Sunday, December 26, 2010

Epistomology by Kent of AvC

This is in response to two recent posts: one by Observer and one by Simon. Now, since at least the beginning of recorded history people have been trying to prove or disprove the existence of God. The results can be summed in two words: They failed. (Simon's recent attempt fall into this category.) If you see a claim of a proof god does or does not exist you can safely ignore it.

One strain of this falls into what is called "Synthetic a priori knowledge", knowledge of the real world obtained by pure thought. This strain was pushed by people like Descartes and Kant.  Descartes got as far as giving an argument that can be used to prove "I exist" everything else may be  figment of my imagination (your use of I and my may vary). Kant's example of synthetic a priori knowledge was Euclidean geometry which ran into problems with general relativity. In general synthetic a priori information does not exist. This includes all proofs of gods existence or non-existence.

Various ways of getting empirical knowledge have been proposed: induction (failed by Hume's argument), verification (a la positivists), falsification and so on. They all have their critics. There is/are various claims that a scientific method does not exist. However, I suggest it does and follows much like Observer's observations.

We have two entities to deal with: observations and models. We make observations and organize them into models (theories, paradigms, world views, etc). Models can not be verified (the failure of induction) or falsified (Duhem-Quine hypotheses). However we can compare models. This is very much in line with what Popper actually said (not the caricature of his work we often encounter).

Knowledge thus consists of model building and testing against observation. We can not say if a model is correct or incorrect but we can say which of two models is better: the one with the best track record of successful predictions for observations. Observer said that the models are fictions. In this he is channeling Einstein who claimed the central ideas in any model are probably wrong (his examples were Newton's fixed space time and the coordinates of general relativity).

Now to the question of does God exist. Simple procedure: construct a model with God and one without God and see if adding God or god increases the ability to make correct predictions. Laplace and Hawkins claim: That hypothesis (god) is unnecessary.

In conclusion: a useful discussion of God's or god's existence must be based on the ability of models containing God to have more predictive power than those without. 

As the raven said: This and nothing more.

No comments: