Your facetious definition of what a universal Schelling point might be
sparked my interest to think about it further. Humor is often employed
in order to glaze over seemingly absurd, paradoxal, or complex ideas;
sort of like a defense-mechanism. If we laugh at something, make humor
of it, we no longer feel necessitated to explore it further. We can move
on without be troubled by it further. I'm not saying this was the case
in this situation, but after reflecting on just what might entail a
Schelling point, I found it to be fairly difficult to pin down. We can
all imagine Schelling points, e.g., if you lost your child at an
amusement park, they would probably turn up at either the entrance or an
information booth (assuming average intelligence), but how do both the
child and parent come to this conclusion if the subject was not brought
up beforehand? The difficulty here seems to lie in the fact that though
many points can be ruled out via reason alone, we would use our
intuition to make the final decision. But then, what is intuition? Seems
like another tool for glazing over a complex issue, sort like the word
'instinctual' or superinducing 'God' into an argument. Intuition in this
case is refering to the empathy we must have for the other individuals
involved. After ruling out the majority of points via reason, we must
imagine which of the remaining points would be most sensical to the
others while keeping in mind that they are probably thinking the same in
relation to us. This sounds a lot like chess to me. Seems to be one of
the fundamental issues that still separates a human player from a
computer. Computers have several ways of attempting to attain intuition,
e.g., probability, fuzzy logic, etc., but all of these methods seem to
lack variables for the human psyche. Even if a computer can learn
certain idiosyncracies about its human opponent, the computer can only
apply what it has learned to the specific "game" it is playing. If the
computer has no record of the player trying to find a Schelling point in
Paris, the computer will not be able to create an isomorphism [as
defined by Hofstadter] between the game it knows and the "game" of
finding a Schelling point in Paris based on what it has learned about
the human player. It seems however that humans do have this power, the
ability to apply learned patterns to new circumstances. I imagine the
Schelling point test would be an incredible way of testing an AI.
-- Gregory Houston vertigo@triberian.com 816.561.1524