Thursday, February 17, 2005

More on Rules and Exceptions

In a previous post, I raised a question about the expression, 'The exception proves the rule.' How can an exception prove a rule? Let the rule be that all bloggers are less than 90 years of age. This rule is not proven, but refuted, by Geriatric Jerry who blogs his reminiscences of times past. All it takes to refute a universal generalization is one counterexample. Now suppose that the rule is that most bloggers are white males. This rule is not refuted by blogging females, but it is not supported by them either, let alone proven by them.

So I suggested that the expression means that the exception to a rule 'proves' the rule in the sense that it throws the rule into a relief, making it stand out as a rule.

Craig Howard of BuffaloG fame then suggested that 'prove' in this context means 'test.'

Commenter Alan then directed us to this page where Michael Quinion writes:

It has often been suggested in reference works that prove here is really being used in the sense of “test” (as it does in terms like “proving ground” or “the proof of the pudding is in the eating”, or in the printer’s proof, which is a test page run off to see that all is correct with the typesetting). It is said that the real idea behind the saying is that the presence of what looks like an exception tests whether a rule is really valid or not. If you can’t reconcile the supposed exception with the rule, there must indeed be something wrong with the rule. The expression is indeed used in this sense, but that’s not where it comes from or what it strictly means.

BV: Origin and meaning are distinct and may diverge. Meaning is tied to usage, so that one may wonder what sense there is in claiming that an expression "strictly means" such-and-such when it is not used to mean such-and-such.

The problem with that attempted explanation is that those putting it forward have picked on the wrong word to challenge. It’s not a false sense of proof that causes the problem, but exception. We think of it as meaning some case that doesn’t follow the rule, but the original sense was of someone or something that is granted permission not to follow a rule that otherwise applies. The true origin of the phrase lies in a medieval Latin legal principle: exceptio probat regulam in casibus non exceptis, which may be translated as “the exception confirms the rule in the cases not excepted”.

Let us say that you drive down a street somewhere and find a notice which says “Parking prohibited on Sundays”. You may reasonably infer from this that parking is allowed on the other six days of the week. A sign on a museum door which says “Entry free today” leads to the implication that entry is not free on other days (unless it’s a marketing ploy like the never-ending sales that some stores have, but let’s not get sidetracked). H W Fowler gave an example from his wartime experience: “Special leave is given for men to be out of barracks tonight until 11pm”, which implies a rule that in other cases men must be in barracks before that time. So, in its strict sense, the principle is arguing that the existence of an allowed exception to a rule reaffirms the existence of the rule.

Despite the number of reference books which carefully explain the origin and true meaning of the expression, it is unlikely that it will ever be restored to strict correctness. The usual rule in lexicography is that sayings progress towards corruption and decay, never the reverse. Unless this one proves to be an exception ...

BV: I would put Quinion's point as follows. There are two senses of 'rule.' In the first sense, a rule is an observed regularity, e.g., women are better than men in personal relationships. A rule in this sense is descriptive rather than prescriptive or proscriptive. In the second sense, a rule is prescriptive/proscriptive: it specifies something that ought to be done or ought to be left undone. To say that the exception proves the rule in the second sense of 'rule' is to say that the existence of an exception reaffirms the bindingness of the rule in the ordinary run of cases.

Question for a classicist: can the Latin regula, regulae be used in both senses lately distinguished, or only in one? I know it is used in the second sense by Descartes. But what about the first sense?

I think Quinion is basically right. Where he goes wrong, however, is in thinking that there is a "true meaning of an expression." There are original meanings, but why should these be the "true meanings"? And note that an original meaning of an expression came to be such because the expression was used in a certain way in a certain context. An expression does not have a meaning the way a rock has hardness. It is more like the way a rock can have different functions in different contexts: to conk someone on the head with, to serve as a paperweight, etc.

And note that Quinion's rule that "sayings progress toward corruption and decay, and not the reverse" the exception to which may be the saying in question, rests on an equivocation on 'rule' that I have just exposed with his help. That is, Quinion is himself not using 'The exception proves the rule' in its "true meaning"!


World Wide Words is copyright © Michael Quinion, 1996–2005. All rights reserved. Contact the author for reproduction requests.Comments and feedback are always welcome.Page created 14 September 2002.