Saturday, December 16, 2006

Journal of Ayn Rand Studies - Fall 2006

The Fall edition of Chris Sciabarra's "Journal of Ayn Rand Studies" has finally been published. It contains, among other interesting articles, my reply to Seddon's critique of ARCHN, wherein I, in effect, restate the case against Rand. Using evidence compiled by cognitive scientists, neuroscientists, and evolutionary psychologists, I challenge not merely Rand's view of man, but also her epistemology, particularly her overestimation of the role of logic in efficacious thinking. Seddon responds in turn, offering a surprisingly feeble rebutal.

In the next few weeks I will offer brief commentary on some of these articles, particularly as the relate to important issues of Randian criticism.

39 comments:

  1. I like Fred Seddon, but after reading his response and his original review, I'm still not sure if he believes Greg Nyquist is correct about Rand's view of human nature. For example, does Seddon think that Rand's theory of concept formation has empirical support?

    Does he believe that Rand's statements in The Objectivist Ethics about emotions, parasitism is wrong because parasites will not survive, don't tell lies because you will get caught, etc. are correct?

    ReplyDelete
  2. And I see that Seddon says that Greg ignored the Thomas/Kelley draft of the book The Logical Structure of Objectivism. On the TOS site, the last draft of the book is 1999. Why should Greg be blamed for failing to engage a draft of a book that even the authors have lost interest in?

    ReplyDelete
  3. ***Why should Greg be blamed for failing to engage a draft of a book that even the authors have lost interest in?***

    First off your reasoning is nothing but a red herring because 1.) The authors have not abandoned the philosophic material in the book and 2.) you are wrong that they have abandoned the book because I specifically asked when the hard copy would be released and William Thomas said it was still a few years out.

    ReplyDelete
  4. Neil wrote: "And I see that Seddon says that Greg ignored the Thomas/Kelley draft of the book 'The Logical Structure of Objectivism.'"

    As with some of Seddon's other points, this falls under the category "true but irrelevant." It's true I don't mention the Kelley draft. But I don't mention any of Kelley's books. In fact, Kelley's name doesn't appear any where in my book. There's a good reason for this: my book is not directed at Kelley. It's directed solely at Rand and her orthodox followers. Kelley's version of Objectivism is one in which the philosophy's excesses and errors are either watered down or revised for the better. Kelley appears to understand the need for empirical responsibility and at least has some idea of what is required in order to be intellectual honest. I don't find that either Rand herself or her followers at ARI have a clue about either intellectual honesty or empirical responsibility.

    ReplyDelete
  5. Gregg Nyquist just wrote:

    "Kelley's version of Objectivism is one in which the philosophy's excesses and errors are either watered down or revised for the better."

    Despite Nyquist's harsh anti-Rand rhetoric, I knew there was some reason I did not dismiss him as a mere hater beyond some of his perceptive insights.

    Hopefully I will have time to go into some of his criticism, but here is a first. It concerns induction.

    Rand came to many of her conclusions using simple induction, which merely means in this sense looking at the world (including her own thought processes) without specialized instruments or knowledge and drawing conclusions based on differences and similarities observed. (This is similar to how Rand separates philosophy from science.) If I remember the beginning of ARCHN correctly, Nyquist denies that induction even exists.

    This would lead him to see dishonesty where a difference in method is used. I do admit that this practice was used by Rand a lot in judging others, so Nyquist's negative judgments of Rand often come off to me more as payback than anything else.

    Interestingly, in the DIM course (which I listened to), if I remember correctly, Peikoff stated that the only valid knowledge is inductive.

    Such arbitrary limitations, denying the existence of induction, or proclaiming it as the only true way, are very strange to me. I am sincerely curious about how highly intelligent people arrive at these formulations.

    Michael

    ReplyDelete
  6. M Kelly:
    >Such arbitrary limitations, denying the existence of induction, or proclaiming it as the only true way, are very strange to me. I am sincerely curious about how highly intelligent people arrive at these formulations.


    Hi Michael,

    Nice to see you at our site, and Merry Christmas. Now, as to this old warhorse...;-) If by "induction" you mean simply "observing" something, then there are few philosophers who have an issue with this.

    "Induction" however, in the philosophic sense, usually means one of two related things:

    1. Generalizing about the properties of a class of objects based on some number of observations of particular instances of that class of objects (for example, "All swans we have seen are white, and therefore all swans are white"

    2.Presupposing that a sequence of events in the future will occur as it always has in the past

    (These are from the Wiki entry on the problem of induction http://en.wikipedia.org/wiki/The_problem_of_induction)

    It was long believed that both the generalisation and the presupposition above were the path to "justified true belief" ie: that philsophic MacGuffin known as "absolute truth."

    Unfortunately David Hume criticised both these beliefs from a logical point of view, and as a result both versions of induction, it turns out, are irrational (ie not logical). Thus it is reasonable to say induction in the above senses "does not exist", just as other logical fallacies such as contradictions don't exist. Hume's criticism has never been successfully countered. Rand herself admits she has not solved Hume's problem - nor has even really thought seriously about it - in a passing comment on p304/5 of the ITOE. I have also read Peikoff's comments on induction in OPAR and find them very confused, like his discussion of the analytic/synthetic dichotomy in the ITOE. I think he's basically in denial...;-)

    So: a reason why intelligent people deny the validity of induction is because they accept the validity of logic. A reason why intelligent people cling to belief that induction is "the only true way" is that they are so emotionally committed to the idea of "the only true way" that they will ignore what their own reason tells them.

    ReplyDelete
  7. I wrote:
    >I have also read Peikoff's comments on induction in OPAR and find them very confused...

    I will expand on this briefly. Peikoff only mentions induction once in OPAR according to the index, and it is about as confused as it gets.

    "What the window of mathematics reveals is not the mechanics of deduction, but of induction." OPAR, 90.

    Now, according to the traditional view of induction one arrives at a theory only after making a number of empirical observations. For example, a naturalist might arrive at the theory that "all swans are white after having observed say 20, or 20,000 or more swans. Or I might arrive at the theory that the sun will rise tomorrow, based on the number of times I have seen it rise on previous days in my life.

    Nothing could be further from the truth in the case of mathematical propositions, however. The truth of the statement "2+2=4" was not arrived at following say 20 or 20,000 or more observations of putting two sets of two objects together! In fact, one knows it is true before any observations are made, just as one knows "1002+1002=2004" without making even one "inductive" observation of it in reality.

    This is because it is deductively true. Just because you can observe it to be true in reality, does not mean it depends on observations to establish it. It just means that deductive reasoning, like mathematics or logic can make useful statements about reality. But it does not require one, 20, 20,000, or even any "inductive" observations to establish that it is true.

    Peikoff's statement conjures up comic images of hordes of scientists dutifully totting up observations of the entire number system to confirm the "inductive" truth of mathematical formulae...;-)

    Anyway, this just illustrates how unclear he is on the basic issues. If mathematics is not "deductive" but "inductive", then what on earth does the man consider "deductive" then?

    ReplyDelete
  8. CORRECTION:
    >p304/5 of the ITOE

    ...should be p303/304.

    ReplyDelete
  9. Im sure I haven't thought about this nearly as much as you three (Greg, Daniel, Michael) have, but it seems to me that induction is valid and reasonable as long as we do not claim more than a degree of probability for our inferences. This degree of probability can be very high, approaching (but not quite reaching) 100%, or it can be very low, depending on the quality and quantity of the evidence and the explanatory value of our generalization. Of course, there will be disagreements about details, but the mere fact of disagreement doesn't mean that induction, per se, is irrational.

    For instance, if I touch the red-hot burner of my stove and get a painful burn, and later I do it again with the same result, it is reasonable for me to infer that touching a red-hot burner will cause a painful burn. I may not be 100% sure of this - if I were to have a film of water over my hand or a heavy callus on my finger, maybe I wouldn't get burned or feel pain - but the probability of getting a painful burn is very high. It is not irrational for me to avoid touching red-hot burners in the future, even though there is some nonzero chance that a particular burner would not hurt me. On the contrary: it would be irrational for me to refuse to draw this inference.

    ReplyDelete
  10. Michael P:
    >it seems to me that induction is valid and reasonable as long as we do not claim more than a degree of probability for our inferences.

    Hi Michael,

    This is where the going gets interesting. Because it turns out that Hume's argument means that you cannot even claim a degree of probability for such inferences! There have been many attempts to do so, for example Bayesianism, and all have failed to date. There is simply no way of predicting any probability that the next sunrise will occur or that the next swan will be white on the basis of previous sunrises or swans.

    This cold, hard logical fact is powerfully counter-intuitive to humans - rather like when we first learned the earth went round the sun. We have a hard time believing it can be true at first - even Hume himself could not quite believe it. But it is. In fact, the logical problem of induction to him formulating the psychological problem of induction.

    That goes like this:
    P1: Induction is irrational
    P2: Humans nonetheless use induction
    C: Humans are irrational

    This is how Hume arrived at the conclusion that reason is in fact "the slave to the passions". It was a devastating result, what has been called the "time bomb" ticking away under philosophy ever since. It effectively opened the door to later irrationalist philosophies (Kant was the first, albeit unsuccessful, attempt to restore reason after Hume).

    The strongest reply to Hume thus far has been from Karl Popper. He accepts the logical problem of induction as Hume formulated it. He regarded it as completely sound. However, he brilliantly challenged the psychological problem as laid out above. What if, he argued, P2 was untrue? What if it only looked like humans used induction? What if they used a different, but logically sound method of learning from experience?

    Popper called induction an "optical illusion". What humans do, he argued, is create theories or conjectures, which we then test by experience. Thus you do not have to keep putting your hand on the stove X number of times to establish an "inductive" knowledge that your hand will be burned! Once is enough.

    A child, for example, will have a theory that she can touch various things without pain. She will accidentally test that theory on the hot stove! This test will lead to a new theory...;-) And so forth.

    Thus Popper reverses the traditional picture of empirical observations somehow inexorably accumulating to produce a theory. In his scheme, the theory or imaginative conjecture comes first*. We then test it using logic and experience. The theories that survive such testing best are the ones we rationally prefer eg: that the sun will rise tomorrow, that you will burn yourself on a red hot stove. This, he claims is the real way humans gain knowledge. Not through inductive accumulation, but through conjecture and refutation, with the fittest conjectures surviving, just like in natural evolution.

    This is a very liberating notion. Because it also solves the centuries-old dispute, caused by the false belief in the inductive method, between reason and imagination - between the scientist and the artist. Under Popper's schema, the scientist is not the dull accountant of reality, dutifully totting up repetitive experiences, but a creative thinker herself, who imagines bold new theories and tests them as severely as she can to get closer to the truth. Thus process of science is creative and never ending, we are always trying to evolve new and better theories.

    Anyway, as you have gathered, there is a lot to this particular problem! I would encourage you to look into it further, particularly Popper's solution.

    *Basically, it turns out we have to have a theory before we can make an observation. To demonstrate this Popper used to open his classes on induction by commanding his students to "Observe!" He would then busy himself at his desk. Eventually a student would raise their hand and say: "Excuse me sir...but observe what?"

    ReplyDelete
  11. Daniel,

    Thanks very much for your reply. I took the time to read a little about Popper. The most common critique of his position is that he sneaked induction in through the back door. It goes like this:

    Popper says that some hypotheses are stronger than others. The stronger ones are those that have been more severely tested - i.e., those that have survived the most attempts at falsification. But why should the number of attempts matter? Each unsuccessful attempted falsification constitutes a new item of empirical data. To say that the strength of the hypothesis depends on the number of failed falsifications, then, is to say that the strength of the hypothesis depends on the number of empirical observations that support it. But this is induction, merely rephrased and presented in a somewhat convoluted or underhanded way.

    If Popper were fully consistent, he would have to say that no undisproven hypothesis is any stronger than any other. But this is clearly untenable. The hypothesis that the sun will rise tomorrow is clearly stronger than the hypothesis that there is buried treasure in my backyard, even though neither hypothesis has been disproven.

    How would you, as a Popperian, reply to this critique?

    ReplyDelete
  12. One more thing: Popper seems to say that some hypotheses are not more probable than others, only better corroborated. But this appears to be a purely semantic distinction, since the "better corroborated" hypothesis is understood to be "stronger," i.e., more probable. Again, we are back in the area of induction and degrees of probability, only using different terms.

    ReplyDelete
  13. MP:
    >How would you, as a Popperian, reply to this critique?

    Firstly, the level of Popper scholarship round and about is quite poor. This line of attack, along with many a supposed "point that Popper missed" is anticipated and replied to in his 1935 "Logic of Scientific Discovery", amusingly usually within the first 100 pages.

    In this case, from page 10:

    ""It should be noticed that a positive decision can only temporarily support the theory, for subsequent negative decisions may always overthrow it. So long as a theory withstands detailed and severe tests and is not superseded by another theory in the course of scientific progress, we may say that it has 'proved its mettle' or that it is corroborated' by past experience. Nothing resembling inductive logic appears in the procedure here outlined(italics DB). I never assume that we can argue from the truth of singular statements to the truth of theories. I never assume that by force of 'verified' conclusions, theories can be established as 'true', or even as merely 'probable.'
    (LScD, p10)

    Obviously this won't satisfy you, as there are many possible objections (the LScD is 400 pages long, and Chapter 10 is devoted to issues of 'corroboration' and the differences between this and probability theory). It took a while to convince me, and to get a handle on the details. But so far I think it holds up. And it also demonstrates that far from being suprised by these issues, as his critics seem to think, Popper is very much aware of them.

    Now, there is a semantic distinction to be noted, and that is the use of the word 'probable.'

    What Popper is tackling is the strict sense of the word "probable", not the loose, informal sense we use when we say "it will probably rain next Tuesday" or "the sun will probably rise tomorrow."

    For when we use such a term, we are not referring to any specific probability. Strictly speaking, the best we can say about calculating the probability of such future events based only on previous observed occurrences*, (ie: the number of previous Tuesdays on which it rained/mornings on which the sun rose) is that it is somewhere between 0% and 100%! In other words, it is a random guess, as there exists neither the formula for making such a calculation nor even a valid argument that such a formula could even exist.

    I know it seems powerfully counterinuitive. And it does take a while to get your head round, just as it must have been difficult for people to understand the earth went round the sun, or that men descended from amoeba. There's a few distinctions to be made too: for example between possessing an absolutely true theory (possible, but difficult) and knowing it is absolutely true (impossible) But trust me - after a while it begins to make sense.

    *We should be careful to distinguish between these events and say, the prediction of a single dice throw at 1/6, as the dice probability is arrived at deductively.

    ReplyDelete
  14. MP:
    >If Popper were fully consistent, he would have to say that no undisproven hypothesis is any stronger than any other.

    If I am not mistaken, this is what he does say.

    ReplyDelete
  15. I wrote:
    > it is a random guess..

    To be clear, it is a random guess that it will rain next Tuesday/the sun will rise tomorrow if and only if we purely use the inductive method of enumerating past instances. We must be careful that we do not let various deductive theories "slip in the back door" eg average rainfall in a given month/Newton's laws of gravitation etc. For using these we can still arrive at a critical preference for what is "likely" in the informal sense. And indeed, this is what we do when we think we are thinking "inductively."

    ReplyDelete
  16. M Kelley: "Such arbitrary limitations, denying the existence of induction, or proclaiming it as the only true way, are very strange to me."

    I should perhaps clarify here. When I say there's no such thing as induction, I am denying the claims of theories of induction—i.e., that consciously reasoned inferences from the particular to the general lead to something called "knowledge." I am not denying that people make such inferences, or even that such inferences can sometimes, by sheer luck, turn out to be true. But the ability to for us to know whether a given so-called "inductive" inference is true does not depend on the inference itself. There are an infinite number of possible inductive inferences, of which only a small handful are true. How is one to distinguish true inductive inferences from false? Well, we know that examining the structure of inductive inference won't help us at all. At least with logical inferences, studying the structure of the inference can help us determine whether the argument is valid, but the concept of validity simply doesn't apply to inductive inferences. In other words, the structure of the inference tells us absolutely nothing as to the truth of that inference's conclusion. How, then, are we to determine the true of an inductive inference? In science, truth about general propositions is determined by testing the empirical content of the proposition, that is, by trying to refute it, preferably through empirical experiments. Meaningful universal statements that have proven their mettle over the years have a strong presumption in their favor. They can generally be regarded as at least very good approxtimations of the truth and as having great practical utility. But note that their cognitive value in no way derives from whatever kind of inductive inference might have been used to formulate them originally. Their status as knowledge has nothing to do with induction itself; and hence to say that our knowledge of universals is based on or derives from induction is not true. There is no such thing as knowledge of universals through induction. Knowledge cannot be attained by that method--only blind guesses. Knoweldge of universals derives from testing and criticizing hypotheses.

    ReplyDelete
  17. >>If Popper were fully consistent, he would have to say that no undisproven hypothesis is any stronger than any other.

    >If I am not mistaken, this is what he does say.

    Now I'm really confused (not an unusual circumstance). Would Popper say that the hypothesis "the sun will rise tomorrow" is no stronger than the hypothesis "there is buried treasure in my backyard"?

    It seems obvious that the first hyothesis is stronger than the second - or, more generally, that some hypotheses are stronger than others. Is he really saying that all conjectures hold equal weight until disproven?

    It seems to me that there is an insuperable difficulty here. Either all nondisproven conjectures have equal weight, or some have more weight than others.

    The first proposition seems obviously incompatible with experience. The second proposition seems to mean that as a conjecture passes various tests (attempts at falsification), it becomes stronger, i.e., weightier, i.e., more probable. But this is inductive reasoning, precisely what Popper wants to avoid.

    I recognize that the issue is complex, and that my objections may be misguided. When I have the time, maybe I'll take a look at The Logic of Scientific Discovery and see how Popper addresses these arguments.

    ReplyDelete
  18. Maquiavel (and Daniel and Michael and others),

    Induction is a very interesting topic to contemplate and discuss. I have found that the purpose of induction is usually missed in discussions of this kind. It is given a charge unsuited to its nature, then that charge is criticized. Induction is not to be used for establishing a fact. A fact exists already without any mental processing, inductive or deductive.

    Induction exists to validate the human mind as a proper tool to abstract knowledge from facts. Nothing more. (This is a long discussion--too long for here. I can elaborate over time as our discussion progresses.)

    Rand was usually vague about this purpose and often went too far in missing it by emphasizing certainty of fact, not certainty of being able to know a fact--and her self-proclaimed "intellectual heir" Peikoff, of course, went whole hog, claiming that all knowledge is inductive.

    My idea is that there can be no "empirical evidence" without a valid mind to process it. Even Daniel stated that an absolute truth can be known. But known by what? A mind? How do you judge the validity of that mind?

    Popper and others in this line of thinking do not provide a basis for even validating "falsification" as an epistemological process. They take logic (and the mind) as the given. To be fair, though, I still have a lot of reading to do. I am merely going by what I have read so far.

    Induction ultimately leads to fundamental axioms. I once stated that axioms are used as an interface between the mind and the rest of reality--sort of like the Assembler language in computing, or even the division of single binary units into groups of bits and bytes. It sets an initial standard for processing. I think this is much more reasonable than claiming that induction does not exist at all for gaining knowledge and accepting the mind itself as the equivalent of axiomatic knowledge, so to speak.

    btw - Maquiavel, are you Mr. Nyquist? Pleased to meet you if you are. If not, I loved your book, The Prince.

    //;-)

    btw - My name is spelled "Kelly." There is another "Kelley," but he is a David. Many people make this mistake from incorrect induction.

    Michael

    ReplyDelete
  19. M Kelly:
    >Even Daniel stated that an absolute truth can be known.

    Whoa there...;-)

    What I said was:
    "There's a few distinctions to be made too: for example between possessing an absolutely true theory (possible, but difficult) and knowing it is absolutely true (impossible)"

    What I mean is as follows: you may in fact have arrived at a final, absolutely true theory. This is just possible (but unlikely, given the infinite amount of theories that can be generated around any given set of facts). But even if you possess it, you cannot know for sure that it is absolutely true.

    They are two different things: possessing a absolutely true theory, and finally knowing that it is absolutely true.

    An example is Newton's theory, which was astonishingly accurate and successful for a couple of hundred years, and seemed unassailable for all time. Then came Einstein, and it turned out Newton's vision of the universe was wrong. Now, it doesn't matter whether the time is 200 years or 2 million. The problem is the same.

    >Induction is not to be used for establishing a fact.

    These and other remarks you make here suggest by "induction" you mean something different quite different from its usual senses. What I'm talking about is the standard versions I mentioned in my initial reply to you i.e.

    1. Generalizing about the properties of a class of objects based on some number of observations of particular instances of that class of objects (for example, "All swans we have seen are white, and therefore all swans are white"

    2.Presupposing that a sequence of events in the future will occur as it always has in the past

    If you disagree with these too, then we're largely on the same page.

    Michael P:
    >Would Popper say that the hypothesis "the sun will rise tomorrow" is no stronger than the hypothesis "there is buried treasure in my backyard"?

    It would depend on the severity of testing that you'd conducted on both hypotheses. Clearly the 'sun rising' hypothesis has survived a huge amount of empirical and deductive testing over the millenia - which have actually produced some falsifying instances, for example around the Poles. Thus it is definitely a better and far more refined theory than your ad hoc, untested theory about your backyard containing treasure.

    Let us then bring a bulldozer into your back yard, and dig down to 150 feet deep, deeper than most pirates can dig, across the entire section and most of the neigbours. We find no treasure. Hence we have tested your theory (and ruined your backyard) severely. We can then quite rationally choose to reject it (while still leaving open the possibility we may have erred and overlooked it).

    But note: we have only done this once. Therefore is nothing "inductive" about this decision.

    ReplyDelete
  20. I should also add it depends on your emotional commitment to a given idea. A truly scientific attitude means ultimately being prepared to give a belief up. If you really really wanted to believe there was treasure in your backyard, no degree of testing would be sufficient...;-)

    ReplyDelete
  21. Mike P, it might be helpful if I concretise this even further.

    Here's what an inductive formula should look like:

    I have seen X number of swans, therefore it is 100% certain that all swans are white.

    Or

    I have seen X number of swans, therefore there is Y probability that all swans are white.

    The thing is to give a specific number for the "X"s and/or give Y a probability more specific than between 0% and 100%.

    This cannot be done. Nor is there any logically valid principle that suggests that it can be done.

    ReplyDelete
  22. I'm afraid this does not solve Popper's problem. In the bulldozer example, you are assuming that we could dig up the backyard, establish that there is no treasure there, and hence falsify the conjecture. But what about conjectures that have not been falsified but are still highly improbable?

    For instance, suppose I advance the hypothesis that some swans are purple. This has not been falsified; nevertheless it is unlikely to be true. Why is it unlikely? Because I cannot present any positive evidence for it. Meanwhile, I can present positive evidence for the hypothesis that the sun will rise every day - namely, the unbroken string of sunrises extending back to the beginning of history.

    It's true that I cannot mathematically quantify the degree of probability of the first conjecture vs. the second. I also cannot mathematically quantify the love I feel for a spouse vs. the love I feel for a pet, but that doesn't mean there aren't different degrees of love involved. It just means that not everything can be reduced to logic and math.

    I agree that induction is problematic in many respects, but to "solve" the problem by saying induction doesn't work at all seems, to me, like throwing out the baby with the bathwater.

    ReplyDelete
  23. MP:
    >It just means that not everything can be reduced to logic and math.

    Well, Hume's criticism of induction is of its invalid logic. Popper's attempt is give an alternative, valid account of the "Logic of Scientific Discovery", given that Hume has not been successfully refuted.

    If we remove logic from the picture then neither Hume nor Popper is a problem.

    ReplyDelete
  24. I'm afraid I'm failing to get my point across, so let me give it one more shot. Let's take two hypotheses. The first is that the sun will rise every day for the next 100 years.

    The second hypothesis is this: Let's say we discover that Titan, a moon of Saturn, has Earthlike conditions. Since similar conditions gave rise to life on Earth, our hypothesis is that Titan also has life on it.

    Nether of these hypotheses can be tested immediately. We can't test the proposition that the sun will rise every day for the next 100 years, or that Titan has life. But both propositions are testable in principle. Given sufficient time, we can find out if the sun will continue to rise for the next century. Given the right technology, we can send a probe to Titan and look for life.

    At the moment, then, we have two testable hypotheses, neither of which has been falsified. As I understand it, according to Popper, no probability can be attached to either one of them, and neither can be said to be more probable than the other.

    This, however, is transparently untrue. The first hypothesis is obviously more probable than the second one. Why? Because centuries of experience and observation have taught us (yes, inductively) that the sun does rise daily, and we have no reason to doubt that it will continue to do so. On the other hand, we have no clearcut evidence of extraterrestrial life.

    Although the probabilities cannot be quantified, it is self-evident that the first hypothesis is more probable, i.e., more likely to be true, than the second one. I don't think anyone would seriously dispute this.

    Popper tries to get around this problem by talking about "severe testing" and "corroboration," which render some hypotheses more battle-hardened and reliable than others; but this is a smokescreen. The testing and corroboration of a hypothesis simply amount to gathering evidence that supports it. Every attempted falsification that fails is another datum of confirmatory evidence. To say that a hypothesis has survived repeated attempts at falsification is simply to say that a growing pile of evidence (inductively) supports the hypothesis.

    No matter how you look at it, induction remains part of the picture.

    >If we remove logic from the picture then neither Hume nor Popper is a problem.

    This is silly. My point is that not everything can be quantified.

    ReplyDelete
  25. MP:
    >I'm afraid I'm failing to get my point across, so let me give it one more shot.

    No, no, I think I'm getting it. This is a reasonably common line of enquiry. I just think there are some distinctions to be clear about if the problem is going to unscramble itself. Firstly, there is the formal "probable" ie:

    >As I understand it, according to Popper, no probability can be attached to either one of them, and neither can be said to be more probable than the other.

    And this is formally true, ok? If you can't quantify it, you can't quantify it. And I'm not just talking about quantifying it exactly within a narrow range. I'm talking about at all, between 0 and 1 probability. No one can seriously dispute the formal truth of this.

    >Although the probabilities cannot be quantified, it is self-evident that the first hypothesis is more probable, i.e., more likely to be true, than the second one. I don't think anyone would seriously dispute this.

    Now in this case, we're talking about the informal use of the word "probable" (or "likely" etc). This is just the same as saying "my best guess". As I say, I agree that centuries of "experience" and "observation" - not to mention deductive theories such as Einstein's - means this theory is well-tested. But this is not the same as "induction" - it just looks like induction.

    >To say that a hypothesis has survived repeated attempts at falsification is simply to say that a growing pile of evidence (inductively) supports the hypothesis.

    Yes we can say evidence "supports" it or "corroborates" it. But we cannot make logically valid claims for its truth on that basis.

    >This is silly. My point is that not everything can be quantified.

    Well you did say "not everything can be reduced to logic and math". I agree. Popper doesn't mind "inductive" intuitions or other non-rational sources of new ideas, such as the imagination. But to repeat - his work is on the logical structure of scientific discovery - ie the best way to test these new ideas.

    I also can't see how we can really say that one theory is more probable than the other, other than in the loosest, most informal sense above (ie based on our "best guess", or what we'd call the best surviving theory)

    Look I'm sure this comes off as a lot of Popperian double talk to you...;-) It did to me for a long while before I finally got my head round it. As it happens I don't recommend LScD as a good Popper introduction. It's a bit hard to chew over as a first bite. I think you should read "The Open Society And Its Enemies" to start with. Its a far more wide-ranging book, one of the best of the last century. It's genius. If you don't like it, I'll pay for your copy...;-) And David Miller's "Popper Selections" puts it all in a nice nutshell.

    Give it time. Let it all soak in. Take your doubts to the limit. Read the people who think Popper is an old faker, like David Stove. Ruminate. Mull. Consider deeply. Consider shallowly! This is one of the big philosophical problems of all time, way more important IMHO than the (non) problem of universals. Enjoy the ride. There's no hurry....;-)

    ReplyDelete
  26. Good advice! I'll look into it further when I have the time. Meanwhile, thanks for bearing with me.

    And oh, yeah ... Happy New Year!

    ReplyDelete
  27. My pleasure. I'm not even sure I've explained it that well, but anyway. A very Happy New Year to you too!

    ReplyDelete
  28. Micheal: "Every attempted falsification that fails is another datum of confirmatory evidence. To say that a hypothesis has survived repeated attempts at falsification is simply to say that a growing pile of evidence (inductively) supports the hypothesis. "

    In one sense, this is true, but it misses the point Popper is trying to make. It may seem like a bit of a quibble, but if one understands why Popper made it, it doesn't seem so unreasonable. The notion that a "growing pile of evidence (inductively) supports the hypothesis" suggests induction by enumerating instances. This form of induction seeks to establish the hypothesis in question by gathering confirming evidence. Popper regards as problematic, because it fosters an uncritical attitude. A person who only looks for evidence that confirms his theory is, in Popper's view, looking for the wrong thing. What instead he should be doing is looking for evidence that refutes the hypothesis. This fosters a more critical attitude and therefore increase the quality of the scientist's empirical tests. And that is really the point Popper is trying to make. He's emphasizing the importance of quality over quantity. Popper's not impressed by a growing pile of evidence, because if that pile is accumulated uncritically, by someone actively looking for evidence to confirm his hypothesis (and who seeks usually finds what they're looking for), it's worth very little in comparison to a significantly smaller pile accumulated by those scientists who are actively seeking to disprove the hypothesis in question.

    Of course, you could argue that since even the smaller pile is a pile that it amounts to induction in the end. But that's more a question of semantics. Why argue about words. You want to call it induction, then call it induction. But there certainly is an important difference between Popper's theory (whether one wants to call it induction or not) and traditional theories of induction. I think the difference is important enough to warrent not calling it induction.

    ReplyDelete
  29. >What instead he should be doing is looking for evidence that refutes the hypothesis. This fosters a more critical attitude and therefore increase the quality of the scientist's empirical tests.

    In theory this is true. In practice, I find that people can be just as careless in seeking to disprove a hypothesis as in trying to prove it. For instance, when Fleischman and Pons announced "cold fusion," other scientists immediately leaped to disprove it. In short order they did so, to their own satisfaction, by failing to replicate the results. However, further experiments by other scientists have shown that some unusual phenomenon is indeed going on. The falsifiers were too quick to find disconfirming evidence.

    Similarly, when Jacques Benveniste announced that chemicals could be diluted out of water but the water would still retain some of the properties of those chemicals (the "memory of water"), a team of skeptics quickly discredited him by failing to reproduce his results. But subsequently, several other labs have verified his work. Again, the rush to find falsifying data led to a premature dismissal of a worthwhile avenue of research.

    Examples could be multiplied, but the point is that bias will always crop up, whether one is seeking to prop up a pet theory or shoot down a rival theory.

    ReplyDelete
  30. Michael, you might want to have a read of this too:

    http://www.la-articles.org.uk/popper.htm

    Here Martin Gardner runs the line you are proposing, more or less, and Jan Lester nicely explains the situation. The bottom line is: you can't get from single instances to universal scientific theories. Chuck in the cognitive bias problem along the way, sure - falsification is a helpful attitude here too. But the key to it is the straightforward logical implications.

    ReplyDelete
  31. MP: "For instance, when Fleischman and Pons announced "cold fusion," other scientists immediately leaped to disprove it. In short order they did so, to their own satisfaction, by failing to replicate the results."

    This sounds like the usual conspiracy thinking by crackpots. Countless scientists just tried to replicate the experiment, as it would be of tremendous significance if the results were valid. It soon turned out however, that the results could not be replicated, and that they were spurious. After that the scientific community soon lost interest, only a small group of True Believers continues to look for the Holy Grail, which they certainly won't find, as it doesn't exist.

    History has shown that if there is really something new discovered, that it is taken seriously by the scientific community, even if it goes against the then current theory: special and general relativity, quantum mechanics, the Big Bang, superconduction, black holes, dark matter, dark energy. There have been enough scientists who were skeptical of these developments, but when the evidence was good enough, the new ideas became quickly accepted.

    Benveniste is really a pathetic case, there isn't a shred of evidence for his theory that water has "memory", let alone that this memory can be stored digitally and sent over the Internet to be transferred to another solution, that is pure crackpottery. No need to blame those evil skeptics for not believing such crap.

    ReplyDelete
  32. Calm down, Dragonfly. They were just examples. :-)

    The crucial MIT paper that "disproved" cold fusion was later shown to have been fudged; the researchers obtained results compatible with Fleischmann and Pons' claims, and misrepresented them.

    http://www.alternativescience.com/cold_fusion.htm

    Benveniste's work was verified by four European laboratories in a series of experiments supervised by a skeptical chemist who was trying to disprove "the memory of water" thesis. More work is being done by Eastern European scientists.

    http://www.guardian.co.uk/Archive/Article/0,4273,4152521,00.htm

    http://thescotsman.scotsman.com/index.cfm?id=1093712003

    In any event, even if my examples are poor, it's hard to deny that people can be just as biased in trying to shoot down an unpopular theory as they are in trying to support a popular one.

    Unfortunately, this kind of thing is probably becoming even more common as the growing expense of scientific experimentation makes researchers increasingly dependent on corporate and government grants. The donors are not likely to fund anyone regarded as a "crackpot" by the establishment, leading to a sad uniformity of thought. Hence we have fifty years of hot fusion experiments that have brough us no closer to generating electric power, thirty years of string theory that has gotten us nowhere, fifty years of assurances that a cancer cure is right around the corner, etc.

    ReplyDelete
  33. MP: "Benveniste's work was verified by four European laboratories in a series of experiments supervised by a skeptical chemist who was trying to disprove "the memory of water" thesis. More work is being done by Eastern European scientists."

    See for example http://tinyurl.com/y29qqk :

    "Recently, an allegedly positive replication of this study was reported. This new experiment was performed in 4 European laboratories, and a member of Benveniste´s original team took part in them. This study, however, was only published as a short, non-peer reviewed article, and, apparently, had methodological flaws that could compromise its results. For instance, the authors reported that ¼ of their data had to be discarded, although no clear justification was presented for that. Also, the statistical techniques used appear inadequate."

    and http://tinyurl.com/yyuyzy :

    "The Noble Countess appears not to have noticed that the first author on both of Ennis's papers was Philippe Belon. who is a director of the huge French homeopathic company, Boiron. In fact the address on the papers is not the University of Belfast (or Louvain), but it is 'Boiron, 20 rue de la Liberation, 69110 Sainte-Foy-Les-Lyon, France.'

    Boiron makes profits from homeopathy of about 20 million euros a year, on net operating revenues of about 300 million euros."

    So Ennis' experiment is hardly an independent replication and Belon is supposedly not biased, yeah, sure...

    It's again the same old song: when you have a closer look at the study that gives positive results, it turns out that it contains many flaws, and as soon as you tighten the controls, all the positive results disappear like snow in summer. And rare is the case that the original experimenters admit that their experiment might have been flawed, no, of course those pesky skeptics are to blame, as they're biased (which they themselves of course are not).

    MP: "In any event, even if my examples are poor, it's hard to deny that people can be just as biased in trying to shoot down an unpopular theory as they are in trying to support a popular one."

    Individual scientists may be biased in favor of their pet theory, but that isn't really a problem, as other scientists may favor a different theory, and it is ultimately the quality of the theory that decides whether it will be generally accepted by the scientific community. The theory that fits the data better and that gives the better predictions is the winner. The system of thousands of different scientists all over the world collaborating and competing works very well in weeding out bad theories and promoting good theories. The idea that there is some kind of conspiracy to keep certain theories at bay, is a fantasy that only exists in the febrile brain of the frustrated quack.

    MP: "Hence we have fifty years of hot fusion experiments that have brough us no closer to generating electric power, thirty years of string theory that has gotten us nowhere, fifty years of assurances that a cancer cure is right around the corner, etc."

    The problem with hot fusion is not the principle that is used(it has been shown that it works, in contrast to that of "cold fusion"), but the enormous technological problems of making an economically feasible reactor. This is in itself nothing special: the principle of the rocket is many centuries old, but it took about fifty years of modern research to build a rocket that could put some men on the moon. The long-term advantages of fusion reactors are so great that such efforts seem worthwile.

    In contrast to the primarily technological problem of fusion reactors, string theory is a purely theoretical subject. The problem with string theory, supersymmetry and GUT's is that they are not directly experimentally verifiable. If such theories make predictions, these could only be observed in conditions that we can't realize in our laboratories. In fact everything that we can test in the laboratory and the big accelerators can perfectly be explained by the so-called "standard model", based on quantum mechanics and quantum field theory. The agreement between theory and experiment is extremely good, in some cases even up to fifteen decimal places! So in a sense the standard model can explain everything that we observe on earth. The reason that we look further is that we want to find a theory behind the standard model, a theory that might explain the values of the free parameters in the theory and that might also give us a quantum theory of gravity. While this is of great theoretical interest, it's unlikely that it will generate practical applications, at least in the short term. The standard theory could be developed in a few decades thanks to the parallel development of the big accelerators, which made testing of the predictions of the model possible. This is no longer the case for the more advanced theories, so these are much more speculative. There may be possibilities to test such theories in cosmology; in that field there are still many puzzling features, like dark matter, dark energy, black holes. However, this is much more difficult than doing an experiment in a laboratorium, you can't just say: put some dark matter in the test tube, then we'll study it. So while it would be nice if we could discover new theories in the same tempo as that of the development of QM and the standard model, it is not surprising that the progress is much slower now. That is no reason however to resort to crackpot theories. Just as Relativity and QM have Newtonian mechanics as limiting case, any new theory should have the standard model as limiting case.

    With regard to the statement "that a cancer cure is right around the corner": I think that this is more hyperbole from ignorant reporters than a viewpoint that is expressed by any responsible scientist in the field, at least in the last decades. The general consensus is now that there will be no single miracle cure, there are many types of cancer, that will have to be treated in different ways. There is progress, but it is slow. That may be regrettable, but that's a fact of life, and certainly no reason to resort to homeopathy, acupuncture or faith healing, as these will certainly not cure cancer.

    ReplyDelete
  34. Thanks very much for the links. They're interesting and give another perspective on Benveniste's work. Btw, Benveniste himself was not interested in homeopathy. (Neither am I.) He was interested in the idea that molecules interact, in part, through vibrational frequencies that would be retained in water even when the chemicals themselves had been diluted out.

    Not being a chemist, I have no firm opinion on this, but it does seem that his work has been replicated by some labs. I recognize the possibility of bias on the part of Benveniste and his associates, but there is also the possibility of bias on the part of mainstream scientists with reputations, grant money, and egos to defend. Bias is found on all sides, and while eventually the good ideas may win out, it can take a long, long time.

    >The idea that there is some kind of conspiracy to keep certain theories at bay, is a fantasy that only exists in the febrile brain of the frustrated quack.

    There is no conspiracy or cabal. There is only the human, all-too-human tendency to defend one's pet theories and cherished opinions ... as you and I are doing right now! :-)

    This, by the way, is the most fundamental problem with Objectivism: in glorifying the ego, Rand makes it very difficult to move past ego-based behaviors and thought patterns. But it's only by detaching from the ego that we can get any perspective on our biases, blind spots, wishful thinking, etc. Objectivism, far from encouraging objectivity, actually makes any approach to objectivity impossible.

    But this is par for the course with a "philosophy for living on earth" that can't actually be lived on earth, and a philosophy of personal happiness that makes people miserable ...

    Cheers, and Happy New Year!

    ReplyDelete
  35. MP:"Objectivism, far from encouraging objectivity, actually makes any approach to objectivity impossible."

    That struck my time and again in my discussions with orthodox Objectivists, the lack of objectivity in their thinking. Often they twist themselves into pretzels trying to defend the obviously indefensible: "no, Rand didn't mean that, you should interpret her statement as...", or "aw, she was speaking off the cuff, so you can't take it too literally". The idea that she just might be wrong in something is apparently unthinkable. I wonder if they secretly fear that by admitting the possibility of error the whole system will soon crumble down - a justified fear, I think.

    ReplyDelete
  36. Hey there. I've been working my way through the archive here, and I've got a question about this problem of induction. I apologize if it's already been addressed in all of these comments or if it's just a question in general.

    Anyway my question: Doesn't our scientific understanding of swans and sunrises have some bearing on predicting these things? By which I mean to say, in the proposed scenario, it sounds like we just have some bumpkin sitting on his porch watching swans go by. He sees 20 white swans, so he predicts on the basis of that alone that the next one will be white. But in reality, we've also learned things about swans and genetics and why they're white and so forth, right? I totally realize that this just sets this issue back and leads to a finer example of the same problem concerning the basis for the scientific evidence in question, but that still must count for something, doesn't it? Otherwise, that would just completely undermine the entire foundation of science.

    I mean, if we're betting on sunrises, it's not JUST on the basis of the fact that the sun has risen every day of our lives, but also on a lot of observation and theory about planetary rotation and gravity, not to mention the astronauts who have witnessed this rotation first hand. If the sun DIDN'T rise tomorrow, it would be more than just a break in a running streak of sunrises. It would upset every single notion we have about the physical workings and nature of the universe. That seems to suggest to me that we have at least SOME fairly solid reason to expect it to happen.

    ReplyDelete
  37. Bryan:
    >If the sun DIDN'T rise tomorrow, it would be more than just a break in a running streak of sunrises. It would upset every single notion we have about the physical workings and nature of the universe. That seems to suggest to me that we have at least SOME fairly solid reason to expect it to happen.

    Hi Bryan,

    This is a widely debated problem as you know.

    But the short answer is that it stems from a problem in logic - that you can't derive a universal conclusion from partial observations, no matter how many.

    This is a specific problem for science, as it really seeks to find universal laws. And we would like to use logic properly as part of the quest - it's an important part of our rational toolkit. So the question becomes "how can we best use logic in science?" At first it was thought the answer was - well, you really can't. Which was something of a shock, as scientists like to think of themselves as logical. But it turns out that there is a logical asymmetry we can take advantage of: while no amount of swans will prove the universal theory "all swans are white", one black swan will prove it is false.

    So we should look to test our theories by trying to falsify them, rather than trying to prove them. All the laws of the universe that underly our modern expectations about the sun rising (and even predict when it won't - say, in Scandanavia in winter, or 5 billion years in the future) are really theories that have survived testing so far. That means, like say your trusty old hammer that's been so useful for all these years, we should definitely keep using them for the time being. However we should not be surprised if they one day confront a problem where they break down, as Newton's did at the beginning of the 20th C after hundreds of years of predictive success. In which case we will need a new theory...

    So that's the very short version!

    ReplyDelete
  38. Ahhhh, I see. Yeah, I can understand that.

    It's always been one of the fundamental attributes of science that theories and even laws could need to be revised in light of new evidence. So I have no problem accepting this.

    Thank you for responding to a comment on such an old post :D

    ReplyDelete