Opiniatrety

The Digital Underground line and (6) both strike me as OK, because they seem posed so as to be obvious word play.

Note that 'OK' is meant to be grammatical, rather than aesthetic. (6) strikes me as dumb word play, whereas the Digital Underground strikes me as clever.

Yeah, there's a big difference between (6) and (5)--(5) is just hard to parse. (1)-(4) aren't as hard to parse in that way, but do they come across as punny like (6)?

They're not punny at all. (4) sounds weird, but (1) - (3) all seem just fine. I think I've uttered nearly identical sentences.

If you're interested, I can try to remember what they were, if I haven't had too much wine to think straight.

(4) scans fine to me. (2) is almost as good. (1) and (3) strike me as odd for reasons I can't quite figure out atm. I'll get back to you if I figure it out.

Good one, Geoff. That definitely scans fine. My guess is that it doesn't cause any trouble for the view that the two uses of "if" have different meanings, since the two "if" clauses are separated, but I'd have to ask a syntactician to make sure.

I'm no semanticist and I'm definitely no syntactician. But it seems to me that the separation of the 'if'-clauses in my example shouldn't make a difference. Consider:

Joey thinks the greatest theater was La Scala, the greatest library was at Alexandria, and the second greatest theater was the Pacific in WWII.

Besides being a sentence nobody would utter, I think it's semantically unacceptable, even though the 'theater' clauses are separated. I would expect that, if 'if' had two meanings, the separated 'if' clauses would sound just as bad as those. If you care what I think.

Two remarks:

For a good account of the relationship between the various senses (or, more accurately, semantic extensions) of the conditional, see Eve Sweetser's book From Etymology to Pragmatics. It provides a convincing account of the unifying conceptual structures underlying not only these different uses of the conditional, but also modal verbs and conjunctions.

As for the linking together of truly distinct senses with ellipsis, I refer you to the rhetorical figure of zeugma (in some taxonomies zeugma is a more general term and syllepsis is the name for this particular variant). The question for some of your test sentences isn't necessarily whether or not they're unacceptable, as such, but whether they result in zeugma.

Zeugma is usually a kind of pun, but it doesn't necessarily equal semantic unacceptability, and it can be hard to make a clear dividing line between zeugmatic and non-zeugmatic sentences:

(1) She was wearing nothing but high heels and a smile.
(2) He lost his keys and his temper, in short succession.
(3) In the background you could hear reports from the newscasters and from the guns.
(4) The dogs are ugly and barking up the wrong tree.

Vera--Oh, someone's answered these questions? Phoo! Sweetser's book is checked out of our library, unfortunately.

The possibility of zeugma seems like it could raise trouble for the whole style of argument in which we try to detect ambiguity by testing sentences. We could try to restore it by consulting our intuitions about whether sentences are zeugmatic/funny-sounding, but those intuitions seem like they'll be a lot more suspect than most (I know many philosophers working in this area disparage semantic intuitions). Note that there's a big difference between my (5) and (6); (6) is a bad pun, (5) is just difficult to make sense of.

Giving rein to my intuitions, I'd order your sentences, most to least zeugmatic, 3-2-1-4. (4) doesn't strike me as zeugmatic at all; is it that it pairs an adjective and a verb?

Geoff--I agree about that example; it's semantically funny. I guess my feeling is that a syntactician might say that in the double-"if" case the two "if"s aren't coordinated; the form is (If A (If B (C))) rather than (If A and If B)(C). But, as I said, I ain't one.