No, that's actually a bad application of logic. "The exception proves the rule" is used when an exception is presented, but on examination, is proven to indeed fit the rule, thus validating the rule. If a rule had to have an exception to be true, it cannot be a rule, because a rule, by definition, must not have exceptions. (it may, however, have nuance and in real life applications often does, but each of those rules which clarify the original rule do not, in fact have exceptions. They are simply a clearer statement of the rule with a fuller application of detail.). Generalization can, in fact, be a valid method of analysis in many situations, as you say. But the exception must also fit the rule for the rule to be true--if it doesn't, the only possibility is that the rule is not encompassing (that is, a bad generalization), or than the shoe does fit but someone isn't willing to admit it.