Environmental affordance (was Freedom)

[From Bruce Gregory 92010.04.29.1235 EDT)]

[From Rick Marken (2010.04.29.0830)]

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;

BG: More precisely, it reminds me of how I used the object, or one similar to it, to exercise control in the past.

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.

BG: I am not sure that I follow you. Are you saying that a memory is not a perception? Or are you saying that memory plays no role in using a mouse to control a cursor in a tracking task? A thermostat is able to control with a memory so perhaps this is true of the PCT model. The memory “I used this mouse to control the cursor yesterday” plays no role in my control of the cursor today. Is this correct?

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.

BG: Let’s leave Hawkins out of this. As far as I know he has nothing to say about affordances. Unless, that is, you mean that Hawkins and I believe memory has a role to play in how we act. In which case, there is no need to tell, by looking at the mouse or the computer, whether the mouse “affords” control over the computer. You may have altered the program so that what worked yesterday no longer works today. You seem to be saying that in the PCT there is no need to invoke memory. That fits with the thermostat example as well as with the PCT tracking models as I understand them.

Bruce

Martin Taylor [2010.04.29.16.05]

[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”.

It has nothing to do with environmental degrees of freedom, except that
if the degrees of freedom are zero, there is no environmental
affordance.

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).

This is true but irrelevant.

Separately [From Rick Marken (2010.04.29.0830)], you said: "
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,…
."
The first part of this is correct, but the second may or may not be
correct in any particular case. The connection is as perceptible as is
any relationship. The influence of the output on the perception is a
relationship, but the relationship[ is not itself an environmental
affordance. The connection as an affordance is at a different level of
perception (typically a program perception), which, to repeat myself
from earlier messages, may be characterized: “If I move the mouse left,
the cursor will move left.” Someone new to computers would not be
expected to start with any perception of that affordance, but some
experimentation with the device is likely to allow the person to
generate it, in much the same way as one develops the perception that
if one sees the clouds thickening and darkening, it is likely to rain
soon (which is not an environmental affordance in itself, but which
could contribute to one).
Continuing, you said: “I presume it would 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.

Simply by looking at the mouse or computer, you can’t tell, as my
comment about the computer-naive person suggests. But if you are
familiar with computers that allow mouse control of the cursoe, and can
see that the mouse is connected to that particular computer, then you
can perceive the affordance. But the environmental affordance is there
whether you perceive it or not, which is why I queries a few messages
ago how to distinguish the affordances that exist (such as Rick’s
jailbreak) from those that are perceived.

Since we are getting into issues of the reality of the external world,
I think it is probably better to take the perception as the default
case, and use special language to deal with environmental affordances
perceived only by an external observer/analyst. The problem is that, to
take an extreme example, a jet plane was a possible way of crossing the
Atlantic in 1750, but nobody had perceived the possibility, or of the
steps needed to make it plausibility. It was, theoretically, an
environmental affordance for someone wanting to cross the ocean, but it
was not a practical one, whereas now it is a practical affordance.

There’s no clear sharp boundary between environmental affordances such
as that of the jet plane in 1750 (obviously impractical) and the plank
that could be placed across the stream (an obviously practical
environmental affordance). The only sharp boundary is perceptual. Those
that are perceived are the ones that could be used if some controlled
perception might use the corresponding environmental feedback path.
Those that are not perceived (or that have not been embodied in
evolution or reorganization) cannot be used.

Martin


···

not__

[From Rick Marken (2010.04.29.1550)]

Bruce Gregory 92010.04.29.1235 EDT)--

BG: �I am not sure that I follow you. Are you saying that a memory is not a
perception? Or are you saying that memory plays no role in using a mouse to
control a cursor in a tracking task?

Let's just say you don't follow me and leave it at that. We've been
through this many times. The answers to all of these questions are in
B:CP and other basic works on PCT.

You seem to be saying that in the
PCT there is no need to invoke memory. That fits with the thermostat example
as well as with the PCT tracking models as I understand them.

You've said this many times before. Of course I don't say this. There
is a whole chapter on memory in B:CP. Why not talk about what PCT is
about instead of what you think PCT is about?

Best

Rick

···

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

[From Rick Marken (2010.04.29.1610)]

Martin Taylor (2010.04.29.16.05)--

Separately [From Rick Marken (2010.04.29.0830)], you said: " 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,..."

The first part of this is correct, but the second may or may not be correct
in any particular case. The connection is as perceptible as is any
relationship.

The connection is not perceptible. The relationship between mouse
movement ad cursor movement is perceptible.

The connection as an affordance is at a different level of
perception (typically a program perception), which, to repeat myself from
earlier messages, may be characterized: "If I move the mouse left, the
cursor will move left." Someone new to computers would not be expected to
start with any perception of that affordance, but some experimentation with
the device is likely to allow the person to generate it, in much the same
way as one develops the perception that if one sees the clouds thickening
and darkening, it is likely to rain soon (which is not an environmental
affordance in itself, but which could contribute to one).

Of course, people have to learn (through experience) the means (lower
level perceptions to control) of controlling a perception. It takes
some experience with the mouse and computer display to learn to
control the cursor by controlling the position of the mouse. It
takes some experience tightening screws to control tightening by
exerting torque (controlling the angular velocity) of a particular
tool. I think the concept of affordance, as you describe it, is relly
unnecessary. You certainly don't need it to build a model that learns
to control a hierarchy of perceptions (as in Bill's
Little Man" model). Actually, the term came up in the concept of a
discussion of freedom. I think you said that freedom increases as the
affordances in the environment increase. This makes sense to me only
if affordances mean something like df. I think the best PCT discussion
of freedom in terms of degrees of freedom available for control is (of
course) Bill Powers' paper entitled "Degrees of Freedom in Social
Interaction" which is reprinted in LCS I, I believe (maybe it's LCS
II). I think that paper is the great start at a discussion of freedom
from a PCT perspective.

Best

Rick

···

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

[From Bruce Gregory (2010.04.29.1910 EDT)]

[From Rick Marken (2010.04.29.1550)]

Bruce Gregory 92010.04.29.1235 EDT)–

BG: I am not sure that I follow you. Are you saying that a memory is not a
perception? Or are you saying that memory plays no role in using a mouse to
control a cursor in a tracking task?

Let’s just say you don’t follow me and leave it at that. We’ve been
through this many times. The answers to all of these questions are in
B:CP and other basic works on PCT.

You seem to be saying that in the
PCT there is no need to invoke memory. That fits with the thermostat example
as well as with the PCT tracking models as I understand them.

You’ve said this many times before. Of course I don’t say this. There
is a whole chapter on memory in B:CP. Why not talk about what PCT is
about instead of what you think PCT is about?

BG: I’ll stop bothering you. I’m sure the answers to my questions are obvious to everyone on CSGnet but me. As far as I can tell PCT works perfectly well with no memory. Feedback control simply does not require memory. Controlling in imagination is not part of any working PCT model that I am aware of and does not, in any case require memory. I can see why you might be reluctant to discuss this, and I will honor your desire to ignore the question. No wonder you are not impressed by Jeff Hawkins. He has nothing to say that of any interest to you. In fact, as you have said on several occasions, you see little of interest in neuroscience. I agree with you, it has nothing to say that would inform PCT.

Bruce