Tuesday, 22 February 2011

Statistics!

Say someone flips a coin and it keeps coming up heads ... how long before you will have to concede that they are using a double-headed coin? It seems most of us will give up after about five or so ... you'd certainly want to grab the coin of that person and take a closer look.

But how reasonable is this? If you only have the run of heads to go on, don't you have to give the person the benefit of the doubt? What if there's money riding on it? You life? Your immortal soul?

I decided to work it out.

This is actually related to the other stuff about religion: we commonly hear that Atheists cannot be sure of the nonexistence of God since we don't know everything. Or variations on that theme. Just because we have disproved a whole bunch of stuff as having to do with God does not mean that God is not true right? Can you really gain proof from negative results?

This is also related to the problem of induction. At stake is the long-term viability of scientific investigation. Science assumes that induction via falsifiability can be used to discover general rules about the way the Universe works. To put it in a theist context: God rarely intervenes in human affairs.

No comments:

Post a Comment