This entire article is hilarious

[whyBish]

Now, *every* system starts out with assumptions.  If working with one subset of those assumptions, we conclude that the probability of a given statement is one value, and working from a different subset of those same assumptions we conclude that the probability of that same statement is some different value, then we know that the assumptions are contradictory.
--Pete
[right][snapback]85509[/snapback][/right]

Agreed.

But what if your
- assumptions have a probability distribution of their probability of correctness
- rules have a probability distribution of their probability of correctness
hence results are not a value but a probability distribution. (This is what I was calling the fuzzy logic one, but obviously it is something totally different...)

Then two 'different' results will only have some level of confidence that they are inconsistent...

Anyway, I've missed my point, which was that people usually:
- don't have consistent assumptions (this pretty much has to be true if you look into any set of assumptions enough, particularly around morality)
- don't usually evaluate them
- don't usually go looking for evidence that they are wrong
- have some beliefs that they hold more strongly than others
- can/do change their minds (learn?) particularly about things that they don't hold strongly, or they are aware of contradictions that this would resolve for their system.