Episode 81: Easy to Varyness vs Ad Hocness

• 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

[00:00:07]  Bruce: Welcome back to The Theory of Anything podcast, eh, Peter?

[00:00:10]  Peter: Hello, Bruce, how are you doing today?

[00:00:12]  Bruce: Good, all right. Well, we have been on a run here of some of, I don't wanna say, I think it's some of the funnest episodes for me. I don't know if they're the funnest episodes for people at home, but we've been talking in depth about the Theory of Knowledge, Constructor Theory of Knowledge. And all of that was actually just a setup for, I mean, it was all stuff in its own right that I wanted to talk about and get off my chest. But in some ways, it was a setup for today's episode, which is gonna be about the problem of easy to varyness. Some of the difficulties of trying to sometimes make sense of what easy to varyness is. And I'm gonna make a proposal on how to improve on the concept. We're gonna start with a real life set of arguments that took place between several of us online critical rationalists. Now, for short, let me just use the term crit-rat and define that.

[00:01:15]  Bruce: Crit-rat is that critical rationalism is Popper's epistemology, okay? So he called it critical rationalism. And crit-rat would be the name of a community that exists especially on Twitter, but also elsewhere on Facebook, et cetera, of people who have adopted aspects, or at least claim to have adopted various aspects of critical rationalism as kind of central to their way of thinking. Now, one of the things that I have argued throughout this podcast is that sometimes the twain don't meet very well, that the crit-rats aren't doing critical rationalism. And we're gonna now start to consider that a little bit as to why it is that sometimes it seems like they've departed from critical rationalism, even though it's like part of their name, crit-rat. What is it that's going on? And I think a part of it, of the mystery, is this concept of easy to varyness, which is why we're going to explore it a little.

[00:02:15]  Bruce: So I'm gonna use some real life and I'm gonna be quoting actual crit-rats, arguments that I have participated in, not always between me and somebody else, sometimes it's like a group of people. And it's gonna cover various subjects. And obviously, I'm picking ones that I feel like are relatively obvious, but you know, Peter, if you disagree, you know, say so, and we'll talk about that, okay?

[00:02:38]  Peter: Well, I'm excited. I feel like you're gearing up to maybe make a few more, let's say enemies or something here, but hopefully people will take it well within the spirit of critical rationalism.

[00:02:53]  Bruce: I hope so. I hope so. So we're gonna, I'm not gonna use this guy's real name. I'm gonna call him Dean. And Dean and Jose Chalupa, who that is his real name, were arguing about whether critical rationalism or Popper's epistemology and libertarianism were compatible. And at some point, Dean got angry, which, you know, we all get angry during online discussions and we sometimes let the insults fly and we shouldn't, but we do. And we all do it. I do it. Everybody does it. It's not the end of the world. Okay. So Dean went for an insult. It was a graphically referenced part of Jose's body. I won't say what it is because I want to keep this a non-explicit podcast. And I called out Dean for using insults instead of engaging in critical discussion. And what followed was a series of arguments from Dean explaining and justifying his use of insults, of what he says is intolerance.

[00:03:49]  Bruce: Okay, that's his term, not mine. Justifying his intolerance. And I'd like us to assess Dean's arguments because I think they're noteworthy for a variety of reasons, okay? Here's Dean. He says, look up paradox of tolerance. There is no need to be tolerant to those who love violence. Okay, so me, Dean, you're literally defining love of violence as any view that isn't mine. Is that really what Popper had in mind with his epistemology? Dean, no, that isn't how I define love of violence. People may disagree with me for other reasons, term other reasons we're going to have to come back to. And I don't describe them as loving violence. I only describe people who love violence as loving violence. Now, what does Dean mean by other reasons other than loving violence, right? Okay, so me, I've been following the conversation and what he's done, what Jose has done, is disagree with you over politics.

[00:04:41]  Bruce: He has never advocated for violence even once. At this point, you may want to see some context as to what the argument was over. So here it is. Going back just before this part of the insult and that part of the discussion broke out, Jose says, Popperian libertarian is a misnomer. Libertarians are against our formal political institutions, but epistemology tells us that they're the only known way to solve problems without violence. So improving them, not demolishing them, is the only option. Okay, so let's be clear. Jose is arguing that our democracies, open societies are formal political institutions that are the currently best known way to resolve problems without violence. That is his argument. Thus he believes we must not abolish these institutions, but instead we should improve them. This is what Dean believes is equivalent to loving violence.

[00:05:34]  Bruce: So Dean, in response to this, says, the reason we shouldn't demolish our political institutions is that changes should be piecemeal because they're— we're fallible. Not that they solve problems without violence, again, they are highly violent. In fact, they're embarrassingly incompetent at being peaceful, claims Dean. Now I'm gonna read that again. He uses the word they several times, and I'm going to replace it with what in context was actually being referred to. Dean is saying, the reason we shouldn't demolish our political institutions is that changes should be piecemeal because we're fallible. Not that our democratic political institutions solve problems without violence. Again, our democratic political institutions are highly violent. In fact, our democratic political institutions are embarrassingly incompetent at being peaceful, claims Dean.

[00:06:26]  Bruce: Jose at this point says, these two reasons are interchangeable. Dean then goes for the insult. Now, Peter, I'm gonna ask you some questions here about this argument so far. So first of all, I mean, it's okay if you got an opinion on this subject. Who do you think's correct, Jose or Dean here?

[00:06:44]  Peter: Well, I got a little lost to who's arguing for what, but I think in general, they seem pretty compatible. One of the things that I think, maybe it was Brett Hall who described himself as a incrementalist libertarian. And that really rang true with me. I like the idea of incrementally changing institutions and moving towards a more free society.

[00:07:13]  Bruce: So you're right that they're, yeah, they're both arguing for incrementalism. So that's kind of an interesting point. So it looks to me like what they're really disagreeing over. They're actually agreeing as far as I can tell. Jose says, Popperian libertarians is a misnomer. Libertarians are against our formal political institutions. So it's true that Dean and Brett favor incrementalism, because they're Popperians. But libertarians, what defines a libertarian? What makes one a libertarian? Nobody would mistake me for a libertarian, even though I'm strongly in favor of the free market. And I'm actually quite open to the possibility that libertarians will someday turn out to be right, that we will get rid of public governments and will replace all government services with privatized government. I don't know for sure. I don't think they know for sure either.

[00:08:05]  Bruce: I think that when they claim they know for sure that they are wrong about that. That's really what separates me from a libertarian. Libertarians claim to know what that final system must look like, where I claim to not know what that final system must look like. Jose is really taking the same stance as me. He's saying Popperian libertarians is a misnomer. He's taking issue with libertarians such as Dean, who claim that we already know that the final system must eventually do away with governments, okay? And Dean's making a stronger claim than that. He's claiming that democratic political institutions do not play a role, that they are highly violent, and that they do not keep the peace, okay? Is that true, in your opinion?

[00:08:52]  Peter: Well, I think when, you know, some of this depends on your definition of libertarian, you know, I think as the term was first put out there, and I'm, geez, who is the guy's name who invented the term?

[00:09:09]  Bruce: It also means something different in Britain than it does in America. In fact, it's drastically different between the two.

[00:09:16]  Peter: My understanding is that it means something closer to anarcho-capitalism.

[00:09:20]  Bruce: That's right, we're really using it to mean, so it doesn't have to mean anarcho-capitalism, but in this case, it does.

[00:09:27]  Peter: Okay, I mean, I kind of like the, I like the libertarian idea in the sense that it's someone who tends to be very, appreciate the free market and what it's done for prosperity and all that. But then, you know, some of the more extreme libertarian ideas sort of lose me, I guess, from the perspective of incrementalism. So I don't, I don't know, I'm a, I consider myself, I don't like the idea of tearing down institutions or anything like that. I'm sorry, what was the question one more time?

[00:10:02]  Bruce: Well, so what, what they're really disagreeing over is whether democratic governments play no role in reducing violence. And if they in fact increase violence overall, if we got rid of them, violence would go down.

[00:10:14]  Peter: That doesn't ring true for me.

[00:10:15]  Bruce: Okay, so you're not, you're closer to Jose on this than you are with Dean. Okay.

[00:10:21]  Peter: I could be.

[00:10:22]  Bruce: Okay, now let me just say though, that maybe let's keep an open mind on this.

[00:10:27]  Unidentified speaker: Okay.

[00:10:28]  Bruce: Here's the real question though. Is Jose's argument equivalent to loving violence?

[00:10:34]  Unidentified speaker: No, no.

[00:10:35]  Bruce: No, okay. So you definitely disagree with Dean over the idea that Jose was in any way loving violence.

[00:10:42]  Peter: Yeah, that seems a bit over the top.

[00:10:45]  Bruce: Okay. So in any case, so let me just say that's exposed my own biases here. I absolutely think democratic governments play an absolutely giant role in the existence of open societies and reduction of violence. And in fact, democracies are the only way we currently know how to do it. It may be that someday we can do away with democracies and replace them with something better. I'm not against that idea, but we don't know how to do that today. Okay, the knowledge to do it does not exist today. Okay, so this is really what Dean and I are now disagreeing over. Is what Jose's argument, is it equivalent to loving violence? Okay, and Dean is arguing, yes, this is equivalent to loving violence, and I therefore now get to behave intolerantly towards Jose. Okay, now here's the rest of the argument, okay, that I'm now arguing to Dean that he's misinterpreting Popper's Paradox of Tolerance, me.

[00:11:49]  Bruce: It's true that you are disagreeing over which political system contains the most violence, but that is not the same as loving violence. And now I go on to argue to him, I say assume that you're right, Dean. Wouldn't that mean that Jose is just mistaken about which political system leads to the most violence rather than that he loves violence? For Jose to love violence, he'd need to first agree with Dean that democratic institutions play no role in keeping the peace, and then knowingly decide, but I love violence, so I'm in favor of democracies anyhow. Now, it seems to me that Dean is arguing that it is the manifest truth that any support for a democratic government as a way to keep peace must be insincere because it's just obvious that they don't keep the peace. So therefore, Jose must be loving violence. Recall, though, that Dean is a critical rationalist and that he spent considerable time reading Popper and Deutsch and learning about epistemology and he's made critical rationalist epistemology a huge part of his life.

[00:12:51]  Bruce: So, of course, at this point, Dean says, oh my gosh, you're right. I'm making a bad argument here and I'm misunderstanding Popper's paradox of intolerance. Thank you for helping me error correct. He then apologized to Jose and the discussion continues.

[00:13:06]  Peter: Except that didn't happen, really.

[00:13:07]  Bruce: It did not happen. No, it absolutely did not happen.

[00:13:10]  Peter: I've been online too much. I know that did not [bring you true/bring you through].

[00:13:13]  Bruce: Okay, here's what actually happened. Dean, libertarians don't want institutionalized violence. Jose is defending institutionalized violence. That isn't that surprising. The vast majority of people love violence and can't imagine living in a society without it. They also rush to defend institutionalized violence. So Jose is in good company. Now let me reread that because again, he's using the term institutionalized violence when he really means democratically elected governments, which he sees as causing violence. So what he's really saying here is libertarians don't want institutionalized violence when he... Sorry, let me read it with the replacement. Libertarians don't want democratically elected governments which only cause violence. Jose is defending democratically elected governments which only cause violence. This isn't that surprising.

[00:14:09]  Bruce: The vast majority of people love democratically elected governments and can't imagine living society without it. They also rushed to defend democratically elected governments. So Jose is in good company. Now, the two readings I gave there are gonna come across drastically differently even though all I'm doing is replacing the word institutionalized violence with what he actually meant, right? Democratically elected governments.

[00:14:35]  Peter: Well, I mean, it's such a straw man. I mean, I love violence if it means keeping criminals off the street. And crime just causes so much damage to our society and especially the most vulnerable people. I tend to be, that's one area where the libertarian ideas kind of lose me when people get into this anti-law and order thing.

[00:15:02]  Bruce: Now, here's the interesting part though. He says, the vast majority of people love violence and can't imagine living in society without it. They also rush to defend institutionalized violence. So Jose is in good company. Now it seems to me that Dean just basically admitted I was completely correct. He is in fact interpreting Popper's paradox of violence to include anyone that believes governments are a net positive. Or in other words, he's saying all non-libertarians love violence. So me, isn't that what I just said? That you are literally defining anyone that politically disagreed with you, you are admitting nearly everyone as loving violence and thus worthy of intolerance. Dean, clever, but no. They can disagree with me for other reasons. He uses that term other reasons again. And make no comment on violence. Well, what does other reasons here mean?

[00:15:49]  Bruce: In context, it apparently means that you can argue with Dean any argument other than that democracies play an important role in how we keep the peace. And he'll not accuse you of loving violence. And treat you tolerantly. Otherwise, he's gonna treat you intolerantly. So me, you literally just said, and I quote, the vast majority of people love violence. Help me understand how that doesn't equate to you allowing yourself to be intolerant to the vast majority of people via the paradox of tolerance. I'm requesting your interpretation of Popper's view of the paradox of violence. Or whether I'm telling you that you are misinterpreting it. I don't see how your interpretation is anything but equivalent to merely being intolerant. Let's now discuss Dean's argument. Is it a good explanation or is it a bad explanation? Why or why not?

[00:16:40]  Peter: Truthfully, that would have been, I would have stepped out of this conversation a few minutes ago. I don't know. I've spent thousands and thousands of hours discussing things with people online. And I really rarely have that much of a problem with people. That's just because when someone starts going down a path that seems so over the top, I really think it's just better to bow out of the conversation. I don't know. That's just never... Critical rationalism should be more about finding truth together, right? Not like getting into this all insulting and straw manning and all that. It's just, I've got too much else I wanna do in this life.

[00:17:28]  Bruce: All right, fair enough. That's very wise advice. Okay. Let me ask you this though. What is, according to Deutsch, a good explanation?

[00:17:37]  Peter: Well, something that is hard to vary.

[00:17:41]  Bruce: So, is his argument easy to vary?

[00:17:44]  Peter: I suspect so.

[00:17:48]  Bruce: Okay, you say suspect so. And this is actually the point I'm trying to make. It's a little bit harder than it first appears. It seems obvious at first. But what exactly is he easily vary here? Right. What is getting easy? What is it about his explanation that is easy to vary? Could you tell me? Like, you probably could come up with something, but it's maybe not super obvious really at first.

[00:18:14]  Peter: So the point is that it's not hard. To vary, but it's still a bad explanation. Is that where you're going with this?

[00:18:23]  Bruce: You know, I'm not sure I'm definitely going somewhere with this. I mean, I am definitely going somewhere with this. I'm working towards my proposal of what I think might be a more obvious way. The point I'm really making is that it's a little hard to tell. Okay. That Easy to Varyness at first seems obvious, but when you actually try to apply it to arguments, it turns out to be way harder to figure out if it applies or not.

[00:18:50]  Peter: I think that's a fair point.

[00:18:52]  Bruce: So you said exactly what I hoped you would say. Yes, this does not ring like it's a good explanation to me. And yet a good explanation is supposed to be hard to vary. And it's not clear that it's easy to vary. Okay. Does that mean it is a good explanation? I mean, I would be really hard pressed to ever say it was a good explanation. Okay. Now, let me ask you the opposite of this. Have I refuted his argument?

[00:19:19]  Peter: In Bruce Nielson's definition of refuted or how most people use it?

[00:19:24]  Bruce: I am intentionally being vague as to what I mean by refuted. Actually think about that for a second. The word might mean different things. Have I refuted his argument? What could I possibly mean by that? In what sense maybe could I have or not have?

[00:19:40]  Peter: Well, I mean, I think probably in your own brain, which I think is pretty, not that that's bad, you've offered a refutation. I mean, is it definitive? And Dean, right? Dean's brain? Now I'm curious if I know this guy. He's probably someone I'm friends with. But yeah, I don't know. What do you think?

[00:20:02]  Bruce: You know, I don't know the answer to that question. I think the word refutation is so vague.

[00:20:06]  Unidentified speaker: Yeah.

[00:20:07]  Bruce: I'm really unclear. I mean, I think if you define refutation the way Popper did, as I did an actual experiment to show that he was wrong, I think I have absolutely not refuted Dean.

[00:20:19]  Peter: Okay.

[00:20:21]  Bruce: But I think a lot of people would say, oh, you just refuted his argument. In other words, you made what I think is a good argument against his argument. You pointed out that he's actually just defining loving violence as equivalent to disagreeing with him. Something along those lines. Okay.

[00:20:35]  Peter: So there's a lot of ambiguity in these concepts.

[00:20:40]  Bruce: There is a lot of ambiguity. So this is why I don't mean to pick on Dean in particular. I actually like Dean a lot. He does get like this sometimes, but we all do.

[00:20:52]  Unidentified speaker: Sure.

[00:20:52]  Bruce: He's often great to talk to. Right? And the point I'm actually trying to make is I'm intentionally picking what I suspect most people listening to this podcast will immediately detect is a bad argument. And yet I'm trying to show it's not so obvious how you apply Popper's epistemology to it upfront. OK, especially if we're just trying to use the terms kind of loosely.

[00:21:15]  Unidentified speaker: Yeah.

[00:21:16]  Peter: That's why I kind of think I think I was saying this to you before. When you're having a conversation with someone, a debate, it rarely is that productive from what I've seen in real life to go like meta into epistemology. It just becomes like almost like increases the ugliness of the exchange. If that makes any sense.

[00:21:42]  Bruce: You know, I have gone meta into epistemology many, many times. Do you know how do you know how many times it has gone well? Zero.

[00:21:53]  Peter: I mean, I mean, if you're having a conversation about epistemology, that's one thing. But if you get into that, you know, people are always going to think you're misrepresenting them or, you know, basically straw manning them, which I get.

[00:22:11]  Bruce: Okay, let me let me ask you something, though. Doesn't it come down to epistemology? I mean, like, why can't we have discussions about epistemology? Why can't we say, okay, I can detect the mistake you made in that argument. It's an epistemological mistake. Here is the epistemological mistake that you made. Why does that get people angry? Why does that cause us to, quote, go meta and derail the conversation? Can you actually explain to me why it shouldn't be a fair argument?

[00:22:43]  Peter: Well, you know, that's maybe in a better world.

[00:22:47]  Bruce: Maybe in a better world.

[00:22:48]  Peter: That would be what we would be talking about all the time. But people get really, I don't know, it just kind of, I can get it if I'm having a conversation with my wife or kids, and I start talking about Karl Popper, well, they don't want to hear it. But at the same time in my own personal life, one of our catchphrases in my family at least is I say, well, if my wife or kids offer some criticism, I say thank you for the feedback. But it's funny, but I strive to actually mean it too, and to be open to criticism from people around me. That's how we move closer to truth. So I think that's the attitude that I strive to. That's what I get from Popper, and what I strive to, imperfectly, of course, strive to adopt, but might be just asking too much of the world to hope everyone will act like that.

[00:24:02]  Bruce: So I agree with everything you just said. Okay, so I'm completely agreeing with you that going meta never goes well. I've never had it go well, like ever. Okay, so I completely agree with you that it derails the conversation. It causes problems.

[00:24:16]  Unidentified speaker: Okay.

[00:24:17]  Peter: Yeah.

[00:24:17]  Bruce: But doesn't it even seem reasonable? Like most conversations are wrong at step one. Like their epistemology is wrong. And so they are already making a mistake at step one. They've lost me at step one, their epistemology, okay? If you're actually going to have a true critical conversation, shouldn't a lot of our conversations be about epistemology? Yeah. Like I can make a rational case for it. I just can't make a human case for it, if that makes any sense.

[00:24:50]  Peter: It's an interesting contradiction. I think you're right. It should be. We should go meta all the time in our conversations. But yeah, it never goes well. I guess we're just too flawed.

[00:25:01]  Bruce: The reason why I have gone meta so many times, and I know this is something people call me out on, is because I have experimented with, and I have extensively experimented with, is there some way to bring up epistemology that doesn't offend people? And I have tried numerous approaches. And the problem was, is that I had to try it to find out, right? I had to literally, there might have been a way to do it. Like maybe people just aren't going about it right, right? Yeah. So if you are super skilled, maybe you can, I'm not a super skilled online discusser, okay? Like I suspect I'm just a hair breath shy of autistic. I'm definitely considered neurotypical, not autistic, but I have like an autistic kid. So I doubt that I'm ever going to be the most super skilled online discusser where I'm in text format and I can't pick up on any of the nuance that I need to be able to pick up on.

[00:25:58]  Bruce: And I feel like text is like my worst, okay? It's like I really just cannot do well with it. But I really felt like I needed to try all sorts of different attempts to see if any of them worked. And I don't think any of them worked. I don't think any of them even came close to working. So at this point, I'm actually in your camp. Like it's basically pointless to try to discuss epistemology, even though that is what we should be discussing, if you actually want to make progress. So let me make a stronger case for that throughout this podcast. Now I've had numerous conversations with Dean in the past. And on more than one occasion, I've pointed out to him that his theories or explanations are quote, easy to vary. I did not do that this time. I never brought up easy to varyness to Dean as part of my criticisms of what he was saying. Dean at this point, though, says, let me guess, my argument is also easy to vary.

[00:26:56]  Bruce: Now, why does he say that? Okay, because I have in the past said, oh no, your argument's easy to vary. And I've done this often enough to him. He has now using that as a way to argue with me. Oh, you just always go to easy to varyness.

[00:27:09]  Unidentified speaker: I'll give you the actual quote where he says that in a different argument, in a different setting.

[00:27:14]  Bruce: Clearly, when he says this though, he intends to scorn the very idea that his argument might be easy to vary. Okay, so is Dean's argument easy to vary? You said kind of yes, but you're not quite sure you could identify exactly why. And I think that's probably right. I think that there's a good case to be made here that he is, well, let's actually take the definition of easy to varyness from Deutsch, beginning of infinity, page 22. When theories are easily variable, I call such theories bad theories. So is Dean's explanation easy to vary? We're not clear. Is it a bad explanation? It is a bad explanation, but it's a little unclear if it's a bad explanation in the Deutsch sense or not, okay? Now in a different setting, let me actually move to why he brought up easy to varyness, where I actually did accuse him of easy to varyness. I swear these two conversations were happening within a day or two of each other, and they kind of got intermingled, I think is what happened.

[00:28:18]  Bruce: And they're both about libertarianism. So it was similar topics, okay? So I had another disagreement with him. And one of the things that Dean advocates for is the idea that you can determine if a law is bad or not based on if it is coercive or not. And he uses an old definition of coercion from TCS to back himself up. So I put up the definition to criticize Dean's use of it, but not to dispute the definition itself. And Sarah Fitz-Claridge, who is the founder of TCS, expressed concern because this definition is an old definition and TCS no longer uses it. However, it was the definition Dean was using. So I don't want, even though it did like originate from TCS some more back in time, let's understand that this definition isn't the current TCS definition of coercion. It was the definition that Dean himself published to explain how to tell if a law is bad or not.

[00:29:20]  Bruce: So I'm only using it for the sake of libertarianism and being I'm not in any way trying to associate it with TCS for Sarah Fitz-Claridge sake here. Here's the definition. By coercion, we mean the psychological state of enacting one idea or impulse while a conflicting impulse is still active in one's mind. This leads to some subsidiary meanings. The action of intentionally or recklessly placing someone in a state of enacting one theory while a rival theory is still active in the person's mind. Behavior that is intended to, behavior that is intended or likely to do this. Now Dean claims we can use this definition of coercion to tell that mask laws during a pandemic are bad laws because they force you to wear a mask when you have a psychological state with a conflicting impulse still active in your mind, i.e. you don't want to wear a mask, but you have to anyhow or else you get punished, okay?

[00:30:15]  Bruce: Ergo, mask laws are immoral and undermine error correction. And then he goes on to argue, using an argument that comes from Deutsch, it is like gluing a piece of a puzzle down so that you can't change it. Now at this point I pointed out to him that you could change it in a democracy via voting. And he argued that took too long so you were still gluing it down for a period of time and therefore he felt his argument still applied. Okay. My argument back to Dean was that his explanation was easy to vary. I pointed out that his definition of coercion is purely subjective. And then in fact, TCS wants you to think of coercion as a sort of broadly subjective thing. The idea from TCS is that you shouldn't make your children feel coerced even if you're just trying to persuade them too hard. Things like that. Thus, it was equivalent to defining coercion as, and this is my definition now, coercion is the word we use to describe the feelings we have when the mere existence of someone else with a different idea than me acts in a way that impacts me in ways I do not enjoy.

[00:31:17]  Bruce: So mine is a purely subjective definition. Dean argued my definition was silly, which, you know, it was. I think it's accurate, by the way, but I think that is what we generally mean by coercion. But obviously, I'm trying to show that it's a subjective definition. And he says, because if Bob wants Sally to leave a public place for no reason other than he didn't like her, Sally's refusal to leave would be considered coercive. And this is clearly wrong. Okay, now I pointed out that if Sally didn't leave, then under his own definition, one, Bob was in a psychological state of enacting one idea or impulse, having to tolerate Sally's presence, while a conflicting impulse is still active in one's mind. He wants her to leave. And two, that Sally was intentionally placing Bob in a state of enacting one theory, forcing him to endure her presence, while a rival theory, not wanting to be around her, was still active in Bob's mind.

[00:32:11]  Bruce: So I argued that Dean's counter example perfectly fit his own definition of coercion. At this point, Dean says, your go-to criticism of anything is that it's easy to vary. Discussing with you has become predictable and boring. Now, at no point does it seem to occur to Dean that maybe the reason why I often point out that his explanations are easy to vary is because they are easy to vary. I would note that later Dean did apologize for this comment and how it was worded. Because Dean's really not that bad a guy. Though he continued to insist, as part of that apology, that I was abusing the easy to vary criteria, that it did not apply to his definition of coercion. Now, here's where things get interesting. Is this really just a subjective matter of opinion? If it is, then I guess I see no particular reason why, how we would resolve which is the more correct idea.

[00:33:07]  Bruce: But I actually suggested we could test between whether or not his definition was easy to vary or not. After all, he agrees with me that my definition is easy to vary. So if we can show that his definition and my definition are equivalent, then his must also be easy to vary. So I sent him an article that I had written, that I actually wrote responding to him sometime in the past, and I sent it to him again. And here is what I wrote in the article. I said, I conjecture that these two definitions are functionally one and the same, that that conjecture is equivalent to a universal statement, saying there are no cases where this old TCS definition of coercion identifies a situation as coercive that my definition doesn't also make the same identification and vice versa. As a universal law, it is now testable. Here is how you test it. Dean or any libertarian can offer me either, A, an example of a scenario that fits my definition of coercion but not his, or B, an example of a scenario that fits his definition of coercion but not mine.

[00:34:11]  Bruce: If he can offer me even one such counter example, my universal law is refuted by counter example. In the meantime, until he offers me a counter example, I argue that this was now our best theory, that his definition of coercion was easy to vary. At this point, the conversation came to an end, and there was never a response to this. Now, again, I want to emphasize, this is someone who has studied Deutsch and Popper extensively. And a quick check of his Twitter feed shows that he has often invoked the easy to vary criteria to criticize ideas that he disagrees with. Okay? So it's not like he has some issue with the concept of easy to vary-ness. And it does seem to me like I was offering a testable way to test whether his idea was easy to vary or not. It just was not accepted. The test was not attempted, if that makes any sense. Let's now talk about how this applies to another topic we've talked about several times.

[00:35:08]  Bruce: Okay? Brett's theory of intelligence, which in episodes 53 to 55 we discussed, you know, at length. Recall that IQ testing started off, started out as several different unrelated tests. So you had various, quote, IQ tests, and they were testing completely different things. And that one of the initial criticisms of IQ theory was that all you're doing is testing what kinds of knowledge you happen to value. And that really people have many different natural abilities. Now, this is all happened way back, like, in the early 20th century, right? Like when IQ tests first came into being. And notice that that's very similar to Bret Hall's own theory. Okay.

[00:35:51]  Peter: To back up for a second, Bret Hall's theory is that all ability is just a matter of time and interest.

[00:35:57]  Bruce: That's correct.

[00:35:59]  Peter: Just for our listeners here.

[00:36:01]  Bruce: We have equivalent intelligence. All universal explainers have equivalent intelligence, and that the only difference is time and interest. This argument that really you're just testing the knowledge that you happen to value, it got refuted because it turned out that all these tests had strong correlations.

[00:36:25]  Unidentified speaker: Okay.

[00:36:26]  Bruce: So if somebody scored well on one IQ test that tested, say, language ability, they would also score well on a totally different IQ test that, say, tested graphic analogies. Okay. And that there was this really strong correlation that existed. Okay. So now you've got this pattern, this exists, something that exists in the world that you have to explain. And they theorized that it was a general intelligence factor.

[00:36:51]  Unidentified speaker: Okay.

[00:36:51]  Bruce: Well, so keep in mind that IQ actually predated this idea of a general intelligence factor, that they kind of just started off assuming that there was a thing called IQ. And then they started to have questions about it. And then they actually had this observation that existed where if you score well on one IQ test, even if they're completely different areas of interest, that the scores tended to correlate.

[00:37:17]  Unidentified speaker: Okay.

[00:37:18]  Bruce: So how do you explain that? So they had to go right back to this idea of a general intelligence factor.

[00:37:24]  Unidentified speaker: Okay.

[00:37:24]  Bruce: Now this is one of the things that I brought up to the defenders of Brett's theory that this had actually taken place. And this is in episode 55 of our podcast, I asked the defender of Brett's theory to explain this refuting observation using their own theory. Okay. So keep in mind why I would ask this. Okay. Because this is what we want to do. We want our theories to explain the actual observations that exist. Okay. And here is the response I got. Correlations between IQ test results is a statistic anywhere between 0 and 1 and where some fuzzy line gets dropped, some fuzzy line gets drawn, where anything above, let's say, for example, 0.3 gets dubbed, dubbed, look, we have a correlation. The kind of correlation one is dealing with when it comes to determining statistical non-fuzzy associations between quantities like exam results is straightforwardly fuzzy.

[00:38:18]  Bruce: If you get a correlation of 0.25 versus 0.4, what different conclusion can you draw? Now, let me just point out that this person is misunderstanding how statistics and correlation work. So let me explain how it actually works, although I'm going to accept his argument even though he's making a mistake here because he's still making a point that I feel like I can interact with. So let's say that you've found, he's confusing two things, the level of correlation and if it's statistically significant. Let's say I found a correlation of 0.1, but that it was highly statistically significant. The conclusion that we'd be tempted to draw is that there is a general intelligence factor but that it plays little role. So you can actually, it's not really a matter of how much of a correlation do you have. If it's over 0.3, then we can draw this conclusion that there's a general intelligence.

[00:39:12]  Bruce: You can draw it with a 0.01 correlation. It's just that you would have to assume that the general intelligence factor is very unimportant. So I want to make that part clear about how that would actually work and how they actually did go about doing this. However, in some sense, this doesn't matter. And the reason why is because I'm not trying to necessarily defend IQ theory. I'm trying to understand his reasoning. So I say, well, this is a quote. Well, I'm interested in your reasoning. So you tell me what correlations you would or would not accept as a refuting observation. Would you accept, for example, a 0.999 correlation as a refuting observation to Brett's Theory of Intelligence? In other words, this is just a thought experiment. You tell me what evidence would cause you to change your mind about Brett's Theory of Intelligence. Now here's his response.

[00:40:04]  Bruce: Correlations as high as 0.99 are exactly what one would not predict given a common sense explanation of what is involved in sitting multiple exams so such a result could indicate cheating. A great example of how high correlations themselves have to be interpreted via explanations. So I offered to him at this point, I said, look, this is a thought experiment. Rethink the thought experiment to include strong anti-cheating measures, okay? And then redo the thought experiment with these anti-cheating measures in place. What I'm really asking you is if there is any combination of factors and observations at all that you would accept as a refuting observation to Brett's Theory of Intelligence. So I'm giving him every opportunity here to explain how his theory can be refuted by experiment. Here's his response. If you could say, here is my explanation of intelligence as a quantity, here is my explanation of why such and such exam results can be proxies of it, and here is how I have calibrated everything to predict what the correlation will be, and, and, and, explanation all the way down.

[00:41:17]  Bruce: Only then would he then accept the results of a refutation. And he explains why. He says correlations have to be connected via explanations. Explanationless statistical associations, apparently even in the case of randomly controlled trials, this is something I'm going to come back to, between exam results tell us nothing about the issue. In this case, a super high correlation between test results would be best explained by fraudulent results or cheating. So let's invert this though. Let's say that we found that the test did not correlate at all. Do you think that there would be any real chance he would not immediately proclaim this as a refutation of IQ theory? I suspect that he absolutely would have. By the way, he'd be right to, because it would be a refutation of IQ theory if we found no correlation between these tests. Because IQ theory is, in fact, an empirically testable theory, although not a strongly empirically testable theory, but it is an empirically testable theory that can, in principle, be contradicted by observations.

[00:42:23]  Bruce: So if he were, if we found no correlation at all, or it wasn't statistically significant, and if he were to use that to say, see, that proves IQ theory is false, he'd be right, okay? So he's saying that unless you can offer a giant chain of causal mechanisms, quote, all the way down, even a randomly controlled trial will not convince him. Now let's talk about randomly controlled trials for a second. They're a tool that we invented to work out causation. And we've been doing them for a very long time. And only really recently has Judea Pearl kind of worked out the mathematics of why they worked, but kind of intuitively we've understood how they work. The idea is that you're going to do an intervention. You're going to actually take two groups. Let's use the typical example of a prescription drug. But it could be anything. So I'm going to take two random groups.

[00:43:22]  Bruce: Why do we make it random? Because we want to make sure that if there's any other causal factors, the random distribution of the two groups breaks the causal factor. So to use the example that Judea Pearl likes to use, there's this idea that smoking causes cancer, which is widely scientifically accepted today. But it took a long time for science to accept this. The reason why is because it was unethical to do a randomly controlled trial. If we could have done a randomly controlled trial, science would have worked it out very quickly. But there was always this chance that there was some other causal factor that caused you both to get lung cancer and also to smoke. And there was no way to refute that idea without a randomly controlled trial. But if you do a randomly controlled trial, what you do is you now are selecting these two groups, the control group and the group that you're going to force to smoke.

[00:44:18]  Bruce: That's why it's unethical. And you're going to break them randomly into two groups. Whatever that other causal factor was, is now broken, you've now intervened, and the causal factor has disappeared. So you're now testing the one causal factor, smoking. This is why we use randomly controlled trials to work out whether something is a cause or not. But he's saying that even that unless you can actually offer an explanation with a giant chain of causal mechanisms, quote, all the way down, that he's not going to be convinced by this randomly controlled trial. That is, unless the result favors his theory, then it will actually work as a [reputation?] for the competing theory. So let me repeat his argument in his own words again. I'm making a point here that's kind of important. Correlations have to be connected via explanations. [Explanational?] statistical associations, even in the case of a randomly controlled trial, between exam results tells us nothing about the issue.

[00:45:15]  Bruce: In this case, a super high correlation between test results, this case being the thought experiment, would be best explained by fraudulent results and cheating. Now, I want to emphasize this is a thought experiment. We're literally trying to see what he would accept as a refuting observation. And he's upfront saying that he's going to declare every lack of correlation as proof he's right and every correlation found as proof of cheating. Is this an easy to vary argument?

[00:45:45]  Peter: I think it's sort of an interesting way to put it though. I mean, that does actually ring true for me a little bit. You can't, you know, if you're just looking at the world and looking for correlations between things, I mean, that it might be, you might find out some interesting information, but, you know, in order to really like make sense of it, you need to link that up with explanations. I think that kind of rings true a little bit.

[00:46:13]  Bruce: Let's talk about that. I think a lot of people do feel that a lot of crit rats do feel that way because this argument comes up all the time.

[00:46:20]  Peter: I mean, they might take it a little extreme. Don't get me wrong, but, you know, it does kind of make sense.

[00:46:26]  Bruce: So what's interesting is I will ask... So I had not this person, but I had somebody else make the exact same argument to me. Another crit rat who said, yeah, if you don't have an explanation, then it's literally just pattern fitting. I then asked him, does smoking cause cancer? What do you think he answered?

[00:46:46]  Peter: Well, I would suspect yes.

[00:46:50]  Bruce: Yes.

[00:46:50]  Peter: I mean, I think that might be an area where you can't... I mean, as far as I know, no one might know exactly why smoking causes cancer, but you...

[00:47:02]  Bruce: In other words, we don't necessarily have an explanation all the way down.

[00:47:06]  Peter: Yeah, but you can kind of assume that the smoke is doing some damage to the cells. I mean, it's not that... It's kind of like the genes and IQ thing in a sense. You might not know exactly how genes influence IQ, but it's reasonable to think that there's some kind of mechanism in the brain or the formation of the brain or something that does influence IQ.

[00:47:34]  Bruce: But we don't necessarily know what the mechanism is.

[00:47:36]  Unidentified speaker: Sure, sure.

[00:47:37]  Peter: But we could say that for just about anything, which I think is kind of your point, I guess.

[00:47:41]  Bruce: I'm going to actually make that point now. There are no theories in existence that can survive this criteria. There is not a single theory in existence where we know the mechanisms all the way down. Which is why this is a bad argument. It literally can be used to dismiss any outcome of any experiment. It is an all purpose argument. Let me tease out what's really going on here. Does smoking cause cancer? I want to emphasize that we have never done a randomly controlled trial of smoking causing cancer. If we could have done it, science could have figured out that smoking caused cancer very quickly. To be able to do that, you would literally have to take two groups of people at random, and you would have to force one group to smoke every day at random, like they may not want to. And you would have to force the other group to never smoke. You have to monitor them to make sure that they're not cheating.

[00:48:46]  Bruce: And unless you actually did that, so you would basically have to put these people in prison to be able to do the experiment. And unless you do that, you haven't eliminated the other possible causal factors that you don't know about. This is why we have never really been able to, by experiment, show that smoking causes cancer. Now, Judea Pearl, in his book, The Book of Why, he goes over this history in detail, and it deserves its own podcast, so I won't go into the whole history here.

[00:49:18]  Unidentified speaker: Okay?

[00:49:19]  Bruce: But what actually happened is, is that science eventually figured out some other interesting ways to criticize the idea that lung cancer wasn't caused by smoking. One of the things they did is they worked out, okay, let's say there's a gene that causes both lung cancer and also causes smoking, and they worked out how effective this gene would have to be, like the degree to which it would cause you to go smoke. Okay? And it was this level of influence that was beyond any genetic factor ever known, like hundreds of times worse, right? You would have to have this overwhelming ability to force you to go smoke, right? And at this point, they started to realize, oh, that's like silly. Like, there's no way that's actually true. And they eventually did come to accept that smoking does cause lung cancer, okay? Because it was just the only explanation that made sense.

[00:50:11]  Bruce: So without doing a randomly controlled trial, they eventually figured out other criticisms they could use that eventually the scientific community adopted, okay? Let me point this out, though. A randomly controlled trial is a very good way to work out causation without knowing all the mechanisms. And in fact, you never know all the mechanisms. All you really need to know is that you're testing if there is some sort of causal link, okay? And that's the whole point of making it random into two groups, okay? And this is something we've worked out quite well. Judea Pearl can actually mathematically, using graph theory, show you what it is we're doing, why it works, okay? And he can show you how you break causal links, and it leaves one link left, and you [can test if/didn't test that] that link is true, and you get totally different results. And just mathematically works out that way.

[00:51:03]  Bruce: There is nothing wrong, nothing wrong with using correlations between IQ tests and using that as a piece of evidence to refute the idea that there is no general intelligence factor, okay? The way you would get out of that isn't by making an all-purpose argument and saying, oh yeah, well, you're not showing me explanations all the way down. It's by offering a counter-explanation, which is what I was asking for. How would Brett's theory that intelligence is based entirely on interests and time, how would they go about trying to explain the existence of this factor? Well, of course, they don't have an explanation. That's why they're trying to instead use all-purpose arguments. Now, let me just point out, though, that this doesn't mean that there is a general intelligence factor. It's just that we don't have a good alternative theory, and Brett's theory isn't a good alternative theory, okay?

[00:51:57]  Bruce: We don't have one that explains this regular pattern in the real world. And maybe you could come up with it, though, okay? It's not that this proves in any way that there is a general intelligence factor. It's just that it eliminates Brett's theory, at least Brett's theory in its current form, as a competitor to it.

[00:52:17]  Peter: No, that really rings true. So the premise that you could practically criticize anything anyone says about life, that, oh, you don't have an explanation for that. I mean, down to the, I don't know what, you could always, you know, down to the, I don't know, the atomic level or something, you know, unless you know what these electrons and protons are doing, you're not, it's not a good explanation.

[00:52:46]  Bruce: And it's just pattern fitting at this point, right? It doesn't make sense.

[00:52:51]  Unidentified speaker: Okay.

[00:52:51]  Bruce: Yeah.

[00:52:52]  Peter: So you, what you should strive to do is come up with a better explanation.

[00:52:57]  Bruce: Yes.

[00:52:57]  Peter: Right. Not, not, not just, oh, that's not a good expla, yeah, okay. That's a good way to think of it.

[00:53:04]  Bruce: So I continue the conversation with this defender of Brett's theory. Okay. And the idea that correlations are meaningless without an explanation, explaining every single mechanism all the way down, that's a very common, or in fact, it's a super common argument amongst crit rats.

[00:53:20]  Unidentified speaker: Okay.

[00:53:21]  Bruce: Although they use it selectively, and that's one of the criticisms I have of the crit rats, is that they only trot that argument out when it's in their favor, and they never trot it out when it's against them. Now, I asked about this. I said, why do you feel correlations are fuzzy in the context of testing an explanation? And I gave the example of studying if grass cures cancer via a randomly controlled trial. And the example I gave was, presumably grass does not cure cancer, is that we do a randomly controlled trial and we find that people with cancer are no more likely to survive if they do or don't eat grass. Okay, so we do the trial and we find out there is no statistical difference between the two groups. Why doesn't that, even though I don't, even though I've got zero explanation, presumably I can't have an explanation for why grass cures cancer, because it doesn't.

[00:54:17]  Bruce: Why would that not work as a refutation of the idea that grass cures cancer? It seems like it would. Okay, in fact, the answer is it does. Okay, here's what the defender said. If I think I have an explanation for why so-and-so cures a cancer, and then I give people the cure and the cancer goes away, then my cure correlated with cancer remission in a sense. That's not the same sense of the word correlation, when it comes to comparing IQ exam results. Okay, so even in this case, where I'm literally arguing grass does not cure cancer, and we can show that through an experiment by showing that there's no correlation, there's no statistical significance at all. He's still claiming you have to have an explanation of why grass cures cancer. Well, how could you have that? Because it doesn't, right? He's saying that without an all-the-way-down explanation for why grass cures cancer, such data would be meaningless.

[00:55:14]  Bruce: Okay, is that actually true? Let's flip this around, okay, and let's see if that's actually true. Let's say that we have no explanation at all for why grass would cure cancer, okay? But we just decide to do a randomly controlled trial anyhow. And what we find is that two groups of people with cancer, that if we feed them grass, the group that was fed grass has a 10 times more likely that their cancer is going to go into remission, okay? Now, I've got zero explanation. I've got no explanation for why this happens. Is it just pattern-fitting, or did we just tentatively refute the theory that grass does not cure cancer?

[00:56:02]  Peter: Well, I mean, it almost sounds like you're moving towards a defense of empiricism or something. I mean, what is a study? A study is a kind of observation, and then you move to an explanation from that, right? So then I guess the logical next step would be to start thinking of an explanation for why we've made this observation.

[00:56:26]  Bruce: Okay, I think you are doing a good job of vocalizing what their concern is, what the crit rat concern is here.

[00:56:36]  Peter: I mean, it's a valid concern. They just take it pretty far, I think.

[00:56:39]  Bruce: Okay, let me actually clarify now. Scientific explanations exist at many levels of explanation, and there is no minimum limit on how explanatory they need to be. And this is the mistake the crit rats are making. So if I have an explanation, and that explanation is that grass causes cancer to go into remission, and that's all I've got. I've got nothing else, no other mechanisms or explanations. That is an idea that can still be tested empirically, scientifically, okay? And it's competing theory. You can only test it against its competing theory, which in this case is that grass does not cause cancer to go into remission, all right? Even just that much is enough for a scientist to start testing. And that's all it takes, okay? Yes, of course, that would be a level that is just unacceptably not very empirical at this point. The only thing that we can really test is to do a test and say, does it cause, in a randomly controlled trial, does it cause remission or does it, is it a causal factor in remission of cancer or not?

[00:57:52]  Bruce: We would immediately, as scientists, want to start digging into what are the causal factors and start making our theory way more empirical. But you do not eliminate theories at this level of empirical. That's what the crit rats are trying to do. They're trying to eliminate certain kind, and again, they do it selectively. That's part of the problem. Smoking does cause cancer, but they're not even going to consider the possibility that IQ could have a general intelligence factor behind it because there's no explanation. This is basically an all-purpose set of explanations. And the real truth is that it doesn't take much to get counted as a scientific theory and to immediately you can start testing it. And if you actually did find that [brass/grass] was causing cancer to disappear, of course you would want to now go start feeding it to people, and that would be the scientifically correct thing to do, even if you had no explanation for why it works.

[00:58:45]  Bruce: Consider the fact that for many, many, many generations, we gave people aspirin with zero idea of how it worked, right? In fact, that is just not uncommon in medical science, where we have a vague idea of this might cause something, and we have no idea what the mechanisms are. They try to dig into mechanisms in medical science. They really do, but they don't always succeed. And sometimes the mechanisms are just made up, and they don't have any way of testing the individual mechanisms. They just give the drug to people, and they see if it gives a statistically significant result. Now I've seen people argue that's not scientific, and I'm saying, no, it absolutely is scientific. This is completely compatible with Popper's epistemology. You do not need a deep explanation to have an explanation, and to have it be empirical, at least a little bit empirical, and you can start testing.

[00:59:38]  Bruce: And that is, I think, the mistake the crit rats are making here.

[00:59:42]  Peter: I mean, how often is it that someone gives someone a drug and they know down to the molecular level exactly what's happening in the body? I suspect it would be kind of rare, right? I mean, it's far more common that they don't.

[00:59:56]  Bruce: Well, okay, but let's go all the way here. Quantum mechanics. Do we understand the causal mechanisms of quantum mechanics all the way down? Sadia, who we've had on the show, she will point out to you that we don't, right? And that's one of her main criticisms of it. Quantum mechanics is our best understood theory. It's the most empirical theory in existence, and there are no counterexamples to it that we know of in terms of observations, okay? And yet it also can't survive this criteria that's being plotted out, okay? There are no explanations that have causal mechanisms all the way down that we understand, okay? All right, I'm emphasizing this for a reason. And particularly in the next podcast, I'm going to come back to this idea that it's okay for scientific explanations to only be barely explanatory. And then we want to make them more explanatory over time, okay?

[01:00:52]  Bruce: It's a difference between speed and acceleration, okay? There's no minimum limit on explanation, but we always want to increase the level of explanation. And we're going to come back to that in the next podcast. But let me move on with the current podcast. So key point, since no explanation has every causal link worked out, this is an all-purpose explanation. Literally any theory that you don't want to accept, you can claim the results are just correlations or pattern fitting. If these examples don't count as bad explanations, I'm seriously at a loss as to what would. Getting back to, but theories are easily, which are easily variable, I call such theories bad explanations. It's less clear to me if these are easily variable or not, okay? I certainly see them as bad explanations, but I'm not so sure if they're easily variable or not. And again, I want to emphasize this is someone who I was, the people who were defending Brett's Theory of Intelligence, they're all people who have studied Popper and critical rationalism extensively, okay?

[01:01:49]  Bruce: And he's literally saying, when it comes right down to it, that unless you can show every single causal mechanism, he doesn't care if it's even a randomly controlled trial or not, he's going to reject the evidence as just pattern fitting or just a correlation. And yet, you know he's not going to actually do that when it comes out in his favor. So what is, if this isn't an easy, this isn't examples of easy to vary explanations, what is, okay? So let's actually dive into this a little bit. What I want to do though, before we move on and try to resolve this problem, is I want to actually read what Deutsch says in Beginning of Infinity about easy to varyness, so that we actually have his explanation of it in mind. I've intentionally been a little vague up to this point, and one of the reasons why I have been is because I think the term easy to varyness brings to mind something somewhat different than what Deutsch actually says it means.

[01:02:48]  Bruce: That's why I haven't defined it up to this point very well. Let me actually do justice to what he says in Beginning of Infinity now. He uses the example of Persephone in Hades, this idea that there's a myth and that we're using this myth to explain why we have seasons. If you're familiar with the myth, Hades takes Persephone into Hades, into Hell, and her mother's upset, so she makes the world be winter. Then when Persephone is allowed to leave, then she makes the world summer, and this explains the seasons. On page 20 of Beginning of Infinity, he says, although the myth was created to explain the seasons, it is only superficially adapted to that purpose. On page 21, the reason those myths are so easily variable is that their details are barely connected to the details of the phenomena. For whenever it is easy to vary an explanation without changing its predictions, one could just as easily vary it to make different predictions if they were needed.

[01:03:46]  Bruce: Page 22, when theories are easily variable in this sense, I have described experimental testing as almost useless for correcting their errors. I call such theories bad explanations. I want to emphasize that. He's claiming that if a theory is easily variable, experimental testing is almost useless for correcting errors. One of the things I'm going to call out is that I don't think this is actually true. It kind of is true. I mean, obviously there's an intuition for it that there's no point in testing Persephone's and Hades as a myth because you can just keep adjusting it to get whatever explanation you want. Page 22 again, by adopting easily variable explanations, a person is ensuring that they will be able to continue fooling themselves no matter what happens. Page 25, it is only when a theory is a good explanation, hard to vary, that it even matters whether it is testable.

[01:04:36]  Bruce: Bad explanations are equally useless whether they are testable or not. So this is how he actually defines Easy to Varyness. I've called out these bad arguments that CritRats have made. And one of the things I pointed out is that it's an all-purpose argument. But that's not really quite the same thing as saying that the parts of the explanation are disconnected. How did Deutsch put that? The details are barely connected to the details of the phenomena. I mean, he's using what I would believe to be a completely all-purpose argument, that there is literally no explanation that can survive the criteria that he's offering. And yet it is connected to the details of the phenomena. At least you could make that argument that it is. So I think that's one of the reasons why it's a little unclear how to connect the Easy to Vary criteria to a bad explanation.

[01:05:32]  Bruce: I now am going to talk about a conversation I had with Jose Chalupa, same guy that was arguing with Dean on Twitter, that really clarified things for me quite a bit. And the rest of the episode, I'm now going to try to resolve this problem. And I think I can mostly resolve the problem. There is one issue that I can't fully resolve, and I will call it out a few times. So in a conversation with Jose, now, he and I agree on this, I think. So I'm not going to be criticizing anything he's saying here. But I asked him, I said, can you give me an example of an Easy to Vary theory and an example of a Hard to Vary theory? And he suggested an Easy to Vary theory would be Spontaneous Generation, and a Hard to Vary theory would be Darwinian Biological Evolution.

[01:06:27]  Peter: Spontaneous Generation is just that life just forms without natural selection?

[01:06:33]  Unidentified speaker: Yeah, that's right.

[01:06:34]  Peter: Okay.

[01:06:36]  Bruce: Now, if I had picked two examples, I would have picked creationism as the Easy to Vary example and quantum mechanics as the Hard to Vary example. Okay. Jose explained to me that why he thought Spontaneous Generation is an Easy to Vary example, which, keep in mind, I agree with Jose, so I'm not trying to argue with him. But here's examples of what he says. He says that really this is an Easy to Vary theory because every experiment is easily defeated by claims like dust mites generate when the invisible mighty dust mite king wills it, or life only spontaneously generates when you aren't watching it. I want to emphasize that one. Life only spontaneously generates when you aren't watching it. And you can easily see how spontaneous generation, it's got no explanation at all behind it. Okay. It's at the level of grass cures cancer, right? There's no deep explanation.

[01:07:34]  Bruce: So it would be very easy to just add things like life only spontaneously generates when you aren't watching it. And there thereby defeat every experimental outcome. Okay. So I can see where he's coming from. I can see why spontaneous generation is a good example of an easy to vary explanation. Go do a Google search of how science refuted spontaneous generation, and you will immediately find the experiments done by Louis Pasteur and Francesco Redi. Okay. Science absolutely did refute spontaneous generation by experiment via testing, exactly like Deutsch says you can't do to an easy to vary explanation. Here's a quote from the Internet. Louis Pasteur provided evidence to disprove spontaneous generation through his famous experiment using swan neck flasks. He filled these flasks with a nutrient rich broth and heated them to kill any existing microorganism.

[01:08:43]  Bruce: The swan neck design allowed air to enter the flask but prevented dust and other particles, which could carry microorganisms from entering. Over time, no new microorganisms grew in the broth, demonstrating that life did not spontaneously generate from nonliving matter. When the necks of the flasks were broken, exposing the broth to the dust and particles, microorganisms began to grow, proving they came from the environment and not through spontaneous generation. Okay. Now I'm going to quote Deutsch again. It is only when the theory is a good explanation, hard to vary, that it even matters whether it is testable. Bad explanations are equally useless whether they are testable or not. How can it be that an even canonical example of an easy to vary theory was in fact refuted by science via testing when Deutsch claims this is impossible? Or is spontaneous generation a hard to vary explanation?

[01:09:37]  Bruce: And Jose and me and you, Peter, are all just wrong when we were mistaken, when we said it was an easy to vary explanation. Which do you think? I mean, obviously I'm asking a question that I suspect is unanswerable. But if you want to give some thoughts on this, I'd actually be happy to hear your thoughts on this.

[01:09:56]  Peter: Well, I don't know. I think you're ruining another one of my go-to talking points with this hard to vary explanation thing. I don't know. I'm not sure I'm ready to give it up yet, but your pushback is interesting.

[01:10:13]  Bruce: Now let's take Darwinian evolution as the hard to vary example. Okay. It turns out we can use the exact same arguments, and in fact, the exact same arguments get used when defending Darwinian evolution. Now I mentioned back in episode 74, which was our problem of open-endedness episode, that I used to be a creationist back when I was a kid, not anytime recent. And I happened to have numerous arguments with Darwinian evolutionists over creationism. So I actually have an intimate familiarity with how Darwinian evolutionists defend Darwinian evolution from a creationist. And I mentioned in episode 74 that I eventually did jump ship and become a Darwinian evolutionist precisely because creationism got refuted for me by observation. Now again, we might ask, how is that possible? If creationism is easy to vary, how can any observation ever refute it?

[01:11:12]  Bruce: Okay, that's one I'm not going to answer for today. I'm going to save that for the next podcast. Despite this, I noticed that evolutionists that I argued with consistently easily varied Darwinian evolution to avoid refutation. Okay, so let's say that I ask you how to explain the existence of life given the second law of thermodynamics using Darwinian evolution. I'll even buy that once life exists, Darwinian evolution works, but you need to explain to me using Darwinian evolution as a theory how life got started in the first place. Okay, this is really...

[01:11:48]  Peter: [Well, not sure anybody can, right?/I'm sure anyone can, right?]

[01:11:49]  Bruce: Right, this is a totally fair question though.

[01:11:51]  Peter: Yeah, oh yeah, yeah.

[01:11:53]  Bruce: Okay, and so the argument against this was always, always, always the Miller-Urey experiment. Okay, and I would get an answer like this. Well, you say Darwinian evolution can't explain the creation of life, but there was this Miller-Urey experiment back in 1952 where they threw a bunch of chemicals believed to be present in the pre-earth together, which by the way, they got wrong supposedly, and presto, out came a bunch of amino acids. So like half the problem is already resolved. My own high school biology teacher used this argument in class. Okay, this is a beyond silly argument. Just because you can show catalysts exist that create amino acids, you cannot use them as an explanation for how life got started, much less declare it half the battle. This is literally like trying to explain airplanes and claiming, well, you see, there's these natural processes called supernovas that create iron and then it goes into mines and we mine the iron out.

[01:12:57]  Bruce: And see, I've already half explained where airplanes come from. So, you know, you just see my explanation works, right? I mean, it's a terrible, terrible explanation.

[01:13:07]  Peter: OK, to be fair, it sounds like kind of an interesting experiment, though. I mean, as looked at as one piece of a puzzle, right?

[01:13:14]  Bruce: Fair enough. And it absolutely is. Just if we needed to explain airplanes, we definitely also need to explain where iron comes from. And supernovas is part of that answer. But it isn't really addressing the question being raised by the creationists. In fact, it's not really even kind of addressing the question being raised by the creationists. OK. And so contextually, it's a bad argument. Even if the experiment itself was a worthwhile experiment.

[01:13:42]  Unidentified speaker: OK.

[01:13:44]  Bruce: When I would point out to the Darwinians that this was kind of a silly argument, I would get a reaction like this. No, this totally shows your argument doesn't refute Darwinian evolution. It shows that there are natural processes that create life on its own. Now, in Episode 74, I also gave the example of the flagellum in a bacteria and how there is not a currently known way to explain how it evolved under the current theory. Although we do have partial explanations. And this usually is very comforting because we can show that some of the steps follow regular neo-Darwinian evolution. OK, even if we don't know what all the steps are. And honestly, there's a certain unfairness in expecting a Darwinian to offer you a complete step-by-step guide as to how it happened. OK, having said that, to some degree, it's still a completely fair question.

[01:14:37]  Bruce: Darwinian evolution requires that every single part of the flagellum of the bacteria has to have had a purpose other than what it's currently being used for. And it's this super complex set of mechanisms. And of course, we have no idea how that actually evolved today. OK. On the other hand, shouldn't that count as a refu... Why doesn't that count as a refuting observation? Right? I mean, you can see why a creationist would raise it and say, Look, I'm refuting your theory. I'm offering you a refutation. You say your theory is testable, and I'm offering you an actual example of where there is no current known way to explain it under the current theory. And you're basically just giving me a hand-wavy, easy-to-vary explanation on par with, it happened when you weren't looking, okay? Almost identical to the whole dust mite thing, the spontaneous generation thing, okay?

[01:15:32]  Bruce: And isn't that really kind of what the Miller-Urey experiments being used as? They're kind of saying, Well, see, I can show you that catalysts exist and some of the stuff orders on its own, and then the rest kind of just happens when you're not looking, okay? It's almost an identical easy-to-vary argument as someone who would be defending spontaneous generation would be using, okay? Does this mean that Darwinian evolution is easy to vary? Well, of course that would be silly to claim that Darwinian evolution is easy to vary. Yet clearly there is something basically indistinguishable from easy-to-varyness taking place when Darwinians defend their theory. There is no set of refuting observations to Darwinian evolution. None. There cannot be because a Darwinian would always just say, well, we don't know how it happened, right? And that would actually be a fair answer.

[01:16:22]  Bruce: And yet it basically dismisses all refuting observations. It is, in a certain sense, makes Darwinian evolution seem very easy to vary, which is, by the way, why Popper considered it not a scientific explanation. I don't agree with Popper on that, like even slightly. That was why Popper felt that way. Now let's take a look at my own example of quantum mechanics vs creationism. So I already mentioned I'm not going to get into how I refuted creationism. But needless to say, I did. I was able to come up with refuting observations to creationism, and that caused me to jump ship and become an evolutionist. Does that mean creationism is hard to vary? What about quantum mechanics? Quantum mechanics is literally the best, most hard to vary theory in existence. And it is also probably the most easily varied theory in existence. So ask any average scientist how to explain quantum mechanics, and you'll get an apocryphal quote from Feynman that says, anyone who claims to understand quantum theory doesn't understand quantum theory.

[01:17:24]  Bruce: The actual quote, by the way, was, I think I can safely say that nobody understands quantum mechanics. And there are various competing interpretations of quantum mechanics that regularly get trotted out by the scientific community to, quote, explain what is going on and explain away the many-worlds interpretation. So Bohm, for example, explains quantum mechanics as a particle that has a definitive position in momentum that you just can't measure. You just can't observe it, according to Bohm. It's happening when you're not looking, basically. Exact same issue with spontaneous generation. And this particle is attached to a giant quantum field, which is of course mathematically equivalent to many worlds, that moves the particle around via non-local influences. Copenhagen does no better, actually does worse. It explains quantum mechanics as a mysterious collapse that takes place for no discernible reason as basically one of just the axioms of the theory.

[01:18:21]  Bruce: If you ask, when does it happen? You're told when you observe it. If you ask what counts as an observation, you get told, we don't know. I mean, we're literally talking about something that happens when you're not looking, right? Or happens when you're looking or not looking. It's the exact same problem as with spontaneous generation.

[01:18:39]  Peter: Are you saying that the hard-to-varyness criteria has some validity in the search for a good explanation about quantum theory? I mean, it kind of seems like it does.

[01:18:55]  Unidentified speaker: The answer is that it does.

[01:18:58]  Bruce: But yes, we're getting to that.

[01:18:59]  Peter: [It's the least hard-to-vary/at least hard to vary]. Okay, that's where you're going.

[01:19:04]  Bruce: Also, very often you'll get told, if you ask these questions, that quantum mechanics is just an instrument. And they'll invoke instrumentalism. And sometimes called the shut up and calculate theory of quantum physics. So quantum mechanics, instrumentalism is literally the single most easy-to-vary theory in existence. It would be impossible to have a more easy-to-vary theory. Because it's by definition, you're not allowed to explain anything, right? So anything, whatever happens, happens, is basically instrumentalism. So quantum mechanics, despite being something like the quintessential example of a hard-to-vary explanation, is consistently and routinely easily varied, almost at a whim. Okay. So apparently, hard-to-vary explanations can become easy-to-vary, and easy-to-vary explanations can be tested and refuted as if they are hard-to-vary.

[01:19:58]  Bruce: So it seems we have a problem. No wonder no one can agree on what counts as easy-to-vary. No wonder Dean and I, I'll say, that theory is easy-to-vary. He'll go, no, it's not. I mean, like, of course. Like, it's really hard to even figure out how to apply easy-to-varyness, you know, to anything. Some wise advice from Vaden on this. He said, you know, you can really just use easy-to-varyness not to criticize other people's theories, but instead use it to just criticize your own theories. Just think about, is your explanation easy to vary? And you can use that to try to improve your explanation. That's very wise advice. Let me say, though, that really this moves the whole thing to the realm of subjectivity, that basically the ultimate deciding of whether a theory is easy to vary or hard to vary just really just boils down to your subjective opinion on the subject.

[01:20:49]  Bruce: That doesn't strike me as what I really want in my epistemology, okay? Even though it's definitely good advice, and I probably should have followed it a bazillion times with Dean because it did no good to point out his theory was easy to vary. I want to try to reform the idea, though, to make it something more objective. And it turns out I don't need to because Popper already did. So what is Popper's equivalent to the easy to vary criteria? I'm going to argue it's his idea of ad hocness. Okay. So let me now actually take you through what Popper says about ad hoc theories. Okay. Versus what he calls bold theories. When we propose a theory, even just an auxiliary theory to save a theory, Popper says we must never make it ad hoc. And he objectively explains what counts as ad hoc. Therefore, ad hocness encompasses and subsumes Easy to Varyness, but in a way that is less subjective.

[01:21:47]  Bruce: And when I say subsumes, it actually exceeds it. Okay. It captures not just Easy to Varyness, but also these other bad arguments we've been talking about that aren't clearly Easy to Varyness. And it does so in a way that's very straightforward. At least in the case of empirical theories, there's still this problem of applying it to philosophical or metaphysical theories, because it's going to be entirely, at least Popper's version of it, is entirely rooted in testability, which, recall, Deutsch denies, applies to Easy to Vary theories. I'm going to show why, in fact, you can test Easy to Vary theories using the concept of ad hocness.

[01:22:26]  Peter: Can we just define, ad hoc would be something just created on the fly to save a theory?

[01:22:32]  Bruce: That is the way you would commonly use the term, but Popper defined, that would still be subjective. Popper defines it in an objective way. I'm going to actually use the quotes from Popper to show how he defines it. The first thing we have to understand is the Easy to Vary vs Hard to Vary dichotomy is first of all not a dichotomy or anything even like a dichotomy. It actually captures three intermingled issues as if they are a single issue, which is why it's a confusing way to try to state things. The first thing it captures is how precisely the theory is stated via axiomatic statements that make it easy to see if you are surreptitiously changing the theory on the fly. Similar to what I previously called in the context of Deutsch's Constructor Theory of Knowledge and all the past episodes, degrees of freedom. Okay, this is episode 75 to 80.

[01:23:22]  Bruce: So that's the first thing that Ad Hoc, that Easy to Varyness captures. The second thing is how much reach does the explanation have? Now in a scientific setting, reach, scientific meaning empirical sciences, reach is identical to saying empirical content, or that the theory has testable consequences separate from the problem you want to solve. It is less clear how you apply the concept of reach to a philosophical or metaphysical theory, so that's still a problem. But for scientific theories, reach can be understood as exactly equivalent to having empirically testable consequences other than the problem you've introduced the theory to solve. Okay, and the third thing that Easy to Varyness captures is whether or not, and this is the important one, whether or not the person defending the theory is willing to ad hoc save the theory from a refuting counter example by introducing untestable auxiliary hypotheses.

[01:24:17]  Bruce: Similar to number two, your auxiliary hypotheses meant to save your theory must also be expected to have reach and have testable consequences other than the consequence that you're trying to solve. In other words, Easy to Vary vs Hard to Vary is partially about the nature of the theory and partially about the critical attitude of the defender of the theory. The theory itself must be stated as precise axioms and must have testable reach, and the critical attitude of the defender requires that all auxiliary hypotheses meant to avoid refutation are themselves precisely axiomatic theories that have testable reach. The two are very deeply intertwined. The critical attitude of the defender and the theory itself. You cannot easily tease them apart under the guise of Easy to Varyness. I'm also going to argue that of these three, the critical attitude of the defender is the most important of the three.

[01:25:16]  Bruce: Now, let me admit up front that this is very much like trying to decide which blade of the scissors is more important, and that really, since number three, the critical attitude, is defined by the first two, that you really can't say one of these points is more important than the other. But I'm going to make a case anyhow, at least a subjective case, that the critical attitude is the most important of the three criteria.

[01:25:39]  Peter: Are we kind of saying that Popper is agreeing with Vaden here, where the concept of hard-to-varyness is a critical attitude that we can apply to ourselves more than other people?

[01:25:55]  Bruce: Well, let's ask that question and let's see how it gets answered as we read through Popper. So when it comes down to it, it's only a proper critical attitude that can stop quantum mechanics from suddenly becoming an easy-to-vary explanation. So that means it's basically impossible to have a theory so precise and with so much reach that an improper critical attitude can't immediately make it easy to vary. In a future podcast, I will also argue the inverse of this, that there is no theory so easy to vary that with the proper critical attitude, you can't decide to axiomatize it in such a way that it becomes testable and thus has reach. This is why spontaneous generation and creationism can be tested and refuted by someone with a proper critical attitude. Again, I'm talking about potential scientific theories here. We still have the problem of how to apply all this to philosophical or metaphysical theories.

[01:26:51]  Bruce: But let's look now at what Popper actually says. A good theory is not ad hoc. The idea of ad hocness and its opposite, which perhaps may be termed boldness, are very important. Ad hoc explanations are explanations which are not independently testable, not independently testable, independently that is, of the effect to be explained. They can be had for the asking and are therefore of little theoretical interest. This is from Objective Knowledge, pages 15 to 16. This is why ad hocness is, unlike easy to varyness, not subjective. Your theory either has other testable consequences or it doesn't. It's up to you to work out what they are. The burden of showing that your theory is not ad hoc is on you, not on the people you're arguing with. This, I think, actually answers the question. While easy to varyness is subjective, ad hocness just is not.

[01:27:50]  Bruce: It is easily detectable whether you're being ad hoc or not. It's by simply asking, what is the other testable consequence? If you don't have any, you're ad hoc. Now, I do want to emphasize that ad hoc explanations aren't necessarily wrong. There's a difference between being wrong and being ad hoc. But ad hocness is very important to understanding critical rationalism. So about independently testable. Here's Popper again, this time in Conjecture and Refutation, page 327. We require that a new theory should be independently testable. That is to say, apart from explaining all the explicanda, which means that which we're trying to explain, which the new theory was designed to explain, it must have new and testable consequences, preferably consequences of a new kind. It must lead to the prediction of phenomena which have not so far been observed.

[01:28:41]  Bruce: This requirement seems to me indispensable, since without it, our new theories might be ad hoc, for it is always possible to produce a theory to fit any given set of explicanda. Okay, now I wonder if we've really given enough thought to this. Let me actually dwell on that just slightly. Well back in time, prior to Newton's theory, we had these theories of how the planets rotated and are they ellipses or are they circles, things like that. We had these theories of planetary rotation and we didn't explain it in terms of gravity. There hadn't yet been a unification of why things fall to the Earth and why planets follow ellipses when they're circling the Sun or whatever. And the way they would quote explain these planets is they used something called epicycles. Now, when they would find an exception case, when they would say, oh, we've now shown that this planet doesn't follow our epicycles, they would just insert more epicycles.

[01:29:45]  Bruce: And you could always do this. You could always come up with an explanation using the epicycle explanation. Any set of data could be fit to the epicycle explanation by just simply adding more epicycles. This is really what we're talking about here, okay? This is the core of the concept of easy-to-varyness versus hard-to-varyness, is this idea that we don't want epicycle explanations because they can always be found. It is always possible, always, always possible to be immune to refutation, to a counter example, if you're willing to use epicycle-style explanations, okay? And this is what I think we really are capturing with the concept of ad hocness versus boldness.

[01:30:32]  Peter: Sounds like it must have a place in the world, too, these ad hoc saves as well in that, I mean, what if epicycles were an explanation? I mean, it's certainly something worth exploring. I'm sure you can think of plenty of examples where someone comes up with an ad hoc save and it ends up being right.

[01:30:50]  Bruce: I think that that's an incredibly important point. And I think you absolutely must understand ad hocness is not being the same as wrong, right? Because ad hocness, an ad hoc save could be right.

[01:31:02]  Unidentified speaker: Okay.

[01:31:03]  Bruce: And it's very important to understand that. That's going to be the topic for the next podcast, by the way. So I will go into that, how to understand the difference between these. Now ad hocness differs from easy to varyness in an important way. It includes our critical attitude as much as it applies to the theory itself. So here's a quote from Popper. It might be said that even if the asymmetry of refutation vs verification is admitted, it is still impossible that any theoretical system should ever be conclusively falsified. For it is always possible to find some way of evading falsification. For example, by introducing ad hoc and auxiliary hypothesis, or by changing ad hoc a definition. It is even possible without logical inconsistency to adopt the position of simply refusing to acknowledge any falsifying experience whatsoever. Now, one of these I want to call out, just a little bit, changing ad hoc a definition.

[01:32:00]  Bruce: This idea that you can immunize your theories by ad hoc changing definitions is a common problem with all online internet discussions and also with crit rats. In episode 70, you brought up the question of is disobedience key to understanding general intelligence? And I argued, I doubt it. Okay, so is disobedience unique to universal explainers? Might be the question. We might then offer a refuting observation. Well, dogs are disobedient and they're not universal explainers. Okay, that's a fair observation, but we can easily, through redefinition, avoid the refutation. Let's just redefine disobedience as doing something new, or basically set it equivalent to human level of creativity. Now disobedience is unique to universal explainers by tautology, problem solved. Don't try to look this definition up in a dictionary, you won't find it. And this causes some problems.

[01:32:55]  Bruce: Einstein is now discovering general relativity through disobedience, and when you rob a bank, you're actually not being disobedient, right? But if we're allowed to define any term any way we want while selectively ignoring the original concept the term pointed to, sure, you can make any problem magically vanish. You can always do this. It is an all-purpose way to immunize your theories. This is what Popper is saying. Okay, so this is an issue for a separate podcast. I won't go any further into it. But here's Popper's point that you should not want to change your definitions ad hoc to immunize your theories. Why? Popper goes on to explain, I must admit the justice of this criticism. So bear in mind, the criticism he's admitting the justice to is that there is no such thing as a conclusive falsification. Okay, because there's always ad hoc auxiliary hypotheses or changing definitions ad hoc that you can do to avoid the criticism or avoid the refutation.

[01:33:51]  Bruce: Okay, but Popper goes on to say, but I need not therefore withdraw my proposal to adopt falsifiability as a criteria of demarcation, for I am going to propose that the empirical method shall be characterized as a method that excludes precisely those ways of evading falsification. Okay, so put another way, critical rationalism, Popper's epistemology, is exactly equivalent to a choice to not introduce post-Ad Hoc auxiliary hypotheses or ad hoc changing definitions to immunize your theories. That is what a critical attitude is. Okay, and notice that Popper is defining critical rationalism less in turn, he's not saying there isn't such a thing as an ad hoc theory. He calls them ad hoc theories, right? But it is partially a nature of the theory. It doesn't have testable consequences. And partially a nature of your attitude that you have to decide not to introduce them.

[01:34:51]  Bruce: Even though they could be true. This does lead to the problem, of course, what if they're true? That's what we're going to do in the next podcast. So I'm not going to go into that right now. Because there is a way to deal with that. Okay, again, note this applies to the empirical method. In a way, for those who claim that there is no scientific method, here Popper is saying, yes, there is an empirical method.

[01:35:10]  Unidentified speaker: Okay.

[01:35:11]  Bruce: But it does not apply, it only applies to empirical methods and theories, not metaphysical ones. So we still have the problem of how to apply this to philosophical theories. Now, here's Popper on Reach. This concept of boldness as the opposite of ad hocness maps very well to the Deutschian concept of reach. In essence, your theory must have some consequences other than the problem you're trying to solve. This is what Popper is saying. So, quote, this is from Objective Knowledge, page 192. In order that the explican- explicans, that is the explanation, should not be ad hoc, it must be rich in content, i.e. have reach. It must have a variety of testable consequences, and among them, especially testable consequences, that are different from the explicandom, i.e. the thing we're trying to explain with the theory. Popper explains that independent is the opposite of ad hoc and in the extreme, circular.

[01:36:04]  Bruce: This is page 192. This is the problem with Dean defining belief in democracies as a means of reducing violence, as upfront loving violence. It is now just a circular argument without testable consequences, which is, by the way, why Dean liked it. Okay. Why is this important that our explanations have testable consequences different than the problem we're trying to solve? This is why corroboration is an integral part of Popper's epistemology. Popper now explains. This is conjecture refutation, page 244. For it is admitted that a theory may be ad hoc if it is not independently testable by experiments of a new kind, but merely explains all the explicanda, including the experiments which refuted its predecessors. Then it is clear that the mere fact that the theory is also independently testable cannot as such ensure that it is not ad hoc. Okay, that was a hard paragraph.

[01:37:08]  Bruce: Let me try to explain it. Popper is saying that even if we continually refuted one theory after another in favor of a better theory via crucial tests, our theories might still be ad hoc if they don't have entirely additional testable consequences that we can independently corroborate. An example. Remember that general relativity, before it was ever tested in the Eddington Expedition, explained the perihelion of Mercury and the Morley and Michelson light speed observations. At this point, you might be tempted to say, oh, this is now our best explanation. Okay, because not only does it explain all previous observations that Newton's did, but it also explains two observations that Newton can't. Okay, at this point, though, the theory has been made to fit those observations, those known observations, and therefore could be entirely ad hoc. The way we test that it's not ad hoc is the Eddington Expedition, okay, that we set up this separate test that nobody's ever thought to do, that we only thought to do because of Einstein's general relativity.

[01:38:14]  Bruce: And we go out and we get this other observation that's completely independent of why the theory was being created. And once it passes that test, at this point we have a principled reason to believe that the theory is not just ad hoc, okay? This is why corroboration is so important to Popper's epistemology. By the way, Popper in one paragraph covered what took me three hours to cover in a single podcast in episode 61. You know, Popper is far pithier than me. Let's just say that much.

[01:38:47]  Peter: It still seems to me that a lot of Popper's, the quotes you're reading from him ring true more as an attitude rather than a methodology, but it sounds like you're reading him as more of a methodology.

[01:39:02]  Bruce: Well, you know, that's an interesting question. What I actually pointed out is that he did refer to it as a method.

[01:39:08]  Peter: Yeah, I was going to say, he even called it an attitude at one point. And even, I mean, that kind of fits in with the Eddington Expedition too. I can't remember if it was you or someone else was saying that the, you know, the Eddington Expedition is framed in a certain way as providing proof for general relativity, but that in reality, it was kind of just like one piece of a very intricate puzzle.

[01:39:33]  Bruce: That is correct. It's Deborah Mayo who used that. And I think that's definitely true. At that point, a problem existed with Newton. The other two problems, the perihelion of Mercury and the Morley and Michelson experiments, were already quote problems for Newton's theory. This was a whole new level of problem, right? That we had now corroborated it through a completely independent test. But at that point, it still wasn't an accepted theory. It had to go through extensive additional corroboration before really we put away Newton's theory and accepted general relativity as a replacement for it. And that was actually the correct critical attitude.

[01:40:13]  Peter: Yeah, well, see, that's why it rings true more for me as an attitude. I mean, if it was an algorithm, I think it would be much more obvious, oh, you have to do this, this, and this to your theory, and then it's true or more true or something. But really, it's more of a, you want to have the right attitude. Maybe the algorithm is the attitude towards these theories.

[01:40:34]  Bruce: So you're raising a question that I'm not prepared to answer because I need to go actually pull all the quotes. But let me try to summarize my view. From my reading of Popper, he does say that there is no scientific method, and he also says that there is an empirical method, which is the same as the scientific method. He does this fairly consistently. He defines science as empirical theories. So those are the same thing. I think the way you resolve this is, and I need to go pull the actual quotes, is that what he's really saying is is that the scientific method includes this idea of conjecture, which has no method. Conjecture is just to creatively come up with stuff. It's true once you have the conjecture, there is more or less a method to how to go about criticizing it, basically by avoiding ad hocness, right? You make everything independently testable.

[01:41:26]  Bruce: So science does have a methodology, but it's more like science is conjecture refutation and it's refutation that has a methodology. Not conjecture. There is no methodology of conjecture, if that makes any sense.

[01:41:41]  Peter: No, that's an interesting way to put it.

[01:41:43]  Bruce: Okay, and I think that's what he's actually saying. Now, I need to go pull the actual quotes. Don't hold me to it. When I actually go pull the quotes, I may find that I changed my mind. But we'll do a podcast on that, where I actually pull the quotes and we read them together and we determine for ourselves what it is we think Popper was actually saying. Now, I want to point out though that Popper is exclusively talking about empirical reach here. Like, whenever he brings up what I'm calling reach using the Deutsch term, it's always empirical testability that he's talking about. He never really differs from this. And that's why trying to apply this to philosophical theories does become a much more difficult thing to do. Now, recall that Brett's Theory of Intelligence has, by his own admission, zero empirical reach, has no testable consequences.

[01:42:29]  Bruce: But that he thought his theory had reach anyhow because of moral consequences that he thought came from his theory. Such as the idea that it is immoral to measure a person's intelligence because all universal explainers have the same intelligence. You're going to discourage them from studying what they really want. And obviously, there's a legitimate moral issue here that he's raising, and I'm not saying that there isn't. The way we treat kids and give them IQ tests is sometimes rather reprehensible. But let me point out that communists make similar claims. They also have a theory with zero empirical reach, but has strong moral reach due to its supposed explanatory content. So communism has got this, as Popper points out, it was very attractive to people because it explained things so well. It was this strong, quote, explanatory theory. And this is what appeals to the defender of Brett's theory of intelligence, is that it's a strong, quote, explanatory theory.

[01:43:29]  Bruce: And it has reach, in that it has moral reach, even if it has no empirical reach. Now, I did point out, though, that Brett also ignored that his theory may have moral reach that he doesn't want. For example, according to his theory, we should probably sit down with a severely mentally challenged individual and give them a good talking to about how they need to use their universal intelligence to not be a burden on others. Now, of course, that would be silly. I'm not advocating for that, nor is Brett. But that would seem to be part of the moral reach of his theory if we took his theory seriously. And it's a part that he doesn't... I don't know how he deals with a problem like that, so I won't speculate.

[01:44:08]  Peter: It opens up a can of worms if we take it really seriously.

[01:44:11]  Bruce: Okay. So here, by the way, Popper only accepted empirical reach as counting as reach, not moral reach. This is the point I'm trying to make. So for Popper, explanatory content was not a good thing unless it was empirically testable. That's why communism was bad. That's why, according to Popper's epistemology, Brett's theory of intelligence is bad. This is why Popper has a low view of communism despite its high explanatory content, all of which was moral and none of which was empirically testable. So I'm going to introduce Bruce's rule of moral reach. This is the rule. If your theory has only moral reach and no empirical reach, you have a meaning meme, not a good explanation. And essentially, from a Popperian standpoint, come back when your theory has empirical reach as well, and at that point, we'll take your moral reach seriously also.

[01:45:03]  Peter: I like that.

[01:45:03]  Unidentified speaker: Could you state the rule one more time?

[01:45:05]  Bruce: If your theory has only moral reach and no empirical reach, you have a meaning meme, not a good explanation.

[01:45:11]  Peter: I like that.

[01:45:12]  Bruce: What does Popper say about good vs bad explanations? He does in fact talk about this and does it in terms of ad hocness instead of easy to varyness and hard to varyness. So, Conjecture and Refutation, page 81. One can show that the methodology of science, notice there, methodology of science, and the history of science also become understandable in its details if we assume that the aim of science is to get explanatory theories that are as little ad hoc as possible. A good theory is not ad hoc, while a bad theory is. In fact, I'm going to go on a limb here. This is where Deutsch is taking the idea of a good explanation is easy to vary and a bad explanation is hard to vary. And notice that the Popperian version is not easy to vary and hard to vary, but ad hoc or not ad hoc. Okay. This is why I actually think ad hoc is the Popper, the Karl Popper equivalent to Easy to Varyness, but that it's a stronger version of it.

[01:46:13]  Unidentified speaker: Okay.

[01:46:14]  Bruce: Recall that ad hoc means has no empirically testable consequences other than the one problem we introduced it to solve. So a good theory is one that has independently testable consequences and a bad theory is one that doesn't. Notice that Popper puts this in terms of testability, Deutsch is claiming that it comes before testability. Now, this is page 193 now. The question, what kind of explanation may be satisfactory thus leads to the reply, this is Karl Popper, obviously, an explanation in terms of testable and falsifiable universal laws and initial conditions. Now, you may have noticed that this is equivalent to what Chiara Marletto and David Deutsch call the traditional conception of fundamental physics and that they claim as an inadequate mode of explanation. Taking this from Chiara's book, I shall call the traditional conception of fundamental physics that all fundamental physics theories must be formulated in terms of predictions about what happens in the universe given the initial conditions and the laws of motion.

[01:47:21]  Bruce: That's on page 26 of her book. Notice that that maps almost exactly to what Popper says is a satisfactory explanation. The definition what kind of explanation is satisfactory. This is an issue I want to bring up, but I'm not going to resolve today. And I think that Chiara and David would here say, look, we're not saying that this is a bad mode of explanation. We're just saying it's an inadequate mode of explanation. That some things can be explained better via other modes of explanation. However, I do want to just really emphasize here that when Popper asked the question what kind of explanation may be satisfactory, in other words, what is a good explanation, he says the reply is an explanation in terms of testable and falsifiable universal laws and initial conditions. So he roots the idea of a good explanation in terms of testability. So to Popper, a good explanation was an empirically testable explanation that has reach, by which we mean, in this case, independently testable consequences, by which we meant empirical consequences other than the problem we're trying to solve.

[01:48:33]  Bruce: Again, this leaves the question hanging, what about metaphysical theories? When critiquing Deutsch's Constructor Theory of Knowledge, I pointed out that the theory has points that aren't explained precisely. Example, how long does adapted information need to keep itself instantiated to be considered knowledge? Thousands of generations, thousands of years? Does it need to be a replicator, or just a variant that outcompeted its other variants, even if it's not a replicator? Does it need to be a recipe in the sense of constructing something, or can it merely say cause something to transform like, say, walking? Okay, I raised all these issues across all those podcasts that we did. The Constructor Theory of Knowledge was imprecise in many ways, and this allowed the defenders of the theory to always find a way to defend the two sources hypothesis, which, recall, is the hypothesis that all knowledge comes from only two sources, Darwinian evolution and human ideas.

[01:49:31]  Bruce: So they would defend the two sources hypothesis by simply applying those imprecise criteria differently depending on the circumstance. If I were to ask you, is Deutsch's Constructor Theory of Knowledge Easy to Vary, it would seem silly to answer yes. What is really going on is that the defenders of the theory are using the degrees of freedom to ad hoc save the two sources hypothesis and to be sure it can't be refuted. But they didn't have to do this. They could have removed the degrees of freedom by making the theory more explicit and precise than taking the consequences of that clarification. For example, they could say, I say it has to be a thousand generations to be counted as knowledge. Take a stance, right? Be precise and take a stance. And then to which we can now respond, okay, then airplane designs aren't knowledge because we haven't yet even gone through a thousand generations of airplane.

[01:50:25]  Bruce: Okay, so at that point, we now have, because we were precise in what our, we removed the degrees of freedom and became precise in what we meant, it's now easier to come up with counter examples. There was a choice being made to defend the theory by refusing to increase the precision of the criteria for what counts as knowledge when confronted with counter example. With this in mind, you can now understand why I keep trying to treat Deutsch's criteria for knowledge as axioms rather than vague illustrations. So now here is from Logic of Scientific Discovery, page 50, Karl Popper again. For a severe test of a theoretical system, i.e. an explanation, presupposes that the theory is at the time sufficiently definitive and final in form to make it impossible for new assumptions to be smuggled in. In other words, the system must be formed sufficiently clearly and definitively to make every new assumption easily recognizable for what it is, a modification and therefore a revision of the system.

[01:51:30]  Bruce: He goes on to say, This, I believe, is the reason why the form of a rigorous system is aimed at in science. It is the form of a so-called axiomatized system. The attempt is made to collect all the assumptions that are needed, but no more, to form the apex of the system. They are usually called axioms. So I was treating Deutsch's Criteria of Knowledge as axioms then working out the consequences of those axioms. And the defenders of the theory were taking issue with me doing that because they felt that they wanted to treat them more as kind of vague illustrations. Sure, it's adapted information that keeps itself instantiated or has causal power, but that doesn't... And yes, the walking robot example fits those criteria, but that's not really what we mean by knowledge. And I would... And I admit to the... I admit to the subjective quality of the argument that they're making.

[01:52:24]  Bruce: But the reason why I kept going back to... But let's treat them as axioms. What then? Is because of what Popper's saying here. That's part of critical rationalism. That you have to make a choice to make your theory precise enough that it's obvious that if you have a refuting example, and it's also obvious whether you've now changed it. Go back to that example of where I said we can imagine someone saying it needs to be a thousand generations to count as knowledge. You can then immediately say, well, airplanes aren't knowledge. And the person then can say, oh, you're right. I'm going to now change that. But now you know you're changing it. You're not accidentally changing it. You're not accidentally moving it around arbitrarily to try to get rid of all refutations. It's forcing you to think more deeply about the theory and to error correct the theory.

[01:53:12]  Peter: It's more refining.

[01:53:13]  Bruce: Yes. You can solve the problems by refining the theory. Or you can solve the problems by being more vague and make the problem disappear. So why do we axiomatize? So this is Karl Popper. In a theory thus axiomatized, it is possible to investigate the mutual dependence of various parts of the theory. For example, we may investigate whether a certain part of the theory, in our example, the two sources hypothesis, is derivable from some part of the axioms, which in this case I'm saying is the criteria for knowledge. So if we treat Deutsch's criteria for knowledge as axioms, then offer a counter example, we can now determine a specific flaw in the Constructor Theory of Knowledge, namely the two sources hypothesis. But if we just refuse to axiomatize them, as the defenders were doing via an intuitive kind of hand-waving, isn't it just obvious the two sources are special?

[01:54:06]  Bruce: Then it becomes impossible to tell what the problem is. This explains why I kept insisting we treat the criteria for knowledge as axioms, and kept insisting that we take seriously the fact that the two sources hypothesis can't be derived from the current axioms, and I refuse to accept the hand-waving ad hoc saves being offered to me, no matter how intuitively obvious they may feel. I should note that Popper admits that science is not fully axiomatized. So this is more like a goal than something that is just always true. And here Popper says this, Although the theories of physics are in general not completely axiomatized, the connections between its various parts may yet be sufficiently clear to enable us to decide which of the subsystems are affected by some particular falsifying observation. This is a little more than I need in terms of the argument I'm making, but what Popper is trying to say here is that one of the reasons why, there's various reasons why we try to make our theories so precise that we treat them like they're axioms of their logic, axioms of some logical system.

[01:55:12]  Bruce: The reason why we, one of the reasons why we do that, is because that's a way we solve the Duhem–Quine problem, where if you understand the system at a detailed enough level as axioms, and you have a falsifying observation, you can then use this detailed system to figure out which part of the subsystems are affected. You might be able to determine, for example, well, actually the theory itself is unaffected, it's one of the auxiliary hypotheses.

[01:55:37]  Unidentified speaker: Okay.

[01:55:38]  Bruce: So he's trying to explain, this is why we axiomatize our theories, this is why we turn them into logical statements and get really precise about them, is because it's directly related to what we can learn by a falsifying example. In fact, it's directly related to if you even can falsify the theory in the first place. Now, I've often pointed out how common it is for crit rats to claim their theories are philosophical or metaphysical, by which they mean not empirically testable, yet still want to claim their theory is a best theory. So that was one of the arguments that got used with the defenders of Brett's theory of intelligence. This is, to some degree, what Dean is arguing with critical rationalism and libertarian. He argues that anarcho-capitalism follows naturally from critical rationalism, and Jose was arguing with him that that wasn't true.

[01:56:27]  Bruce: Now, it's well known that the Austrian School of Economics is famously untestable, right, and that this is actually often treated as a good thing. You follow from these axioms and you work stuff out, but you can't actually test it. This is not the same as critical rationalism. This is, in some sense, the opposite of critical rationalism. Okay? Now, that doesn't mean these theories are worthless. And Popper always emphasizes this, that metaphysical theories are not worthless, right? They often have a lot of uses. But we do need to understand that you don't get to just make up a metaphysical theory and offer it as a competitor to an empirical theory and then claim, well, mine's more explanatory, by which we really just mean, you know, it's got no testable consequences, but it seems like a good explanation to me. And this is not what Popper's epistemology is about.

[01:57:16]  Bruce: In fact, it's the opposite of what Popper's epistemology is about. I would argue that the Constructor Theory of Knowledge's Two Sources Hypothesis is also an example of a basically philosophical or metaphysical theory that's untestable the way that it's getting treated today. When I point this problem out, I routinely get told, oh, so you're saying that every single empirical theory is more important than every single metaphysical theory, including Popper's own epistemology. Is that it? You know, cue lobster meme here. I've had this argument numerous times, and I'm always told this by the crit rats the moment I point out, look, you're trying to advance a metaphysical theory as a competitor to an empirical theory. You're not supposed to do that in Popper's epistemology, and this is always the argument that they use. But here's what Popper actually says.

[01:58:07]  Bruce: A theory which is not refutable by any conceivable event is non-scientific. Irrefutability is not a virtue of a theory, as people often think, but a vice. Every genuine test of a theory is an attempt to falsify it or to refute it. Testability is falsifiability. But there are degrees of testability. Some theories are more testable, more exposed to refutation than others. They take, as it were, greater risks. That's why he calls it boldness instead of hard-to-varyness. Some genuinely testable theories, when found to be false, are still upheld by their admirers, for example, by introducing ad hoc some auxiliary assumptions, or by reinterpreting the theory ad hoc in such a way that it escapes refutation. Such a procedure is always possible, but it rescues the theory from refutation only at the price of destroying, or at least lowering, its scientific status.

[01:59:01]  Bruce: This is what I'm really trying to say to the crit-rats I'm arguing with. When I bring up an empirical theory, IQ theory, which is empirical at least a little bit, and you bring up a metaphysical theory as a competitor, that's not a good thing, right? That absolutely must be seen as a vice under critical rationalism. This is all, by the way, conjecture refutation, page 48. So yeah, Popper is pretty harsh on the idea that you can just trot out a metaphysical theory as a competitor to an empirical theory. So how does ad hocness solve our problems here? So let's take a look at Darwinian evolution again. Yes, large parts of the theory are highly empirical due to how the explanation fits, details strongly connected to the details of the phenomena to invert Deutsch's statement of Easy to Varyness. And that is why I feel that Popper is wrong, that Darwinian evolution is not testable.

[01:59:56]  Bruce: I could give you many, many examples of where we have tested and corroborated evolutionary theory. I will admit that there is no way to refute Darwinian evolution. Now this is an- I think it's an issue that we're going to have to not cover in this podcast because it's already getting too long. The fact is, is that Darwinian evolution really isn't refutable, but it can be corroborated.

[02:00:18]  Peter: At what point in his career does he say that about Darwinian evolution? I'm just curious. Was this the 30s or 40s or?

[02:00:26]  Bruce: You know what? It was in his last book. Saadia quoted it to me, where he had to actually ask her. So he went apparently to his grave.

[02:00:34]  Peter: So it was in the 80s or 90s then?

[02:00:36]  Bruce: Yeah, he was still claiming Darwinian evolution. But I'm going to explain this in the next podcast, I swear. Why is it that Darwinian evolution can be corroborated but not refuted? I'm cheating when I say that. There's a certain sense in which it can be refuted. And that's what we're going to cover in the next podcast. But there is a certain sense in which it's just true. There are no refuting observations to Darwinian evolution in the sense that we would actually drop it as a theory. But you can corroborate it through experiment. And we have numerous, numerous times. You think about when we discovered DNA, and then we discovered how DNA has to play a role in Darwinian evolution. And then we discovered, I mean, think about how once they discovered DNA, they can start to say, this is how these two animals or these two plants are related to each other on the evolutionary chain.

[02:01:27]  Bruce: Why can you even do that? Like, if creationism, just kind of a naive creationism was true, that evolution played no role, there would be no reason for these two different plants to have the same DNA. You absolutely are testing Darwinian evolution in a very deeply empirical way when you start looking at DNA. This is why I really have a problem with trying to call Darwinian evolution a metaphysical theory like Popper sometimes did. And yet, I understand why he had this concern, because it was more than a little obvious that you couldn't just refute Darwinian evolution using observation. Okay, and that's going to be the topic for the next podcast, so I'm going to move on. So Darwinian evolution, taking a look at it as to how ad hocness solves the problem. So large parts of Darwinian evolution are highly empirical due to how the explanation quote fits the details strongly connected to the details of the phenomena.

[02:02:24]  Bruce: That's an inversion of Deutsch's statement as to what Easy to Varyness is. So it shows that it's a hard to vary theory. But this is exactly identical to saying the theory has a lot of empirical content or it has reach. And scientific theories need not have high empirical content or would be impossible to get scientific theories off the ground. We've talked about that extensively as part of this podcast, but I've brought up the whole idea of oranges stop scurvy. There is no mechanism or explanation being offered. That is still enough to start to have an empirical content, right? We can test the idea oranges stop scurvy very easily through a scientific experiment. Now I'm not an idiot. Of course, back when I was a creationist as a kid, I went for where I knew evolution couldn't explain things yet. And I offered counter examples that I knew had no hard to vary explanation yet.

[02:03:17]  Bruce: Offering the Miller-Urey experiment as an explanation for how life got started is an explanation that explains basically nothing. What it does is it pumps your intuitions from, well, a tiny part of the process took place on its own. So probably the entire freaking process can happen on its own. It's induction, pure and simple. It is an ad hoc explanation because it has no empirical content of its own. Here's the distinction I want to make, okay? Darwinian evolution is not an ad hoc theory, per se. It's got its own testable consequences. It may not have high empirical content as high as, considering what a best-in-class theory it is, Popper's right that it does not have a ton of empirical content compared to say a physics theory, okay? And I think this is why he has this intuition that there's something wrong with Darwinian evolution is because he's trying to think of it like a physics theory.

[02:04:13]  Bruce: But it does have some empirical content and there is no minimum amount of empirical content you need to have to be able to start being a scientific theory, okay? When the defenders of Darwinian evolution offered the Miller-Urey experiment as an explanation for how life got started, that itself is an ad hoc explanation or an ad hoc theory. It's not Darwinian evolution that is. And this is the problem, is that theories are theoretical systems, they're all interconnected. This is why an easy-to-vary theory can be treated like a hard-to-vary theory. You can simply start to axiomatize it. This is why a hard-to-vary theory like quantum physics or Darwinian evolution can be treated like an easy-to-vary theory. You simply introduce ad hoc auxiliary theories to try to deal with problems. Let me put this in a somewhat different way. It might be easier to speak of not easy-to-vary theories, but theories being easily varied.

[02:05:24]  Bruce: And this gets closer to what I think the truth is, which is that it's more about the attitude of the defender and less about the theory, okay? All the examples that we've talked about, where we said these are bad arguments, it's not clear that their theories are easy to vary. It's very clear that they are easily varying their theories. And this is what ad hocness I think really is, how it connects to easy-to-varyness. So there are no easy-to-vary theories. So likewise, spontaneous generation is clearly a case of a theory that has very low empirical content, but not none, okay? Crit rats are quick to dismiss competing theories on the ground that it doesn't explain, by which they really mean you can't offer me a comprehensive explanation of every mechanism involved in the process all the way down. But no theory can do this, not even Darwinian evolution or quantum mechanics.

[02:06:20]  Bruce: So this argument is itself easy to vary, or it's ad hoc, because you just trot it out when you don't like the theory and you ignore it when you do. This is a clear case of subjectively deciding where to draw the line of what counts as lacking explanatory content, depending on whether you like the theory or not. So this is the distinction. Some theories are easier to vary, but all theories can be easily varied. Some theories are harder to vary, but all theories have some spots that are not currently testable, quantum mechanics in many worlds, for example, or that aren't axiomatized yet and thus is open to interpretation, such as with the constructor theory of knowledge and the two sources hypothesis. And these spots, these degrees of freedom, can be easily varied. This is true of all theories. It is true that if a defender of a theory is being dogmatic, they will prefer a theory that can't be easily criticized or tested.

[02:07:21]  Bruce: Thus, some theories are much easier to vary than others, that is, they have more spots that can be easily varied. But ultimately what matters is not if the theory itself is easy to vary, but if the defender is willing to easily vary it. No theory is entirely protected from a dogmatic attitude, where the defenders of the theory may choose to routinely immunize the theory from criticism or refutation. Not even best-in-class theories like Darwinian evolution and quantum mechanics are immune to Easy to Varyness or rather Ad Hocness. So in summary, this explains each of the examples that I believe it explains each of the examples that we have looked at. Two sources hypothesis, the defenders of the hypothesis rely on vague intuitions about isn't it just obvious that there's something special about the two sources? Instead of treating the criteria of knowledge as axioms and taking the consequences, I think that is a bad choice.

[02:08:18]  Bruce: They should not be doing that. It would be okay if they were upfront about it. If they'd say, here's our axioms, I point to a problem. They go, okay, let me try to redo the axioms. In fact, they should be allowed to have any number of retries with their axioms forever to see if they can get it to a set of axioms that do define knowledge only as the two sources hypothesis. That isn't what they're doing though. What they're doing is, is they're hand waving. They're saying, isn't it just obvious? I can't believe you have a problem with this. That's where they've gotten off the tracks with critical rationalism. Brett's Theory of Intelligence, the defenders of the theory use an all-purpose argument. IQ theory doesn't explain things all the way down. Without that explanation, all statistical tests are to be considered meaningless. Even though no theory can do this.

[02:09:05]  Bruce: And except when it comes to theories they favor, then they accept it as evidence. They freely endorse a theory that they claim has no empirical reach, can't be refuted, but instead point to its moral reach. These are all examples of ad hocness. Dean's Libertarianism. Dean's Libertarianism relies on the Austrian School of Economics, which is famously non-testable. He endorses a circular argument, i.e. it has no reach. That loving violence is exactly equivalent to arguing against libertarianism. When challenged, he claims, Jose could have made a different argument and I would have accepted it, rather than offering how we can test his claims that democracies cause violence. In other words, his whole set of arguments was ad hoc. There were no independently testable consequences that we could talk about. All of these examples violate Popper's epistemology via ad hoc reasoning.

[02:09:57]  Bruce: But I know from experience that all of these defenders will simply say, and have actually said to me at times, yes, but this is a philosophical theory, so it isn't held to Popper's scientific epistemology. Which brings me to the question I've asked a couple times. What is the difference between a bad explanation and a philosophical theory? And of course, I can't answer that question today, because that's going to have to be its own separate podcast. And it's not exactly the easiest question to answer, let me admit. It is not that Easy to Varyness is an incorrect concept per se, but it is a misleading and maddeningly subjective concept compared to Popper's concept of ad hoc reasoning. No one ever feels like they're choosing to vary their theory to save it. They always believe they're being rational and that their proposed save just follows naturally from their theory.

[02:10:51]  Bruce: Even a gambler, prophet, or mystic, to use Deutsch's terms, feels this way. They don't see themselves as varying their theories for the purpose of saving them. They see it as raising real possibilities. And maybe they're right because ad hoc saves can be correct. We won't know until they try to formulate their theories testably. But the moment you choose to hold yourself accountable to make sure all your immunizations for refutation have their own empirical content that can be independently tested, it becomes impossible to keep fooling yourself. This is why I believe Easy to Varyness is an entirely subjective concept, whereas ad hoc reasoning subsumes and exceeds the concept in an objective way. Okay, that's my argument. And that's my summary.

[02:11:36]  Peter: So you're you're preferring Popper here to Deutsch.

[02:11:39]  Bruce: I am. I'm not really saying that Deutsch is wrong, though. Like I think he took the concept of ad hocness and he called it Easy to Varyness. Yeah, two concepts are really the same. But that he didn't define it objectively like Popper did as having independently testable. Now, why would he do that? Well, it's because he wanted to show that Easy to Varyness comes before testability. And I'm going to argue in the next podcast that he made a mistake when he claimed that that he had misunderstood something important about Popper's epistemology. And that's going to be the subject for the next. This has all been a setup for the next episode, where we're going to talk about what I understand Popper's methodology to actually be.

[02:12:25]  Peter: Well, I will look forward to that, Bruce. And this has been another fascinating podcast. And I love listening to you. And I can't wait to do it next time.

[02:12:38]  Bruce: All right. Thank you very much. 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.

[02:13:26]  Bruce: If you are interested in financially supporting the podcast, we have two ways to do that. 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.