Deutsch: reading list

One of the more interesting of the scientists working today is David Deutsch. David Deutsch takes philosophy seriously. This is a refreshing change from the mood of anti-philosophy prevalent in many areas of science. Though I am in full agreement with many of the criticisms of how contemporary philosophy is undertaken, it is not the case that philosophy is unimportant or bad. philosophy is a very important subject and for one main reason -- bad world-views, which are invariably suffused with bad philosophy, can have disastrous consequences. This is the same apology that Popper gives for the practice of philosophy. Philosophical investigation and discussion of our expectations and world-view can help us eliminate those ones that are false and retain those that are true and so remove bad view that will lead to those outcomes. We should let theories die in our stead.

Rant over. Now to the goal of this post. I recently stumbled on a blog called Apple Butter Dreams, the author of which gives an extensive reading list on Deutsch (and incidentally, Karl Popper) with some interesting commentary on the chapters he chooses, he also gives suggestions for reading around the subjects.

This is the author, Joe Kern's, appraisal of DD's work.

Professional physicist and semi-pro philosopher David Deutsch’s works—most of The Fabric of Reality (1997) and a few parts of The Beginning of Infinity (2011)—have completely transformed my view of life and the universe, at a comparable level to probably only two other works I’ve encountered, Richard Dawkins’ The Selfish Gene (1976) and part 3 of Derek Parfit’s Reasons and Persons (1984). (Go Oxford!) I feel that anyone who regularly wrestles with the big ideas of science and/or philosophy should have Deutsch on their plate.

He also has many other topics that should be of interest, ranging from economics to relief work.

Quote of the day (and commentary)

Critics of critical rationalism should at least try to state clearly what it is that they are attacking. Is what is under fire the method that we propose, or is it our frank refusal to garnish our product with spurious and deceitful claims of reliability? The two are not the same, and to dress up a criticism of the second kind as a criticism of the first is simply mischievous.

Miller, D. Critical Rationalism: A Restatement and Defence


A criticism then of the type that "your method is not reliable", is not actually a criticism of the method, because critical rationalists do not claim reliability for their method. "But", you might interject "why should we adopt the method?" The only answer to that is: "you do not have to accept it; do what you want." Then we would simply ask: "have you got a criticism of the method?" Or "Have you got an alternative method that does not immediately contravene its own demands?"

If you have either of these two things we are then in a position to learn something new and to enter it into critical discussion.

This post relates to an earlier post (here). I made a defence of critical rationalism that Miller thinks is true but smacks of casuistry. I did not know he had already considered this rebuttal and put it aside. It relates to the fact that critical rationalism, although it begs the question, it does not beg the same question at each step.

Wilson On Popper On Science

I like it when I find one of my favourite writers talk about another of my favourite writers, especially when they come from what seem like totally different world views (or reality tunnels, as Wilson would say.) In this instance the writer is Wilson and the one written about is Popper. Since, I guess, I know a lot more about Popper's philosophy than Wilson does, I wish to elaborate on and also correct the passage. In the following passage,  It might seem, by association, that Popper holds the views expressed in the latter half of this paragraph, not just the first sentence that he is mentioned in. Many people might assume that the rest of the passage is actually Wlsons's representation of Popper's views: I am going to be a little unfair to Wilson by  assuming that this is exactly the case. I will assume that all of this passage is a representation of Popper's views and criticize Wilson accordingly. This is so I can elucidate Popper's philosophy, and also quote from one of my favourite authors. The Passage is as follows:

Sir Karl Popper has argued, plausibly I think, that no proposition can ever reach the level of absolutely proven true for all time (1/1) because that would require an infinite number of experiments and we haven't done that many experiments yet, nor does it seem likely that we can do them in any foresee able future . Any of our theories, however, can reach the level of proven false (0) very quickly, since any failed experiment raises doubt, and a long rigorous series of failed experiments must either indicate that (a) the theory has no relation to experimental/experiential data, or else (b) some god or demon has rigged the results just to mislead us.  The latter choice does not rank as a meaningful proposition in science, although it might keep theologians (or some academic multi-culturalists) busy with debate for centuries. 


That being proven true is an aim of science is misleading, since most people who believe in (scientific) proof (i.e induction), do not think that science is looking for truth, but likelihood.  So the 1/1 is not to do with truth, but with certainty. Truth and certainty are often confounded, and many writers, even very smart ones (as Wilson surely is), mix up these very different notions. Compare a similar statement made by Rafe Champion:

It is a commonplace observation that according to the version of the probability calculus that Carnap spent some years of his life to perfect, the probability of a universal Law such as "all ravens are black" is zero.

This does not mean the statement is false.

Certainty has to do with our belief; truth has to do with correspondence of our statements to the world. What Popper was arguing was that we cannot ever confirm our hypotheses to any amount, especially not certainty. With this correction in the passage, the first sentence is now an accurate representation of Popper. Also Popper argued a logical point, it is irrelevant whether it is plausible.

Now, as well as being totally unfair to Wilson, I am going to be, in this next criticism, completely pedantic.  Popper did not argue that something can be proven false. Popper argued on  two subtly different levels which Wilson is not taking into account. There is the level of logic and the level of methodology. On the Logical level there is an asymmetry between verification and falsification. Logically (i.e, in principle) we can, through Modus Tollens, refute a theory (it only takes one piece of counter-evidence for this), but we cannot verify a theory¹, because we need an infinite amount of evidence to verify a theory. Russell argued the same thing, the claim that we have verified a theory, is itself a universal statement that cannot be verified. On the methodological level it is not so straightforward, because a counter-instance itself is tentative (this is why theories cannot be proven false), what goes into a refutation includes not just counter-evidence, but also an explanation of why the theory fails in the light of this counter-evidence.² This is because we cannot just abandon a theory, even if it is false, because we need something work with, and if we only have this theory, we can still use it, because even false theories entail true consequences (and we have found out in which domain this theory gives false consequences).

We will now turn to '(a)' and '(b)' in the text, keeping in mind that I am being totally unfair to Wilson, in order to elucidate what Popper argued for. (a) states that consistent experimental failure means that the hypothesis has no relation to  experimental data, but this is simply false. It does have a relation, two in fact: The hypothesis has a logical relation and subsequently an informative relation. The counter-evidence of the theory counts as a refutation (which is the logical modus tollens) and all counter-evidence is the refuted theories information content , since counter-evidence is not something that theory predicts and so in repsect to the theory we are testing it is informative i.e it is not contained in the class of its logical consequences. Wilson also mentions a similar thing about what information is, and strikes gold in his book promethes rising, but misses its actual value, because he is a subjectivist about information, (for the discussion about the subjectivist view of information, see  Quantum Theory and the Schism in Physics, by Popper).

'(b)' states that the Cartesian Demon is 'meaningless' in science. This is a very old and stable misconception of the demarcation between science and metaphysics (even those who purport to understand it) The demarcation has to lie inside meaning, if not then we end up with a principle which is meaningless, simply because it cannot be scientific. So therefore the demarcation has to be between meaningful scientific and meaningful metaphysical statements. This demarcation comes in the form of the falsificationist principle outlined above: something is scientific in so far as it can be, in principle, refuted. The Cartesian demon, therefore, is not meaningless, it is untestable (i.e unfalsifiable); it is perfectly meaningful, even scientifically, It is just not scientific.

Another further thing in relation to cartesian doubt: Popper argued  that doubt does not count a criticism of a theory or a statement. A theory can be true, even if we cannot prove it and even if we doubt it - doubt is always possible, and so empty as a criticism of anything. 

Popper argued that we should aim for the truth, even if it is unattainable, because the search for truth is what drives us to eliminate error. If there is no truth, there is no error and conversely. I have never come across a criticism of Popper that takes this it account, or a defense of the search for likelyhood that takes this sufficiently into account. The abandonment of truth, also means that we should abandon the search for error. This mistaken view can be read into the slogan often quoted by wilson:


All statements are true in some sense, false in some sense, meaningless in some sense, true and false in some sense, true and meaningless in some sense, false and meaningless in some sense, and true and false and meaningless in some sense.

This statement itself is just false.

I will be doing a longer critique of Wilson's epistemological views at some point, showing its strengths and weaknesses. Where he draws on Korzybski and in turn on Wittgenstein.

Footnotes

1. This is often referred to as the 'Popper Asymmetry' between falsification and verification

 

2.  it is not saved even if we weaken the demand to confirmation, for in order for evidence to confirm, the evidence itself has to be assumed to be certain)

3. This explanation might come from a further theory (as long as it is testable, consistent,  explains all the data the first theory did and has predictions which the first theory did not have), if we do not have this further theory, then the older theory, even in light of counter-evidence (although rendered problematic and labeled false) is kept.

General Comments on John Gray's view on progress.

These are left here for general commenting (at some point they will be made into a proper blogpost)

The distinction he is trying to make ulimately leads to the idea that something, if it can be lost, was never improved. The problem of progress is not one of seeing where preogress is and is not. It is one about changing our society in ways that can preserve the things that we improve and allows to correct the errors that we make, the fact that this is tentative and not guarenteed, is no argument against the fact that it is the best method we have, or that progress happens. It is also not to the point to suggest that just because new errors are made in every advance, that no progress was made. Problems are always going to be with the human race; it is utopian to think that every problem that we solve, should lead to no further problems. Gray is just a crypto-utopian attacking piece-meal engieering for presumptions that only utopianists (which gray shares) hold; utopianists believe that as soon as you solve all the problems of current human life, that there will never be any more problems and therefore if there are still problems, there has been no progress or if the progress can be lost, there has also been no progress. This is grays view also. It is mistaken and deeply problematic viewpoint, it denies human ability to correct his own errors, and if this is true, then we cannot even try, because it is futile.


  The first enlightenement happened in Greece. Much of the knowledge there was lost for years and not just their ethical or moral views, when we rediscovered this knowledge we had an advance, in logic, in science. It is not set in stone that we will keep these things.

 Another confusion he makes seems to be his discussion on post-modernism, as though relativism was relevant to the idea that some knowledge can be lost and others cannot be. He seems to be saying that there are no moral ideas that solve problems better than other moral ideas, and that therefore there is no objective way to evaluate them. This is mistaken. Some moral views are just better and solving moral problems than others. Just because it is hard to preserve these moral ideas does not mean that there is no progress. Progress itself can be lost. But progress can also happen, and it has happened. I can speak of two progresses in moral thinking (there are probably more). Socrates argument against euthyphro about goodness and Popper's political innovation which turned political thinking on its head (especially the contractarian tradition) from "who should rule" to "how do we get rid, without violence, of those in power who misrule " these two innovations, have striking consequences for moral and political theorising, and they are better ideas objectively. Just because they have been largely ignored, or confined to academic squabbling, does not mean that these ideas aren't an improvement on earlier ideas, and even current ones.

Society is difficult to make more civilized, but it will always be that way. Does not mean there is not progress.

  There are resurgences of atrological thinking. The fact that people still think that truth is manfiest, or that data speaks for itself, was the same error that astrologers made. Francis Bacon's own methods of science would make astrology a science (even though he himself did not like astrology) this also goes for logical positivism and any variation thereof. SO there was resurgence in the fundemental misconceptions, and the only reason these people could fight against astrology was not for any rational reason, but simply because there was a concensus on it being bogus, and so they could get away with infinging their own standards of rationality, because who would argue for astrology right?

A Clarification of 'A Rejoinder to Phillip Gregory.'

I blogged what I thought was a very clear essay yesterday.  But Philip Gregory offered  a rebuttal that, although did not show weaknesses in my intended position, has shown weaknesses in my formulation of that position. To give some background, the following is the paragraph to which Philip Gregory is responding:

"To get back to the problem  at hand: The issue of justification is not solved, because evolution, although it gives us a universal set of expectations, it does not give us a universally valid set of expectations (or innate cognitive abilities). So even if something fits in with our expectations and theories, ie it is coherent with the system, that does not mean they are true, or become more or less true, because they are more or less coherent with that system. If there is something wrong with one of our evolutionary expectations and our background theories, and our theory is coherent with it, then the error contained in the expectation or theory carries over to the hypothesis, and is therefore not justified, we can never know fully whether there is not an error in our expectations or theories. This is just to say that our expectations, and evolution itself, are fallible, i.e not justificatory. There is fallibitilty at the core of human knowledge and experience."

And this is Philip Gregory's rebuttal:

"As a quick rebuttal, I will say that you are missing the point because you have not accepted my revision of epistemology, and your rebuttal amounts to: I don't like coherentism because it can result in some falsities whilst having a coherent web of beliefs. To that I say: oh well, such is life. Whether you like the fact that coherentism can result in false beliefs or not is not a sufficient rebuttal "

I think a relatively clear rebuttal, but it does not quite hit the target.

But first, to clear up any confusion about what I do and don't have to accept, If he claims that I have to accept his epistemology to "get the point", he is merely saying that unless I accept his epistemology as true, he will not believe that I understand it. This is an illegitimate burden to place on somebody and a burden I will not own.

Secondly I did not say I disliked (I never made any claims about my likes or dislikes) coherentism because it leads to false conclusions. I said that human knowledge is fallible at the core and therefore no justificationist stance could work, and in this particular case, coherentism. I further delineated (or tried to) that coherentism fails in its own goals. That goal is (this time I am going to try to make it more clear) is to confer reliability to our theories, statements hypotheses etc. Coherentism says that something increases in reliability if it is logical and it fits in with our already conjectured knowledge and expectations (This, in psychology, is known as confirmation bias). We can attack this view in the same way we can attack other justificationist views by turning it upon itself. in this way we can ask how reliable or justified, or even coherent, coherentism is? To make it clear what I mean, consider the following: If the background knowledge is false, but we believe it to be true, and we confer an increase in reliability to a hypotheses because it fits in with our current theories expectations or whatever and it is (logically) valid, has the reliability of the hypothesis actually increased? if not then coherentism is false and has not done the job that it is intended to do. Therefore we can abandon coherentism, and because all justificatory epsitemologies are false (they founder on the horns of agrippa's/munchausen's trillemma), we can adopt robust fallibalism (not just the weakening to partial justifcatory methods also known perversly as fallibalism) as the only currently rational option (this does not mean I consider it a certain method or even a partially reliable method, I am merely claiming it is the only one that is not logically questionable).  So I am not claiming that I dislike Coherentism because it leads to falsities; I am objecting to it because it does not do the thing it claims to do, namely increase the reliability of our hypotheses, theories or statements.