Quote:
Originally Posted by well named
I had not seen any of this but I agree with you about the 0.5.
Quote:
I do think MB is correct that presumably, if the 0.5 is justified because we're proceding from maximum ignorance, which I think requires that we have no pre-conceived notion of what the word "God" refers to whatsoever, then it is possible that by adding to it we might actually increase the probability. Like if it turned out that "God" meant "The Mississippi River". When the blogger objects to P_before because he doesn't think P_before(Yahweh) should be greater than P_before(Zeus) I think he's missing what maximum ignorance actually means.
I am almost positive that to have the 0.5 starting point, both Yahweh and Zeus are contained within the original definition of "God". That is why any specification beyond a "conscious creator(s)" for God is narrowing and lowers the probability. If you add "the Mississippi River", you are no longer talking about God so the probability goes up but you have changed the question. That is unless the Mississippi River is actually the conscious creator hiding in the form of a river, in which case it is already in the 0.5. In other words, 0.5 probability of a God which includes a very small probability that God is hiding in the form of the Mighty Mississippi. Likewise FSM.
Quote:
But given the kinds of attributes we'd actually like to add to "God" so that it has something more to do with what we actually think we're talking about, it does seem reasonable enough to think that adding them will decrease the probability. In the same way that you are less likely to encounter a panda with a specific pattern of markings than just any panda.
Could not agree more.
Quote:
For MB, the very rough version of this is as simple as expanding from one completely undefined concept to the P(Yahweh) vs P(Zeus) vs P(...) where, even with "Zeus" and "Yahweh" minimally defined, and without actually talking about the probabilities of the attributes, now supposedly we go from P=1/2 to P=1/N (where N is the number of hypothetical deities). That is what RLK basically means when he says adding specification reduces the probability.
I agree except that P=1/N for N hypothetical deities assumes that each characterization is equally likely. That is probably not true.
Quote:
There seems to be some tension between the one process (specificity reduces P going from "panda" to "panda with marking pattern X") and the idea of proceeding from maximum ignorance though, unless we're capping the probability of God existing at being .5 max.
I think that the way I have defined things above, the probability is capped at 0.5. It could move up or down based on our observations of our reality.
Quote:
The first process should require a starting conception of God which is maximally broad for "God" but not complete ignorance, and there is not really any reason why the prior probability for the most generic conception of God should be 0.5, because even the most generic definition of God that is still specified enough to be reasonably named "God", is specified enough that we're not talking about complete ignorance.
I agree that we are not in complete ignorance. For one thing, if there is no God (using the broad Creator definition) than there might not have been a universe or self-aware beings occupying it. But that is built into the starting point of 0.5. It is like a question that was argued in this forum some time ago. There exists a card and it has a dot painted on it. The dot is painted either blue or red. You know absolutely nothing about how that color came to be. What is the probability of red? In that state of ignorance about the card, the probability is 0.5. It is not perfect ignorance because you know that there is a card. That gives the question of the color meaning. There is a universe so the question of a creator has meaning. If there was nothing created than the question of a creator becomes very hard to structure. But in that case we would not be having this conversation.
Quote:
But saying P_before=0.5 implies complete and absolute ignorance, but at that point it's no longer really like moving from "panda in the most general sense" to "specific panda".
I still believe that this is a good analogy. I hope I have clarified definitions to the point that you agree.