree maybe misrepresented Ray's thesis because I don't think he really in any way denies that human beings are irrational at times. He often says things like, well, once we learn about our irrationalities, we can come up with institutions to correct them, which I think is exactly the right answer. His book emphasizes over and over this idea that often a theory makes one little immunization and it feels like nothing has changed. But over time, the immunizations add up and eventually the theory is completely a completely different theory. So David Deutsch has also wisely, so I want to point out that that's kind of what happened to my friend, that he finally admitted that there were certain counter examples. And he went back to believing something similar to his original belief, but it was actually a modified version where it was okay for there to be some exception cases now. That really is kind of an interesting change to his theory, even though outwardly he thought of himself as holding the same theory. He was now actually holding a different theory than he did a moment ago. David Deutsch has also wisely suggested that we look at critical discussions, not so much as a chance to change other people's minds, which he admits rarely happens, but as a chance to improve your own arguments.

[01:05:38]  Bruce: Let me actually quote him, so I thought this was such a great quote. The point is that by having a debate, you don't improve your agreement necessarily with the other side, but you improve your own understanding of the other side. If you are right, you improve your own arguments so as to be better. If you think about real life, about people changing their minds about things, you can very rarely remember a case where somebody has changed their mind during a debate. Yet, if you look at the big picture, if you look at an opinion poll about, say, quote, would you live next to a person with a different race? A generation ago, that would be like 20 percent of people, and now 95 percent of people would. And in that time, you can't find anybody who says, oh, right now I've changed my mind about that, or hardly anybody anyhow. What has happened, this is continuing quoting David Deutsch, what has happened is that they've changed their view on a larger scale and on a deeper scale, including in the first instance, the type of reasons that they give themselves for their ideas. Notice that that is what happened to my Crit-Rat friends. I can't find this quote, he says, but there's a marvelous quote from some moral philosopher saying, the reason we need moral philosophy is that people change their reasons for their behavior before they change their behavior. You justify it in a different way. Now, what you are justifying in a different way and better way, even though you are justifying the same view and the same behavior as before, you're now changing.

[01:07:07]  Bruce: And eventually, this will lead to you changing your actual behavior in that little way that we were talking about. But you never see that because it happens as a result of a deeper shift. So this in practice is what happens. And in theory, and in theory of it was first understood, I think, by Popper and expressed in that essay. So that was all from the Joseph Walker podcast episode 139, by the way. Now, I strongly agree with this argument. I think this argument is precise precisely why, even though human nature includes an uncanny ability to self-deceive that ultimately the cause isn't hopeless on the societal scale. But I do want to point out that essentially both Ray and Deutsch are surreptitiously arguing that when people do change their minds, they really only do so unconsciously.

[01:07:59]  Unidentified speaker: That isn't a small point worth glossing over.

[01:08:01]  Bruce: It's a huge and important point about human nature. Humans really do have a powerful and uncanny ability to self-deceive. So powerful, in fact, they need to discover special institutions to escape from 200,000 years of static societies.

[01:08:17]  Unidentified speaker: And that's the end of my TED talk.

[01:08:21]  Unidentified speaker: Okay.

[01:08:22]  Unidentified speaker: Well, another good one, Bruce.

[01:08:23]  Peter: I must say that I think one of the things that makes both the strengths and maybe weaknesses of our podcast, or at least in the weakness, in the sense that it makes it kind of hard for people to, I don't know, digest is that you're just too nuanced. Like your points about, you know, I mean, I think it's actually a very profound point that, that I hear you making here and elsewhere, that these things like dogmatism and moralism, or moralizing, I should say, are dangerous on one hand, but they also must like, I think if they were just dangerous, it would be...

[01:09:16]  Unidentified speaker: It makes sense.

[01:09:16]  Unidentified speaker: And [evolution?] would have removed them.

[01:09:18]  Unidentified speaker: Yeah.

[01:09:18]  Peter: It would be very easy to see that they're dangerous, but they're not. And that like, as you say that, as you've said before, moralizing is always dangerous, but it's also the source of so much good in the world.

[01:09:35]  Unidentified speaker: Right.

[01:09:35]  Unidentified speaker: And same with dogmatism.

[01:09:37]  Peter: Obviously, dogmatism is very dangerous. But as you've said, after Einstein had his happy thought that he held on to for 10 years, he wasn't looking at moral studies. I mean, moral studies, meta studies for guidance. He was dogmatically holding on to pretty much a feeling. I mean, it was a little more than a feeling, but it was a complicated feeling. But it was still, it was a feeling, an emotional connection to an idea that he had.

[01:10:12]  Unidentified speaker: So, you know, it's really a very realistic and compelling way to look at this stuff.

[01:10:24]  Peter: But maybe it makes it hard for our audience to digest it sometimes, and it's hard for me to digest sometimes. But you know, that's where we're at in this.

[01:10:35]  Unidentified speaker: So thank you, Bruce.

[01:10:37]  Bruce: Maybe let's just take a moment to summarize my point, because you're right that I'm very nuanced about it. I know that a lot of my audience is critical rationalists. And I know that the idea of dogmatism, even though Popper did speak positively about dogmatism, the idea that human beings are dogmatic and irrational doesn't seem to fly very well with most critical rationalists. I think probably a lot because of Deutsch's influence, correct influence, like I'm not even arguing he's wrong, right? I do think human beings must in general be rational. I just don't see how we would survive if that wasn't the case. But I think that you can take that idea, human beings are generally rational. And you can miss the degree to which human beings may just be unable to understand your arguments because they're dogmatic, right? And it's something I feel like I have bumped into quite a number of times. That's why it was on my mind and I decided to do an episode on it. I wanted to get my thoughts down. But you're right, I don't want to in any way imply, therefore humans are bad because it's clear that there's some sort of survival value for it, right? And it can be very positive. It seems to be one of the driving forces. Even in science, scientists are very dogmatic about their theories, and they don't try to refute their theories, they try to support their theories, right?

[01:12:07]  Bruce: And that dogmatism causes them to really mine out the true, since theories can be partially true and partially false, verisimilitude, that causes them to mine out the true parts of their theory, which then can get wired into some future theory because of the work that they do. I—it may be that, that—I think there's several reasons why dogmatism is something that we are evolved to do, including self-deception. I don't think all of them are positive, but I think a lot of them are. And it does kind of change the way you look at this, right? To some degree, that's what I'm trying to do. I'm trying to come to grips with the fact that I will make an argument and nobody can understand what I'm saying. And I go, you know, it's probably best I just let this person believe what they want, right? It's if they've got a strong dogmatic bent, let's say, motivated bent towards something, you know, so what?

[01:13:03]  Unidentified speaker: Right?

[01:13:04]  Bruce: And it's not necessarily a bad thing. Often a false theory is partially true. Often a false theory is partially true. Often a false theory is partially true. So I do wish, I mean, of course, I wish we could be a little more open with each other, right? I can imagine a Spock-like character who were able, you know, maybe a best of both worlds, partially Spock and partially Kirk or something. And they have dogmatic beliefs, but they're aware when they're doing it, right? Like I can imagine a sort of perfect utopia where humans are like that, right? Kind of the way Gene Roddenberry kind of imagined humans in the future, they didn't have lots of contention anymore and things like that in Star Trek. We can imagine that sort of future, and there's a part of me that yearns for that, so that it's easier to have conversations and we don't get mad at each other, and people make sincere efforts to try to understand what the other side is saying. And then there's just kind of the reality. That's probably just not the way this works, right? It's probably mostly that we, as Deutsch put it, we change our view on a larger and on a deeper scale, so that we're not aware we're changing it. But I don't think that's, it doesn't, it doesn't change the fact that we do change our views, right? Even if there's not a moment where we go, oh, I've just changed my mind. Even if that's rare for us on certain beliefs, we still do change our minds in very subtle ways, as we re-imagined our theories, as we improve them.

[01:14:42]  Bruce: And one of the things I'm afraid to even say this, one of the things that I'm going to do an episode on, sometime way into the future, because I'm a long ways off from doing this one, is I want to talk about Marcus Hutter's view of induction, because induction is such a dirty word for critical rationalists. Marcus Hutter's view of induction, I say Marcus Hutter's, but it actually comes from Solomonoff, is really interesting because he actually took the concept of induction and formalized it in a way that may not actually be that bad. One of the things I would argue is it's because it's very close to critical rationalism and it's questionable whether you should still call it induction. But he came up with a way to take this vague concept of induction and to keep adding to it and adjusting it and refuse to accept that it was wrong. And he eventually came up with a certain view of induction that's probably much closer to the truth that I think critical rationalists would have thought possible, right? That starts to impinge upon critical rationalist ideas because he had to recreate a lot of the same ideas. So I think that's a very interesting example of why it doesn't make sense to become so dogmatic over terms because we can stretch them. We can take something like induction, which was wrong, and we can stretch the term until it starts to become something true. And I wonder if that's what happens in a lot of cases, that we don't feel like we're changing our mind because we're just subtly changing our theory.

[01:16:12]  Bruce: But the theory that we wind up with is actually now aligned with the theory we think we're fighting. And if you look just a little more carefully, you think, I don't know how good an example this is, but let's say you're a critical rationalist and you're arguing with a Bayesian over epistemology. I have wondered if Bayesians keep improving their epistemology and keep really paying attention to what they're getting wrong. And if critical rationalists do the same thing with maybe their negative views of probability, I suspect the two epistemologies would eventually intertwine and become one and the same, even though they're conceptually said in different ways.

[01:16:51]  Unidentified speaker: I don't know that for sure.

[01:16:53]  Bruce: It's one of the things that I want to explore as a possibility when we talk about Bayesianism, which is something we're going to probably be doing as a topic soon, a number of podcasts on the topic. But it's interesting because concepts are vague and fuzzy. You can stretch them in ways until two concepts that were believed to be at odds with each other become the same concept, or at least overlaps so much that they're not the horrifying alternative that we thought they were, if that makes any sense.

[01:17:24]  Unidentified speaker: Okay.

[01:17:25]  Unidentified speaker: Well, thank you, Bruce.

[01:17:26]  Unidentified speaker: Have a great day.

[01:17:27]  Unidentified speaker: All right.

[01:17:28]  Unidentified speaker: You too.

[01:17:36]  Unidentified speaker: Hello again.

[01:17:37]  Peter: If you've made it this far, please consider giving us a nice rating on whatever platform you use, or even making a financial contribution through the link provided in the show notes. As you probably know, we are a podcast loosely tied together by the Popper Deutsch Theory of Knowledge. We believe David Deutsch's Four Strands tie everything together. So we discuss science, knowledge, computation, politics, art, and especially the search for artificial general intelligence. Also, please consider connecting with Bruce on X at bnielson01. Also, please consider joining the Facebook group, The Many Worlds of David Deutsch, where Bruce and I first started connecting.

[01:18:20]  Unidentified speaker: Thank you.