Epistemology

7/2/2014

Mistakes are ubiquitous. Error is always with us. The whole purpose of creating new knowledge is to correct errors. We cannot have certainty - about anything, ever. But at the same time this does not mean that all claims to truth are on equal footing. Relativism - the idea that all claims to truth are equal (that everyone's opinion has equal value) is false. Science, mathematics, philosophy and morality uncover objective truth about reality. It's just that objective truth does not mean what most people think it means.

If something is objectively true, this does not mean it is certainly true. Pause for a moment. What does "certain" mean? If I'm "certain" - what does that mean? It just means I feel some way. I feel certain. Certainty is an emotion! It's a feeling you get (or at least a feeling some people seem to want to have). But a feeling - the feeling that you are definitely, without a doubt correct is no guarantee that you are. You should have doubts. Doubt is good. Someone without doubt is dogmatic. And dogmatism only ever leads to evil.

Do you remember the time when you were certain that through any two points a only a single line can be drawn? Perhaps you still are, depending on how much maths you remember. So try it now: draw two random points anywhere on a piece of paper and draw a straight line through those two points. Can you draw more than one line through those two points? Have you tried? Well think a bit. It's said only one line can be drawn through those two points. Indeed Euclid who first laid down our mathematical foundations of geometry provided a rock solid mathematical proof of this. And this is what you are taught in school geometry class.

So there's a proof. But are you certain of so simple a truth? Most people are certain at least for some time that only one such straight line can be drawn.

Now bend the piece of paper. Or wrap the paper around a ball. Suddenly infinitely more straight lines can be drawn. Your prior certainty is undermined by a simple change of perspective.

But maybe you have objections now. You still refuse to accept this. "That's cheating" or some such. "The line is bent". The point here is: do you doubt it? Should you doubt it? Did you misunderstand the instruction as I first posed it to "draw a straight line through those two points"? Could I have phrased it in such a way as not to be misunderstood by you?

Karl Popper once said "It is impossible to speak in such a way that you cannot be misunderstood".

So it is with all knowledge. We always have doubts. We can always misunderstand. We can never be certain.

So when someone says "I know X" where "X" is any claim at all, what they're not - in actuality saying is: "I am certain that X is the case" or even: "I am justified in believing that X". No, what they actually mean is "The best idea I have at the moment is X and I've got no good criticism of X". When people say they know X they're not saying they can't (actually!) change their mind. Yes, of course: some people are dogmatists and claim not to doubt "X" when "X" is either something really obvious or something that is really important to them. But their thinking these things does not change the true structure of knowledge - or that they are, in fact, actually fallible humans. People are always fallible. We should know that even if they claim to have no doubts: we should doubt that very claim. Or better yet: understand that if they thought better about things then they would, like us, understand that although knowledge is genuinely possible the feeling that one is "certain" is no guarantee at all that one is in possession of the final, ultimate truth. That final, ultimate truth is something we aim for - but it's not something we can ever know to have obtained.

The "Justified True Belief" mistake handed down to us from Plato remains deeply ingrained in our language and ways of thinking about knowledge. But it's just that - a mistake. Knowledge isn't anything to do with being "justified true" (which is to say: in some way finally shown to be not possibly false) or even about the beliefs of people. Knowledge instead is very much a real thing: abstract and yet instantiated in physical objects. In other words: physical structures are arranged in particular ways in order to "code" the knowledge. So, for example brains encode knowledge, but so too do computers and even telescopes. All those structures encode causal relationships between objects so that something useful can be done with information.

Popper understood that knowledge is real and possible. But many people think he rejected knowledge because they think his entire philosophy is about falsificationism. So they think that Popper's philosophy says something like: you can be sure about those things you have falsified but not about anything else. Or something like that. But this is wide of the mark completely. Popper did not think you could be sure (as in certain, or absolutely 100% without any doubts for all time satisfied) that anything was true. Or false. No. He didnt believe in "confidence levels" at all. Everything we know is always conjectural - which is to say "tentative". It contains truth, but it's not the final truth. Knowledge claims are explanations about what is true. Falsification - or refutation broadly - rules out some explanations (theories, ideas, hypotheses) as being good explanations. David Deutsch sharpened up this idea even more: an experimental refutation makes a theory almost impossible to vary in such a way as to account for the refutation. In other words: if your favourite beautiful theory is slain by an ugly fact, then if the theory is any good, you shouldn't be able to just change it a little bit in order to now explain what went wrong. For example if you think eating a kilogram of grass cures the common cold, and someone tries and it doesn't work (so your theory is falsified) then if you come back with "oh - actually it's probably more like 1.1kg - try again" - this is called an "ad-hoc" modification - making your theory "easy to vary" and so we can reject the "grass cures the cold" theory and all infinite variants of it on that basis. We don't have to keep testing it. And it's fair for us to say "We know that grass doesn't cure the common cold". Does this mean we're sure it doesn't? That we're certain one day someone won't come along and show how if you isolate some chemical in grass and add it to some other chemical in alcohol and take it as a pill that this won't kill cold viruses? No. We can never be sure. But we've no reason whatever to think this is, or ever will be, the case.

Popper was never a skeptic in the philosophical sense of the word. Quite the opposite! Skeptics believe in the "Justified true belief" theory of knowledge. They rightly conclude one can never be absolutely certain about knowledge claims but then wrongly conclude this means knowledge is not possible. Popper was a critic of this idea - he rejected it completely. He threw it away root and branch and started from the ground up. The idea here is to divorce knowledge from being about "certainties" or "justified truth". Instead knowledge is possible because there is an objective difference between knowledge claims. None of them are final absolutes - but in almost all cases, always, where there are two (or more) competing claims, one of them is better than the others. The process of criticism rules out all but one of the ideas. The one that survives: we say "I know that". That claim comes to be called a part of our "knowledge" of the world. But it's always revisable, always improbable. Popper was no skeptic about knowledge - he was a critic. And these are quite different things. One assumes knowledge is not possible because certainty isn't. The other says almost exactly the opposite: the pursuit of knowledge and its objective growth and improvement is only possible because certainty isn't. (To unpack this: if we were able to be certain then this would mean we'd have direct access to final truth and this would mean the quest for knowledge would stop. We'd just "tap-in" to that ultimate source of truth and that would be that. Game over. We'd never improve anything because we'd have the final, complete answer (and all the answers). So the pursuit of knowledge would stop.) But happily, the pursuit of knowledge isn't like this at all. Knowledge is hard won. And it comes to us with much misconception. Almost all we know contains misconception - and we can never be sure which is misconception and which is true. But that doesn't matter. As long as our process of error correction keeps on sifting the misconceptions from the kernels of truth, so that we make progress - this is all we need to be able to say "we are creating knowledge" or "we are learning" and "making objective progress". We can't know we've got the final truth, ever: but what we can know is that our process of criticism: experimental falsification and other kinds of refutation - makes the ideas that survive this process objectively better than the ideas that do not.

The purpose of science is to correct our knowledge about physical reality. Here we see improvements most obviously perhaps - objective progress seems to be made all the time. New medical discoveries improve our health and lengthen our lives and cure disease previously a death sentence. Each new iPhone is better than the one which went before and each new car engine is more efficient than previous models. One great theory is overturned by another even better theory. New explanations replace old ones. The news ones are not certainly true - but just closer to the "ontological" truth than its rival(s). Karl Popper called this verisimilitude. This means "closeness to truth" - where by "truth" we mean the ontological truth. What does that mean?

Ontological truth is a description of what is really there. We do not have a method for finding a final, once and for all description of what is really there but this is a wonderful thing. It means that the quest for knowledge is infinite. We make progress: forever. We don't find a final fact about reality and just stop and say "well that's it. No more needs to be known about this. We now know everything there is to know about this." That day, happily, will never come - we can always improve things.

This means we must try to show which of our ideas are wrong and how. And so this is the best thing you can do to help progress along. You don't need to be a great philosopher, a quick thinking mathematician or a creative scientist - you just need to be a critic. Critics have a bad reputation. And they don't deserve it. There are two, equally important ways to be in the world when it comes to progress; you can be a creator and you can be a critic. We are all both of these at various times and both are absolutely essential to creating knowledge.

Creation can result in improvements - and that's great. Producing something new can be exhilarating. But that again is just a feeling. Just because you're excited about a new creation doesn't mean you've actually produced something of value. You can be wrong, remember? You might very well be right about your new creation - you'll earn fame or money or social cache - your new idea will spread: some new art more beautiful or incisive might make you rich or famous. Some new scientific theory to overturn an old theory could solve a problem previously a stumbling block to progress might improve the world. But sometimes new creations lead to dead ends. That's just part of the creative process - creating stuff that just doesn't do what you want it to: art that's just not that good - music no one will listen to, a painting everyone says is derivative. A scientific theory quickly slain by an experiment that shows it to be wrong. A computer game no one wants to play. Criticism is how we sift the good, the beautiful, the true, the useful - from the bad, the ugly, the false, the damaging.

And this is true for ideas generally and your own ideas personally. Criticise your own ideas and don't be upset when others criticise your ideas. They are not criticising you personally. You are not your ideas. You merely have ideas. You can discard ideas. And you can criticise criticisms. You can defend your ideas. That's good too. If you want to get better, as fast as possible - create, criticise and repeat. It's the way to make progress.

7 Comments

Erik Johansson

1/23/2017 09:59:44 am

Great post! I really like this passage: "There are two, equally important ways to be in the world when it comes to progress; you can be a creator and you can be a critic." Since I come from the world of software development I translate this to mean that developers and testers are both equally important. I wonder if the software we are creating (and criticising) is part of the first or third world according to Poppers model. What do you think?

Brett link

10/30/2017 08:38:03 pm

Because the information that is the software can be written on paper or coded in a computer (in the form of magnetic domains or electric potentials, etc, etc) - in fancy philosophy speak because it can be "instantiated" in lots of different "physical substrates" (things made of matter) but is itself not identical to matter, it's not world 1 as such. So the software itself - the pattern - is "World 3". It's independent of its particular World 1 instantiation. World 1 is necessary for the existence of, but not identical to world 3.
As an aside, I don't know that Popper fully understood the universality of computation and therefore I'd argue world 2 and world 3 have perhaps a more complex, harder to differentiate boundary than perhaps Popper might have been thinking. I don't know. But mental events must arise out of the software that is our minds. So something about World 3 can give rise to World 2 in kind of the same way World 1 gives rise to World 3. There's probably something foundational lurking there :)

Bruce Nielson

4/10/2018 01:51:23 pm

I just discovered your website today and started to read it. It's clear you've been heavily influenced by David Deutsch like I have. One of these days, if you need someone to talk to about these subjects, I'd love to chat with you more. I must say, you might be able to help me understand your views on 'morality' better.

I have spent years researching what was in Duetsch's two books and what I basically found so far was:

1. He's right, many worlds is the only explanation we currently have for quantum physics (a point I only with great horror have to admit to and really wish wasn't the case.)
2. Popper's epistemology (at least with Deutsch's further clarifications) are a spot on description of scientific rationality and the rest of philosophy's epistemological really are just wrong.
3. Biological natural selection really is incredibly important as an explanation
4. Computational theory is, in fact, a branch of physics and explains what can or cannot be computed under the laws of physics.

In other words, I'm apparently now 100% on on board with the 4 strands.

But I've failed to pick up Deutsch's views of morality or his over all optimism. I have found criticisms his arguments on these subjects that (unlike the 4 strands) seem fairly strong to me and suggest his arguments were perhaps flawed.

This puts me in an unfortunate circumstance. What I most like about Deutsch is his optimism and how it stems from his rationality. But this is the very part of his arguments that don't work for me yet. In short, I can't find any criticisms with his 4 strands, but it's not at all clear to me at this time (despite his best arguments in his books) how they equate to objective morality and an optimistic view of reality. In fact, on the contrary, so far it seems to me that his 4 strands are rather un-optimistic long term (i.e. heat death) and suggest entirely parochial morality. (Perhaps 'objective' yes, but parochial, i.e. species specific.)

Would be interesting to talk to you some time if you're interested.

Brett

4/10/2018 04:18:41 pm

We have a community - of sorts. I first began interacting with that community - which included David Deutsch himself - via an email list during the late 90s. We've talked through these things for 2 decades now, since the publication of The Fabric of Reality - and continue to - to this day. Twitter is only a small part. But yes, I'm happy to talk to new people all the time. I'll note what you say in point 2 right there as it comes to bear on your next comment. Morality is about solving moral problems. What part of "morality" do you object to? I say something brief about morality here: http://www.bretthall.org/morality We could discuss that, if you like.

4/10/2018 03:37:18 pm

I saw your discussion online with someone about what does 'to know' mean and does it imply certainty?

You argued that it never implied that (it was impossible) and he argued that for better or worse, the word is used at large to mean that, and therefore Popper's ideas were really about beliefs and there was no such thing as knowledge.

I want to suggest another way to understand this.

Yes, we can never be certain of any particular conjecture. But there are in fact all sorts of things we can be 'certain' of and therefore I don't think we need to divorce knowledge from 'certainty' if we use a bit of nuance to explain it.

What I have in mind here is the idea that while I can never be certain that, say, Einstein's theory is true, maybe I can with some some nuance claim I'm certain it's not true. (i.e. it will be revised.)

Likewise, I can, with some nuance, claim I'm certain that there is no better alternative theory available to me today. So I think perhaps knowledge and certainty aren't divorced. It's more a matter of what we can feel certain about vs. what we can't. i.e. we can never feel certain of specific theories, but we can feel certain that there is no better theory currently available to us, and thus there is a kind of 'certainty' available to us about the theory after all. Popper's views only require that we know (are certain) we know of no better theory at the moment. If we can know that, then progress is a possible and knowledge is real. If we can't know that, then the skeptical view is correct.

I keep mentioning 'nuance' because I would anticipate someone who is skeptical saying "oh yeah! Well what if there is some guy in his basement with a better theory and no one but he knows it!"

But that is really an objection that doesn't get to the truth bottom of the certainty we're talking about. A guy in his basement who hasn't published his theories, for all intents and purposes, isn't available to me any more so than if it didn't exist any where on earth.

I believe we could play this game with any boundary a skeptic cares to draw (i.e. what if he has published, but it was so revolutionary that everyone ignored it and it was forgotten entirely. Again, this is really for all intents and purposes not available to me, so it doesn't affect the concept of certainty I'm suggesting. Repeat.)

4/10/2018 04:25:24 pm

Remember point 2, above. Well that rejection of certainty is simply a part of that epistemology, which you said you agreed with. Certain, to me - and I think I can speak for most fallibilists - generally is interpreted as "cannot possibly doubt" or "in principle is irrefutable" or something like that. Moreover it's a feeling - the emotion that such a state has been obtained. I reject all of that. Now why did you pick Einstein's theory as the thing that you are certain will be overturned? Is it because you considered scientific theories and thought: I know Einstein overturned Newton and I understand how this process works and so I conclude that, therefore, General Relativity is going to be overturned as well? In other words: a rather long argument got you to that conclusion. Could you be mistaken about it? Could you be mistaken about fallibilism? Yes! It could be the case (I don't expect it to be, I'm just saying, logically, it's possible) Einstein's general theory is indeed exactly the final word. After all, I'm fallible so when I say "It's not the final word" - that could be wrong. So I'm not certain that Einstein's theory will be overturned. I simply expect it won't be because I've an explanation - as you do - of why it should be overturned. But all these things are open to the possibility of being false. Because I'm a fallible human. And I don't know if, indeed David Deutsch is wrong about progress: it is logically possible there is a limit and Einstein reached the final theory of gravity. I expect that not to be the case and have an argument against that position. But whatever the case I am not certain in either direction. I never am, about anything. Your phrase "some nuance" seems to me to add nothing but an attempt at qualification of arguing "Well I don't mean certain like *that* - I mean something rather less, well, certain". In which case, certain just means "know fallibly" and there was nothing to disagree with to begin with. :)

4/10/2018 05:28:23 pm

"Now why did you pick Einstein's theory as the thing that you are certain will be overturned?"

Wondered if you'd pick on this. Regretted using it because it distracted from my real point.

Yes, I actually meant, "I'm certain it will be over turned given certain explanations that I know of no better alternative to." (which, note, is in fact a certainly true statement -- see below.) We'll ignore that example for now.

In many ways, what I'm suggesting is that "given certain explanations that I know of no better alternative to" might be a universal not worth stating in most cases and ought to be assumed as part of every statement. At least between us falliblists.

What I'm really suggesting is that maybe we can be certain of ignorance, and maybe Popperism can be thought of as stating that certainty of ignorance is itself a kind of knowledge that allows meaningful progress.

Think of it this way: I can be certain there is no better alternative, that I currently know of, to General Relativity?

Is that a fallible statement? Can I be certain of it? 100% certain? Even if I used to know of a better alternative and then forgot it, it's still, at the moment, a true statement with -- it seems to me -- 100% certainty. The reason why is because I made it a statement about my own personal ignorance at a specific point in time.

I think one could argue that it's not a particularly powerful statement when I say it -- after all, I'm not a physicist and therefore even a *certain* statement about my lack of knowledge means very little. And certainly that is the case. (No pun intended)

But imagine David Deutsch making the same statement. It's still really just a statement about the certainty of his own personal ignorance at a specific point in time. But in that case, it means a heck of a lot more because he's made a much more expansive attempt to find an alternative including education from the best currently known in the world, etc.

Both statements are 'certain' because they are really both just statements of our personal ignorance. We don't happen to know of a better theory. But in the case of David Deutsch it's a much more powerful statement because he's made a much stronger attempt to criticize his own ignorance than I have. In fact, he's made such a powerful attempt to do so that we can actually feel fairly comfortable with the idea that likely no one in the world currently knows of a better theory (though that is NOT certain, as per my previous statements, but also doesn't matter.)

Could this be what Popperism really is? That's what I'm asking. If this is a true understanding of Popperism then Popperism is NOT about *nothing* being certain. Popperism is really (under this explanation) about how personal ignorance isn't equivalent to never being able to know (even in the sense of being certain) of nothing at all. Popperism, in this light, is about how not all ignorances are equivalent or even close to equivalent. Some contain a great deal more knowledge that others.

I'm just playing with this idea. So I'll not be offended at all if you continue to disagree. But really think about the example I'm giving here and try to find our strongest criticisms of it.

Is there some sense in which my statement is still uncertain, even though it's only about my personal ignorance at a specific moment in time? I can't, at the moment see how. It really does strike me as a form of certain knowledge and has a direct relationship to Popperism -- i.e. the degree with which we've attempted to find an alternative makes the statement of certainty about what we don't know more meaningful.

Also, you are defining falliblism as including "rejection of certainty is simply a part of that epistemology" but there is no particular reason why can't decide to refine the theory to mean something more true if there is something to what I'm saying. So what it's currently defined as sort of doesn't matter. in my mind. Sure, you're right, it currently means that. So what? What matters is the argument itself. Give me your best criticisms of it. Is it really true that my statement of my own personal ignorance at a moment in time is fallible or uncertain in some way?

knowledge and its creation

Epistemology

Leave a Reply.

Author

Archives

Categories