Photos

30 Apr 2014

Criticism of Critical Rationalism.

Critical rationalism is the idea that all our conjectures (guesses, theories, explanations) about the world are fallible, radically so. It is the idea that there is no such thing as justification. This is because justification itself leads to what is seen as an infinite regress of justifications of our justifications.

Recently I have been discussing the merits of CR with a facebook user, one who has kindly offered many  criticism on the what he calls the methods and foundations of CR. Criticisms that might appear at first to be reasonable, but later turn out to be trivial points that lead to no problematic conclusions. The first one he terms the "recursive falsificatory loop" later he say he means spiral rather than loop, but for our purposes this is irrelevant (His view can be found here). The recursive loop/spiral is this:

"If the Critical Rationalist’s falsification attempts are themselves fallible, shouldn’t he subject his falsification attempts to criticism as well? And shouldn’t he then further criticise the criticisms of his criticisms?"

A critical rationalist in a sense is not obligated to do anything other than accept that his explanation is fallible and keep it open to criticism from others. This is exactly inline with a critical rationalist point-of-view.

The reason that a person does not indefinitely criticize his own theory is because he is trying to solve a particular problem and at some point we need to apply  our explanation to the problem in order to solve it.

The other idea is that if there are two solutions to a problem which of the solutions should we use? We can use whichever one we want, this might be based on many factors in  a person or groups environment and preferences, but does not necessarily have anything to do with the merits of the theories in question.

The fact that we have two solutions leads to further problems-situations, if we have two explanations where we can further elaborate on the problem and create hypothetical situations in which each one would fail we can do so, but this creates a problem situation we can put aside in order to solve the problem at hand.

 The recursion here is just the idea that one problem-situation might lead to discovery of more problem situations, in which we have to criticize our current theories and the deeper theories they are embedded in. Whether a  theory "needs" to be criticized is determined by the people involved in a problem-situation.

The reason a person is not obligated to criticise his own theories or his own refutations indefinitely is because this would lead to him to not apply a solution to the problem he is ostensibly trying to solve, and therefore, if he has a theory, not putting his theory to the severest possible test ie the empirical test.  If there are refutations that we have of a particular theory, there will be a little dogmatism from people who hold the theory, So therefore if a person does not involve himself in criticizing his own refutation this role will be filled by others who wish to show that the refutation was not actually a refutation. This is because a person engaged in a problem-situation is not trying to get information necessarily beyond his own problem-situation, but knowledge itself has a way of encroaching on other people's problem situations, and in this way all a critical rationalist expects is that his solution will come under severe criticism from other people in different problem situations, and maybe at a later date a new problem situation he is in that has lead to him reconsidering attributes of his earlier solution that are based on faulty assumptions. This is the idea that although our theories are fallible, there is no point in assuming they are false, until we have a situation in which the explanation itself seems inadequate, even if there are two solutions (it is true that at least one of them is false, but we don't know which or if in fact both are false). This is why each theory that falsifies another theory explains why the early theory predicted the way it did and also explains its failure to solve the current problem at hand.

The reason this situation is different from the justificationist stance, is simply because the recursive function of justification does not lead to any new problem-situations, it leads to one problem: a regressive justification of justifications. Which is a non-problem, because justification is false.

6 comments:

  1. This is great too: http://io9.com/how-bayes-rule-can-make-you-a-better-thinker-471233405

    ReplyDelete
  2. Ugh, had some difficulty setting up the commenting account. Anyway, here it is.

    The other idea is that if there are two solutions to a problem which of the solutions should we use? We can use whichever one we want, this might be based on many factors in a person or groups environment and preferences, but does not necessarily have anything to do with the merits of the theories in question.

    The issue is not so much that there are two solutions. The important thing to note here is that if there are two theories either of which could be true, it makes more sense to look at the chances of each theory being true. If you need to make a gamble anyway, it's best to gamble with both theories in mind, instead of arbitrarily choosing one theory and gambling based on that theory.

    Also, here's a relevant page: http://lesswrong.com/lw/1to/what_is_bayesianism/

    ReplyDelete
  3. "The important thing to note here is that if there are two theories either of which could be true, it makes more sense to look at the chances of each theory being true."

    What would indicating the probable truth of a theory solve, and how would we measure the potential truth of it?

    Critical rationalism says that truth is not a function of evidence, but a function of whether what the theory says matches with reality, this designation is either 1 or 0. If we have two theories at least one of them is true, we can't say anymore than that.

    As far as I know Bayesian epistemology does not say anything about truth, but about how we should update our beliefs in response to evidence.

    ReplyDelete
    Replies
    1. Either 1 or 0 is a false dilemma. Yes, if you're thinking in terms of whether a theory is either true or false, then you end up with either 1 or 0. But if you think in terms of how likely a theory is true, then you have a range of 0% to 100%.

      Well, I'll leave it up to you to explore Bayesian theory on your own. It helps a lot to first study how Logic works. Stuff like logical fallacies, modus ponens, modus tollens, etc.

      Delete
    2. This comment has been removed by the author.

      Delete
    3. "Yes, if you're thinking in terms of whether a theory is either true or false, then you end up with either 1 or 0."

      That is what we (well at least I am) are doing.

      "Either 1 or 0 is a false dilemma." You agreed that it was not given that we are thinking about whether a theory is in fact true, which is exactly my point, so therefore it is not a false dilemma, unless you think that a theory is not trying to be true.

      "But if you think in terms of how likely a theory is true, then you have a range of 0% to 100%."

      What is the relevance of such a thing. The logic of this fails, anyway, because judging how likely a theory is is based on judging how likely the theories it is based on are, which means that either Bayesian theory is about adjusting belief about a theory in response to evidence (which seems like an exercise in futility) or it leads to an infinite regress of likelihood, which means no likelihood of anything being true. Bayesian theory is thus useless or false.

      Delete