Episode 44: Clarifying David Deutsch’s Views of “Knowledge”
- Links to this episode: Spotify / 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
[00:00:10] Bruce: Welcome to The Theory of Anything podcast. Hey guys, how's it going?
[00:00:15] Unidentified speaker: Going good.
[00:00:16] Unidentified speaker: Good morning.
[00:00:17] Bruce: Good morning. So we're doing an emergency podcast today. We had a different plan for what we were gonna do, but there's a bit of a story as to how this one came about. I was talking with Bart, who by the way is gonna probably be joining us, if he can, in some upcoming episodes. But he and I have been talking about knowledge creation for quite a while, and the potential difference between Campbell's ideas and Deutsch's ideas. Is it just a linguistic difference, or is it an actual difference of opinion? So, Deutsch on, David Deutsch on Twitter said something that really sounded like it was a difference of opinion. So I actually sent him a message and asked him. And it was really hard to explain stuff through text. And here I am working, it's during my working day. And he suddenly goes, [opening words unclear] talk online right now for a second on Zoom or on Skype?
[00:01:19] Bruce: And I'm suddenly like fanboying, oh my gosh. So I'm freaking out. I didn't have any way of recording it because Cameo does the recordings. I didn't want to lose the chance to ask him questions because I actually wanted to get an answer to my questions as to his opinion on this. So I had to just take notes. So this podcast is going to be about me describing, which is unfortunate. It would have been better if I could have recorded it. But me describing my conversation with him where I got a chance to ask him some of the questions that we've had from our past podcast episodes. Now, keep in mind, I didn't have the questions prepared. It was really hard to figure out on the fly how to give him good examples of say, Byrne's theory, things like that. So I concentrated on the stuff that he was more familiar with. So I didn't talk very much about animal intelligence.
[00:02:22] Bruce: I'll tell you what he did say about that. But I concentrated more on AI knowledge creation. And I felt like I got some really good answers on that front. And so that's what I'm really going to talk about. And then right after I'm done with him, I immediately start thinking of, oh, I should have asked this. Oh, I should have asked that. So now that I know, I mean, like I had no idea that this world famous scientist was just this accessible, right? So I guess I had kind of heard rumors, but I just didn't believe it. So I think what we should probably do is, invite him to the show and actually record it, and collect all our questions in advance so that we can ask intelligent questions.
[00:03:07] Cameo: Be prepared and have notes and be ready for it.
[00:03:10] Bruce: Right.
[00:03:11] Tracy: Great.
[00:03:12] Unidentified speaker: Yes.
[00:03:12] Unidentified speaker: I love it.
[00:03:13] Bruce: So I think we'll try to do that, but let's talk about it because he really answered one of the questions well, that we had made a guess back in episode 38. I had said that I thought I had now figured out what he was really trying to say. Or it was actually maybe episode 39. But anyhow, in one of those episodes, when we were talking about animal intelligence, I had found a quote from David Deutsch that talked about, it really left me with the impression that when he talks about like knowledge creation, he really means that which is created by an open-ended knowledge creation process. This turns out to be really close to the truth. I have a bit of a tweak from that, that maybe makes the answer better than what I had thought. But it comes really close to being exactly that. So I'm of the opinion now, people can form their own opinions. And of course, I'm interpreting, so I understand that maybe I'm misinterpreting him or something.
[00:04:13] Bruce: But I'm of the opinion that the difference is primarily linguistic, but there is at least one legitimate difference between his view and Campbell's view. So let me take you through what we talked about. So first of all, he was not familiar with Donald Campbell. I'm not too surprised about that. Deutsch has obviously been a huge fan of Karl Popper, but he's by his own admission never read all Karl Popper's books. He's not a scholar of Popper. Like, say, Danny Frederick, who before he passed away, I chatted with on Facebook. Actually, just a shout out to my friend, Andrew Crenshaw, who I don't think listens to the show, but he's on Facebook, and he's someone who's looking to become a PhD in Popper's philosophy. And he just knows tons of stuff. Like he can just quote stuff like you wouldn't believe. These people who have really become scholars of Popper.
[00:05:14] Bruce: And on the Facebook group, there's a number of Popper's actual students that are still alive, that will talk about Popper, and some of them are really deep scholars of Popper. That isn't who Deutsch is or is even trying to be. It's no knock at all. It just isn't who he is. So it's not too surprising that he hadn't really heard of Donald Campbell or had any familiarity with Donald Campbell's theory. Now, of course, this puts me in a difficult position, where I have to now try to explain it really briefly off the cuff. And I didn't know how to do that. So I did the best I could. And he had actually listened to a little bit of our podcast, interestingly enough. So when he came on, he goes, yeah, I was listening to a little bit of your podcast. Since we were discussing things, I'd send him some links, and he says, I really liked what you had to say about, you know, when Karl Popper was talking about the evolution of things similar to an autopilot.
[00:06:16] Bruce: He like totally agreed with me that what Karl Popper really meant to say was that evolution, that the software in the mind evolves separate from the physical hardware. And he says, yeah, you're totally right that just, you know, we've got this culture where now we think in terms of software. Software was around back in Popper's day, but it wouldn't have been part of the culture. He wouldn't have had the lingo or the cultural memes necessary to really just say, look, what we're talking about here is evolution of software in the mind. That was why Popper was struggling with like an autopilot and things like that. So I thought that was pretty cool. And then he right away, right off the bat, he goes, look, the real truth is that anytime I use a word, anytime anyone uses a word, that we use it differently in different circumstances. It doesn't always mean the same thing.
[00:07:14] Bruce: I don't even think it could be any other way. I think that when you use a word, you've got this context and you [mean it] in one context like this, and you [mean it] in a different context like that. And then he used the example of creativity. So in his book, he uses the word creativity, which he defines as, let me read it. So I get it exactly. The capacity to create new explanations. So that's maybe a really interesting, albeit non-standard understanding of the word creativity. I mean, you think about when we talk about someone who's a creative artist, a lot of their knowledge is implicit. They don't necessarily do it based on explanations. So this definition of creativity, it may exclude art from being creative to some degree. I don't think that's entirely true because we do use explanations as part of our art. But it's not solely explanations.
[00:08:12] Bruce: A lot of it is just you practice, you gain a skill. Almost every skill.
[00:08:17] Cameo: I'm going to disagree with you. I think that you could use that explanation to say that art is explaining things using typically non-verbal means. And that the act of making art or creating art is exactly explanation. It just happens to be non-verbal or non-spoken explanation.
[00:08:50] Unidentified speaker: It's interpretation.
[00:08:52] Bruce: So let me clarify. I'm not talking about the meaning of art. I'm talking about the creation of art. There's this idea of implicit knowledge versus explicit knowledge. An explanation is explicit. That's what an explanation really is. That's pretty much what we mean by explanation. But we have implicit knowledge too, which is almost like a tacit explanation. It's not a true straightforward, I can explain this to you. If I go out and I ride a bike, I learn to ride this bike and I gain a skill, and it exists somewhere inside my brain. When I first start to ride the bike, I literally have to explain to myself, oh, I got to turn here, I'm supposed to do this under this circumstance, or maybe driving a car is a better example. But at some point, you don't even think about it anymore. It just sort of becomes part of the implicit knowledge in your brain on how to move your body.
[00:09:45] Bruce: And there's still error correction going on, but it's not even conscious anymore at this point. I don't think it would be right to call that explanatory knowledge. That just isn't what we mean by explanatory knowledge. It's a different kind of knowledge. There's many kinds of knowledge. There's heuristic knowledge, which is something totally different, that is not explanatory knowledge. And there's probably relationships between all the different kinds of knowledge. So trying to say this one's definitely explanatory, this one definitely isn't, probably doesn't work. Bart has argued with me numerous times that heuristic knowledge is a kind of explanatory knowledge. And I kind of agree with him. The reason why I can't accept that is because at that point, the word explanatory knowledge means all kinds of knowledge, and it ruins the distinction and makes the term not useful anymore.
[00:10:32] Bruce: So a lot of times, you're trying to protect the term, not because you particularly care about the term. It's just you want it to have some sort of meaning. It can't mean everything, right? So when I think of explanatory knowledge, I'm thinking of a subset of knowledge. And I think a painter's skill isn't something that is theoretical. You don't teach a painter most of their skill theoretically. You teach it by having them do it until the brush strokes come easily. It's like riding a bike, right? Now, they do teach theoretical knowledge for art. There is theoretical knowledge for art. But if I were to go out and ever to learn all the theoretical knowledge for art that exists, all the explanatory knowledge for art, I would still be a terrible artist. In fact, it wouldn't help me much at all until I actually go and start practicing with it and gain that implicit knowledge of how to move my arm with the paintbrush and what colors look the best under the right circumstance and things like that.
[00:11:28] Bruce: Stuff that I may not even be able to explain to somebody else. Or even if I tried, I may explain it completely wrong because I don't understand it myself in terms of an explanation. It's kind of the old thing about trying to explain to someone how to tie your shoes. You may know very well how to do it, but it is tough to explain. And that was what I was getting at. I don't think art is primarily explanatory theoretical knowledge, if that makes sense. I think it's primarily implicit knowledge that you gain from practice. Maybe I'm wrong about that. I have three artist children. I'll go ask them what they think. But I would expect that it would be really tough to explain to people, this is how you'd be a good artist, right?
[00:12:13] Unidentified speaker: Yeah.
[00:12:14] Bruce: And that's really what I was trying to get at is it's just not theoretical knowledge. I'm kind of equating explanatory knowledge with theoretical knowledge.
[00:12:22] Cameo: Okay, I can get behind that.
[00:12:25] Bruce: I agree with you, though, that art is explaining something. A lot of times, you may not even be aware what it's explaining. I think it often captures this idea of implicit knowledge. But when you have really good art, it evokes something in you. And sometimes you have to really stop and think, what is it that it's invoking in me? And then you have to make up an explanation that you didn't have. You were feeling something, but you didn't have the explanation yet. In some cases, you may explain it and then go, wow, now I get why this is evoking something in me. So anyhow, that was kind of an aside, and maybe that's a bad example. Let me use the example he used. He said, you know, when I use the term creativity for biological evolution, I'm using it in a different way. It's a slightly different definition than when I use it, referring to human knowledge, because biological evolution is not explanatory knowledge.
[00:13:21] Bruce: And we've actually brought this up in the past on the podcast. So it was really interesting that he brought that up. We had discussed that fact in the past, and he's well aware that that's the case, that if you're trying to hold to a single definition of creativity as the ability to create better explanations, you couldn't claim that biological evolution is creative. And yet he clearly calls biological evolution creative. But when he does that, he's using a slightly different definition of creativity. He's not trying to be consistent. He's relying on the context of circumstances, how I would say it. That's not what he said. And I think that makes perfect sense, right? Is that we shouldn't expect that the word always has exactly the same definition. Now, why is this important? This is really what I was claiming way back in the AI episode, the AI and knowledge creation episode, episode 25, the one on universal Darwinism, Campbell's theory, basically.
[00:14:23] Bruce: So David Deutsch says, I've explained enough of Campbell's theory to him. He goes, yeah, you know, you know, let's say I have like a robot lawn mower, and by trial and error, it creates a kind of model of your lawn, so it knows how to mow your lawn. Says, I have no problem calling that knowledge. Says, sure, that is, it comes out of a trial and error process. It's a kind of knowledge. Sure, I've got no problem using the word in that way. So to me, that kind of answers the question. He is not necessarily trying to say anything at odds with Campbell's theory. It's just a matter of how you happen to use the word knowledge. And I've mentioned that several times in past podcasts, and it sounds like he completely agrees with that. Now, here's where things get interesting, though. So he went on and he said, okay, and then he started to give me the example of perspiration versus inspiration.
[00:15:14] Bruce: He says, but that, you know, the trial and error for AI, it just tries lots of different things. So it's just perspiration, whereas with humans, you know, it's inspiration. I said, you know, I'm not sure I can buy that argument. I mean, like, I'm not saying it's wrong, but like, how do I know that human creativity doesn't use perspiration? You know, I don't know what the algorithm is in my head, for one thing. It may be doing all sorts of perspiration. Secondly, I forgot to say these. So here are two things I wish I had said, but I didn't get to say. Biological evolution has overwhelming perspiration. I mean, like, it tries all sorts of things, right, over billions and billions of years, and, you know, billions of tries and things like that. So I have a hard time understanding how biological creativity isn't a form of perspiration. And then there's just the fact that if you really look at human creativity, I mean, Einstein didn't have some flash of inspiration in a vacuum.
[00:16:08] Bruce: He put in a lot of perspiration of learning stuff first, not knowing what was going to be relevant and what wasn't. And a lot of it was just happenstance. He, for example, happened to know about Riemannian geometry, non-Euclidean geometry, which most people probably didn't know. He went to some lecture and had learned about it. And when he started to work on his problem, realized, oh, this, I can use that to model what I'm trying to get at with my theory. And if he hadn't happened to have known about that, he wouldn't have been capable of coming up with his theory. So there's this huge perspiration element that exists with human creativity, too. So Deutsch kind of accepted that criticism. And we started talking about something else. And then he came back and he goes, I got a better way of explaining this. And this is the brilliant thing he said that really I found the most useful.
[00:16:58] Bruce: He goes, you got to think of it in terms of the problem. To understand Popper, it's about problems. Always starts with problems. AI doesn't have problems. And I didn't understand what he meant by that at first, because of course it does have a problem it's working on. He goes, what I mean is that the problem is determined by a human. The human decides, here's the problem that you're going to work on. And then it just trial and error, try stuff. That's what I mean by perspiration. It doesn't have the creative aspect where you have some sort of contradiction of ideas and you decide how you're going to go about working on that contradiction of ideas. You decide what problem you're even going to work on in the first place. And that's the creative side to him, from what I can understand. And AI does not have that at all, which is, by the way, completely true.
[00:17:49] Bruce: He's nailed an important difference between Popper's epistemology and Campbell's theory that I hadn't really fully noticed. I had been calling it open-ended versus non-open-ended, which doesn't really capture it as well as what he's saying, but it's clearly related. It's this ability to change which problem you're working on. And so then he explained a lot of things. He said these before, but now I understood him in better context. He said, yeah, you know, I was talking with the AI guy Demis Hassabis. I guess he's friends with. I don't know if I pronounced that name correctly or not. And he says, yeah, you know, I'll know that it's real intelligence when AlphaGo decides it doesn't want to play Go anymore and it wants to go do something else instead. You know, it can actually pick its problem at that point. And Demis's like, oh, yeah, totally agree with you on that, you know.
[00:18:41] Bruce: And I do think that this is a good way of trying to explain what is wrong with current AI. And what is something that's missing from Campbell's theory that Popper's theory has? OK, Popper's theory has this idea of contradictions that lead to, that become problems. You select which problems you're going to work on. That is itself a creative process. So I totally buy into that. And I think that is a really big difference. Now, in defense of Campbell's theory on this front, I don't know that Campbell is trying to entirely take Popper's theory and apply it to other things. I think his interest was in finding an umbrella that many different things fell under, including Popper's theory. So Popper's theory has additional elements that Campbell's doesn't, but it doesn't lack any of the elements that Campbell's have. So one's a superset, one's a subset.
[00:19:37] Bruce: Now, here's where things kind of get interesting. And this is where I did not think of any of this at the time, or I would have asked him more questions about this. Is biological creativity an example of picking your own problem? Stop and think about that for a second. What do you think? What would be your kind of off-the-cuff answer to that question?
[00:19:57] Tracy: I think yes, probably is my answer.
[00:20:00] Bruce: Tell me why you think that would be a good answer.
[00:20:02] Tracy: It just seems like all the capabilities and traits that animals and people, humans have evolved, it seems somehow to be aware of a problem and being able to go, oh, I'm in this environment. I need to have better eyesight or I don't know.
[00:20:19] Bruce: No, no, you're thinking exactly what I was thinking. I think there is a sense in which we could say that biological evolution chooses its problems. And it's exactly what you're talking about. Biological evolution isn't optimizing a single thing. That's probably how we would normally think of it. In fact, it actually is. And I'll explain in a second what I mean by that. But it's going to create wings as a type of locomotion and try that out. Right. And that may lead to better survival value. It is, in another sense, always working on exactly the same problem. And this is why I think it's difficult to say what the correct answer is. The problem it's always working on is the ability to replicate the genes. Everything biological evolution ever does is always because the thing that's successful is always because it replicated the genes better compared to its competitors.
[00:21:14] Bruce: So you could make the case biological creativity does not have this element, by concentrating on one level, that it's always optimizing a single problem, replication of genes. Or you could go one level down and you could say, oh yeah, it absolutely is showing this ability to pick its own problems by looking at the myriad of different ways that animals create survival. They're not all in direct competition with each other. They pick different survival strategies. They may be in non-competition with each other. Particularly if it's two animals living on the opposite side of the world in different climates, they may have zero competition with each other. So I actually think that that's an interesting question and one that needs a better answer. It's too easy to make either answer at this point. We need to understand how that concept of picking your problem applies to biological creativity in some better way than we currently understand it, to get a more definitive answer on this topic.
[00:22:14] Bruce: Now, given that, at least we can see how it's related. So let's take the stance for a moment that biological creativity and Popper's epistemology both have this element of picking your own problem. Campbell's blind variation and selective retention does not have that element. Now, like I said, you could think of it as it's an umbrella. He's trying to say, look, I'm looking at all blind variation selection processes, and I'm looking at the fact that they solve problems, and therefore, in my mind, they create a certain kind of knowledge. And Deutsch is saying he can accept that that's a valid word for a trial and error process like Campbell's concentrating on. But Deutsch is saying, yeah, but it's not the same. Yes, you can call that knowledge, and there's nothing wrong with that, and he's making a valid point. But it's not the same as being able to pick your own problem.
[00:23:09] Bruce: And that would be what Deutsch is calling knowledge. That's what he's calling creativity, okay, to kind of use that term. So we might, therefore, have a difference between Campbell's theory and possibly, obviously, there's a difference between Campbell's theory and Popper's theory, but there may even be a difference between the commonalities between biological evolution, Darwin's theory, and Popper's theory, and Campbell's theory. So this is something that's an interesting idea for further research or thought. So I wanted to call that out. Now, here's a question for you. Are animals creative in this sense? Can animals pick their own problems?
[00:23:48] Unidentified speaker: Not sure.
[00:23:49] Cameo: I'm not sure either. It's an interesting question. I mean, [unclear], somewhat, maybe.
[00:23:56] Unidentified speaker: I don't know.
[00:23:57] Bruce: It's a tough question, isn't it? So what makes it tougher is this idea that you can always claim that the quote, the knowledge was in the genes. So let's clarify what we mean by that now. If you're thinking of, if you're Deutsch and you're thinking of biological evolution as being a kind of picking your own problem and therefore creative, and you say the knowledge is in the genes, what you really mean is the open-ended knowledge creation where you pick your own problem happened at the gene level and the animals don't have that. That's what he's claiming. OK, it's not that they in no sense have knowledge or can create knowledge. Of course, they have trial and error learning. They can just like an AI, they can create models of things or whatever. Right. He's not denying any of that. In fact, he even went so far as to say, what animals do is amazing compared to what our current AIs do, something that I've called out several times.
[00:24:54] Bruce: He says, look, we have no idea how animals are really going about doing this because we just don't understand the algorithm they'd be using. They're clearly way more open-ended or general on how they, you know, go about doing things compared to our current AIs. So I thought it was interesting that he brought that out since I've mentioned that several times. But I think by saying all the knowledge is in the genes, by which you really mean the creative aspect, the choosing of a problem was in the genes, that might be a true statement. It could be. But I don't think it's clear. I don't think it's clearly a true statement. If we were to use the example of the dolphin that saw the man smoking and went off and got milk from the mother and then swam back and put a puff of milk out so that it could join in smoking with the man, you could always make some sort of explanation.
[00:25:47] Bruce: Oh, well, that problem was actually chosen in the genes because of, and you can always come up with some ad hoc explanation for why that is, quote, knowledge in the genes. But boy, we're talking about a choice of a problem that really seems entirely outside of the ancestral environment of a dolphin, right? I would have a hard time saying that this isn't a kind of creativity, but I don't know, right? Without being able to [define it] better, it's really hard to say. And then I gave David Deutsch the example of Alex the parrot, where Alex the parrot can actually count, right? Understands the concept of numbers, where the way we know that is an animal can seem to learn the concept of numbers through trial and error. If you give it four and it learns, oh, four of this type of item, I'm going to answer four, right? And by trial and error, it could seem to have knowledge of numbers.
[00:26:44] Bruce: And, you know, if a dog has knowledge of numbers, most likely, that's all it's doing. Alex the parrot could be given novel items and say, there's four of these items and there's two of one kind that it's never seen before in its life and two of another kind. That requires the abstraction of numbers, right? [Unclear; possibly: “Would seem to.”] Again, this seems like it's a problem completely outside its ancestral environment. You know, who knows, though, right? Maybe they come with a pre-programmed concept of numbers for some reason and it wasn't actually taught. I don't believe that. I think parrots in general do not have a concept of numbers and Alex was taught the concept of numbers. I think there's probably good evidence of that. This isn't something that we have to guess at. We could probably perform experiments to determine if parrots normally have a concept of numbers.
[00:27:35] Bruce: And I suspect the answer is going to be no, they don't. They got no concept of numbers. But you could always make the case because animals, they're hard to test. It's hard to replicate the conditions. As I talked about in the animal episodes, you can always make the case. It's just a coincidence. You can always come up with some sort of ad hoc saying and you might be right, right? It's a tough thing to be sure. So I think this one's an open question, right? Clearly, though, animals don't have open-ended creativity. So even if Alex the Parrot really did learn numbers, which I suspect he did, Alex the Parrot's never going to start composing new kinds of poetry or music or invent rocket engines, right? I mean, there's clearly a limit to a creative bound that exists on animals. And I think this is why Deutsch really is hesitant to see animals as being creative.
[00:28:32] Bruce: And I can totally see where he's coming from on this. The very fact that they have this bound really seems like it maybe calls into question. Is it really creativity? Clearly, it can't be creativity in the full human sense. So does it make sense to call it creativity? Now, however, let me point out that that's also true for biological evolution, which he does call creativity sometimes. Biological evolution is clearly very open-ended on the problems it can select and then try to solve. But I don't think we will ever see biological evolution create rockets to go to the moon. I mean, there is a clear boundary on biological evolution also. It's much further out than a dolphin's intelligence or creativity. It's a much larger boundary. But there's clearly a boundary on them, on biological evolution also. It can't do anything. Absolutely anything is possible for it [unclear wording about limits].
[00:29:26] Bruce: There's limits to what it can do. What those limits are, we don't really entirely understand, but it has limits. I think most of us would accept that it has limits. So even with this further definition of creativity and accepting that there is somewhat of a difference between human creativity and biological creativity, I think there's still a lot of open questions here. And I think that's probably the most I can intelligently say at this time, is that we need to understand this concept of creativity. And I feel like the idea of picking your own problems is a really interesting way to maybe go about trying to explore these thoughts further. And that's probably the most I can say at this point. I've got to give it more thought myself as to how you might be able to utilize this. Now, by the way, when I brought up Alex the Parrot, knowing numbers, he didn't like contradict me or anything, but I could, and he was like, Oh, yeah, that's interesting.
[00:30:25] Bruce: But I could totally tell from the look on his face that he was thoroughly skeptical of that. So I don't think it changed his mind in the slightest. And I asked him if he was aware. I said I read some of Byrne's books, and he says, Well, I've read two of Byrne's papers, but I've never read any of his books. So you already know more than me about Byrne's Theory because you've read his books. And again, this is another thing that's tough. How in the world do I explain Byrne's Theory to him in a few words? I mean, I don't have a ton of time. He's undoubtedly a busy person.
[00:30:58] Unidentified speaker: Sure.
[00:30:58] Cameo: I mean, it took us four, three episodes to talk through Byrne's.
[00:31:04] Bruce: Four, if you count the last one. Yeah. And that was my attempt to summarize, right? And I'm trying very hard to really help people understand what Byrne's Theory and Viewpoint are. And it's not the easiest to explain. So I tried to explain to him, well, insights about building models. And he'd say, well, I mean, like all animals build models. It's like, okay, that's true. I didn't know how to explain it well to him. That's one of the ones where if I were to bring him back, I might be able to figure out how to summarize it a little bit better. But I think realistically, he would have to have enough interest to go read Byrne's books. And there's no reason why he should have that interest, right? And then it would be really interesting to see what he thought at that point. And maybe he would have legitimate criticisms that I haven't thought of, right?
[00:31:51] Bruce: I mean, like, I'm seriously interested in if David Deutsch read Byrne's books, what would he think of them? And what would he say about those? Would it to some degree change his mind? Would it not change his mind? And would he have reasons for why it didn't change his mind? Things maybe that Byrne hasn't thought of. These are interesting questions, and, you know, maybe hopefully someday he'll read Byrne's books and we'll find out, but he may never, right? It just may not be an area of interest for him, and it's just not where he's going to spend his time. That's totally legitimate, like, completely legitimate. And I actually, we talked about that. We talked about the fact that I feel like no matter how much I read, I'm barely scratching the surface. And he goes, yeah, you got to think about Popper's statement that we're kind of all equal in our infinite ignorance.
[00:32:37] Bruce: There's just always just too much to learn. There's just always, always too much to learn. So you're totally going to be ignorant on many, many, many, many subjects, and there's no way around that. So it's no knock on David Deutsch that he doesn't want to understand that theory better. Maybe he's looked into it enough and he thinks he understands enough, and maybe he's even right. Maybe, you know, I'm buying into Byrne's theories too much. You know, that's a possibility. But because of that, I think it would be tough to ask him detailed questions about Byrne's theories. That would be unfair to some degree, because Byrne's theories aren't simple. You have to actually understand his methodology, why he believes what he does, why, you know, how he went about experimenting, coming up with these things. It's not a simple subject. It's something that you have to take the time to actually go understand.
[00:33:34] Bruce: Let me use an analogy here that has been bouncing around in my mind. You know, growing up religious, back in the, [Cameo] and I have talked about this in the podcast, back in the 70s and 80s, I grew up in an environment where evolution was wrong. I was actually at one point told by my parents, you know, science has it wrong about evolution. You know, you're a kid, you believe it. It's not like you don't know what evolution is. Richard Dawkins often says, the reason why people don't buy evolution is because they're not educated about it. We've done studies on this. You can go to Christians who do not believe in evolution, or religious people, not just Christians, that do not believe in evolution, and you can quiz them on evolution, and they absolutely do understand the theory when you do a study on it. They are not ignorant of the theory. That is just not the case.
[00:34:27] Bruce: So knowing the theory is insufficient to convince people that think it's wrong, that it's right. That is not, does not work, you know, plain and simple. Now, as someone who came from that environment and then eventually changed my mind, what is it that changed my mind? Well, I can tell you what it is that changed my mind. As I started to read about it on my own, whatever I was taught in class never was sufficient to convince me of anything. They're going over it too fast. They explain the theory well, but explaining the theory does not change your mind. It just doesn't, right? It's totally insufficient to change your mind. But when I actually, since I love science and I'm reading about science, I would read about the problems that evolution solved and how they went about using evolution to solve these problems. This makes perfect sense now that I understand Popper's epistemology.
[00:35:17] Bruce: You have to really understand the problem space. You have to understand the history of problems that led to the theory to really get why the theory is important. And you would read about, I would read about stuff, and they would give this example of how there's this sort of situation, the whale has these things that look like legs that used to be legs, but they've atrophied through evolution and they're not needed anymore. So they don't actually show up, but you can still see it in the bones. That's just one example, right? And you can see there's this problem, you can see how evolution solves the problem. Now, when you have the first problem and you're told the first solution, it is not convincing. It's not even slightly convincing. And the reason why I think is because it's really easy to come up with an alternative explanation at that point.
[00:36:07] Bruce: And this is why corroboration, the Popperian concept of corroboration, is such an important concept. I know some Popperians don't believe in Popper's concept of corroboration, but it's a totally important concept, the idea that we need to stress test our theories as best as we can. And every time a theory comes out positive, that does, in some sense, strengthen the theory. Okay, that's why some Popperians don't buy Popper's corroboration, is because they can't buy the idea that a theory can ever be strengthened. They only see it as you refute them and that's it. And you never make the theory stronger. It just is not true. And actually, David Deutsch in the Lunar Society podcast recently talked about this. He was asked, you know, if you actually get more tests and more convincing tests, doesn't that change the epistemic status? And he goes, yeah, it does, because it makes it harder to come up with a good alternative explanation.
[00:37:02] Bruce: And that is exactly the right answer. When you have a single problem with evolution and you get it resolved, it doesn't convince you of anything. The second one doesn't convince you of anything. The tenth one doesn't convince you of anything. By the 100th one, you know that there's not a good alternative explanation anymore. But until you actually reach the 100th problem that evolution solves, you are not convinced by it. And that's really how it works. As each corroboration happens, your alternative explanations that you're using in your mind, you start to understand their weakness, that there really are just not very good alternative explanations that you can work with. And at that point, it kind of falls away and you go, okay, I can accept this now. But in defense of the Christians that have a problem with evolution, I've never had a problem with people having a problem with evolution.
[00:37:53] Bruce: Let me say that. If somebody tells me they don't believe in evolution, I'm not going to try to talk them out of that. I'll tell them what I believe, and if they ask me why, I'll explain it to them. But I don't really feel any particular need to say, oh, you're wrong or something like that. I know there are people who do feel that way. Even there's Christians who feel that way. I might point them to Francis Collins' book, The Language of God, which he is an atheist turned Christian, devout Christian, who was in charge of the Human Genome Project, wrote a book defending evolution to Christians. I might suggest that book to them or something along those lines, and then let them just decide for themselves. The thing that I think needs to be understood, though, is that I've talked to many people who are pro-evolution, who have not reached the hundredth problem either, and have never really earned the knowledge for themselves.
[00:38:48] Bruce: And as much as they do believe in evolution, and as much as I believe they are correct on that, they are really ignorant of the theory of evolution, as much as, sometimes even more so, than someone who doesn't believe in evolution. When you, for example, a Christian might bring up a counter example to evolution. They might say, Bart translated this whiptail. I don't remember what it's actually called in English, but there are certain things that exist that we don't know how neo-Darwinian evolution could have formed it. We don't have a good theory for how it could have formed, because there are two things that have to exist, and each one individually would not be useful on its own. It's like this tail that whips on single-celled animals. I'll have to look it up and give a better example in a future podcast. And there is no current known—like Francis Collins mentions—there is no current known answer to some of these questions.
[00:39:44] Bruce: They are, in a certain sense, not in a certain sense, they are Popperian refutations. That's precisely what they are. Now you have to understand that a Popperian refutation does not mean the theory is false. That's a misunderstanding of Popper, a common misunderstanding of Popper. But evolution has Popperian refutations that exist. It would be natural for someone who believed in evolution to believe someday we're going to understand it. I think they're right. I think that there is an explanation that exists that we just don't have yet that's going to explain why this, how this evolved and how it's still consistent with Neo-Darwinian evolution. But a Christian is at least aware that there are problems. And a lot of people I've talked to who are pro-evolution really and truly don't understand that there are problems that need to be worked on.
[00:40:28] Bruce: Now, of course, there's problems. Every theory has problems. People misunderstand the importance of this and how it fits into things. The issue with Darwinian evolution is that it has no competitor. It doesn't matter if it has problems or not. It doesn't have a scientific competitor and one that's an actual good explanation. This is the real reason why we embrace Darwinian evolution. It's got nothing to do with it does or doesn't have problems. It's got nothing to do with we've corroborated it. You can't corroborate something as true. You can't. That's just not what corroboration means. To really understand why we should embrace evolution, you have to understand Popper's epistemology. And the vast majority of people I've talked to who believe in evolution and just learned about it in school, they haven't earned that knowledge for themselves.
[00:41:19] Bruce: And they're in essence believing on it with faith, just like a Christian might believing against it with faith. I don't know if this makes sense or not.
[00:41:30] Unidentified speaker: It's making sense.
[00:41:31] Bruce: Yeah. Now, this is a tough thing though, because it does mean that to really understand a theory, you really do have to put some effort into it, right? You have to understand the problem space. Now, I think in some cases, you don't necessarily have to reach the hundredth problem. In evolution, you probably do. I think that's a tough enough theory. You may have to read deeply enough that you reach the hundredth problem to really understand why evolution matters and why it's such a good theory. But like quantum mechanics and many worlds, I was able to find a single problem that allowed me to move from I can't buy many worlds to oh my gosh, it might be true. And it took a little bit more than that. I had to then go look at the alternative explanations. It took me a while to seek those out. But sometimes there's like the one problem that's so good that you can use the one problem.
[00:42:28] Bruce: And if you can really get people to understand that problem and how this theory solves that problem, it will be very convincing, right? And I think that does happen. Definitely for quantum mechanics, by the way, for anyone who wants the one problem, go look up the Elitzur–Vaidman bomb experiment. That is the one that for me really made me realize I've got to start taking that theory seriously. Because there was just no other way. The math for it is so simple. It's not that hard to understand. You can actually walk through the math for yourself. We'll do it for a podcast sometime. And you can actually understand the theory at the theoretical mathematical level using a very simple thought experiment. And you can immediately see that there is no way to explain this without referencing many worlds as of today. So it does depend on what the theory is and what its problem space looks like.
[00:43:19] Bruce: But I think it always does come down to you're not convinced until you see it solve problems and you understand the problems it solves. And you really have to understand that problem space before you get the theory. So this is all kind of a large aside, and it's somewhat of a setup for a future podcast episode that we're going to do. So let me go back to my conversation with David Deutsch at this point. And the way this ties into my conversation with David Deutsch is we shouldn't expect that David Deutsch is necessarily going to immediately have answers to a competing theory like Byrne's Theory. He would need to go through that problem space and understand the problems, understand how it solves the problems, before it's going to be at all convincing. And that takes time. And with Byrne's Theory, it's a particularly tough one. It's more like evolution than like quantum mechanics.
[00:44:13] Bruce: The problems are always a little bit vague. They're always a little bit fuzzy. I mean, we talked about this. We talked about, and Cameo was very skeptical of some of the examples that I used. And you should be skeptical of them. Every single individual example that Byrne uses, you should be highly skeptical of. It's only after you read the 100th one that you start to realize Byrne might be on to something. And you start to say, you know what, I'm kind of feeling stupid with all my little alternative explanations. They just aren't working as a coherent theory anymore. And that's how I feel about Byrne's theory. I feel like until you read his example after example, I gave several examples in the podcast to try to give a feel for that. He gives even more in the books, right? Until you've gone through enough of those examples and you've gone through enough of his methodology and really checked them out, and you understand the problems that it solves, you don't really understand Byrne's theory.
[00:45:11] Bruce: And it's very easy to come up with an off the cuff, oh, this is probably this. You'll feel comfortable with that. So I do think this is one where I wouldn't expect that unless you've actually gone and done the homework, that it's going to seem like you can just make a really simple answer. Does that make sense? Do you understand what I'm trying to say here?
[00:45:31] Cameo: Yeah, and it also means that we may never get a really great answer from Deutsch about this particular question, because, like you say, he might not, if he doesn't have interest, and that's a lot of problem-space work to do just to be able to refute, you know, somebody's competing theory.
[00:45:52] Bruce: That's right. I think that I think this is a tough one. And you know what? Even after we've gone through this, I've now read two of Byrne's books. I've read a couple of his papers. I've explained it to you guys. It's clear at this point that I'm favoring Byrne's theory. But even then, I still feel a certain degree of skepticism towards his theories. And I think Byrne feels a certain degree of skepticism towards his theories, right? It's the best theory he's got at this point. But we really are dealing with a particularly difficult to test space when we're talking about animal intelligence. So I think that there is room to feel comfortable with an alternative explanation still, if that makes any sense, maybe shrinking room. And I suspect that we're on our path to eventually the alternative theories are just going to die. But I don't know that for sure, right?
[00:46:45] Bruce: It could be that there's something we're missing. But I don't blame anybody for being skeptical of Byrne's theories. Let me just put it that way. I feel like I totally get skepticism of Byrne's theories at this point. Just as I get skepticism of evolution from most people who haven't yet gone through the problem space well enough, I completely understand why evolution, honestly, on the surface, sounds like a completely bafflingly stupid theory. It really does. And it just really goes against our intuitions in a lot of ways. If you weren't raised with it, it should cause you to feel skeptical. And that's why people were so skeptical of it for years. On both sides, left and right were originally really skeptical of Darwinian evolution. And it had to win very slowly people over as it continued to become a better and better explanatory theory.
[00:47:39] Bruce: And it was more and more clear there's no alternative theory available. And that's why it eventually won the day. And at this point, even most religious people accept biological evolution. The vast majority of religious people I know accept biological evolution in one form or another. But that's a very slow... It takes maybe centuries. How long has it been since Darwin? I don't know. It's been a long time since Darwin.
[00:48:02] Unidentified speaker: Yeah, it's been a while.
[00:48:02] Unidentified speaker: 300, 400 years.
[00:48:04] Bruce: Yeah, I mean, it's... So we should expect it to take that long, I guess is what I'm saying. We should be understanding of people's skepticism of even good theories. It takes a while for it to sink in. It does. And if Byrne's theory is correct, it will take a while for it to sink in, but it will eventually sink in. At this point, I don't think we're anywhere near Darwin's level, right? I mean, I think we're way earlier in the process with Byrne's theory than we are with Darwin's theory. So we owe David Deutsch at least another 100 years before he's really in any reasonable denial of a theory that's correct, if that makes any sense. At this point, we should accept that Byrne's theory could be wrong for reasons we just don't understand yet. And we should keep an open mind towards that idea. Okay, now getting back to my conversation with David Deutsch.
[00:48:58] Bruce: So I felt like this boundary line that he talked about with picking your own problems was this really ingenious boundary line, and it really got me thinking. So now he brought up AlphaGo. Now I probably would have if he hadn't, because it's a really good example of what we're talking about. AlphaGo invented a whole new style of gameplay. I would be hard pressed to not call that a form of creativity. But clearly it's not creativity in the way Deutsch is using the term, because AlphaGo is solving exactly one problem, and a human picked the problem. And it's doing it through what he calls perspiration, just playing billions of games with itself. Furthermore, the creative aspect of AlphaGo, and I told Deutsch this, so now he's aware of this, really comes from the fact that it optimizes the problem differently than the human would, because of the way the math works.
[00:49:52] Bruce: And to some degree, even the creative aspect of AlphaGo really came from the fact that the humans gave it a fitness function that is different than the one humans were using. They didn't realize that was what they were doing, but that was what they were doing. So from the Deutsch viewpoint, the creative aspect came from the human, so I can completely see where he's coming from. To say, oh, the creative aspect of AlphaGo actually came from the programmer, once you understand that he means in terms of picking the problem, that is accurate now. Clearly, that's accurate now. Is this the best way to define creativity? I'm not sure it is. But it is the way he's defining it, and I can see that he's right for the way he's defining it, even in the case of AlphaGo. So we talked about that. He thought that was interesting, by the way, that it was a matter of how it optimized.
[00:50:44] Bruce: Basically, what it comes down to from our AlphaGo episode is that it optimized... Humans optimize by how much they win, and AlphaGo optimizes that it wins. And it turns out that that's an ingenious, different way to play the game that improves upon how every human has played before. Okay, so I see, therefore, Deutsch's real concern with AI is really this problem space thing, that the problems are all determined by the human. I can see that that means that Campbell's Theory is not identical to Popper's. I never thought it was, but now I have a really obvious distinction that has been made between Popper's Theory, possibly even between the connection between Darwinianism and Popper's Theory and Campbell's Theory. This doesn't mean Campbell's Theory is wrong. This doesn't mean Campbell's Theory isn't valuable. But it's something important to understand about Campbell's Theory, that he's abstracting out part of evolutionary theory, not necessarily all of it, not even necessarily the most important parts.
[00:51:52] Bruce: Now, Deutsch points out, and this is from the conversation, that even human knowledge is not always creative in the sense that he's using the term. So he gave me an example that I never heard before. He said that Newton didn't know what the gravitational constant was, because he didn't know what the size of the Earth was back then. So somebody worked it out later, and what he instead did is he knew what the combination of two values were. I don't remember what the two values were. And then somebody later worked out what the size of the Earth was, and then they were able to work out from that what the gravitational constant was. In Deutsch's opinion, let's not call this, we don't want to say that isn't knowledge. Of course, it's knowledge in the lesser sense, right? Of course that we would want to call that a growth of scientific knowledge.
[00:52:39] Bruce: But it's not creative knowledge. It's not knowledge in the sense Deutsch is interested in. And this is actually his point, right? This is more the perspiration, as he calls it. I still don't think that's the best way to do it. I think it's got more to do with the problem space. The creative problem came from Newton. This other guy just took the theory and then said, well, now that I can figure out this, I can figure out that, right? Now, I would argue that's still a problem. That's still a human picking a problem. So I think there's maybe a little bit of fuzziness, whether we would call this creative or not, under Deutsch's definition. But Deutsch does not see this as creative. This is not what he means. And I think it's always important to understand, what is the person trying to say, not did they pick the right word, right? We always are just grasping at words to use.
[00:53:26] Bruce: We have an idea in our mind, we're grasping at the words to use to try to explain ourselves. And we just pick a word. We have to. We have to pick a word that's close enough. And then we have to kind of hope people will be charitable in trying to read us and understand us and that they're going to try to understand what the point was. So in his opinion, then, a lot of human knowledge isn't creative. That a lot of it, the creativity pre-exists, that knowledge already existed. So this is an interesting point, though, is if you want to say AI doesn't create knowledge, human knowledge isn't knowledge either in many, many cases. And this was Deutsch's point to me. Now, OK, we talked about animal intelligence and Byrne's theory. I really do wish that he would get interested in that because I'm really curious what his answer would be to that. Now, we talked about, so I'm almost done, actually, but we talked about what is it that creates, why do humans have creativity?
[00:54:24] Bruce: Now, one of Deutsch's theories that he mentions in his book is that humans may not be born universal explainers, that they may need to pick up a meme first. They come equipped to pick up memes, and then once they pick up certain memes, then they become universal explainers. Now, this makes a distinction that I think you could argue. You could say, well, something that is an algorithm that is potentially a universal explainer, does it make sense to not call it a universal explainer? I think it depends on what you mean. Let me make a defense, though, of this idea. There are real-life cases of humans being raised without language, raised by monkeys or raised by wolves.
[00:55:06] Unidentified speaker: I mean, that actually has happened in real life at least once.
[00:55:09] Bruce: Or just they had neglectful parents. There was some well-meaning parents once where they didn't want their child... They were deaf. They didn't want their child... Their child could hear. So they wanted their child to learn normal language. So they did not interact with their child with sign language, thinking that that would hinder their ability to learn spoken language. And instead, they tried to put them in front of a TV, which is a completely wrong way to teach child language. It has to be interactive. So the child learned no language at all. And they were effectively severely mentally challenged by the time the scientists came in, the doctors and scientists came in to try to save the day. And from sitting and talking with the scientists, their intelligence changed from mentally challenged to almost normal intelligence as they started to pick up language.
[00:56:01] Bruce: So one of the things that I've wondered about is if the meme that you have to pick up to be a universal explainer is itself language. Yes, I brought that up to David Deutsch and I asked him about that. And he goes, well, I'm in no position to argue with that theory. I'm also in no position to agree with that theory. He says, you know, that's an interesting idea that could be. And then he goes, he goes, Daniel Everett, the linguist, he believes that. I'm like, I've never heard of Daniel Everett. So I'm like looking him up online. And he goes, yeah, Daniel Everett believes that homo erectus evolved language first before speech. And that was what led to kind of the explosion of intelligence, what we might say. So language being the ability to represent things symbolically. So I'm going to have to go maybe read up about Daniel Everett's theories now.
[00:56:55] Bruce: But that sounds really interesting. Sounds very similar to my completely wild speculation that language might be the meme you have to learn to be a universal explainer. My thinking on this is there's a book called, I'll have to look up the name of the book. There's a book about a man who never learned language. And he was deaf. He existed. He came from a Spanish-speaking country and was brought to America. And so he didn't have, when I say didn't have any language, I don't know what that means. I mean, maybe he did have some language in some sense, but he certainly didn't have it in the way you would normally think of it. You could not talk with him about stuff. And the person who wrote the book interacted with him a lot, was a scientist, interacted with him a lot, and writes this book about him. A person who does not have language is not a universal explainer.
[00:57:47] Bruce: They're clearly not going to be able to comprehend anything like a universal explainer. Obviously, this is a trick, the way I'm saying this, because if they learn language, then they can, because they got the potential to be a universal explainer. And that's why I've wondered, I mean, we don't represent things necessarily in spoken language in our minds, but sometimes we do. And I think that the ability to understand language allows us to conceptualize concepts into a word, and then we can think in terms of that concept easily, because now we have a word for it. We've got a single symbolic representation that we can point to that now means that to us. I think it's also why people have such a hard time understanding that words have multiple meanings in different contexts, is because they're so used to using that word to represent a certain concept that they have a really tough time buying this word in a different context, has a different meaning.
[00:58:43] Bruce: And Douglas Hofstadter gives a really good example of this, which I think I've used in a past podcast, where he says, what does the word grow mean? And when you ask someone who's an English speaker, what does the word grow mean? You know, they say it means to change in size to be larger. Almost everyone would say that. And he'd say, well, yes, but sometimes it means just to change in size, not necessarily larger. And people will deny it. They'll say, no, absolutely not. The word grow does not mean to change in size, but not necessarily larger. And so he'll suddenly quote Alice in Wonderland. She ate the mushroom and she grew smaller and smaller and smaller. And then the person's like in shock, like they understand exactly what that sentence means. They've used sentences like that multiple times, and it's never struck them that the word grow has a completely different meaning in some contexts.
[00:59:36] Bruce: And I think that's why people end up dipping into essentialism is because it's really hard to break yourself out of the fact that words don't have meanings, that we use a word as a marker for something else in our mind, a concept. But they're actually differentiable, that every word out there has multiple meanings in different contexts. Hofstadter actually goes so far as to claim that every word has hundreds of meanings, that when you go to a dictionary, and it gives you five different definitions, that it's misleading you on all the different subtle definitions that word actually has.
[01:00:10] Cameo: So my family actually has a game that we play that's a word definition game. It doesn't have any physical manifestation of the game, but we sit and somebody picks a word, and this is in big groups, like 12 people, 15 people. Everybody takes turns defining the word, trying to come up with another definition that nobody else has used yet. And it's fascinating because some words, I mean, there are almost any word that you can come up with, especially words like grow that are adverbs or adjectives, modifying words, have just unbelievable amounts of meanings that are accepted amongst people. And so it's a fun game just to think of how many ways you can think of the same word in different contexts.
[01:01:08] Bruce: Yes. And then Douglas Hofstadter's point is that it's a growing list, that we have a word, it has a meaning, we then find a new analogous situation that's close enough to the previous situation, and we reapply the word, and now that word has a new meaning. And then from that point, we can jump to a new definition. And eventually, you can end up with a completely different definition than the original meaning. And that now this word possibly has all those meanings, or possibly has lost some of its original meanings. It's very fascinating because language has this evolutionary context because of that. Every child and parent has been able to talk with each other, and yet you go enough generations back and they can't talk, you know, if you could take someone from today, and then go back in time, you know, a thousand years, you wouldn't be able to understand each other, even if you're speaking the same language.
[01:02:00] Bruce: They would sound like a completely different language. Old English sounds like Welsh or something to me, right? It's not even recognizable as English to me. By the way, when I say old English, that's very different than, say, Shakespeare's English. Some people might think that's old English, but that's actually considered modern English. We're talking about something further back in time. I had an English teacher who actually played old English for us. It is literally a foreign language, because the language had changed so much. And then she showed us something in between, modern English, Shakespeare, and old English, and it was weird. You could make out what the person was saying sort of, because it was just close enough to modern English, and yet it clearly wasn't modern English. It was something else, right? So old English, you couldn't understand at all.
[01:02:44] Bruce: In between, you could understand if you concentrated really hard, but it was clearly different words, but the words were close enough you could make out what they were saying. And then you had modern English with Shakespeare, which a lot of us can't understand. A lot of us can't understand even modern English, because it's changed so much since Shakespeare's time. But the words have meanings very similar or the same as we use them today, even if they sometimes use words that we don't use anymore. So anyhow, I thought all that was kind of fascinating. I wanted to pursue the idea that language is what creates universal explainers. But I don't know how. We'll have to do a podcast on Hofstadter's theory, because I think that's probably one of the most interesting theories I've heard where he makes that case. He believes that there's this connection between language and what he calls analogy.
[01:03:38] Bruce: I feel like that's misleading, because what we're really talking about is abstraction. The fact that we can abstract concepts, and then we'll reuse a word for it, is this sign that we're able to abstract things. One of the examples he uses is a desktop. A desktop, I'm sitting in front of a desk right now, a physical desktop, a desktop has a certain meaning. Then suddenly computers get invented with point and click with a mouse, and we take the word desktop. There's not that much in common between a computer desktop and a real desktop, but there was enough that when people started talking about a computer desktop, you got the analogy immediately, and you were able to start using that computer desktop, using your knowledge of a physical desktop. You could reapply the knowledge that you had learned about physical desktops, and you didn't require as much instruction, because you would pick up on the analogy, and you would figure out, oh, this is what I have to actually do to do things on this virtual desktop.
[01:04:44] Bruce: And that's what humans do, right? That is kind of our creative element, that we reuse knowledge and abstract it and then use it in new circumstances. What we're doing there, I don't know. If I knew, that would be AGI.
[01:04:57] Cameo: It's interesting, too, to see sometimes how quickly we lose the analogous item. I'm thinking of the... I saw something that a kid had said, oh, look, they 3D printed the save button. And really, it was just a hard disk from back in the 80s. The kid hadn't seen the disk before, doesn't really understand why the save icon looks like a disk.
[01:05:30] Bruce: Right, right, right, right.
[01:05:34] Cameo: And then only says, oh, well, how cool someone would print out the save icon.
[01:05:40] Bruce: Right. You know, that is an interesting point. And Hofstadter gives tons of examples of words that we've lost the original meaning. He talks about, like, the word handsome. That's actually a combination of hand and some. It's hard to even figure out how the original words relate to the current word. I don't even know, right? We've completely lost whatever the connection was. And I think he explains what it actually is, but it's not something I know in any case. And there's a lot of words like that. And he talks about how there's actually degrees. There are some words where we use the word, and if we really stop and think about it, we can figure out what the original analogy was, because the words that make up that word are close enough to words we still use that we can say, oh, you know, I can see that this word originally meant this, and they were combining two words.
[01:06:36] Bruce: And there's other cases where we just use the word and we don't even know what the original meaning was. We've lost it entirely. The word is just itself now at this point. And then there's others where we think of it as an analogy. Like, there are many words that are analogies that we don't think of as analogies. If I were to talk about having a higher IQ, or the idea of up being good and down being bad, that's clearly an analogy if you really stop and think about it. But you don't stop and think about it. To you, up is good and down is bad. That's just the way it is. That's the way we use terms. That's what the terms mean. And yet there's clearly some sort of analogy there way back in the past. And you can still pick it out. You can still understand, well, actually, up isn't actually good and down isn't actually bad, but I can see how those relate to the concepts.
[01:07:30] Cameo: You know, actually, because I've played this, that word game I told you about a whole lot, up is a word that has more meanings in how we use it than—it's baffling the myriad ways we use the word up. And only a very small percentage of them have anything to do with rising or height. It is a very conceptual word within our language. It's a fascinating word.
[01:08:06] Bruce: Yes. All right. Well, that's actually the end of my conversation with David Deutsch. I wish I had recorded it. I think the main conclusion I would take away from this would be, first of all, the concept of creativity as he understands—as he... not understands the term—as he uses the term in some circumstances, is about selecting your own problem. I think that's brilliant, by the way. I think that's a really strong clarifying factor. And I feel like we've kind of resolved that. We've been asking this question for episodes now. Is David Deutsch disagreeing with Campbell, or is he simply using different lingo? The answer is he's using different lingo. I feel like we know that now. It's not that there isn't a valid distinction to be made. There is. But Campbell's theory, the idea of trial and error creating knowledge, is fine. That is a completely valid way of speaking.
[01:09:02] Bruce: The fact that Popper and Campbell use the terms in that way does not in any way say anything about Deutsch being right or wrong. He's just using a different lingo. He's simply concentrating on what he's interested in, which is this element of creativity of switching the problem, of being able to switch one's problem. That's what he's looking at. When he talks about the knowledge is in the genes, if by knowledge you mean the information, it is not in the genes. The vast majority of information from an animal or from an AI comes from its trial and error process. What he really means is all the creative elements, the choosing of the problem, came before that trial and error process. It wasn't part of it. Completely valid. Completely valid. Makes perfect sense to me. So I was glad to get that cleared up. It'll be interesting to... I know I have tons of other questions I wish I could ask him.
[01:09:58] Bruce: We're going to have to collect those. I know that Tracy and I have been talking forever about my questions around heat death. I really want to ask him questions about heat death in more detail. I would have to figure out how to be able to get that down to a compact set of questions, and that's a little harder than it first sounds, but I think I could. So I would really like to collect a bunch of those questions and then invite him on the show and then have a discussion with him. Why does he not believe in heat death? How is he looking at that? I understand some of his theory, but I have many questions about how he's conceptualizing things. Just as an example, his main explanation for why he doesn't accept heat death comes from computational universality. And I do not understand the argument. I think I understand what he's trying to say. I think I can even put it in a very convincing way, but it doesn't seem like a valid argument to me.
[01:10:53] Bruce: I always feel like I'm missing something. And the other one that is... And I've talked to other people who are fans of David Deutsch, and I've had multiple people say, yeah, I didn't get that argument either. So I don't think I'm the only one who feels like I don't get that argument. And the other one that comes up is, it's not super clear why he rejected the Omega Point Theory. Now, I've rejected the Omega Point Theory for other reasons than the ones he gives. He says because the universe is growing, he rejects the Omega Point Theory. I don't see why that would count as a rejection of the Omega Point Theory. And the reason why I say that is because Frank Tipler, who is the one who created the Omega Point Theory, he has given a very good answer to that, namely that we can choose to not let the universe grow, that we've got theories that allow us to decide we won't let the universe grow, which means that our intelligence and our knowledge creation will determine if the universe grows or not.
[01:11:44] Bruce: You can't just reference that as a refutation of the Omega Point Theory. Now, there's other problems with the Omega Point Theory that I feel are really making a problematic theory. I just don't think that happens to be one of them. So I've really been curious about why he rejected that theory, and I'm sure he's got deeper reasons for why he did, and probably just didn't want to go into it in his book. So he referenced kind of the more popular reason, the fact that the universe is growing, that at least throws some doubt on the theory, which I think it does. So I know I've got a lot of questions like that, so maybe we'll have to collect those together and do a future episode and see if we can get him to come on the show and try to ask him some... get some more answers like this, because I feel like it's clarified a ton. I've been struggling with a lot of these topics.
[01:12:32] Bruce: What did Deutsch really mean? I think he means this. And you just ask him and he goes, Oh, yeah, this is what I mean. You know, it's way faster to do that. So anyhow, that is our episode for the day.
[01:12:46] Unidentified speaker: Well, that was fun.
[01:12:47] Unidentified speaker: It was great.
[01:12:49] Unidentified speaker: I'm glad you got that opportunity.
[01:12:50] Unidentified speaker: Really cool.
[01:12:51] Unidentified speaker: Yeah, very cool.
[01:12:53] Tracy: Also, just a happenstance. Since we were talking about Darwin earlier, it actually happens to be International Darwin Day today.
[01:13:01] Unidentified speaker: International Darwin Day.
[01:13:03] Unidentified speaker: Yeah, for real.
[01:13:04] Bruce: Awesome. Well, I don't doubt that Darwin was one of the great minds, right? His theory was a groundbreaking change in the way we thought of things, and goes so much deeper than merely to explain animals, right? I mean, like, there's just, we've seen how it got reapplied to Popper's theory, what it meant with Popper's theory. I think it's one of the strongest reach theories we've ever seen, right? It's an amazing theory. We'll have to do an episode sometime on Leslie Valiant. Oh, I actually talked with Deutsch about that. It was really brief. But we were talking about the fact that Darwin's theory is an incomplete theory. Most people don't know that, right? Most people think of Darwin's theory as this complete theory. But Leslie Valiant, who's a big name in AI, has pointed out that we don't know how to turn it into an algorithm, so we don't really understand it.
[01:13:58] Bruce: And so Leslie Valiant has kind of pioneered this field called computational Darwinism. And it's trying to find what is the computational algorithm that Darwinian evolution is doing. The fact that we don't know it's kind of a big deal, right? It's one of the reasons why I feel like it's so easy to have doubts about theory. If you haven't solved the 100th problem, it's really easy to have doubts about the theory because the theory is so imprecise in so many ways. And Valiant talks about how it doesn't even seem like it's tractable. It says every theory we've tried to come up with that would be similar is completely intractable. Even for billions of years, it's intractable. It says what we really need is we need to find this algorithm that is equivalent to Darwin's theory that's tractable. And says we don't know what it is. It's just an incomplete theory at this point.
[01:14:49] Bruce: David Deutsch and I were talking about that, and I brought up Leslie Valiant. I don't know if you knew who Leslie Valiant was, but I said, oh yeah, Leslie Valiant in his book talks about that in Probably Approximately Correct, that it's just an incomplete theory. And I've actually talked with people where I've said, well, Darwinism is an incomplete theory. And I've had people go, oh, no, it's not. I've had people who are fans of David Deutsch say that to me, who are giant fans of David Deutsch, not realizing that David Deutsch said the same thing in his book. He talks about how Darwinian evolution is this incomplete theory. I have to go find the exact reference, but he really does say that. He's one of the scientists who helped me realize that it was an incomplete theory. And yet, even though it's this incomplete theory, it's this powerful reach.
[01:15:36] Bruce: Sometimes you don't have to have all the details for that explanation to immediately start being powerful. And I think that's a good example with Darwinian evolution. All right. Well, thank you, guys.
[01:15:49] Unidentified speaker: Thank you.
[01:15:50] Unidentified speaker: Thank you, everybody.
[01:15:51] Unidentified speaker: See you next time.
[01:15:52] Bruce: All right.
[01:15:53] Unidentified speaker: Bye-bye.
[01:15:57] Bruce: The Theory of Anything podcast could use your help. We have a small but loyal audience and we'd like to get the word out about the podcast to others so others can enjoy it as well. To the best of our knowledge, we're the only podcast that covers all four strands of David Deutsch's philosophy, as well as other interesting subjects. If you're enjoying this podcast, please give us a five star rating on Apple podcasts. This can usually be done right inside your podcast player, or you can Google The Theory of Anything podcast Apple or something like that. Some players have their own rating system and giving us a five star rating on any rating system would be helpful. If you enjoy a particular episode, please consider tweeting about us or linking to us on Facebook or other social media to help get the word out. If you are interested in financially supporting the podcast, we have two ways to do that.
[01:16:42] Bruce: The first is via our podcast host site, Anchor. Just go to anchor.fm slash four dash strands F-O-U-R dash S-T-R-A-N-D-S. There's a support button available that allows you to do reoccurring donations. If you want to make a one time donation, go to our blog which is fourstrands.org. There is a donation button there that uses PayPal. Thank you.