Saturday, January 12, 2013

Justificationism and Critical Rationalism


Justificationism says that positive arguments can make a theory true (or more probably true).

Some justificationists are also Bayesians. Bayesianism says that better (or more probable) truths can be calculated using arbitrarily-assigned initial values representing the "weight" of a positive argument. [1]

Critical Rationalism (aka Popperism)

Critical Rationalism says that all knowledge is created by (1) guesses and (2) criticism:

(1) A guess is a new theory. Another name for this is a positive argument. Note that if the theory can be criticized using physical evidence, then we call it a scientific theory.

(2) A criticism is also a guess. It rules out a theory, i.e. make it false. Another name for this is a negative argument. Note that if the negative argument uses physical evidence, then we call it a scientific experiment.

In other words, a theory is (conjecturally) true so long as there is no negative argument acting against it.

A consequence of this is that Justificationism is wrong, since positive arguments cannot make a theory true. All a positive argument can do is propose a new theory. And a theory is true only so long as no negative argument is acting against it.

Critical Rationalism also says that:

Truth is objective. This means that truth exists independent of what people think about what the truth is. This is called Objective Knowledge.

People are fallible. This means that people cannot know which of their ideas are objectively true -- any one of their ideas could be wrong. What we do have is conjectural knowledge (aka fallible knowledge). A consequence of this is that people do not have access to infallible sources of knowledge (or more trust-worthy sources of knowledge), like intuition, emotion, justifications, induction, etc.

So human knowledge is created by guessing ideas and ruling them out with criticism.

In short-hand philosophy-speak: (1) guesses, and (2) criticism.

In epistemology-speak: (1) a positive argument creates a new idea and by default its truth-value is set to true (aka +1), and (2) a negative argument has the same qualities as positive arguments and it has one extra quality which is that it acts against another idea, thus immediately (and without hesitation) setting the target idea's truth-value to false (aka 0).

In morality, which is the body of knowledge concerning how people should make choices, aka decision-making/choice-theory, we do the same thing. We guess ideas and criticism them. Some commonly known moral ideas that have passed the test of criticism are *murder is evil* and *rape is evil*. By "evil" I mean that it shouldn't-be-done/immoral/bad/hurtful/harmful/non-mutually-beneficial.

In science, there is an extra component which is physical evidence. Physical evidence is used as part of a negative argument. Its important to note a typical justificationist mistake which is to think that physical evidence can be used as part of a positive argument. This is actually a special case of the more general mistake of thinking that positive arguments can be used to act in support of other ideas, thereby adding truth-value to the target idea, giving it a truth-value great than +1, which is impossible.

The truth-value of an idea is either true or false. Its boolean. 1 or 0. Yes or no. There are no other meaningful options.



The question that Bayesianism can't answer, which Popperism does answer, is this: If I have 3 rival theories, and they don't refute each other, and if I use Bayesian math to figure out that T1 is 90% likely to be true, and T2 is 9% likely to be true, and T3 is 1% likely to be true, then how does a person act on this? Which theory does he enact? T1 because its got the most chance of happening?

Also this raises a question. Why did T1 get a higher score? By what criteria was that judged? How do we know that criteria is right? Did you submit that to external criticism?

Consider how Bayesian math involves initial values and then running that through an algorithm that updates the values while also producing probabilities. Do you know how those initial values are chosen? Arbitrarily. That means it doesn't correspond with reality.

So think about that for a moment. What does it mean for those initial values to be arbitrary? It means that its mysticism. It means that there is no way to test those values against reality because it doesn't correspond with reality. It means that there is error and that the error is not being corrected. And that error in the system (referring to the Bayesian algorithm and the data too) will grow without limit. So its useless.


Join the discussion group or email comments to

No comments:

Post a Comment