[Martin Taylor 2017.05.28.14.41]
[From Rick Marken (2017.05.28.1045)]
If you are talking about Hans Blom, I don't think he was at all
opposed to PCT. What he was opposed to was a voluble and dogmatic
refusal to permit expression of the thought that it is appropriate
to use with PCT some mathematical tools that he believed could be
useful. His opposition was to dogmatism, not PCT.
As has been the case, to my first-hand knowledge, for several others
who tried to use CSGnet to improve their understanding of PCT or to
offer different viewpoints on it, but left in frustration over the
same dogmatic opposition to looking at PCT in anything but the
officially approved way. None were “opposed to PCT”, or they
wouldn’t have joined CSGnet and tried to contribute. People come to
CSGnet presumably to learn and possibly to contribute. They don’t
come to “oppose” PCT, but all too often their attempt to contribute
have been met not with reasoned analysis of why the contribution is
misguided, but with the moral equivalent of “We don’t want your
nasty ideas here. Forget them or go away.” I think, for example, of
Peter Small, who was interested in the dynamical implications of
perceptual control, but was sneeringly blown off every time he tried
to explain what he was trying to get across.
I realize it must be difficult for you to allow the possibility that
people might usefully apply to PCT tools you don’t understand, but
it’s no benefit to the future development of PCT to treat them the
way such people have been treated on CSGnet (and I include myself as
one who has persevered in CSGnet despite dogmatic opposition to
using common mathematical tools).
Over the last decade or two, I have refrained from posting to CSGnet
a lot of interesting aspects of PCT that I have worked on – most
recently, for example, the tensegrity properties of the hierarchy
that give it resilience – because of my expectation of being
subjected to mindless criticism, which basically comes down to “It
isn’t PCT, because I don’t understand it”. I have, however,
discussed many of them with small groups of researchers interested
in PCT who also don’t want to talk about interesting novel points on
CSGnet.
Quite recently you told Bruce and me that we were S-R behaviourists
opposed to PCT because we pointed out the mathematical fallacy in
your supposed demonstration that the curvature power law was a
behavioural illusion. Suggesting you made a mistake in maths was
sufficient to declare us to be opposed to PCT – like Hans Blom,
whose suggestion that some mathematics you didn’t understand might
be useful in analyzing control systems was enough for you to say he
was “a vigorous opponent” of PCT. “Opposed to PCT” appears to be a
synonym for “questioning Rick Marken”.
SM:
I would argue that the engineering student would be the
easiest to convince because he understands how the math
works underneath the theory, he has no doubts about the
theoretical foundation.
I agree. I guess I am further evidence against Sean's thesis,
because I nearly became a graduate student in control engineering,
and you frequently call me an opponent of PCT. So is Bill Powers,
come to that, since control engineering was part of his background.
I guess you would have predicted that he could never “get” PCT.
Control engineers obviously are not the kind of people one should
try to attract. They might try to use what they know in analyzing
the behaviour of hierarchic feedback systems. Since Official PCT
already understands all there is to know about that, their efforts
would necessarily be opposed to PCT, by definition.
I think you need both. They aren't opposites. They are complements.
To rebuild a house that is perfectly livable, but perhaps a little
rickety, requires the flexibility to see how a redesign could
improve its livability while avoiding more of the patch jobs you
have for a long time been doing to keep the house standing. Along
with that flexibility, perseverance is required, to learn all the
building codes to which the new house must conform, not to mention
the techniques needed to design and construct it when you have
discovered some of the possibilities of the site. If you know
something of carpentry, you don’t throw that away when you are
framing your new house. The real problems arise, however, when you
have a friend that says “houses must be built exactly like this” and
is inflexible in insisting that no other kind of house or
construction technique is admissible, no matter how useful the
technique or how well built the resulting house.
I would call that, as Sean does, "unlearning".
Not everyone's understanding of PCT is the same, even among those
who do not get it wrong. Fred’s understanding seems correct to me,
but there’s lots about his application that I don’t understand, and
lots about the mathematics and dynamics of control that he doesn’t
understand. Bill not infrequently let us know that there was a lot
about PCT he didn’t understand, and he thought it might take decades
or centuries of patient “persevering” research to find all its kinks
and hidden possibilities. Is there anyone who claims to really
understand PCT? Perhaps there is – just one!
PCT is dead simple at heart. Where it gets complicated and difficult
is when you get into the details of the behaviour of interlinked
feedback systems. For example, how does the tolerance zone affect
the speed with which a control loop corrects for a step change in
the disturbance? That’s not easy to analyze, and complex to
parameterize by a series of experiments. That’s a simple question.
Others are more complicated. But the basic idea is so simple that
one of the biggest difficulties I had when I first learned about PCT
was that I couldn’t believe it was as simple as it is. That may be a
characteristic of powerful theories; they have to be simple, because
any complexity would indicate a dependence on special circumstances,
limiting the range of their application.
Feedback systems are not obvious to the general public. For example,
everyone want lower taxes, but they don’t want to lose what those
taxes pay for because if they lost those, they would have less money
available to pay taxes. It’s a feedback loop, but people act as
though money paid in taxes vanishes from the economy, and their
success in having a good lifestyle is due to their own hard work –
and so they vote for politicians who advocate for lower taxes (but
never mention the resulting impoverishment of living conditions for
everyone except the rich).
That's just one example. It's not hard to find examples where people
think of something as cause and effect without noticing that the
“effect” is also a cause of the “cause”. It is very hard for most
people to understand that higher taxes just might mean less total
expenditure and a better quality of life. That’s because they can’t
imagine the effects of feedback loops.
I would hope it would be more scientifically accurate than your
“peer reviewed” publication on the power law effect of curvature. By
the way, have you yet asked the editor of that journal to get
someone mathematically capable to check the specific point where the
problem occurs – the confusion between derivatives with respect to
time and derivatives with respect to distance?
Martin
···
RM: In the early years of CSGNet (early 1990s) we had
a control engineer join the discussion. He became one of
the most vigorous opponents of PCT.
RM. About 10 years ago Bill and I engaged with some
control engineers, demonstrating how a hierarchical
control system could account for the apparent
“adaptation” that was observed to changes in the order
of the control loop (proportional to integral, for
example). The control engineers vigorously opposed the
elegant PCT solution, which, by the way, became Chapter
5 in Powers’ Living Control Systems III: The Fact of
Control. I think the evidence, limited as it is, argues
strongly against your prediction.
RM: Some
individuals (like me) have been found to be more
convincable regarding PCT than others but those
individuals are few and far between.
SM: Th at’s
because you are an aberration mate. A special
case. PCT is complex, you may have forgotten
because you are an old hand now, but for a naive
learner it is a very challenging theory to pick
up. It requires unlearning a lot. You have to
have a high level of motivation to persevere.
RM: In my experience, it's not perseverance but just
the opposite, maybe call it flexibility, that allows
some individuals to “get” PCT.
Everyone I know who gets PCT (and there are not
very many of them) gets it, not because of a willingness
to persevere but because of a willingness to abandon all
their cherished preconceptions that conflicted with an
understanding of PCT.
SM: If the goal is
understanding, then PCT creates high levels of
disturbance. When a novice population at any
discipline meets a high level of disturbance,
you get massive drop out rates. They cannot
control the disturbances so they cope using
avoidance and then make up rationalizations to
reduce that avoidance error.
RM: Avoidance is a way to protect existing ideas from
the effects of disturbance.
SM: I believe the low
numbers using PCT are not due to the theory, or
intelligence, they are due to the sharp
learning curve.
RM: I think that's highly likely. PCT is no more
difficult to learn than, say, computer programming.
RM: I think it would be great to have such a course.
… It would have to be scientifically accurate so a
small board of experts on PCT have to have final say on
the content.