Re: The Conjunction Fallacy Fallacy [WAS Re: Anti-singularity spam.]

From: Richard Loosemore (rpwl@lightlink.com)
Date: Thu May 04 2006 - 12:53:20 MDT


Eliezer S. Yudkowsky wrote:
> Richard Loosemore wrote:
>>
>> Human minds are designed for immensely sophisticated forms of
>> cognitive processing, and one of these is the ability to interpret
>> questions that do not contain enough information to be fully defined
>> (pragmatics). One aspect of this process is the use of collected
>> information about the kinds of questions that are asked, including the
>> particular kinds of information left out in certain situations. Thus,
>> in common-or-garden nontechnical discourse, the question:
>>
>> Which of the following is more probable:
>> 1) Linda is a bank teller and is active in the feminist movement.
>> 2) Linda is a bank teller.
>>
>> Would quite likely be interpreted as
>>
>> Which of the following is more probable:
>> 1) Linda is a bank teller and is active in the feminist movement.
>> 2) Linda is a bank teller and NOT active in the feminist movement.
>
> Old, old, old alternative hypothesis disconfirmed a dozen ways from
> Tuesday. I was very quickly summarizing an extensive literature with
> thousands of papers. These are not my ideas, these are the mainstream
> conclusions of an experimental science. Go forth and read the
> literature before you make up your own interpretations. I suggest
> starting with "Judgment Under Uncertainty" and moving on to "Heuristics
> and Biases".
>

You just attacked an argument that has nothing whatsoever to do with the
claim that I actually made.

The claim that I made was *not*:

(1*) Tversky and Kahneman (etc.) failed to prove their case that
       human reasoning behavior is flawed, because there are
       good alternative explanations that can completely account
       for all of their data.

In fact, I explicitly disavow this conclusion:

[Yudkowsky:]
>> In each of these experiments, human psychology fails to follow the
>> rules of probability theory.

[Loosemore:]
> Correct: because in normal discourse, human psychology is required to
> carry out far more complex, broad-spectrum cognitive processing than
> the mere calculation of probabilities.
>
> People are not very good at doing strict probability calculations,
> because those calculations require mechanisms that have to be trained
> into them rather carefully, in order to avoid the problem of
> triggering all those other mechanisms, which in the normal course of
> being a thinking creature are actually a lot more useful.

You will notice that I clearly acknowledge here that human reasoning
ability is flawed. I am not challenging this basic point and saying
that the experimental data supporting it can be explained away. That
question (about the validity of claim 1*) was thoroughly addressed by
the literature, and what those studies tried to do was to eliminate all
the other factors that were contributing to the effects, to see if there
was any residual effect of pure irrationality.

My interpretation of the particular experiments that you cite (and
remember, you were the one who chose to cite experiments where the
effect was confounded with other factors) is that *other* *cognitive*
*mechanisms* played at least some role in interfering with the subjects'
interpretation of the questions, or interfered with their ability to
execute a probability calculation. Those weak studies were a good
illustration that other things were at work, and that is the reason why
  I used them to illustrate possible interpretations of what was going
on. For what it is worth, you will note (if you read the literature)
that the "irrationality" effects are diminished when more trouble is
taken to eliminate these other things .... evidence that they were
indeed responsible for part of the original effect.

Are you clear on this distinction so far? It is not that [Other
mechanisms interfered with the reasoning, and without those other
mechanisms the reasoning would have been perfect], it is that [Other
mechanisms interfered with the reasoning and *partially* accounted for
the imperfect reasoning].

If you read my post in detail, you can see that the point I was trying
to focus on was the *nature* of the mechanisms that were causing some of
the reasoning errors: unlike the people in the literature, whose put a
lot of effort into eliminating these other effects to see what remained,
my goal was to focus on what those other mechanisms were doing and how
important they might be in the larger scheme of things.

Specifically, there are two interpretations of their role:

1) The interfering mechanisms were just dumb, maladaptive strategies.
Basically, systematic biasses and mistakes.

2) These other mechanisms were not just systematic biasses, but may
actually have been components of very powerful, sensible, adaptive
cognitive mechanisms that do not use logical reasoning, and without
which the system as a whole could not function. A number of people have
arguments that are equal or closely related to this point, so

Interpretation (1) is the default assumption in the literature. To the
extent that the literature looked at what was going on in these
experiments, it tended to treat the situation as one of rationality
corrupted by mistakes.

Interpretation (2) is hard to investigate. These other mechanisms
are extremely hard to pin down, and if you know anything at all about
the nature of the experimentation/theorization cycle in the science of
cognitive psychology (something that I have a few concerns about, given
your armchair experience in the field), you will know that what is
difficult to address by experiments tends to be ignored.

But just because they are ignored, that does not mean they do not exist,
and their existence would have profound implications for the position
you take vis a vis the role of rationality in human thought.

*That* was the argument in my post. Not the silly argument that Tversky
and Kahneman's results can be completely explained away.

In the face of the great difficulty involved in trying to assess the
relative importance of those other mechanisms, I stuck my neck out and
offered some examples of what they might be doing, in an attempt to show
that they might not have been mistakes or dumb strategies so much as
powerful mechanisms that govern pragmatics. I used the weakness in the
Linda studies to highlight these possibilities.

So when I said:

> (my suggested interpretation would have to be tested: I am
> only giving an existence proof for an alternative explanation
> -- I could do the experiment, or maybe somebody already did
> do the experiment.

I did not mean doing experiments to test whether these mechanisms could
be used to explain away the irrationality effects (the totally naive and
uninteresting Claim 1*), I was talking about the almost impossible kinds
of experiments that might establish that exactly *what* those mechanisms
might be. People have done some work along these lines, but it is not
very theoretically deep or empirically sound, in my opinion.

And then, of course, I got to the important conclusion of my post, the
way that some people apply their theories of human irrationality to the
larger processes of cognition (in this case, judgments of future
scenarios), and come to conclusions as if teh Conjunction Fallacy, and
the general lack of logical reasoning skills, were the main determinants
in those analyses.

[BEGIN META-COMMENT]

You say that your analysis represents the "mainstream conclusions of an
experimental science". And you tell me that I should "Go forth and read
the literature before [I] make up [my] own interpretations". You think
I just made up my own interpretations? I have had extensive discussions
about these and other issues with people like Mike Oaksford (he and I
were at UCNW Bangor together, then at Warwick), and to a lesser extent
with Nick Chater. I have also benefitted from many discussions with Tom
Ormerod, who does experimental work in the same field. In 1987 I did an
extensive analysis of Phil Johnson-Laird's work, and although I never
got to challenge him in person, I did talk to Alan Garnham, one of his
students, and got a great deal of agreement from him about my
understanding of the issues and my specific criticisms of J-L.

Strangely, Oaksford and Chater disagree with your analysis, as do Evans
and Over. From my limited understanding of them, Maule and Hodgkinson
(2002), and Tetlock and Mellers (2002) would also be in sympathy with
the position I just took. As would many others that I have talked with
over the years, who do not work directly in this area. Given all of
this, it hardly sounds like you presented the mainstream conclusions of
an experimental science.

I disagree with you intellectually in some quite fundamental ways. This
is intellectual, not personal.

[END META-COMMENT]

Richard Loosemore.



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT