Cognitive Biases and Errors in Logic

[Sunstone]

Proposition: Humanity can be divided into two groups. (1) Those who, when reasoning, make little or no effort to take into account and compensate for their natural tendency to errors in logic and their cognitive biases, and (2) those who, when reasoning, make a meaningful effort to take into account and compensate for their natural tendency to errors in logic and their cognitive biases.

Does the proposition have any substantial truth to it? If so, how? If not, how not?

I myself would argue that most of us fall into either category at one time or another, depending on such factors as how seriously we take something, how energetic or tired we are, and so forth. So, humanity, for the large part, is not really divided into two camps: Those who think skilfully, and those who don't. Rather, most people sometimes think skilfully and sometimes don't. But what do you think?

 

[Falvlun]

I think the skill is taught, rather than innate. Most people do not innately know that they are making a logical error; it must be pointed out to them, and then explained why it should be considered an error. The same thing with biases; we often make them unconsciously and take them for granted. It is not until someone says "Hey, why do you make that assumption?" that we realize it is an assumption at all.

So, perhaps the world can be split into two groups: Those who have been taught to think skillfully and those who haven't. Which is why I think that logic should be a mandatory class in highschool.

 

[idav]

There is a lot of truth to that, the people in the first camp would be the more confident or even arrogant. As we know being confident doesn't mean being right but how likely is someone who is confident to admit any error. Depends on what kind of attitude you go into the question with. Like are we here to teach or learn, most of us realizing that we are often times both teacher and student but may prefer one over the other.

 

[Quintessence]

I wouldn't group people like this at all. I would say that the vast majority of the time, all people do not take into account their cognitive biases. It's part of how human psychology works: through lots and lots of heuristic shortcuts. It is only when our heuristics fail to be accurate and cause problems or when a special situation demands we abandon them that we take into account our cognitive biases.

Alternatively, there is what Falvlun points out. If you are trained to think critically, you're perhaps a bit more likely to do it. I know that as a trained scientist, I can snap into that mode fairly readily when the situation demands it, whereas a person who doesn't have that training has difficulty or cannot. But if I'm honest with myself, most of the time throughout the day I'm the same heuristics-driven animal as everyone else.

 

[LegionOnomaMoi]

I think the skill is taught, rather than innate.

I'd agree with that for some reasoning skills. I'm not a big fan of evolutionary psychology (even a well-controlled experiment with an outcome that is as good as any could be still suffers from the problem of explaining how it might have been beneficial in any number of possible ways). That said, there are clearly some ways in which biases help us survive or at least perform better in most cases, while causing us to fail almost all the time in others.

For an innate judgment that is an advantage, my go-to example is coin flips. Given a fair coin and 20 tosses (H representing heads and T tails), which of the two is more likely?

HHTHTTTHTTTHHHHHTTHT
TTTTTTTTTTTTTTTTTTTTT

Human intuition tells us correctly that the second toss sequence is extremely unlikely, while the first looks more like what we'd expect. We look for patterns, which is absolutely essential for us, but in this case tends to give most people the wrong answer. It is true that we'd expect something more like the first sequence, but that wasn't the question. Both sequences are equally likely. However, without that ability to abstract from specific sequences language and cognition would be incredibly limited.

Computers run completely according to precise operations. The architecture is literally a physical instantiation of Boolean logic. As such, a computer will easily give you the probability of both sequences. However, it takes an enormous amount of work to get a computer to recognize a mix of H's and T's as more likely than all or nearly all H's or T's.

Or there's conditionals: "if you're hungry, there's food on the table." Logically, this means that if you aren't hungry, there isn't food. Clearly, that's not what it actually means. In fact, the statement is logically true if you aren't hungry and there isn't food on the table.

The reason that cognitive psychologists, computer scientists, mathematicians, linguists, and logicians have all participated in a 50+ year quest to get computers to learn more than just procedures is because some of these people study how humans learn, while others try to get computers to do this (and others both). If you practice a skill over and over again, but you were given poor instructions, it becomes advantageous to be able to unlearn that. However, the more quickly one can unlearn whatever it is they learned inaccurately, the more quickly they can unlearn everything.

Logic, precision, deduction & induction, and all other formal reasoning practices are, to some degree, counter to what enables humans to learn languages and think abstractly. We understand "car" such that if we see one we've never seen before, we require no rational processes and application of fuzzy Bayesian inference to calculate the odds that it's a car. We just know. We know because our perceptual system evaluates input using a conceptual network and classifies it accordingly. I don't need measure every branch to determine whether something is a bush or a tree, I don't need to count the number of windows and doors to distinguish a home from a building, all because human cognition is imprecise.

Which is great...right up until you need to use logic, precision, and deduction/inductions.

Which is why this:

I think that logic should be a mandatory class in high school.

couldn't be more true.

 

[Kilgore Trout]

I suspect more education focusing on critical thinking would help. I also suspect that the type of internal sense of consistency and honesty required to apply such thinking to one's self is largely a matter of personality and one's fundamental nature. I also suspect that there are a certain number of people who do not have the ability and/or desire to ever understand or apply such thinking skills.

 

[LuisDantas]

I think it is mostly a social acceptance issue. Rational thinking can make one very unpopular.

 

[LegionOnomaMoi]

It's part of how human psychology works:

And part of how human cognitive psychologists get paid! 2 very readable and interesting books on this issue for those interested:
How We Know What Isn't So: The Fallibility of Human Reason in Everyday Life

Predictably Irrational: The Hidden Forces That Shape Our Decisions

 

[InvestigateTruth]

Proposition: Humanity can be divided into two groups. (1) Those who, when reasoning, make little or no effort to take into account and compensate for their natural tendency to errors in logic and their cognitive biases, and (2) those who, when reasoning, make a meaningful effort to take into account and compensate for their natural tendency to errors in logic and their cognitive biases.

Does the proposition have any substantial truth to it? If so, how? If not, how not?

I myself would argue that most of us fall into either category at one time or another, depending on such factors as how seriously we take something, how energetic or tired we are, and so forth. So, humanity, for the large part, is not really divided into two camps: Those who think skilfully, and those who don't. Rather, most people sometimes think skilfully and sometimes don't. But what do you think?

These are good points. I think there are those who only defend what their current belief is and in doing so they do not hassitate any illogical fallacy. They only care about winning in arguments and avoid loosing.
Then there are extremely rare few who only want to see and say the Truth. They don't care about loosing or winning. They put aside the ego and curiously search for the Truth.

 

[LegionOnomaMoi]

I'm not a big sci fi fan, but when I was a teenager one of my favorite books was Orson Scott Card's Speaker for the Dead. Something said by one of the characters has always stuck with me:

"This is how humans are: We question all our beliefs, except for the ones we really believe, and those we never think to question."

I think the OP's categories and how everyone fits into both is pretty good. I would add, though, something more. We are all confronted with views, opinions, positions, etc., all the time, from sources ranging from other people (in person, online, on TV, etc.) to billboards and books. To me, the mark of an intellectual (of wisdom) has nothing to do with education, and not much at all to do with knowledge. The intellectual (if that word means anything at all) is the person who always considers the question: "To what extent am I agreeing/disagreeing with x position because of an emotional commitment without as much of a rational basis as I'd like to think?" There's nothing wrong with being passionate about a topic (I'd say its essential in many cases), but the more one is able to question even those beliefs they "really believe", and force themselves to ask what logic, what reasoning, is behind such a belief, the more one (IMO) deserves to be called an intellectual.

Knowledge is a bunch of memorization, intelligence is all kinds of different ways in which a person can integrate new information with stored information, but one can have 5 doctorates and be the expert in a dozen fields, and still be as dogmatic as a brain-washed cult member.

 

[PolyHedral]

Computers run completely according to precise operations. The architecture is literally a physical instantiation of Boolean logic. As such, a computer will easily give you the probability of both sequences. However, it takes an enormous amount of work to get a computer to recognize a mix of H's and T's as more likely than all or nearly all H's or T's.

Any sequence of n coin flips is as likely as any other.

 

[Falvlun]

Any sequence of n coin flips is as likely as any other.

I remember when I realized that for the first time. The term "revelation" never felt more appropriate.

 

[LegionOnomaMoi]

Any sequence of n coin flips is as likely as any other.

This is true. And it is cognitive-perceptual abilities which, so useful most of the time, tend to yield the wrong answer to the question "which sequence is more likely?" However, the intuitive sense that makes most people get the wrong answer is actually "better" in a very particular sense: people get the question wrong because they correctly realize that one sequence is highly unlikely, and think of the other in terms of patterns. Why is this intuition better than getting the right answer? Because although all sequences are just .5^n, certain distributions of H's and T's are more likely than others.

Ideally, one should (whether because they are Gauss or because they took a class) be able to answer the question correctly and still realize that there is a good reason for the feeling that the sequence which appears more random also appears more likely.

Given the total probability space, how many sequences are there which will be all heads, all tails, or at most all of either with the exception of 2 given 20 fair tosses? Compare that number of sequences relative to the power set of all permutation sets of sequences that contain 10 H's and 10 T's.

Humans get the question wrong because they inherently and automatically judge as if they knew what a power set and a permutation set were. They realize intuitively that some mixture is far more likely than another, and thus they are calculating the probability of an abstract pattern or configuration rather than restricted to evaluating each sequence. They're still wrong, but it is this kind of abstraction that is so essential for higher level cognitive processes.