Quotes are from the book Karl Popper, by Anthony O'Hear (AOH). It's in "The Arguments of the Philosophers" series edited by Ted Honderich. (Be careful, AOH has two other books with titles beginning with "Karl Popper".)

AOH says:

Popper's attempt to dispense with induction is unsuccessful. [ch. 4, p. 57]

AOH says his reason, which he'll attempt to show, is:

any coherent conceptualization of the experience requires the assumption of a stable order in the world. [ch. 4, p. 58, emphasis added]

Previously, AOH wrote:

But, argues Popper, we can see on logical grounds that there is no such thing as a perfect repetition of any event. Similarity in all respects would mean that the two events were really identical, and so there would actually be only one event. So the repetitions we experience are only approximate. But this means that some features of repetition B of event A will be different from some features of A. Thus B is to be seen as a repetition of A only to the extent that we discount those features in which B differs from A. [ch. 2, p. 13]

So AOH ought to address the question: "Stable in which respects?" He ought to know that the world is stable in some respects and not others, just as the future resembles the past in some ways and not others, and any two observations are similar to each other in some ways and not others.

Saying the world is "stable" means just as little as saying two observations are "similar". Claiming a stable world means claiming some things stay the same over time (or at least only change a small amount, according to some suitable measure). Of course not all things stay the same over time.

So AOH needs to say what type of stability he's talking about for his claim to mean anything.

One of the standard problems with inductivists is their routine failure to understand this general problem (that when we compare non-identical things they're always both similar and different, and you have to specify what sort of comparison you're doing). What does AOH do about this issue? Nothing. After the "stable" claim I quoted, he immediately changes the subject to solipsism. He's apparently unaware of this issue, even though he discussed it earlier in the book.

AOH proceeds (p. 59) to talk about regularities and patterns of experience without talking about which ones. Of course there are some regularities and some non-regularities in the world. AOH's approach to epistemology is basically "We live in a stable world, so recognize regularities and project them into the future." This is standard inductivist, and misses the point in the standard ways, such as the issue of which regularities to project into the future and how to find them (how does thinking work? AOH just takes for granted that we find these regularities somehow – that is, his epistemology presupposes intelligent thought and fails to explain how thinking actually works. He starts in the middle.) Then:

Our notion of an objective world, then, is reflected by the degree of continuing order and regularity that is to be found within our perceptions. [ch. 4, p. 59]

But Popper already explained the problem with this, and AOH already included that in this book. There is no such thing as "order" or "regularity" out of context. You have to first say which things you want to be the same which you'll count as being orderly or regular. Different aspects of the world are always similar (orderly, regular) in some ways and different in other ways. AOH doesn't address this.

I also found this bizarre statement:

That a belief in induction is not something which can be dropped without substantial alterations elsewhere in our conceptual scheme is why the failure of Popper to develop a truly non-inductive science is not a chance result, but one with deep roots. [ch 4, p. 60]

But Popper was aware of this issue, and wrote about it, and did develop substantial alterations in our conceptual scheme. I would understand if someone thought Popper's substantial alterations were mistaken, or if someone was unfamiliar with Popper's writing. But AOH has studied Popper a lot, and then is apparently unaware this substantial alterations even exist. AOH even quotes and discusses some of them, but apparently(?) doesn't recognize their meaning and importance. This is just like the similar in which respects issue, where AOH quoted Popper about it and discussed it – but then later on he writes as if he was unaware of it (which I take to mean he doesn't fully understand it).

the assumption that the world is not going to [suddenly become chaotic] [ch 4. p. 61]

The world is already chaotic in some ways and not others. So what does this mean? AOH doesn't say.

Does it mean the world won't suddenly become chaotic in all respects? But what would a world that is chaotic in all respects even mean? AOH doesn't address the issue and it's highly problematic.

One fairly technical way to approach the matter is via the theory of computation: consider whether there exist long bitstrings which can't be compressed by any compression algorithm (or, equivalently, can't be the output of any computer program, in any language, which is much shorter than the bitstring). Such a bitstring would be chaotic in all respects. But the answer is no, such a bitstring doesn't exist.

AOH might imagine that, all of a sudden, all the ways the world is stable stop working, and some new ones take their place. But that doesn't make sense, because no matter what happens, you can always retrospectively find regularities in the bigger picture including both before and after the so-called descent into chaos. All that's happened is this: from the infinitely many regularities compatible with the data you have, you favored some (why those? how were they chosen?), and found out those favored regularities were mistaken. (Meanwhile this so-called descent into chaos is fully compatible with some of the other data-compatible claims about regularities you could have made before it happened.)

So the assumption of the world's stability really means assuming your favored theories are correct. Why did you favor them over other theories, compatible with the same data, which make different predictions about the future? From the perspective of those rival theories, the future you predict is a descent into chaos. So when you say the world won't descend into chaos, you just mean the future will happen as you expect and not as your rivals expect – you mean the world will descend into chaos for the people who disagree with you, just not for yourself.

Thus, I am not simply saying that our ability to distinguish between true experience and illusions depends on our once having experienced an orderly world, but that it depends on the continuance of whatever order we had previously recognized. But to assume this is just what, according to Popper, is deeply irrational, and which should be eliminated from our conceptual scheme. [ch 4. p. 61]

Yes, it is irrational. Because it consists of assuming you're right.

What does "whatever order we had previously recognized" refer to? There are infinitely many theories compatible with the data you've observed previously. To recognize some order means to choose some of those of those theories (why those? why not others?) to provide order to your thinking. Then to assume the continuance of that order means to assume that your choice of which theories to prefer won't turn out to be mistaken in the future.

The solution to all this is what Popper said: critical and explanatory thinking (which is literally evolution). We can only conjecture which of the infinite regularities (or, preferably, explanatory theories) compatible with our data are correct. And we can correct errors with criticism, which is how progress is made. (Part of this is explained by AOH, pp. 171-177)

AOH also objects to Popper's corroboration, and I agree that corroboration is a mistake. I have fixed that aspect of Critical Rationalism. You can find my solution here. For a quick overview, I also offer a free short argument.

My new philosophy education product is ready!

Yes or No Philosophy

The link explains everything, has screenshots, etc. In short, it's about how to judge ideas, in particular by using yes or no judgements. It has a criticism of Critical Rationalism and a big improvement. (The same criticism also applies to standard philosophies of knowledge.) This is an epistemology breakthrough.

I put a short argument from the product online. It gives an idea of what I'm saying. The overall product is more focused on explaining things and helping people learn, rather than arguing, but I do include quotes from Karl Popper, David Deutsch and others along with criticism.

I put a lot of work into this and I'm really happy with it. It's going to help people learn more about philosophy! Especially if you've been interested in philosophy but find it difficult to get into, then this could help get you unstuck. I put a lot of effort into making it accessible.

Mac software used in this project:

• Screenflow
• Keynote
• Final Cut Pro X
• Compressor
• Ulysses
• Lightpaper
• Textmate
• Affinity Designer
• Numbers
• Adobe Acrobat
• KindleGen

Feel free to ask questions about Yes or No Philosophy in the comments below.

Alan blogged criticizing a "pro-truth pledge" which is actually a pro-authority pledge:

Defer: recognize the opinions of experts as more likely to be accurate when the facts are dispute

which experts? how much more likely? what about when, as usual, some experts disagree with each other? should we believe the more numerous or more famous experts, as the Global Warming lobby recommends?

as a philosophy expert, should everyone defer to me about everything? my field deals with how to think, reason, judge ideas, evaluate conclusions, etc. so i have expertise on whether, for example, a global warming expert is reasoning correctly about climate science, or not.

as a philosopher of science, do my claims trump the claims of all scientists? i've studied the scientific method itself and can evaluate whether they used it correctly or not.

please no. i don't trust my fellow philosophers to make wise kings ;)

saying to defer to experts changes the debate from

1) the debate about the actual issues

to

2) the debate about which people are experts, how much of an expert they are, and which type of expertise has priority in this case.

But it's way more productive to talk about (1), not (2)!

You often ask me for information/arguments that I have already given you

We're partially misunderstanding each other because communication is hard and we have different ways of thinking. I'm trying to be patient, and I hope you will too.

Please address these two questions about induction. Answering with page numbers from a book would be fine if they directly address it.

I've read lots of inductivist explanations and found they consistently don't address these questions in a clear, specific way, with actual instructions one could follow to do induction if one didn't already know how. I've found that sometimes accounts of induction give vague answers, but not actionable details, and sometimes they give specifics unconnected to philosophy. Neither of those are adequate.

1) Which general laws, propositions, or explanations should one consider? How are they chosen or found? (And whatever method you answer, how does it differ from CR's brainstorming and conjecturing?)

2) When and why is one idea estimated to have a higher weight of observational evidence in favor of it than another idea? Given the situation that neither idea is contradicted by any of the evidence.

These are crucial questions to what your theory of induction says. The claimed specifics of induction vary substantially even among people who would agree with the same dictionary definition of "induction".

I've read everything you wrote to me, and a lot more in references, and I don't yet know what your answers are. I don't mind that. Discussion is hard. I think they are key questions for making progress on the issue, so I'm trying again.

As a fallibilist, you acknowledge that the 'real world' is a contingent one and there are no necessary truths. But is not 1+1=2 a necessary truth? Is not 'All tall men are men' a necessary truth since its negation is self-contradictory?

I'll focus on the math question because it's the easier case to discuss first. If we agree on it, then I'll address the A is A issue.

I take it you also think the solution to 237489 * 879234 + 8920343 is a necessary truth, as well as much more complex math. If instead you think that's actually a different case than 1+1, please let me know.

OK, so, how do you know 1+1=2? You have to figure out what 1+1 sums to. You have to calculate it. You have to perform addition.

The only means you have to calculate sums involve physical objects which obey the laws of physics.

You can count on your fingers, with an abacus, or with marbles. You can use a Mac or iPhone calculator. Or you can use your brain to do the calculation.

Your knowledge of arithmetic sums depends on the properties of the objects involved in doing the addition. You believe those objects, when used in certain ways, perform addition correctly. I agree. If the objects had different properties, then they'd have to be used in different ways to perform addition, or might be incapable of it. (For example, imagine an iPhone had the same physical properties as an iPhone-shaped rock. Then the sequences of touches the currently sum 1 and 1 on an iPhone would no longer work.)

Your brain, your fingers, computers, marbles, etc, are all physical objects. The properties of those objects are specified by the laws of physics. The objects have to be used in certain ways, and not other ways, to add 1+1 successfully. What ways work depends on the laws of physics which say that, e.g., marbles don't duplicate themselves or disappear when arranged in piles.

So I don’t think 1+1=2 is a truth independent of the laws of physics. If there's a major, surprising breakthrough in physics and it turns out we're mistaken about the properties of the physical objects used to perform addition, then 1+1=2 might have to be reconsidered because all our ways of knowing it depended on the old physics, and we have to reconsider it using the new physics. So observations which are relevant to physics are also relevant to determining that 1+1=2.

This is explained in "The Nature of Mathematics", which is chapter 10 of The Fabric of Reality by David Deutsch. If you know of any refutation of Deutsch's explanation, by yourself or others, please let me know. Or if you know of a view on this topic which contradicts Deutsch's, but which his critical arguments don't apply to, then please let me know.

I believe that Einstein is closer to the truth of what you call the real world than was Aristotle. So when I'm told by this type of fallibilist that we don't know anymore today than we did 400 years ago, I demur.

Neither Popper nor I believe that "we don't know anymore today than we did 400 years ago".

Given your comments on LSD and the a-s dichotomy, after reading this I conclude that you are a fan of late Popper (LP) and I prefer early Popper (EP).

Yes.

You think EP is wrong, and I think LP is right, so I don't see the point of talking about EP.

(I disagree with your interpretation of EP, but that's just a historical issue with no bearing on which philosophy of knowledge ideas are correct. So I'm willing to concede the point for the purpose of discussion.)

Gellner argued that Popper is a positivist in the logical positivist rather than the Comtean positivist sense. His discussion proceeded from the contrasting of positivists and Hegelians and so he put (early) Popper in the positivist camp - Popper was certainly no Hegelian. Of course, Popper never tired of reminding us that he destroyed the positivism of the Vienna Circle and went to great pains to declare himself opposed to neo-positivism. For example, he says that he warmly embraces various metaphysical views which hard positivists would dismiss as meaningless. Moderate positivists, however, accept metaphysical views but deny them scientific status. Does not Popper do this too, even if some of these views may one day achieve scientific status?

Yes: (Late) Popper accepts metaphysical and philosophical views, but doesn't consider them part of science.

CR (late-CR) says non-science has to be addressed with non-observational criticisms, instead of what we do in science, which is a mix of observational and non-observational criticism.

If by fallibilism you mean searching for evidence to support or falsify a theory, I'm a fallibilist. If, however, you mean embracing Popper's view of 'conjectural knowledge' and the inability, even in principle, or arriving at the truth, then I'm not. I believe, against Popper, Kuhn and Feyerabend, that the history of science is cumulative.

No, fallibilism means that (A) there are no guarantees against error. People are capable of making mistakes and there's no way around that. There's no way to know for 100% sure that a proposition is true.

CR adds that (B) errors are common.

Many philosophers accept (A) as technically true on logical grounds they can't refute, but they don't like it, and they deny (B) and largely ignore fallibilism.

I bring this up because, like many definitions of knowing, yours was ambiguous about whether infallibility is a requirement of knowing. So I'm looking for a clear answer about your conception of knowing.

Philosopher Justin Kalef joined FI list, then quickly left. I emailed him about leaving:

I see that you decided to leave without offering a criticism of any point, sharing any of your positions, or attempting discussion once. Sad. I take it you disagreed with some things I said and changed your mind about your supposed interest in open discussion. Presumably you mentally categorized your disagreements as something other than disagreement (e.g. you judged me unserious, ignorant, arrogant, or whatever). And somehow that let you think there was no need to allow the possibility of counter-arguments or clarifications-of-misunderstandings regarding your your negative judgements.

Since you won't argue your case (including by linking anything prewritten), do you know anyone who will discuss and could represent your views? If not, aren't you shutting down rational enquiry regarding your views? If you care about the truth, how can you simply silently refuse to argue your side by any means (including link or proxy)?

If you don't know how to fit discussion into a busy schedule, there are methods which I can explain to you. That's a solvable problem, not a reason one has to give up. Or if there was some other problem, you could have stated it briefly in case there was a misunderstanding or solution.

He replied 5 weeks later:

Hi, Elliot. Apologies for taking so long in getting back to you: I'm in transit this summer and had many administrative issues to deal with, also.

The main reason I decided to leave the group was that I am very busy and need to be highly selective with how I spend my free time. I occasionally participate in online discussions, but usually I do this on widely read blogs where I think a point needs to be made to stop large-scale consensus from shifting uncritically in a certain direction, or where I expect that the other participants in the conversation are likely to offer arguments, information, and ideas that will be particularly useful to me in the pursuit of my (admittedly quite narrow) range of interests.

I appreciate what you're trying to do, but a) my inbox is already too full, and it was filling up with more and more posts from your website; and b) frankly, I think the questions your group was discussing were not in general being addressed very helpfully. For instance, there was (as I recall) a discussion on artificial intelligence. There is a wealth of important material to digest on that subject. If this were a conversation in which some genuine experts on robotics, logic, computer languages, etc. were taking part, I could see some optimism for helping to resolve it: but as it stood, it seemed that the crowd was trying to tackle some a posteriori issues from the armchair. That's just one example.

I don't agree that, by my leaving the group, I am shutting down rational enquiry regarding my views. I am far from the only person in the world who could explain the problems with many of Ayn Rand's ideas; and even if I were the only one, you could pursue these conversations on your own without me. I have in no way prohibited you from having your own conversations. Surely, what I do with my own time is my business, particularly when I have the sorts of constraints on my time that I do. Do you really think that, if a thousand groups wanted me to participate and talk about my ideas, I would be doing about my ideas, I would be morally obligated to participate in all of them?

That being said, I would be willing to address one of Rand's arguments that I disagree with. If you like, please send me a number of her conclusions: your choice. If any of them are both interesting and dubious to me, I will ask you to reproduce a summary of her argument for the conclusion (with short quotes if you like). I find that Randroids tend to refer excensively to her novels. That's okay, but a novel is not an argument. If you can extract from her novels a coherent argument for an interesting view, I will address that argument (again, if the conclusion is interesting and if I find the argument dubious).

I replied:

OK, Rand conclusions:

VoS ch. 4:

there are no conflicts of interests among rational men

RotP, The “Inexplicable Personal Alchemy”:

nobody builds sanctuaries for the best of the human species

VoS ch. 8:

One must never fail to pronounce moral judgment.

CUI ch. 14:

1. In any conflict between two men (or two groups) who hold the same basic principles, it is the more consistent one who wins.

2. In any collaboration between two men (or two groups) who hold different basic principles, it is the more evil or irrational one who wins.

3. When opposite basic principles are clearly and openly defined, it works to the advantage of the rational side; when they are not clearly defined, but are hidden or evaded, it works to the advantage of the irrational side.

AS:

money is the root of all good

(all italics omitted)

Regarding being busy and discussion: I think, if you wish to be a serious, rational, public intellectual with an opinion on Ayn Rand, who is very busy, then you ought to have a website (or equivalent) where you've posted your position on Ayn Rand (references are fine for unoriginal points). Otherwise your view of Ayn Rand is inaccessible which prevents both learning from it and criticizing it.

You say there are other critics. Sure. But you haven't chosen specific criticisms which you will take responsibility for. Just pointing in the general direction of a bunch of arguments – many of which contradict each other – isn't a meaningful take on the matter.

If you choose to focus on other matters, that's fine. I and others can pursue the matter without you. In that case, I'd say that you should not make claims about the matter you've chosen to stay out of.

It's interesting that you bring up the scenario of 1000 forum invitations. When we first met, I asked if you knew of a single good forum meeting certain criteria. You said that you didn't know of any, so I invited you to FI. There is a shortage of invitations to worthwhile forums, not a surplus! In this context, you shouldn't give up on a forum so fast. You should instead write a public criticism and see if there's any answer or misunderstanding – especially a criticism of a way in which FI fails to meet the criteria we'd discussed (which I understood you to think were good criteria).

Regarding AI: If lack of expertise caused any mistakes, you could point them out rather than comment on the credentials of the speakers. Especially when you don't know, and didn't ask, anyone's credentials. (A lot of the discussion participants do have expertise relating to computation).

It's also generally a difficult and debatable problem to judge which credentials matter – e.g. for AI discussion, is it more important to be good at programming or good at epistemology? And which credentials shall we defer to for that meta-debate? That's another reason to focus on the issues, not credentials.

Which Rand conclusions would you have used? Share them in the comments below!

a Charles H Bennett paper mentioned (search for "simple counting argument". it's on page 3) that there's a counting argument that most strings can't be compressed, or can only be compressed a little bit. Something like that (from memory). He didn't explain it.

So I thought about what it is.

Consider a bitstring of length 50. What does compressing it mean? It means you have some shorter bitstring which maps to it.

Suppose you were going to compress it to a length 40 bitstring. Could you compress each length 50 bitstring to a length 40 bitstring to save 10 bits of space? No because there are 1024 times more length 50 bitstrings than length 40 bitstrings. So each length 40 bitstring has to, on average, be mapped to 1024 length 50 bitstrings. So by removing 10 bits you're also losing 10 bits of information (you're specifying the target with 1/1024 ambiguity). Given a 40 bit string, you don't know WHICH 50 bit string it's shorthand for b/c there are 1024 it could be shorthand for, if you use a method to compress ALL the 50 bit strings uniformly (for example, by simply cutting off the last 10 bits. or you could cut off the first 10 bits or some other selection of 10 bits that don't even have to be in a row).

so at most we could compress 1/1024th of all 50-bit strings to 40-bit strings. that's a theoretical limit otherwise you have multiple 50-bit strings that you're trying to compress to the same 40-bit string.

you can take a 40-bit string, compress 2 50-bit strings to it, and then add an extra bit saying which one of the multiple mapped strings you mean. like dealing with hash table collisions with a linked list. but that's stupid in this case. it makes more sense to just compress to a 41-bit string. (maybe in some more complex approach it's useful? but it won't fundamentally change the conclusions.)

so what if we were to set up a compression algorithm which uses ALL smaller strings to compress to? we can compress to a 1 bit string, a 2 bit string, etc, all the way up to 49 bits.

well there's half as many 49 bit strings as 50 bit, and there's a quarter as many 48bit strings as 50bit, and an eighth as many 47 bit strings as 50bit, etc. so 1/2 + 1/4 + 1/8 ... = 1. so ALL the prior bitstrings have exactly the right number of possibilities. roughly. but not quite:

a 2-bit string has 4 possibilities. a 1-bit string has 2 possibilities. not enough. even if we count 1 more for a 0-bit string we're still 1 short.

so at least 1 (2 if we can't use a 0bit string) of the 50bit strings can't be compressed.

also, how many can be compressed a lot? well half of them are only getting compressed 1 bit. 1/4 are getting compressed 2 bits. so if we define "a lot" as more than N bits, then it's only 2^-N which get compressed a lot. like if you want 5+ bits of compression, then it's only 1 out of 32 bitstrings that can be compressed that much.

so significant compression is sparse/rare. this is important b/c the field defines random by the content (uncompressible) rather than by the source (how you generated the data). which is interesting b/c it's a "content not source" thing like Critical Rationalism cares about, and it's also not how common sense thinks of random data. common sense says if you flip a coin a bunch, whatever you get is random. but computation theory stuff says if you get 50 tails and that's it, that sequence isn't random. but MOST sequences of randomly generated coinflips will be random (mostly uncompressible) sequences because that's how most sequences are as i was just proving (e.g. at best only 1 in 32 sequences is compressible by 5+ bits).

there's another issue i think is interesting. what if you make 1024 different compression algorithms? each one could compress a different 1/1024th of all the 50bit strings to 40bit strings. however, the overhead for specifying which of the 1024 compression algorithms a particular 40-bit number was encoded with would be 10 bits b/c you need to specify a number from 1-1024 to say which compression algorithm it was. so you don't gain anything. this kinda of overhead is an important issue which gets in the way of attempts to cheat.

There are two reasons bitcoin is worth money. Neither is because it's so wonderfully libertarian.

1) Bitcoin facilitates crime. it's really helpful for paying for illegal goods and services, money laundering, etc

2) Bitcoin is a ponzi scheme. It's like amway where the people who get in first get money from the people who get in later.

PS both bitcoin and etherium are absolute and total shit on technical grounds. here's one example from a few days ago. i've read many, many more. issues with bitcoin include limits on transactions, transaction fees, and the ridiculous problem that when someone pays you, you both have to wait and see if the payment actually happens or not, and you can be pretty confident the payment will be permanent after like an hour (but you're still not actually 100% sure).

Richard Hammond's Jungle Quest is a documentary where he visits the Amazon rainforest and takes photographs. He doesn't really know what he's doing. There's other people there to make things work out OK, he doesn't have to be responsible or competent, he just has to be charismatic and popular. So it's pretty silly.

He says he wants to inspire people to save the rainforest and that something like 20% of it has been destroyed over the last 30 years. (I forget the exact numbers.) I recommend the South Park episode about the rainforest.

What the documentary doesn't say a word about is why the rainforest is being destroyed. What's the upside of destroying the rainforest? What's the downside of preserving it?

People destroy the rainforest on purpose. And it's not because it simply never occurred to them to save it...

People destroy the rainforest because they want land to live on and farm on. They want to be less god damn poor. They want their lives to suck less. They clear areas of forest to make space for humanity and to improve their lives. "Save the rainforest" means making human lives worse.

There is, of course, a way to save the rainforest without screwing anyone over. But it's not what environmentalists have in mind. The win/win solution goes like this: dramatically reduce building regulations in the U.S. (so there can be way more homes), make the U.S. dramatically more capitalist (so there can be way more jobs), dramatically reduce welfare in the U.S. (so immigrants won't be a drain on the current American taxpayers), and then, after doing all of those, let in immigrants who want a better life and are willing to assimilate. A lot of people would rather assimilate to be Americans than work hard to clear the rainforest to make their lives slightly less miserable. Low paying American jobs pay way more (are way more economically productive) than lots of the reasons people clear the rainforest.

There's no need for there to be a shortage of productive jobs in America. That's caused by destructive government actions.