[From Rick Marken (931213.1800)]
Avery Andrews (931214.1110) --
Me:
The essential -- crucial -- difference between Chapman and Avery
(now that I look carefully) is that Chapman talks about "deciding
what to do" where Avery talks about "control". "Deciding what to
do" suggests (to me) "selecting an output". So I read Chapman as
saying "output selection is facilitated by perception (knowing
what's in front of you)".
Avery:
I myself wouldn't be so sure that Chapman meant `selecting output'
in the sense that Rick wants to interpret it. If I say `what are
you going to do', and you say `telephone the property manager',
you're not describing the outputs you're going to produce, but
the effect (or, rather, an aspect of the effect, other aspects, such
as what you're going to say, being understood from the context) you
intend to achieve.
You know that and I know that. But does Chapman know that? I admit
that I'm reading an awful lot into a few words, but let's try to
understand Chapman's phrase from your point of view. Chapman said:
`deciding what to do is easy when you know what's in front of you'.
I presume that the phrase "what's in front of you" refers to the state
of a perceptual variable. You say "deciding what to do" was understood
by Chapman to mean "deciding what (perceptual) effect to produce" (it
has to be a perceptual effect -- if it's just an effect on the environment
then we are talking about a generated output). So now we can restate
Chapman as follows:
'deciding what perception (state of a perceptual variable) to produce is
easy when you know the state of the perceptual variable'
Bill Powers (931213.0930 MST) must have heard the phrase this way
because he explained the problem with it; perceptual variables don't
specify the states in which they should be maintained (unless they are
the kind of perceptions with information in them); perceptions just ARE;
control systems specify the intended level of a perceptual variable;
ie. the perception to be "produced"; the perceptions themselves can't
do that, of course.
Avery:
This is how `do this/that/what/something'
constructions are understood in ordinary English, and I see no
reason for interpreting the Chapman statement by any other scheme.
And I beleive that I have shown that even if you interpret the "do
this" construction" as "produce this perceptual result" Chapman
is still not in the ballpark. Face it, Avery, you've got PCT
runnin' in your veins now; your restatement of Chapman may have
been an attempt to show that a nonPCTer can understand the gist of
PCT, but in the process you showed just the opposite. Chapman's
statement only makes sense if you make precisely the little
alterations in it that you made. You said:
'PCT says that control will be easier to achieve if the controller has
the right perceptions'
This restatement of Chapman makes sense because "control" is "stabilization
of some aspect of the environment"; "easier to achieve" means (to me)
easier than calculating inverse kinematics; "right perceptions" are
perceptual signals that are a function of the aspect of the environment
that is to be stabilized (the controlled variable). So you are saying
something quite sensible:
'control is easier (and usually only possible) when the perceptual
signal is a function of the variable to be controlled'.
Leave Chapman to his confusion.
Best
Rick