Logical hygiene, or keeping the brane clean
There's a reason we spend a bit of time in 4400-nee-110 going over some of the widely known errors in reasoning. Actually, there are several reasons:
* So you can be the life of the party with cool phrases like "post hoc, ergo propter hoc."
* So you don't say "begs the question" (Vox, p. 7) when you mean "raises the question." Begging the question would be supporting an assertion with another unsupported assertion, as in: "This is Fiona Apple's most adventurous album because she takes chances she's never taken before."
* And, of course, the real reason, which is to exercise good traffic-copping over what gets into the newspaper. From the editor's perspective, this comes in two main flavors, so pay attention.
1) Faulty reasoning (accidental or deliberate, it doesn't really matter) on the part of the source. Unfortunately, we don't get to put big red stickers saying NON SEQUITUR or ARGUMENTUM AD MISERICORDIAM on offending passages. But we might think about moving a contrasting paragraph higher in the story, asking the originating desk whether it sought out a rebuttal, suggesting that a passage be deleted or even asking to hold a story for a day or two in the interest of getting an appropriate comment.
2) Faulty reasoning on our own part. This too calls for consultation, but it's even more important to raise the flag; people who see it in the paper can't tell if we just don't know any better or if we're stacking the deck. Herewith a case in point, from Wednesday's 1A piece on traffic-stop statistics:
"The statistics also showed that on average, all blacks age 16 and older had a 60 percent of being stopped for a violation."
The statistics don't show that, for a couple of reasons. First, as the chief suggests, the stats measure total traffic stops against the population. They don't tally stops of nonresidents or account for the possibility anyone's being stopped more than once. That's not to say the overall percentage is inaccurate, it's to say we have no way of knowing whether it's accurate. If it's correct, it's correct only by accident -- not by "statistics."
Of equal concern, though, is the logical fallacy of turning this stat -- I think it's (traffic stops/black residents)(100), but this machine won't let me open two windows at the same time -- into a generalized percentage. It's called the fallacy of division: assuming that what's true of the whole is true in equivalent proportions of every part of the whole. If you're an undergraduate J-student reading this, you don't have a 60 percent chance of being female. Your chances are either 0 or 100 percent.
Circumstances beyond accidents of birth are likely to have an effect too. If you peacefully finish your shift at the Missourian, obey all the traffic laws while driving home, pour a glass of lemonade and settle in to memorize the stylebook, you have a pretty low chance of being arrested. On the other hand, if you get drunk, trash the place, ram a police cruiser on the way out of the parking lot, squeeze off a few rounds from a stolen handgun and and impugn the responding officers' heritage, your chances of free room and board for the evening go up substantially. You can't add those two chances, divide by two and get a number generalizable to the undergraduate J-population.
None of this is to say that the disparities in traffic-stop numbers aren't interesting and important. They are. It's a serious issue worth serious attention. But we're going to have a hard time reporting on that issue properly if we can't distinguish between "statistics say" and "Missourian says." And when we put a thumb on the metaphorical scales, we're not just suggesting that we have different rules for causes we like and causes we don't. We're giving everybody who doesn't believe there's racial profiling a chance to say: "See? All the reports on this are distorted" -- a Rush Limbaugh trick of the first order.
Speaking of whom, ff you want to sharpen your fallacy-hunting skills, tune Limbaugh in. In the brief time it takes to get from the Stewart-Providence intersection to the Hitt Street deck, he indulged in:
* Argument ad hominem: Gleefully quoting Ben Stein's idiotic observation that Mark Felt "looks like one of those old Nazi war criminals they find in Bolivia or Paraguay."
* Tu quoque: Lefties think Nixon was awful, but look at what Clinton did!
* Fallacy of composition: Whatever Dan Rather says is what The Meedja does (the singular usage is Rush's).
There might well have been more, but, y'know, keep your mind on your driving....
* So you can be the life of the party with cool phrases like "post hoc, ergo propter hoc."
* So you don't say "begs the question" (Vox, p. 7) when you mean "raises the question." Begging the question would be supporting an assertion with another unsupported assertion, as in: "This is Fiona Apple's most adventurous album because she takes chances she's never taken before."
* And, of course, the real reason, which is to exercise good traffic-copping over what gets into the newspaper. From the editor's perspective, this comes in two main flavors, so pay attention.
1) Faulty reasoning (accidental or deliberate, it doesn't really matter) on the part of the source. Unfortunately, we don't get to put big red stickers saying NON SEQUITUR or ARGUMENTUM AD MISERICORDIAM on offending passages. But we might think about moving a contrasting paragraph higher in the story, asking the originating desk whether it sought out a rebuttal, suggesting that a passage be deleted or even asking to hold a story for a day or two in the interest of getting an appropriate comment.
2) Faulty reasoning on our own part. This too calls for consultation, but it's even more important to raise the flag; people who see it in the paper can't tell if we just don't know any better or if we're stacking the deck. Herewith a case in point, from Wednesday's 1A piece on traffic-stop statistics:
"The statistics also showed that on average, all blacks age 16 and older had a 60 percent of being stopped for a violation."
The statistics don't show that, for a couple of reasons. First, as the chief suggests, the stats measure total traffic stops against the population. They don't tally stops of nonresidents or account for the possibility anyone's being stopped more than once. That's not to say the overall percentage is inaccurate, it's to say we have no way of knowing whether it's accurate. If it's correct, it's correct only by accident -- not by "statistics."
Of equal concern, though, is the logical fallacy of turning this stat -- I think it's (traffic stops/black residents)(100), but this machine won't let me open two windows at the same time -- into a generalized percentage. It's called the fallacy of division: assuming that what's true of the whole is true in equivalent proportions of every part of the whole. If you're an undergraduate J-student reading this, you don't have a 60 percent chance of being female. Your chances are either 0 or 100 percent.
Circumstances beyond accidents of birth are likely to have an effect too. If you peacefully finish your shift at the Missourian, obey all the traffic laws while driving home, pour a glass of lemonade and settle in to memorize the stylebook, you have a pretty low chance of being arrested. On the other hand, if you get drunk, trash the place, ram a police cruiser on the way out of the parking lot, squeeze off a few rounds from a stolen handgun and and impugn the responding officers' heritage, your chances of free room and board for the evening go up substantially. You can't add those two chances, divide by two and get a number generalizable to the undergraduate J-population.
None of this is to say that the disparities in traffic-stop numbers aren't interesting and important. They are. It's a serious issue worth serious attention. But we're going to have a hard time reporting on that issue properly if we can't distinguish between "statistics say" and "Missourian says." And when we put a thumb on the metaphorical scales, we're not just suggesting that we have different rules for causes we like and causes we don't. We're giving everybody who doesn't believe there's racial profiling a chance to say: "See? All the reports on this are distorted" -- a Rush Limbaugh trick of the first order.
Speaking of whom, ff you want to sharpen your fallacy-hunting skills, tune Limbaugh in. In the brief time it takes to get from the Stewart-Providence intersection to the Hitt Street deck, he indulged in:
* Argument ad hominem: Gleefully quoting Ben Stein's idiotic observation that Mark Felt "looks like one of those old Nazi war criminals they find in Bolivia or Paraguay."
* Tu quoque: Lefties think Nixon was awful, but look at what Clinton did!
* Fallacy of composition: Whatever Dan Rather says is what The Meedja does (the singular usage is Rush's).
There might well have been more, but, y'know, keep your mind on your driving....
posted by fev at 5:24 PM
1 Comments:
I think we should just start saying "petitio principi" instead of "begging the question". It's a crappy translation, no one who's never studied logic can guess what it means - due to the huge number of English phrases with "beg" in them that do mean what people think they mean - and we seem happy to use, let's see here, Argument ad hominem, Tu quoque, non sequitur, argumentum ad misericordiam, and post hoc, ergo propter hoc.
Just my little suggestion to help end some confusion.
Post a Comment
<< Home