“[T]he winner might actually represent the story which has actually alienated the least number of voters, rather than excited them,”

This is something Instant Run-off Voting does. In an ideal situation, it drives to the most commonly popular candidate. A candidate who is offensive to a majority of the electorate will be driven out. In a less-than-ideal situation, it will drive to a candidate the the electorate can at least be satisfied with, if not thrilled with.

In modern politics, the last Oakland mayoral election displayed this. The ostensible front-runner was swept under by a tide of second preferences; the folks voting for other candidates wanted anybody but Don Perata and they voted their preferences accordingly. There’s indication that the current San Francisco election has the same kind of groundswell against Lee, and a similar result is expected.

In the 2011 Hugos, there was a groundswell against Randall Munroe in the Fan Artist category (his subject matter is fannish, but it’s exploited commercially and makes him his living) and against StarShipSofa in the Fanzine category (even SSS listeners have said that the subject matter isn’t as fannish as it used to be, but many more voters seem to agree the presentation is fundamentally different than a print periodical).

the award might be given to someone who already has prestige as a re-invocation of respect for them

This is an “the electorate is made up of idiots” argument, but it’s one that I strongly support. It’s been made numerous times. Whether it’s a mediocre work by a great (and popular) writer or an artist who hadn’t produced (or published/sold depending on the rules at the time) any new work in the year of eligibility, I’m disappointed in the electorate voting for a comfortable relationship on those merits alone. But it does happen sometimes.

The awards become a structure for conversations about the literature and overshadow the discourse on both literary and social subjects.

We’re too tied up in the latter. The former is ignored or discounted, particularly the nominations process and shortlist driving reading and discussion. Mike Glyer had a great post at File770 about the perennial “I’ve never heard of most of these nominees” gripe. You can, as the pessimists do, discount the nomination process as being out of touch with everything that’s important in the world. Or you can, as the optimists do, look at the short list and say “Neat. New things to read. I wonder what the nominators saw in these works.”

As James English pointed out in his work, there is also a paradox in selecting works soon after publication and labeling them with an unqualified designation of excellence, but this too may be necessary for many awards because of the need for temporally-ritualized social reproduction.

The Worldcon is going to select some timely and topical works that, under the glare of history and reflection, become horribly dated. It might be amusing to do a “Hugos: 20 years later” cycle, look at the long list (every nomination), re-nominate and re-vote for what we now think was the best work of the year. I expect we would see a lot of “Well, the original winner was good, but it wasn’t the outstanding classic of the year.”

Of course, I hear a Hugo administrator screaming at me about the associated workload…

And that dovetailed nicely with my point in your last column, John. I agree, its a conundrum.