Environmental affordance (was Freedom)

[Martin Taylor 2010.04.26.10.14]

[From Rick Marken (2010.04.25.2310)]

External reasons come down to not having the necessary
environmental affordances.
     

Oops. There is no such thing as an environmental affordance, according
to control theory.

Maybe we have a difference in the use of words. Let me define what I mean by "environmental affordance" by some examples. It is much the same as "environmental feedback path" except that "environmental feedback path" has the connotation that the path is being used for control, whereas "environmental affordance" means the path could be used for control but is not necessarily being used.

1. Everyday language: Suppose I am on one side of a river. There is a bridge. I could walk across the bridge and find myself on the other side.
     PCT rewording: If I ere to be controlling a perception of my location, with a reference to perceive myself on the other side of the river and a current perception of being on this side of the river. I might perceive the existence of a bridge, and perceive that if I used the bridge I could act so as to reduce the error in my perception of my location.
     The bridge is an environmental affordance.

2. I am doing a pursuit tracking experiment, controlling my perception of the location of a cursor with respect to a target. My computer has a mouse and a joystick and a trackpad, any of which could be used to influence the cursor. Each of these mechanical systems is an environmental affordance for controlling the perceived location of the cursor. Only one of them is in the environmental feedback pathway.

3. I find the current policies of my government differ from my perceptions of what I would prefer them to be. What environmental affordances do I have? I can vote just once every few years, I can write to my representative(s), I can join a protest movement on the streets, I can try to make my views known publicly, and so forth. All of these are environmental affordances, but few of them are environmental feedback pathways.

4. I am in a jail cell, and have a reference to be at home. I have no environmental affordances for controlling that perception.

So when we understand "Freedom" as the ability to control, we see that
the true enemies of freedom (as Dickens realized so long ago) are
ignorance (lack of education) and want (poverty).
   
One question in my mind is whether an environmental affordance is a possibility for the existence of an environmental feedback path as seen by some arbitrary observer, or whether an environmental affordance exists only if the potential control system has been reorganized so that the pathway can be used or is perceived and used in program-level analysis-type control. In other words, is it a property of the environment or of the developed control structure that includes its environment? If environmental affordances are properties of the environment, then "want" is a restriction of them, but ignorance is not. If they are determined by the reorganization, knowledge, and intellectual ability of the controller, then both ignorance and want are restrictions of environmental affordances.

Consider the case when I want to control my perception of my location. I am on this side of a deep, fast flowing stream, but have a reference to be on the other side. There is no bridge, but a strong plank is lying near me, and it is long enough to lie from bank to bank of the stream. I am inclined to say that the plank is an environmental affordance for my control of my location, but only if I am smart enough to realize that I could lay it across the stream to form a bridge. A more Newtonian approach would be to consider it an environmental affordance regardless of my skills.

   

A lack of environmental affordances probably could be characterized
in many ways, but I think a major split is between social affordances
(the kind Martin L listed in his response) and non-social affordances.
     

So even though we disagree about the existence of affordances we get
to the same place. Nice.

The availability of money provides the opportunity to change and
enhance environmental affordances. In other words, a rich person
has a lot more freedom than does a pauper.
     

Aside from the fact that I don't consider money an "affordance"
   
Do you, now, consider it an affordance?
   

"Freedom" can be split in another dimension, "freedom to" and>"freedom from".
     

I think both are implied when we see freedom as a person being "in
control". "Freedom to" simply means the ability to bring a controlled
variable to it's reference state; "Freedom from" is the ability to do
this while protecting the variable from disturbance.
   
I split them by environmental affordances. "Freedom to" means that environmental affordances exist for the controlled perceptions in question, whereas "Freedom from" means what you say, the non-existence of overwhelming disturbances.
   

That's a very crude first cut at what I think "freedom" means in PCT.
     

And a very nice one at that.
   
Thanks.
   

I suspect we all would wish to maximize our ability to control such
perceptions as we choose to control. Where I suspect we disagree is
in whether we would like to extend this wish to other people.
     

I presume you are talking to Martin L. here.

No, I don't think so, though his list inspired my comment. I remember a long-ago comment by an unremembered commentator that the difference between right-wing and left-wing politicians was in the list of things they wanted to prohibit (reduce or remove the environmental affordances for). Libertarians are a different breed. I think that my differences with Libertarians is analytic. I believe that if the kinds of "intrusions" of which they complain were eliminated, my own freedom, as well as that of most people including the complaining Libertarians, would be reduced rather than enhanced. I agree with them on objectives, not on means.

If we can agree that the term "freedom" is equivalent to what we mean
by being "in control" in PCT,

I don't agree with that, because we are actually "in control" of very little at any moment. To me "freedom" has more to do with the ability to choose what perceptions to control and the ability to control those perceptions than with the plain ability to control those perceptions we currently control. If I'm a billionaire, I could control to be taking a hoilday in a fancy resort on Tahiti, but maybe I prefer staying home. If I'm living hand to mouth, I could not have that choice. It's the fact that I could choose a wide range of perceptions to control at a wide range of reference values that to me signifies freedom.

  then maybe we can then deal with what
this means in the context of a society where each person's individual
ability to control depends on the controlling done by virtually
everyone else.

It always does, and that's the point of "just" laws.

  I think this is where people like Martin L. start
feeling "oppressed" by government.

You will have to ask him about that. I admit to feeling much as he does about many laws. I don't extrapolate it as far as he does, and in particular I think quite oppositely to him about taxation. But that's a matter for analysis, not a difference of basic philosophy.

Martin

[From Bruce Gregory (2010.04.26.1250 EDT)]

[Martin Taylor 2010.04.26.10.14]

Consider the case when I want to control my perception of my location. I am on this side of a deep, fast flowing stream, but have a reference to be on the other side. There is no bridge, but a strong plank is lying near me, and it is long enough to lie from bank to bank of the stream. I am inclined to say that the plank is an environmental affordance for my control of my location, but only if I am smart enough to realize that I could lay it across the stream to form a bridge. A more Newtonian approach would be to consider it an environmental affordance regardless of my skills.

BG: I am not sure how “smart enough” translates into the language of PCT. I assume that it involves the ability to exercise control in imagination. Is that what you have in mind?

Bruce

[From Rick Marken (2010.04.26.1030)]

Martin Taylor (2010.04.26.10.14) --

Let me define what I mean by
"environmental affordance" by some examples. It is much the same as
"environmental feedback path" except that "environmental feedback
path" has the connotation that the path is being used for control,
whereas "environmental affordance" means the path could be used for
control but is not necessarily being used.

OK, then I have no particular problem with it, though I think it's not
really very useful (see below). I thought you were talking about
Gibsonian "affordances" -- perceptual features of the environment that
inform the controller about how to use them -- which are, of course,
nonsense.

One question in my mind is whether an environmental affordance is a
possibility for the existence of an environmental feedback path as
seen by some arbitrary observer, or whether an environmental
affordance exists only if the potential control system has been
reorganized so that the pathway can be used or is perceived and
used in program-level analysis-type control.

I wouldn't waste my time on it if I were you. I think this notion of
"affordance" really isn't that useful. For example, in your jail
example you say:

4. I am in a jail cell, and have a reference to be at home. I have no
environmental affordances for controlling that perception.

Well, maybe, but maybe not. You just can't tell in advance whether or
not there are such affordances. If there is a successful jail break
then the affordances were there after all. Affordances can only be
detected post hoc; after the controlling is successful. If control is
not successful that doesn't necessarily mean that there were no
affordances; the controller might not have been successful because he
didn't discover these "affordances". Before 1903 it seemed like there
were no affordances for powered flight. But the Wrights showed that
such affordances were there all along.

Since affordances can't be identified in advance, so that we can know
whether or not an environment "affords" the ability to control some
perception (like getting home from jail) or not, it seems to me that
the concept of "affordances" is useless.

Do you, now, consider it [money] an affordance?

Sure, you can call it that. But I still don't think the concept is very useful.

If we can agree that the term "freedom" is equivalent to what we mean
by being "in control" in PCT,

I don't agree with that, because we are actually "in control" of very little
at any moment. To me "freedom" has more to do with the ability to
choose what perceptions to control and the ability to control those
perceptions than with the plain ability to control those perceptions
we currently control.

I don't know what you mean by "choosing perceptions to control". There
are two things in PCT that I think this could refer to. One is higher
order systems varying ("choosing") the references sent to lower order
ones. This is just part of the process of control so the ability to do
this is implicitly part of my definition of freedom as control. The
other is willfully (consciously) selecting a particular setting for a
reference when there is a conflict (two or more higher level systems
are trying to move that reference to two different states, as when the
system controlling for perceiving a sweet taste tries to set the
reference for apple pie to "large slice" and the system controlling
for perceiving thinness tries to set the reference for apple pie to
"none"). In a conflict situation like this where an arbitrary choice
is made there is actually no control. I don't see why you would want
to call that freedom. So I think I'll stick with my definition:
freedom is being in control.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.26.1400) EDT]

[From Rick Marken (2010.04.26.1030)]

Martin Taylor (2010.04.26.10.14) --

Let me define what I mean by
"environmental affordance" by some examples. It is much the same as
"environmental feedback path" except that "environmental feedback
path" has the connotation that the path is being used for control,
whereas "environmental affordance" means the path could be used for
control but is not necessarily being used.

OK, then I have no particular problem with it, though I think it's not
really very useful (see below). I thought you were talking about
Gibsonian "affordances" -- perceptual features of the environment that
inform the controller about how to use them -- which are, of course,
nonsense.

BG: One man's nonsense is another man's obvious fact. When I open my toolbox I find a collection of tools, each of which "informs me," via pattern recognition, as to how I might make use of the tool to exercise control. I pick the tool which constitutes the relevant affordance to the task at hand. I presume you call this an "environmental variable." But since the class of environmental variables or potential environmental disturbances is virtually unlimited, I'll stick with "affordances."

Bruce

[From Rick Marken (2010.04.26.1400)]

Bruce Gregory (2010.04.26.1400) EDT) --

Rick Marken (2010.04.26.1030)

OK, then I have no particular problem with it, though I think it's not
really very useful (see below). I thought you were talking about
Gibsonian "affordances" -- perceptual features of the environment that
inform the controller about how to use them -- which are, of course,
nonsense.

BG: One man's nonsense is another man's obvious fact. When I open > my toolbox I find a collection of tools, each of which "informs me," via > pattern recognition, as to how I might make use of the tool to exercise > control. I pick the tool which constitutes the relevant affordance to the > task at hand. I presume you call this an "environmental variable." But > since the class of environmental variables or potential environmental
disturbances is virtually unlimited, I'll stick with "affordances."

Good for you. But let's hope that the affordance of that Phillips
screwdriver doesn't inform you that it can be used to clear the wax
out of your ears.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.26.1706 EDT)]

[From Rick Marken (2010.04.26.1400)]

Good for you. But let’s hope that the affordance of that Phillips
screwdriver doesn’t inform you that it can be used to clear the wax
out of your ears.

It’s not the best tool for that purpose, but it would work. Unlike a hammer, for example.

Bruce

[From Rick Marken (2010.04.26.2220)]

Bruce Gregory (2010.04.26.1706 EDT)]

Rick Marken (2010.04.26.1400)--

Good for you. But let's hope that the affordance of that Phillips
screwdriver doesn't inform you that it can be used to clear the wax
out of your ears.

It's not the best tool for that purpose, but it would work. Unlike a
hammer, for example.

These affordances sound pretty amazing. You say that when you open
your toolbox (which, I presume, you do because it affords opening) you
find a collection of tools, each of which "informs you," via pattern
recognition, as to how you might make use of the tool to exercise
control. Could you explain how pattern recognition does this? How does
it inform you how you might use a tool to exercise control? A working
model of this would be really interesting to see as well.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.27.0703 EDT)]

[From Rick Marken (2010.04.26.2220)]

Bruce Gregory (2010.04.26.1706 EDT)]

Rick Marken (2010.04.26.1400)–

Good for you. But let’s hope that the affordance of that Phillips
screwdriver doesn’t inform you that it can be used to clear the wax
out of your ears.

It’s not the best tool for that purpose, but it would work. Unlike a
hammer, for example.

These affordances sound pretty amazing. You say that when you open
your toolbox (which, I presume, you do because it affords opening) you
find a collection of tools, each of which “informs you,” via pattern
recognition, as to how you might make use of the tool to exercise
control. Could you explain how pattern recognition does this? How does
it inform you how you might use a tool to exercise control? A working
model of this would be really interesting to see as well.

BG: How do you know that you are looking at a phillips-head screwdriver? A hammer? How do you know that each is a tool? How do you know what each is suited to do? The literature on pattern recognition is extensive. You might want to start with the Wikipedia, http://en.wikipedia.org/wiki/Pattern_recognition

Jeff Hawkins’s excellent book On Intelligence is a very nice introduction to the field of prediction and memory and pattern recognition in the context of neuroscience. Jeff’s company Numenta is pursuing the development of artificial intelligence software based on pattern recognition.

There are an abundance of working pattern-recognition models, if you have the time to explore them. For some reason I doubt your sincerity, but I would be pleased to discover that my doubts are groundless. I don’t think CSGnet is the appropriate forum for an extensive tutorial on pattern recognition, but I may be misjudging the level of interest.

Bruce

[Martin Taylor 2010.04.27.10.19]

[From Bruce Gregory (2010.04.27.0703 EDT)]

[From Rick Marken (2010.04.26.2220)]

Bruce Gregory (2010.04.26.1706 EDT)]

Rick Marken (2010.04.26.1400)–

Good for you. But let’s hope that the
affordance of that Phillips
screwdriver doesn’t inform you that it
can be used to clear the wax
out of your ears.

It’s not the best tool for that purpose,
but it would work. Unlike a
hammer, for example.

These affordances sound pretty amazing. You say that when you open

your toolbox (which, I presume, you do because it affords opening) you

find a collection of tools, each of which “informs you,” via pattern

recognition, as to how you might make use of the tool to exercise

control. Could you explain how pattern recognition does this? How does

it inform you how you might use a tool to exercise control? A working

model of this would be really interesting to see as well.

BG: How do you know that you are looking at a phillips-head
screwdriver? A hammer? How do you know that each is a tool? How do you
know what each is suited to do? The literature on pattern recognition
is extensive. You might want to start with the Wikipedia, http://en.wikipedia.org/wiki/Pattern_recognition


There are an abundance of working pattern-recognition models, if
you have the time to explore them.

Every perceptual input function is a pattern-recognition process,
whether its output is a continuous value for “the degree of redness” or
a logical value for “this is a hammer”. Higher-order (e.g.
program-level) perceptual functions employ logical analysis, such as
“Is this is a hammer I see before me? If so, I can use it to drive this
nail; if not I must seek a hammer”. It is at that level that I see
environmental affordances as being a useful concept.

Consider my example of the plank lying by the bank of a stream you want
to cross. At what level might the perception arise that “The plank is,
if moved, a possible means to provide an environmental feedback pathway
that does not at present exist, enabling the control of my perception
of location, which has a reference value of “on the other side””? That
complicated sentence is a possible description of a perception, is it
not? It is a perception of an environmental affordance that at present
does not have its reference value, but that by action can be brought to
its reference value.

Martin

[From Bruce Gregory (2010.04.27.1045 EDT)]

[Martin Taylor 2010.04.27.10.19]

Consider my example of the plank lying by the bank of a stream you want to cross. At what level might the perception arise that "The plank is, if moved, a possible means to provide an environmental feedback pathway that does not at present exist, enabling the control of my perception of location, which has a reference value of "on the other side""? That complicated sentence is a possible description of a perception, is it not? It is a perception of an environmental affordance that at present does not have its reference value, but that by action can be brought to its reference value.

BG: It is a perception in the language I call PCT-speak, where everything that is not an action is a perception. Are you saying that every tool has "a reference value?" If so, there is a neurological disorder characterized by the inability to resist the urge to pick up any tool in sight and attempt to make use of it.

Bruce

[Martin Taylor 2010.04.27.11.08]

This has to be quick. I intended to be on the road an hour ago!

[From Bruce Gregory (2010.04.27.1045 EDT)]

[Martin Taylor 2010.04.27.10.19]

Consider my example of the plank lying by the bank of a stream you want to cross. At what level might the perception arise that "The plank is, if moved, a possible means to provide an environmental feedback pathway that does not at present exist, enabling the control of my perception of location, which has a reference value of "on the other side""? That complicated sentence is a possible description of a perception, is it not? It is a perception of an environmental affordance that at present does not have its reference value, but that by action can be brought to its reference value.
     

BG: It is a perception in the language I call PCT-speak, where everything that is not an action is a perception.

Not so. In PCT everything is a signal. What we name a particualar signal depends on where in the control loop it occurs. One signal is an error value, another is a reference value, another is a perceptual value, and so forth. You know the rest. Bill P and I differ on whether to call the effects through the environment "signals". I do, he doesn't, so you might expect a slight difference of language depending on who is writing. I would call what is emitted from the output function an "output signal", and only loosely an "action". I tend to think of an "action" as an effect on the environment, rather than as the signal value.

  Are you saying that every tool has "a reference value?"

No. However, if you want to perceive yourself to be handling a tool, that perception has a reference value. To perceive something as being a tool requires only that there be some perceptual function that produces output that feeds into the "tool" category perceptual function in such a way as to cause it to produce a positive output (the "tool" category perceptual fucntion outputs a perceptual signal, for which the analysts says "he perceives a tool" when the signal value goes sufficiently positive).

What may have a reference value is the perception of an environmental affordance. "I want to perceive a way to cross that stream", which might be satisfied by perceiving (a logic or program level perception) "If I move that plank I can cross the stream". Perhaps in the process the person might perceive the plank as being a tool, perhaps not.

If so, there is a neurological disorder characterized by the inability to resist the urge to pick up any tool in sight and attempt to make use of it.

I don't understand this comment.

Martin

[From Rick Marken (2010.04.27.0910)]

Bruce Gregory (2010.04.27.0703 EDT)–

Jeff Hawkins’s excellent book On Intelligence is a very nice
introduction to the field of prediction and memory and pattern
recognition in the context of neuroscience…

I see that Martin Taylor (2010.04.27.10.19) has already replied to this and I agree with him when he says:

Every perceptual input function is a pattern-recognition process, whether its output is a continuous value for “the degree of redness” or a logical value for “this is a hammer”. Higher-order (e.g. program-level) perceptual functions employ logical analysis, such as “Is this is a hammer I see before me? If so, I can use it to drive this nail; if not I must seek a hammer”. It is at that level that I see environmental affordances as being a useful concept.

I’ve read Hawkin’s book and I’m not nearly as impressed with it as you are. But I agree with Martin that his efforts to build perceptual functions may prove to be quite valuable. Though I think the PCT approach to “pattern-recognition” as a hierarchy of more and more abstract perceptual variables, the higher level perceptions being functions of lower level ones, might prove to be a useful architecture.

I have nothing against “pattern-recognition”, by the way. I just want to figure out how the brain does it (I doubt that the amazing patten-recognition system that reads my hand written checks when I deposit them in the ATM, for example, recognizes the numbers written on the checks the way my brain does it).

Pattern -recognition is, as Martin says, a perceptual process and is, therefore, a central feature of control. But obviously perception alone can’t explain control; it’s got to be part of a closed negative feedback loop.

I think the concept of “affordance” – the idea that pattern recognition can detect aspects of things that can tell the behaving system (inform it about) how to use them – is simply an effort to explain purposeful behavior as driven (caused) by perceived properties of the environment (affordances). These affordances are just part of the old causal model of behavior that sees perception as the control of behavior. Every piece of research I have done proves that this view of behavior is wrong. What is actually happening is, of course, that behavior is the control of perception.

Best

Rick

···


Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.27.1247 EDT)]

[From Rick Marken (2010.04.27.0910)]

I’ve read Hawkin’s book and I’m not nearly as impressed with it as you are. But I agree with Martin that his efforts to build perceptual functions may prove to be quite valuable. Though I think the PCT approach to “pattern-recognition” as a hierarchy of more and more abstract perceptual variables, the higher level perceptions being functions of lower level ones, might prove to be a useful architecture.

BG: I am not sure what you mean by “abstract perceptual variables.” Higher level patterns are certainly patterns of patterns. All there are, are patterns.

I have nothing against “pattern-recognition”, by the way. I just want to figure out how the brain does it (I doubt that the amazing patten-recognition system that reads my hand written checks when I deposit them in the ATM, for example, recognizes the numbers written on the checks the way my brain does it).

Pattern -recognition is, as Martin says, a perceptual process and is, therefore, a central feature of control. But obviously perception alone can’t explain control; it’s got to be part of a closed negative feedback loop.

I think the concept of “affordance” – the idea that pattern recognition can detect aspects of things that can tell the behaving system (inform it about) how to use them – is simply an effort to explain purposeful behavior as driven (caused) by perceived properties of the environment (affordances). These affordances are just part of the old causal model of behavior that sees perception as the control of behavior. Every piece of research I have done proves that this view of behavior is wrong. What is actually happening is, of course, that behavior is the control of perception.

BG: I am beginning to suspect that this is a distinction without a difference. The question is not whether perception controls behavior or behavior controls perception. The question is whether behavior and perception are linked open loop or closed loop. If we agree that it is closed loop, then what is causing what takes on the aura of a theological dispute. Or so it seems to me.

Bruce

[From Rick Marken (2010.04.27.1040)]

Bruce Gregory (2010.04.27.1247 EDT)–

RM: These affordances are just part of the old causal model of behavior that sees perception as the control of behavior. Every piece of research I have done proves that this view of behavior is wrong. What is actually happening is, of course, that behavior is the control of perception.

BG: I am beginning to suspect that this is a distinction without a difference. The question is not whether perception controls behavior or behavior controls perception. The question is whether behavior and perception are linked open loop or closed loop. If we agree that it is closed loop, then what is causing what takes on the aura of a theological dispute. Or so it seems to me.

It makes a difference because “control” and “cause” are two different things.The distinction is subtle but crucially important. X can be a cause of Y and yet not control Y. Cause means something like “variations in X have an influence on variations in Y”. Control means something like “X specifies the value of Y”. The difference can be seen by considering a Y variable that is influenced by another variable besides X: Say Y = X + Z. If X is causally related to Y then variations in X will influence the value of Y but not determine what value Y will have, because the value of Y is also influenced by Z. But if X is controlling Y, then Y will equal X regardless of the value of Z, because the actions of the system – another variable, R – will null out the effect of variations in Z (the disturbance) on Y: so Y-Z+R = X and R is varied so that R+Z ~ 0; control theory is about how the system does that).

People who say “perception controls behavior” are actually saying perception “causes” behavior. If you agree that behavior is closed-loop – that perception causes behavior (output) and behavior causes perception simultaneously – then it is simply incorrect to say that perception controls behavior (Bill Powers has a nice paper on this called the “Asymmetry of Control” in LCS I). Also, if behavior occurs in a closed loop then what is causing what is not a theological dispute at all; perception causes behavior and behavior causes perception; both causal links exist simultaneously. The result is control of perception.

Best

Rick

···


Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.27.1425 EDT)]

[Martin Taylor 2010.04.27.11.08]

This has to be quick. I intended to be on the road an hour ago!

[From Bruce Gregory (2010.04.27.1045 EDT)]

[Martin Taylor 2010.04.27.10.19]

Consider my example of the plank lying by the bank of a stream you want to cross. At what level might the perception arise that “The plank is, if moved, a possible means to provide an environmental feedback pathway that does not at present exist, enabling the control of my perception of location, which has a reference value of “on the other side””? That complicated sentence is a possible description of a perception, is it not? It is a perception of an environmental affordance that at present does not have its reference value, but that by action can be brought to its reference value.

BG: It is a perception in the language I call PCT-speak, where everything that is not an action is a perception.

MT: Not so. In PCT everything is a signal. What we name a particualar signal depends on where in the control loop it occurs. One signal is an error value, another is a reference value, another is a perceptual value, and so forth. You know the rest. Bill P and I differ on whether to call the effects through the environment “signals”. I do, he doesn’t, so you might expect a slight difference of language depending on who is writing. I would call what is emitted from the output function an “output signal”, and only loosely an “action”. I tend to think of an “action” as an effect on the environment, rather than as the signal value.

BG: Fair enough.

If so, there is a neurological disorder characterized by the inability to resist the urge to pick up any tool in sight and attempt to make use of it.

I don’t understand this comment.

BG: I was simply describing a phenomenon. Most of us do not experience compulsive urges of the sort obsessives do. There is a disorder where the compulsion appears to involve anything that can serve as a tool. The sufferer feels compelled to pick the object up and use it. There is some evidence that whenever we see a tool, the brain recalls the use of the tool, but inhibits motor activity associated with actually using it, unless that is what we “intend” to do. The inhibitory mechanism appears to fail with certain obsessive disorders.

Bruce

[From Bruce Gregory (2010.04.27.1430 EDT)]

[From Rick Marken (2010.04.27.1040)]

People who say “perception controls behavior” are actually saying perception “causes” behavior. If you agree that behavior is closed-loop – that perception causes behavior (output) and behavior causes perception simultaneously – then it is simply incorrect to say that perception controls behavior (Bill Powers has a nice paper on this called the “Asymmetry of Control” in LCS I). Also, if behavior occurs in a closed loop then what is causing what is not a theological dispute at all; perception causes behavior and behavior causes perception; both causal links exist simultaneously. The result is control of perception.

Point taken.

Bruce

[Martin Taylor 2010.04.28.15.52]

[From Rick Marken (2010.04.27.0910)]

Bruce Gregory (2010.04.27.0703 EDT)–

Jeff Hawkins’s excellent book On Intelligence is a very nice

introduction to the field of prediction and memory and pattern

recognition in the context of neuroscience…

I see that Martin Taylor (2010.04.27.10.19) has already replied to this
and I agree with him when he says:

Every perceptual input function is a
pattern-recognition process, whether its output is a continuous value
for “the degree of redness” or a logical value for “this is a hammer”.
Higher-order (e.g. program-level) perceptual functions employ logical
analysis, such as “Is this is a hammer I see before me? If so, I can
use it to drive this nail; if not I must seek a hammer”. It is at that
level that I see environmental affordances as being a useful concept.

I’ve read Hawkin’s book and I’m not nearly as impressed with it as you
are. But I agree with Martin that his efforts to build perceptual
functions may prove to be quite valuable.

I’ve never seen the book, so I have no opinion on it.

I think the concept of “affordance” – the idea that pattern
recognition can detect aspects of things that can tell the behaving
system (inform it about) how to use them –

I’m surprised you got this notion of “environmental affordance” from my
writing. It seems to derive more from the Gibsonian idea of perceptual
affordances than from the concept I’ve been trying to get across.

is simply an effort to explain purposeful behavior as
driven (caused) by perceived properties of the environment
(affordances).

No, No NOOO!!! That’s in direct opposition to the idea of an
environmental affordance.

These affordances are just part of the old causal model
of behavior that sees perception as the control of behavior. Every
piece of research I have done proves that this view of behavior is
wrong. What is actually happening is, of course, that behavior is the
control of perception.

Yes, and if there is no environmental feedback path suited to the
current state of reorganization, a perception will not be controlled.
At higher (e.g. program) perceptual levels, substitute “known to the
perceiver” for “state of reorganization”. Think of it this way: “If I
am to achieve this, I will need something that allows me to do that”.
“Something that allows me to do that” is an environmental affordance.
As with many perceptions, the perception of an environmental affordance
may be controlled. When you think “I want to do that” along with “To do
that I need something that will let me do this”, the environmental
affordance “something that will allow me to do this” is a controlled
perception. If the something is already available, the environmental
affordance perception has no error. If not, that control loop has
error, and the person is likely to act to reduce the error by creating
the needed environmental affordance.

Does that help?

Martin

[From Rick Marken (2010.04.28.1520)]

Martin Taylor (2010.04.28.15.52)

Rick Marken (2010.04.27.0910)]
I think the concept of “affordance” – the idea that pattern
recognition can detect aspects of things that can tell the behaving
system (inform it about) how to use them –

I’m surprised you got this notion of “environmental affordance” from my
writing. It seems to derive more from the Gibsonian idea of perceptual
affordances than from the concept I’ve been trying to get across.

I know that your concept of affordance is different than Gibson’s. I was responding to Bruce Gregory, whose concept of affordance (along, apparently, with Hawkins’) is precisely Gibson’s. To prevent confusion I would suggest that you use a different name for the concept you’ve been calling affordance. I would suggest that you call what you are talking about “environmental degrees of freedom”. These are the characteristics of the environmental feedback path that either allow or inhibit control (as in my “Cost of Conflict” demo at http://www.mindreadings.com/ControlDemo/Conflict.html, where, when the degrees of freedom in the environment are reduced from 2 to 1 it is impossible to control the cursor in two dimensions).

Best

Rick

···


Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bruce Gregory (2010.04.28.1837 EDT)]

[From Rick Marken (2010.04.28.1520)]

Martin Taylor (2010.04.28.15.52)

Rick Marken (2010.04.27.0910)]
I think the concept of “affordance” – the idea that pattern
recognition can detect aspects of things that can tell the behaving
system (inform it about) how to use them –

I’m surprised you got this notion of “environmental affordance” from my
writing. It seems to derive more from the Gibsonian idea of perceptual
affordances than from the concept I’ve been trying to get across.

I know that your concept of affordance is different than Gibson’s. I was responding to Bruce Gregory, whose concept of affordance (along, apparently, with Hawkins’) is precisely Gibson’s. To prevent confusion I would suggest that you use a different name for the concept you’ve been calling affordance. I would suggest that you call what you are talking about “environmental degrees of freedom”. These are the characteristics of the environmental feedback path that either allow or inhibit control (as in my “Cost of Conflict” demo at http://www.mindreadings.com/ControlDemo/Conflict.html, where, when the degrees of freedom in the environment are reduced from 2 to 1 it is impossible to control the cursor in two dimensions).

BG: I have no idea how you arrived at this conclusion. I know nothing of Hawkins’s views on this topic, but I would be surprised if he differed with Martin. I know that I do not. Since you prefer to refer to this phenomenon in terms of environmental degrees of freedom, I will use that terminology.

Bruce

[From Rick Marken (2010.04.29.0830)]

Bruce Gregory (2010.04.28.1837 EDT)--

Rick Marken (2010.04.28.1520)]

I was responding to Bruce Gregory, whose concept of affordance (along,
apparently, with Hawkins') is precisely Gibson's.

BG: I have no idea how you arrived at this conclusion.

Because I used to be a fan of J. J. Gibson and his basic idea was that
perception is the detection of "invariances" in the sensory array. I
didn't read his "Theory of Affordances" but I believe that he saw an
"affordance" as something like an "invariance" inasmusch is is a
detectable aspect of the sensory array. What is being detected in the
case of an "affordance" is a "action possibility". This seemed to be
the way you saw "affordance", too, when you said that it was an aspect
of an object (like a tool) that "informs" a person about how it can be
used.

The difference between your definition of "affordance" and Martin's
seemed to be that you (like Gibson) treat an "affordance" as a
perception that tells you how to use an object as a component of
control; Martin, on the other hand, said that he was using
"affordance" simply to describe the action possibilities of the
environment -- the possibly non-perceptible physical characteristics
of the environment that make control possible or impossible. So for
Martin, the connection between the mouse and the cursor in a tracking
task is an "affordance" which makes control possible (if the
connection exists) or impossible (if it's broken). But the connection
itself cannot be perceived, so this is not an "affordance" in the
Gibson sense. I presume it would _not_ be an "affordance" for you or
Hawkins because you can't tell, by looking at the mouse or computer,
whether the mouse "affords" control over the cursor.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com