From: Richard Loosemore (rpwl@lightlink.com)
Date: Tue Aug 29 2006 - 19:14:02 MDT
Jeff,
I am glad that you are focussing on the issues, but I have only enough
energy to try to deal with one, highly focussed point in your argument
below.
I am going to be extremely careful, and by being that careful I am going
to show that you are making the most incredible mistake in what you say
below. I am going to be careful enough that you can actually go and
make a simple check and verify what I say for yourself, with an outside,
independent source.
Jeff Medina wrote:
> On 8/28/06, Richard Loosemore <rpwl@lightlink.com> wrote:
>> Jeff: one of your biggest concerns is with the irresponsibility of
>> people who spout their opinions without bothering to read the relevant
>> evidence, so how could you have missed the fact that I so explicitly
>> rejected this claim you impute to me, back there in the original post.
>> Did *you* read it thoroughly? And what did you make of it?
>
> I noted that you never disavowed the alternative interpretation of the
> feminist bank teller results, and that Eliezer *stayed focused on
> replying to that claim of yours, rather than your main point*, and so
> quite independent of whether your main point is right or wrong, the
> issue of your interpretation of the feminist bank teller results
> stayed open and your protestation meant roughly nothing.
>
> To summarize my view of this whole mess, before going back to
> non-participant mode:
>
> You (Richard) said, (cf. http://sl4.org/archive/0605/14748.html)
> "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."
>
> This is a pretty straightforward claim.
>
> It has been disconfirmed in the h&b lit.
Let's concentrate on this, in the hope that some progress can be made by
being very, very precise on at least one point.
What exactly was disconfirmed in the h&b literature here? There was a
WEAK claim, implicit in a lot of the attacks made against experiments
like the first Linda ones, and that weak claim could be paraphrased as
follows:
Weak Claim: The Linda results can be explained away (i.e. they would
vanish completely) if subjects were not misled or confused by the nature
of the question. In other words, the weak claim is that the Linda
result is purely a consequence of asking the question in such a way that
the subject was tricked into a misunderstanding.
Are you with me so far? This claim is an attempt to say that, in
effect, the experimental result was an artifact of a bad design. A
better design would ask the question in such a way that the subjects
were not misled, and then (the weak claim would say), we would expect
the effect to disappear.
To people familiar with the way that this subfield played itself out,
this is the Familiar Story: first the Linda experiments, then the
attempts to claim they were flawed, then the subsequent experiments
which .... well, we won't jump the gun just yet, but to someone like
myself or another cognitive scientist familiar with the literature, the
story DID involve this Weak Claim, and then the subsequent attempts to
see if it was an adequate counter to the original interpretation of the
experimental results.
To someone who was in a hurry, reading those paragraphs of mine that you
quoted above - and *not* bothering to read further into the message -
might give them the impression that all I was doing was writing out a
simpleminded restatement of that Weak Claim.
So, now: what happened to that Weak Claim. Well, guess what? It was
disproved, as Eliezer put it, every whcih way from Sunday. The simple
version of the story is that they re-did the first Linda experiments,
and then re-did them again as OTHER alternative explanations of the
results were offered, until eventually it was shown that the effect was
real: people were not simply misunderstanding an ambiguous question,
they were doing something wrong inside their heads. The Weak Claim was
not supported, at the end of the day.
But now I want to point something out to you. In my message I make a
very precise statement: I say that the "the [Linda] question would
quite likely be interpreted as..." and I give the alternative that was
first proposed as a misreading of the experimental question.
This is the crucial point: you (and Eliezer before you) jump on this
statement and *immediately* say:
> This is a pretty straightforward claim.
>
> It has been disconfirmed in the h&b lit.
Jeff: do you *know* that the effect was much weaker when the confusion
was eliminated? Do you *know* that when the protocol was tightened up
to exclude that alternative explanation, the effect diminished?
Do you understand what that means? It means that the alternative
explanation quite likely *did* play a role in the initially observed effect.
And the subsequent history of the series of experiments was that when
other alternatives were suggested by those advocating the Weak Claim,
many of those, too, when they were eliminated as possibilities, caused
the effect to go down. Not all of them, but many of them.
And so this means that, whatever was happening, the alternative
explanations offered DID turn out to be something that was going on in
the minds of the subjects.
This is important, because I *knew* this perfectly well when I wrote
what I wrote: I knew it because I am a cognitive scientist steeped in
the experimental and theoretical results of this field, and I have known
it for a long time. When I wrote my words in that first message, I was
generous enough to give Eliezer credit for also knowing the field, and
so I did not hand-hold him through all the tedious details, assuming
that he was smart enough to know them already.
What is the significance of this? Quite simple: my words "the [Linda]
question would quite likely be interpreted as..." introduced one of the
mechanisms that DID kick into play .... this was my introduction of an
example mechanism that is at least partially responsible for what goes
on in the head of the subject when confronted by the Linda question.
Did I say (as the people who advocated the weak claim tried to do) that
this alternative could EXPLAIN AWAY ALL OF THE RESULT? I most certainly
did not: that is the Weak Claim itself, and I took it for granted that
nobody except an outsider who did not know the field would attempt to
defend it. I wouldn't have dreamed of trying to do that. I'm no dummy.
So what that means is, that when you, in your words above, quote my
statement and then say "This is a pretty straightforward claim." and
then "It has been disconfirmed in the h&b lit." you are guilty of not
looking at the precise words that I used, and perhaps guilty of not
knowing the literature: when I said that subjects would quite likely
have made that mistake, I was making a precisely correct claim: the
literature *did* support that claim.
By saying that my words were a "pretty straighforward claim" you are
saying, in effect, that YOU interpreted this to be an obvious statement
of the Weak Claim. It was not obviously such a thing, but unless you
knew the literature pretty well, you might not have realized that fact.
But even if you had not been familar with the literature, you are also
guilty of taking that statement of mine and cutting it off from its
context: by chopping it away from the rest of the post, and pretending
that it should be considered on its own, you distorted its meaning in
every bit as heinous a way as the tabloid journalist who selectively
edits a quote to turn the meaning on its head.
Because after I introduce that statement I go on to make it clear that
my interest is not in the Weak Claim at all, and that the whole focus of
my essay was on the mechanisms which are *known* to be involved. And
just in case anybody might have doubted it, I stated very clearly that I
disavowed the Weak Claim.
So even if someone (an expert, say) were reading my essay and came
across that first statement of mine, and - because they were not
paying careful attention - thought that I was making the discredited
Weak Claim, that person need only read a little further and see that a
sentence or two later I reject the Weak Claim explicitly, at which point
a smart person would realize their mistake and look carefully at what I
said..... at which point they would discover that I was making a far
more sophisticated claim than any of those old ones that were in the
early literature.
So when you say below that this was my first claim, I say: look at the
words and you will see that even if you cut the text in half at that
point, my words were carefully chosen and were not equivalent to the
weak claim at all ... and if you had been knowledgeable of the field you
would certainly have held your horses at that point and said to yourself
"Well... he is right to say that the subjects could well have
interpreted the question in that way, because the results showed that
the first batch of subjects were indeed doing this at least part of the
time..." and you would have continued reading to find out where I was
going to take it.
In your own case, Jeff, I guess my only criticism would be that you have
jumped into this argument not knowing the field in enough detail (and
you have never claimed to know the field, I should add) ... so my only
complaint is that you are not in any position to make the statements
that you make (above and below) about which one of my paragraphs
constitutes a separate claim. In my opinion you have embarrassed
yourself by stepping into the fray and judging the argument by looking
at the words alone, without knowing the background facts to which those
words were referring, and by being careless about interpreting those
words (again, due to lack of knowledge).
You have an excuse. Eliezer does not. He claims to be an expert. His
supposed knowledge *should* have led him to interpret my words correctly
(saying to himself: "True enough, later research did show that the
subjects sometimes did this in the first experiment: now what of it?").
Reading onward in my post he would have seen my claim laid out carefully
and precisely. With the (supposed) knowledge at his command he should
have understood the detailed character of my argument, and quickly
realised that I could not possibly be talking about anything as foolish
as the Weak Claim.
(At that point, I still gave him the benefit of the doubt and assumed
that he *did* understand the field, but had just boxed himself into a
corner from which he was too embarrassed to extricate himself.)
He may well have had questions at that point. I am not perfect: I do
my best, but still I may not get my point across perfectly the first
time. So ask questions!
[This is what you are implictly starting to do in your words below
(mixed with restatements of the false claim that I have just analyzed).
Fair enough: but this is the discussion that should have taken place,
back at the beginning. I am not going to have it now, with you, for
obvious reasons.]
Instead, I was treated, back in his first response, to a pathetically
weak and supercilious piece of garbage, recommending that I do some
reading that I actually did nineteen years ago. And that was just his
*first* response: after that one I wrote a carefully constructed reply
to explain what the situation was, in the most thoughtful detail. It
took a lot of effort to write that explanation: he didn't deserve it,
but I did it anyway.
When he finally replied to that essay, the other day, I found myself on
the receiving end of the most amateurish display of petulance, stupidity
and arrogance that I have ever seen in someone who claimed to be a
scientist, which then blossomed into fantasy allegations that showed
rather conclusively that I had been wrong in giving him the benefit of
the doubt. He clearly knows the h&b subfield of the human reasoning
studies, but he knows nothing outside that area.
So I end with your attempt to put a charitable reading on the situation
when you talk about Eliezer's misunderstanding of Johnson-Laird, below.
I'm sorry Jeff: it would be nice if things were as you say, but you
could only make your statement if you simply did not know the cognitive
psychology field. Johnson-Laird's book "Mental Models" is one of the
classics of the human reasoning literature - it is his one claim to
fame, not just some little bit of work among hundreds of other papers.
Mental Models may not be on the same sub-branch as the h&b stuff (and
note: I never said *anything* about it being here, there or anywhere!),
but if you were to ask any cognitive psychologist who is familar with
the stuff in the "Human Reasoning" chapters of elementary cog psy
textbooks, whether it might be possible for an expert in that area to
simply not know who Johnson-Laird is, or to not immediately know that
the phrase "mental models" in the context of Johnson-Laird meant
something very specific, they would laugh at you. Try it if you don't
believe me. They'll probably think it is some kind of trick question.
Sincerely,
Richard Loosemore.
>
> This is what Eliezer commented on in his reply here:
> http://sl4.org/archive/0605/14749.html.
>
> It's not your main claim. That's great. It doesn't matter, because
> it's what Eliezer actually commented on. He, like everyone, gets to
> pick which subclaims he wishes to focus on, right? Right.
>
>> From your later comment in the above referenced post…
> "> [Eliezer said:] In each of these experiments, human psychology
> fails to follow the rules
>> of probability theory.
>
> [to which Richard replied:] 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."
>
> … it appears you thought Eliezer thinks the human brain does or should
> merely calculate probabilities, rather than function in the complex,
> specialized, modular fashion that it does.
>
> He doesn't.
>
> It's pretty obvious that he doesn't, given how big on evolutionary
> psychology he is, and seeing as the EP folks heavily emphasize the
> specialization of different components in the brain for particular
> tasks in particular contexts. If you'd ever taken a look at his
> "Levels of Organization in General Intelligence" paper, you'd know
> this.
>
> And we all already know you disagree with Eli's
> non-human-psychology-based, decision-theoretic approach to building a
> thinking thing. So, on that point (re how to build an AGI), you've
> shared nothing new, and one wonders why you spoke up in the
> Conjunction Fallacy thread at all.
>
> Paring away the ad hominem that came later, and the admittedly unusual
> fact that Eliezer had to look up Johnson-Laird (though there's a
> difference between having to look him up because he didn't know who he
> was or what he did generally, and having to look up his large
> bibliography of work to verify whether or not J-L did any work on
> H&B), you two have just been talking past one another.
>
> You, Richard, took Eli's reply to be a criticism of your main point,
> when he hadn't responded to your main point at all. Since Eli's
> comment was so very clearly directed as applying to the
> "interpretation" subclaim I quoted above (and which *he* quoted in his
> reply, so as to be clear what he was replying to), your repeated
> protestations that Eli wasn't addressing your claim fell (and continue
> to fall) on deaf ears – he wasn't *trying* to reply to your main
> claim, so why are you complaining that he's doing something wrong or
> unfair? You did a poor job of reading his email, viz. as if it, by
> itself, were proffered as a wide-swath rejection of your overall claim
> about how to build a (human-like) AGI, and responded in accordance
> with that misreading, causing all this hullaballoo.
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:57 MDT