Self Interest

CSGnet Archive CSG2004
41

[Martin Taylor 2004.04.27.0.20]

From[Bill Williams 26 April 2004 10:20 AM CST]

[Martin Taylor 2004.04.26.0222]

>From[Bill Williams 25 April 2004 6:20 PM CST]
>
> > Martin Taylor 2004.04.25.1530]
>
> >> In [Bill Williams 25 April 2004 12:20 PM CST]
>
> >> I ask, Perhaps it might be well to clarify a point.
>Is >>reorganization the same thing as learning?

Martin in reply would prefer to start from first principles rather than
what has been said about this question-- which is OK with me.

But it doesn't really matter what Bill or I say about it. Anyone can
work out the implications of the theory for themselves.

Perhaps there is as much truth to the assertion that "Anyone can
work out the implications" as there is to the assertion that
pedophiles would make good police officers. Besides I wonder
just "How long it would take me to work out the implications?"

"Can" can mean permission, or it can mean ability. I intended to
suggest permission.

But one more question, if I may, when you use the term "rule" would I be
making a mistake to substitute "filter" where you say rule?

You would. At least as I understand "filter". A "filter" passes some
things and rejects others. A "rule" could do that, but it could also
do just about anything else, such as "replace black with white".

Martin

42

[Martin Taylor 2004.04.26.0.22]

I prematurely sent off my partial response to

From[Bill Williams 26 April 2004 10:20 AM CST]

> [Martin Taylor 2004.04.26.0222]

Is my understand correct that an "optimal orthogonal set of perceptual
input functions" would generate perceptions free from extraneous
confusions resulting from something category mistakes.

"Orthogonal" means that each sees what the others don't. "Optimal" is
in the information-theoretic sense of concentrating the variance in
the smallest number of dimensions. It says nothing about categories
or mistakes, or even confusions. It just says that the neural
structure won't be any bigger than is needed to deal with the
variation that exists in the incoming sensory data.

J.G. Taylor 1963 -- I haven't encountered this work.

"The Behavioural basis of Perception". I can't find it on my shelves,
though it must be there somewhere. J.G.had a feedback loop, but his
notions were not the same as PCT. His intellectual background was
Clark Hull, and perceptions guided actions through reinforcement,
rather than actions controlling perceptions.

I was assuming that a good perceptual system would generate perceptions
that made sense of the perceved environment-- that is presented the world
as a more or less orderly place. I am assuming that the world is more or
less
orderly or organism couldn't exist. So, it seemed to me reasonable to me
to assume that we have an innate disposition to work at understanding what
order there is in an environment. I think I attempt to say approximately
the
same thing later on, and it seeded to make sense to you.

There are two aspects to this. One is that the visual world consists
more of similar regions separated by distinct edges than of randomly
coloured with no spatial correlation. Christian von der Malzberg (I
think--and I think it was in Kybernetica) modelled the development of
layers of pseudo-neurons exposed to naturalistic scenes, and found
that they organized themselves into the same kind of receptive fields
as the primate visual system does -- on-center/off-surround,
orientation, and the like.

But then there's the otehr aspect about the "reliability" of the
world--making sense. That is a question of how the perceptual world
changes when you push it. J.G.T. argued that you only see those
things that are (or have been) affected by your actions. PCT would
say that reorganization links actions to specific perceptions that
the actions affect. Neither could work for patterns that don't
correlate with action.

> There's something missing here, however. One of the

levels in the Powers hierarchy involves something like
algorithms--processes that act on perceived categories. Logical
processes (rule-based operations) can produce many different kinds of
results (perceptions) from the same data, and rules for working on
rules also can develop perceptions of methods. That's a kind of
learning that is more Hebbian than reorganizational, though it is
actually neither. It has quasi-instantaneous shifts of perception as
different rules come into play,

Would perceptual shifts like necker (sp?) cubes be an instance of this?

I don't know. It's possible. I hadn't thought about things that way.

Martin

43

From[Bill Williams 27 April 2004 2:00 AM CST]

[Martin Taylor 2004.04.27.0.20]
>But one more question, if I may, when you use the term "rule" would I be
>making a mistake to substitute "filter" where you say rule?

You would. At least as I understand "filter". A "filter" passes some
things and rejects others. A "rule" could do that, but it could also
do just about anything else, such as "replace black with white".

OK. This clarifies the meaning of "rule" for me.

Now that I understand a "rule" it occurs to me that a rule may be the same
or nearly the same thing as a "concept." Or, at least the same thing as one
specification of a concept.

Bill Williams

44

From[Bill Williams 27 April 2004 2:15 AM CST]

[Martin Taylor 2004.04.26.0.22]

I prematurely sent off my partial response to

>From[Bill Williams 26 April 2004 10:20 AM CST]
>
> > [Martin Taylor 2004.04.26.0222]
>
>Is my understand correct that an "optimal orthogonal set of perceptual
>input functions" would generate perceptions free from extraneous
>confusions resulting from something category mistakes.

"Orthogonal" means that each sees what the others don't.

I understood this part.

"Optimal" is

in the information-theoretic sense of concentrating the variance in
the smallest number of dimensions. It says nothing about categories
or mistakes, or even confusions. It just says that the neural
structure won't be any bigger than is needed to deal with the
variation that exists in the incoming sensory data.

I took this to mean that the number of categories used would be the smallest
number of categories required to handle the situation. If more categories
were used than is neccesary then the same phenomena might be assigned to
either of two categories when one would do. This mistake, if it was noticed,
could be corrected by adopting new rules that would assign the contents of
the two categories to one new category.

If there were too few categories then two distinct phenomena might be
assigned to the same category. When this error became apparent, it would set
into motion a process of assigning the content of the one category to two
new categories by way of new rules.

The nominclature I'm using is a bit different, but it seems to me that the
transformations described would be the same.

Bill Williams

45

[From Bill Powers (2004.04.27.0233 MST)]

Martin Taylor 2004.04.26.0.22 --

> There's something missing here, however. One of the

levels in the Powers hierarchy involves something like
algorithms--processes that act on perceived categories. Logical
processes (rule-based operations) can produce many different kinds of
results (perceptions) from the same data, and rules for working on
rules also can develop perceptions of methods. That's a kind of
learning that is more Hebbian than reorganizational, though it is
actually neither. It has quasi-instantaneous shifts of perception as
different rules come into play,

B.W.: Would perceptual shifts like necker (sp?) cubes be an instance of this?

I don't know. It's possible. I hadn't thought about things that way.

Necker cubes are not really cubes. We do not get those flips with real
cubes unless we view them with one eye, or at a distance so large that
binocular vision fails to resolve the ambiguity. A Necker cube is a
2-dimensional drawing in which either of two three-dimensional
interpretations is equally valid (and Martin claims he can see more than
two). The missing information has to be imagined, namely the distance of
each segment and vertex from the viewer. As soon as you imagine one corner
to be nearer than another, the orientation of the cube is settled and you
see it that way. This goes for all the other ambiguous drawings of 3d
objects in which depth information is missing -- the Old Masters had all
sorts of tricks for adding depth information (like the haze on distant
parts of scenes, or variations in texture or size of known objects) to
their two-dimensional paintings.

I don't think that logic comes into play here -- at least it's not required.

Best,

Bill P.

46

[From Dick Robertson,2004.04.27.0545CDT]6

Mary Powers wrote:

from Mary Powers 2004.04.26

>[From Boss Man (2004.04.25)]
>
>Bill Williams 25 April 2004 7:00 PM CST
> >
>
>Probably the best way to learn this skill is to observe several examples,
>discern a pattern, and
>formulate a rule that allows you to reproduce the pattern. Is this a PCT
>account? No, not as far as I
>know. Is this account consistent with PCT? Yes, as far as I can tell.

I also think it is consistent, though incomplete. It may be that you can
formulate a rule on your own, or have the rule told to you. A rule, PCP
suggests, is a principle level concept, and can set reference signals at
the program level (if-then). That in turn sets reference signals for
sequences, and so on down to actually writing on a piece of paper or
pedalling the bike. However, I don't think that riding a bike requires
starting at such a high level, at least not consciously, unless you are
interested in the physics of it. Following the pattern is probably enough,
or maybe even simply reproducing a sequence. But learning to do math does
involve learning the rules.

As for whether learning is reorganization, I'm inclined to think that it is
not, unless what you are learning conflicts with something you already know
or believe. Then you must either reorganize or reject what is to be learned.

Yeah, but, when reorganization is successful (reduces intrinsic error) the person
can now do something he/she previously could not do. Then we call it learning. I
think that point was already made in B:PC.

Best,

Dick R.

···

At 06:08 PM 4/25/2004, you wrote:

.

47

[From Kenny Kitzke (2004.04.27)]

<Bill Powers (2004.04.26.1739 MST)>

What a wonderful and worthwhile post, Bill. At least for me, it is the kind that keeps me on CSGNet and participating here. What a privilege it is to have the man who conceptualized HPCT and a separate “reorganization” system in humans answer questions about his viewpoints. It feels like getting a personal answer from Einstein.

It is also feels like President Bush personally writing to me to explain the things he does and even considering my ideas without any need to insist I agree with his or dismissing mine out of hand or condemning me for having the nerve to disagree.

This is the kind of model and example of professional and scientific dialogue that would make coming to CSGNet a pleasant, helpful and worthwhile experience; one you could recommend to a friend as a prospective PCTer.

Thank you. Thank you. And, it was very nice to see Mary Powers contribute again too.

But, this is not intended to affirm your speculations of human behavior or nature. It is intended to show appreciation for you explaining your theories. It should not be a big surprise that the more you try to explain the details, the more questions are engendered. :sunglasses:

I have a number of concerns and doubts about your theories and conjectures in the abstract or general, including those more specifically about golf. But, there were some great observations as well and some points we seem to comprehend (perceive) the same way. I sense the dialogue is a way to grow our understanding–a way to learn. I am not sure any aspect of it qualifies as “reorganization.”

I think reorganization, as you define it, is a rather rare thing. But, I am not sure what you suspect. Can you estimate how many times so far in 2004 has your reorganization system been engaged and has successfully reorganized your hierarchy? I am just trying to understand what we agree upon and where the differences are and why. Hopefully, others might benefit from the dialogue.

48

[From Boss Man (2004.04.27)]

Kenny Kitzke (2004.04.27)]

<Bill Powers (2004.04.26.1739 MST)>

It is also feels like President Bush personally writing to me to explain the
things he does and even considering my ideas without any need to insist I
agree with his or dismissing mine out of hand or condemning me for having the
nerve to disagree.

I never noticed the similarity between Bill Powers and George W. Bush. Are you sure about this?

49

[From Richard Kennaway (2004.04.27 16:18 BST)]

From[Bill Williams 26 April 2004 3:40 PM CST]

When you assume that an economic agent maximizes and the agents goals
are free from interdependencies then as it has been widely claimed it can be
proven that market solutions are best. There is no need to look and any

                                                                   ^^^ at?

evidence.

This is in contrast to the usual practice in scientific disciplines,
where the consequences of one's assumptions are subjected to
experimental test.

-- Richard Kennaway

50

[From Bill Powers (2004.04.27.0947 MST);

Kenny Kitzke (2004.04.27)--

Thanks, Kenny. Yes, we disagree about some things, but fortunately there is
some common ground where we can meet.

I think reorganization, as you define it, is a rather rare thing. But, I
am not sure what you suspect. Can you estimate how many times so far in
2004 has your reorganization system been engaged and has successfully
reorganized your hierarchy?

I'd say hundreds if not thousands. I'm spending a lot of time at the
computer these days, struggling to get old DOS programs running in Delphi
so they will work on modern computers. I couldn't count how many times I
have sat looking at something happening on the screen and thinking, "Oh,
no, that HAS to work, so why doesn't it work?" You might say that I'm
saying a prayer to some technological god, but from in here is seems that
I'm just waiting for an idea to pop into my head that might help with the
current problem. And ideas do pop up fairly reliably, some of which work
and most of which don't. I either discard them after a moment's thought, or
try them out to see what happens. They come out of nowhere; I can't take
credit for them. Eventually an idea works, and I go on to the next thing,
no big deal. I'd been doing that all my life, and I suspect that most
people do this fairly continuously. I think it's a background level of
reorganization that fluctuates as little problems come and go, tossing up
variations to be tried out and then subsiding again.

Then there are other things that happen not quite so often, but still often
enough. Somebody says something that ticks me off and I reply defensively
and in haste, which leads to a crisis, which ends up with bad feelings all
around. Yet in a day or two, what mattered to me then changes, and the bad
feelings go away, and I'm trying some different way of dealing with the
same situation. The person can say the same things to me again, and oddly
enough they've lost their sting. They're just words.

I call that reorganization, too. It's a little unpleasant while it's
happening, but over lo, these multitudinous years, I've learned that the
process has a beginning, a middle, and an end, like an event, and that the
outcome will probably leave me able to function again.

So the answer to your question, Kenny, is that I think I see
reorganizations going on all the time in my life, with big ones happening
less often than small ones.

Best,

Bill P.

51

[From Bill Powers (2004.04.27.1007 MST)]

Richard Kennaway (2004.04.27 16:18 BST)--

This is in contrast to the usual practice in scientific disciplines,
where the consequences of one's assumptions are subjected to
experimental test.

I have to admit that I really can't detect whether you're being sarcastic.
It would be believable either way.

Best,

Bill P.

52

[From Richard Kennaway (2004.04.27 17:30 BST)]

[From Bill Powers (2004.04.27.1007 MST)]

Richard Kennaway (2004.04.27 16:18 BST)--

This is in contrast to the usual practice in scientific disciplines,
where the consequences of one's assumptions are subjected to
experimental test.

I have to admit that I really can't detect whether you're being sarcastic.
It would be believable either way.

I am serious. (I prefer to avoid sarcasm and irony altogether: they
invariably obscure communication, as do practically every other
rhetorical device in the book.)

In a scientific discipline, by definition, one does not start from
assumptions, derive conclusions from them, and then assume that those
conclusions must be true. Every experiment is an examination of how
well the conclusion of an argument compares with reality: that is
what an experiment is.

I have to admit that I can't tell what Bill Williams attitude to
assuming the consequences of one's assumptions is, from the message
of his that I quoted, and I'm not imputing any particular attitude to
him.

-- Richard Kennaway

53

[From Rick Marken (2004.04.27.1055)]

Richard Kennaway (2004.04.27 17:30 BST) --

Bill Powers (2004.04.27.1007 MST)]

Richard Kennaway (2004.04.27 16:18 BST)--

This is in contrast to the usual practice in scientific disciplines,
where the consequences of one's assumptions are subjected to
experimental test.

I have to admit that I really can't detect whether you're being sarcastic.
It would be believable either way.

I am serious.

It was obvious to me that you were serious, Richard. And that you were
making an excellent point.

In a scientific discipline, by definition, one does not start from
assumptions, derive conclusions from them, and then assume that those
conclusions must be true. Every experiment is an examination of how
well the conclusion of an argument compares with reality: that is
what an experiment is.

Scientific experiment is also an extremely nice alternative to warfare,
which is the eventual approach taken by those who find proof of their
conclusions in the persuasiveness of their assumptions.

Best regards

Rick

···

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400

54

From[Bill Williams 27 April 2004 12:40 PM CST]

[From Richard Kennaway (2004.04.27 17:30 BST)]

>[From Bill Powers (2004.04.27.1007 MST)]
>
>Richard Kennaway (2004.04.27 16:18 BST)--
>
>>This is in contrast to the usual practice in scientific disciplines,
>>where the consequences of one's assumptions are subjected to
>>experimental test.
>
>I have to admit that I really can't detect whether you're being

sarcastic.

>It would be believable either way.

I am serious. (I prefer to avoid sarcasm and irony altogether: they
invariably obscure communication, as do practically every other
rhetorical device in the book.)

In a scientific discipline, by definition, one does not start from
assumptions, derive conclusions from them, and then assume that those
conclusions must be true. Every experiment is an examination of how
well the conclusion of an argument compares with reality: that is
what an experiment is.

I have to admit that I can't tell what Bill Williams attitude to
assuming the consequences of one's assumptions is, from the message
of his that I quoted, and I'm not imputing any particular attitude to
him.

Richard, and Bill Powers,

I regret having been a possible source of confusion. The over-whelmingly
dominant school in economics based upon maximization doesn't always
speak with one voice, and often the memebers of this school have had
sufficient good sense to remain silent. However, sometimes they let their
guard down and say things that are excessively candid.

Ludwig Von Mises at one point stated that economic theory had been
established upon a foundation that further evidence or reasoning "can
not shake."

If you agree with dear old Ludwig, that labour unions are criminal
conspiricies
and democraticly elected governments are facist, this might be OK.

But the fact that there is a new specialty in economics-- the field of
Experiemental
Economics, ought to be an indicator of something.

Consider a statment made by Lionnell Robbins,

      "The efforts of economists during the last hundred

      and fifty years have resulted in the establishment

      of a body of generalizations whose substantial

      accuracy and importance are open to question only

      by the ignorant or perverse." p. 1.

Robbins, Lionell. 1937 _The Nature and Significance of Economic

     Science_ Macmillian

There has been a quite serious discussion going on here for several

years as to whether there are such things in economics as synthetic

a priori truths. I do not believe in the existence of _a priori_

truths at all. But, these people take Kant quite seriously.

Even Karl Popper had the idea that economics was a science. How he

came to this conclusion will require further research.

The problem has been that it can be genuinely difficult to determine

what, if anything, the economists are saying. And, when Joan Robinson

says, the chief reason for studying economics is so that you can tell

when the economists are lying, a cynic might well decide "Why bother

making the effort?" Maybe, so it might be possible to determine when

we are lying to ourselves? Would that be a valid reason?

Bill Williams

55

From[Bill Williams 27 April 2004 1:20 PM CST]

[From Rick Marken (2004.04.27.1055)]

> Richard Kennaway (2004.04.27 17:30 BST) --
>
>> Bill Powers (2004.04.27.1007 MST)]
>>
>>> Richard Kennaway (2004.04.27 16:18 BST)--
>>
>>> This is in contrast to the usual practice in scientific disciplines,
>>> where the consequences of one's assumptions are subjected to
>>> experimental test.
>>
>> I have to admit that I really can't detect whether you're being

sarcastic.

>> It would be believable either way.
>
> I am serious.

It was obvious to me that you were serious, Richard. And that you were
making an excellent point.

> In a scientific discipline, by definition, one does not start from
> assumptions, derive conclusions from them, and then assume that those
> conclusions must be true. Every experiment is an examination of how
> well the conclusion of an argument compares with reality: that is
> what an experiment is.

Scientific experiment is also an extremely nice alternative to warfare,
which is the eventual approach taken by those who find proof of their
conclusions in the persuasiveness of their assumptions.

Too bad Rick still doesn't see that starting from the assumption that all
economists think the same thing, hasn't been a good starting point for a
discussion of economic theory. But then Rick tends to make some very
obviously implausible assumptions about the world in general regarding
social policy-- such as the one that pedophiles would make good police
officers.

Rick says,

Scientific experiment is also an extremely nice alternative to warfare,

So, I might add is the avoidance of paranoid delusions.

Bill Williams

56

From[Bill Williams 27 April 2004 2:15 AM CST]

[Martin Taylor 2004.04.26.0.22]

I prematurely sent off my partial response to

>From[Bill Williams 26 April 2004 10:20 AM CST]
>
> > [Martin Taylor 2004.04.26.0222]
>
>Is my understand correct that an "optimal orthogonal set of perceptual
>input functions" would generate perceptions free from extraneous
>confusions resulting from something category mistakes.

"Orthogonal" means that each sees what the others don't.

I understood this part.

"Optimal" is

in the information-theoretic sense of concentrating the variance in
the smallest number of dimensions. It says nothing about categories
or mistakes, or even confusions. It just says that the neural
structure won't be any bigger than is needed to deal with the
variation that exists in the incoming sensory data.

I took this to mean that the number of categories used would be the smallest
number of categories required to handle the situation.

Not categories, but dimensions in the sense Peter Small uses the
term. You can specify soemthing as, say, 3 cm high, 5 cm wide, 6 cm
deep, 3/4 red 1/4 white, 0.4 hardness ... For sure, these are
categories to the person who puts labels on them, but for the
orthogonal perceptual input functions, they are just magnitudes, and
there's no guarantee that the PIF would correspond to any particular
category label. That's just soemthing that makes it easy to talk
about them.

If more categories
were used than is neccesary then the same phenomena might be assigned to
either of two categories when one would do. This mistake, if it was noticed,
could be corrected by adopting new rules that would assign the contents of
the two categories to one new category.

When you get into what would be categories to the perceiver, the
usual question is whether an object is or is not a category member.
Is it a bird, is it a plane, ... but it isn't both. But it can be a
bird, a black thing, a corw, a thing that is sitting on the branch
... Those aren't incompatible categories, and they all apply at once
to the same entity. More than is necessary? perhaps, but not more
than is useful in different contexts.

If there were too few categories then two distinct phenomena might be
assigned to the same category.

What determines when two phenomena are distinct? From one viewpoint,
no two situations can ever be the same, so they are all distinct.
Distinctness has to have some relevance to the observer. I think
that's where feedback through action on the environment comes in.

Martin

57

[Martin Taylor 2004.04.27.23.23]

[From Richard Kennaway (2004.04.27 16:18 BST)]

From[Bill Williams 26 April 2004 3:40 PM CST]

When you assume that an economic agent maximizes and the agents goals
are free from interdependencies then as it has been widely claimed it can be
proven that market solutions are best. There is no need to look and any

                                                                  ^^^ at?

evidence.

This is in contrast to the usual practice in scientific disciplines,
where the consequences of one's assumptions are subjected to
experimental test.

"When you assume" is a simple dig at the classical economists against
whom Bill W so often inveighs. He is saying that there exist people
who claim their discpline to be scientific, and so rigorous that one
must know the results to be true without looking at any evidence...
Hah!

But there are such disciplines, which one must nevertheless treat
with respect. We collect them under the rubric of "mathematics."
Presumably "classical" economics belongs there.

Using mathematical disciples, "when we assume" some fact or other,
consequences do appear and need no test against reality. But if what
we assume is intended to match reality, and the hope is that the laws
of the particular branch of mathematics do also, then one gets an
experimental-theoretical science like physics, where the consequences
can be and should be checked against reality.

The problem with economics is that politicians tend to treat as
reality the abstract mathematical results that are based on
assumptions and rules that have failed checks against reality. It's
not a problem with economics as mathematics. It's a problem with
misuse of a branch of mathematics by people who don't understand that
it is only a branch of mathematics and not an
experimental-theoretical science.

Martin

58

[Martin Taylor 2004.04.27.23.32]

[From Bill Powers (2004.04.27.0947 MST);

Kenny Kitzke (2004.04.27)--

Thanks, Kenny. Yes, we disagree about some things, but fortunately there is
some common ground where we can meet.

I think reorganization, as you define it, is a rather rare thing. But, I
am not sure what you suspect. Can you estimate how many times so far in
2004 has your reorganization system been engaged and has successfully
reorganized your hierarchy?

I'd say hundreds if not thousands.

In corridor talk at the 1993 CSG conference, I described what I
called "The Bomb in the Machine". I wrote it up for a presentation
later in 1993, which is now on my Web site. You can see it at
<Bomb in the Hierarchy > The principle of it
was that a reorganizing hierarchy is in a state where a trivial
reorganization at one point in the hierarchy could occur by itself,
or might have consequences in switching other feedback loops from
negative to positive, requiring more reorganization.

Reorganization events, therefore, might be of any magnitude, tiny
ones being orders of magnitude more common than organism-scale
conversion like that of St Paul. Such big ones may not happen in the
lifetime of any one person, but tiny ones may be happening thousands
of times a day, little ones daily, moderate ones a few times a year,
and so forth. One can never tell when a tiny one will induce a huge
one, but it hardly ever happens.

Without intending to say that a neural synaptic reconnection or
disconnection is a reorganization event, because it probably isn't,
just think how many of them occur in an hour or a day. Reorganization
events may well occur at a similar rate.

Martin

59

From[Bill Williams 28 April 2004 12:10]

[Martin Taylor 2004.04.27.23.23]

>[From Richard Kennaway (2004.04.27 16:18 BST)]
>
>>From[Bill Williams 26 April 2004 3:40 PM CST]

"When you assume" is a simple dig at the classical economists against
whom Bill W so often inveighs. He is saying that there exist people
who claim their discpline to be scientific, and so rigorous that one
must know the results to be true without looking at any evidence...
Hah!

Then there is the problem of "evidence." The "evidence" is collected
corrected and reported based upon some considerations that people
"assume" happens to be true. So, it is to some extent possible to say
that what is being called "evidence" is actually theory. When "evidence"
is really theory then it can become confusing.

But there are such disciplines, which one must nevertheless treat
with respect. We collect them under the rubric of "mathematics."
Presumably "classical" economics belongs there.

Using mathematical disciples, "when we assume" some fact or other,
consequences do appear and need no test against reality. But if what
we assume is intended to match reality, and the hope is that the laws
of the particular branch of mathematics do also, then one gets an
experimental-theoretical science like physics, where the consequences
can be and should be checked against reality.

The problem with economics is that

People sometimes assume that there is a single tradition "economics."

politicians tend to treat as

reality the abstract mathematical results that are based on
assumptions and rules that have failed checks against reality.

I think the danger comes from politicians who make up things like
"lockboxes"
which have no reality at all.
It's

not a problem with economics as mathematics. It's a problem with
misuse of a branch of mathematics by people who don't understand that
it is only a branch of mathematics and not an
experimental-theoretical science.

Part of the problem is that many of these people have no idea what an
"experimental-theoretical science" is supposed to do. So to make the
mistake Martin is attributing to them would first required that they learn
what an "experimental-theatrical science" is, then they could go on to
make the mistake of confusing math with science.

Bill Williams

60

[From Rick Marken (2004.04.28.0930)]

Martin Taylor (2004.04.27.23.23) --

The problem with economics is that politicians tend to treat as
reality the abstract mathematical results that are based on
assumptions and rules that have failed checks against reality. It's
not a problem with economics as mathematics. It's a problem with
misuse of a branch of mathematics by people who don't understand that
it is only a branch of mathematics and not an
experimental-theoretical science.

If economics were a branch of mathematics then whether or not its
assumptions and rules have failed checks against reality (perception) would
be irrelevant. The assumptions of mathematics are the untestable axioms from
which theorems (what you call the "abstract mathematical results") are
derived. The correctness of these results is determined, not by checking the
validity of the axioms, but by determining whether these results (theorems)
can be derived from the axioms using the rules of mathematics. This is
called theorem proving and it is the main occupation of mathematicians. If,
therefore, economics were really a branch of mathematics then the
politicians would be right to ignore tests of the validity of its
assumptions and treat the mathematical results of economics as correct (just
as the Pythagorean x^2 = y^2 + z^2 is correct) if those results were
properly derived from the axioms.

But, of course, economics is not a branch of mathematics any more than
physics is. Economics uses mathematics as a model of actual observable
behavior. These models, like the models of mathematics, are built on
assumptions. But the correctness of these models is tested not by theorem
proving, but, as Richard Kennaway (2004.04.27 16:18 BST) says, by subjecting
"the consequences of one's assumptions to experimental test." That is, one
compares the behavior of the model to the behavior of data that the model
purports to explain.

To the extent that economists test their models only by evaluating the
plausibility of the assumptions on which these models are built they are
doing neither mathematics nor science. To the extent that economists test
their models by evaluating the correctness of their derivations, they are
doing mathematics (theorem proving) but not science.

Best regards

Rick

···

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400