You win the prize! You have hit on the reason "fuzzy logic" is not
logic.
You can attempt to rescue the situation by using the "rule for rules"
that when two rules conflict, you should use the one that is more
specific. In this case "high(a) AND med(b)" is more specific. If
however, you have two rules that conflict, and neither predicate is a
subset of the other, you have a problem. You need a rule that applies
to the intersection (conjunction) of the two predicates, and there is
no reasonable way to synthesize one. May I recommend Judea Pearl's
awe-inspiring _Probabilistic Reasoning in Intellegent Systems: Networks
of Plausible Inference_?
Jive
------------------------------