November 03, 2006

Failure of Closure...

...of something or other. Chris Bowers projects that a Democrat will win the Senate race in Montana, and that a Democrat will win the Senate race in Missouri, and that a Democrat will win the Senate race in Virginia, but not that Democrats will win the Senate races in Missouri, Montana, and Virginia. Which (given his background assumptions) seems entirely sensible. If he had the Democrats at odds of, say, .8, .7, and .6 in each race, then there would only be a .336 chance that they would win all three. It's routine that the property of having a probability greater than p isn't deductively closed.

The interesting question is what Bowers's attitude toward the propositions is. Has he asserted them? I'm inclined to say that this sort of projection is a weak form of assertion, in which case willingness to assert doesn't seem to be closed under known entailment. Does he believe them? Well, if you say "What does Chris Bowers think will happen in the Missouri Senate race?" the answer would be "The Democrat will win," so there's at least a little evidence for saying he believes them. In which case belief isn't closed under known entailment. But there's also a lot of use for a more robust notion of belief on which Bowers wouldn't count as believing any of these propositions, but only that they're more likely than not. Interestingly, on Mark Kaplan's conception of belief Bowers might count on believing them all and their conjunction to boot; each proposition is part of the story he tells, as is their conjunction, even if he believes that the conjunction is less likely than not.

[In other news, I disagree with Bowers about TX-22 as a toss-up. I just voted, and it would be extraordinarily patience-taxing to write in "Shelley Sekula-Gibbs" on one of those machines. Don't see it happening. Also, TPM reader ZL sings a familiar tune.]

Posted by Matt Weiner at November 3, 2006 02:26 PM
Comments

I was just going to ask if you'd seen that TPM post.

Posted by: teofilo at November 3, 2006 03:08 PM

I had to think about how to put it, since I'm a little reluctant to put "Weiner-pwned" on the front page. (Actually, technically I"m Farbing.)

Posted by: Matt Weiner at November 3, 2006 03:59 PM

Is the distinction between Weiner-pwning and Farbing just the time between postings?

Posted by: teofilo at November 3, 2006 05:57 PM

Is this ad more entertaining if one knows that the President of the United States was just in Montana claiming that Sen. Tester would raise taxes?

Posted by: at November 3, 2006 10:27 PM

Is the distinction between Weiner-pwning and Farbing just the time between postings?

You're making a category error -- the initial post is the Weiner-pwn, while the Farb is the announcement that you posted about this a couple of months ago. But yeah, I think the distinction between announcing a Weiner-pwn and a Farb is in time and proximity of posts.

In particular, the classic and original Weiner-pwn took place when the first poster commented while the second poster was typing. Also, the farther away the first comment is from the second (for instance, on another blog, especially one that the second poster may not always read) the closer the announcement is to a Farb. Note also Standpipe's use of retropwned.

No-name, I have to say I don't find it that amusing even with that knowledge, though I support it. There's kind of a frustrating Kabuki dance here; even after perhaps the most irresponsible tax cuts in the country's history, we can't say that we'll raise taxes to bring the country back to some semblance of fiscal sanity. But ah well.

Posted by: Matt Weiner at November 4, 2006 10:14 AM

Sorry about the no name thing -- computer error. Or more likely operator error.

Calling the President one of John Cornyn's 'crooked pals in Washington' would be a big damn deal in Lubbock, wouldn't it?

Posted by: CharleyCarp at November 4, 2006 03:13 PM

I just voted, and it would be extraordinarily patience-taxing to write in "Shelley Sekula-Gibbs" on one of those machines.

Does the whole state use the same machines?

Posted by: teofilo at November 4, 2006 05:14 PM

Charley, probably (Hutchison is actually the one who's up for reelection), but Lubbock is waaaaay to one extreme here. Second most pro-Bush city in the 2004 election. I don't think we could get a popular local Democratic politician to run such an ad; in fact I'm not positive that there are any Democrats in elected office here, up to the statewide level. Maybe a judge or two.

Teofilo, I don't think so, but the machines I used seemed to fit the descriptions of voting machines in the TX-22 race.

Posted by: Matt Weiner at November 4, 2006 06:48 PM

The Lubbock Democratic party has a pretty slick website (as does the state party), but it doesn't look like they have anyone in office right now; all the people on the "candidates" page seem to be challengers.

Posted by: teofilo at November 4, 2006 08:41 PM

It keeps making tinkly noises at me!

Posted by: Matt Weiner at November 4, 2006 09:56 PM

That's the sign of a really high-class political party.

Posted by: teofilo at November 4, 2006 10:04 PM

You don't have to punch in all of "Shelley Sekula-Gibbs" - the machine's got auto-completion built in, so just start tapping and it knows what you mean to say. In fact, no matter what button you push, it enters "Shelley Sekula-Gibbs." It's very advanced.

Your Texas Pal,
Karl R.

Posted by: Ben at November 5, 2006 06:33 AM

Bowers now seems to be actively asserting a bunch of propositions and denying their conjunction; if we take "she really should have it now" and "This should be good enough for Webb to win" as assertions that McCaskill and Webb should win, which seems not implausible. Well, maybe it's not outright assertion.

Posted by: Matt Weiner at November 6, 2006 09:22 AM