Episode 123: Campbell vs Deutsch: Incremental vs Cosmic Significance

Links to this episode: Spotify / Apple Podcasts
This transcript was generated with AI using PodcastTranscriptor.
It may contain mistakes. Please check against the actual podcast.
Speakers are denoted as color names.

Transcript

[00:00:00] Blue: This week on the Theory of Anything podcast, Bruce compares Donald Campbell’s evolutionary epistemology and David Deutch’s ideas on infinite knowledge growth. Is knowledge created in human minds an incremental process, as Campbell might emphasize, based on blind variation in selection? Or as Deutch might emphasize, is this a cosmically significant open -ended process based on creativity and explanation? I enjoyed this short conversation with Bruce a lot, and I hope someone else out there gets something out of it.

[00:00:47] Red: Welcome to the Theory of Anything podcast. Hey, Peter. Hello, Bruce. How are you doing today? Good. We’re gonna do another me reading one of my blog posts and then from back in the day and then discussing it. Me discussing it and then you and I can discuss it. Sound good?

[00:01:01] Green: Okay. Well, it’s good to get these ideas out there. Is your blog still up, by the way? No, it got hacked.

[00:01:08] Red: Now, somebody sent me a lot of the old posts. I need to take the time to put them up on Substack or something, but I’ve just never found the time to do it.

[00:01:18] Green: Okay, malicious Bayesian hackers or something. It’s inductivists, for sure. Okay, okay.

[00:01:26] Red: So I’m gonna read a blog post that I published back in October 29, 2020, and it was called Campbell vs. Deutsch, Ubiquitous Knowledge Creation. It was my first stab at trying to make sense of the contradiction between Deutsch’s theory of knowledge and Campbell’s evolutionary epistemology, which is his theory of knowledge. The article is important to me because it marked the point at which I started to realize that Deutsch had made a mistake in his understanding of knowledge. It also played a big role in my slow realization that the modern crit -rack community didn’t accept Popper’s no ad hoc rule and how I came to believe that that was why bad explanations tend to move through that network fairly quickly. Now, in the past, I wrote about how one of my most important disagreements with David Deutsch is on the subject of what counts as knowledge creation. I also noted in passing this is like prior to me writing this blog post. In my post summarizing Donald Campbell’s theories, I made a post where I summarized Donald Campbell’s theories, that Campbell’s views are at odds with Deutsch’s. So, side note, I had written previously a post summarizing Campbell’s theory. Episode 114 is similar, but is far more detailed than that post was. So if interested, Campbell’s theory of evolution epistemology is discussed in depth in episode 114. The key disagreement comes from the beginning of infinity, where Deutsch argues that artificial evolution algorithms do not create knowledge. Note, some will immediately respond, recognize this is the walking robot example that I criticized in episode 75, Deutsch’s theory of knowledge, the walking robot.

[00:03:01] Green: I will read Deutsch’s arguments almost in full here.

[00:03:04] Red: So, Deutsch claims, recall Edison’s idea that progress requires alternating inspiration and perspiration phases, and that because of computers and other technology, it has increasingly become possible to automate the perspiration phase. This welcome development has misled those who are overconfident about achieving artificial evolution and AI. For example, suppose that you are a graduate student in robotics, hoping to build a robot that walks on legs better than previous robots do. The first phase of the solution must involve inspiration. That is to say, creative thought attempting to improve upon previous researchers’ attempts to solve the same problem. This is from beginning of infinity, page 158. Deutsch goes on to describe how this researcher will use existing knowledge to come up with ways to solve this problem. Ultimately, the researcher creates a language of sorts that solves most of the problems involved in the robot walking. When you have identified and solved as many of these sub problems as you can, you will have created a code or language that is highly adapted to making statements about how your robot should walk. Each one of these subroutines is a statement of command in that language. This is all Deutsch still. So, so far, most of what you have done comes under the heading of inspiration. It required creative thought, but now perspiration looms. Once you have automated everything that you know how to automate, you have no choice but resort to some sort of trial and error to achieve any additional functionality. However, you do now have the advantage of a language that you have adapted for the purposes of instructing the robot in how to walk.

[00:04:50] Red: So you can start with a program that is simple in that language, despite being very complex in terms of elementary instructions of the computer. Then you can run the robot with that program and see what happens. When it falls over or anything else undesirable happens, you can modify your program, still using the high level language that you have created to eliminate the deficiencies as they arise. This method will require less inspiration and ever more perspiration. So an alternative approach is open to you. You can delegate the perspiration to a computer but using a so -called evolutionary algorithm. That was, by the way, all Deutsch beginning of infinity pages 159 to 160. Now, Deutsch admits this is a successful technique and that it certainly constitutes, this is a quote now, it certainly constitutes evolution in the sense of alternating variation in selection. But is evolution in the more, is it evolution in the more important sense of creation of knowledge by variation in selection? This is, this will be achieved one day, but I doubt that it has been yet. For the same reason that I doubt chatbots are intelligent, even slightly. The reason is that there is a much more obvious explanation for their abilities, namely the creativity of the programmer. This is why I doubt that any artificial evolution in scare quotes has ever created knowledge. That was page 160 to 161 of beginning of infinity. This idea that existing computer algorithms can’t create knowledge has gone on to become a key part of arguments made by Deutsche and Critrats today. This is still true even today, although it was like really even more true back then.

[00:06:29] Red: Although Deutsche didn’t explicitly state this, it has been become commonly accepted in the Critrat community that quote perspiration has the ability to mimic the appearance of knowledge creation. But in reality, the human programmer created this new knowledge. Citing perspiration has become a way to explain a way seeming knowledge creation by algorithms. There is some difference of opinion in the community as to what Deutsche intended in the quotes above. I will cover other possibilities how to interpret Deutsche in future posts. I don’t think I ever actually did. But the most straightforward reading of the above is that Deutsche is claiming that knowledge creation has never been accomplished by any existing computer algorithm even when they utilize Popper’s epistemology of alternating variation selection of improved variations as a way to optimize and improve the variants over time. As previously mentioned, there are many people in the community that read Deutsche in this way. They make bold claims based on their reading that no existing computer algorithms create knowledge today. So assessing this version, even if it wasn’t what Deutsche intended, seems necessary at this point. I will refer to this view as the pseudo -Deutch theory of knowledge. That’s what I’ve called it in the past in this podcast also. For the moment, though, let’s start with the idea that this is a correct theory and follow it through to its logical conclusions. If this view is correct, then one thing we can conclude is that Donald Campbell’s theory, which is actually just Popper’s epistemology applied more broadly than Popper initially did, must be incorrect. Given Popper’s enthusiastic agreement with Campbell, as discussed in episode 114, by the way, this must also mean that Popper was incorrect about the implications and applications of his own epistemology.

[00:08:23] Red: I believe this therefore potentially represents a massive disagreement among these three great thinkers. The disagreement could be stated as, first, the pseudo -Deutch view. Knowledge creation is rare. Only biological evolution and human minds have currently achieved it. We don’t currently know how to create a knowledge -creating algorithm versus two, Campbell’s and Popper’s view. Knowledge creation is common and ubiquitous. Any process that follows Popper’s epistemology of variations where you can differentiate between better and worse variations over time and you select the better variations is included in what we call knowledge creation. Nature consists, as part of this theory, nature consists of a vast ubiquitous overlapping hierarchy of evolutionary algorithms that are constantly creating knowledge. Computer algorithms that create knowledge are common and well -known. In fact, knowledge -creating algorithms existed before the invention of the digital computer. Can Campbell’s theory be reconciled with Deutch’s? That’s the question I want to ask. Campbell explicitly argues that Herbert Simon’s logic theorist program did create knowledge using his process of blind variation and selective retention. Campbell made this argument despite the fact that Simon himself didn’t believe his program utilized blind variation. This became one of Campbell’s key arguments as to how common blind variation and therefore knowledge creation actually is, even if we don’t always recognize it. So right off the bat, Campbell and Deutch can’t be reconciled on this point because they are mutually exclusive views. Simon’s logic theorist program and the invention of artificial intelligence. Simon’s logic theorist is a program that existed prior to the corning of the term artificial intelligence. As Wikipedia notes, in 1955, when Newell and Simon began to work on the logic theorist, the field of artificial intelligence did not yet exist.

[00:10:22] Red: Even the term itself, artificial intelligence, would not be coined until the following summer. That’s from Wikipedia under logic theorist. In fact, this program was one of the main programs represented at that conference where the term artificial intelligence was coined, leading many to refer to it as the first artificial intelligence program. The program famously uses a search tree to try to find a useful logic theorem. Reviewing the code, I actually did review the code, it appears to be entirely deterministic and non -random, which isn’t uncommon for artificial intelligence algorithms that rely on heuristics as logic theorist does. Certainly logic theorist does, as Deutch allows, follows the logic of variation, the nodes of the search tree, and selection, and thus does follow Popper’s epistemology. This is why Campbell saw it as a clear case of knowledge creation. So Campbell is arguing that narrow AI is often knowledge creating. It was less clear if logic theorist did blind variation or not, as per Campbell’s version of Popper’s theory. But Campbell argued that it did do blind variation because it was based on heuristics that might fail. See episode 114, where we discussed this. Therefore, it was blind in some legitimate sense and was therefore discovering new knowledge using its search algorithm. According to the suit of Deutch view, these proofs were existing knowledge. When, to a little context, logic theorist found unique new proofs that, of things we already knew, but proofs that were simpler and better than the ones that were published by, say, Bertrand Russell. Okay, so it actually created new proofs.

[00:12:06] Red: So according to the suit of Deutch view, these proofs were existing, these new proofs, these novel new proofs were a quote unquote existing knowledge, including the previously unknown proof of theorem 2.85 that had already been injected by the programmer. Yet the program itself contains none of these theorems in any direct sense. But presumably suit of Deutch would argue that the knowledge embodying the program in some unexplained sense already contained this knowledge and perhaps he’d say that the program was simply using perspiration to somehow uncover this already existing knowledge. Now I don’t personally find this a compelling argument, but since we’re currently starting with the assumption that the suit of Deutch view is correct, we’ll accept this argument for now. So logic theorist created no knowledge according to our starting assumption. Even on the elegant new theorem it discovered that no humans previously knew. What are the implications of this view and how are we prepared to accept them? One obvious outcome is that Campbell’s hierarchy of knowledge creation must be entirely wrong, except in the cases of biological evolution and human memes or human ideas, I should have said. Keep in mind that Campbell gives many examples of knowledge creation at various levels of the hierarchy. To Campbell, a paramecium randomly trying to move in various directions and then finally retaining the direction that allows it to move forward is an example of knowledge creation using Popper’s epistemology. But if logic theorist doesn’t count as knowledge creation, as Deutch would have, or at least theory suit of Deutch would have it, it must be that this simple paramecium algorithm doesn’t create any knowledge either. Why?

[00:13:53] Red: Because if it did, it would be trivial to program, and thus we could falsify Deutch’s pseudo -Deutch theory by easily demonstrating that we do have knowledge -creating algorithms that are artificial evolution. In fact, this is true of every example Campbell uses in his essay. All of them, except of course biological evolution and human culture and ideas, the ones that pseudo -Deutch considers to be true knowledge creation are not hard to program. We already today use computer vision as a replacement for locomotion, for example. Campbell included the example of a computer playing chess as an example of artificial evolution. Even seemingly benign ideas like the idea that B -language is a vicarious selector, this is from Campbell’s theories, for motion can’t be considered knowledge creation or we could easily program something equivalent to it and thus falsify the pseudo -Deutch view. Therefore, pseudo -Deutch, to be correct, all of Campbell’s theory must be false. It must be the case that knowledge creation through Popper’s epistemology is not ubiquitous, like Campbell and Popper thought was the case. So it would seem that we are forced to choose between the pseudo -Deutch view and the Campbell -Popper view of knowledge creation, because they can’t both be correct. Is knowledge creation common, like Popper and Campbell believe, or is it rare, and we don’t really understand it yet, as Deutsch or pseudo -Deutch believes? But note one, as a nod to those that believe Deutsch meant something different than the straightforward reading, I’m going to call this theory I’m about to discuss the pseudo -Deutch theory of knowledge.

[00:15:35] Red: In the original version of this post I instead just used the term Deutsch’s view in scare quotes to emphasize that we’re making the assumption for the sake of the post that the straightforward reading of Deutsch is the correct one. But what I found was that people were usually not paying attention to my actual argument and instead were trying to defend Deutsch’s honor, usually arguing that this is not his actual view and he meant something else. So I’m now calling it the pseudo -Deutch view. By calling it the pseudo -Deutch view, I’m hoping to avoid that problem. I’m referring to it as such based on something from biblical scholarship. Bible scholars have some letters from Paul that they believe Paul wrote and some that they are not sure about. The ones that they’re not sure about are sometimes called the pseudo -Pauline letters. So I’m leaving open the possibility that this is or is not Deutsch’s actual view. But it should be more obvious than mere scare quotes that I’m not claiming to know for sure that this is Deutsch’s actual view. End of footnotes. Update to that note. This is me today. Today there is still a bit of doubt around what Deutsch’s intentions were around how crit -rats cite perspiration as a way to discount apparent knowledge creation by say AlphaGo inventing a novel new way to play Go. See episode 34, AlphaGo and creativity. A Deutsch and crit -rat will simply slough off a refutation like this and say, oh, that’s just perspiration, not actual knowledge creation. Whereas Deutsch will admit that maybe AlphaGo does create some limited knowledge.

[00:17:05] Red: But most of what I’m calling the quote pseudo -Deutch theory of knowledge ended up getting added to Deutsch’s two sources hypothesis, which is now clearly considered central to his theory of knowledge today. End of that note. This was the end of the original article. For those that think I’m wrong when I claimed Campbell used chess playing as an example of artificial evolution, here’s the actual quote. In biological evolution and in thought, the number of variations explored is greatly reduced by having selective criteria imposed at every step. Thus mutant variations on non -adaptive variations of the previous generation are never tested, even though many wonderful combinations may be missed therefore. Some of the heuristics currently employed in logic and chess playing machines have the similar effect of evaluating all next possible moves in terms of immediate criteria and then of exploring further variations upon only those stems passing the screen of each prior state. That’s from page 106 of the book Evolutionary Epistemology, Rationality, and the Sociology of Knowledge. So Campbell did consider chess playing machines as an example of this artificial evolution that he’s talking about, this blind variation in selective retention, evolution epistemology that he’s arguing for. There is an interesting comment made on this blog post by Ella Hopner. I’m going to discuss my disagreements with Ella in a future podcast over how to interpret Campbell’s theory, but this particular comment is one I strongly agree with her on and I think is really relevant to the problems of the whole discussion and how to actually solve the problems. She says, I agree there is a contradiction between Campbell’s and Deutsch’s view of knowledge and I come down on Campbell’s side with this disagreement.

[00:18:53] Red: I’m convinced by Campbell’s identification of knowledge creation with blind variation in selective retention and this means that knowledge creation processes are ubiquitous. What are ubiquitous, however, are open -ended knowledge -creating processes. Most knowledge -creating processes, examples, vision, logic, Simon’s logic theorist, and many modern machine learning algorithms are severely limited in the kind or amount of knowledge they can create. Open -ended knowledge -creating processes, by which I mean processes which can go on creating new novel knowledge in a wide domain endlessly or at least for an arbitrarily long period of time, are quite rare. The only two examples of open -ended knowledge -creation that we know of are biological evolution and the human mind. No artificial process has yet managed open -endedness, though this is the goal of the field of artificial life and artificial general intelligence. Note how Ella properly understands Deutsch’s argument as confusion between open -ended knowledge -creation, the problem of open -endedness, as discussed in episode 74, by the way, and knowledge -creation in general. And this is the argument that I have made throughout this podcast is that he’s confusing those two things.

[00:20:03] Green: I think that’s a pretty satisfying way of removing the goalpost. That makes perfect sense to me that really what’s special about human knowledge is the open -endedness of it. I mean, that’s just another thing you’ve 100 % convinced me on. But you know, Deutsch did repost that podcast that we did on it. And I mean, it’s not clear to me that even though he doesn’t use that language, it’s not clear to me that that’s not basically what he’s saying is that, I mean, we can nitpick about what knowledge is and what isn’t, but that, you know, that it’s open -ended knowledge that really is special in the world. You know,

[00:20:57] Red: it’s not too hard to show Deutsch coming down on both sides of the argument in various Twitter posts. And I think that is part of what makes this also confusing.

[00:21:11] Green: Okay.

[00:21:13] Red: The reason why I just read, I’m going to go on to give a discussion now and I’m going to actually show that this is a deeper problem than it first appears. But I am not arguing that I know what Deutsch’s exact theory is. And I’m therefore not arguing against quote, unquote, Deutsch’s theory. I’m arguing against a certain interpretation of Deutsch’s theory, which I’m calling the pseudo -Deutsch theory of knowledge. I do think that the pseudo -Deutsch theory of knowledge takes a lot from things that Deutsch absolutely did say and that there’s a certain common way, easy understanding of what he says that is equivalent to what I’m calling the pseudo -Deutsch theory of knowledge. But I do think you also find counter examples within Deutsch. So it’s always a little unclear what his view is. Let me say that I completely disagree with his two sources hypothesis that when he says that these are the only two sources of knowledge, the only sense in which I could accept that is if I understand the word knowledge to mean explicitly open -ended knowledge. If you mean by knowledge, open -ended knowledge, then of course, tautologically, you’re now correct that there’s only two sources of knowledge. And I’ve brought this up in past podcasts. I’m going to now make a series of arguments where I’m going to show even though that’s true, this isn’t just pedantic. There’s a very real problem here. Okay. So back when I wrote this post, I didn’t fully understand the degree to which Deutsch’s theory was problematic. And I was simply trying to show that there was this disagreement between Popper and Deutsch, Campbell and Deutsch, that needed to be attacked and resolved in some way.

[00:22:58] Red: And I was very naive back then. I was very new at all this. I honestly thought that there was an obvious answer that Deutsch had simply confused. Eventually I thought this. I got back when I wrote this article. I’m not sure I had thought this through all the way yet. But in conversations with Ella and others, I helped clarify my own thoughts. And I started to realize there’s an easy answer here. He’s confusing knowledge and open -ended knowledge, which means he’s confusing physically what knowledge is and the process that creates knowledge. I honestly thought I could just explain that to people and they’d go, oh, you’re right. You’re properly capturing the good part of what Deutsch was trying to say, but you’ve clarified it. That is absolutely not what happened. Not at all, other than someone like Ella who clearly agrees with me on this. There are a few exceptions. There was a complete shutdown to this idea within the crit -rat community when I tried to discuss it with them. And that is an important part of my own journey here. The fact that something that seems to me to be a problem but with an easy solution was complicated into an increasingly larger problem over time. I didn’t miss that fact. That was what was going on. Let me see if I can explain this a bit further as I go forward here.

[00:24:13] Blue: Okay.

[00:24:15] Red: Consider these quotes from Deutsch that I read a moment ago. Quote, once you have automated everything that you know how to automate, you have no choice but to resort to some sort of trial and error to achieve any additional functionality. And then another quote, it certainly constitutes evolution in the sense of alternating variation selection. But is it evolution in the more important sense of creation of knowledge by variation in selection? This will be achieved one day, but I doubt that it has been yet. So Deutsch is not denying that this is a form of trial and error, nor is he denying it is a form of variation in selection. Now, let me earmark this because it will become important in a future podcast when I explain that it’s hackathols and veins disagreements with Popper and Campbell over evolutionary epistemology. Their arguments differing in a significant way from Deutsch’s disagreements with Popper and Campbell. But let’s be clear. The genetic programming algorithm that allows a robot to walk does do trial and error problem solving via an algorithm of variation in selection. This isn’t in doubt in Deutsch’s mind, according to what I just quoted him saying. Deutsch does not doubt this by his own admission. Yet despite this, Deutsch claims it is not knowledge creation. Again, let me be clear. You have to make these things explicit or they slip out of your mind and you can’t criticize them. Deutsch is saying artificial evolution does solve real problems such as the walking robot or as we’ll see the immune system and does so via trial and error or variation in selection. At the beginning of the genetic programming algorithm, the robot can’t walk. At the end of the process, it can walk.

[00:26:02] Red: So a problem was solved. Or using the example of the immune system, another one that I’ve debated extensively with the quit route community. The immune system discovers a previously unknown recipe for an antibody that kills off a novel pathogen it’s never seen before. So it solved a real -life problem using a solution that it had to invent. But Deutsch is arguing that the solution found in these problems are not knowledge. But then what is it? Something got created here. No matter what you personally want to call it, the word does not matter. Moreover, what is the relationship of this, whatever it is that got created to poppers epistemology? Most crit rats I know sincerely believe that AI and machine learning have no relationship at all to poppers epistemology. And I’ve had Deutsch and crit rats extensively argue to me that the immune system creates no knowledge because Deutsch said that was the case in an interview and because it’s quote -unquote perspiration and perspiration does not create knowledge, only inspiration does. But consider that the genetic algorithm literally attempts different solutions and keeps the portion of the code that works and lets the parts that don’t work that are less useful die off. And keep in mind the immune system literally does a miniature version of natural selection to find the right antibody, trying out different variants and then allowing only the successful ones to replicate. This is more than a passing resemblance to poppers evolutionary theory of knowledge. There is something weird going on here, the way this is getting instantiated in people’s brains.

[00:27:49] Red: Deutsch does not seem to be disagreeing that the immune system and the walking robot algorithm are doing something clearly Popperian, namely finding solutions to problems via trial and error or variation in selection of some sort. He simply doesn’t happen to call it knowledge. But isn’t poppers epistemology specifically about how knowledge grows? Popper says this is from logic of scientific discovery in the introduction, the central problem of epistemology has always been and still is the problem of the growth of knowledge. Doesn’t even Brett Hall call himself Toke teacher? TOK is short for theory of knowledge, precisely because poppers epistemology is explicitly a theory of knowledge growth. Yet Deutsch is seemingly arguing that there are some applications of poppers theory of knowledge that for reasons not fully explained do solve real problems via variation in selection but aren’t a growth of knowledge but instead something else. Something else that never actually gets explained or even given a name. Note that we’re not talking about say a failure of evolution epistemology here. That would be a different case. This isn’t a case of doing variation and selection and nothing comes from it. The robot starts out not walking, the trial and error algorithm runs and then it walks at the end because the best variant is selected and retained or the immune system starts out without the knowledge of how to create the antibody to kill the novel pathogen and by the end of the variation and selection algorithm it knows how to kill the pathogen. So this is a successful application of poppers evolutionary epistemology where we do in fact find a solution to a problem via trial and error.

[00:29:34] Red: So what are we going to call this apparently not knowledge that gets created by the genetic programming algorithm or by the immune system? There is something obviously interesting here that needs to be connected to poppers epistemology in some way. Shouldn’t we become curious and interested in understanding what this not knowledge actually is and how it relates to poppers epistemology? Yet the critrack community takes no interest at all in this strange phenomenon because to them it’s not knowledge and thus uninteresting. I’m going to argue that the critrack community is making a clear cut epistemological error one that we want to learn from. So let’s introduce two new rational fallacies that are interrelated. I’m going to call the first one argument by tautology. A tautology is a statement that is true by necessity. Popper points out that tautologies when treated as theories have zero content so they make bad theories. This is from objective knowledge. He says, we may describe the consequent class of tautological statements as the zero class so that tautological statements have zero content. Since we can choose to define words any way we wish it is always trivial to quote -unquote, win an argument by simply defining one’s terms in such a way that one is necessarily correct. But the cost of this is that your theory now has zero content. A price of paparian should not find a fair exchange. When you fall into this rational fallacy you may even sincerely feel you’ve made a fair rational point when in fact you’ve really said nothing at all since your conclusion is now built into the assumptions of your definition. A famous crit -rat example of this fallacy is defining creativity as explicitly the human ability to create new explanations.

[00:31:23] Red: Then claiming AI is quote -not -creative completely missing that the definition now guarantees the conclusion. Now of course it’s never put as straightforward as I just put it, but it would be put like this, which is really the same thing. For example, a crit -rat might argue AlphaGo is not creative. Now to popper a word is just a label that we use as a shorthand for some concept or theory. Deutsch used creativity as a shorthand for quote, the capability to create new explanations. That’s a quote straight from beginning of infinity. But what counts as an explanation? Can animals create explanations? Now according to Richard Burns well corroborated research the answer is yes. In experiments two elephants know how to work together to pull chains to obtain food, demonstrating that they understand the physics of the situation, not just simple association learning. By the way,

[00:32:11] Unknown: Burns

[00:32:11] Red: is careful in the design of the experiment so that the elephants cannot learn the prior association. They’ve never been given a chance to prior to the experiment. Does this refute Deutsch’s theory? Well, not according to crit -rats. Instead, they redefine explanation subtly to mean the kind of open -ended explanation humans can do but animals cannot. By this definition, AlphaGo is now necessarily not creative. Making the entire claim circular. This illustrates the fallacy of argument by tautology. Also considered the term creative to find this way now makes a biological evolution also not creative. Oops. So of course we now need a second definition of creativity that points to whatever biological evolution is doing that is not creative. So adjusting the definition to try to get out of calling AlphaGo creative all that really does is force you to make a second definition of creative that allows biological evolution to be a kind of creativity. Now as Dworkisch -Pertell put this in his blog post arguing against Deutsch on this, he says if a non -creative AI earns you a million dollars by inventing a new drug, you’re still a million dollars richer regardless of whether you decide to call it. And if AlphaGo creates a new playstyle for go, sure under your narrow definition of the word creativity it is not tautologically creative but who cares? It still created a new playstyle for go which is presumably what the term creative normally means in the first place. So this is why argument by tautology gains you nothing. At best it causes you to fool yourself with words. What do I mean by fool yourself with words? This is the second rational fallacy we’re going to discuss today.

[00:33:57] Red: Consider what happens when we choose to define knowledge the term so that genetic programming algorithms or immune system somehow don’t count as creating knowledge under the definition. For example in real life when I raised this with Dworkisch -Pertell apologists, they might respond but the program created by the genetic algorithm doesn’t persist for millions of years like DNA does. I understand what they’re trying to say when they’re saying this but that wasn’t actually part of Dworkisch’s theory and in fact it’s easy to point out they’re adding something ad hoc to Dworkisch’s theory as ad hoc save and it’s easy to point out well actually most human knowledge doesn’t last that long either. Does that make human knowledge not knowledge? Now clearly this is unacceptable and they will continue to keep adding arbitrary criteria to their definition until genetic programings and immune systems by tautology don’t create knowledge and human human knowledge and biological DNA knowledge do count as knowledge. Yet no matter how you gerryman the definition of knowledge a walking robots genetic program still contains something that causes it to walk or an immune system to fight a pathogen so the Krutrat theory still has a major explanation gap at Krutrat making this argument may sincerely believe they’ve defended their theory but what they’ve really done is they’ve cloaked the problem whatever the robots genetic program is they simply deny it a label out of sight out of mind. I call this rational fallacy cloaking meaning defining terms so that inconvenient observations or problems have no label. Without a label it becomes difficult to discuss the problem and it can be forgotten entirely. As with all rational fallacies there is an objective mistake being made.

[00:35:49] Red: It is undeniable that the immune system and the walking robot genetic algorithm creates something that requires explanation. That is an objective fact. Therefore the pseudo -Deutch theory of knowledge has at a minimum a significant explanation gap. That something the genetic programming causing the robot to walk or the immune system to fight pathogens is right in front of me. I can’t pretend it doesn’t exist. This fallacy is easy to spot by the use of the term not. AlphaGo is not creative and the immune system or the walking robot algorithm supposedly produce not knowledge but nowhere will you find a discussion of what it does create. It will be deemed uninteresting and not worthy of discussion. So defenders of the pseudo -Deutch theory of knowledge always stop there as if there’s a boundary they refused to cross. My argument is that the pseudo -Deutch theory of knowledge has caused the crit -rat community to think poorly around Popper’s evolutionary theory of knowledge which Popper intended to be a much farther reaching theory. To Popper and Campbell all seeming inductive achievements was to be considered actually a form of Popper’s evolutionary theory of knowledge. That’s a bold, wild claim. Indeed it is in a future podcast I’m going to argue that the pseudo -Deutch theory of knowledge has turned many crit -rats today into crypto of inductivists. That’s an argument for another time, however. Be warned here. Don’t make the mistake of taking my criticism of the pseudo -Deutch theory of knowledge as a view that I know it to be wrong or refuted. Indeed, in its current form it is impossible to refute it. So I’d never make such a claim.

[00:37:22] Red: In fact as far back as episode 26 I pointed to examples that might prove Popper’s evolutionary theory of knowledge wrong. For example I pointed out that the normal equations can be used to induce a machine learning model from observations without any sort of trial and error or variation in selection process. This does seem to me to outright refute Campbell’s and thus Popper’s evolutionary epistemology. So perhaps there is something to the pseudo -Deutch theory of knowledge. Who knows? I don’t know. And I won’t know until the problem is solved. What I’m arguing is not the pseudo -Deutch theory of knowledge is wrong though it may well be wrong. But rather that it is in current form a very bad explanation and will remain so until its defenders take the problem with the theory seriously. But this is also an argument for a future podcast. That’s actually the end of my initial comments. Any questions or comments about that?

[00:38:19] Green: Well my only comment is that you’ve raised a lot of provocative questions about this. You’re thinking in the right way. I believe asking questions, criticizing you know we’re not part of a cult. I hope I’m not part of a cult. I never signed up for that. I think the correct Popperian attitude should be to criticize our best theories and move closer to truth together. So I thank you Bruce.

[00:38:57] Red: All right, thank you.

[00:39:05] Orange: Hello again. 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 -Deutch theory of knowledge. We believe David Deutch’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 B Nielsen 01. Also please consider joining the Facebook group, The Many Worlds of David Deutch, where Bruce and I first started connecting. Thank you.

Links to this episode: Spotify / Apple Podcasts

Generated with AI using PodcastTranscriptor. May contain mistakes; please verify against the actual podcast.