What is wrong with Critical Rationalism?

Miscellaneous General
1

This is the thread I promised to make here:

First I will set up Logical Positivism as a foil to demonstrate how Critical Rationalism is supposed to work. Half of what makes Logical Positivism what it is is the verification criteria. If any statement is to have cognitive meaning it must have some step (or procedure) that would (in principle) verify if it is true or false. How science is supposed to work is that scientists will put forward theories (preferably in a commonly understood logical language) that would have a procedure by which its truth be confirmed or denied. Almost all of science can be reconstructed, so as to show that scientists are doing the following kind of reasoning:

If theory P is true, then we will witness experimental result Q.

We witnessed experimental result Q.

Therefore, theory P has been confirmed.

Remember that confirmed doesn’t mean proven or even likely. If something is confirmed many times in different conditions we can safely have confidence in the truth of the theory.

Those of you who are familiar with logic will have already noticed the problem with the above form of reasoning. It is deductively invalid and commits the formal fallacy of affirming the consequent. Yes boys and girls, modern science was built on a logical fallacy. This is the first problem with Logical Positivism. The second, is that it itself doesn’t not meet its own criteria for cognitive meaning and therefore, is self-refuting. Logical Positivists must admit that Logical Positivism is meaningless.

Next up is Critical Rationalism. It’s logical form looks very similar to the confirmation that takes place in the above example, but has an extra feature: it is deductively valid (a nice feature indeed!). The reasoning goes like this:

If theory A is true, then we will witness experimental result B.

We did not witness experimental result B.

Therefore, theory A isn’t true.

Those familiar with logic will recognize this as the valid inference rule, modus tollens. This escapes the problem of invalidity, but gives us a theory of science that doesn’t have any confirmation of theories at all (only falsification). In this way, no theory can be shown to be true, but a theory CAN be shown to be false. Interestingly, you could have the One True Theory of the Universe, but you could never know for sure. All you could know is that you haven’t falsified it yet.

To show why Critical Rationalism runs into trouble I’ll symbolize my reasoning so that everything is crystal clear:

  1. A ⊃ B

  2. ~B

∴ ~A

In proposition 1, A can’t just be “the theory is true”. It must be “the theory is true” and other conditions. To see why this is so, think about this:

If the moon is gray, then when I point my telescope at the moon I will see gray.

I didn’t see gray.

Therefore, the moon isn’t gray.

The problem is that there may have been an extra condition that was falsified and not the theory itself. What if I left the cap on the lens of the telescope? Would my theory that the moon is gray still be falsified? No, it wouldn’t. All it would show is that when you leave the cap on the telescope you see black (and not gray). So instead of symbolizing it the way I did above I would symbolize it like this:

  1. (A · B) ⊃ B

  2. ~B

∴ ~(A · B)

As long as it takes a conjunction and not a simple proposition to make the kind of conditional statements that could be falsified, we can never be sure that it is the theory that is false and not one of the other conditions. All we can say (in the case of falsification) is that one or more parts of the long conjunction that makes up the antecedent of our conditional proposition is false. You can’t really test a theory in isolation, but entire belief systems.

Positivism, Logical Positivism, etc..
2

This seems to me to be a limiting factor (with empirics in general and not Critical Rationalism), but while the conjunction you made does fail to ‘falsify’ the proposition A, it did succeed in falsifying (A · B).

Critical Rationalism broadly construed seems to me like a better research program than any other, if empirical research is to be done at all. People aren’t really looking for certain truths when it comes to empirical science, but bayesian relative certainty.

So it seems that your critique is that of being unable to experiment in a controlled setting, which certainly plagues falsificationism as much as anything else. With that we are in agreement.

3

I skimmed this a while back, YMMV: Debunking Popper: A Critique of Karl Popper’s Critical Rationalism

4

My example of (A · B) being the antecedent of the conditional is the most convenient possible world. In our world, the conjunctions can be made of hundreds or thousands of simple propositions (one or more of which are falsified by any one experiment). You basically understood my point, but I fear the simpleness of my illustration might have made things look easier than they really are. You can’t really falsify a theory through experiment. At best you will know that something is wrong somewhere in your “web of belief”. I do acknowledge that this is a valuable thing to know, but you shouldn’t overstate its usefulness. Personally, I think like a Critical Rationalist much of the time, because I think the degree of falsifiability a theory has is important. If theory A and B are equal in all or most respects except that one is more falsifiable, then I will spend more time testing the more falsifiable theory.

5

That’s a pretty good paper. It reminds me of the reason why Popper went in the modus tollens only direction in the first place. He basically accepted Hume’s view on induction (that it can’t be rationally justified), and so didn’t let induction play a role in what scientists were actually doing (to do so would be to admit that scientists acted irrationally in their work).

BTW, I’m a huge fan of Karl Popper’s work and would place him along Hayek in my top ten thinkers of the 20th century.

6

Right, I think we are in agreement. Thanks for your analysis, I don’t know what I was expecting but I was pleasantly surprised!

7

I read up to the critique of Hume’s problem of induction. I think the author is deconstructing a fiction, as the position that is critiqued was never Hume’s position. What did you think of it? I’ll give a defense of Hume if needed, but it will have to wait till tonight.

8

No.

If theory O is true then we will witness P → Q.

Experiment A demonstrates P and we get experimental result Q.

This adds statistical confidence that P → Q.

9

How do you witness a conditional proposition?

Perhaps you mean?:

  1. O ⊃ (P ⊃ Q)

  2. (P • Q)

∴ O

This is still a deductively invalid argument, but depending on the specifics can be a strong inductive argument (making it unacceptable to Popper, who wanted to banish induction from science).

10

My reading of that section is that it is a little oversimplified, but not contradictory to Hume’s actual argument. What do you think is wrong with his version of Hume? I am curious, because I read Hume so long ago that I may remember more of what people say Hume’s argument against induction was instead of Hume’s argument as he wrote it.

11

Are you guys trying to use this methodology in the social sciences?

12

I don’t think it makes sense to use different methodologies depending on if the science is physical or social. Perhaps the case can be made that different methodologies should be used for simple or complex phenomena, but even this seems rather ad hoc.

13

This is the Misesian argument for why teleological (prexeological) analysis is appropriate in the realm of the social sciences, whereas the traditional causal analysis is appropriate in the real of the physical sciences…

There are for man only two principles available for a mental grasp of reality, namely, those of teleology and causality. What cannot be brought under either of these categories is absolutely hidden to the human mind. An event not open to an interpretation by one of these two principles is for man inconceivable and mysterious. Change can be conceived as the outcome either of the operation of mechanistic causality or of purposeful behavior; for the human mind there is no third way available [9]. It is true, as has already been mentioned, that teleology can be viewed as a variety of causality. But the establishment of this fact does not annul the essential differences between the two categories.

The panmechanistic world view is committed to a methodological monism; it acknowledges only mechanistic causality because it attributes to it alone any cognitive value or at least a higher cognitive value than teleology. This is a metaphysical superstition. Both principles of cognition–causality and teleology–are, owing to the limitations of human reason, imperfect and do not convey ultimate knowledge. Causality leads to a regressus in infinitum which reason can never exhaust. Teleology is found wanting as soon as the question is raised of what moves the prime mover. Either method stops short at an ultimate given which cannot be analyzed and interpreted. Reasoning and scientific inquiry can never bring full ease of mind, apodictic certainty, and perfect cognition of all things. He who seeks [p. 26] this must apply to faith and try to quiet his conscience by embracing a creed or a metaphysical doctrine.

If we do not transcend the realm of reason and experience, we cannot help acknowledging that our fellow men act. We are not free to disregard this fact for the sake of a fashionable prepossession and an arbitrary opinion. Daily experience proves not only that the sole suitable method for studying the conditions of our nonhuman environment is provided by the category of causality; it proves no less convincingly that our fellow men are acting beings as we ourselves are. For the comprehension of action there is but one scheme of interpretation and analysis available, namely, that provided by the cognition and analysis of our own purposeful behavior.

The problem of the study and analysis of other people’s action is in no way connected with the problem of the existence of a soul or of an immortal soul. As far as the objections of empiricism, behaviorism, and positivism are directed against any variety of the soul-theory, they are of no avail for our problem. The question we have to deal with is whether it is possible to grasp human action intellectually if one refuses to comprehend it as meaningful and purposeful behavior aiming at the attainment of definite ends. Behaviorism and positivism want to apply the methods of the empirical natural sciences to the reality of human action. They interpret it as a response to stimuli. But these stimuli themselves are not open to description by the methods of the natural sciences. Every attempt to describe them must refer to the meaning which acting men attach to them. We may call the offering of a commodity for sale a “stimulus.” But what is essential in such an offer and distinguishes it from other offers cannot be described without entering into the meaning which the acting parties attribute to the situation. No dialectical artifice can spirit away the fact that man is driven by the aim to attain certain ends. It is this purposeful behavior–viz., action–that is the subject matter of our science. We cannot approach our subject if we disregard the meaning which acting man attaches to the situation, i.e., the given state of affairs, and to his own behavior with regard to this situation.

**It is not appropriate for the physicist to search for final causes because there is no indication that the events which are the subject matter of physics are to be interpreted as the outcome of actions of a being, aiming at ends in a human way. Nor is it appropriate for the praxeologist to disregard the operation of the acting being’s volition and intention; they are undoubtedly given facts. If he were to disregard it, he would cease to study human action. Very often–but not always–the events concerned can be investigated both [**p. 27] from the point of view of praxeology and from that of the natural sciences. But he who deals with the discharging of a firearm from the physical and chemical point of view is not a praxeologist. He neglects the very problems which the science of purposeful human behavior aims to clarify.

14

Oh, so you are? lol. That’s a real shame. If you read any of the below you’ll realise trying to apply the methodology of the natural sciences to that of the social sciences, is bunk.

The Mantle of Science by Murray Rothbard

Praxeology: The methodology of the Austrian School by Murray Rothbard

What is the Proper Way to Study Man? by Murray Rothbard

In Defense of “Extreme Apriorism” by Murray Rothbard

Toward a Reconstruction of Utility and Welfare Economics by Murray Rothbard

15

It’s bunk because Rothbard sez so? What a surprise.

At least there’s google for a larger variety of sources. This seems to be a nice initial find: http://www.cscs.umich.edu/~crshalizi/notebooks/social-science-methodology.html

16

No, who said that? It’s bunk because the arguments for it are…

I gave the best sources I know of that deal with this issue, and they all happen to be from Rothbard. Got a problem with that? Seems like it, as you took issue with it.

Care to point me in the direction of anything similar written by other austro-libertarians?

Have you read any of those, or do you just feel like throwing out links you found on google? What a surprise.

17

I’m am rereading that paper on Popper and Hume, and plan to expand my defense of Hume into a critique of the article. Not only does it oversimplify Hume’s argument but I feel confident that I can show that his whole argument via the law of identity being contradictory to Hume’s problem of induction is completely unwarranted and betrays a deep misunderstanding of Hume. I’ll probably have it posted by tonight or tomorrow.

18

I await your defense.

I won’t be able to think about these until after Thanksgiving. I’ll make a comment sometime after that.

19

I thought this was an excellent paper pertaining to the topic here.

What Empiricism Can’t Tell Us, and Rationalism Can by Mark Crovelli

20

If theory O is true we witness p->q, theory O does not cause p → q, it explains it. Theory O stands on no solid ground and is ready to be toppled as soon as what it predicts does not occur, or something occurs within its domain that it does not predict.