Tags:

Share:
Facebook

The Argumentative Theory of Reasoning
May 5, 2011 9:08 AM Subscribe

Reasoning was not designed to pursue the truth. Reasoning was designed to help us win arguments.

The Argumentative Theory of Reasoning ... maintains that there is an asymmetry between the production of arguments, which involves an intrinsic bias in favour of the opinions or decisions of the arguer whatever their soundness, and the evaluation of arguments, which aims at differentiating good arguments from bad ones and thereby genuine information from misinformation. This asymmetry is often obscured in a debate situation (or in a situation where a debate is anticipated). People who have an opinion to defend don't really evaluate the arguments of their interlocutors in search for genuine information but rather consider them from the start as counter-arguments to be rebuked. Still, as shown by the evidence reviewed in section 2, people are good at assessing arguments, and are quite able to do so in an unbiased way, provided they don't have a particular axe to grind. In group reasoning experiments where participants share an interest in discovering the right answer, it has been shown that truth wins.

The paper: "Why Do Humans Reason? Arguments for an Argumentative Theory," Hugo Mercier and Dan Sperber.

A conversation with Mercier.

posted by AceRock (60 comments total) 38 users marked this as a favorite

No, it wasn't.
posted by hal9k at 9:11 AM on May 5, 2011 [5 favorites]

no, YOU shut up
posted by dismas at 9:12 AM on May 5, 2011 [1 favorite]

(AceRock: This looks really interesting! Thanks!)
posted by dismas at 9:13 AM on May 5, 2011

I'll play devil's advocate here and just agree with the entire premise of this article.
posted by ShutterBun at 9:15 AM on May 5, 2011 [11 favorites]

Sorry, time's up. I'm afraid you'll have to pay for another session if you'd like to continue.
posted by likeso at 9:19 AM on May 5, 2011 [5 favorites]

The one thing I don't understand is why it's upsetting that reason doesn't help us find the truth. It helps us explain our findings. Doesn't that mean our findings must be reasonable in order to be explained?

For example. Christianity doesn't make sense to most non-Christians. But once you accept its initial premise, it makes sense as a system based on that premise. The problem for non-Christians isn't the internal logic...it's the external jump one has to make to get to a place where it might be believed that the sacrificial death of a man 2000 years ago has anything to do with one's life. Which isn't reasonable, never has been reasonable, and isn't portrayed by Christianity itself as reasonable. Christianity does not advertise itself as a religion you can get to solely by reason, or for the sake of reason, or for any kind of scientific proof of anything at all.

But when you start with the premise that He did sacrificially die and it did mean something for the rest of humanity, all kinds of things that didn't make sense before suddenly matter and get sorted out. This is the kind of thing this article seems to be talking about. Reason won't help you get to the truth, but it will help you explain it once you're there. Which doesn't bother me at all...so...help me out. Where is the gap?
posted by FunkyStar at 9:28 AM on May 5, 2011 [3 favorites]

I'd be really interested to know what Harry Reasoner would've had to say about all this.
posted by flapjax at midnite at 9:31 AM on May 5, 2011

So remember back when Plato made a big deal about the difference between rhetoric and philosophy?
posted by oddman at 9:32 AM on May 5, 2011 [13 favorites]

I think it's just that reason is a tool, like a sword or a gun. Having a bigger, better weapon doesn't actually make the person wielding it any stronger, nor does it inherently validate their cause -- but it can still help them win.
posted by hermitosis at 9:34 AM on May 5, 2011 [2 favorites]

Your favorite theory of the mind sucks.
posted by Pirate-Bartender-Zombie-Monkey at 9:35 AM on May 5, 2011 [2 favorites]

Sorry, time's up. I'm afraid you'll have to pay for another session if you'd like to continue.

That was never five minutes just now!!
posted by Mister Fabulous at 9:36 AM on May 5, 2011 [2 favorites]

Reasoning was designed?
posted by DU at 9:36 AM on May 5, 2011 [2 favorites]

Seems like confirmation bias on the part of stats researchers.
posted by GuyZero at 9:42 AM on May 5, 2011 [1 favorite]

And you know what's really freaky? Laws aren't written to decide what's right or wrong - laws are written to end disagreements.
posted by Curious Artificer at 9:42 AM on May 5, 2011 [6 favorites]

Let's say that this is an accurate ~~argument~~ theory. Even then, it's relevant in some situations more than others. A novel academic or scientific theory may develop along these lines and turn out to be useful or productive, for instance; many won't, and that's fine too. I think the situation at which the developers of this theory are looking is an informal one, and I don't really see how the pursuit of Truth especially deserves to be foregrounded in such situations.
posted by clockzero at 9:45 AM on May 5, 2011 [1 favorite]

The one thing I don't understand is why it's upsetting that reason doesn't help us find the truth. It helps us explain our findings.

More than that -- reasoning is how we explain our own motivations to ourselves after our brain has already made a decision. We're good at using reasoning to mislead others, because we first use reason to mislead ourselves.
posted by empath at 9:45 AM on May 5, 2011 [8 favorites]

So remember back when Plato made a big deal about the difference between rhetoric and philosophy?

Exactly. Probably the most important reason they killed Socrates was that he was "making the weaker argument the stronger", i.e. that he was teaching people to win arguments they had no business winning on the facts, which is something that sounds like lots of fun until you use your brilliant arguments to destroy a democratic government.

This is old, old stuff they're talking about.
posted by Copronymus at 9:46 AM on May 5, 2011 [12 favorites]

Btw-- bookmark this article, because it will come in handy in future internet arguments as a way to attack the sincerity of the other opinion.
posted by empath at 9:47 AM on May 5, 2011 [1 favorite]

@empath: totally agree; though I do think we should not mislead others, I believe we can't really help ourselves.
posted by FunkyStar at 9:47 AM on May 5, 2011

Wow, good point, Copronymus.
posted by clockzero at 9:52 AM on May 5, 2011

The author's "asymmetry" of reasoning looks more like an inconsistency in their account.

Reasoning on the one hand is designed to get at the truth (reliable communication) and weed out falsehoods (misinformation).

On the other it ignores information, isn't interested in avoiding bias, and is mostly dedicated to indiscriminately rebuking arguments.

... sounds inconsistent to me, but I was pretty primed to refute it's misinformed ass.
posted by airing nerdy laundry at 9:54 AM on May 5, 2011

Falsity refers to errors of a formal logical nature in linking a set of ideas. But here we are concerned with the basic material principles of which there can only be two; either that of truth or that of superstition. The latter cannot be termed an error; for to those steeped in superstition it operates as the perverted elemental principle which, unlike an error, is incapable of being corrected. Furthermore, the treatment it receives at the hands of society is vastly different. For logical errors are not being preached, at least not everywhere, and not for generation after generation. Neither do the teachings of individuals that are founded on stupidity and error ever enjoy the sanction of the entire community. Errors of this type which do succeed in spreading widely are soon rectified by the laws of logic, which, being universally valid, will periodically reassert their validity. All this is quite different in regard to material untruths. These keep centuries in darkness and are tenaciously preserved through thousands of years.--Constantin Brunner

posted by No Robots at 9:56 AM on May 5, 2011 [1 favorite]

Interesting. Thanks for posting this.
posted by painquale at 9:58 AM on May 5, 2011

Probably the most important reason they killed Socrates was that he was "making the weaker argument the stronger"

From Aristophanes' parody of Socrates, The Clouds:

Socrates enters the Thoughtery; a moment later the JUST and the UNJUST DISCOURSE come out; they are quarrelling violently.

JUST DISCOURSE
Come here! Shameless as you may be, will you dare to show your face to the spectators?

UNJUST DISCOURSE
Take me where you will. I seek a throng, so that I may the better annihilate you.

JUST DISCOURSE
Annihilate me! Do you forget who you are?

UNJUST DISCOURSE
I am Reasoning.

JUST DISCOURSE
Yes, the weaker Reasoning."

UNJUST DISCOURSE
But I triumph over you, who claim to be the stronger.

JUST DISCOURSE
By what cunning shifts, pray?

UNJUST DISCOURSE
By the invention of new maxims.

posted by empath at 10:00 AM on May 5, 2011 [4 favorites]

More than that -- reasoning is how we explain our own motivations to ourselves after our brain has already made a decision.

This may be overstating the case or imputing a single concatenation of mental events to a more diverse set of possibilities. Not all reasoning is rationalization. One might reason oneself out of a particular decision; people do this every day.
posted by clockzero at 10:01 AM on May 5, 2011

I think they're separating "reasoning" from its semi-synonym "logic". If used properly, logic should only lead you to facts and truths. Reasoning can take you wherever you want to go.
posted by oneswellfoop at 10:03 AM on May 5, 2011

"Reason is both a suit of armor and a sword. But, it protects and is wielded by the child within."
posted by Debaser626 at 10:06 AM on May 5, 2011 [1 favorite]

"I think they're separating "reasoning" from its semi-synonym "logic"."

Yep - "... it is a common but costly mistake to confuse the causally and temporally related steps of an inference with the logically related steps of an argument." - p.7
posted by airing nerdy laundry at 10:07 AM on May 5, 2011

This may be overstating the case or imputing a single concatenation of mental events to a more diverse set of possibilities. Not all reasoning is rationalization. One might reason oneself out of a particular decision; people do this every day.

I'm sure that's right, but even when you reason about something as simple as 'should i eat this 10 day old slice of pizza?' there's probably a significant element of subconscious decision making going on. There are a lot of cases where reason only narrows down the choices and the yes or no still comes from 'the gut'. But you'll still rationalize an excuse for why you did one or the other, in most cases.
posted by empath at 10:10 AM on May 5, 2011

Reason won't help you get to the truth, but it will help you explain it once you're there. Which doesn't bother me at all...so...help me out. Where is the gap?

What gap are you referring to?
posted by EmpressCallipygos at 10:12 AM on May 5, 2011

Ok, this is a really clever way to reconcile domain-general reasoning with a massive modularity thesis. I've definitely wondered before how the two of those are supposed to be compatible. The idea is that we have a domain-specific module for argument evaluation, but given that arguments can be about anything, the conclusions have domain-general consequences. So we get the benefits of a central processor without actually having one.

Can the argument module handle inductive or abductive reasoning, or just deductive? (continues reading)

Lots of you guys talking about Plato and such pretty clearly have not read the paper.
posted by painquale at 10:14 AM on May 5, 2011 [1 favorite]

I think the article could have made a much better opening statement by sticking to purely mechanical/random events (hot slot machines, roulette numbers, etc.) than by basketball field goal percentages. While it's perfectly fine for them to introduce the notion of confirmation bias & stubborn beliefs, the article seems to regard making a shot as being both A: a solely individual endeavor, unaffected by external forces, and B: able to be evaluated in foresight and hindsight on identical terms.

In their example, when Toney had made 3 shots in a row, his percentage dropped down to 34%. Would their recommendation be that "once Toney has made 3 shots, don't give him the ball?" I doubt it. They only know about the 34% in hindsight, because it's *what actually happened*. Unless they can demonstrate some reason why a 4th shot is intrinsically more difficult than a 3rd shot (by the same player) the statistic is as meaningless as a player having an uncanny knack of scoring triple-doubles on Thursdays. In short, they are handicapping based on a single yes/no metric.

Scenario:

Person A: I'm gonna roll this die of (indeterminate) sides. What number do you want?
Person B: What numbers have been coming up lately?
Person A: #1 has come up for the last 6 rolls in a row.
Person B: I'll take #1

Person A: Sorry! It's number 4.
Person C: See! I told you there's no such thing as a streak! You only had a 1-in-4 chance! Math!
Person B: Thanks for the heads up, and by the way, what do you call the previous 6 rolls?

The article also fails to point out one of the chief benefits of confirmation bias (as far as it applies to discovering the "truth"): it's right a lot of the time.
posted by ShutterBun at 10:23 AM on May 5, 2011

Sorry, time's up. I'm afraid you'll have to pay for another session if you'd like to continue.
posted by AceRock at 10:29 AM on May 5, 2011

Reasoning was designed?

Intelligently, I'm sure you'll agree!
posted by nickmark at 10:32 AM on May 5, 2011

@painquale - Sperber tries to reconcile his massive modularity with a "content"-general logic module by saying the domain of inputs is nevertheless "form"-restricted.

He proposed this kind of module a while ago, which processes simple inferences and checks for validity.

... I think he's running the danger of watering the massive modularity thesis down to lacking content tho ...
posted by airing nerdy laundry at 10:37 AM on May 5, 2011

Well I'm shocked. Shocked I tell ya.

Next you'll be telling me that water is wet or something.
posted by seanyboy at 10:43 AM on May 5, 2011

I think he's running the danger of watering the massive modularity thesis down to lacking content tho ...

I can see that. It's not obvious what the difference between a reasoning module and a central processor would be.

The claims he's making in this paper make it a lot more contentious though. Even if you don't think there's a difference between a reasoning module and a central processor, claiming that it had evolved for communication purposes, and only in creatures with metarepresentational capacities... anti-modularity folks aren't gonna buy that.
posted by painquale at 10:49 AM on May 5, 2011

Lots of you guys talking about Plato and such pretty clearly have not read the paper.

I read the linked article, but I didn't follow all the links in the links or all the links in the links in the links.
posted by empath at 10:50 AM on May 5, 2011

Mercier's homepage has a more layman friendly summary.
posted by AceRock at 11:13 AM on May 5, 2011

... anti-modularity folks aren't gonna buy that.

Because... umm... everyone's trying to win the argument for their own side?
posted by philipy at 11:16 AM on May 5, 2011

Let's say that this is an accurate ~~argument~~ ~~theory~~ hypothesis.

If I had a dollar for every time someone promoted their half baked hypothesis to theory without experiments, or searching for contrary data, or even really looking very hard at the data they already had, I'd be all, "You kids get off my internet!" because, hey, I'd own the whole thing. Instead I just grind my teeth in meetings.

I don't think, though, that this is about winning an argument. I think they're mostly trying to ballance what the effort required to know what's going on with their desire to reallyto know what's really going on, comming up short and finding a comfortable "truth". And there you are ruining it for them. It's really the product of apathy times sloth.

When you really want to understand things, you'll set upon your own hypotheses with all the paternal love of a misanthropic Spartan. (As if being a misanthropic Spartan wasn't it's own reward.)
posted by Kid Charlemagne at 11:22 AM on May 5, 2011 [1 favorite]

Let's say that this is an accurate ~~argument~~ ~~theory~~ hypothesis.

If I had a dollar for every time someone promoted their half baked hypothesis to theory without experiments, or searching for contrary data, or even really looking very hard at the data they already had, I'd be all, "You kids get off my internet!" because, hey, I'd own the whole thing. Instead I just grind my teeth in meetings.

I don't think, though, that this is about winning an argument. I think they're mostly trying to ballance what the effort required to know what's going on with their desire to reallyto know what's really going on, comming up short and finding a comfortable "truth". And there you are ruining it for them. It's really the product of apathy times sloth.

When you really want to understand things, you'll set upon your own hypotheses with all the paternal love of a misanthropic Spartan.

As if being a misanthropic Spartan wasn't it's own reward.
posted by Kid Charlemagne at 11:22 AM on May 5, 2011

I got yer hot hand right here. Just can't seem to keep it out my zone....
posted by Benny Andajetz at 11:22 AM on May 5, 2011

@EmpressCallipygos : sorry for rambling on so.

I am wondering why it genuinely seems to bother people to say that the purpose of reason is justification or rationalization of an idea, not truth-divining.
posted by FunkyStar at 11:55 AM on May 5, 2011

Conversational terrorism.
posted by mullingitover at 12:02 PM on May 5, 2011 [1 favorite]

...Rambling on? I'm not sure that I accused you of that; I just didn't understand what you meant by "gap," that was all.

I suspect -- and this is a hunch -- is that "reason" is held up as this impartial thing, kind of like the scientific method, and so perhaps the implication that reason has had a bias from the get-go runs counter to what people perceive reason to be.
posted by EmpressCallipygos at 12:05 PM on May 5, 2011

Where, there is 'reason' and logic as an idealized process, which really is about finding truth, and 'reason' as it's actually used in every day life which is mostly about persuasion and winning arguments.

You obviously can use reason to determine a great many true things. In fact, I doubt that you can discover very much that is true without it (as well as experimentation and observation).

That most people don't use reason for that doesn't change that.
posted by empath at 12:10 PM on May 5, 2011

Funkystar: I think the concept of "reason" has a somewhat broad meaning here, varying along a spectrum of "why am I doing/saying/believing this" and "why IS this." (granted, it depends on what your definition of "is" is)
posted by ShutterBun at 12:23 PM on May 5, 2011

As others have noted, this article seems to hinge on an odd (and I think misleading) use of the word 'reason.' It seems like Mercier's point is, human beings aren't so good at objective logical thought; what we're good at is a bundle of cognitive biases that, when we accept them, help us win arguments.

Which I totally buy. What's weird is that he refers to that second thing, the bundle of biases, as 'reason'. But conventionally, isn't 'reason' actually the first thing, the thing we almost never do? A better summary, I think, would be: humans almost never reason, they just believe what their biases tell them. Which is way a less provocative statement, which is perhaps why the author didn't put it that way. :)

Incidentally, it's not that we never reason, or are incapable of doing so. The first set of researchers cited, the ones who discovered that "hot hands" are a myth, were doing just that, right? It was reason that allowed them to avoid the trap of their own confirmation bias.
posted by molybdenum at 1:37 PM on May 5, 2011 [1 favorite]

No, it wasn't.

Sorry, time's up. I'm afraid you'll have to pay for another session if you'd like to continue.

That was never five minutes just now!!

I, too, have seen an episode of a TV show.
posted by decagon at 2:57 PM on May 5, 2011

I was thinking I read something recently that was awfully familiar to this stuff this morning and then went to Mercier's hompage linked above and BAM! There's a bust of Protagoras.

Yay, sophistry!
posted by P.o.B. at 3:29 PM on May 5, 2011 [1 favorite]

I hate these kind of things, because I'm not that great at math, but....

Does this kind of analysis really answer the "streaky" question? If you are good at an activity, but do that activity a bazillion times, won't most of your performance metrics move to the middle of your ability? Sure, you can show that if you've made three baskets you have X% chance of missing the next one. But that's in aggregate - you really need all the total instances of X number of baskets in a row to show streakiness. Am I off base here?

( I've also seen mathematical studies that show that Tiger Woods is waaaay more likely to sink a bogey putt than the exact same putt for birdie. But Tiger is streaky - as are all top-level golfers- in my world.)
posted by Benny Andajetz at 3:45 PM on May 5, 2011

So what you're describing is recognition of a streak after it happens, e.g. he hit 12 field goals in a row 15 times during 2009 and missed 12 field goals in a row 8 times during the same period. That's not quite the same as the "hot hand" theory, which is saying that a player who hits three shots of some kind is more likely to hit the fourth, and a player who misses three shots is more likely to miss the fourth. In fact, the study demonstrates basically the opposite, that streaks are unlikely to happen because shot-making suffers during the streak of hits and improves during the streak of misses. Note that this isn't just "you have x% chance of missing after hitting 3", it's "you have x% chance of missing, which is higher than your average". So the regressed mean does come into play as a baseline metric.

That isn't to say that streaks as we perceive them don't happen; obviously, people do hit 12 field goals in a row or whatever the equivalent is. What the study disproves is the notion that streaks are self-maintaining, which is what's being claimed when people talk about being "streaky" or "hot and cold". Actually, the research suggests that streaks are largely self-defeating.
posted by Errant at 4:37 PM on May 5, 2011

The problem with reason is that it took a while to show up in history, and also takes a while to emerge in development.

It's not inevitable and it's not the only way to think, in other words.

Therefore, all cultures and all individuals come from unreason to reason, so reason must have unreasonable roots. This paper is an attempt to investigate what those roots might be, perhaps.

But even once you've got reason, it's not enough, because it's unreasonable to be reasonable. To say otherwise is to argue in a circle, which reason forbids.

Reason, then, depends on some other, more basic way of reaching a conclusion for it's claim to be a valid way of thinking.

I find it hard not to conclude that whatever this more basic way of reaching a conclusion is, it must be more valid than reason as well as more powerful than reason.
posted by jamjam at 4:59 PM on May 5, 2011

Obviously gambler's fallacy is a real thing, but so is the fact that someone's perception can affect their skill and how well they play. There's a whole field dedicated to the idea that if someone believes they are playing better than they will play better. See also Flow.
posted by P.o.B. at 5:18 PM on May 5, 2011

Right. "Hot hand" is not actually the gambler's fallacy, because performance is controllable by the player in the first and not in the second, and elements return to neutral state in the second and not in the first. But it is interesting that runs of hits induce increased miss chance and vice versa.

That isn't to contradict sport psychology or flow mentality, which are more concerned with overall performance and not with specific strings of result. It's arguable that "hot hand" induces the kind of confirmation bias in which results counter to perception are discarded, so while someone believing they have a hot hand is not really streaky, it improves overall performance to believe that they are.
posted by Errant at 5:53 PM on May 5, 2011

I had always thought that 'Logic' and 'Rhetoric' were two different things...

...and who gives a f*ck about f*ckin' 'Grammar' these days...
posted by ovvl at 8:26 PM on May 5, 2011

...as an introduction to 'Mathematics', 'Geometry', 'Music', and ... 'Astronomy'!
posted by ovvl at 8:27 PM on May 5, 2011

There's a mystery behind the Forer effect, and subjective validation, and confirmation bias, or basically why everyone hears their own horoscope when it is read to them. But it isn't difficult. Nature is so random that it doesn't make sense to the immediate situation. It rains on our weddings, etc. So we assign a meaning to the information ("The gods must not want me to get married", etc). This helps us cope with adversity and, ironically, allows us to better conquer the odds of survival by not thinking about them, because until we can change the odds in our favor in a very modern sense, they wouldn't matter so much. That's probably why we go around saying in our pseudo-wisest moments, "There must be a reason for this trial in my life." That's a sign that it's not a reason we can change.
posted by Brian B. at 6:52 AM on May 6, 2011

« Older My God is doin' a *pant* brand nu *wheeze* thang! | A History of the Library as Seen Through Notable... Newer »

This thread has been archived and is closed to new comments

苍井空番号迅雷链接

The Argumentative Theory of Reasoning May 5, 2011 9:08 AM Subscribe

The Argumentative Theory of Reasoning
May 5, 2011 9:08 AM Subscribe