And/or or both
April 19, 2008 @ 12:31 pm · Filed by Arnold Zwicky under Semantics
« previous post | next post »
Language Log brings little surprises into my life — this week, the report by Geoff Pullum that there are people who believe that the English conjunction or is always understood exclusively (so that and/or and or both might be useful expressions to have around, to convey an inclusive reading). Geoff countered with an argument that the meaning of or is in fact just inclusive disjunction, though the conjunction can be used to convey exclusivity via conversational implicature.
The usage literature largely agrees with this position, and so and/or is widely (but not universally) reviled, as unnecessary. But there's more to the story.
The complaints about and/or in the usage literature tend to focus on its perceived ugliness — an aesthetic judgment — rather than on semantics or even on communicative considerations like brevity (and/or is one word, and one syllable, longer than plain or), and by and large these criticisms aren't extended to or both (though X or Y or both is one word, and one syllable, longer than X and/or Y). MWDEU (p. 94) reports that most of the criticism in their files is aesthetic, "although a few consider it confusing or ambiguous, and a few (Johnson 1982, Shaw 1970, 1975, Reader's Digest 1983) find it compact and convenient." The usual advice is to use, where necessary, or both to force an inclusive reading (and but not both to force an exclusive reading).
Here's Bryan Garner (Garner's Modern American Usage, p. 45) on and/or, with a dubious claim about the way or is used:
A legal and business expression dating from the mid-19th century, and/or has been vilified for most of its life–and rightly so. To avoid ambiguity, don't use it. Many writers–especially lawyers–would be surprised at how easy and workable this solution is.
Or alone usually suffices. If you are offered coffee or tea, you may pick either (or, in this case, neither), or you may for whatever reason order both. This is the ordinary sense of the word, understood by everyone and universally accommodated by the simple or.
Garner's illustration is badly chosen. Disjunctive offers are frequently, and in some circumstances always, intended as providing exclusive alternatives. If you offer me coffee or tea and I ask for both, you'll be at least startled by my response, and you probably won't be satisfied by an explanation of the meaning of or from me. A disjunctive offer in a restaurant must be understood exclusively; if the menu says the main items come with soup or salad, you can get them both only by paying extra, and a lecture on semantics won't help things.
Perhaps it was such common situations that led some people to think that or is always exclusive. (Most people seem to take the common-sensical view that or sometimes functions inclusively and sometimes exclusively, and this is the position that the OED takes.)
In any case, writers sometimes want to make an inclusive reading explicit, and and/or is one way to do this; MWDEU notes that it occurs not only in "legal, commercial, technical, or bureaucratic contexts" but much more widely. Of course, or both is also available for this purpose, and will work for those who find and/or ugly.
A final note: or both presents an interesting problem for those who maintain that or is always exclusive. (The critique that follows is not original with me, though I can't now recall where I first heard it.) Start by considering how X or Y or Z is interpreted if or is always exclusive: exactly one of the disjuncts is true, both the others false. Now look at X or Y or both, that is, X or Y or (X and Y).
Suppose that X, the first disjunct, is true. Then Y and (X and Y) must both be false. No problem, since if Y is false, (X and Y) is also false. Parallel reasoning applies if Y, the second disjunct, is true: X and (X and Y) are both false.
But now suppose that it's the third disjunct, (X and Y), that is true. Then the two other disjuncts, X and Y, must both be false — but from that it follows that (X and Y) is false. Contradiction. The third disjunct cannot be true: if or is always exclusive, then X or Y or both must be understood exclusively; or both doesn't get you inclusive disjunction. Put another way: if or both is to allow for X and Y both being true, then the or of or both must be understood inclusively. So or cannot always be exclusive. (Just in case you were still hoping it could be.)
April 19, 2008 @ 12:31 pm · Filed by Arnold Zwicky under Semantics