Re: Is the Curt net a kind of decision tree?

Glen M. Sizemore wrote:
Michael Olea wrote:
Curt Welch wrote:

And that, sorry Curt, is all I have time for at the moment.

* Sisemore The Odious, Glen, is the shining light in this newsgroup. It is
his expertise that keeps me here.

AC: If I get some time over the next couple of days, I may try to reply to
the link you posted, but I DO want to reply to this ...

I have the same problem with Glen that I have with Wolf and that I have
with Curt ... their claims overstep their arguments, and yet they are
all insistent that they are correct. Ultimately, Glen ends up arguing
that everyone should share his interests and his research projects[]



GS: No, not everyone. You're not invited Cybulski. Don't take that as an
insult, I just think that is much more valuable for you to continue thinking
about non-physical "forces" that, nonetheless, have physical effects.

For someone who later accused me of not being able to produce reasoned
arguments, it is a bit inconsistent to use an "ad hominem" argument
here: attacking one argument of mine based on an irrelevant argument
that I make elsewhere. My dualism has nothing to do with my belief in
folk psychology, nor is my belief in folk psychology dependent on my
dualism.

As an aside, you seem to have forgotten my many comments that I'm not
necessarily a substance dualist. That depends on what is meant by
"physical", which I allow all the materialists to define. My argument
is nothing more than I see no reason to claim that mind is neurons,
since neurons seem utterly unqualified to produce conscious experience.

AC: Look at the article you posted. What he ends up doing is "translating"
cognitive or mentalistic terminology into what he thinks are
behaviouristic correlates: it's all about the environment and it's all
behaviour, it's the same as in a pigeon, etc, etc. To ME, doing that
loses a lot of detail that I want in order to discuss intelligence.



GS: But an important question is why you want these alleged details. Not
because the answers to that question logically defeats some of your alleged
details, but because it casts suspicion on them.

I want those details because they are interesting details and
distinctions in human behaviour, and between humans and animals. In
short, based on observations, but internal and external. How is that
suspicious?

BTW, if you want to argue that I have philosophical prejudices that
interfere here, note that the same claim can be made of you. In short,
your strong desire for uber-scientific methodology makes your
rejections of things that you consider not scientifically rigourous
suspect.




AC: First, what it does is eliminate the distinction between automatic or
instinctive responses and reasoned responses.



GS: No it doesn't. With the caveat added that your nomenclature is
needlessly vague, I can say that the distinction is rather important to
behaviorism. The distinction is between contingency-shaped and rule-governed
behavior.

If it's so important, how come you and other behaviourists never
mention it? How come the article that Michael says you provided him
never uses it in translating the mentalistic or cognitive viewpoints?

As for my nomenclature being "needlesly vague", I fail to see how yours
is any better. In my case, I may need to define reasoned. In YOUR
case, you need to define what a contingency is, what a rule is, what it
means to be shaped by a contingency or governed by a rule, and address
why a rule-governed behaviour CAN'T be contingency-shaped (f it can be,
then you wouldn't have a distinction at all).




AC: While it may be the case
that the underlying implementation is not different -- a claim that
behaviourism CANNOT make, since it studiously avoids the level of
detail on implementation that is required to make that claim[]



GS: Half right (which is about the best I have ever seen you do). One can
make the claim that rule-governed behavior is ultimately contingency-shaped.

Is that what you claim?

That is, that "language use" is behavior that can be explained by pointing
to the contingencies that shape and maintain it, even when speaker and
listener are in the same skin.

One can also claim that language use is influenced by contingencies
without commiting oneself to the idea that that's ALL language use is
or is influenced by. The latter commitment is what you and other
behaviourists TEND to make, even if you backpedal on it later.

One can hardly talk about human behavior
without talking about thinking, but that doesn't mean that thinking is not
behavior.

Who said it wasn't? Under the broad definition of behaviour that you
use, it's clearly behaviour. I'm just arguing that it's a special type
of behaviour that has interesting unique functions and qualities.




AC: -- we
wouldn't call automatic responses themselves INTELLIGENT. Habits
especially can lead to incorrect behaviours that I consciously knew
were wrong.



GS: To the contrary, contingency-shaped behavior falls into your category of
"automatic" but such behavior is frequently called intelligent. The dog that
walks around the fence to get the food instead of jumping at the fence would
be called, by many, intelligent, or at least more intelligent than one that
doesn't.

You are overextending the meaning of the term "automatic", and also
making the same mistake that most people do: classifying the ACTION as
being intelligent as opposed to classifying the METHOD OF PRODUCING
that action as being intelligent.

First, this is not a clear example (odd, since I GAVE you a very clear
example of what I meant). Who says that the dog is not "thinking"
about the action in your example? Let's expand on it a bit. Imagine
that you place some food behind the fence, but that the dog can see it
through small cracks in the fence. The dog comes upon it and sees that
it could get the food if it could just get to the other side of the
fence. It looks around and sees that the fence is too high for it to
jump over, so it starts to walk down the fence to see if it can get
around it. It can, so it does so. This I would indeed call
intelligent, but it was hardly automatic.

For the automatic case, imagine that you repeat this daily for several
weeks. The dog comes in, sees the food, and learns to automatically
walk around the fence to get the food. In short, it does it without
pausing or even considering other options. Then imagine that you open
a few boards along the fence away from the food on the dog's normal
path, enough so that the dog could walk through. Then further imagine
that the dog doesn't notice and still walks around the whole fence to
get the food. This is a CLEAR example of an automatic response; the
dog takes that action REGARDLESS of the global circumstances or the way
things are. And surely in this case no one would argue that it was
intelligent.

Where the problem comes in is that many people conflate "right" with
"intelligent". One can be accidently and unintelligently right and
intelligently wrong.

And here's something to ponder, Cybulski, what is "behind thought"?
Certainly not "more thought" right? That means, in a curious sense (showing
the paucity of your distinctions) that thought must "be automatic" because
it certainly can't be the product of thought. No?

Again, you overextend at least MY usage of the term "automatic" here.
If thought is involved and has an impact on the action, the action is
reasoned and isn't automatic -- even if at the implementation level
"automatic" responses are in play. Or, at least that's my view for
now. Searle's Chinese Room does often get me rethinking that ...

Now, a smart guy like you
might argue that nervous system behavior is, in some sense, "behind thought"
and I would have to agree. The nervous system is "behind thought" just like
it is "behind" all behavior.

And your point would be ... what, exactly? You'll note that I've never
claimed that neurology was not involved in all human behaviours ...




AC: For example, locking my car door with the keys inside
because when I'm anywhere else I always lock the door "automatically"
(and also have the keys on me). We don't want to call such behaviours
"intelligent" because they are, generally, independent of the "big
picture" environmental details AND my internal beliefs ABOUT those
details. So if those behaviours aren't "intelligent", then simply
relating intelligent behaviours and those behaviours simply loses this
important distinction.



GS: Rule-governed, "thought-driven," behavior can also go astray. The
distinction is useless. The distinction between rule-governed behavior and
contingency-shaped behavior is important, but your distinction concerning
the particular outcome on particular occasions is worthless.

That's because you didn't look at the examples closely enough to
understand my claim. My claim is not that "thought-driven" behaviour
doesn't make mistakes. It's that automatic behaviour makes mistakes in
different ways. Generally, when a reasoned behaviour fails it's
because we forget to take something specific into account, and that the
action would have made perfect sense if we had only considered that one
piece of data. When automatic behaviour fails, we generally end up
arguing that the mistake we made was in not INTERRUPTING the automatic
response to consider the big picture. In short, automatic behaviour
ends up being wrong because it was spawned only by one portion of the
environment which normally indicates that such an action is useful but
in this case absolutely wasn't. The "big picture" is TOTALLY missing
in automatic behaviour. In general, we would claim that if we had
thought about it FOR EVEN A SECOND we would have known that the action
was wrong (such as the keys in the car example).

That's why the distinction is CRITICAL. "Thought-driven" behaviour is
the WATCHDOG of automatic behaviour. An automatic behaviour is
problematic and needs to be changed if it cannot be judged by reasoning
to provide the right action in most of the cases in which it activates.
"Thought-driven" behaviour generates the right action if done properly
and things like memory also provide all the relevant details, but
automatic behaviour fires off even if my memory and beliefs would
insist that the action is incorrect.

And this difference in error-making and reliability is simply ONE
difference that is of interest.

A question for you: Based on your behaviouristic analysis, which should
be trusted more: automatic or thought-driven behaviour?


AC: This leads to the second point: I learn differently than pigeons do --
and the mechanism I use is much more flexible and faster than the
pigeon's. The pigeon learns to peck the object with the "white" text
to get food, whereas I can be taught to do the same thing merely by
being told "press the object labelled 'white" to get food".



GS: We have had this discussion before. The fact that you can do this is not
inconsistent with the definition operant behavior.

I never claimed it was. I simply claimed that it was a different
mechanism. One involves much repetition, the other does not. You seem
to define "operant behaviour" so broadly that EVERYTHING can fit into
it. Fair enough. But that's exactly what I'm arguing against, because
to me it seems unlikely that we can just stick more memory or more
processing power into the pigeon's algorithm and get the human learning
mechanism, meaning that we need to care more about the differences than
the similarities.

The distinction between
contingency-shaped and rule-governed behavior is an important detail, but
both are operants, products of a history of exposure to contingencies and
the current environment.

Since I agree that they are impacted by those things, we aren't really
disagreeing here. I simply argue that what most people -- not you, it
seems -- calls "conditioning" is how the pigeon learns, but humans seem
to have an additional mechanism that does not rely on repetition the
way the pigeon's mechanism does. (I also agree, BTW, that humans DO
use the pigeon's mechanism to "learn". I just don't think that that's
the ideal way for humans to learn, nor that it's a way to learn
"intelligently").




AC: When we
want to change the mechanism to be the "white" object, the pigeon must
be retrained. I merely have to be told "Press the 'white' object now"
and my behaviour is changed. While one can argue that this is an
artifact of language, that does not seem credible, since it seems that
reasoning could do this as well (given sufficient non-language "clues"
to the change).



GS: Huh? OK, what is "reasoning" then? I have been discussing rule-governed
and contingency-shaped behavior, and I will let that part stand because it
is closely related to what many people "mean" by reasoning. But you want to
argue that reasoning is not connected to "language"? Then what is reasoning?

That isn't my argument. My argument is that someone could reply to my
example that the "retraining" is nothing more than knowing how to parse
as sentence to produce specific behaviour, just as it is in the first
case. Thus, all it is is the mechanism of UNDERSTANDING what is said
to you in language. Thus, my comment is that if you left around
non-verbal clues, the untraining and retraining through repetition that
the pigeon does would still be avoided, and hence the mechanisms still
seem distinct FUNCTIONALLY.

And this is the problem with behaviourism and why I think it unduly
hides distinctions: it tends to argue that if the same end behaviour is
produced how it got there is unimportant.

AC: []and the behaviouristic language hides that crucial
distinction that casts doubt on the idea that we just have a faster
processor for what the pigeon does (or more memory or ...).



GS: Well, neither of the things you point to are things that I would say,
but that is secondary to the fact that you have said nothing to defend the
notion that the sort of behavior that you are contrasting are not, a the
most fundamental level, different at all.

Where did I claim that they were different at the most fundamental
level? At the most fundamental level, two humans are no different
because their atoms have the same structure. That doesn't mean that
talking about how two humans ARE different is invalid.

BTW, I know you wouldn't argue about that because you don't talk about
AI. But those who take behaviourism into AI end up saying exactly
that.

AC: As was seen in the
first point (or should be obvious from it) I believe that internal
behaviour is CRITICAL in behaviour that can be classified as
intelligent.



GS: I don't know what you mean by "internal behavior,"

So I misstated your term. I meant "private behaviour".

but I would say that
I think that to explain some aspects of human behavior one must point to
private behavior.

But do you go further? No, you don't. You simply say that private
behaviour is involved and leave it at that. Hardly useful in any
guidance for how to live my life. Do you have any opinion on whether
or not private behaviour is to be privileged as being more useful or
whether automatic behaviour -- without private behaviour -- is
superior? If thought-driven and automatic behaviour clash, which
should be followed, or how should we determine which should be
followed?

In short, you're missing all the normativity in your claims.

I have already pointed out that some behavior (that does
not involve private behavior) would be called "intelligent" by lay-speakers,

And I have shown how that is wrong ...

and the word "intelligence" is, at least part of the time, a colloquial
term.

Such things haven't stopped you BEFORE from calling it wrong and asking
for its change or removal ...




AC: In short, if you don't think about it, it isn't
intelligent.



GS: Ok. If that is the distinction you want to make, that's fine.

Then why do you constantly get upset with me when I talk like that?
Most of your rants about how stupid I am are about such claims.

AC: And so if I pursue my project about considering
what are the qualities of behaviours that INCLUDE internal behaviour, I
end up having to reintroduce mentalistic terms because that idea is set
to DELIBERATELY examine the critical components of behaviours that
include internal behaviour. So for MY project, mentalistic terminology
is useful and behaviouristic terminology is not. Yet Glen insists that
all interests and all theories are wrong if they even BORROW from
mentalistic terminology -- or, at least he does that when he's being
grumpy [grin].



GS: Mentalism is sufficient for ordinary activities but it is a hemorrhoid
on the rectum of behavioral science.

And who says that all projects care about being behavioural science?
Folk psychology -- being about ordinary activities -- would WANT to
maintain mentalism. Psycholgists in most cases would want to avoid
some of behaviourisms more stringent concerns -- like the idea that
past history controls much of behaviour, that small changes in past
history can change behaviour, and therefore to understand current
behaviour (to alter it or see what caused a certain problematic
behaviour) one would have to examine in detail a patient's past
history.

You've basically just justified my claim about what I didn't like about
you. Who says that anyone outside of behavioural science should CARE
about what behavioural science thinks is good or useful? You can have
clear conceptual theories WITHOUT taking on all the assumptions and
normative concerns of behavioural science. Folk psychology AND
mentalism are examples of this that are -- in my experience -- as clear
and conceptually sound as behaviourism (if not as scientifically
rigourous).

It is simply animism.

Yet I have shown you examples of areas where mentalistic terms are
useful without any reference to animism.



AC: I suspect that you -- and Curt -- approve of Glen's ideas because they
do indeed hide these distinctions, allowing you to avoid questions that
would complicate what you are interested in looking at. Which is fine,
but doesn't mean anything about the views of others, nor does it make
the ideas RIGHT.



GS: As I have said, there are two issues concerning behavioristic
extrapolations of known facts. 1.) is the extrapolation logically
consistentand cogent? 2.) will empirical methods verify it? You want to
argue that it violates the first, but it doesn't. The second is much more
difficult to evaluate because we don't get to do the critical experiments
for ethical reasons.

Actually, I don't want to argue the first. I do not claim that
behaviourism is not logically consistent. I simply argue that its
narrow focus makes it problematic for some issues that it seems to
consider outside of its area of interest.

.

Loading