Good guys, Bad guys

<[Bill Leach 950704.04:21 U.S. Eastern Time Zone]

[From Rick Marken (950703.1510)]

Into the fray (again)...

It seems that the effort to express EAB in math terms is essential.

What Bruce is saying (as I perceive it), is not that the EAB
explainations for all observed behaviour are correct and internally
consistent but rather that the can be made and are made, verbally, and
sound sufficiently plausible to be generally accepted in the field.

Building an EAB math vocabulary may well prove to be a monumental
undertaking (though it also may well not be necessary for such a work to
be complete). A careful start with well defined methodology provided for
extension should do.

It is necessary, I think, to show that many "formal" EAB explainations
are internally inconsistent and/or at variance with one or more
fundamental principles of the field.

It is not and will not be adequate to verbally point such things out,
there will always be "other" explainations... work of different
researchers to consider and all of it verbal and incomplete.

The "incomplete" part is a problem for HPCT just as much as for EAB and
an astute EABer should be expected to notice that fact. We are faced
with the quite serious fact that we have not modeled specific system
behaviour that is the object of EAB research. We have only "talked"
about non-linear and discontinous behaviour.

While I (and I think everyone else here) admits that Control Theory
itself provides ample assurance that the ability to model such systems
behaviour is a virtual certainty, such confidence, from an EABer's
perspective, looks little different than their own confidence.

What I think has to occur is that a sufficient precise symbolic EAB
system has to be created that exposes the existance of: "Albus type
models" (magic occurs here), use of "standard" terms that can not
possibly have the standard meaning when examined closely, use of
identities as causal explainations, hidden logical absurdities, and the
like.

The fact that EAB people already like to run "hard" data controlled
experiments and are uncomfortable with data anomolies is I think, a sign
that this is the one branch of behavioural science where a dispasionate,
solid, logically, and mathematically based challenge could be taken
seriously. It is not going to be accomplished without first producing a
common precise language for discussion.

It may indeed prove to be all but impossible to produce such a common
language with existing level of formality of EAB terms but even an honest
effort at just that should still prove to be very useful. As I
understand the situation now, many EAB researchers are already
dissatisfied with some of the ambiguities and vagueness that exist for
many of their "standard" terms now (this is not meant to claim that EAB
researchers are somehow horribly remiss but rather that ALL of the "soft"
science suffer from this problem of not having relatively complete formal
definitions for their terms).

-bill

[From Rick Marken (950703.1510)]

Bill Powers (960703.1315 MDT) --

Are you sure you want to call our old pal Bruce Abbott a "merchant of
obfuscation?" Sounds unnecessarily nasty to me.

No. I did not want to call Bruce Abbott a "merchant of obfuscation?" Whatever
I say is selected by previous consequences of what I said;-)

But what if Bruce wants to do it his way instead of yours?

Then he would not only be a control system, but one that controls relative to
its own reference signals. I can live with that.

You want to say "Behavior is the control of perception. Period. End of
discussion."

Nope. I want someone to say "reinforcement theory predicts this; PCT predicts
that; now lets see what happens."

Me:

If, however, these two explanations of behavior [PCT and reinforcement
theory] are fundementally the same (as Bruce seems to be saying)

Bill:

I don't recall Bruce ever saying anything like that.

As I said in the post, Bruce has always been quick to explain how
reinforcement theory can handle every apparently contradictory finding we
present -- ratio data, random consequences, changed reference state. That's
why it SEEMS to me that he thinks that PCT and reinforcement theory are
basically the same. But I don't really care if he thinks this or not -- all I
want is an experiment that will distinguish PCT from reinforcement theory.

I think the ratio, random consequences and changed reference state data
DO distinguish reinforcement theory from PCT and they clearly reject
reinforcement theory. But Bruce keeps saying that these data don't
distinguish reinforcement theory from PCT and, thus, they don't reject
reinforcement theory. If he knows this, then he must already know enough
about reinforcement theory and PCT to know what kind of data WOULD
distinguish the two theories. That's why all this effort at "comparing" the
two theories looks pretty suspicious to me.

You are casting him in the role of the bad guy (for some reason I don't
comprehend, other than that he's available),

I LOVED my mother and wanted to kill my father. I confess.

Of course you are perfectly right in defining the end-point. But by
demanding one-step perfection, you are very effectively making it
impossible to get there.

I am not demanding one step perfection; it's just that something seems real
fishy to me. Here's what I know:

1. Bruce Abbott (not a bad guy;-)) knows enough about PCT and reinforcement
theory to know what data DOESN'T distinguish the two theories.

2. Bruce wants to go through a long and laborious process (with Bill Powers,
also not a bad guy;-)) of translating the verbalisms of reinforcement theory
into a set of system equations that allow a comparison of the predictions of
teh two theories.

3. Bruce is an advocate of PCT because he knows that PCT is a better theory
than reinforcement theory.

4. Bruce has not presented (or acknowledged) one piece of data that can be
accounted for by one theory (PCT or reinforcement theory) but not the other.

I'm having trouble putting all this together. I can see how (2) fits in
with (4). But I don't see how (1) and (3) fit with (2).

In Bruce's language, it seems like Bruce (a nice guy;-)) is rejecting
Ptolomy's theory 1) before he knows what it predicts and 2) before he's
looked through the telescope. And it seems like he's getting upset at me
because I'm saying that Ptolomy's theory doesn't predict some things I've
seen through the telescope.

Is it really that difficult to find a prediction of reinforcement theory (a
theory that has been around for over 50 years; people should know what it
predicts by know) that differs from a prediction of PCT?

Best

Rick