What's in a name (was Analyzing Feedback Paths)

[MartinT]

[Martin Taylor 2012.11.14.00.10]

Since I got back on the Western side of the Atlantic, I have been looking through the twin threads "Analyzing feedback paths" and "Affordances -- some added thoughts" with some bemusement. Things seem to have gone far away from what Kent was originally talking about, and I would never recognize the concept of "environmental affordance" from the discussions that have revolved around the phrase (usually omitting "environmental").

Much of the discussion seems to have considered single fixed feedback loops, whereas Kent was concerned about the liberty of an individual who acts in a world that contains a lot of stuff that both is used in feedback loops for control of that individual's many perceptions and is affected by natural processes and the actions of other control systems. control systems affects how multiple individuals affect each others' freedoms to act, such discussion is irrelevant to Kent's issues (and to that facet of my interests as well). So I won't make many specific comments on the past discussion.

Similarly on environmental affordances, I find it hard to reconcile most of the comments in the thread with my conception of an environmental affordance, Much of the criticism seems to be that Gibson used the same word for a different concept in the same semantic ballpark. I would be quite happy to use another word, if I could find one that fit the concept equally well, but so far I have not found an existing word or a created a neologism that satisfies me. Perhaps "environmental feedback affordance" could be shortened to "effordance", which is what I will use in what follows (if I remember).

A control system consists of several distinct path segments connected in series to form a loop. Some of these segments are inside the individual control hierarchy, normally inaccessible to an external observer, some are outside in the publicly accessible environment, and two -- effectors and sensors -- form the interface between inside and outside. The segments that are in the publicly accessible environment pass through physical objects and other independent control hierarchies. The fact that an object is used as part of the environmental feedback path is not an intrinsic property of the object, any more than the fact that a wire connects this switch to that light is an intrinsic property of the wire. The relevant property of the wire is that it can carry current if it is connected to the appropriate terminals.

An effordance is analogous to the wire except that it is a wire that must be connected to specific terminals. It is a possible carrier of the signal from a particular kind of effector output to sensory inputs that form part or all of a particular controlled perception -- a part of a possible environmental feedback path. It need not actually be used in practice, but it could be used if required. An effordance is not an intrinsic property of an object, but if the object is appropriately placed, it could form part of some particular feedback pathway.

For example, a plank does not have the property of allowing a person to cross a brook with dry feet unless it is placed across the brook. The ability of a person to use the plank to keep the feet dry and to control a perception of being on the other side of the brook is not an intrinsic property of the plank, but when placed across the brook, the plank provides an effordance for that particular control loop. It matters that the output effectors are producing "walking", and it matters that correcting error in the controlled perception involves crossing the brook. They all are part of the effordance of the plank. The plank provides the possibility of control, specifically for that control loop. The plank may lie across the brook but its effordance may not be used if the person instead decides to use the footbridge 50m upstream.

Effordances are important when considering social interactions because it is only through the environment that individuals interact. However, when one is interested only in a single isolated control loop, what matters is the ability to act effectively on the environmental complex that gives rise to the controlled perception.

For isolated control loops, it really does not matter how the linkage between output and the perceived environmental complex is achieved, provided that the parameters are understood. The concept of effordance is therefore not of much use if you are working only with a single loop for which the environmental feedback path varies only in its parameter values. The effordance concept is of use in conjunction with reorganization. Reorganization finds new ways for control to work, which includes (as the analyst-observer would see it) finding new paths through the environment through which the output can influence the perception. All of this can be analyzed without talking about effordances, although new environmental pathways discovered by the reorganization process necessarily take advantage of the effordances then present in the environment (obviously without any explicit representation of the fact in the newly functional control system).

The concept of effordance becomes useful when one is dealing with a changing environment, particularly if it is an environment changed by the actions of other control hierarchies. If someone takes the plank away to use in a construction, that effordance is lost to the system controlling the perception of being on the other side of the brook. This is the heart of Kent's "Exclusivity-Inclusivity" dimension in his discussion [From Kent McClelland (2012.11.02.1250)] of the characteristics of feedback paths as related to individual liberty. If another person can take an effordance from you, you have less liberty than you have if you can be sure of keeping that effordance available to you.

We have discussed the definition of freedom or liberty (they have slightly different connotations) more than once over the years. To me, an increase in freedom occurs if any of the following increase while the others hold constant:
(1) the number of perceptions that one might be able to control (in sport, to be an all-rounder);
(2) the ease of shifting which perceptions one is controlling from moment to moment ("keeping many balls in the air");
(3) the speed, accuracy, or range of disturbance that can be effectively countered in controlling a perception (these three might better be listed separately, but they fall together in Kent's concept of "Bandwidth");
(4) the ability to change environmental feedback paths for controlling any one perception (flexibility, not being constrained by habit, the availability of many parallel effordances).

Of these criteria, all but the second are constrained by the effordances in one's environment. For example, in social life, the number of perceptions one might be able to control and the range of disturbances against which one is able to maintain control are both strongly constrained by how much money one can access for use in transactions. Rich people have much more freedom than do those who must use most of their money if they are to be housed and fed.

For another example, the speed with which one can walk to correct an error of a few metres in one's location is very different in muddy, swampy ground than on solid turf. The "plank and the footbridge" provide an example of the parallel effordances considered in the fourth criterion. When the plank is removed by someone, the brook can still be crossed dryshod by using the bridge and walking a few metres along the bank (with luck, the walk is not muddy).

Just as the control of higher-level perceptions is effected by controlling a variety of lower-level perceptions, so an effordance for control of a high-level perception can usually be broken down into a set of effordances for control of the lower level variables. Walking to the bridge, crossing the brook, and walking back on the other side to get where you want to be provides an example of a feedback path that uses three effordances in series to wind up with the same end result as would be achieved by using the single effordance provided by the plank across the brook. Each of these three can also be seen as the environmental feedback path for a sequence of three controlled perceptions: to be at the near end of the bridge, to be at the other side of the bridge, and to be where the other end of the plank would have been. Each of these three controlled perceptions uses a different effordance, but "seeing myself over there on the other side" has one effordance, just as a connection between two terminal may consist of several wires connected in series.

I think that's enough to be going on with, but I think I might conclude with a quote:

"Perceptual Control Theory (PCT) shows that a �behavioral illusion� can occur when studying closed-loop control systems. The illusion is that an observed relationship between environmental inputs and behavioral outputs reflect characteristics of the system itself when it actually reflects properties of the feedback connection between the system�s output and a controlled perceptual input." (R. S.Marken, The Power Law: An Example of a Behavioral Illusion?, 2008)

I take "the properties of the feedback connection between the system's output and a controlled perceptual input", or at least that part of the connection accessible to outside observers, to be the properties of the effordance that is actually being used by the control system at the time of analysis.

Martin

···

From my point of view, unless the generic properties of individual

[Bill_Powers1]

[From Bill Powers (2012.11.19,1340 MST)]

Martin Taylor 2012.11.14.00.10

Much of the discussion seems to have considered single fixed feedback
loops, whereas Kent was concerned about the liberty of an individual who
acts in a world that contains a lot of stuff that both is used in
feedback loops for control of that individual’s many perceptions and is
affected by natural processes and the actions of other control systems.
from my point of view, unless the generic properties of individual
control systems affects how multiple individuals affect each others’
freedoms to act, such discussion is irrelevant to Kent’s issues (and to
that facet of my interests as well). So I won’t make many specific
comments on the past discussion.

Similarly on environmental affordances, I find it hard to reconcile most
of the comments in the thread with my conception of an environmental
affordance, Much of the criticism seems to be that Gibson used the same
word for a different concept in the same semantic ballpark. I would be
quite happy to use another word, if I could find one that fit the concept
equally well, but so far I have not found an existing word or a created a
neologism that satisfies me. Perhaps “environmental feedback
affordance” could be shortened to “effordance”, which is
what I will use in what follows (if I remember).

BP: I think my main point has been that one can’t determine the
affordances or effordances of any part of the environment just by
examining the environment. The reason is that *ffordances are expressed
in terms of the ideosyncratic purposes of the user or the inventor, not
in terms of universal physical properties of the environment. While I
would agree that some ways of organizing the environment favor the
achievement of certain purposes while others are neutral or detrimental
in that regard, I don’t think it is possible to predict what any
individual’s purposes will be in regard to available properties of the
environment.

As far as I can see, I don’t think you would disagree with any of that.
So that leaves me wondering why we need either term represented by
“*ffordance.” That, of course, may be just a matter of personal
preferences, so it would be inappropriate to speak prescriptively about
the term. But I dislike terms that according to ordinary usages imply
assumptions we think are false. That is what I see in this
passage:

MT: An effordance is analogous
to the wire except that it is a wire that must be connected to specific
terminals. It is a possible carrier of the signal from a particular kind
of effector output to sensory inputs that form part or all of a
particular controlled perception – a part of a possible environmental
feedback path. It need not actually be used in practice, but it could be
used if required. An effordance is not an intrinsic property of an
object, but if the object is appropriately placed, it could form part of
some particular feedback pathway.

BP: As far as I can see, the “feedback path” through the
environment has physical properties, but there are no constraints on how
any given set of properties might be used to achieve various ends. The
specificity of which you speak seems to me to be a rarity.

MT: For example, a plank does
not have the property of allowing a person to cross a brook with dry feet
unless it is placed across the brook. The ability of a person to use the
plank to keep the feet dry and to control a perception of being on the
other side of the brook is not an intrinsic property of the plank, but
when placed across the brook, the plank provides an effordance for that
particular control loop.

BP: And for an infinite variety of others, such as indicating to those
who care where the boundary between two properties lies in midstream or,
providing a place for people who like fishing to sit, or for lovers, a
place on which to engrave or paint sentimental messages, or for mothers,
an elevated place on which to stand while looking for a child. And there
is probably a rather large number of other items that would provide the
same *ffordance if placed appropriately – a rope stretched between trees
on opposite banks of the stream, a rowboat or lighter-than-air balloon
attached by ropes on both banks, and so on.
What I’m saying would be nitpicking but for the fact that some
sociologists want to claim that *ffordances influence us to carry out
specific behaviors. For that claim to hold true, it would have to be true
that only those behaviors are favored. It isn’t very useful to
claim that some abstract property of the environment-organism
relationship encourages only some unpredictable behavior among an
infinite set of behaviors. That would leave the probability of
encouraging any one specific behavior at essentially zero.

Obviously, if an organism succeeds in controlling some perception by
producing some organized effect in the environment, it follows that the
link from that organized effect to the state of the controlled perception
must *fford that causal relationship. There may be dozens of possible
links that would work, or if we include different constants of
proportionality as different *ffordances, an infinite number. But that
rules out the idea that specific objects have particular
*ffordances.

Best,

Bill P.

[Kent]

[From Kent McClelland (2012.11.19.1650)]

Martin’s message (2012.11.14.00.10) made a lot of sense to me, and Bill’s reply (2012.11.19,1340 MST–see below) seems a bit beside the point.

Bill is undoubtedly correct that any part of the physical environment could be used at any moment by a person in that physical space as part of a feedback loop for controlling an almost infinite number of perceptions, so that we can’t determine in advance
the whole list of “*ffordances” it might offer.

As Bill puts it, “While I would agree that some ways of organizing the environment favor the achievement of certain purposes while others are neutral or detrimental in that regard, I don’t think it is possible to predict what any individual’s purposes
will be in regard to available properties of the environment.”

A little further on he says, “What I’m saying would be nitpicking but for the fact that some sociologists want to claim that *ffordances influence us to carry out specific behaviors. For that claim to hold true, it would have to be true that only those
behaviors are favored.”

I guess that I’m one of the sociologists that he’s talking about here, because I do think that some *ffordances influence us to carry out specific behaviors, because they have been intentionally designed to do so by the people who have manufactured those
particular artifacts or who have set up the physical environment that way.

Furthermore, only those particular behaviors are favored, because the people whose active efforts to control their own perceptions have given shape to the parts of the environment that we are talking about, and who then make it their business to maintain
the physical properties of those parts of the environment according to their own specifications, intend to limit the range of behaviors (possible perceptions to control) by other people who use those parts of the environment. They want to make their favored
behaviors, whatever they are, far more likely to occur at that spot in the environment than any of an infinity of disfavored behaviors. Their control of other people’s behaviors is never absolute, but it’s often good enough for government purposes, as they
say.

Case in point. The seating options in McDonalds restaurants are designed, we’re told by a sociologist who has studied the matter, not to be too cosy and comfortable, so that patrons will eat their food and move on, rather than lingering as they might in
a pleasant coffee house. The avowed purpose of the design is to promote the rapid turnover of patrons and thus the profits of the restaurant chain.

Nothing would prevent a patron from lingering in a McDonalds for hours using the free wifi, except that his seat might begin to get sore. And nothing about the seating arrangements would prevent two patrons from taking off their clothes and using the uncomfortable
seats to make love, or doing something equally unanticipated by the designers. But most of the time most patrons will use the space simply to eat their happy meals and move on, exactly as the designers of the space intended.

To insist on the existential freedom of people as living control systems to do an infinite variety of things with the environmental objects available at any place at any moment may make sense from a PCT point of view, but it doesn’t seem very helpful as
a way of trying to understand the regularities that appear in the social world around us. It seems to me more like a theory of social chaos than a theory that might interest sociologists, who know very well that the social world is pretty chaotic, but not
entirely so. And sociologists would like to be able to explain the patterns and regularities that do emerge. I think PCT can offer some insights that can be useful to them, particularly if we start looking at the (manufactured) physical environment as a complex
of socially designed feedback paths.

Kent

···

On Nov 19, 2012, at 3:39 PM, Bill Powers wrote:

[From Bill Powers (2012.11.19,1340 MST)]

Martin Taylor 2012.11.14.00.10 –

Much of the discussion seems to have considered single fixed feedback loops, whereas Kent was concerned about the liberty of an individual who acts in a world that contains a lot of stuff that both is used in feedback loops for control of that individual’s
many perceptions and is affected by natural processes and the actions of other control systems. From my point of view, unless the generic properties of individual control systems affects how multiple individuals affect each others’ freedoms to act, such discussion
is irrelevant to Kent’s issues (and to that facet of my interests as well). So I won’t make many specific comments on the past discussion.

Similarly on environmental affordances, I find it hard to reconcile most of the comments in the thread with my conception of an environmental affordance, Much of the criticism seems to be that Gibson used the same word for a different concept in the same semantic
ballpark. I would be quite happy to use another word, if I could find one that fit the concept equally well, but so far I have not found an existing word or a created a neologism that satisfies me. Perhaps “environmental feedback affordance” could be shortened
to “effordance”, which is what I will use in what follows (if I remember).

BP: I think my main point has been that one can’t determine the affordances or effordances of any part of the environment just by examining the environment. The reason is that *ffordances are expressed in terms of the ideosyncratic purposes of the user or the
inventor, not in terms of universal physical properties of the environment. While I would agree that some ways of organizing the environment favor the achievement of certain purposes while others are neutral or detrimental in that regard, I don’t think it
is possible to predict what any individual’s purposes will be in regard to available properties of the environment.

As far as I can see, I don’t think you would disagree with any of that. So that leaves me wondering why we need either term represented by “*ffordance.” That, of course, may be just a matter of personal preferences, so it would be inappropriate to speak prescriptively
about the term. But I dislike terms that according to ordinary usages imply assumptions we think are false. That is what I see in this passage:

MT: An effordance is analogous to the wire except that it is a wire that must be connected to specific terminals. It is a possible carrier of the signal from a particular kind of effector output to sensory
inputs that form part or all of a particular controlled perception – a part of a possible environmental feedback path. It need not actually be used in practice, but it could be used if required. An effordance is not an intrinsic property of an object, but
if the object is appropriately placed, it could form part of some particular feedback pathway.

BP: As far as I can see, the “feedback path” through the environment has physical properties, but there are no constraints on how any given set of properties might be used to achieve various ends. The specificity of which you speak seems to me to be a rarity.

MT: For example, a plank does not have the property of allowing a person to cross a brook with dry feet unless it is placed across the brook. The ability of a person to use the plank to keep the feet
dry and to control a perception of being on the other side of the brook is not an intrinsic property of the plank, but when placed across the brook, the plank provides an effordance for that particular control loop.

BP: And for an infinite variety of others, such as indicating to those who care where the boundary between two properties lies in midstream or, providing a place for people who like fishing to sit, or for lovers, a place on which to engrave or paint sentimental
messages, or for mothers, an elevated place on which to stand while looking for a child. And there is probably a rather large number of other items that would provide the same *ffordance if placed appropriately – a rope stretched between trees on opposite
banks of the stream, a rowboat or lighter-than-air balloon attached by ropes on both banks, and so on.
What I’m saying would be nitpicking but for the fact that some sociologists want to claim that *ffordances influence us to carry out specific behaviors. For that claim to hold true, it would have to be true that
only those behaviors are favored. It isn’t very useful to claim that some abstract property of the environment-organism relationship encourages only some unpredictable behavior among an infinite set of behaviors. That would leave the probability of encouraging
any one specific behavior at essentially zero.

Obviously, if an organism succeeds in controlling some perception by producing some organized effect in the environment, it follows that the link from that organized effect to the state of the controlled perception must *fford that causal relationship. There
may be dozens of possible links that would work, or if we include different constants of proportionality as different *ffordances, an infinite number. But that rules out the idea that specific objects have particular *ffordances.

Best,

Bill P.

[rsmarken]

[From Rick Marken (2012.11.1710)]

Kent McClelland (2012.11.19.1650)--

KM: To insist on the existential freedom of people as living control systems to
do an infinite variety of things with the environmental objects available at
any place at any moment may make sense from a PCT point of view, but it
doesn't seem very helpful as a way of trying to understand the regularities
that appear in the social world around us.

RM: Actually, I don't think that is what is being insisted. What is
being insisted, I think, is that what an environmental object "makes
available" (affords) depends on the purposes of the person using that
object. What is being insisted, in other words, is that you have to
take _controlled variables_ into account before you can have any idea
what any particular "object" in the environment "makes available."

KM: It seems to me more like a theory
of social chaos than a theory that might interest sociologists, who know
very well that the social world is pretty chaotic, but not entirely so. And
sociologists would like to be able to explain the patterns and regularities
that do emerge.

RM: And they can, once they know what variables people control:
controlled variables. I think I can make this clearer using your own
example:

KM: Case in point. The seating options in McDonalds restaurants are designed,
we're told by a sociologist who has studied the matter, not to be too cosy and
comfortable, so that patrons will eat their food and move on, rather than
lingering as they might in a pleasant coffee house. The avowed purpose of
the design is to promote the rapid turnover of patrons and thus the profits
of the restaurant chain.

RM: And to the extent the design works it is because most people
control for lingering in cozy, comfortable places. That is, it is a
disturbance to a controlled variable that is corrected for by not
lingering. The design has no effect on people who are not controlling
for comfortably lingering (like me, when I had to wait for my car to
get fixed and stayed there for a couple hours playing with my iPhone).

KM: Nothing would prevent a patron from lingering in a McDonalds for hours using
the free wifi, except that his seat might begin to get sore. And nothing about
the seating arrangements would prevent two patrons from taking off their
clothes and using the uncomfortable seats to make love, or doing something
equally unanticipated by the designers. But most of the time most patrons will
use the space simply to eat their happy meals and move on, exactly as the
designers of the space intended.

RM: Only because most customers control for lingering comfortably (if
the design actually works). But to the extent that it works it is
because people are controlling for something for which the design is a
disturbance that is most easily corrected by leaving quickly. The most
important thing to take into consideration here is the _controlled
variable_. It's the thing that always gets left out of conventional
analyses of behavior (social or individual). It is the most central
concept in the control theory approach to understanding behavior.
Disturbances and feedback functions don't even exist until there is a
controlled variable. Once there is a variable that is under control --
a controlled variable -- then you get regular output (o) from regular
variations in the environment (disturbances, d) and a particular
feedback function, f(), relating output to controlled variable per
the disturbance-output relationship: o = -1/f(d) . So control theory
does explain why regular disturbances (d) and feedback functions (f())
produce regular behavior. But it all depends on the existence of
controlled variables; so you have to know what people are controlling
in order to understand why regular disturbances/feedback functions are
producing regular outputs

Best

Rick

···

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

[MartinT]

[Martin Taylor 2012.11.19.23.30]

[From Rick Marken (2012.11.1710)]

Kent McClelland (2012.11.19.1650)--
KM: To insist on the existential freedom of people as living control systems to
do an infinite variety of things with the environmental objects available at
any place at any moment may make sense from a PCT point of view, but it
doesn't seem very helpful as a way of trying to understand the regularities
that appear in the social world around us.

RM: Actually, I don't think that is what is being insisted. What is
being insisted, I think, is that what an environmental object "makes
available" (affords) depends on the purposes of the person using that
object.

Here you restate the main point of my long message (signed [Martin Taylor 2012.11.14.00.10] but posted today). Effordances incorporate the control action, the object, and the input to the perceptual signal being controlled. They are not intrinsic properties of objects.

  What is being insisted, in other words, is that you have to
take _controlled variables_ into account before you can have any idea
what any particular "object" in the environment "makes available."

Not as I read Bill's comment to which Kent was responding. I read Bill as saying that the range of purposes (controlling particular perceptions) that would have to be considered for any object is always effectively infinite. I read Kent as saying something that agrees better with my thinking, that most objects are ordinarily considered useful for only a few purposes, although an ingenious person might well find a few other uses for them. A plank is not normally considered useful if one wants to mow the lawn, but a bladed lawn-mower is, while the lawnmower would not ordinarily be considered useful for crossing a brook dryshod. Now it is probably true that one could use each as an effordance for control of the other perception (the plank for lawnmowing, the mower as means of crossing the brook), but it is not something that one would normally do if there were other objects that provided easier ways of controlling those perceptions.

As I emphasised in my long message, the concept of effordance is likely to be more useful in a social context than when considering single isolated control loops. If I want to perceive the lawn to have been mown by me, I would find it more difficult to execute the necessary control actions if someone has stolen my mower than if someone took away my plank. Vice-versa if the perception I want controlled is my location, with a reference value of "on the other side of the brook". Bill's "infinite availability" suggestion seems to equate the usability of both objects as part of the environmental feedback pathway for both controlled perceptions (and an indefinite number of others).

Martin

···

On 2012/11/19 8:10 PM, Richard Marken wrote:

[rsmarken]

[From Rick Marken (2012.11.20.1100)]

Martin Taylor (2012.11.19.23.30)--

RM: Actually, I don't think that is what is being insisted. What is
being insisted, I think, is that what an environmental object "makes
available" (affords) depends on the purposes of the person using that
object.

MT: Here you restate the main point of my long message (signed [Martin Taylor
2012.11.14.00.10] but posted today). Effordances incorporate the control
action, the object, and the input to the perceptual signal being controlled.
They are not intrinsic properties of objects.

RM: Great. So "effordances" is not only an ugly word, it is also an
unnecessary concept. It's already part of PCT.

RM: What is being insisted, in other words, is that you have to
take _controlled variables_ into account before you can have any idea
what any particular "object" in the environment "makes available."

MT: Not as I read Bill's comment to which Kent was responding. I read Bill as
saying that the range of purposes (controlling particular perceptions) that
would have to be considered for any object is always effectively infinite. I
read Kent as saying something that agrees better with my thinking, that most
objects are ordinarily considered useful for only a few purposes, although
an ingenious person might well find a few other uses for them. A plank is
not normally considered useful if one wants to mow the lawn, but a bladed
lawn-mower is, while the lawnmower would not ordinarily be considered useful
for crossing a brook dryshod. Now it is probably true that one could use
each as an effordance for control of the other perception (the plank for
lawnmowing, the mower as means of crossing the brook), but it is not
something that one would normally do if there were other objects that
provided easier ways of controlling those perceptions.

RM: So the concept of "effordance" tells us that a plank is better
than a lawnmover for crossing a stream and a lawnmower is better than
a plank for mowing the lawn? Is that what the new science of
effordances is about? We document all the variables that people might
want to control and list (in order of effectiveness) the objects that
can be used to control those variables? I must have that wrong. How
does this new science of effordances work?

MT: As I emphasised in my long message, the concept of effordance is likely to
be more useful in a social context than when considering single isolated
control loops.

RM: It doesn't seem so to me. I'm currently all alone and I can tell
you that this nice new computer of mine "effords" writing this email a
heck of a lot better than do the radio, lamp or printer that are also
within reach.

MT: If I want to perceive the lawn to have been mown by me, I
would find it more difficult to execute the necessary control actions if
someone has stolen my mower than if someone took away my plank. Vice-versa

RM: You will also find it harder to mow the lawn if the mower breaks
down all on its own. It seems that the concept of effordance is
equally useful (or useless) in social and non-social contexts.

Best

Rick

···

if the perception I want controlled is my location, with a reference value
of "on the other side of the brook". Bill's "infinite availability"
suggestion seems to equate the usability of both objects as part of the
environmental feedback pathway for both controlled perceptions (and an
indefinite number of others).

Martin

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

[Bill_Powers1]

[From Bill Powers (2011.11.21.1540 MST)]

Kent McClelland (2012.11.19.1650)]

KM: As Bill puts it, “While
I would agree that some ways of organizing the environment favor the
achievement of certain purposes while others are neutral or detrimental
in that regard, I don’t think it is possible to predict what any
individual’s purposes will be in regard to available properties of the
environment.”
A little further on he says, “What I’m saying would be nitpicking
but for the fact that some sociologists want to claim that *ffordances
influence us to carry out specific behaviors. For that claim to hold
true, it would have to be true that only those behaviors are
favored.”

I guess that I’m one of the sociologists that he’s talking about here,
because I do think that some *ffordances influence us to carry out
specific behaviors, because they have been intentionally designed to do
so by the people who have manufactured those particular artifacts or who
have set up the physical environment that way.

BP: My point is that the presence of an *ffordance does not, by itself,
influence anyone to break off from what he is doing and begin behaving in
a way that uses the *ffordance. The person must first want to achieve the
end that the *ffordance provides when used. If I don’t think the lawn
need mowing right now, the sight of the lawnmower will not urge me to
start using it (though it might remind me to have a look at the lawn). To
use Martin Taylor’s example, a plank does not actually *fford crossing a
stream unless one wants to cross the stream. Otherwise it could be picked
up and used for construction, to use Martin’s example a little further.
It *affords that use, too. It affords any use to which it is actually
put.

KM: Furthermore, only those
particular behaviors are favored, because the people whose active efforts
to control their own perceptions have given shape to the parts of the
environment that we are talking about, and who then make it their
business to maintain the physical properties of those parts of the
environment according to their own specifications, intend to limit the
range of behaviors (possible perceptions to control) by other people who
use those parts of the environment. They want to make their favored
behaviors, whatever they are, far more likely to occur at that spot in
the environment than any of an infinity of disfavored behaviors. Their
control of other people’s behaviors is never absolute, but it’s often
good enough for government purposes, as they say.

BP: The wild card in this view is the unpredictability of any one
person’s intentions. The object in question has been carefully designed
to make it easy to produce one particular outcome when one particular way
of using the object (i.e., a particular behavior) is employed. But
whether the person uses the object at all depends on whether the outcome
it produces when properly used is wanted. If the object is in fact used
for the engineer-imagined purpose, then of course the behavior required
for using it is carried out. So there will be some correlation between
the properties of the object and the way it is actually used, even if the
object has no influence on anyone’s using it. The actual driving force is
the reference signal that specifies that the outcome is to occur. In
order for someone to achieve that outcome, a means of producing it must
be used. Presumably, a carefully-engineered tool will be easier to use
than some chance object, so after some experience it will probably be the
tool that is chosen. And once that tool is chosen, its engineering
determines what physical actions are required in order to use it most
effectively.

Will you hand me that nutcracker? I have to get the screw-on cap off this
bottle.

So we’re looking at another instance of the behavioral illusion, aren’t
we? It seems as if the object is causing the behavior of using it in a
particular way, where in PCT we would say instead that it is an error
signal in a control system that determines how the object will be used.
The actions used to wield the object, the tool, will be varied until they
have the greatest effect on the controlled variable. If the tool designer
has done a good job, that will involve using the tool in a certain way,
which the user can discover and with practice make the default way of
behaving.

That latter way of describing the situation makes the organism, not the
object, the initiator of the behavior.

I think my present conclusion is that *ffordance is indeed an example of
the behavioral illusion. It reverses the actual direction of causation,
in the same way that Skinner did when he said that behavior is controlled
by its consequences, when it is perfectly clear that the consequences are
caused by the behavior.

KM: Case in point. The seating
options in McDonalds restaurants are designed, we’re told by a
sociologist who has studied the matter, not to be too cosy and
comfortable, so that patrons will eat their food and move on, rather than
lingering as they might in a pleasant coffee house. The avowed purpose of
the design is to promote the rapid turnover of patrons and thus the
profits of the restaurant chain.

BP: So if an *ffordance influences behavior, it is the design of the
seating that influences the customer to leave quickly. That’s how it
looks, but I claim again that this is the behavioral illusion. I would
say that the seating is engineered to cause discomfort if used for a long
time, meaning that it will disturb a controlled variable somewhere in the
nether regions. The control system will, when the discomfort becomes
great enough, take some action such as getting up and leaving that has an
equal and opposite effect on the discomfort. This interpretation would
mean that there is no direct connection between the design of the seating
and the subsequent behavior. Between those two elements there is a
controlled variable, and the design of the seating generates a
disturbance (when sat upon) that the subsequent behavior
opposes.

Is there any example of *ffordance that doesn’t fit this
pattern?

KM: Nothing would prevent a
patron from lingering in a McDonalds for hours using the free wifi,
except that his seat might begin to get sore. And nothing about the
seating arrangements would prevent two patrons from taking off their
clothes and using the uncomfortable seats to make love, or doing
something equally unanticipated by the designers. But most of the time
most patrons will use the space simply to eat their happy meals and move
on, exactly as the designers of the space intended.

BP: Yes, agreed. But is that a property of the seats or a property of the
patrons? If I stick a pin in your hand and you withdraw the hand, can I
conclude that you are hooked up inside so the sensory signals generateed
by the pin end up in the muscles that move the hand? That’s how it looks,
and it’s what every real scientist believed for a long time. But that’s
not how it works. The pin causes pain, and the pain results in an error
signal, and the control system acts to bring the pain signal to zero, its
reference level. You may sometimes notice that there are patrons in
McDonald’s who seem to have actually pushed a pin through a sensitive
body part like a tongue or a nostril and left it there: their reference
level for pain is different from ours.

This is actually a little taste of what I would like to see in the
“joint project.” I would outline the PCT explanation and
propose that *ffordance is an incorrect interpretation of an ordinary
control process. You could then refute my proposition.

Considering that Colorado has recently decided to legalize the possession
and use by pot by adults, perhaps we should stop referring to the joint
project.

The *ffordance is an illusion; it reverses cause and effect. Considering
that circular causality is somewhat hard to understand, I will generously
admit that nobody can be blamed for this (unless they now go on believing
in *ffordance).

KM: … sociologists would like
to be able to explain the patterns and regularities that do emerge. I
think PCT can offer some insights that can be useful to them,
particularly if we start looking at the (manufactured) physical
environment as a complex of socially designed feedback
paths.

I think the same thing. And I think our explanation is better than
Gibson’s. It is the organism, not the object that appears to be doing the
affording, that decides how the object will actually be used.

Best,

Bill P.

[MartinT]

[Martin Taylor 2012.11.23.09.24]

Kent, I'm with Bill on this. Granted, it is often the case that one

sees something that leads one to do something for which one had not
been controlling. For example, the sight of some Lindt chocolate
bars on sale at $2.99 (the ones I am talking about usually sell for
$7.99) might induce me to buy chocolate even though I had had no
thoughts about chocolate when I left home. You might say that the
effordance of cheap chocolate influenced me to buy, and in a sense I
suppose it did. But that’s not the way I would see it through PCT
spectacles. If I did not like chocolate, I would not buy it, even at $2.99.
Suppose the offer had been of roasted mashed maggots for $2.99 when
they usually sell for $7.99. Would you buy them? Some people might,
but I wouldn’t. When I say I “like chocolate”, what do I mean? I
think I mean that I have a positive reference value for perceiving
the taste and texture of chocolate in my mouth. I am not currently
experiencing that perception, so there is an error. But the error is
not big enough to allow that control system to win the conflict with
a control system with a reference value for having money in hand
when the money control system would increase its error by the
perceptual equivalent of $7.99. It is big enough to win the conflict
when the money control system would increase its error by $2.99.
Likewise, If I cannot see anywhere selling chocolate, I have no way
to reduce the error in my “liking chocolate” control system.
I think that most of these apparent influences of effordances can be
seen as examples of conflict, and all of them are either that or the
appearance of a new effordance where none were previously available
to some control system that was experiencing error. A prisoner in
jail has no effordance that would allow him to experience a holiday
on a Samoan beach. But usually the adage “Where there’s a will,
there’s a way” holds true. Whether to use that way usually depends
on how its use would conflict with other perceptions being
controlled. The conflict might be within the person or between that
person and some other(s), as when I want to cross the brook over the
plank and you want to use the plank for construction.
To repeat Bill’s statement by way of agreeing with it: “the presence
of an *ffordance does not, by itself,
influence anyone to break off from what he is doing and begin
behaving in
a way that uses the *ffordance.” However, the appearance of a new
effordance may change the balance of a conflict in a way that to an
outside observer looks as though the effordance did influence the
person to “break off from what he is doing and begin behaving in
a way that uses the *ffordance.” Alternatively, it may provide a
means for some control system to reduce its error where none was
previously available. As I said in the post that introduced this
subject heading [Martin Taylor 2012.11.14.00.10, posted 11.19]
, this appearance is an example of the behavioural illusion, albeit
in a form different from the way the behavioural illusion is usually
presented.
Here’s where I part from Bill, in the use of the word, but not in
the science.
As I would use the word, the plank does efford crossing a brook
dryshod, whether one uses it that way or not. To me, if something
can be used as part of an environmental feedback path for control of
some particular perception, it provides that effordance. As Bill has
pointed out previously, most objects could, in principle, have an
indefinite number of effordances. I don’t find that to be an
objection to the concept, any more than I find the fact that an
object has an indefinitiely large number of properties to be an
objection to describing one of those properties. It is only when a
control system exists that has a perception that could be influenced
by some action on the object and has the means to effect that action
that the effordance becomes actual – whether or not the control
system does in fact use it. To get to the shops, my car and my bike
both provide the necessary effordance, because I know how to ride a
bike and to drive a car.
As for Bill’s use of “*ffordance”, I have come to appreciate a
reason why “effordance” is a PCT-suitable word whereas “affordance”
is not. Initially, I thought of “effordance” as a shortening of
“environmental affordance → e-affordance”, and as rather ugly.
But consider the differences between these pairs of words: affect
vs. effect, adit vs. exit, accept vs. except, etc. In each case, the
“a-” word has an inward connotation of something done to, movement
toward, inclusion, whereas the corresponding “e-” word has an
outward connotation, doing to something, going somewhere, actively
avoiding. I am passively affected by something, but I actively
effect some change. Likewise, an “effordance” is something one uses
in the environment, whereas an “affordance” has a connotation of
being imposed by the environment.
Yes, indeed. And I agree with most of what follows in Bill’s posting
until …
Here Bill comes to a point from which I started but use it to come
to the opposite conclusion. I used a Marken quote to define the
behavioural illusion: “The illusion is that an observed relationship
between environmental inputs and behavioral outputs reflect
characteristics of the system itself when it actually reflects
properties of the feedback connection between the system’s output
and a controlled perceptual input.” An outside observer might see
someone use an effordance and conclude that the effordance
influenced the person to emit a certain behaviour. The observer has
succumbed to the behavioural illusion. The effordance itself is not
an illusion of any kind. It is just a condition in a person’s
environment through which a person can control a perception.
It does no such thing. Your imagined theorist might reverse cause
and effect, but the effordance does not. You can put up all the
straw theorists you like so that you can easily knock them down, but
that does not change the situation to be explained. The connotation
of “Affordance” may well seem to reverse cause and effect.
“Environmental affordance”, now shortened to “effordance” does not,
and never did, either in this series of threads or in earlier times
when the usage was also disputed.
Yes, exactly. But why bother to bring up Gibson yet again? Has
anyone in any of these related threads tried to support his views?
Having agreed to all that, and if Kent also agrees with both of us,
the next step should be to address what Kent was targeting in his
five characteristics of feedback paths that increase an organism’s
liberty [From Kent McClelland (2012.11.02.1250)]. To refresh the
memory, they were entitled:
I would like to address these in the language of effordances rather
than of actualized feedback paths.
Martin

···

[From Bill Powers (2011.11.21.1540 MST)]

  Kent McClelland (2012.11.19.1650)]

    KM: As Bill puts it,

“While
I would agree that some ways of organizing the environment favor
the
achievement of certain purposes while others are neutral or
detrimental
in that regard, I don’t think it is possible to predict what any
individual’s purposes will be in regard to available properties
of the
environment.”

    A little further on he says, "What I'm saying would be

nitpicking
but for the fact that some sociologists want to claim that
*ffordances
influence us to carry out specific behaviors. For that claim to
hold
true, it would have to be true that only those behaviors
are
favored."

    I guess that I'm one of the sociologists that he's talking about

here,
because I do think that some *ffordances influence us to carry
out
specific behaviors, because they have been intentionally
designed to do
so by the people who have manufactured those particular
artifacts or who
have set up the physical environment that way.

  BP: My point is that the presence of an *ffordance does not, by

itself,
influence anyone to break off from what he is doing and begin
behaving in
a way that uses the *ffordance. The person must first want to
achieve the
end that the *ffordance provides when used.

  If I don't think the lawn

need mowing right now, the sight of the lawnmower will not urge me
to
start using it (though it might remind me to have a look at the
lawn). To
use Martin Taylor’s example, a plank does not actually *fford
crossing a
stream unless one wants to cross the stream.

  Otherwise it could be picked

up and used for construction, to use Martin’s example a little
further.
It *affords that use, too. It affords any use to which it is
actually
put.

  I think my present conclusion is that *ffordance is

indeed an example of
the behavioral illusion. It reverses the actual direction of
causation,
in the same way that Skinner did when he said that behavior is
controlled
by its consequences, when it is perfectly clear that the
consequences are
caused by the behavior.

  The *ffordance is an illusion; it reverses cause and effect.

Considering
that circular causality is somewhat hard to understand, I will
generously
admit that nobody can be blamed for this (unless they now go on
believing
in *ffordance).

    KM: ... sociologists

would like
to be able to explain the patterns and regularities that do
emerge. I
think PCT can offer some insights that can be useful to them,
particularly if we start looking at the (manufactured) physical
environment as a complex of socially designed feedback
paths.

  I think the same thing. And I think our explanation is better than

Gibson’s. It is the organism, not the object that appears to be
doing the
affording, that decides how the object will actually be used.

1.	Exclusivity—Inclusivity.
2.	Flexibility of Purpose
3.	Bandwidth (which I separated into PoPS -- Power, Precision, and Speed)
4.	Fixity—Evanescence
5.	Manipulability.
Hypothesis: An individual’s liberty and flexibility of action is increased when the individual has access to feedback paths that are exclusive, multi-purpose, high bandwidth, and fixed but highly manipulable.

[Kent]

[From Kent McClelland (2012.11.23.2124 CST)]

Martin Taylor (2012.11.23.09.24)

Hello again all,

Having roused myself from my illness-induced lethargy, and having read back over the posts in this thread, I’ve decided that, in fact, I agree with both Bill and Martin, and with regard their points of difference, which seem to be mainly terminological, I don’t
have any strong opinion.

My argument, as I had phrased it, was incorrect. I said that “some *ffordances influence us to carry out specific behaviors,” and that’s obviously wrong. Whether an object in the environment is used by an individual as a feedback path for controlling a particular
perception depends entirely, as Bill and Rick have been at pains to point out, on the perception that the individual wants to control. It’s not the object influencing the person, but the person using the object for his own purposes.

I guess you could describe my argument as an instance of the behavioral illusion, but that’s not exactly the problem that I see with it. A central feature of PCT is its reliance on closed-loop analysis. A circle of causality stretches from the person’s perceptual
input functions to reference values in the HPCT hierarchy to the person’s output functions and then back through a feedback function in the physical environment, combining with disturbances in that environment, to affect the person’s input functions. In that
negative-feedback loop, the active agent of control is the person, as he attempts to maintain a match between his reference values and his current perceptions of what is happening in the environment. To suggest that something in the environment “influences”
the person to behave in a certain way is to locate the agency in the environment, not the person.

However, it still seems to me that there’s a disconnect between the argument that I’ve been attempting to make in this thread (not very successfully, it appears) and Bill and Rick’s responses to it. What Bill and Rick have been saying is exactly right as long
as your frame of reference is the individual. But, as Martin has been pointing out, and very eloquently in the initial post of this thread, my frame of reference is social interaction. I’m interested as a sociologist in questions of social influence, manipulation,
and social power more generally. I’ve been talking about the environment, but what I’m ultimately interested in are relationships between people, how people influence and manipulate each other when they want to control their own perceptions of other people’s
behavior. My frame of reference is not the individual, but multiple individuals interacting with each other in a shared environment.

The only way that a person can influence or manipulate another person’s behavior is for the first person to control his or her perception of some variable in their shared physical environment. A classic rubber-band experiment describes the situation well. The
“experimenter” moves one end of the dual rubber bands, and the “subject” moves the other end while attempting to control the perception of keeping the jointure of the two bands over some fixed point. The experimenter can manipulate the physical actions of
the subject by moving her own end in some geometric pattern, and in order to keep his own perception in control the subject begins (perhaps unconsciously) to move his end in a geometric pattern that mirrors the experimenter’s. My point is that the manipulator
must control something about their shared environment, in this case the manipulator’s own physical movements, in order to control her perceptions of the other person’s behavior.

When Bill or Rick have talked about social interactions in the past, they have usually described it, if my memory is correct, in terms of one person providing disturbances to perceptions controlled by another. In the case of the kinds of interactions that I
am talking about, however, disturbances seems like not quite the right metaphor. The word ‘disturbance’ carries with it the connotation of fluctuation, something unpredictably changing, although of course a disturbance could also be simply a steady pressure
in one direction or another. The physical arrangements that I am talking about are more like steady pressures than fluctuating disturbances. They are portions of the environment that have been stabilized by one set of people, the manipulators, as they control
their own perceptions in hopes of influencing the actions of another set of people, just as the experimenter in the rubber band experiment creates a stable pattern of movement of her own hand motions in order to have the perception of the subject’s hand moving
in a similar pattern.

Let’s get back to my McDonald’s example. Although I’m not particularly fond of the food at those restaurants, I often find myself stopping there when I’m traveling on the highway. I have a predictable set of perceptions that I would like to control. After a
certain number of hours behind the wheel, I’m hungry and would like to perceive some food in my stomach. Even more, at my age, I need to visit a toilet. Lots of other travelers on that highway are likely to be dealing with similar sets of bodily disturbances.
I know that there is likely to be a McDonald’s at the next interchange, and I know that if I stop there I will be able to get some food quickly, relieve my bladder, and get back on the road in short order. At any given interchange, there is likely to be more
than one feedback path available for controlling these perceptions, and in almost every case I would be able to find better food and a more elegant bathroom were I willing to take the time to venture farther away from the highway interchange. But the McDonald’s
is likely to be located right at the interchange. It’s predictable, easy, and cheap. And the higher-level perception that I’m controlling is to reach my destination as quickly as possible. The McDonald’s for me is “the (feedback) path of least resistance.”
When my controlling a given perception is simply a means to an end, not my primary purpose in view, I often take the feedback path that is closest to hand and easiest to use. And most others do the same, I would argue.

We can analyze this situation sociologically as an interaction between me and the owners of the McDonald’s chain. I think it is reasonable to see them as having manipulated me, although I’ve just been controlling my own perceptions, as always happens in cases
of manipulation, because manipulation from a PCT point of view involves taking advantage of a person’s focus on maintaining control of one set of perceptions to change the person’s actions in other ways. By locating their restaurants at the interchanges of
busy highways, by having a standardized menu and standardized bathrooms, as well as a standardized architectural appearance and signage, they manipulate me and travelers like me to choose their restaurants as the feedback path for controlling predictably recurring
perceptions we are likely to have.

In this social interaction between owners and patrons, the McDonald’s owners are clearly the more powerful party, as is evidenced by their ability to modify large parts of our shared physical environment in accordance with their architectural plans, as
well as their ability to get a host of employees to carry out the routines of these restaurants according to their routinized specifications (or, in other words, the owners’ reference values). Patrons of McDonald’s are individually powerless, although if they
organized to boycott the chain they might have some countervailing force. It matters only infinitesimally to McDonald’s bottom line whether I choose their restaurant some other as a feedback path for controlling my perceptions on my trip. And from these transactions
between owners and patrons, the owners, arguably, have more to gain more than the patrons. The owners gain tidy profits, while the patrons only gain bigger waistlines from all the fattening food they sell. (Does this analysis sound Marxist enough for you,
Bill?)

Now, I’ve never had any direct person-to-person contact with these owners. I don’t move in their social circles. But in our complex social system they interact with me indirectly by locating their restaurants at interchanges of busy highways as a predictably
easy feedback path for controlling perceptions that I and other travelers are predictably likely to want to control. And thus, even though I’m a free agent who can attempt to control any perception I want to control, I’m subject to their manipulation. (Don’t
get me started on McDonald’s ads.)

Enough for now. Best to all.

Kent

···

Bill Powers (2011.11.21.1540 MST)

Rick Marken (2012.11.1710)

[From Bill Powers (2011.11.21.1540 MST)]

Kent McClelland (2012.11.19.1650)]

KM: As Bill puts it, “While I would agree that some ways of organizing the environment favor the achievement of certain purposes while others are neutral or detrimental in that regard, I don’t think it
is possible to predict what any individual’s purposes will be in regard to available properties of the environment.”
A little further on he says, “What I’m saying would be nitpicking but for the fact that some sociologists want to claim that *ffordances influence us to carry out specific behaviors. For that claim to hold true, it would have to be true that
only those behaviors are favored.”

I guess that I’m one of the sociologists that he’s talking about here, because I do think that some *ffordances influence us to carry out specific behaviors, because they have been intentionally designed to do so by the people who have manufactured those particular
artifacts or who have set up the physical environment that way.

BP: My point is that the presence of an *ffordance does not, by itself, influence anyone to break off from what he is doing and begin behaving in a way that uses the *ffordance. The person must first want to achieve the end that the *ffordance provides when
used.

If I don’t think the lawn need mowing right now, the sight of the lawnmower will not urge me to start using it (though it might remind me to have a look at the lawn). To use Martin Taylor’s example, a plank does not actually *fford crossing a stream unless
one wants to cross the stream.

Otherwise it could be picked up and used for construction, to use Martin’s example a little further. It *affords that use, too. It affords any use to which it is actually put.

I think my present conclusion is that *ffordance is indeed an example of the behavioral illusion. It reverses the actual direction of causation, in the same way that Skinner did when he said that behavior is controlled by its consequences, when it is perfectly
clear that the consequences are caused by the behavior.

The *ffordance is an illusion; it reverses cause and effect. Considering that circular causality is somewhat hard to understand, I will generously admit that nobody can be blamed for this (unless they now go on believing in *ffordance).

KM: … sociologists would like to be able to explain the patterns and regularities that do emerge. I think PCT can offer some insights that can be useful to them, particularly if we start looking at
the (manufactured) physical environment as a complex of socially designed feedback paths.

I think the same thing. And I think our explanation is better than Gibson’s. It is the organism, not the object that appears to be doing the affording, that decides how the object will actually be used.

1.	Exclusivity—Inclusivity.
2.	Flexibility of Purpose
3.	Bandwidth (which I separated into PoPS -- Power, Precision, and Speed)
4.	Fixity—Evanescence
5.	Manipulability.
Hypothesis: An individual’s liberty and flexibility of action is increased when the individual has access to feedback paths that are exclusive, multi-purpose, high bandwidth, and fixed but highly manipulable.

[Bob_Hintz]

bob hintz 11.23.12

Very nice discussion Kent. How would you describe the process by which the owners discovered/decided which intersections were busy enough to warrant creating a McDonald’s at that location rather than some other in order to manipulate/facilitate your ability to control potty/food breaks on your trip?

It seems to me that learning what to do in order control a perception is different from doing it once you know how to control that perception. When I am attempting to learn how to control perception A, my perceptions of the consequences of my activity on perception A certainly help me decide what to continue doing. If my activity makes no difference in my error signal, I either try harder, give up or change my activity until I discover what I can do to achieve my reference. When I discover what I can do to reduce my error, I typically quit trying new things and just control the perception when I experience that error. Building the first McDonalds was surely quite a different activity than building the millionth one.

bob

···

On Fri, Nov 23, 2012 at 9:54 PM, McClelland, Kent MCCLEL@grinnell.edu wrote:

[From Kent McClelland (2012.11.23.2124 CST)]

Martin Taylor (2012.11.23.09.24)
Bill Powers (2011.11.21.1540 MST)

Rick Marken (2012.11.1710)

Hello again all,

Having roused myself from my illness-induced lethargy, and having read back over the posts in this thread, I’ve decided that, in fact, I agree with both Bill and Martin, and with regard their points of difference, which seem to be mainly terminological, I don’t
have any strong opinion.

My argument, as I had phrased it, was incorrect. I said that “some *ffordances influence us to carry out specific behaviors,” and that’s obviously wrong. Whether an object in the environment is used by an individual as a feedback path for controlling a particular
perception depends entirely, as Bill and Rick have been at pains to point out, on the perception that the individual wants to control. It’s not the object influencing the person, but the person using the object for his own purposes.

I guess you could describe my argument as an instance of the behavioral illusion, but that’s not exactly the problem that I see with it. A central feature of PCT is its reliance on closed-loop analysis. A circle of causality stretches from the person’s perceptual
input functions to reference values in the HPCT hierarchy to the person’s output functions and then back through a feedback function in the physical environment, combining with disturbances in that environment, to affect the person’s input functions. In that
negative-feedback loop, the active agent of control is the person, as he attempts to maintain a match between his reference values and his current perceptions of what is happening in the environment. To suggest that something in the environment “influences”
the person to behave in a certain way is to locate the agency in the environment, not the person.

However, it still seems to me that there’s a disconnect between the argument that I’ve been attempting to make in this thread (not very successfully, it appears) and Bill and Rick’s responses to it. What Bill and Rick have been saying is exactly right as long
as your frame of reference is the individual. But, as Martin has been pointing out, and very eloquently in the initial post of this thread, my frame of reference is social interaction. I’m interested as a sociologist in questions of social influence, manipulation,
and social power more generally. I’ve been talking about the environment, but what I’m ultimately interested in are relationships between people, how people influence and manipulate each other when they want to control their own perceptions of other people’s
behavior. My frame of reference is not the individual, but multiple individuals interacting with each other in a shared environment.

The only way that a person can influence or manipulate another person’s behavior is for the first person to control his or her perception of some variable in their shared physical environment. A classic rubber-band experiment describes the situation well. The
“experimenter” moves one end of the dual rubber bands, and the “subject” moves the other end while attempting to control the perception of keeping the jointure of the two bands over some fixed point. The experimenter can manipulate the physical actions of
the subject by moving her own end in some geometric pattern, and in order to keep his own perception in control the subject begins (perhaps unconsciously) to move his end in a geometric pattern that mirrors the experimenter’s. My point is that the manipulator
must control something about their shared environment, in this case the manipulator’s own physical movements, in order to control her perceptions of the other person’s behavior.

When Bill or Rick have talked about social interactions in the past, they have usually described it, if my memory is correct, in terms of one person providing disturbances to perceptions controlled by another. In the case of the kinds of interactions that I
am talking about, however, disturbances seems like not quite the right metaphor. The word ‘disturbance’ carries with it the connotation of fluctuation, something unpredictably changing, although of course a disturbance could also be simply a steady pressure
in one direction or another. The physical arrangements that I am talking about are more like steady pressures than fluctuating disturbances. They are portions of the environment that have been stabilized by one set of people, the manipulators, as they control
their own perceptions in hopes of influencing the actions of another set of people, just as the experimenter in the rubber band experiment creates a stable pattern of movement of her own hand motions in order to have the perception of the subject’s hand moving
in a similar pattern.

Let’s get back to my McDonald’s example. Although I’m not particularly fond of the food at those restaurants, I often find myself stopping there when I’m traveling on the highway. I have a predictable set of perceptions that I would like to control. After a
certain number of hours behind the wheel, I’m hungry and would like to perceive some food in my stomach. Even more, at my age, I need to visit a toilet. Lots of other travelers on that highway are likely to be dealing with similar sets of bodily disturbances.
I know that there is likely to be a McDonald’s at the next interchange, and I know that if I stop there I will be able to get some food quickly, relieve my bladder, and get back on the road in short order. At any given interchange, there is likely to be more
than one feedback path available for controlling these perceptions, and in almost every case I would be able to find better food and a more elegant bathroom were I willing to take the time to venture farther away from the highway interchange. But the McDonald’s
is likely to be located right at the interchange. It’s predictable, easy, and cheap. And the higher-level perception that I’m controlling is to reach my destination as quickly as possible. The McDonald’s for me is “the (feedback) path of least resistance.”
When my controlling a given perception is simply a means to an end, not my primary purpose in view, I often take the feedback path that is closest to hand and easiest to use. And most others do the same, I would argue.

We can analyze this situation sociologically as an interaction between me and the owners of the McDonald’s chain. I think it is reasonable to see them as having manipulated me, although I’ve just been controlling my own perceptions, as always happens in cases
of manipulation, because manipulation from a PCT point of view involves taking advantage of a person’s focus on maintaining control of one set of perceptions to change the person’s actions in other ways. By locating their restaurants at the interchanges of
busy highways, by having a standardized menu and standardized bathrooms, as well as a standardized architectural appearance and signage, they manipulate me and travelers like me to choose their restaurants as the feedback path for controlling predictably recurring
perceptions we are likely to have.

In this social interaction between owners and patrons, the McDonald’s owners are clearly the more powerful party, as is evidenced by their ability to modify large parts of our shared physical environment in accordance with their architectural plans, as
well as their ability to get a host of employees to carry out the routines of these restaurants according to their routinized specifications (or, in other words, the owners’ reference values). Patrons of McDonald’s are individually powerless, although if they
organized to boycott the chain they might have some countervailing force. It matters only infinitesimally to McDonald’s bottom line whether I choose their restaurant some other as a feedback path for controlling my perceptions on my trip. And from these transactions
between owners and patrons, the owners, arguably, have more to gain more than the patrons. The owners gain tidy profits, while the patrons only gain bigger waistlines from all the fattening food they sell. (Does this analysis sound Marxist enough for you,
Bill?)

Now, I’ve never had any direct person-to-person contact with these owners. I don’t move in their social circles. But in our complex social system they interact with me indirectly by locating their restaurants at interchanges of busy highways as a predictably
easy feedback path for controlling perceptions that I and other travelers are predictably likely to want to control. And thus, even though I’m a free agent who can attempt to control any perception I want to control, I’m subject to their manipulation. (Don’t
get me started on McDonald’s ads.)

Kent

Enough for now. Best to all.

On Nov 23, 2012, at 9:47 AM, Martin Taylor wrote:

[Martin Taylor 2012.11.23.09.24]

[From Bill Powers (2011.11.21.1540 MST)]

Kent McClelland (2012.11.19.1650)]

KM: As Bill puts it, “While I would agree that some ways of organizing the environment favor the achievement of certain purposes while others are neutral or detrimental in that regard, I don’t think it
is possible to predict what any individual’s purposes will be in regard to available properties of the environment.”
A little further on he says, “What I’m saying would be nitpicking but for the fact that some sociologists want to claim that *ffordances influence us to carry out specific behaviors. For that claim to hold true, it would have to be true that
only those behaviors are favored.”

I guess that I’m one of the sociologists that he’s talking about here, because I do think that some *ffordances influence us to carry out specific behaviors, because they have been intentionally designed to do so by the people who have manufactured those particular
artifacts or who have set up the physical environment that way.

BP: My point is that the presence of an *ffordance does not, by itself, influence anyone to break off from what he is doing and begin behaving in a way that uses the *ffordance. The person must first want to achieve the end that the *ffordance provides when
used.

Kent, I’m with Bill on this. Granted, it is often the case that one sees something that leads one to do something for which one had not been controlling. For example, the sight of some Lindt chocolate bars on sale at $2.99 (the ones I am talking about usually
sell for $7.99) might induce me to buy chocolate even though I had had no thoughts about chocolate when I left home. You might say that the effordance of cheap chocolate influenced me to buy, and in a sense I suppose it did. But that’s not the way I would
see it through PCT spectacles.

If I did not like chocolate, I would not buy it, even at $2.99. Suppose the offer had been of roasted mashed maggots for $2.99 when they usually sell for $7.99. Would you buy them? Some people might, but I wouldn’t. When I say I “like chocolate”, what do I
mean? I think I mean that I have a positive reference value for perceiving the taste and texture of chocolate in my mouth. I am not currently experiencing that perception, so there is an error. But the error is not big enough to allow that control system to
win the conflict with a control system with a reference value for having money in hand when the money control system would increase its error by the perceptual equivalent of $7.99. It is big enough to win the conflict when the money control system would increase
its error by $2.99. Likewise, If I cannot see anywhere selling chocolate, I have no way to reduce the error in my “liking chocolate” control system.

I think that most of these apparent influences of effordances can be seen as examples of conflict, and all of them are either that or the appearance of a new effordance where none were previously available to some control system that was experiencing error.
A prisoner in jail has no effordance that would allow him to experience a holiday on a Samoan beach. But usually the adage “Where there’s a will, there’s a way” holds true. Whether to use that way usually depends on how its use would conflict with other perceptions
being controlled. The conflict might be within the person or between that person and some other(s), as when I want to cross the brook over the plank and you want to use the plank for construction.

To repeat Bill’s statement by way of agreeing with it: “the presence of an *ffordance does not, by itself, influence anyone to break off from what he is doing and begin behaving in a way that uses the *ffordance.” However, the appearance of a new effordance
may change the balance of a conflict in a way that to an outside observer looks as though the effordance did influence the person to “break off from what he is doing and begin behaving in a way that uses the *ffordance.” Alternatively, it may provide a means
for some control system to reduce its error where none was previously available. As I said in the post that introduced this subject heading [Martin Taylor 2012.11.14.00.10, posted 11.19] , this appearance is an example of the behavioural illusion, albeit in
a form different from the way the behavioural illusion is usually presented.

If I don’t think the lawn need mowing right now, the sight of the lawnmower will not urge me to start using it (though it might remind me to have a look at the lawn). To use Martin Taylor’s example, a plank does not actually *fford crossing a stream unless
one wants to cross the stream.

Here’s where I part from Bill, in the use of the word, but not in the science.

As I would use the word, the plank does efford crossing a brook dryshod, whether one uses it that way or not. To me, if something can be used as part of an environmental feedback path for control of some particular perception, it provides that effordance. As
Bill has pointed out previously, most objects could, in principle, have an indefinite number of effordances. I don’t find that to be an objection to the concept, any more than I find the fact that an object has an indefinitiely large number of properties to
be an objection to describing one of those properties. It is only when a control system exists that has a perception that could be influenced by some action on the object and has the means to effect that action that the effordance becomes actual – whether
or not the control system does in fact use it. To get to the shops, my car and my bike both provide the necessary effordance, because I know how to ride a bike and to drive a car.

As for Bill’s use of “*ffordance”, I have come to appreciate a reason why “effordance” is a PCT-suitable word whereas “affordance” is not. Initially, I thought of “effordance” as a shortening of “environmental affordance → e-affordance”, and as rather ugly.
But consider the differences between these pairs of words: affect vs. effect, adit vs. exit, accept vs. except, etc. In each case, the “a-” word has an inward connotation of something done to, movement toward, inclusion, whereas the corresponding “e-” word
has an outward connotation, doing to something, going somewhere, actively avoiding. I am passively affected by something, but I actively effect some change. Likewise, an “effordance” is something one uses in the environment, whereas an “affordance” has a connotation
of being imposed by the environment.

Otherwise it could be picked up and used for construction, to use Martin’s example a little further. It *affords that use, too. It affords any use to which it is actually put.

Yes, indeed. And I agree with most of what follows in Bill’s posting until …

I think my present conclusion is that *ffordance is indeed an example of the behavioral illusion. It reverses the actual direction of causation, in the same way that Skinner did when he said that behavior is controlled by its consequences, when it is perfectly
clear that the consequences are caused by the behavior.

Here Bill comes to a point from which I started but use it to come to the opposite conclusion. I used a Marken quote to define the behavioural illusion: “The illusion is that an observed relationship between environmental inputs and behavioral outputs reflect
characteristics of the system itself when it actually reflects properties of the feedback connection between the system’s output and a controlled perceptual input.” An outside observer might see someone use an effordance and conclude that the effordance influenced
the person to emit a certain behaviour. The observer has succumbed to the behavioural illusion. The effordance itself is not an illusion of any kind. It is just a condition in a person’s environment through which a person can control a perception.

The *ffordance is an illusion; it reverses cause and effect. Considering that circular causality is somewhat hard to understand, I will generously admit that nobody can be blamed for this (unless they now go on believing in *ffordance).

It does no such thing. Your imagined theorist might reverse cause and effect, but the effordance does not. You can put up all the straw theorists you like so that you can easily knock them down, but that does not change the situation to be explained. The connotation
of “Affordance” may well seem to reverse cause and effect. “Environmental affordance”, now shortened to “effordance” does not, and never did, either in this series of threads or in earlier times when the usage was also disputed.

KM: … sociologists would like to be able to explain the patterns and regularities that do emerge. I think PCT can offer some insights that can be useful to them, particularly if we start looking at
the (manufactured) physical environment as a complex of socially designed feedback paths.

I think the same thing. And I think our explanation is better than Gibson’s. It is the organism, not the object that appears to be doing the affording, that decides how the object will actually be used.

Yes, exactly. But why bother to bring up Gibson yet again? Has anyone in any of these related threads tried to support his views?

Having agreed to all that, and if Kent also agrees with both of us, the next step should be to address what Kent was targeting in his five characteristics of feedback paths that increase an organism’s liberty [From Kent McClelland (2012.11.02.1250)]. To refresh
the memory, they were entitled:

1.	Exclusivity—Inclusivity.
2.	Flexibility of Purpose
3.	Bandwidth (which I separated into PoPS -- Power, Precision, and Speed)
4.	Fixity—Evanescence
5.	Manipulability.
Hypothesis: An individual’s liberty and flexibility of action is increased when the individual has access to feedback paths that are exclusive, multi-purpose, high bandwidth, and fixed but highly manipulable.

I would like to address these in the language of effordances rather than of actualized feedback paths.

Martin

[rsmarken]

[From Rick Marken (2012.11.24.1250)]

Kent McClelland (2012.11.23.2124 CST)--

KM: Having roused myself from my illness-induced lethargy, and having read back
over the posts in this thread, I've decided that, in fact, I agree with both
Bill and Martin, and with regard their points of difference, which seem to
be mainly terminological, I don't have any strong opinion.

RM: There seems to be a puzzling intransitivity (or asymmetry, or
something) here. I disagree with Martin, Martin disagrees with me (so
far so good). I disagree with Kent, Kent disagrees with me (still
ok). Bill agrees with me, I agree with Bill (we're on a roll). Martin
agrees with Kent, Kent agrees with Martin (all making sense so far).
Bill disagrees with Martin, Martin agrees with Bill (oops). Bill
disagrees with Kent, Kent agrees with Bill (hmm).

All I can say is "I want to be like Bill"! Since I can't, I think the
best idea for me is to just stay out of this;-)

I'm glad you're feeling better Kent!

Best

Rick

···

My argument, as I had phrased it, was incorrect. I said that "some
*ffordances influence us to carry out specific behaviors," and that's
obviously wrong. Whether an object in the environment is used by an
individual as a feedback path for controlling a particular perception
depends entirely, as Bill and Rick have been at pains to point out, on the
perception that the individual wants to control. It's not the object
influencing the person, but the person using the object for his own
purposes.

I guess you could describe my argument as an instance of the behavioral
illusion, but that's not exactly the problem that I see with it. A central
feature of PCT is its reliance on closed-loop analysis. A circle of
causality stretches from the person's perceptual input functions to
reference values in the HPCT hierarchy to the person's output functions and
then back through a feedback function in the physical environment, combining
with disturbances in that environment, to affect the person's input
functions. In that negative-feedback loop, the active agent of control is
the person, as he attempts to maintain a match between his reference values
and his current perceptions of what is happening in the environment. To
suggest that something in the environment "influences" the person to behave
in a certain way is to locate the agency in the environment, not the person.

However, it still seems to me that there's a disconnect between the argument
that I've been attempting to make in this thread (not very successfully, it
appears) and Bill and Rick's responses to it. What Bill and Rick have been
saying is exactly right as long as your frame of reference is the
individual. But, as Martin has been pointing out, and very eloquently in the
initial post of this thread, my frame of reference is social interaction.
I'm interested as a sociologist in questions of social influence,
manipulation, and social power more generally. I've been talking about the
environment, but what I'm ultimately interested in are relationships between
people, how people influence and manipulate each other when they want to
control their own perceptions of other people's behavior. My frame of
reference is not the individual, but multiple individuals interacting with
each other in a shared environment.

The only way that a person can influence or manipulate another person's
behavior is for the first person to control his or her perception of some
variable in their shared physical environment. A classic rubber-band
experiment describes the situation well. The "experimenter" moves one end of
the dual rubber bands, and the "subject" moves the other end while
attempting to control the perception of keeping the jointure of the two
bands over some fixed point. The experimenter can manipulate the physical
actions of the subject by moving her own end in some geometric pattern, and
in order to keep his own perception in control the subject begins (perhaps
unconsciously) to move his end in a geometric pattern that mirrors the
experimenter's. My point is that the manipulator must control something
about their shared environment, in this case the manipulator's own physical
movements, in order to control her perceptions of the other person's
behavior.

When Bill or Rick have talked about social interactions in the past, they
have usually described it, if my memory is correct, in terms of one person
providing disturbances to perceptions controlled by another. In the case of
the kinds of interactions that I am talking about, however, disturbances
seems like not quite the right metaphor. The word 'disturbance' carries with
it the connotation of fluctuation, something unpredictably changing,
although of course a disturbance could also be simply a steady pressure in
one direction or another. The physical arrangements that I am talking about
are more like steady pressures than fluctuating disturbances. They are
portions of the environment that have been stabilized by one set of people,
the manipulators, as they control their own perceptions in hopes of
influencing the actions of another set of people, just as the experimenter
in the rubber band experiment creates a stable pattern of movement of her
own hand motions in order to have the perception of the subject's hand
moving in a similar pattern.

Let's get back to my McDonald's example. Although I'm not particularly fond
of the food at those restaurants, I often find myself stopping there when
I'm traveling on the highway. I have a predictable set of perceptions that I
would like to control. After a certain number of hours behind the wheel, I'm
hungry and would like to perceive some food in my stomach. Even more, at my
age, I need to visit a toilet. Lots of other travelers on that highway are
likely to be dealing with similar sets of bodily disturbances. I know that
there is likely to be a McDonald's at the next interchange, and I know that
if I stop there I will be able to get some food quickly, relieve my bladder,
and get back on the road in short order. At any given interchange, there is
likely to be more than one feedback path available for controlling these
perceptions, and in almost every case I would be able to find better food
and a more elegant bathroom were I willing to take the time to venture
farther away from the highway interchange. But the McDonald's is likely to
be located right at the interchange. It's predictable, easy, and cheap. And
the higher-level perception that I'm controlling is to reach my destination
as quickly as possible. The McDonald's for me is "the (feedback) path of
least resistance." When my controlling a given perception is simply a means
to an end, not my primary purpose in view, I often take the feedback path
that is closest to hand and easiest to use. And most others do the same, I
would argue.

We can analyze this situation sociologically as an interaction between me
and the owners of the McDonald's chain. I think it is reasonable to see them
as having manipulated me, although I've just been controlling my own
perceptions, as always happens in cases of manipulation, because
manipulation from a PCT point of view involves taking advantage of a
person's focus on maintaining control of one set of perceptions to change
the person's actions in other ways. By locating their restaurants at the
interchanges of busy highways, by having a standardized menu and
standardized bathrooms, as well as a standardized architectural appearance
and signage, they manipulate me and travelers like me to choose their
restaurants as the feedback path for controlling predictably recurring
perceptions we are likely to have.

In this social interaction between owners and patrons, the McDonald's owners
are clearly the more powerful party, as is evidenced by their ability to
modify large parts of our shared physical environment in accordance with
their architectural plans, as well as their ability to get a host of
employees to carry out the routines of these restaurants according to their
routinized specifications (or, in other words, the owners' reference
values). Patrons of McDonald's are individually powerless, although if they
organized to boycott the chain they might have some countervailing force. It
matters only infinitesimally to McDonald's bottom line whether I choose
their restaurant some other as a feedback path for controlling my
perceptions on my trip. And from these transactions between owners and
patrons, the owners, arguably, have more to gain more than the patrons. The
owners gain tidy profits, while the patrons only gain bigger waistlines from
all the fattening food they sell. (Does this analysis sound Marxist enough
for you, Bill?)

Now, I've never had any direct person-to-person contact with these owners. I
don't move in their social circles. But in our complex social system they
interact with me indirectly by locating their restaurants at interchanges of
busy highways as a predictably easy feedback path for controlling
perceptions that I and other travelers are predictably likely to want to
control. And thus, even though I'm a free agent who can attempt to control
any perception I want to control, I'm subject to their manipulation. (Don't
get me started on McDonald's ads.)

Enough for now. Best to all.

Kent

On Nov 23, 2012, at 9:47 AM, Martin Taylor wrote:

[Martin Taylor 2012.11.23.09.24]

[From Bill Powers (2011.11.21.1540 MST)]

Kent McClelland (2012.11.19.1650)]

KM: As Bill puts it, "While I would agree that some ways of organizing the
environment favor the achievement of certain purposes while others are
neutral or detrimental in that regard, I don't think it is possible to
predict what any individual's purposes will be in regard to available
properties of the environment."

A little further on he says, "What I'm saying would be nitpicking but for
the fact that some sociologists want to claim that *ffordances influence us
to carry out specific behaviors. For that claim to hold true, it would have
to be true that only those behaviors are favored."

I guess that I'm one of the sociologists that he's talking about here,
because I do think that some *ffordances influence us to carry out specific
behaviors, because they have been intentionally designed to do so by the
people who have manufactured those particular artifacts or who have set up
the physical environment that way.

BP: My point is that the presence of an *ffordance does not, by itself,
influence anyone to break off from what he is doing and begin behaving in a
way that uses the *ffordance. The person must first want to achieve the end
that the *ffordance provides when used.

Kent, I'm with Bill on this. Granted, it is often the case that one sees
something that leads one to do something for which one had not been
controlling. For example, the sight of some Lindt chocolate bars on sale at
$2.99 (the ones I am talking about usually sell for $7.99) might induce me
to buy chocolate even though I had had no thoughts about chocolate when I
left home. You might say that the effordance of cheap chocolate influenced
me to buy, and in a sense I suppose it did. But that's not the way I would
see it through PCT spectacles.

If I did not like chocolate, I would not buy it, even at $2.99. Suppose the
offer had been of roasted mashed maggots for $2.99 when they usually sell
for $7.99. Would you buy them? Some people might, but I wouldn't. When I say
I "like chocolate", what do I mean? I think I mean that I have a positive
reference value for perceiving the taste and texture of chocolate in my
mouth. I am not currently experiencing that perception, so there is an
error. But the error is not big enough to allow that control system to win
the conflict with a control system with a reference value for having money
in hand when the money control system would increase its error by the
perceptual equivalent of $7.99. It is big enough to win the conflict when
the money control system would increase its error by $2.99. Likewise, If I
cannot see anywhere selling chocolate, I have no way to reduce the error in
my "liking chocolate" control system.

I think that most of these apparent influences of effordances can be seen as
examples of conflict, and all of them are either that or the appearance of a
new effordance where none were previously available to some control system
that was experiencing error. A prisoner in jail has no effordance that would
allow him to experience a holiday on a Samoan beach. But usually the adage
"Where there's a will, there's a way" holds true. Whether to use that way
usually depends on how its use would conflict with other perceptions being
controlled. The conflict might be within the person or between that person
and some other(s), as when I want to cross the brook over the plank and you
want to use the plank for construction.

To repeat Bill's statement by way of agreeing with it: "the presence of an
*ffordance does not, by itself, influence anyone to break off from what he
is doing and begin behaving in a way that uses the *ffordance." However, the
appearance of a new effordance may change the balance of a conflict in a way
that to an outside observer looks as though the effordance did influence the
person to "break off from what he is doing and begin behaving in a way that
uses the *ffordance." Alternatively, it may provide a means for some control
system to reduce its error where none was previously available. As I said in
the post that introduced this subject heading [Martin Taylor
2012.11.14.00.10, posted 11.19] , this appearance is an example of the
behavioural illusion, albeit in a form different from the way the
behavioural illusion is usually presented.

If I don't think the lawn need mowing right now, the sight of the lawnmower
will not urge me to start using it (though it might remind me to have a look
at the lawn). To use Martin Taylor's example, a plank does not actually
*fford crossing a stream unless one wants to cross the stream.

Here's where I part from Bill, in the use of the word, but not in the
science.

As I would use the word, the plank does efford crossing a brook dryshod,
whether one uses it that way or not. To me, if something can be used as part
of an environmental feedback path for control of some particular perception,
it provides that effordance. As Bill has pointed out previously, most
objects could, in principle, have an indefinite number of effordances. I
don't find that to be an objection to the concept, any more than I find the
fact that an object has an indefinitiely large number of properties to be an
objection to describing one of those properties. It is only when a control
system exists that has a perception that could be influenced by some action
on the object and has the means to effect that action that the effordance
becomes actual -- whether or not the control system does in fact use it. To
get to the shops, my car and my bike both provide the necessary effordance,
because I know how to ride a bike and to drive a car.

As for Bill's use of "*ffordance", I have come to appreciate a reason why
"effordance" is a PCT-suitable word whereas "affordance" is not. Initially,
I thought of "effordance" as a shortening of "environmental affordance -->
e-affordance", and as rather ugly. But consider the differences between
these pairs of words: affect vs. effect, adit vs. exit, accept vs. except,
etc. In each case, the "a-" word has an inward connotation of something done
to, movement toward, inclusion, whereas the corresponding "e-" word has an
outward connotation, doing to something, going somewhere, actively avoiding.
I am passively affected by something, but I actively effect some change.
Likewise, an "effordance" is something one uses in the environment, whereas
an "affordance" has a connotation of being imposed by the environment.

Otherwise it could be picked up and used for construction, to use Martin's
example a little further. It *affords that use, too. It affords any use to
which it is actually put.

Yes, indeed. And I agree with most of what follows in Bill's posting until
...

I think my present conclusion is that *ffordance is indeed an example of the
behavioral illusion. It reverses the actual direction of causation, in the
same way that Skinner did when he said that behavior is controlled by its
consequences, when it is perfectly clear that the consequences are caused by
the behavior.

Here Bill comes to a point from which I started but use it to come to the
opposite conclusion. I used a Marken quote to define the behavioural
illusion: "The illusion is that an observed relationship between
environmental inputs and behavioral outputs reflect characteristics of the
system itself when it actually reflects properties of the feedback
connection between the system�s output and a controlled perceptual input."
An outside observer might see someone use an effordance and conclude that
the effordance influenced the person to emit a certain behaviour. The
observer has succumbed to the behavioural illusion. The effordance itself is
not an illusion of any kind. It is just a condition in a person's
environment through which a person can control a perception.

The *ffordance is an illusion; it reverses cause and effect. Considering
that circular causality is somewhat hard to understand, I will generously
admit that nobody can be blamed for this (unless they now go on believing in
*ffordance).

It does no such thing. Your imagined theorist might reverse cause and
effect, but the effordance does not. You can put up all the straw theorists
you like so that you can easily knock them down, but that does not change
the situation to be explained. The connotation of "Affordance" may well seem
to reverse cause and effect. "Environmental affordance", now shortened to
"effordance" does not, and never did, either in this series of threads or in
earlier times when the usage was also disputed.

KM: ... sociologists would like to be able to explain the patterns and
regularities that do emerge. I think PCT can offer some insights that can be
useful to them, particularly if we start looking at the (manufactured)
physical environment as a complex of socially designed feedback paths.

I think the same thing. And I think our explanation is better than Gibson's.
It is the organism, not the object that appears to be doing the affording,
that decides how the object will actually be used.

Yes, exactly. But why bother to bring up Gibson yet again? Has anyone in any
of these related threads tried to support his views?

Having agreed to all that, and if Kent also agrees with both of us, the next
step should be to address what Kent was targeting in his five
characteristics of feedback paths that increase an organism's liberty [From
Kent McClelland (2012.11.02.1250)]. To refresh the memory, they were
entitled:

1. Exclusivity�Inclusivity.
2. Flexibility of Purpose
3. Bandwidth (which I separated into PoPS -- Power, Precision, and Speed)
4. Fixity�Evanescence
5. Manipulability.
Hypothesis: An individual�s liberty and flexibility of action is increased
when the individual has access to feedback paths that are exclusive,
multi-purpose, high bandwidth, and fixed but highly manipulable.

I would like to address these in the language of effordances rather than of
actualized feedback paths.

Martin

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

[Kent]

[From Kent McClelland (2012.11.24.1635)]

bob hintz 11.23.12

Very nice discussion Kent. How would you describe the process by which the owners discovered/decided which intersections were busy enough to warrant creating a McDonald's at that location rather than some other in order to manipulate/facilitate your ability to control potty/food breaks on your trip?

It seems to me that learning what to do in order control a perception is different from doing it once you know how to control that perception. When I am attempting to learn how to control perception A, my perceptions of the consequences of my activity on perception A certainly help me decide what to continue doing. If my activity makes no difference in my error signal, I either try harder, give up or change my activity until I discover what I can do to achieve my reference. When I discover what I can do to reduce my error, I typically quit trying new things and just control the perception when I experience that error. Building the first McDonalds was surely quite a different activity than building the millionth one.

bob

The first location choices were probably trial-and-error. I remember a McDonald's from my childhood that was stuck in a nondescript location, miles from the only super highway in the area, along a moderately important local road in the north hills suburbs of Pittsburgh.

After the chain had expanded to hundreds of outlets, they had access to lots more information--sales data from the franchises--to use in forming their perception of where the most profitable locations might be.

I would expect that today the company has a team of statisticians grinding out "location desirability" formulas based on daily traffic along interstate highways, population density and income statistics of surrounding localities, and who knows what else. They will surely have rationalized the system as much as they can, since that's their basic business approach.

Once we've learned (by imitation or reorganization or whatever) to control a perception, we typically control it by habit. For an organization to control something by habit means to build in a set of procedures for handling recurring events.

Kent

[Kent]

[From Kent McClelland (2012.11.24.1650)]

[From Rick Marken (2012.11.24.1250)]

Kent McClelland (2012.11.23.2124 CST)--

KM: Having roused myself from my illness-induced lethargy, and having read back
over the posts in this thread, I've decided that, in fact, I agree with both
Bill and Martin, and with regard their points of difference, which seem to
be mainly terminological, I don't have any strong opinion.

RM: There seems to be a puzzling intransitivity (or asymmetry, or
something) here. I disagree with Martin, Martin disagrees with me (so
far so good). I disagree with Kent, Kent disagrees with me (still
ok). Bill agrees with me, I agree with Bill (we're on a roll). Martin
agrees with Kent, Kent agrees with Martin (all making sense so far).
Bill disagrees with Martin, Martin agrees with Bill (oops). Bill
disagrees with Kent, Kent agrees with Bill (hmm).

All I can say is "I want to be like Bill"! Since I can't, I think the
best idea for me is to just stay out of this;-)

KM: It seems to show that it's possible for someone to listen to the logic of someone else's argument and then change his own mind.

You ought to try it sometime . . .

(Sorry, Rick. That was just too easy a setup.)

RM: I'm glad you're feeling better Kent!

KM: Thanks!

My best,

Kent

[rsmarken]

[From Rick Marken (2012.11.24.1800)]

Kent McClelland (2012.11.24.1650)--

RM: There seems to be a puzzling intransitivity (or asymmetry, or
something) here. I disagree with Martin, Martin disagrees with me (so
far so good). I disagree with Kent, Kent disagrees with me (still
ok). Bill agrees with me, I agree with Bill (we're on a roll). Martin
agrees with Kent, Kent agrees with Martin (all making sense so far).
Bill disagrees with Martin, Martin agrees with Bill (oops). Bill
disagrees with Kent, Kent agrees with Bill (hmm).

KM: It seems to show that it's possible for someone to listen to the logic of someone else's argument and then change his own mind.

RM: I don't see what changing minds has to do with it. It's just a
logic problem, as far as I can tell. Martin says X is true. Bill and I
say X is false. You say you agree with Bill and Martin, which implies
that you agree that X is true and that X is false. But you say you
don't agree with me, which implies you believe X is not false. So
there seems to be some contradictions here, which would exist even if
I changed my mind and decided that X was actually true. Which is fine;
no law says you have to be logical.

But it's interesting to think what you might be controlling for that
is "efforded" by agreeing with people espousing two completely
incompatible points of view. Same goes for Martin. Why does he agree
with Bill and not me when both Bill and I have been disagreeing with
him? My guess is that it has more to do with who is being disagreed
with than with what is being disagreed about.

KM: You ought to try it sometime . . .

RM: I've actually done it quite a bit (changed my mind) but it's
almost always been in the face of modeling and data. The last time I
changed my mind was about a year ago when Bill told me that a control
model is equivalent to an S-R model with a linear filter when there is
no disturbance to the controlled variable. I thought I had developed a
demonstration where I showed that a control model did better
(accounted for more variance in output) than an S-R model of the same
situation. Alas, I set up the linear filter, as suggested by Bill,
and, low and behold, the two models did equally well. Very
disappointing but facts is facts. I do change my mind if confronted
with facts that demand it.

KM: (Sorry, Rick. That was just too easy a setup.)

RM: Oh, that's OK. I didn't even notice whatever it was you were set up to do.

Best

Rick

···

RM: I'm glad you're feeling better Kent!

KM: Thanks!

My best,

Kent

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

[MartinT]

[Martin Taylor 2012.11.25.00.23]

[From Rick Marken (2012.11.24.1800)]

  Kent McClelland (2012.11.24.1650)--

RM: There seems to be a puzzling intransitivity (or asymmetry, or
something) here. I disagree with Martin, Martin disagrees with me (so
far so good). I disagree with Kent, Kent disagrees with me (still
ok). Bill agrees with me, I agree with Bill (we're on a roll). Martin
agrees with Kent, Kent agrees with Martin (all making sense so far).
Bill disagrees with Martin, Martin agrees with Bill (oops). Bill
disagrees with Kent, Kent agrees with Bill (hmm).

KM: It seems to show that it's possible for someone to listen to the logic of someone else's argument and then change his own mind.

RM: I don't see what changing minds has to do with it. It's just a
logic problem, as far as I can tell. Martin says X is true. Bill and I
say X is false.

Could you point to the place where this happened? So far as I can see, was untrue was a position that contradicted the position I took. You were apparently agreeing with me.

Quote [From Rick Marken (2012.11.1710)]:

What is
being insisted, I think, is that what an environmental object "makes
available" (affords) depends on the purposes of the person using that
object. What is being insisted, in other words, is that you have to
take_controlled variables_ into account before you can have any idea
what any particular "object" in the environment "makes available."

That is, and has been, my position.

I suspect that what you thought you were contradicting was my earlier argument that although the above is true for any particular effordance, you don't have to consider individual ones or specific control systems to understand the ways in which characteristics of effordances matter in respect of a person's overall liberty. As I had said earlier by way of analogy, you can examine all the individual cases you want in which the sum of squares of two numbers is a particular constant, but you need a wider point of view and a different technique to determine that for ALL cases in which x^2+y^2 = k, the points lie on a particular circle. Your only response to that was that you were never any good at math, which I suppose may be considered in some circles to be a counterargument. But the point remains valid.

When you are seeking generalizations, you need more than suggestive particular cases. For example, it is generally true that a feedback loop is unstable if the loop gain is positive. You can try hundreds of special cases and intuit that since has proved true for all your test cases it is likely to be true always, but working out the maths that describe all control systems is a better way to prove the point. Likewise, for a given loop gain greater than 1.0 (negative), there exists a loop transport lag beyond which the loop will be unstable. Again, you might guess this to be true after trying a few hundred loops with different parameters, controlling different perceptions. But although each individual control system must control some perception, the general statement will hold true no matter what the perception is.

In discussing *ffordances, I have tried always to make the point that each example of the concept requires that there be a specific perception to be controlled as well as specific actions that involve the object in a way that results in those actions influencing that perception. So you were not disagreeing with me on this point. But the parameters that affect Kent's hypothesis "An individual�s liberty and flexibility of action is increased when the individual has access to feedback paths that are exclusive, multi-purpose, high bandwidth, and fixed but highly manipulable" do not depend on any particular perceptions being controlled.

I do not remember Bill having claimed to be false anything that I actually asserted. He has, however, persistently created straw men by using the word *ffordance in a way that is distinctly different from the meaning I associate with effordance (and earlier with "environmental affordance", which is the same thing). He uses the term to mean something that is imposed upon someone, which I do not accept. So I do disagree with Bill on the usage of the word "effordance", though not necessarily on the usage of "affordance" when used alone without the preceding "environmental". I do not think I disagree with Bill on the science. I did disagree with Kent on the science, but Kent changed his mind so that he now accepts the position taken by Bill and me (and perhaps you, if only we knew what your position is).

Does this help your pseudo-logical conundrum?

Martin

···

at least in your last two messages responding to me, all that you said

[rsmarken]

[From Rick Marken (2012.11.25.0940)]

Martin Taylor (2012.11.25.00.23)--

RM: I don't see what changing minds has to do with it. It's just a
logic problem, as far as I can tell. Martin says X is true. Bill and I
say X is false.

MT: Could you point to the place where this happened?

Let X = *ffordance. My overall impression has been that Bill and I
have been saying things that show that X is either wrong, misleading
or useless, ie. false.

MT: So far as I can see...You were apparently agreeing with me.

Quote [From Rick Marken (2012.11.1710)]:

What is
being insisted, I think, is that what an environmental object "makes
available" (affords) depends on the purposes of the person using that
object. What is being insisted, in other words, is that you have to
take_controlled variables_ into account before you can have any idea
what any particular "object" in the environment "makes available."

MT: That is, and has been, my position.

RM: Then you agree with Bill and I that the concept of *ffordance is
unnecessary because it refers to things that are already included in
PCT as disturbances or feedback functions.

MT: In discussing *ffordances, I have tried always to make the point that each
example of the concept requires that there be a specific perception to be
controlled as well as specific actions that involve the object in a way that
results in those actions influencing that perception. So you were not
disagreeing with me on this point.

RM: If we agree with each other then you would see that there is no
need to introduce the concept of *ffordance. I think both Bill and I
agree that the concept of *ffordance is (as I said above) either
wrong, misleading or useless. But maybe I'm wrong. Maybe Bill agrees
with you that *ffordance is a useful new concept. If so, then I am
indeed the odd man out here.

MT: But the parameters that affect Kent's
hypothesis "An individual�s liberty and flexibility of action is increased
when the individual has access to feedback paths that are exclusive,
multi-purpose, high bandwidth, and fixed but highly manipulable" do not
depend on any particular perceptions being controlled.

RM: I would say that this is not true at all. As I said in an earlier
post, feedback paths don't even exist until there is a controlled
variable. There is no way to tell anything about a feedback path
until you know what variable(s) is(are) under control. I believe Bill
has been saying the same thing, but, again, I may be wrong. If so, I
am again odd man out. But I will certainly change my mind if you (or
Bill or Kent or anyone else) can explain how I could know things about
a feedback path (*ffordance?), such as whether it is exclusive,
multi-purpose, high bandwidth, or fixed but highly manipulable,
without knowing what variable is under control.

Best

Rick

···

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

[Kent]

[From Kent McClelland (2012.11.25.1315)]

Rick Marken (2012.11.25.0940)

MT: But the parameters that affect Kent's
hypothesis "An individual�s liberty and flexibility of action is increased
when the individual has access to feedback paths that are exclusive,
multi-purpose, high bandwidth, and fixed but highly manipulable" do not
depend on any particular perceptions being controlled.

RM: I would say that this is not true at all. As I said in an earlier
post, feedback paths don't even exist until there is a controlled
variable. There is no way to tell anything about a feedback path
until you know what variable(s) is(are) under control. I believe Bill
has been saying the same thing, but, again, I may be wrong. If so, I
am again odd man out. But I will certainly change my mind if you (or
Bill or Kent or anyone else) can explain how I could know things about
a feedback path (*ffordance?), such as whether it is exclusive,
multi-purpose, high bandwidth, or fixed but highly manipulable,
without knowing what variable is under control.

KM: It seems to me, Rick, that one can't even determine whether a variable is under control without knowing something about the feedback path that is being used to control it. Observing a possible feedback path, I would say, comes first as we try to figure these things out.

Think about the classic coin-game example of The Test. We can determine that a variable is under control by doing the test, but how can we begin to tell that a variable might be under control?

The first step is for us to observe some stability or regularity in our shared physical environment, where instead we might expect to see some fluctuation, and to observe actions on the part of the putative controller that appear to relate to that stability. In the coin game, we see coins sitting on a table in some physical arrangement, whereas people don't usually just leave coins sitting around for no reason, and we see that the person sitting across from us has arranged the coins in that way.

We make the hypothesis that the person is trying to control some perceptual variable, and we make a guess at what the controlled variable might be. At the same time, merely to carry out the test, we must make the additional hypothesis that the physical arrangement of coins on the table is (part of) the feedback path that the person is using to control that perceptual variable.

We proceed to disturb the arrangement of coins and observe the person's actions in rearranging them, or not. When we are satisfied that we can predict whether the person will respond to a given disturbance or not, we claim that we have found the controlled variable (though our claim is always tentative, because, as often happens in the coin game, the supposedly controlled perception that we have imagined in terms of our own perceptual hierarchy may not precisely match the perception that the person is controlling in terms of their hierarchy).

We never have certain knowledge of what the other person's controlled variable is, because we don't have direct access to the other person's brain. We just give it our best guess based on our perceptions of the person's actions and mutually observable patterns of stability or regularity in our shared physical environment, patterns that may serve as feedback paths if our hypothesis is correct. Our knowledge of what has occurred in the shared physical environment is more nearly certain than is our guess as to what variable might be under control.

If we see millions of Americans spending the Thanksgiving weekend sitting in front of their television sets with the stations set on broadcasts of football games, and millions of others queueing up to get into retail stores in the wee hours of the morning of "Black Friday", can we fairly conclude from these physical arrangements of bodies and patterns of human actions that some perceptual variables might be under control? If we see lots of these people doing the same thing at the same time, is it fair to assume that the perceptual variables they seek to control might be similar? Would you wish to assert that there is nothing at all about controlled perceptions to be concluded from these massive and widespread patterns of physical and (possibly) behavioral regularity?

How might you suggest going about testing the hypothesis that the millions of Americans in front of their TV sets are controlling for, say, the perceptual variable of "watching a football game"? Would you propose going into people's homes, grabbing the remote control, changing channels, and seeing if the guy with the beer sitting in the easy chair gets up and slugs you (and, of course, then switches the channel back to the game)? How many homes would you need to invade? If that plan seems unreasonable, how would you advise the sociologist interested in this phenomenon to proceed?

Is it always necessary to take a know-nothing attitude about whether any perceptual variable is being controlled when the possible feedback paths for controlling a variable are in plain view and people appear to be taking advantage of those feedback paths? (Now, I'll grant you that there may be an open question about how the instrumental actions we observe might fit into the control of higher-order variables in a person's perceptual hierarchy�whether they're rooting for the Redskins or the Cowboys or just want to look macho, for example). From my point of view, easily observable physical stabilities that appear to result from human actions and can serve as possible feedback paths are prima facie evidence that some kind control must be taking place, and they seem a reasonable place to start our investigations, if we wish to find out more about what is being controlled.

Back to some earlier posts in this sequence:
Rick Marken (2012.11.24.1800)
Martin Taylor (2012.11.25.00.23)

KM: I haven't been taking Rick's little puzzle in logic very seriously, because it didn't make much sense to me in the context of the earlier posts in the thread that he drew it from.

If each of our posts had consisted only of a single unambiguous proposition, the puzzle might have seemed relevant. But the posts contained complex, nuanced arguments (at least that's what I was aiming at), so it wouldn't be surprising if each of us agreed with some of the things the others were saying, disagreed with other things, and agreed with yet other things, but only up to a point.

from a PCT perspective, as I understand the theory, people are self-guided control systems with perceptual hierarchies constructed out of our own unique experiences. Thus, in effect, each of us lives in his own perceptual world. Some disagreement is to be expected in that kind of social situation, and lock-step agreement would be the only surprising outcome (an outcome that would require some careful sociological analysis, I might add). From the PCT perspective, to argue over who agrees with whom seems like a waste of our time, not to mention the time of the other 150 or so people who monitor messages in this forum.

To me, PCT does not represent a finished, settled body of knowledge, but rather an extremely useful perspective for understanding and relating to the world around us, especially our own and other people's actions and relationships. In my understanding, PCT is an open-ended theory, still in development. Certainly, Bill's writings have always given me that impression. I regard my participation in this forum as an opportunity to contribute to the collective endeavor of articulating the theory more fully and extending its application in new and creative ways.

In that spirit, I've been contributing partly-formed ideas that are at the very edges of my understanding of PCT, and I don't expect to get everything right or that everybody will agree with me. Sometimes I even disagree with myself when I look later at what I've written. We learn from correcting errors (and PCT would suggest that life is one long process of error correction!).

Sorry for the preachy tone of this post. I really value your critical comments when they focus on substantive ideas.

My best,

Kent

···

On Nov 25, 2012, at 12:01 AM, Martin Taylor wrote:

[Martin Taylor 2012.11.25.00.23]

[From Rick Marken (2012.11.24.1800)]

Kent McClelland (2012.11.24.1650)--

RM: There seems to be a puzzling intransitivity (or asymmetry, or
something) here. I disagree with Martin, Martin disagrees with me (so
far so good). I disagree with Kent, Kent disagrees with me (still
ok). Bill agrees with me, I agree with Bill (we're on a roll). Martin
agrees with Kent, Kent agrees with Martin (all making sense so far).
Bill disagrees with Martin, Martin agrees with Bill (oops). Bill
disagrees with Kent, Kent agrees with Bill (hmm).

KM: It seems to show that it's possible for someone to listen to the logic of someone else's argument and then change his own mind.

RM: I don't see what changing minds has to do with it. It's just a
logic problem, as far as I can tell. Martin says X is true. Bill and I
say X is false.

Could you point to the place where this happened? So far as I can see, at least in your last two messages responding to me, all that you said was untrue was a position that contradicted the position I took. You were apparently agreeing with me.

Quote [From Rick Marken (2012.11.1710)]:

What is
being insisted, I think, is that what an environmental object "makes
available" (affords) depends on the purposes of the person using that
object. What is being insisted, in other words, is that you have to
take_controlled variables_ into account before you can have any idea
what any particular "object" in the environment "makes available."

That is, and has been, my position.

I suspect that what you thought you were contradicting was my earlier argument that although the above is true for any particular effordance, you don't have to consider individual ones or specific control systems to understand the ways in which characteristics of effordances matter in respect of a person's overall liberty. As I had said earlier by way of analogy, you can examine all the individual cases you want in which the sum of squares of two numbers is a particular constant, but you need a wider point of view and a different technique to determine that for ALL cases in which x^2+y^2 = k, the points lie on a particular circle. Your only response to that was that you were never any good at math, which I suppose may be considered in some circles to be a counterargument. But the point remains valid.

When you are seeking generalizations, you need more than suggestive particular cases. For example, it is generally true that a feedback loop is unstable if the loop gain is positive. You can try hundreds of special cases and intuit that since has proved true for all your test cases it is likely to be true always, but working out the maths that describe all control systems is a better way to prove the point. Likewise, for a given loop gain greater than 1.0 (negative), there exists a loop transport lag beyond which the loop will be unstable. Again, you might guess this to be true after trying a few hundred loops with different parameters, controlling different perceptions. But although each individual control system must control some perception, the general statement will hold true no matter what the perception is.

In discussing *ffordances, I have tried always to make the point that each example of the concept requires that there be a specific perception to be controlled as well as specific actions that involve the object in a way that results in those actions influencing that perception. So you were not disagreeing with me on this point. But the parameters that affect Kent's hypothesis "An individual�s liberty and flexibility of action is increased when the individual has access to feedback paths that are exclusive, multi-purpose, high bandwidth, and fixed but highly manipulable" do not depend on any particular perceptions being controlled.

I do not remember Bill having claimed to be false anything that I actually asserted. He has, however, persistently created straw men by using the word *ffordance in a way that is distinctly different from the meaning I associate with effordance (and earlier with "environmental affordance", which is the same thing). He uses the term to mean something that is imposed upon someone, which I do not accept. So I do disagree with Bill on the usage of the word "effordance", though not necessarily on the usage of "affordance" when used alone without the preceding "environmental". I do not think I disagree with Bill on the science. I did disagree with Kent on the science, but Kent changed his mind so that he now accepts the position taken by Bill and me (and perhaps you, if only we knew what your position is).

Does this help your pseudo-logical conundrum?

Martin

[Bill_Powers1]

[From bill Powers (2012.11.25.0955 MST]

Kent McClelland (2012.11.23.2124 CST) –

KM: … To suggest that
something in the environment “influences” the person to behave
in a certain way is to locate the agency in the environment, not the
person.

However, it still seems to me that there’s a disconnect between the
argument that I’ve been attempting to make in this thread (not very
successfully, it appears) and Bill and Rick’s responses to it. What Bill
and Rick have been saying is exactly right as long as your frame of
reference is the individual. But, as Martin has been pointing out, and
very eloquently in the initial post of this thread, my frame of reference
is social interaction. I’m interested as a sociologist in questions of
social influence, manipulation, and social power more generally. I’ve
been talking about the environment, but what I’m ultimately interested in
are relationships between people, how people influence and manipulate
each other when they want to control their own perceptions of other
people’s behavior. My frame of reference is not the individual, but
multiple individuals interacting with each other in a shared
environment.

BP: The disconnect is what I am trying to eliminate. What we say about
individual behavior should never be contradicted by what we say about
many individuals interacting with each other, though the latter may
include measures that have no meaning for individuals (such as averages
or group tendencies).

KM: … My point is that the
manipulator must control something about their shared environment, in
this case the manipulator’s own physical movements, in order to control
her perceptions of the other person’s behavior.

BP: Physical movements and the environmental effects of such movements –
the person affected may never see the movements themselves, as in your
discussion of your relationship to the managers of McDonalds.

KM: When Bill or Rick have
talked about social interactions in the past, they have usually described
it, if my memory is correct, in terms of one person providing
disturbances to perceptions controlled by another. In the case of the
kinds of interactions that I am talking about, however, disturbances
seems like not quite the right metaphor. The word ‘disturbance’ carries
with it the connotation of fluctuation, something unpredictably changing,
although of course a disturbance could also be simply a steady pressure
in one direction or another.

BP: Yes. If you look at the canonical diagram, you’ll see that the effect
of a disturbance simply adds to the effect of the control system’s own
action on the controlled variable, the input quantity or the perception
derived from input quantities. This effect could actually aid the efforts
of the control system: it will still be opposed by the control system,
because the reduction in error will result in the control system’s
producing less output, so most of the reduction is canceled out. And as
you point out, there is nothing that restricts the behavior of the
disturbance through time: it could be something transient like a gust of
wind or something continuously present like gravity.

KM: The physical arrangements
that I am talking about are more like steady pressures than fluctuating
disturbances. They are portions of the environment that have been
stabilized by one set of people, the manipulators, as they control their
own perceptions in hopes of influencing the actions of another set of
people, just as the experimenter in the rubber band experiment creates a
stable pattern of movement of her own hand motions in order to have the
perception of the subject’s hand moving in a similar
pattern.

BP: Martin Taylor would say that you’re talking, in part, about
disturbances that alter the form of the environmental feedback function
(the “effordance”), like moving the position of a fulcrum
relative to a lever and thus changing the mechanical advantage. This is a
more complex kind of disturbance because it changes the loop gain and can
change the linearity or any other property of the feedback function. It
may not cause any actual change in the controlled variable, but the next
action by the control system will have a different effect on the
controlled variable. Now the behaving system may have to reorganize to
recover its original ability to control, or it may find that it can now
control much better with less effort.

KM: When my controlling a given
perception is simply a means to an end, not my primary purpose in view, I
often take the feedback path that is closest to hand and easiest to use.
And most others do the same, I would argue.

BP: Yes, but again you have to take the user’s controlled variables into
account. The farmer who sold the land on which the McDonald’s was built
may benefit from the cash but other farmers could lose because of the
competition with the farmer’s own food products.

So you can’t just speak about the effects of McDonald’s on farmers. The
basic problem with a purely sociological analysis is that measures over a
whole population add up effects on many people with different desires and
intentions, so the final measure doesn’t really apply to anybody. Pluses
for some people cancel minuses for others, so all you’re measuring is the
excess of one over the other – and then assuming that this excess is
typical of all members of that population.

The only solution to this problem that I can see is to apply the model to
individuals first, and measure their parameters, and only then look for
averages and tendencies and other statistical quantitites with regard to
those parameters. For example, an important datum is the way the
parameters are distributed over the population. Rather than reducing the
collection of measurements to a single mean value, we look at all the
data and try to find a pattern in the way the distribution explains
global properties of the social system, such as amount of violence or
communication or cooperation. The biggest mistake I see is to measure
properties of individuals and then used their means as if they were
properties of the society. As per HTML, social variables are of different
types from variables that refer to individuals, and the boundary should
never be crossed. The old apples-oranges problem.

KM: We can analyze this
situation sociologically as an interaction between me and the owners of
the McDonald’s chain. I think it is reasonable to see them as having
manipulated me, although I’ve just been controlling my own perceptions,
as always happens in cases of manipulation, because manipulation from a
PCT point of view involves taking advantage of a person’s focus on
maintaining control of one set of perceptions to change the person’s
actions in other ways.

BP: Being a manipulator may be true of you and not true of the person
next to you. Is the truth then the average of manipulating and not
manipulating?

KM: By locating their
restaurants at the interchanges of busy highways, by having a
standardized menu and standardized bathrooms, as well as a standardized
architectural appearance and signage, they manipulate me and travelers
like me to choose their restaurants as the feedback path for controlling
predictably recurring perceptions we are likely to
have.

BP: What do you mean by “travelers like me?” Do you mean a
subset of all the people who see or go to a McDonald’s? In that case only
half (or fewer) of the travelers may be enough like you to fall into the
same behavioral category. Or are you saying that if you define people
like you as having certain characteristics, and find that half of them
respond as predicted, everyone having those characteristics has a
“tendency” to show a given behavior, with 50% actually showing
it?

If social measures were giving us correlations of 0.95 and up, there
might be some excuse for thinking of individual and group behavior in the
same terms. But as you know the actual correlations obtainable in
sociological research are drastically lower than that in most cases – or
don’t you agree?

KM: In this social interaction
between owners and patrons, the McDonald’s owners are clearly the more
powerful party, as is evidenced by their ability to modify large parts of
our shared physical environment in accordance with their architectural
plans, as well as their ability to get a host of employees to carry out
the routines of these restaurants according to their routinized
specifications (or, in other words, the owners’ reference values).
Patrons of McDonald’s are individually powerless, although if they
organized to boycott the chain they might have some countervailing force.
It matters only infinitesimally to McDonald’s bottom line whether I
choose their restaurant some other as a feedback path for controlling my
perceptions on my trip. And from these transactions between owners and
patrons, the owners, arguably, have more to gain more than the patrons.
The owners gain tidy profits, while the patrons only gain bigger
waistlines from all the fattening food they sell. (Does this analysis
sound Marxist enough for you, Bill?)

BP: Not quite. The problem here is treating interactions among
individuals as if they were properties of the groups to which the
individuals belong. Patrons of McDonald’s are powerless to influence the
owners, but they can easily escape the influence of the owners by not
going to McDonald’s.

If you’re going to try to represent the relationships between owners and
patrons, you have to avoid the behavioral illusion. How the owners place
their stores does not affect the patrons’ behavior, and what the patrons
do does not control the owners’ behavior. Instead, what each one does
disturbs the variables that the other is trying to control, and the
behavior of each group acts against that disturbance (whether the
disturbance is helpful or harmful). Come to think of it, this isn’t very
Marxist, either, is it? Marx saw only social groups as
interacting.

KM: Now, I’ve never had any
direct person-to-person contact with these owners. I don’t move in their
social circles. But in our complex social system they interact with me
indirectly by locating their restaurants at interchanges of busy highways
as a predictably easy feedback path for controlling perceptions that I
and other travelers are predictably likely to want to control. And thus,
even though I’m a free agent who can attempt to control any perception I
want to control, I’m subject to their manipulation. (Don’t get me started
on McDonald’s ads.)

BP: I’m not saying that there are no influences at work here. But they
are not as direct and simple as they look, nor as reliable as one might
wish they were. And influences have to be assessed at the same level of
control, not across levels. Social variables are those that apply ONLY to
groups; individual variables apply ONLY to individuals. When you mix them
together, I think you get mainly nonsense, like designing doors to
provide three inches of clearance for the average person’s head.

Also, I’m not saying that I think sociologists have never thought of
problems like these. I’m sure they have. But they have never had PCT to
offer a better alternative. That’s what we’re here for.

Best,

Bill P.

[Bill_Powers1]

[From Bill Powers (2012.11.25.1755 MDT)]

Kent McClelland (2012.11.25.1315)]

KM: The first step is for us to
observe some stability or regularity in our shared physical environment,
where instead we might expect to see some fluctuation, and to observe
actions on the part of the putative controller that appear to relate to
that stability. In the coin game, we see coins sitting on a table in some
physical arrangement, whereas people don’t usually just leave coins
sitting around for no reason, and we see that the person sitting across
from us has arranged the coins in that way.

BP: The problem here is that the controlled variable is not just an input
quantity, it is some function of a set of input quantities. In the coin
game, the input quantities in the environment are all the properties of
the coins: their denominations, their sizes, their coordinates on (or
off!) the table, and so on. The different possible controlled variables
are different functions of different subsets of all the input
quantities.

This means that you can’t see the effective feedback path from output to
controlled perception until you know (a) which properties of the input
quantities are sensed, and (b) what function of those quantities is being
controlled.

In the coin game that uses four coins of different denominations, one
controlled variable that people have successfully controlled, and others
have discovered, is the relationship “three coins in a straight
line.” Note that the feedback paths relate only to the coordinates
of the coins in relation to each other: the x,y coordinates of any three
of the four coins are such that the ratios of x to y are all the same.
The denominations and sizes of the coins are irrelevant, as are the dates
on them, their colors, and which three coins are to form the straight
line. The sizes of the numbers representing the three sets of coordinates
are irrelevant as long as they remain in the right ratio for x and y. If
the reference value of the ratio x:y is 3:4, then the reference condition
is matched when a given coin’s coordinates might be (33,44) or
(3000,4000).

When we speak of the feedback path, we don’t mean just a line drawn on a
piece of paper to show where the path is. We mean the quantitative
properties of the environment along that path that determine exactly how
much effect there is. If there are two quantities in the environment with
values of x and y, a person might act on them to generate a particular
sum, x + y, or a particular difference, x - y (or both at the same time
independently, if there are two control systems). And there may be many
different physical paths through which a given action might affect the
values of those quantities. Each different path would be likely to give
the same action a different amount of effect, yet any one of those paths
could be used to achieve the same reference condition. If one path
created a smaller effect on the controlled variable than another, then
the control system would produce more output when using the first path
than when using the other.

We aren’t speaking here of qualitative identification of a controlled
variable. We have to consider the quantitative identification, because
that determines how much behavior we will see, and what direction of
behavior, when control is successful.

The coin game is a good example to use in speaking of affordances
(Taylor’s effordances, meaning properties of the environmental feedback
function). There are huge numbers of effordances for any controlled
variable – actually infinite numbers of them when we think
quantitatively. Many quantitatively different effordances make no
important difference in controlling a given controlled variable, since
control systems can handle wide variations in the loop gain without
materially altering the quality of control.

I think rick Marken is right when he keeps insisting that you can’t even
define an *ffordance until you know what the controlled variable is. It’s
only when one thinks qualitatively, in categories rather than quantities,
that the concept makes any degree of sense. And then that degree seems
pretty small.

Best,

Bill P.

[Kent]

[From Kent McClelland (2012.11.25.2005 CST)]

Thanks, Bill, for your cogent replies [(2012.11.25.1755 MDT) and (2012.11.25.0955 MDT)] to my latest posts. You’ve given me lots to think about, and unfortunately I won’t have much time over the next week to put any of those thoughts into pixels, since
I’ll be crossing the Atlantic the following week for the holiday season, and many other things need to be done before that trip.

When I’ve been able to put together something reasonable to say, I’ll be back to you.

My best,

Kent

···

On Nov 25, 2012, at 7:52 PM, Bill Powers wrote:

[From Bill Powers (2012.11.25.1755 MDT)]

Kent McClelland (2012.11.25.1315)]

KM: The first step is for us to observe some stability or regularity in our shared physical environment, where instead we might expect to see some fluctuation, and to observe actions on the part of the
putative controller that appear to relate to that stability. In the coin game, we see coins sitting on a table in some physical arrangement, whereas people don’t usually just leave coins sitting around for no reason, and we see that the person sitting across
from us has arranged the coins in that way.

BP: The problem here is that the controlled variable is not just an input quantity, it is some function of a set of input quantities. In the coin game, the input quantities in the environment are all the properties of the coins: their denominations, their sizes,
their coordinates on (or off!) the table, and so on. The different possible controlled variables are
different functions of different subsets of all the input quantities.

This means that you can’t see the effective feedback path from output to controlled perception until you know (a) which properties of the input quantities are sensed, and (b) what function of those quantities is being controlled.

In the coin game that uses four coins of different denominations, one controlled variable that people have successfully controlled, and others have discovered, is the relationship “three coins in a straight line.” Note that the feedback paths relate only to
the coordinates of the coins in relation to each other: the x,y coordinates of any three of the four coins are such that the ratios of x to y are all the same. The denominations and sizes of the coins are irrelevant, as are the dates on them, their colors,
and which three coins are to form the straight line. The sizes of the numbers representing the three sets of coordinates are irrelevant as long as they remain in the right ratio for x and y. If the reference value of the ratio x:y is 3:4, then the reference
condition is matched when a given coin’s coordinates might be (33,44) or (3000,4000).

When we speak of the feedback path, we don’t mean just a line drawn on a piece of paper to show where the path is. We mean the quantitative properties of the environment along that path that determine exactly how much effect there is. If there are two quantities
in the environment with values of x and y, a person might act on them to generate a particular sum, x + y, or a particular difference, x - y (or both at the same time independently, if there are two control systems). And there may be many different physical
paths through which a given action might affect the values of those quantities. Each different path would be likely to give the same action a different amount of effect, yet any one of those paths could be used to achieve the same reference condition. If one
path created a smaller effect on the controlled variable than another, then the control system would produce more output when using the first path than when using the other.

We aren’t speaking here of qualitative identification of a controlled variable. We have to consider the quantitative identification, because that determines how much behavior we will see, and what direction of behavior, when control is successful.

The coin game is a good example to use in speaking of affordances (Taylor’s effordances, meaning properties of the environmental feedback function). There are huge numbers of effordances for any controlled variable – actually infinite numbers of them when
we think quantitatively. Many quantitatively different effordances make no important difference in controlling a given controlled variable, since control systems can handle wide variations in the loop gain without materially altering the quality of control.

I think rick Marken is right when he keeps insisting that you can’t even define an *ffordance until you know what the controlled variable is. It’s only when one thinks qualitatively, in categories rather than quantities, that the concept makes any degree of
sense. And then that degree seems pretty small.

Best,

Bill P.