Environmental variables (was: Why we fight...)

[Martin Taylor 2014.08.22.22.57]

The problem with this, and it is a bone I have picked with Bill as

well, is that the same is true of EVERY perception. Bill seemed to
take an internally inconsistent view, saying on the one hand that
all we ever know is our perception, no matter what kind of
perception it might be, and on the other hand that there were some
privileged environmental variables that are “real” and some that are
not. Since all perceptions are equally constructed by the operation
of layers of perceptual functions on the firing patterns of neurons
connected to sensors, on what grounds can one say a perception such
as “intensity” is of an environmental variable more real than the
taste of lemonade? The argument that the taste of a given glass of lemonade is
different for every person cannot hold water :slight_smile: There is no way
that we can say that any perception in one person is the same as a
perception in someone else. My perception of “blue” may not be
yours, just as my perception of the trustworthiness of the
neighbourhood loan shark or municipal politician may not be yours.
My preferred position on this is, I think, more consistent than
Bill’s. I assume that there exists a “real world” that influences my
sense organs and on which I can act in ways that influence the
inputs to my sense organs. I know not what is in that world, but my
nervous system constructs a perceived reality by operations on those
sensory inputs (plus all the history of operations on earlier
sensory inputs). Every perception thus produced defines some
function of present and past states of the environment that have and
had some influence on my sensors. Those functions describe a
presumed reality “out there”, no one function, no one presumed
environmental variable thus constructed having any priority over any
other.
In this, I believe internally consistent, view, the brightness of a
light, the taste of lemonade, the willingness of my dog to follow
me, the autocratic tendencies of my political leaders, all have
equal validity as existing states of the external environment. I
call them Complex Environmental Variables, because they are
environmental variables defined by complex perceptual functions.
There’s no way to say that second, fifth, or ninth-level perceptions
are of environmental variables more or less real than each other.
As for the taste of lemonade having no special physical effects on
anything else, except the person tasting the mixture, that is true
of every uncontrolled perception. If the person is not controlling
the perception of the taste of the lemonade, any taste will do. But
if the taste is a controlled perception and it doesn’t match its
reference value, the person will act, and those actions definitely
will have effects on something else, perhaps only by adding sugar to
the mixture, perhaps by throwing the glass at the poor waiter who
brought this foul tasting mixture.
I have disagreed with Bill on this for a couple of decades without
changing his opinion, and I don’t imagine my argument will change
anyone else’s opinion. All the same, I thought it worth putting “out
there” for the consideration of anyone who might be interested.
Martin

···
        [From

Rick Marken (2014.08.22.1940)]

            Martin Taylor

(2014.08.22.14.41)–

                        RM:  I think I mistakenly read this to

mean that you were saying that control also
involves reducing the variability of the external
environment. But, in fact, you never said
that; you were correctly implying that
control involves reducing the variability of
perceptual aspects of the environment ,
keeping them from varying around their
“desired states”.

            MT: That's true, but as you often point out, controlling

a perceptual (internal) variable also implies reducing
the variability of the corresponding environmental
variable.

          RM: I would rather say that controlling a perceptual

variable implies controlling the environmental correlate
of the perceptual variable. The reason for this slight
change in terminology is to emphasize the fact that there
may be no environmental variable that corresponds to the
controlled perceptual variable. For example, in B:CP (p.
112 of 2nd edition) Bill gives the example of controlling
for the perception of the taste of lemonade. That
perception is a construction (by a perceptual function)
“…derived from the intensity signals generated by sugar
and acid (together with some oil smells)… the mere
intermingling of these physical components has no special
physical effects on anything else, except the person
tasting the mixture.” The same is true of many other
perceptions that we control, such as the perception of
variability (if we can perceive it). The variability of
independent events that happen with different
probabilities is an aspect of these events that we can
perceive but it has no more physical significance than
some other aspect of these events that we can perceive,
such Morse code patterns.

[Philip 8/22/14 9:09]

Martin Taylor:

There is no way that we can say that any perception in one person is the same as a perception in someone else. My perception of “blue” may not be yours…

Philip:

In the realm of colors, perhaps there’s a way to ambiguate the name of a color from the experience of the color itself.

But if we both pointed to, say, the 6-inch, or perhaps the 7.416407865-inch mark on a ruler? What is the experience of a number?

···

On Fri, Aug 22, 2014 at 8:26 PM, Martin Taylor mmt-csg@mmtaylor.net wrote:

[Martin Taylor 2014.08.22.22.57]

        [From

Rick Marken (2014.08.22.1940)]

The problem with this, and it is a bone I have picked with Bill as

well, is that the same is true of EVERY perception. Bill seemed to
take an internally inconsistent view, saying on the one hand that
all we ever know is our perception, no matter what kind of
perception it might be, and on the other hand that there were some
privileged environmental variables that are “real” and some that are
not. Since all perceptions are equally constructed by the operation
of layers of perceptual functions on the firing patterns of neurons
connected to sensors, on what grounds can one say a perception such
as “intensity” is of an environmental variable more real than the
taste of lemonade?

The argument that the taste of a given glass of lemonade is

different for every person cannot hold water :slight_smile: There is no way
that we can say that any perception in one person is the same as a
perception in someone else. My perception of “blue” may not be
yours, just as my perception of the trustworthiness of the
neighbourhood loan shark or municipal politician may not be yours.

My preferred position on this is, I think, more consistent than

Bill’s. I assume that there exists a “real world” that influences my
sense organs and on which I can act in ways that influence the
inputs to my sense organs. I know not what is in that world, but my
nervous system constructs a perceived reality by operations on those
sensory inputs (plus all the history of operations on earlier
sensory inputs). Every perception thus produced defines some
function of present and past states of the environment that have and
had some influence on my sensors. Those functions describe a
presumed reality “out there”, no one function, no one presumed
environmental variable thus constructed having any priority over any
other.

In this, I believe internally consistent, view, the brightness of a

light, the taste of lemonade, the willingness of my dog to follow
me, the autocratic tendencies of my political leaders, all have
equal validity as existing states of the external environment. I
call them Complex Environmental Variables, because they are
environmental variables defined by complex perceptual functions.
There’s no way to say that second, fifth, or ninth-level perceptions
are of environmental variables more or less real than each other.

As for the taste of lemonade having no special physical effects on

anything else, except the person tasting the mixture, that is true
of every uncontrolled perception. If the person is not controlling
the perception of the taste of the lemonade, any taste will do. But
if the taste is a controlled perception and it doesn’t match its
reference value, the person will act, and those actions definitely
will have effects on something else, perhaps only by adding sugar to
the mixture, perhaps by throwing the glass at the poor waiter who
brought this foul tasting mixture.

I have disagreed with Bill on this for a couple of decades without

changing his opinion, and I don’t imagine my argument will change
anyone else’s opinion. All the same, I thought it worth putting “out
there” for the consideration of anyone who might be interested.

Martin
            Martin Taylor

(2014.08.22.14.41)–

            MT: That's true, but as you often point out, controlling

a perceptual (internal) variable also implies reducing
the variability of the corresponding environmental
variable.

          RM: I would rather say that controlling a perceptual

variable implies controlling the environmental correlate
of the perceptual variable. The reason for this slight
change in terminology is to emphasize the fact that there
may be no environmental variable that corresponds to the
controlled perceptual variable. For example, in B:CP (p.
112 of 2nd edition) Bill gives the example of controlling
for the perception of the taste of lemonade. That
perception is a construction (by a perceptual function)
“…derived from the intensity signals generated by sugar
and acid (together with some oil smells)… the mere
intermingling of these physical components has no special
physical effects on anything else, except the person
tasting the mixture.” The same is true of many other
perceptions that we control, such as the perception of
variability (if we can perceive it). The variability of
independent events that happen with different
probabilities is an aspect of these events that we can
perceive but it has no more physical significance than
some other aspect of these events that we can perceive,
such Morse code patterns.

                        RM:  I think I mistakenly read this to

mean that you were saying that control also
involves reducing the variability of the external
environment. But, in fact, you never said
that; you were correctly implying that
control involves reducing the variability of
perceptual aspects of the environment ,
keeping them from varying around their
“desired states”.

Martin,

I’m not a good reader of your »English«, so I don’t know whether I understood some thoughts of yours, as you wanted to be understaood.

MT :

The problem with this, and it is a bone I have picked with Bill as well, is that the same is true of EVERY perception. Bill seemed to take an internally inconsistent view, saying on the one hand that all we ever know is our perception, no matter what kind of perception it might be, and on the other hand that there were some privileged environmental variables that are “real” and some that are not.

HB :

Where did Bill wrote that some privileged environmental variables are more real than others ? I really don’t recall »meating« this in his books. But I’m not some specialy good »English« reader. It’s possible that I missed something.

MT :

The argument that the taste of a given glass of lemonade is different for every person cannot hold water :slight_smile: There is no way that we can say that any perception in one person is the same as a perception in someone else. My perception of “blue” may not be yours, just as my perception of the trustworthiness of the neighbourhood loan shark or municipal politician may not be yours.

HB :

I really don’t see the point. In the first sentence it seems to me, that you are saying »that the taste of a given glass of lemonade is different for every person cannot hold water« so I understand it, that different persons have the same perception of taste (can not »hold water«). It seems to me that you are denying the existance of different taste of lemonade. In the other senstences I understand that perceptions are different for any person. Did I misssed something ?

MT :

My preferred position on this is, I think, more consistent than Bill’s. I assume that there exists a “real world” that influences my sense organs and on which I can act in ways that influence the inputs to my sense organs. I know not what is in that world, but my nervous system constructs a perceived reality by operations on those sensory inputs (plus all the history of operations on earlier sensory inputs). Every perception thus produced defines some function of present and past states of the environment that have and had some influence on my sensors. Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

HB :

I never felt that Bill had different opinion from yours. But he did once have a discours of »solipsism« but he asked himself whether his thinking was »solipsist« view or not. Well whatever I never have a feeling that he is solipsist or something different from your perceptual understanding of »real world«.

MT :

In this, I believe internally consistent, view, the brightness of a light, the taste of lemonade, the willingness of my dog to follow me, the autocratic tendencies of my political leaders, all have equal validity as existing states of the external environment. I call them Complex Environmental Variables, because they are environmental variables defined by complex perceptual functions. There’s no way to say that second, fifth, or ninth-level perceptions are of environmental variables more or less real than each other.

HB :

I really don’t imagine what »Complex Environemntal Variable« (CEV) can be ? If I assume that intensities as first level controlled variables which »integrate« into second level »sensations« and so on through the whole heirarchy, what is CEV ? Maybe your »perceptual mosaic« of the world which you assume to be »outside«. As you said. You beleive that your view is internally consistent. Others beleive that Bill’s view or their view is internally consistent. How can we know which »view« is internally consistent ? But it seems interesting what’s your theory  of how we perceive the »world« outside. How does »Complex Perceptual Functions« (CPF) work ?

MT :

Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

HB :

So if I choose to control some variable in »environment«, are you saying that, it has no priority over other perceived variables in »environment« in that moment ? We can also decide not to control any perceived variable in »external environemnt«, because we control only »internal« environmental variables. In this case maybe »presumed external environmental variables« have no prioority. Anyway I have quite a problem with understanding this view.

MT :

As for the taste of lemonade having no special physical effects on anything else, except the person tasting the mixture, that is true of every uncontrolled perception.

HB :

If I recall right you said that Bill’s model assume only controlled perceptions in the diagram. So where do »uncontrolled perceptions« go in your model ?

MT :

Since all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors,…

HB :

Do I understand right that all perceptions (controlled and uncontrolled) are equally constructed in nervous system ? So I suppose that we can assume that all perceptions go though the same perceptual »chanell« what is from physilogical view right. So how do we distinguish the pathway of »controlled and uncontrolled« perceptions in your theory ?

MT :

I have disagreed with Bill on this for a couple of decades without changing his opinion, and I don’t imagine my argument will change anyone else’s opinion. All the same, I thought it worth putting “out there” for the consideration of anyone who might be interested.

HB:

I’m always interested. But as far as I was talking to Bill and read his books I really don’t understand what was your and Bill’s disagreement ? It seems to me a little bit confusing. I think I always understood Bill’s theory and your understanding of it as something very similar, specialy as you frquently pointed out, that you talked with Bill for days and hours and that he »filled« some of your diagrams with his knowledge.

Boris

···

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Saturday, August 23, 2014 5:26 AM
To: csgnet@lists.illinois.edu
Subject: Environmental variables (was: Why we fight…)

[Martin Taylor 2014.08.22.22.57]

[From Rick Marken (2014.08.22.1940)]

Martin Taylor (2014.08.22.14.41)–

RM: I think I mistakenly read this to mean that you were saying that control also involves reducing the variability of the external environment. But, in fact, you never said that; you were correctly implying that control involves reducing the variability of perceptual aspects of the environment, keeping them from varying around their “desired states”.

MT: That’s true, but as you often point out, controlling a perceptual (internal) variable also implies reducing the variability of the corresponding environmental variable.

RM: I would rather say that controlling a perceptual variable implies controlling the environmental correlate of the perceptual variable. The reason for this slight change in terminology is to emphasize the fact that there may be no environmental variable that corresponds to the controlled perceptual variable. For example, in B:CP (p. 112 of 2nd edition) Bill gives the example of controlling for the perception of the taste of lemonade. That perception is a construction (by a perceptual function) “…derived from the intensity signals generated by sugar and acid (together with some oil smells)… the mere intermingling of these physical components has no special physical effects on anything else, except the person tasting the mixture.” The same is true of many other perceptions that we control, such as the perception of variability (if we can perceive it). The variability of independent events that happen with different probabilities is an aspect of these events that we can perceive but it has no more physical significance than some other aspect of these events that we can perceive, such Morse code patterns.

The problem with this, and it is a bone I have picked with Bill as well, is that the same is true of EVERY perception. Bill seemed to take an internally inconsistent view, saying on the one hand that all we ever know is our perception, no matter what kind of perception it might be, and on the other hand that there were some privileged environmental variables that are “real” and some that are not. Since all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors, on what grounds can one say a perception such as “intensity” is of an environmental variable more real than the taste of lemonade?

The argument that the taste of a given glass of lemonade is different for every person cannot hold water :slight_smile: There is no way that we can say that any perception in one person is the same as a perception in someone else. My perception of “blue” may not be yours, just as my perception of the trustworthiness of the neighbourhood loan shark or municipal politician may not be yours.

My preferred position on this is, I think, more consistent than Bill’s. I assume that there exists a “real world” that influences my sense organs and on which I can act in ways that influence the inputs to my sense organs. I know not what is in that world, but my nervous system constructs a perceived reality by operations on those sensory inputs (plus all the history of operations on earlier sensory inputs). Every perception thus produced defines some function of present and past states of the environment that have and had some influence on my sensors. Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

In this, I believe internally consistent, view, the brightness of a light, the taste of lemonade, the willingness of my dog to follow me, the autocratic tendencies of my political leaders, all have equal validity as existing states of the external environment. I call them Complex Environmental Variables, because they are environmental variables defined by complex perceptual functions. There’s no way to say that second, fifth, or ninth-level perceptions are of environmental variables more or less real than each other.

As for the taste of lemonade having no special physical effects on anything else, except the person tasting the mixture, that is true of every uncontrolled perception. If the person is not controlling the perception of the taste of the lemonade, any taste will do. But if the taste is a controlled perception and it doesn’t match its reference value, the person will act, and those actions definitely will have effects on something else, perhaps only by adding sugar to the mixture, perhaps by throwing the glass at the poor waiter who brought this foul tasting mixture.

I have disagreed with Bill on this for a couple of decades without changing his opinion, and I don’t imagine my argument will change anyone else’s opinion. All the same, I thought it worth putting “out there” for the consideration of anyone who might be interested.

Martin

[From Rick Marken (2014.08.23.1900)]

Martin Taylor (2014.08.22.22.57)--

RM: I would rather say that controlling a perceptual variable implies controlling the environmental correlate of the perceptual variable....

MT: The problem with this, and it is a bone I have picked with Bill as well, is that the same is true of EVERY perception. Bill seemed to take an internally inconsistent view, saying on the one hand that all we ever know is our perception, no matter what kind of perception it might be, and on the other hand that there were some privileged environmental variables that are "real" and some that are not.

RM: I think Bill's point was that the environment in PCT is simply the current model of physics and chemistry. This is the model of what is on the "other side" of our sensory receptors. The sensory receptors are the first step in the perceptual process. The physics model of the environment does suggest that the sensory receptors transduce certain physical variables into neural impulses; for example, the cone cells in the retina seem to transduce electromagnetic energy of a particular frequency, into neural impulses. Perhaps this is what you think of as the "privileged" physical variables; the variables that are in the physics model.

MT: Since all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors, on what grounds can one say a perception such as "intensity" is of an environmental variable more real than the taste of lemonade?

RM: I don't think it's that some perceptions are considered "more real" than others. It's that some perceptions -- the intensity level perceptions -- are assumed to be a direct function of the physical (environmental) variables that are part of the physics/chemistry model: rods and cones transduce electromagnetic energy, hair cells transduce mechanical energy, olfactory sensory cells transduce chemical energy, etc. So it's true that the lowest level sensory receptors are assumed to produce outputs that are directly proportional to the amount of environmental energy impinging on the cells; so the outputs of these lowest level sensory receptors can be considered perceptions that are proportional to the "real" level of energy impinging on them. But that "real" level of energy is, itself, a model -- part of the current physics model, so it's really not correct to call it "real".

MT: The argument that the taste of a given glass of lemonade is different for every person cannot hold water :slight_smile:

RM: I don't think that argument is being made. The argument is simply that the taste of lemonade is a construction -- a function of lower level perceptual signals. Whether a perception is constructed the same way by all individuals is not relevant; just that it is constructed. And even if it were constructed the same way by all individuals it will never be know whether the experience of this perception is the same for everyone. All PCT says is that perceptions above the level of the intensity perceptions are constructions -- functions of lower level perceptual signals. I suppose these constructions could be considered a "map" of external reality; what is constructed certainly depends on the "raw materials" provided by the variations in the first order intensity signals. But there are, in principle, many different equally valid maps that could be constructed from the same raw intensity signal variations.
RM: According to PCT, the perceptions that are constructed are those that result in the best control. An excellent description of how this works is given by Bill in a paper called PCT and Engineering Control Theory. It comes along with a demo program. I have put both the paper and program in a folder and am making them available through dropbox at:
<https://www.dropbox.com/sh/bgdztitng5uohqm/AACg7DzKk_ZfSKb0Upoq_hzba?dl=0&gt;https://www.dropbox.com/sh/bgdztitng5uohqm/AACg7DzKk_ZfSKb0Upoq_hzba?dl=0

The whole paper is great but the most relevant section to what we're discussing here is near the end called A DEMONSTRATION OF MULTIDIMENSIONAL CONTROL. This section describes what the the Delphi program does, which is to learn how to construct up to 500 control systems, each controlling a weighted vector of the same 500 scalar environmental inputs. Basically the system learns to construct perceptual functions of these environmental variables that allows each of the control systems to simultaneously control it's perception without interfering with the other systems' ability to control their own perceptual constructions based on the same 500 variables.

MT: My preferred position on this is, I think, more consistent than Bill's. I assume that there exists a "real world" that influences my sense organs and on which I can act in ways that influence the inputs to my sense organs. I know not what is in that world, but my nervous system constructs a perceived reality by operations on those sensory inputs (plus all the history of operations on earlier sensory inputs). Every perception thus produced defines some function of present and past states of the environment that have and had some influence on my sensors. Those functions describe a presumed reality "out there", no one function, no one presumed environmental variable thus constructed having any priority over any other.

RM: That is an excellent description of my own way of looking at it!

MT: In this, I believe internally consistent, view, the brightness of a light, the taste of lemonade, the willingness of my dog to follow me, the autocratic tendencies of my political leaders, all have equal validity as existing states of the external environment.

RM: I would say that they all have equal validity as _functions_ of existing states of the external environment -- the states of the intensity signals that are proportional to the environmental energy fluctuations at your sense receptors. The validity of a perception is defined in PCT by the controllability of that perception in the context of all the other perceptions that you are controlling.

MT: I call them Complex Environmental Variables, because they are environmental variables defined by complex perceptual functions. There's no way to say that second, fifth, or ninth-level perceptions are of environmental variables more or less real than each other.

RM: The only problem here is that any particular CEV exists only because it is defined by a perceptual function. Many CEVs, such as the taste of lemonade or the oppressiveness of the weather (the combinatoin of heat and humidity called the "discomfort index" and discussed by Bill on pp. 7 and 8 of the PCT and Engineering paper above) have no "physical significance; that is, they exist only as mathematical functions of "real" (according to the physics model) physical variables. So I would rather refer to the environmental correlates of these perceptions as aspects of the environment rather than as variables in the environment to make it clear that a perception defines an aspect of the environment that has significance for the control system but not necessarily anything "out there" of "real" physical significance.

MT: I have disagreed with Bill on this for a couple of decades without changing his opinion, and I don't imagine my argument will change anyone else's opinion. All the same, I thought it worth putting "out there" for the consideration of anyone who might be interested.

RM: Well, thank you for doing so. I always profit from discussions with you. I learned from this one that we are probably closer than you might think on the perception/reality issue. But most important you led me to the PCT and Engineering paper and the accompanying program (which I think Bill sent to me in about 2007) and I now have a better idea of what he was trying to show. The paper is printed in the format Dag uses for his work so I wonder if that paper (and the program) has already been published or is available at Dag's website. I couldn't find it there but maybe Dag could tell us more about that.
Best regards
Rick

···

--
Richard S. Marken, Ph.D.
Author of <http://www.amazon.com/Doing-Research-Purpose-Experimental-Psychology/dp/0944337554/ref=sr_1_1?ie=UTF8&qid=1407342866&sr=8-1&keywords=doing+research+on+purpose&gt;Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[Martin Taylor 2014.08.23.23.52]

Rick references B:CP, 2nd Ed. p 112 as an example. On rereading it,

I can see that the writing is ambiguous. I read it (as Rick seems to
have done) as saying that there are environmental variable
corresponding to first-level perceptions, but not to higher-level
perceptions, which are constructed.
No. I’m afraid my English was not clear enough. It’s “the argument”
(for the non-existence in the environment of a variable
corresponding to the taste of lemonade) that cannot hold water. I
made no claim about whether different people, or the same person at
different times, would or would not perceive the same taste from a
particular glass of lemonade.
Solipsism doesn’t come into this discussion, except that we have to
dismiss it before we can have a discussion that includes talk about
an external environment. Basically, it’s totally irrelevant. We
agree it’s wrong, and then deal with our theories.
Every perceptual function defines a CEV. If you happen to have a
perceptual function that produces an output proportional to 3 times
the temperature in degrees Kelvin plus twice the number of minutes
since last midnight, then 3*(deg K)+2*(minutes) is the CEV for that
perception. It’s a nonsense one, of course, but nonsense sometimes
is useful to illuminate sense.
Are not all perceptual functions complex? I don’t mean in the
mathematical sense of using numbers with real and imaginary
components, but in the everyday sense of being rather complicated.
It matters not whether you choose to control a perception. Any
function you define creates a value. That value is a variable whose
magnitude depends on the values of the inputs to the function.
Perceptual functions are no different. If you have a perceptual
function, by definition its output is a perceptual signal that
changes when the environmental variables that enter into it change.
No. I have often said to you that Bill’s DIAGRAMs contain only
controlled perceptions, because they are diagrams of how those
perceptions are controlled. Uncontrolled perceptions have values
that may or may not contribute to the values of controlled
perceptions. It’s irrelevant in a diagram intended to show how
control works. I don’t have any particular model for uncontrolled
perceptions, nor do I think one is needed. You don’t. The only difference is whether at any particular moment
the perception is being controlled – whether you are actively doing
something to keep it near its reference value, if it has one.
I have always believed that my understanding of Bill’s theory was
very close to his. Our differences, as I perceived them, were always
on the margins. For example, I have never been convinced by the
specific levels Bill describes in his hierarchy. Neither was he, to
judge from the way he inverted the order of levels in his models,
and the number of times he explicitly said that his set of levels
was always subject to revision. There have been many times I have
disagreed with Bill and sometimes one of us changed our mind, other
times when we agreed to disagree. But on the big things, I don’t
think we disagreed much, if at all. This issue of whether all or
only some environmental variable are constructed by perceptual
functions is a very little thing.
Martin

···

On 2014/08/23 10:36 AM, Boris Hartman
wrote:

Martin,

Â

        I'm

not a good reader of your »English«, so I don’t know whether
I understood some thoughts of yours, as you wanted to be
understaood.

Â

        MT

:

      The problem with this, and it is a bone I

have picked with Bill as well, is that the same is true of
EVERY perception. Bill seemed to take an internally
inconsistent view, saying on the one hand that all we ever
know is our perception, no matter what kind of perception it
might be, and on the other hand that there were some
privileged environmental variables that are “real” and some
that are not.

Â

        HB

:

        Where

did Bill wrote that some privileged environmental variables
are more real than others ? I really don’t recall »meating«
this in his books. But I’m not some specialy good »English«
reader. It’s possible that I missed something.

Â

        MT

:

      The argument that the taste of a given

glass of lemonade is different for every person cannot hold
water :slight_smile: There is no way that we can say that any perception
in one person is the same as a perception in someone else. My
perception of “blue” may not be yours, just as my perception
of the trustworthiness of the neighbourhood loan shark or
municipal politician may not be yours.

Â

        HB

:

        I

really don’t see the point. In the first sentence it seems
to me, that you are saying » that the taste of a given
glass of lemonade is different for every person cannot hold
water« so I understand it, that
different persons have the same perception of taste (can
not »hold water«). It seems to me that you are denying the
existance of different taste of lemonade. In the other
senstences I understand that perceptions are different for
any person. Did I misssed something ?

Â

        MT

:

      My preferred position on this is, I think,

more consistent than Bill’s. I assume that there exists a
“real world” that influences my sense organs and on which I
can act in ways that influence the inputs to my sense organs.
I know not what is in that world, but my nervous system
constructs a perceived reality by operations on those sensory
inputs (plus all the history of operations on earlier sensory
inputs). Every perception thus produced defines some function
of present and past states of the environment that have and
had some influence on my sensors. Those functions describe a
presumed reality “out there”, no one function, no one presumed
environmental variable thus constructed having any priority
over any other.

Â

        HB

:

        I

never felt that Bill had different opinion from yours. But
he did once have a discours of »solipsism« but he asked
himself whether his thinking was »solipsist« view or not.
Well whatever I never have a feeling that he is solipsist
or something different from your perceptual understanding of
»real world«.

Â

        MT

:

      In this, I believe internally consistent,

view, the brightness of a light, the taste of lemonade, the
willingness of my dog to follow me, the autocratic tendencies
of my political leaders, all have equal validity as existing
states of the external environment. I call them Complex
Environmental Variables, because they are environmental
variables defined by complex perceptual functions. There’s no
way to say that second, fifth, or ninth-level perceptions are
of environmental variables more or less real than each other.

Â

        HB

:

        I

really don’t imagine what »Complex Environemntal Variable«
(CEV) can be ? If I assume that intensities as first level
controlled variables which »integrate« into second level
»sensations« and so on through the whole heirarchy, what is
CEV ?

        Maybe

your »perceptual mosaic« of the world which you assume to be
»outside«. As you said. You beleive that your view is
internally consistent. Others beleive that Bill’s view or
their view is internally consistent. How can we know which
»view« is internally consistent ? But it seems interesting
what’s your theory  of how we perceive the »world« outside.
How does »Complex Perceptual Functions« (CPF) work ?

Â

        MT

:

      Those functions describe a presumed reality

“out there”, no one function, no one presumed environmental
variable thus constructed having any priority over any other.

Â

        HB

:

        So

if I choose to control some variable in »environment«, are
you saying that, it has no priority over other perceived
variables in »environment« in that moment ?

Â

        MT

:

      As for the taste of lemonade having no

special physical effects on anything else, except the person
tasting the mixture, that is true of every uncontrolled
perception.

Â

        HB

:

        If

I recall right you said that Bill’s model assume only
controlled perceptions in the diagram. So where do
»uncontrolled perceptions« go in your model ?

Â

        MT

:

      Since all perceptions are equally

constructed by the operation of layers of perceptual functions
on the firing patterns of neurons connected to sensors,…¦

Â

        HB

:

        Do

I understand right that all perceptions (controlled and
uncontrolled) are equally constructed in nervous system ? So
I suppose that we can assume that all perceptions go though
the same perceptual »chanell« what is from physilogical view
right. So how do we distinguish the pathway of »controlled
and uncontrolled« perceptions in your theory ?

Â

        MT

:

      I have disagreed with Bill on this for a

couple of decades without changing his opinion, and I don’t
imagine my argument will change anyone else’s opinion. All the
same, I thought it worth putting “out there” for the
consideration of anyone who might be interested.

Â

HB:

        I'm

always interested. But as far as I was talking to Bill and
read his books I really don’t understand what was your and
Bill’s disagreement ? It seems to me a little bit confusing.
I think I always understood Bill’s theory and your
understanding of it as something very similar, specialy as
you frquently pointed out, that you talked with Bill for
days and hours and that he »filled« some of your diagrams
with his knowledge.

Â

Boris

Â

Â

From:
csgnet-request@lists.illinois.edu
[mailto:csgnet-request@lists.illinois.edu ]
On Behalf Of Martin Taylor
Sent: Saturday, August 23, 2014 5:26 AM
To: csgnet@lists.illinois.edu
Subject: Environmental variables (was: Why we
fight…)

Â

[Martin Taylor 2014.08.22.22.57]

Â

              [From

Rick Marken (2014.08.22.1940)]

Â

                  Martin Taylor

(2014.08.22.14.41)–

                              RM: Â I think I

mistakenly read this to mean that you
were saying that control also involves
reducing the variability of the external
environment. But, in fact, you never
said that; you were correctly implying
that control involves reducing the
variability of * perceptual aspects
of the environment* , keeping them
from varying around their “desired
states”.

Â

                  MT: That's true, but as you

often point out, controlling a perceptual
(internal) variable also implies reducing the
variability of the corresponding environmental
variable.

Â

                RM: I would rather say that

controlling a perceptual variable implies
controlling the environmental correlate of the
perceptual variable. The reason for this slight
change in terminology is to emphasize the fact that
there may be no environmental variable that
corresponds to the controlled perceptual variable.
For example, in B:CP (p. 112 of 2nd edition) Bill
gives the example of controlling for the perception
of the taste of lemonade. That perception is a
construction (by a perceptual function) “…derived
from the intensity signals generated by sugar and
acid (together with some oil smells)… the mere
intermingling of these physical components has no
special physical effects on anything else, except
the person tasting the mixture.” The same is true of
many other perceptions that we control, such as the
perception of variability (if we can perceive it).
The variability of independent events that happen
with different probabilities is an aspect of these
events that we can perceive but it has no more
physical significance than some other aspect of
these events that we can perceive, such Morse code
patterns.

      The problem with this, and it is a bone I have picked with

Bill as well, is that the same is true of EVERY perception.
Bill seemed to take an internally inconsistent view, saying on
the one hand that all we ever know is our perception, no
matter what kind of perception it might be, and on the other
hand that there were some privileged environmental variables
that are “real” and some that are not. Since all perceptions
are equally constructed by the operation of layers of
perceptual functions on the firing patterns of neurons
connected to sensors, on what grounds can one say a perception
such as “intensity” is of an environmental variable more real
than the taste of lemonade?

      The argument that the taste of a given glass of lemonade is

different for every person cannot hold water :slight_smile: There is no
way that we can say that any perception in one person is the
same as a perception in someone else. My perception of “blue”
may not be yours, just as my perception of the trustworthiness
of the neighbourhood loan shark or municipal politician may
not be yours.

      My preferred position on this is, I think, more consistent

than Bill’s. I assume that there exists a “real world” that
influences my sense organs and on which I can act in ways that
influence the inputs to my sense organs. I know not what is in
that world, but my nervous system constructs a perceived
reality by operations on those sensory inputs (plus all the
history of operations on earlier sensory inputs). Every
perception thus produced defines some function of present and
past states of the environment that have and had some
influence on my sensors. Those functions describe a presumed
reality “out there”, no one function, no one presumed
environmental variable thus constructed having any priority
over any other.

      In this, I believe internally consistent, view, the brightness

of a light, the taste of lemonade, the willingness of my dog
to follow me, the autocratic tendencies of my political
leaders, all have equal validity as existing states of the
external environment. I call them Complex Environmental
Variables, because they are environmental variables defined by
complex perceptual functions. There’s no way to say that
second, fifth, or ninth-level perceptions are of environmental
variables more or less real than each other.

      As for the taste of lemonade having no special physical

effects on anything else, except the person tasting the
mixture, that is true of every uncontrolled perception. If the
person is not controlling the perception of the taste of the
lemonade, any taste will do. But if the taste is a controlled
perception and it doesn’t match its reference value, the
person will act, and those actions definitely will have
effects on something else, perhaps only by adding sugar to the
mixture, perhaps by throwing the glass at the poor waiter who
brought this foul tasting mixture.

      I have disagreed with Bill on this for a couple of decades

without changing his opinion, and I don’t imagine my argument
will change anyone else’s opinion. All the same, I thought it
worth putting “out there” for the consideration of anyone who
might be interested.

      Martin

[From Rick Marken
(2014.08.23.1900)]�

[From Dag Forssell (2014.08.24.1610 PDT)]

Rick, rather than post the pdf and program to Dropbox, where it becomes
difficult to download the executable, why not just point to your source,

http://www.livingcontrolsystems.com/demos/tutor_pct.html
, last
item.

This will also serve as a test whether I can post to CSGnet now.

Best, Dag

Martin Taylor (2014.08.22.22.57)–

RM: I would rather say that controlling a perceptual variable implies
controlling the environmental correlate of the perceptual
variable…
MT: The problem with this, and it is a bone I have picked with Bill
as well, is that the same is true of EVERY perception. Bill seemed to
take an internally inconsistent view, saying on the one hand that all we
ever know is our perception, no matter what kind of perception it might
be, and on the other hand that there were some privileged environmental
variables that are “real” and some that are not.

RM: I think Bill’s point was that the environment in PCT is simply the
current model of physics and chemistry. This is the model of what is on
the “other side” of our sensory receptors. The sensory
receptors are the first step in the perceptual process. The physics model
of the environment does suggest that the sensory receptors transduce
certain physical variables into neural impulses; for example, the cone
cells in the retina seem to transduce electromagnetic energy of a
particular frequency, into neural impulses. Perhaps this is what you
think of as the “privileged” physical variables; the variables
that are in the physics model.�

�
MT: Since all perceptions are equally constructed by the operation of
layers of perceptual functions on the firing patterns of neurons
connected to sensors, on what grounds can one say a perception such as
“intensity” is of an environmental variable more real than the
taste of lemonade?

RM: I don’t think it’s that some perceptions are considered “more
real” than others. It’s that some perceptions – the intensity level
perceptions – are assumed to be a direct function of the physical
(environmental) variables that are part of the physics/chemistry model:
rods and cones transduce electromagnetic energy, hair cells transduce
mechanical energy, olfactory sensory cells transduce chemical energy,
etc. So it’s true that the lowest level sensory receptors are assumed to
produce outputs that are directly proportional to the amount of
environmental energy impinging on the cells; so the outputs of these
lowest level sensory receptors can be considered perceptions that are
proportional to the “real” level of energy impinging on them.
But that “real” level of energy is, itself, a model – part of
the current physics model, so it’s really not correct to call it
“real”. �

�
MT: The argument that the taste of a given glass of lemonade is
different for every person cannot hold water :slight_smile:

RM: I don’t think that argument is being made. The argument is simply
that the taste of lemonade is a construction – a function of lower level
perceptual signals. Whether a perception is constructed the same way by
all individuals is not relevant; just that it is constructed. And even if
it were constructed the same way by all individuals it will never be know
whether the experience of this perception is the same for everyone. All
PCT says is that perceptions above the level of the intensity perceptions
are constructions – functions of lower level perceptual signals. I
suppose these constructions could be considered a “map” of
external reality; what is constructed certainly depends on the “raw
materials” provided by the variations in the first order intensity
signals. But there are, in principle, many � different equally valid maps
that could be constructed from the same raw intensity signal variations.�

RM: According to PCT, the perceptions that are constructed are those that
result in the best control. An excellent description of how this works is
given by Bill in a paper called PCT and Engineering Control Theory. It
comes along with a demo program. I have put both the paper and program in
a folder and am making them available through dropbox at:�


https://www.dropbox.com/sh/bgdztitng5uohqm/AACg7DzKk_ZfSKb0Upoq_hzba?dl=0

The whole paper is great but the most relevant section to what we’re
discussing here is near the end called A DEMONSTRATION OF
MULTIDIMENSIONAL CONTROL. This section describes what the the Delphi
program does, which is to learn how to construct up to 500 control
systems, each controlling a weighted vector of the same 500 scalar
environmental inputs. Basically the system learns to construct perceptual
functions of these environmental variables that allows each of the
control systems to simultaneously control it’s perception without
interfering with the other systems’ ability to control their own
perceptual constructions based on the same 500 variables.�

�
MT: My preferred position on this is, I think, more consistent than
Bill’s. I assume that there exists a “real world” that
influences my sense organs and on which I can act in ways that influence
the inputs to my sense organs. I know not what is in that world, but my
nervous system constructs a perceived reality by operations on those
sensory inputs (plus all the history of operations on earlier sensory
inputs). Every perception thus produced defines some function of present
and past states of the environment that have and had some influence on my
sensors. Those functions describe a presumed reality “out
there”, no one function, no one presumed environmental variable thus
constructed having any priority over any other.

RM: That is an excellent description of my own way of looking at it!�

�
MT: In this, I believe internally consistent, view, the brightness of
a light, the taste of lemonade, the willingness of my dog to follow me,
the autocratic tendencies of my political leaders, all have equal
validity as existing states of the external environment.

RM: I would say that they all have equal validity as functions of
existing states of the external environment – the states of the
intensity signals that are proportional to the environmental energy
fluctuations at your sense receptors. The validity of a perception is
defined in PCT by the controllability of that perception in the context
of all the other perceptions that you are controlling.

�
MT: I call them Complex Environmental Variables, because they are
environmental variables defined by complex perceptual functions. There’s
no way to say that second, fifth, or ninth-level perceptions are of
environmental variables more or less real than each other.

RM: � The only problem here is that any particular CEV exists only
because it is defined by a perceptual function. Many CEVs, such as the
taste of lemonade or the oppressiveness of the weather (the combinatoin
of heat and humidity called the “discomfort index” and
discussed by Bill on pp. 7 and 8 of the PCT and Engineering paper above)
have no "physical significance; that is, they � exist only as
mathematical functions of “real” (according to the physics
model) physical variables. So I would rather refer to the environmental
correlates of these perceptions as aspects of the environment rather
than as variables in the environment to make it clear that a
perception defines an aspect of the environment that has significance for
the control system but not necessarily anything “out there” of
“real” physical significance.�

�
MT: I have disagreed with Bill on this for a couple of decades
without changing his opinion, and I don’t imagine my argument will change
anyone else’s opinion. All the same, I thought it worth putting “out
there” for the consideration of anyone who might be
interested.

� RM: Well, thank you for doing so. I always profit from discussions with
you. I learned from this one that we are probably closer than you might
think on the perception/reality issue. But most important you led me to
the PCT and Engineering paper and the accompanying program (which I think
Bill sent to me in about 2007) and I now have a better idea of what he
was trying to show. The paper is printed in the format Dag uses for his
work so I wonder if that paper (and the program) has already been
published or is available at Dag’s website. I couldn’t find it there but
maybe Dag could tell us more about that.�

Best regards

Rick

···

At 06:57 PM 8/23/2014, you wrote:

Richard S. Marken, Ph.D.

Author of� �

Doing Research on Purpose
.�

Now available from Amazon or Barnes & Noble

[From Rick Marken (2014.08.24.1700)]

···

Dag Forssell (2014.08.24.1610 PDT)–

DF: Rick, rather than post the pdf and program to Dropbox, where it becomes
difficult to download the executable, why not just point to your source,

http://www.livingcontrolsystems.com/demos/tutor_pct.html
, last
item.

RM: I certainly would have done that if I had been able to find it at your site.But my (probably too superficial) search of your site didn’t turn it up so I used what I had on my disk, which I assumed came from Bill. Thanks for showing me where it is. Now I’ll take it down from my dropbox site! Thanks.

DF:This will also serve as a test whether I can post to CSGnet now.

RM: It apparently works now! Welcome back!

Best

Rick


Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

Martin,

sorry for the delay…J. It’s the beggining of school year…

MT :

Every perceptual function defines a CEV. If you happen to have a perceptual function that produces an output proportional to 3 times the temperature in degrees Kelvin plus twice the number of minutes since last midnight, then 3*(deg K)+2*(minutes) is the CEV for that perception. It’s a nonsense one, of course, but nonsense sometimes is useful to illuminate sense.

HB :

Do I understand right, that CEV is something what happens in sensor (afferent) nerves - output of »input function« ? In your case if I understand right temperature receptors »transformed« some physical stimuli from outer environment into »CEV«, that »coresponds« to the temperature outside ?

MT :

If you have a perceptual function, by definition its output is a perceptual signal that changes when the environmental variables that enter into it change.

HB:

Well I think I don’t understand this one clearly. Are you assuming that there is linear (precise) relationship between changes in values of environmental variables that enters in »input function« and changes in perceptual signal ? So that changes in »outside physical variables« are  varying in the same time and intensity with changes in perceptual signal ? Â

MT :

I have often said to you that Bill’s DIAGRAMs contain only controlled perceptions, because they are diagrams of how those perceptions are controlled. Uncontrolled perceptions have values that may or may not contribute to the values of controlled perceptions. It’s irrelevant in a diagram intended to show how control works. I don’t have any particular model for uncontrolled perceptions, nor do I think one is needed.

HB :

So If I understand right in first sentence you say that Bill’s diagram contain only controlled perception and in second sentence you say that uncontrolled perception have values that may or may not contribute to the values of controlled perceptions. So if I try to make a cocnlusion »uncontrolled perception« which contribute to the values of controlled perception go through the same »perceptual channel« as controlled goes. How could they otherwise contribute to controlled perception ? O.K. maybe we could make some example to release misunderstanding.

Suppose that you are driving a car (it’s top example in PCT) and you are controlling the perception of front side of the car in just enough close range to reference perception of the middle right side of the street, that we say you have it »under control«. And you are looking also arround. I assume that beside »controlled perception«, you accept also »uncontrolled perception« of surrounding (maybe trees, fields, buildings, people and so on). I suppose that all these perception goes through the same pathway as you describe it : »…all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors,… ». Â

»Suddenly a «bump« of wind »pushed« your perception of the front side of the car »away« from reference (right side of the street). Is this »bump of the wind« in first moment controlled or uncontrolled perception ? Where does it go ?

MT :

There have been many times I have disagreed with Bill and sometimes one of us changed our mind, other times when we agreed to disagree. But on the big things, I don’t think we disagreed much, if at all.

RM:

Well, thank you for doing so. I always profit from discussions with you.

MT :

I think we understand one another, and it is such a small “angels on a pin-head” point that I don’t feel like digging into the very slight difference in emphasis that I think does remain. It will probably remain a disagreement that shows up from time to time in the wording of some statement or other, but probably not in an understanding of PCT.

HB :

Everything is good what ends good J. I took a long time thinking for this quite complex environmental variable problem of why you two agree and profit from each other. And on the end occured to me that possible answer is, that you have common, great genius »teacher« J.

Boris

···

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Sunday, August 24, 2014 6:21 AM
To: csgnet@lists.illinois.edu
Subject: Re: Environmental variables (was: Why we fight…)

[Martin Taylor 2014.08.23.23.52]

On 2014/08/23 10:36 AM, Boris Hartman wrote:

Martin,

I’m not a good reader of your »English«, so I don’t know whether I understood some thoughts of yours, as you wanted to be understaood.

MT :

The problem with this, and it is a bone I have picked with Bill as well, is that the same is true of EVERY perception. Bill seemed to take an internally inconsistent view, saying on the one hand that all we ever know is our perception, no matter what kind of perception it might be, and on the other hand that there were some privileged environmental variables that are “real” and some that are not.

HB :

Where did Bill wrote that some privileged environmental variables are more real than others ? I really don’t recall »meating« this in his books. But I’m not some specialy good »English« reader. It’s possible that I missed something.

Rick references B:CP, 2nd Ed. p 112 as an example. On rereading it, I can see that the writing is ambiguous. I read it (as Rick seems to have done) as saying that there are environmental variable corresponding to first-level perceptions, but not to higher-level perceptions, which are constructed.

MT :

The argument that the taste of a given glass of lemonade is different for every person cannot hold water :slight_smile: There is no way that we can say that any perception in one person is the same as a perception in someone else. My perception of “blue” may not be yours, just as my perception of the trustworthiness of the neighbourhood loan shark or municipal politician may not be yours.

HB :

I really don’t see the point. In the first sentence it seems to me, that you are saying »that the taste of a given glass of lemonade is different for every person cannot hold water« so I understand it, that different persons have the same perception of taste (can not »hold water«). It seems to me that you are denying the existance of different taste of lemonade. In the other senstences I understand that perceptions are different for any person. Did I misssed something ?

No. I’m afraid my English was not clear enough. It’s “the argument” (for the non-existence in the environment of a variable corresponding to the taste of lemonade) that cannot hold water. I made no claim about whether different people, or the same person at different times, would or would not perceive the same taste from a particular glass of lemonade.

MT :

My preferred position on this is, I think, more consistent than Bill’s. I assume that there exists a “real world” that influences my sense organs and on which I can act in ways that influence the inputs to my sense organs. I know not what is in that world, but my nervous system constructs a perceived reality by operations on those sensory inputs (plus all the history of operations on earlier sensory inputs). Every perception thus produced defines some function of present and past states of the environment that have and had some influence on my sensors. Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

HB :

I never felt that Bill had different opinion from yours. But he did once have a discours of »solipsism« but he asked himself whether his thinking was »solipsist« view or not. Well whatever I never have a feeling that he is solipsist or something different from your perceptual understanding of »real world«.

Solipsism doesn’t come into this discussion, except that we have to dismiss it before we can have a discussion that includes talk about an external environment. Basically, it’s totally irrelevant. We agree it’s wrong, and then deal with our theories.

MT :

In this, I believe internally consistent, view, the brightness of a light, the taste of lemonade, the willingness of my dog to follow me, the autocratic tendencies of my political leaders, all have equal validity as existing states of the external environment. I call them Complex Environmental Variables, because they are environmental variables defined by complex perceptual functions. There’s no way to say that second, fifth, or ninth-level perceptions are of environmental variables more or less real than each other.

HB :

I really don’t imagine what »Complex Environemntal Variable« (CEV) can be ? If I assume that intensities as first level controlled variables which »integrate« into second level »sensations« and so on through the whole heirarchy, what is CEV ?

Every perceptual function defines a CEV. If you happen to have a perceptual function that produces an output proportional to 3 times the temperature in degrees Kelvin plus twice the number of minutes since last midnight, then 3*(deg K)+2*(minutes) is the CEV for that perception. It’s a nonsense one, of course, but nonsense sometimes is useful to illuminate sense.

Maybe your »perceptual mosaic« of the world which you assume to be »outside«. As you said. You beleive that your view is internally consistent. Others beleive that Bill’s view or their view is internally consistent. How can we know which »view« is internally consistent ? But it seems interesting what’s your theory of how we perceive the »world« outside. How does »Complex Perceptual Functions« (CPF) work ?

Are not all perceptual functions complex? I don’t mean in the mathematical sense of using numbers with real and imaginary components, but in the everyday sense of being rather complicated.

MT :

Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

HB :

So if I choose to control some variable in »environment«, are you saying that, it has no priority over other perceived variables in »environment« in that moment ?

It matters not whether you choose to control a perception. Any function you define creates a value. That value is a variable whose magnitude depends on the values of the inputs to the function. Perceptual functions are no different. If you have a perceptual function, by definition its output is a perceptual signal that changes when the environmental variables that enter into it change.

MT :

As for the taste of lemonade having no special physical effects on anything else, except the person tasting the mixture, that is true of every uncontrolled perception.

HB :

If I recall right you said that Bill’s model assume only controlled perceptions in the diagram. So where do »uncontrolled perceptions« go in your model ?

No. I have often said to you that Bill’s DIAGRAMs contain only controlled perceptions, because they are diagrams of how those perceptions are controlled. Uncontrolled perceptions have values that may or may not contribute to the values of controlled perceptions. It’s irrelevant in a diagram intended to show how control works. I don’t have any particular model for uncontrolled perceptions, nor do I think one is needed.

MT :

Since all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors,…

HB :

Do I understand right that all perceptions (controlled and uncontrolled) are equally constructed in nervous system ? So I suppose that we can assume that all perceptions go though the same perceptual »chanell« what is from physilogical view right. So how do we distinguish the pathway of »controlled and uncontrolled« perceptions in your theory ?

You don’t. The only difference is whether at any particular moment the perception is being controlled – whether you are actively doing something to keep it near its reference value, if it has one.

MT :

I have disagreed with Bill on this for a couple of decades without changing his opinion, and I don’t imagine my argument will change anyone else’s opinion. All the same, I thought it worth putting “out there” for the consideration of anyone who might be interested.

HB:

I’m always interested. But as far as I was talking to Bill and read his books I really don’t understand what was your and Bill’s disagreement ? It seems to me a little bit confusing. I think I always understood Bill’s theory and your understanding of it as something very similar, specialy as you frquently pointed out, that you talked with Bill for days and hours and that he »filled« some of your diagrams with his knowledge.

I have always believed that my understanding of Bill’s theory was very close to his. Our differences, as I perceived them, were always on the margins. For example, I have never been convinced by the specific levels Bill describes in his hierarchy. Neither was he, to judge from the way he inverted the order of levels in his models, and the number of times he explicitly said that his set of levels was always subject to revision. There have been many times I have disagreed with Bill and sometimes one of us changed our mind, other times when we agreed to disagree. But on the big things, I don’t think we disagreed much, if at all. This issue of whether all or only some environmental variable are constructed by perceptual functions is a very little thing.

Martin

Boris

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Saturday, August 23, 2014 5:26 AM
To: csgnet@lists.illinois.edu
Subject: Environmental variables (was: Why we fight…)

[Martin Taylor 2014.08.22.22.57]

[From Rick Marken (2014.08.22.1940)]

Martin Taylor (2014.08.22.14.41)–

RM: I think I mistakenly read this to mean that you were saying that control also involves reducing the variability of the external environment. But, in fact, you never said that; you were correctly implying that control involves reducing the variability of perceptual aspects of the environment, keeping them from varying around their “desired states”.

MT: That’s true, but as you often point out, controlling a perceptual (internal) variable also implies reducing the variability of the corresponding environmental variable.

RM: I would rather say that controlling a perceptual variable implies controlling the environmental correlate of the perceptual variable. The reason for this slight change in terminology is to emphasize the fact that there may be no environmental variable that corresponds to the controlled perceptual variable. For example, in B:CP (p. 112 of 2nd edition) Bill gives the example of controlling for the perception of the taste of lemonade. That perception is a construction (by a perceptual function) “…derived from the intensity signals generated by sugar and acid (together with some oil smells)… the mere intermingling of these physical components has no special physical effects on anything else, except the person tasting the mixture.” The same is true of many other perceptions that we control, such as the perception of variability (if we can perceive it). The variability of independent events that happen with different probabilities is an aspect of these events that we can perceive but it has no more physical significance than some other aspect of these events that we can perceive, such Morse code patterns.

The problem with this, and it is a bone I have picked with Bill as well, is that the same is true of EVERY perception. Bill seemed to take an internally inconsistent view, saying on the one hand that all we ever know is our perception, no matter what kind of perception it might be, and on the other hand that there were some privileged environmental variables that are “real” and some that are not. Since all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors, on what grounds can one say a perception such as “intensity” is of an environmental variable more real than the taste of lemonade?

The argument that the taste of a given glass of lemonade is different for every person cannot hold water :slight_smile: There is no way that we can say that any perception in one person is the same as a perception in someone else. My perception of “blue” may not be yours, just as my perception of the trustworthiness of the neighbourhood loan shark or municipal politician may not be yours.

My preferred position on this is, I think, more consistent than Bill’s. I assume that there exists a “real world” that influences my sense organs and on which I can act in ways that influence the inputs to my sense organs. I know not what is in that world, but my nervous system constructs a perceived reality by operations on those sensory inputs (plus all the history of operations on earlier sensory inputs). Every perception thus produced defines some function of present and past states of the environment that have and had some influence on my sensors. Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

In this, I believe internally consistent, view, the brightness of a light, the taste of lemonade, the willingness of my dog to follow me, the autocratic tendencies of my political leaders, all have equal validity as existing states of the external environment. I call them Complex Environmental Variables, because they are environmental variables defined by complex perceptual functions. There’s no way to say that second, fifth, or ninth-level perceptions are of environmental variables more or less real than each other.

As for the taste of lemonade having no special physical effects on anything else, except the person tasting the mixture, that is true of every uncontrolled perception. If the person is not controlling the perception of the taste of the lemonade, any taste will do. But if the taste is a controlled perception and it doesn’t match its reference value, the person will act, and those actions definitely will have effects on something else, perhaps only by adding sugar to the mixture, perhaps by throwing the glass at the poor waiter who brought this foul tasting mixture.

I have disagreed with Bill on this for a couple of decades without changing his opinion, and I don’t imagine my argument will change anyone else’s opinion. All the same, I thought it worth putting “out there” for the consideration of anyone who might be interested.

Martin

[Martin Taylor 2014.08.27.13.48]

The Complex ENVIRONMENTAL Variable is a function in the environment

of some variables in the environment. It corresponds to a perception
inside an organism (a person, perhaps). If the variables are the
environmental correlates of what the peripheral sensors send into
the organism, then that’s what they are. But they don’t have to be.
The point is that changing the value of the CEV will also change the
value of the corresponding perceptual signal. It is something you
perceive to be in the environment.
The temperature sensors produced a signal. If that signal
corresponds to a feeling of hot or cold, then the perception is that
there exists outside the body something that is hot or cold. That
“something” has a property called “temperature”, which is the CEV
corresponding to the perception of the temperature of the
“something” in the environment.
No. Any function will do. I always assume that a perceptual function
is nonlinear, but that doesn’t usually matter much when dealing with
control unless you are doing very precise simulation.
[Maybe the following is an unnecessary complication at this point,
but it’s worth remembering.] We know that most of these functions change over time. For example,
perceived brightness doesn’t have a fixed relationship with the flux
of photons onto the retina, because we adapt to the overall level of
illumination. The adaptation is slow compared to the rate at which
illumination can change, so over short time periods there is a
function relating changes in photon flux to changes in perceived
brightness. The same seems to be true of any perceptual function I
can think of at the moment. They all depend on history as well as
what is currently coming from the sensors, but if the effect of
history changes slowly compared to the changes in the sensors, it is
still reasonable to say there is a functional relationship between
perception and changes in the environmental variable.
[end complication]
There are always delays and time-averaging incorporated in the
perceptual process, and the “intensity” won’t vary in the same ratio
as the perceptual signal, but you have the core of the idea correct.
There is exactly one outside physical variable defined by the
processes that lead to one perceptual signal. That outside physical
variable may be a very complicated function of the basic physical
variables of mass, distance, force, voltage, time, etc. Even the
response of a sensor is a very complex function of the basic
variables. Sometimes we take the CEV of a sensor to be a basic
physical variable, but that’s just as arbitrary as saying that the
perceived taste of lemonade is not a physical variable. Sometimes we
call “momentum” and “energy” basic variables even though they are
functions of other basic variables. It’s all models and functions
and what you want to do as an analyst, after all. To the perceiver,
there’s just something “out there”, and for each perceptual variable
we call what is “out there” a Complex Environmental Variable.
You don’t need to add “looking around”. All you need to say is
something like: “There are shadows across the road that affect your
perception of the edge, and maybe the sun gets into your eyes.
Perhaps there is rain on the windscreen. You aren’t controlling
those, but they do influence your perception of the relation of the
car to the edge of the lane.”
Yes. But say “pathways”, plural. There are at least millions, maybe
billions of such pathways.
You don’t perceive a “bump of the wind” unless you have mounted a
detector for such things onto the hood of your car. You perceive the
location of the car on the road, which has departed from its
reference value. An outside observer might know that the error was
due to a “bump of the wind”, but you can’t tell whether it was a dip
in the road surface – at least I can’t. When I’m on the highway, I
often wonder whether a deviation is cause by wind, road surface, or
(I hope not) a defect in the car such as a tire going flat.
Martin

···

On 2014/08/27 12:51 PM, Boris Hartman
wrote:

Martin,

Â

        sorry

for the delay…J .
It’s the beggining of school year…

<

Â

        MT

:

      Every perceptual function defines a CEV. If

you happen to have a perceptual function that produces an
output proportional to 3 times the temperature in degrees
Kelvin plus twice the number of minutes since last midnight,
then 3*(deg K)+2*(minutes) is the CEV for that perception.
It’s a nonsense one, of course, but nonsense sometimes is
useful to illuminate sense.

        HB

:

        Do

I understand right, that CEV is something what happens in
sensor (afferent) nerves - output of »input function« ?

        Â 

In your case if I understand right temperature receptors
»transformed« some physical stimuli from outer environment
into »CEV«, that »coresponds« to the temperature outside ?

Â

MT :

      If you have a perceptual function, by

definition its output is a perceptual signal that changes when
the environmental variables that enter into it change.

Â

HB:

        Well

I think I don’t understand this one clearly. Are you
assuming that there is linear (precise) relationship between
changes in values of environmental variables that enters in
»input function« and changes in perceptual signal ?

        So that changes in »outside physical variables« are  varying

in the same time and intensity with changes in perceptual
signal ?Â

Â

MT :

      I have often said to you that Bill's

DIAGRAMs contain only controlled perceptions, because they are
diagrams of how those perceptions are controlled. Uncontrolled
perceptions have values that may or may not contribute to the
values of controlled perceptions. It’s irrelevant in a diagram
intended to show how control works. I don’t have any
particular model for uncontrolled perceptions, nor do I think
one is needed.

Â

        HB

:

        So

If I understand right in first sentence you say that Bill’s
diagram contain only controlled perception and in second
sentence you say that uncontrolled perception have values
that may or may not contribute to the values of controlled
perceptions. So if I try to make a cocnlusion »uncontrolled
perception« which contribute to the values of controlled
perception go through the same »perceptual channel« as
controlled goes. How could they otherwise contribute to
controlled perception ?

        O.K. maybe we could make some example to release

misunderstanding.

Â

        Suppose

that you are driving a car (it’s top example in PCT) and you
are controlling the perception of front side of the car in
just enough close range to reference perception of the
middle right side of the street, that we say you have it
»under control«. And you are looking also arround.

        I

assume that beside »controlled perception«, you accept also
»uncontrolled perception« of surrounding (maybe trees,
fields, buildings, people and so on). I suppose that all
these perception goes through the same pathway as you
describe it : »… all perceptions are equallly
constructed by the operation of layers of perceptual functions
on the firing patterns of neurons connected to sensors,…¦
».Â

Â

        »Suddenly

a «bump« of wind »pushed« your perception of the front side
of the car »away« from reference (right side of the street).
Is this »bump of the wind« in first moment controlled or
uncontrolled perception ? Where does it go ?

Â

From:
csgnet-request@lists.illinois.edu
[mailto:csgnet-request@lists.illinois.edu ]
On Behalf Of Martin Taylor
Sent: Sunday, August 24, 2014 6:21 AM
To: csgnet@lists.illinois.edu
Subject: Re: Environmental variables (was: Why we
fight…)

Â

[Martin Taylor 2014.08.23.23.52]

        On 2014/08/23 10:36 AM, Boris Hartman

wrote:

Martin,

Â

          I'm

not a good reader of your »English«, so I don’t know
whether I understood some thoughts of yours, as you wanted
to be understaood.

Â

          MT

:

        The problem with this, and it is a bone I

have picked with Bill as well, is that the same is true of
EVERY perception. Bill seemed to take an internally
inconsistent view, saying on the one hand that all we ever
know is our perception, no matter what kind of perception it
might be, and on the other hand that there were some
privileged environmental variables that are “real” and some
that are not.

Â

          HB

:

          Where

did Bill wrote that some privileged environmental
variables are more real than others ? I really don’t
recall »meating« this in his books. But I’m not some
specialy good »English« reader. It’s possible that I
missed something.

      Rick references B:CP, 2nd Ed. p 112 as an example. On

rereading it, I can see that the writing is ambiguous. I read
it (as Rick seems to have done) as saying that there are
environmental variable corresponding to first-level
perceptions, but not to higher-level perceptions, which are
constructed.

Â

        MT

:

      The argument that the taste of a given

glass of lemonade is different for every person cannot hold
water :slight_smile: There is no way that we can say that any perception
in one person is the same as a perception in someone else. My
perception of “blue” may not be yours, just as my perception
of the trustworthiness of the neighbourhood loan shark or
municipal politician may not be yours.

Â

        HB

:

        I

really don’t see the point. In the first sentence it seems
to me, that you are saying » that the taste of a given
glass of lemonade is different for every person cannot hold
water« so I understand it, that
different persons have the same perception of taste (can not
»hold water«). It seems to me that you are denying the
existance of different taste of lemonade. In the other
senstences I understand that perceptions are different for
any person. Did I misssed something ?

      No. I'm afraid my English was not clear enough. It's "the

argument" (for the non-existence in the environment of a
variable corresponding to the taste of lemonade) that cannot
hold water. I made no claim about whether different people, or
the same person at different times, would or would not
perceive the same taste from a particular glass of lemonade.

Â

        MT

:

      My preferred position on this is, I think,

more consistent than Bill’s. I assume that there exists a
“real world” that influences my sense organs and on which I
can act in ways that influence the inputs to my sense organs.
I know not what is in that world, but my nervous system
constructs a perceived reality by operations on those sensory
inputs (plus all the history of operations on earlier sensory
inputs). Every perception thus produced defines some function
of present and past states of the environment that have and
had some influence on my sensors. Those functions describe a
presumed reality “out there”, no one function, no one presumed
environmental variable thus constructed having any priority
over any other.

Â

        HB

:

        I

never felt that Bill had different opinion from yours. But
he did once have a discours of »solipsism« but he asked
himself whether his thinking was »solipsist« view or not.
Well whatever I never have a feeling that he is solipsist
or something different from your perceptual understanding of
»real world«.

      Solipsism doesn't come into this discussion, except that we

have to dismiss it before we can have a discussion that
includes talk about an external environment. Basically, it’s
totally irrelevant. We agree it’s wrong, and then deal with
our theories.

Â

        MT

:

      In this, I believe internally consistent,

view, the brightness of a light, the taste of lemonade, the
willingness of my dog to follow me, the autocratic tendencies
of my political leaders, all have equal validity as existing
states of the external environment. I call them Complex
Environmental Variables, because they are environmental
variables defined by complex perceptual functions. There’s no
way to say that second, fifth, or ninth-level perceptions are
of environmental variables more or less real than each other.

Â

        HB

:

        I

really don’t imagine what »Complex Environemntal Variable«
(CEV) can be ? If I assume that intensities as first level
controlled variables which »integrate« into second level
»sensations« and so on through the whole heirarchy, what is
CEV ?

      Every perceptual function defines a CEV. If you happen to have

a perceptual function that produces an output proportional to
3 times the temperature in degrees Kelvin plus twice the
number of minutes since last midnight, then 3*(deg
K)+2*(minutes) is the CEV for that perception. It’s a nonsense
one, of course, but nonsense sometimes is useful to illuminate
sense.

        Maybe

your »perceptual mosaic« of the world which you assume to be
»outside«. As you said. You beleive that your view is
internally consistent. Others beleive that Bill’s view or
their view is internally consistent. How can we know which
»view« is internally consistent ? But it seems interesting
what’s your theory  of how we perceive the »world« outside.
How does »Complex Perceptual Functions« (CPF) work ?

      Are not all perceptual functions complex? I don't mean in the

mathematical sense of using numbers with real and imaginary
components, but in the everyday sense of being rather
complicated.

Â

        MT

:

      Those functions describe a presumed reality

“out there”, no one function, no one presumed environmental
variable thus constructed having any priority over any other.

Â

        HB

:

        So

if I choose to control some variable in »environment«, are
you saying that, it has no priority over other perceived
variables in »environment« in that moment ?

      It matters not whether you choose to control a perception. Any

function you define creates a value. That value is a variable
whose magnitude depends on the values of the inputs to the
function. Perceptual functions are no different. If you have a
perceptual function, by definition its output is a perceptual
signal that changes when the environmental variables that
enter into it change.

Â

        MT

:

      As for the taste of lemonade having no

special physical effects on anything else, except the person
tasting the mixture, that is true of every uncontrolled
perception.

Â

        HB

:

        If

I recall right you said that Bill’s model assume only
controlled perceptions in the diagram. So where do
»uncontrolled perceptions« go in your model ?

      No. I have often said to you that Bill's DIAGRAMs contain only

controlled perceptions, because they are diagrams of how those
perceptions are controlled. Uncontrolled perceptions have
values that may or may not contribute to the values of
controlled perceptions. It’s irrelevant in a diagram intended
to show how control works. I don’t have any particular model
for uncontrolled perceptions, nor do I think one is needed.

Â

        MT

:

      Since all perceptions are equally

constructed by the operation of layers of perceptual functions
on the firing patterns of neurons connected to sensors,…¦

Â

        HB

:

        Do

I understand right that all perceptions (controlled and
uncontrolled) are equally constructed in nervous system ? So
I suppose that we can assume that all perceptions go though
the same perceptual »chanell« what is from physilogical view
right. So how do we distinguish the pathway of »controlled
and uncontrolled« perceptions in your theory ?

      You don't. The only difference is whether at any particular

moment the perception is being controlled – whether you are
actively doing something to keep it near its reference value,
if it has one.

Â

        MT

:

      I have disagreed with Bill on this for a

couple of decades without changing his opinion, and I don’t
imagine my argument will change anyone else’s opinion. All the
same, I thought it worth putting “out there” for the
consideration of anyone who might be interested.

Â

HB:

        I'm

always interested. But as far as I was talking to Bill and
read his books I really don’t understand what was your and
Bill’s disagreement ? It seems to me a little bit confusing.
I think I always understood Bill’s theory and your
understanding of it as something very similar, specialy as
you frquently pointed out, that you talked with Bill for
days and hours and that he »filled« some of your diagrams
with his knowledge.

      I have always believed that my understanding of Bill's theory

was very close to his. Our differences, as I perceived them,
were always on the margins. For example, I have never been
convinced by the specific levels Bill describes in his
hierarchy. Neither was he, to judge from the way he inverted
the order of levels in his models, and the number of times he
explicitly said that his set of levels was always subject to
revision. There have been many times I have disagreed with
Bill and sometimes one of us changed our mind, other times
when we agreed to disagree. But on the big things, I don’t
think we disagreed much, if at all. This issue of whether all
or only some environmental variable are constructed by
perceptual functions is a very little thing.

      Martin

Â

Boris

Â

Â

From:
csgnet-request@lists.illinois.edu
[mailto:csgnet-request@lists.illinois.edu ]
On Behalf Of Martin Taylor
Sent: Saturday, August 23, 2014 5:26 AM
To: csgnet@lists.illinois.edu
Subject: Environmental variables (was: Why we
fight…)

Â

[Martin Taylor 2014.08.22.22.57]

Â

              [From

Rick Marken (2014.08.22.1940)]

Â

                  Martin Taylor

(2014.08.22.14.41)–

                              RM: Â I think I

mistakenly read this to mean that you
were saying that control also involves
reducing the variability of the external
environment. But, in fact, you never
said that; you were correctly implying
that control involves reducing the
variability of * perceptual aspects
of the environment* , keeping them
from varying around their “desired
states”.

Â

                  MT: That's true, but as you

often point out, controlling a perceptual
(internal) variable also implies reducing the
variability of the corresponding environmental
variable.

Â

                RM: I would rather say that

controlling a perceptual variable implies
controlling the environmental correlate of the
perceptual variable. The reason for this slight
change in terminology is to emphasize the fact that
there may be no environmental variable that
corresponds to the controlled perceptual variable.
For example, in B:CP (p. 112 of 2nd edition) Bill
gives the example of controlling for the perception
of the taste of lemonade. That perception is a
construction (by a perceptual function) “…derived
from the intensity signals generated by sugar and
acid (together with some oil smells)… the mere
intermingling of these physical components has no
special physical effects on anything else, except
the person tasting the mixture.” The same is true of
many other perceptions that we control, such as the
perception of variability (if we can perceive it).
The variability of independent events that happen
with different probabilities is an aspect of these
events that we can perceive but it has no more
physical significance than some other aspect of
these events that we can perceive, such Morse code
patterns.

      The problem with this, and it is a bone I have picked with

Bill as well, is that the same is true of EVERY perception.
Bill seemed to take an internally inconsistent view, saying on
the one hand that all we ever know is our perception, no
matter what kind of perception it might be, and on the other
hand that there were some privileged environmental variables
that are “real” and some that are not. Since all perceptions
are equally constructed by the operation of layers of
perceptual functions on the firing patterns of neurons
connected to sensors, on what grounds can one say a perception
such as “intensity” is of an environmental variable more real
than the taste of lemonade?

      The argument that the taste of a given glass of lemonade is

different for every person cannot hold water :slight_smile: There is no
way that we can say that any perception in one person is the
same as a perception in someone else. My perception of “blue”
may not be yours, just as my perception of the trustworthiness
of the neighbourhood loan shark or municipal politician may
not be yours.

      My preferred position on this is, I think, more consistent

than Bill’s. I assume that there exists a “real world” that
influences my sense organs and on which I can act in ways that
influence the inputs to my sense organs. I know not what is in
that world, but my nervous system constructs a perceived
reality by operations on those sensory inputs (plus all the
history of operations on earlier sensory inputs). Every
perception thus produced defines some function of present and
past states of the environment that have and had some
influence on my sensors. Those functions describe a presumed
reality “out there”, no one function, no one presumed
environmental variable thus constructed having any priority
over any other.

      In this, I believe internally consistent, view, the brightness

of a light, the taste of lemonade, the willingness of my dog
to follow me, the autocratic tendencies of my political
leaders, all have equal validity as existing states of the
external environment. I call them Complex Environmental
Variables, because they are environmental variables defined by
complex perceptual functions. There’s no way to say that
second, fifth, or ninth-level perceptions are of environmental
variables more or less real than each other.

      As for the taste of lemonade having no special physical

effects on anything else, except the person tasting the
mixture, that is true of every uncontrolled perception. If the
person is not controlling the perception of the taste of the
lemonade, any taste will do. But if the taste is a controlled
perception and it doesn’t match its reference value, the
person will act, and those actions definitely will have
effects on something else, perhaps only by adding sugar to the
mixture, perhaps by throwing the glass at the poor waiter who
brought this foul tasting mixture.

      I have disagreed with Bill on this for a couple of decades

without changing his opinion, and I don’t imagine my argument
will change anyone else’s opinion. All the same, I thought it
worth putting “out there” for the consideration of anyone who
might be interested.

      Martin

Â

[From Rick Marken (2014.08.27.2040)]

···

Martin Taylor (2014.08.27.13.48)

MT: The Complex ENVIRONMENTAL Variable is a function in the environment

of some variables in the environment. It corresponds to a perception
inside an organism (a person, perhaps). If the variables are the
environmental correlates of what the peripheral sensors send into
the organism, then that’s what they are. But they don’t have to be.
The point is that changing the value of the CEV will also change the
value of the corresponding perceptual signal. It is something you
perceive to be in the environment.

RM: The only disagreement I have is that it is not necessarily true that what you call the Complex Environmental Variable is a function in the environment of some variables in the environment. The function that relates the variables is not necessarily in the environment. The variables are in the environment but the function that combines them into a perception may exist only in the control system itself. I think that’s what Bill meant when he said that the combination of variables that makes up a perception, such as the taste of lemonade, may have no “physical significance”; the combination doesn’t necessarily exist in the environment.

RM: I actually built a little demo to illustrate this point – well, actually to illustrate the difficulty of doing the TCV but 'twill serve – using the Arduino board that Adam Matic told us about some time ago. I built a little control system that controls a perception of a combination of sound and light stimuli. The sound is produced by a buzzer and the light is produced by two LEDs, one representing an independent source of light in the environment – the disturbance LED – and the other being the output of the control system – the output LED. The control system senses the sound using a tiny microphone and it senses the light using a photoresistor. The outputs of these two sensors are combined by the control system (a computer program in the Arduino) into a perceptual signal which is compared to a reference specification, also in the computer. Deviations of this perception from the reference specification drive the output LED, increasing its intensity if the perception is below the reference and decreasing its intensity if the perception is above the reference.

RM: So this control system controls a perception that is a combination of light and sound by varying the intensity of a single output; the output LED. The level of output from the photoresistor (call it s1) depends on the intensity of light produced by the input and output LEDs; the level of output from the mic (s2) depends on just the amplitude of the sound produced by the buzzer. The perception controlled by the control system is a linear combination of s1 and s2: p = k1s1+k2s2. The intensity of the sound produced by the buzzer and the intensity of the light produced by the disturbance LEDs are disturbances to this perception. The magnitude of these two disturbances can be varied independently by variable resistors.

RM: I stopped working on this a couple months ago for various reasons, mainly having to do with having problems making it work properly (and not fall apart) but I still have it set up (in the state it was last in) and I took a little video of what I have so you can get a better idea of what I’m talking about. The video is at:

https://www.dropbox.com/s/ro12arm81ksgtip/2014-08-27%2017.24.06.mov?dl=0

RM: There are two LEDs next to each other on the right; the leftmost of the two is the disturbance LED and the rightmost is the output LED; so the rightmost LED is the only “behavior” of the control system that is visible. There are two variable resistors on the left. The one on the far left, which I turn first, varies the amplitude of the disturbance LED. It’s a little hard to tell but when I turn the intensity of that LED down the intensity of the output LED goes up; and vice versa; when I turn the disturbance LED intensity up the intensity of the output LED goes down.

RM: While the intensity of the disturbance LED is off I then turn the other variable resistor, which varies the amplitude of the buzzer sound disturbance. This leads to a decrease in the intensity of the output LED; increased sound seems to cause a decrease in output. But this happens because the system is controlling a perception that is a combination of light and sound; when the light is low with no sound present the perception is maintained at the reference state by the increased output. But when the sound comes on, it makes up for the absence of light in the perception and the intensity of the output decreases.

RM: As I said, I was developing the little demo to use as an illustration (in a class on control theory eventually) the difficulty of determining the purpose being carried out by a control system; ie, the difficulty of determining controlled variables. This is because the perceptual variables the system is controlling may not be perceptions that the person studying the control system can have, as is the case here. We don’t perceive the peculiar combination of sound and light that this little control system is controlling. So it would be difficult to do much more than determine that the system does seem to be controlling a combination of sound and light; but it would be impossible to figure out exactly what perception it is controlling without doing some modeling; that is, you would have to have the computer compute various combinations of light and sound and see which one seems to be controlled the best.

RM: But this demo also serves as a nice demonstration of the point that the perception that a control system controls does not necessarily correspond to a variable in the environment. In this demo, the only relevant variables in the environment are the sound vibrations and light waves. The linear combination of sensed values of these variables is not in the environment any more than any other linear or non linear combination of these variables is out there. The environmental variables – sounds and light – provide the opportunity for the system to construct a perception based on these “raw materials”; but there is no variable in the environment that corresponds to the perceptual variable that is constructed by the control system.

RM: I would be interested in hearing your (or anyone’s) opinion about this demo to see if it’s worth trying to get back into it and make it work better. And, of course, any suggestions that might make it better would certainly be welcome as well. I think I like it, even though it’s not nearly as exciting as the things Adam and Rupert are doing. But I do think it could be a nice demo of the “real world” difficulties involved in determining what a purposeful system is doing (what perceptions it is controlling).

Best

Rick


Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

Martin,

I’ll start from the end, as problem there is somehow continuing »story« throughh the whole your post.

MT :

You don’t perceive a “bump of the wind” unless you have mounted a detector for such things onto the hood of your car.

HB :

I perceive what I, we call the »wind« (from my previos experiences) if my car window is opened (sensors for pressure), or usualy I can hear (perceive) it as »bump« or feel specific »trembling« of the car. So I can distinguish special sounds and »tremblings« (at least what I call sound and trembling) when wind »bump« into a car. And beside that I also perceive deviation of the front side of the car from reference. It’s »Complex perception«.

If you don’t perceive such a »CEV« than I understand why you have to mount devices for »detecting« such »complex stimuli« J. It seems to me, that difference between your and my perception is in the way we perceive. I for example use CEP (Complex Environmental Perception). It’s something like a »multi-media« perception. There are numerous perceptual inputs through which I perceive at the same time many »analogs« of environmental variables which form some »complex perceptual constructions« of environment on different levels in hierarchy. So I don’t need to mount detector for any physical quantity as my own numerous Input functions are effective enough J.

MT :

You perceive the location of the car on the road, which has departed from its reference value.

HB :

If you would read some rows above in my past post you could see my description of what you described as »deviation from reference« caused by a wind. It’s quite similar. So you didn’t need to repeat explanation. I’ve already know it. J

MT :

An outside observer might know that the error was due to a “bump of the wind”, but you can’t tell whether it was a dip in the road surface – at least I can’t. When I’m on the highway, I often wonder whether a deviation is cause by wind, road surface, or (I hope not) a defect in the car such as a tire going flat.

HB :

I don’t see how outside observer »might know« that »error« was due to »bump of the wind« ? What kind of »remote« sensors did he mount on the car to know it ? Martin I’m sorry to say, but we are exaggerating. J.

MT :

….but you can’t tell whether it was a dip in the road surface – at least I can’t. When I’m on the highway, I often wonder whether a deviation is cause by wind, road surface, or (I hope not) a defect in the car such as a tire going flat.

HB :

Martin, maybe you will not beleive me, but I do clearly distinguish the sound and trembling of what we call wind when it »bump« into a car, from the sound and trembling when I perceive my car drive into something you call »dip in the road surface« or sound and trembling caused by tyre defect. As I said I use »CEP«. But If you don’t distinguish that sounds or »tremblings« with »CEV« I can teach you to use my »CEP«. It’s easy to use. If I happen to make a trip to Canada and if you’ll alove me to visit you J we can drive arround and try to find windy area and area with dips on the road and than we can put nails on the road and drive over them to cause tyre defect. Every event will have at least specific sound and specific trembling (perceptions)so some specific overall sensation. But I’m wondering how do you know for wind, surface defect or tyre defect as you don’t distinguish them on the highway ? You don’t have self-experiences ? Did somebody tell you about that ?

MT :

The Complex ENVIRONMENTAL Variable is a function in the environment of some variables in the environment. It corresponds to a perception inside an organism (a person, perhaps). If the variables are the environmental correlates of what the peripheral sensors send into the organism, then that’s what they are. But they don’t have to be. The point is that changing the value of the CEV will also change the value of the corresponding perceptual signal. It is something you perceive to be in the environment.

HB : I don’t know if Rick made a mistake when sending his post several times, or he wanted to tell us something about CEV. I see you alredy answered so I’ll let you chat about this one.

MT :

The temperature sensors produced a signal. If that signal corresponds to a feeling of hot or cold, then the perception is that there exists outside the body something that is hot or cold. That “something” has a property called “temperature”, which is the CEV corresponding to the perception of the temperature of the “something” in the environment.

HB : Agree…

MT :

[Maybe the following is an unnecessary complication at this point, but it’s worth remembering.]

We know that most of these functions change over time. For example, perceived brightness doesn’t have a fixed relationship with the flux of photons onto the retina, because we adapt to the overall level of illumination. The adaptation is slow compared to the rate at which illumination can change, so over short time periods there is a function relating changes in photon flux to changes in perceived brightness. The same seems to be true of any perceptual function I can think of at the moment. They all depend on history as well as what is currently coming from the sensors, but if the effect of history changes slowly compared to the changes in the sensors, it is still reasonable to say there is a functional relationship between perception and changes in the environmental variable.

[end complication]

HB :

On contrary. I see it as very necessary complication as it shows how vague perceptions can be. I can add to your »complication«, that not only adaptation and history of sensors activity (input function) but also structure of input function, »min and max trasholds«, refractory period… influence (non)linear transduction of »external stimuli« to »perceptual signal«. So we can see how »deformed« can be perceptual signal in comparison to »external physical stimuli«. Andi if we add upgrade of perceptual hierarchy to perceptual intensities, we see that quite odd »CEV« or »CEP« can be formed. It seems to be entiraly subjective, depending from all those specific characteristics of input functions and central nervous system.

So sometimes it’s hard to say what is »real« and what’s »illusion«, »hallucination« or what we didn’t want to perceive and what we wanted to perceive and what we missed to perceive…. But I agree with you : »it is still reasonable to say there is a functional relationship between perception and changes in the environmental variable«. We perceive and control quite efficially to survive.

From the rest of our discussion I concluded that we agree about »controlled and uncontrolled perceptions« that they go through the same »perceptual chanel« or as you wrote : »…all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors,… ». So I concluded that all perceptions are formed in the same input functions and end in comparator.

So my conclusion is that Bill’s generic diagram is right showing that all perceptions (uncontrolled and controlled) travel through the same perceptual input function to the comparator. Now I need to apologize to you as I was misleading you with saying that all these perceptions are controlled. You were right. They are not. But I affirm now that all perception (controlled and uncontrolled) are compared in comparator. Some of them are controlled and some are not. Some values of uncontrolled perceptions are added to controlled as you said, or uncontrolled perception are »turned« into controlled (for ex. sudden deviation of the car)… I have quite strong physiological evidence for this affirmation.

It seems also that you are right about Bill’s diagram that shows only controlled perceptions. I don’t know whether he ever mentioned what’s happening to uncontrolled or how uncontrolled are added to controlled perceptions, or otherwise dismissed or included in futher process of hierarchical control.

Boris

image00169.jpg

···

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Wednesday, August 27, 2014 11:53 PM
To: csgnet@lists.illinois.edu
Subject: Re: Environmental variables (was: Why we fight…)

[Martin Taylor 2014.08.27.13.48]

On 2014/08/27 12:51 PM, Boris Hartman wrote:

Martin,

sorry for the delay…J. It’s the beggining of school year…

MT :

Every perceptual function defines a CEV. If you happen to have a perceptual function that produces an output proportional to 3 times the temperature in degrees Kelvin plus twice the number of minutes since last midnight, then 3*(deg K)+2*(minutes) is the CEV for that perception. It’s a nonsense one, of course, but nonsense sometimes is useful to illuminate sense.

HB :

Do I understand right, that CEV is something what happens in sensor (afferent) nerves - output of »input function« ?

The Complex ENVIRONMENTAL Variable is a function in the environment of some variables in the environment. It corresponds to a perception inside an organism (a person, perhaps). If the variables are the environmental correlates of what the peripheral sensors send into the organism, then that’s what they are. But they don’t have to be. The point is that changing the value of the CEV will also change the value of the corresponding perceptual signal. It is something you perceive to be in the environment.

In your case if I understand right temperature receptors »transformed« some physical stimuli from outer environment into »CEV«, that »coresponds« to the temperature outside ?

The temperature sensors produced a signal. If that signal corresponds to a feeling of hot or cold, then the perception is that there exists outside the body something that is hot or cold. That “something” has a property called “temperature”, which is the CEV corresponding to the perception of the temperature of the “something” in the environment.

MT :

If you have a perceptual function, by definition its output is a perceptual signal that changes when the environmental variables that enter into it change.

HB:

Well I think I don’t understand this one clearly. Are you assuming that there is linear (precise) relationship between changes in values of environmental variables that enters in »input function« and changes in perceptual signal ?

No. Any function will do. I always assume that a perceptual function is nonlinear, but that doesn’t usually matter much when dealing with control unless you are doing very precise simulation.

[Maybe the following is an unnecessary complication at this point, but it’s worth remembering.]

We know that most of these functions change over time. For example, perceived brightness doesn’t have a fixed relationship with the flux of photons onto the retina, because we adapt to the overall level of illumination. The adaptation is slow compared to the rate at which illumination can change, so over short time periods there is a function relating changes in photon flux to changes in perceived brightness. The same seems to be true of any perceptual function I can think of at the moment. They all depend on history as well as what is currently coming from the sensors, but if the effect of history changes slowly compared to the changes in the sensors, it is still reasonable to say there is a functional relationship between perception and changes in the environmental variable.

[end complication]

So that changes in »outside physical variables« are varying in the same time and intensity with changes in perceptual signal ?

There are always delays and time-averaging incorporated in the perceptual process, and the “intensity” won’t vary in the same ratio as the perceptual signal, but you have the core of the idea correct.

There is exactly one outside physical variable defined by the processes that lead to one perceptual signal. That outside physical variable may be a very complicated function of the basic physical variables of mass, distance, force, voltage, time, etc. Even the response of a sensor is a very complex function of the basic variables. Sometimes we take the CEV of a sensor to be a basic physical variable, but that’s just as arbitrary as saying that the perceived taste of lemonade is not a physical variable. Sometimes we call “momentum” and “energy” basic variables even though they are functions of other basic variables. It’s all models and functions and what you want to do as an analyst, after all. To the perceiver, there’s just something “out there”, and for each perceptual variable we call what is “out there” a Complex Environmental Variable.

MT :

I have often said to you that Bill’s DIAGRAMs contain only controlled perceptions, because they are diagrams of how those perceptions are controlled. Uncontrolled perceptions have values that may or may not contribute to the values of controlled perceptions. It’s irrelevant in a diagram intended to show how control works. I don’t have any particular model for uncontrolled perceptions, nor do I think one is needed.

HB :

So If I understand right in first sentence you say that Bill’s diagram contain only controlled perception and in second sentence you say that uncontrolled perception have values that may or may not contribute to the values of controlled perceptions. So if I try to make a cocnlusion »uncontrolled perception« which contribute to the values of controlled perception go through the same »perceptual channel« as controlled goes. How could they otherwise contribute to controlled perception ?

O.K. maybe we could make some example to release misunderstanding.

Suppose that you are driving a car (it’s top example in PCT) and you are controlling the perception of front side of the car in just enough close range to reference perception of the middle right side of the street, that we say you have it »under control«. And you are looking also arround.

You don’t need to add “looking around”. All you need to say is something like: “There are shadows across the road that affect your perception of the edge, and maybe the sun gets into your eyes. Perhaps there is rain on the windscreen. You aren’t controlling those, but they do influence your perception of the relation of the car to the edge of the lane.”

I assume that beside »controlled perception«, you accept also »uncontrolled perception« of surrounding (maybe trees, fields, buildings, people and so on). I suppose that all these perception goes through the same pathway as you describe it : »…all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors,… ».

Yes. But say “pathways”, plural. There are at least millions, maybe billions of such pathways.

»Suddenly a «bump« of wind »pushed« your perception of the front side of the car »away« from reference (right side of the street). Is this »bump of the wind« in first moment controlled or uncontrolled perception ? Where does it go ?

You don’t perceive a “bump of the wind” unless you have mounted a detector for such things onto the hood of your car. You perceive the location of the car on the road, which has departed from its reference value. An outside observer might know that the error was due to a “bump of the wind”, but you can’t tell whether it was a dip in the road surface – at least I can’t. When I’m on the highway, I often wonder whether a deviation is cause by wind, road surface, or (I hope not) a defect in the car such as a tire going flat.

Martin

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Sunday, August 24, 2014 6:21 AM
To: csgnet@lists.illinois.edu
Subject: Re: Environmental variables (was: Why we fight…)

[Martin Taylor 2014.08.23.23.52]

On 2014/08/23 10:36 AM, Boris Hartman wrote:

Martin,

I’m not a good reader of your »English«, so I don’t know whether I understood some thoughts of yours, as you wanted to be understaood.

MT :

The problem with this, and it is a bone I have picked with Bill as well, is that the same is true of EVERY perception. Bill seemed to take an internally inconsistent view, saying on the one hand that all we ever know is our perception, no matter what kind of perception it might be, and on the other hand that there were some privileged environmental variables that are “real” and some that are not.

HB :

Where did Bill wrote that some privileged environmental variables are more real than others ? I really don’t recall »meating« this in his books. But I’m not some specialy good »English« reader. It’s possible that I missed something.

Rick references B:CP, 2nd Ed. p 112 as an example. On rereading it, I can see that the writing is ambiguous. I read it (as Rick seems to have done) as saying that there are environmental variable corresponding to first-level perceptions, but not to higher-level perceptions, which are constructed.

MT :

The argument that the taste of a given glass of lemonade is different for every person cannot hold water :slight_smile: There is no way that we can say that any perception in one person is the same as a perception in someone else. My perception of “blue” may not be yours, just as my perception of the trustworthiness of the neighbourhood loan shark or municipal politician may not be yours.

HB :

I really don’t see the point. In the first sentence it seems to me, that you are saying »that the taste of a given glass of lemonade is different for every person cannot hold water« so I understand it, that different persons have the same perception of taste (can not »hold water«). It seems to me that you are denying the existance of different taste of lemonade. In the other senstences I understand that perceptions are different for any person. Did I misssed something ?

No. I’m afraid my English was not clear enough. It’s “the argument” (for the non-existence in the environment of a variable corresponding to the taste of lemonade) that cannot hold water. I made no claim about whether different people, or the same person at different times, would or would not perceive the same taste from a particular glass of lemonade.

MT :

My preferred position on this is, I think, more consistent than Bill’s. I assume that there exists a “real world” that influences my sense organs and on which I can act in ways that influence the inputs to my sense organs. I know not what is in that world, but my nervous system constructs a perceived reality by operations on those sensory inputs (plus all the history of operations on earlier sensory inputs). Every perception thus produced defines some function of present and past states of the environment that have and had some influence on my sensors. Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

HB :

I never felt that Bill had different opinion from yours. But he did once have a discours of »solipsism« but he asked himself whether his thinking was »solipsist« view or not. Well whatever I never have a feeling that he is solipsist or something different from your perceptual understanding of »real world«.

Solipsism doesn’t come into this discussion, except that we have to dismiss it before we can have a discussion that includes talk about an external environment. Basically, it’s totally irrelevant. We agree it’s wrong, and then deal with our theories.

MT :

In this, I believe internally consistent, view, the brightness of a light, the taste of lemonade, the willingness of my dog to follow me, the autocratic tendencies of my political leaders, all have equal validity as existing states of the external environment. I call them Complex Environmental Variables, because they are environmental variables defined by complex perceptual functions. There’s no way to say that second, fifth, or ninth-level perceptions are of environmental variables more or less real than each other.

HB :

I really don’t imagine what »Complex Environemntal Variable« (CEV) can be ? If I assume that intensities as first level controlled variables which »integrate« into second level »sensations« and so on through the whole heirarchy, what is CEV ?

Every perceptual function defines a CEV. If you happen to have a perceptual function that produces an output proportional to 3 times the temperature in degrees Kelvin plus twice the number of minutes since last midnight, then 3*(deg K)+2*(minutes) is the CEV for that perception. It’s a nonsense one, of course, but nonsense sometimes is useful to illuminate sense.

Maybe your »perceptual mosaic« of the world which you assume to be »outside«. As you said. You beleive that your view is internally consistent. Others beleive that Bill’s view or their view is internally consistent. How can we know which »view« is internally consistent ? But it seems interesting what’s your theory of how we perceive the »world« outside. How does »Complex Perceptual Functions« (CPF) work ?

Are not all perceptual functions complex? I don’t mean in the mathematical sense of using numbers with real and imaginary components, but in the everyday sense of being rather complicated.

MT :

Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

HB :

So if I choose to control some variable in »environment«, are you saying that, it has no priority over other perceived variables in »environment« in that moment ?

It matters not whether you choose to control a perception. Any function you define creates a value. That value is a variable whose magnitude depends on the values of the inputs to the function. Perceptual functions are no different. If you have a perceptual function, by definition its output is a perceptual signal that changes when the environmental variables that enter into it change.

MT :

As for the taste of lemonade having no special physical effects on anything else, except the person tasting the mixture, that is true of every uncontrolled perception.

HB :

If I recall right you said that Bill’s model assume only controlled perceptions in the diagram. So where do »uncontrolled perceptions« go in your model ?

No. I have often said to you that Bill’s DIAGRAMs contain only controlled perceptions, because they are diagrams of how those perceptions are controlled. Uncontrolled perceptions have values that may or may not contribute to the values of controlled perceptions. It’s irrelevant in a diagram intended to show how control works. I don’t have any particular model for uncontrolled perceptions, nor do I think one is needed.

MT :

Since all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors,…

HB :

Do I understand right that all perceptions (controlled and uncontrolled) are equally constructed in nervous system ? So I suppose that we can assume that all perceptions go though the same perceptual »chanell« what is from physilogical view right. So how do we distinguish the pathway of »controlled and uncontrolled« perceptions in your theory ?

You don’t. The only difference is whether at any particular moment the perception is being controlled – whether you are actively doing something to keep it near its reference value, if it has one.

MT :

I have disagreed with Bill on this for a couple of decades without changing his opinion, and I don’t imagine my argument will change anyone else’s opinion. All the same, I thought it worth putting “out there” for the consideration of anyone who might be interested.

HB:

I’m always interested. But as far as I was talking to Bill and read his books I really don’t understand what was your and Bill’s disagreement ? It seems to me a little bit confusing. I think I always understood Bill’s theory and your understanding of it as something very similar, specialy as you frquently pointed out, that you talked with Bill for days and hours and that he »filled« some of your diagrams with his knowledge.

I have always believed that my understanding of Bill’s theory was very close to his. Our differences, as I perceived them, were always on the margins. For example, I have never been convinced by the specific levels Bill describes in his hierarchy. Neither was he, to judge from the way he inverted the order of levels in his models, and the number of times he explicitly said that his set of levels was always subject to revision. There have been many times I have disagreed with Bill and sometimes one of us changed our mind, other times when we agreed to disagree. But on the big things, I don’t think we disagreed much, if at all. This issue of whether all or only some environmental variable are constructed by perceptual functions is a very little thing.

Martin

Boris

From: csgnet-request@lists.illinois.edu [mailto:csgnet-request@lists.illinois.edu] On Behalf Of Martin Taylor
Sent: Saturday, August 23, 2014 5:26 AM
To: csgnet@lists.illinois.edu
Subject: Environmental variables (was: Why we fight…)

[Martin Taylor 2014.08.22.22.57]

[From Rick Marken (2014.08.22.1940)]

Martin Taylor (2014.08.22.14.41)–

RM: I think I mistakenly read this to mean that you were saying that control also involves reducing the variability of the external environment. But, in fact, you never said that; you were correctly implying that control involves reducing the variability of perceptual aspects of the environment, keeping them from varying around their “desired states”.

MT: That’s true, but as you often point out, controlling a perceptual (internal) variable also implies reducing the variability of the corresponding environmental variable.

RM: I would rather say that controlling a perceptual variable implies controlling the environmental correlate of the perceptual variable. The reason for this slight change in terminology is to emphasize the fact that there may be no environmental variable that corresponds to the controlled perceptual variable. For example, in B:CP (p. 112 of 2nd edition) Bill gives the example of controlling for the perception of the taste of lemonade. That perception is a construction (by a perceptual function) “…derived from the intensity signals generated by sugar and acid (together with some oil smells)… the mere intermingling of these physical components has no special physical effects on anything else, except the person tasting the mixture.” The same is true of many other perceptions that we control, such as the perception of variability (if we can perceive it). The variability of independent events that happen with different probabilities is an aspect of these events that we can perceive but it has no more physical significance than some other aspect of these events that we can perceive, such Morse code patterns.

The problem with this, and it is a bone I have picked with Bill as well, is that the same is true of EVERY perception. Bill seemed to take an internally inconsistent view, saying on the one hand that all we ever know is our perception, no matter what kind of perception it might be, and on the other hand that there were some privileged environmental variables that are “real” and some that are not. Since all perceptions are equally constructed by the operation of layers of perceptual functions on the firing patterns of neurons connected to sensors, on what grounds can one say a perception such as “intensity” is of an environmental variable more real than the taste of lemonade?

The argument that the taste of a given glass of lemonade is different for every person cannot hold water :slight_smile: There is no way that we can say that any perception in one person is the same as a perception in someone else. My perception of “blue” may not be yours, just as my perception of the trustworthiness of the neighbourhood loan shark or municipal politician may not be yours.

My preferred position on this is, I think, more consistent than Bill’s. I assume that there exists a “real world” that influences my sense organs and on which I can act in ways that influence the inputs to my sense organs. I know not what is in that world, but my nervous system constructs a perceived reality by operations on those sensory inputs (plus all the history of operations on earlier sensory inputs). Every perception thus produced defines some function of present and past states of the environment that have and had some influence on my sensors. Those functions describe a presumed reality “out there”, no one function, no one presumed environmental variable thus constructed having any priority over any other.

In this, I believe internally consistent, view, the brightness of a light, the taste of lemonade, the willingness of my dog to follow me, the autocratic tendencies of my political leaders, all have equal validity as existing states of the external environment. I call them Complex Environmental Variables, because they are environmental variables defined by complex perceptual functions. There’s no way to say that second, fifth, or ninth-level perceptions are of environmental variables more or less real than each other.

As for the taste of lemonade having no special physical effects on anything else, except the person tasting the mixture, that is true of every uncontrolled perception. If the person is not controlling the perception of the taste of the lemonade, any taste will do. But if the taste is a controlled perception and it doesn’t match its reference value, the person will act, and those actions definitely will have effects on something else, perhaps only by adding sugar to the mixture, perhaps by throwing the glass at the poor waiter who brought this foul tasting mixture.

I have disagreed with Bill on this for a couple of decades without changing his opinion, and I don’t imagine my argument will change anyone else’s opinion. All the same, I thought it worth putting “out there” for the consideration of anyone who might be interested.

Martin