SciAm article about rationality corresponding only weakly with IQ

post by DavidPlumpton · 2014-12-27T20:56:25.686Z · LW · GW · Legacy · 28 comments


Comments sorted by top scores.

comment by Gunnar_Zarncke · 2014-12-28T19:47:09.272Z · LW(p) · GW(p)

Maybe before you read the article you'd like to do the following test (mentioned in the article):

Jack is looking at Anne, but Anne is looking at George. Jack is married, but George is not. Is a married person looking at an unmarried person?


Replies from: ChristianKl, FeepingCreature, Unnamed, Luke_A_Somers, HungryHippo, bbleeker
comment by ChristianKl · 2014-12-29T21:04:36.373Z · LW(p) · GW(p)

The poll lacks an option for those people who have already read the article and want to see the results.

comment by FeepingCreature · 2014-12-29T06:13:43.709Z · LW(p) · GW(p)

Haha. The second I read the first sentence of that bit in the article I knew my mistake.

comment by Unnamed · 2014-12-30T00:36:02.430Z · LW(p) · GW(p)

This question was included on the 2012 LW Census/Survey.

You may want to take a guess at the results from that survey (perhaps after reading the SciAm article, but before following my link to the survey results). What percent of LWers who took the 2012 census/survey answered the question correctly? And, what percent answered it correctly among the subset of LWers who said that they had read "Nearly all of the Sequences"?

Answers: gur erfhyg jnf gung sbegl fvk creprag bs yrff jebatref tbg vg pbeerpg, vapyhqvat svsgl rvtug creprag bs gubfr jub'q ernq gur frdhraprf.

Replies from: satt
comment by satt · 2014-12-31T00:00:11.980Z · LW(p) · GW(p)

Uru, nccneragyl jr'er vzcebivat.

comment by Luke_A_Somers · 2014-12-30T01:13:01.965Z · LW(p) · GW(p)

The only two things that even slowed me down were, first, "This is supposed to be hard? Where's the catch?", and second, "Are they pointed the wrong way?"

comment by HungryHippo · 2014-12-29T19:48:25.991Z · LW(p) · GW(p)

Damn, I got it wrong.

Semi-spoiler below.

It did not occur to me to "check all cases". Had this been a math problem about the parity of numbers or some such, I would immediately think "well, A can be either even or odd. If A is odd, then ...; and if A is even then ...; QED".

However, my conscious thought process went more like "We can't tell whether Anne is married, since Jack does not have to be married to her if more than three people exist. We don't know who George is looking at, so the answer must be C."

For this problem you can also get the right answer by reasoning wrongly: "Jack must be married to Anne, so the answer is A."

comment by bbleeker · 2014-12-29T12:04:59.063Z · LW(p) · GW(p)

Dammit, I got it wrong. :(

Replies from: Gunnar_Zarncke
comment by Gunnar_Zarncke · 2014-12-29T16:32:38.681Z · LW(p) · GW(p)

But the fraction is still much better than in the study.

Replies from: bbleeker
comment by bbleeker · 2014-12-30T12:33:09.555Z · LW(p) · GW(p)

Yes, and I did get some of the other questions in the article right, and I'm 100% sure it's because of reading LW.

comment by buybuydandavis · 2014-12-27T21:34:37.048Z · LW(p) · GW(p)

Less intelligence can render you immune to a lot of the anti epistemology running around out there. A lot of very stupid ideas take some intelligence to consume.

I like the concept of cognitive miserliness, though I've thought of it as cognitive aversion.

While I'm cognitively compulsive, and expect most people to have a greater aversion to thought than I do, they do seem compulsive in their aversion, paying huge costs to avoid putting out even the most trivial cognitive effort. "No, I don't wanna think, and you can't make me!"

I'd note the guy's "rational analysis of the Jack->Anne->George problem left much to be desired. Just list out the options and test them. Notation matters.


Similarly, the bat and ball prices are trivial if you just write out the equation.

The way to be a cognitive miser is the use the right tools and notation. He might have demonstrated effective "mindware" in these problems.

The author also needs to work on his own rationality. The car example is just bad start to finish. You need a lot more information to even estimate net deaths from the car in question.

His gratuitous imposition of his own moral assumptions are worse.

We weigh evidence and make moral judgments with a myside bias that often leads to dysrationalia that is independent of measured intelligence.

Preferring your side is not necessarily dysrational. What rationality has to do with moral judgments is a non trivial topic.

Replies from: gjm, somervta, ChristianKl, the-citizen
comment by gjm · 2014-12-28T00:48:14.974Z · LW(p) · GW(p)

I think there's a better way to think about the Jack-Anne-George problem, which generalizes more readily. You've got a chain with "married" at one end and "unmarried" at the other: so of course at some point along it there has to be a transition from the former to the latter, QED.

Replies from: buybuydandavis, solipsist
comment by buybuydandavis · 2014-12-30T23:49:09.208Z · LW(p) · GW(p)

That is a very tidy analysis.

Easier than enumerate and evaluate, but much less general.

Replies from: gjm
comment by gjm · 2014-12-31T00:12:54.531Z · LW(p) · GW(p)

much less general.

It instantly gives you the answer to cases that look like A->B->C->D->...->Z where enumerate-and-evaluate requires you to consider 2^24 possibilities.

Let's think a moment about further generalization. So you have an arbitrary directed graph where a->b means a is looking at b; some vertices are coloured white ("married") and some black ("unmarried"), and the question is: is there a way to colour all the vertices black and white that has no instance of white->black?

Well, if there is any chain of arrows starting at a white vertex and ending with a black one, then the reasoning I described tells you that in any colouring there must be a white->black edge.

On the other hand, if there isn't then we can start at every white vertex, walking along arrows and whitening every vertex we reach; since there is no W->...->B chain this will never produce a clash. At this point we have whitened every vertex reachable from a white one, so now we can colour all the rest black; we've coloured the whole graph without any W->B edges.

So we have our (maximally generalized) answer: the answer to "must a married person be looking at an unmarried person?" is "yes, if there is a married->...->unmarried chain somewhere; no, otherwise". And (so it seems to me) this is just following the path of least resistance using the approach I described. So I'm going to stand by my claim that it "generalizes more readily".

comment by solipsist · 2014-12-29T05:09:30.118Z · LW(p) · GW(p)

Assuming the only states are married and unmarried. I'm not sure if I would call a widow unmarried, in the same way I'm not sure if I would call a man with a surgically reattached foreskin uncircumcised.

Replies from: gjm, buybuydandavis, ChristianKl
comment by gjm · 2014-12-29T15:08:35.254Z · LW(p) · GW(p)

Sure. (I think it's pretty obvious in the "puzzle" context that you're supposed to take "married" and "unmarried" as exhausting the possibilities, though.)

comment by buybuydandavis · 2014-12-30T21:16:07.423Z · LW(p) · GW(p)

I'd call a widow unmarried if she wasn't currently married.

I suppose the language usage might get complicated.

Is she still a widow, in the present tense, after she has remarried? Looking at a few definitions, it appears so, but the archtype of widow is one who has yet to remarry.

comment by ChristianKl · 2014-12-30T22:52:53.873Z · LW(p) · GW(p)

Marriages vows are "Till Death Do Us Part".

comment by somervta · 2014-12-29T00:34:28.666Z · LW(p) · GW(p)

The author also needs to work on his own rationality. The car example is just bad start to finish. You need a lot more information to even estimate net deaths from the car in question.

Which has nothing to do with the point being made.

Replies from: Jiro, buybuydandavis
comment by Jiro · 2014-12-29T03:37:50.431Z · LW(p) · GW(p)

Also, in the car example, the second version names the American car and says that it is 8 times as likely to injure another car's occupants as a typical family car. The readers would probably take this to mean "a typical American family car". Combined with the fact that the other car is named and that the reader knows that this named car is considered safe enough for American roads, this provides the information that a spread of 8 times is not dangerous. This information is absent when a German car is compared to an American car, especially when the German car is not named.

Replies from: buybuydandavis
comment by buybuydandavis · 2014-12-30T23:43:53.626Z · LW(p) · GW(p)

Showing that they were sloppy even in their sloppiness.

comment by buybuydandavis · 2014-12-29T01:26:49.676Z · LW(p) · GW(p)

He was trying to make a point about bias undermining rationality, and so was completely sloppy in asking a question that didn't have an answer determined by rationality.

Replies from: ChristianKl
comment by ChristianKl · 2014-12-30T22:41:32.322Z · LW(p) · GW(p)

The point isn't that the Yes/No answer is determined by rationality but that the nationality shouldn't matter. Do you think the nationality should matter?

Replies from: buybuydandavis
comment by buybuydandavis · 2014-12-30T23:34:52.062Z · LW(p) · GW(p)

The point isn't that the Yes/No answer is determined by rationality

So ask a question not determined by rationality, then complain that dispositions other than a will to rationality and ability to execute rationally determine the answer?

I don't see that as accomplishing the intended end here, or any end, but making the author look extremely intellectually lazy, at best.

Do you think the nationality should matter?

Sure. Government action should be dependent on the nationality of the parties involved, with a particularly relevant distinction between foreign and domestic.

Of course, my answer isn't determined by rationality, but context, preferences, and a theory of government, among other things.

comment by ChristianKl · 2014-12-30T22:34:16.524Z · LW(p) · GW(p)

Preferring your side is not necessarily dysrational. What rationality has to do with moral judgments is a non trivial topic.

If you fail to use Bayes rule properly because you are faced with a political question and using it would be a win for the other side, that dysrationalia.

comment by the-citizen · 2014-12-29T14:30:58.517Z · LW(p) · GW(p)

His gratuitous imposition of his own moral assumptions are worse.

I don't see the problem with moral assumptions, as long as they are clear and relevant. I think generally the myside effect is a force that stands against truth-seeking - I guess its a question of definition whether you consider that to be irrational or not. People that bend the truth to suit themselves distort the information that rational people use for decision making, so I think its relevant.

*The lack of an "unretract" feature is annoying.

Replies from: buybuydandavis
comment by buybuydandavis · 2014-12-29T21:02:52.253Z · LW(p) · GW(p)

There's not a problem if he is aware of the dependency on his moral premises and discloses that dependency to his readers. I don't see evidence of either.

The lack of an "unretract" feature is annoying.

Yeah. Interesting that in my inbox, it is not showing as retracted.

Replies from: the-citizen
comment by the-citizen · 2015-01-02T08:22:06.037Z · LW(p) · GW(p)

Yeah I think you're right on that one. Still, I like and share his moral assumption that my-side-ism is harmful because it distorts and is often opposed to the truth in communication.

I retracted an earlier incorrect assertion and then edited to make this one instead. Not sure how that works exactly...