Entropy production and emergence in behavior LO20309

Jon Krispin (jkrispin@prestolitewire.com)
Wed, 06 Jan 1999 17:39:19 -0500

Replying to LO20232 by At de Lange (incorporating a plethora of other

Happy New Year to everyone!

In LO20232, At de Lange wrote the following in response to my proposal (in
LO20211) that the antecedents for behavior were the entropic force and
that the consequences for behavior were the entropic flux:

>The one thing that complicates matters very much, is that we have to
>seek for a common pattern throught the entire web-of-reality. In other
>words, what I have to do now, is to take your suggestions and see how
>far I can go with them in the physical world (inanimate as well as
>living) and the spiritual world (creating, learning and believing).
>You will have to give me time to think them carefully over. In the
>mean time, keep up with the exploring and the dialog!

Greetings At,

The following is an attempt on my part to keep up the dialog. I have been
thinking continually on this stuff, and I have made some new connections
and have some additional questions that might stimulate further dialog.
In your Primer on Entropy, part III C, you defined entropic force and
entropic flux in the following manner:

Entropic force is a kind of "being" and an intensive quantity.
Entropic flux is a kind of "becoming" and an extensive quantity.

You also gave a criteria the 'halving the system' criteria for
distinguishing between the two - when a system is halved, some properties
(the intensive quantities, or the entropic force) stay the same (for
example, voltage in an electrical energy system), while others (the
extensive quantities, or the entropic flux) will also be divided in half
(for example, amperage in an electrical energy system).

In LO20120, you wrote the following:

>The antecedents (A) of Jon Krispin are nothing else than the entropy
>producing force-flux (tension flow) pairs. The consequents C
>themselves can and usually do act in the entropy producing force-
>flux pairs of a successive phase, thus giving rise to a string of
>ABC-ABC-ABC's. ...snip...

I tentatively tried to refine this by suggesting that the antecedents were
the entropic force and the consequences were the entropic flux. As you
mentioned in LO20232, it is difficult to consider whether or not this is
the case, because it is much more difficult to parse out the intensive and
extensive quantities in behavioral systems and systems of the mind than it
is for physical systems. In lieu of this, you suggested conducting
thought experiments as one alternative that might help us sort this out.

>From LO20232:

>Here is a thought experiment. The truth values "true" and "false" set
>up between them an entropic force. In sharp logic it is a discrete
>force and in fuzzy logic it is a continuous force. When I think away
>(like in phenomenology) half of the mathematical theorems which hover
>in my mind, the force set up by the difference between "true" and
>"false" stays the same for me. ...snip.. In other words, the two truth
>values "true" and "false" express an intensive quantity which we
>not even yet has given a name. ...snip... Whereas "true"
>and "false" stays the same, it is not the case with "truth". In other
>words, if I remove half the mathematical theorems from my mind, the
>mathematical "truth" within me is less (half?). This means that truth
>is an extensive quantity.

In this thought experiment, you recommend halving what might be considered
as the structure of a system of truth (consisting of the mathematical
theorems). This halving of the system (thinking away half of the
mathematical theorems) allowed you to distinguish between the intensive
properties (the true and false differences which make up the entropic
force) and the extensive properties ("truth" which is the entropic flux).

I had suggested a similar thought experiment in terms of behavior in
LO20211, where halving a behavioral system (the antecedents and the
consequences for a given behavior) lead to a change in the rate of a
behavior. [Here was my first clue that the behavior is actually
representative of the entropic flux because it is the property of the
behavioral sytem that is halved when the system is halved, indicating that
this is an extensive quantity.] I had proposed that this "thought
experiment" suggested that my definition of antecedents as entropic forces
and consequences as entropic fluxes had some merit. I am now fairly sure
that this is not correct, a conclusion that is based on a number of
contributions to the list.

First among these was a comment that you wrote later in LO20232:

>What you are doing, in terms of my own understanding (which might be
>utterly wrong), is to articulate our sensitivity to the first
>manifestation of entropy production as "increased diversity of
>BECOMING (chaos)". Since entropic fluxes are also "becomings", you
>identify this first manifestation (first consequent) as the entropic
>fluxes. Further more, your articulation of the antecedents as entropic
>forces (rather than as force-flux pairs) is because the force-flux
>pairs do not exist on their own, but emanate from the existing
>being (structure) at that point of time (the present). In other words,
>RISE TO THEM. [caps and emphasis mine, not At's] ...snip...

Second was a comment made by Leo Minnigh (Hello, Leo!) in his post LO20246:

>What causes all these flows? One needs something to flow, and a difference
>in level (high-low), in other words a force (in the case of water, it is
>the gravitational force). But how does this force acts? Does it push, or
>does it pull? Or can't we speak in these terms? there are some great
>differences between these two, although in strict vector calculations
>there is no difference. However, the ultimate example of 'pushing' is the
>big bang in the beginning of our universe. This explosion spreads outward
>from a centre. ...snip...

>If we think of pulling or attracting, instead of pushing or spreading, we
>think of concentration. Pulling concentrates. Again, the example from
>cosmology is the black hole.

Third was my own comment from later in the very post where I suggested the
antecedent=force and consequent=flux distinction (LO20211):

>In a system of behavioral energy, the antecedents prompt (or "cause"
>in the Kantian sense, see my post in LO20152) the behavior into "being".
>The occurrence of every individual behavior is a potential bifurcation
>point in the system. The consequences which follow act like attractors
>in the system that influence, or shape the becoming of the behaviors.

At, I think your observation that I was reducing my awareness of the
force-flux pairs to the structure which gives rise to them is correct.
Even looking at my own words (especially after reading Leo's post) I
became conscious of this. It seems more plausible that the antecedents
and the consequences are both entropic forces, and that the entropic flux
to which they give rise is evidenced in the behavior that results. Each
discrete behavior is a "being" in its own right, and is representative of
the entropy in the behavioral system at that moment in time. However, by
observing the "flow", or change, in behavior over time, we can see the
immediate results of the entropy production and the "becoming" in the
behavioral system.

The distinction between antecedents and consequences may better be
understood as the different types of forces that Leo describes, the push
and the pull. Antecedents may be considered as an entropic force that
PUSHES a behavior (prompts it into existence) and consequences may be
considered as an entropic force that PULLS the behavior and shapes the
form of its flux/flow, or the direction that the organization of the
behavioral system takes.

antecedents --> force --> push/prompt --> spread
consequences --> force --> pull/attract --> concentrate

Together, the antecedents and the consequences work to form the structure
within which the behavioral energy system exists.

Leo further illustrated this difference in the types of entropic forces
using an example from the flow of water when he wrote the following in

>Where is our water-analogy? Think of a bathtub. If we could fill the bath
>with water by a tap which is mounted in the bottom of the bath, we will
>see something in the water which resembles the mushroom cloud of a nuclear
>bomb: the water spreads towards all sides from the centre in over-rolling
>'clouds'. After a while the bath is filled and we close the tap. Now we
>take the plug off. We introduced a 'puller' or attractor. We soon will see
>a maelstorm, a singular vortex spiralling towards a centre. After a while,
>the water level in the bath diminishes, the speed of flow decreases, and
>the last water runs in a laminar way towards the hole.

And you also reinforced the concept of the push and pull as two forms of
entropic force in your response to Leo (LO20274):

>Leo, I want to express my gratitude because you have articulated
>something with which I was only tacitly aware of - the push-pull
>effect in the entropic force. The force is made up by a difference in
>some intensive factor. Two different values of the intensive factor
>are needed to make up the difference. The one value will push
>dispersingly while the other value will pull concentratingly. This
>dual action set up a whole field of possible flows. ...snip...

Actually, as I am writing this and thinking about this difference between
the push and the pull, I think that we have to even subdivide the types of
consequences. In earlier posts (such as LO19894 and LO20081), I
differentiated between consequences that increase the future likelihood of
the occurrence of a behavior (reinforcement) and those that decrease the
future likelihood of the occurrence of a behavior (punishment and
extinction). Using the lens of the push and the pull, we can reconsider
especially the two areas of reinforcment (positive reinforcement and
negative reinforcement), and also understand punishment and extinction in
the flow of behavior. Positive reinforcement results when a behavior is
followed by a desired outcome (consequence), "causing" an increase in the
future likelihood that the behavior will occur again. This type of
reinforcement is the most powerful for increasing the rate of a behavior
and sustaining it over time. It will pull behavior, like the attractors
that we have described above.

On the other hand, negative reinforcement acts much more like a push. As
I have mentioned, negative reinforcement is also commonly known as escape
and avoidance behavior, due to the fact that the behavior that results is
"driven" (or pushed) by the desire to avoid or escape an undesired outcome
(or the potential occurrence of one). In either of these cases, an
ANTECEDENT condition exists that the individual will work to alleviate. In
many organizations, the creation of the possibility of an undesired
outcome is often used as a means for "motivating" and even "controlling"
behavior. These potential undesired outcomes may take the form of an
unhappy customer, a looming deadline in a project, or the threat of a poor
performance appraisal. It is these threats of undesired consequences (or
their actual appearance) that causes fear to exist in organizations (If we
want to drive out fear, as Deming suggested was necessary, here is our
clue as to how to do this). And I think that the fact that the condition
that the individual works to alleviate is an antecedent condition is
significant and explains why negative reinforcement also takes on the
characteristics of a pushing force.

The problems with this type of reinforcement as a means for causing
behavior are manyfold. One is the fact that the behavior that results may
not be exactly what is needed or wanted (the spread goes in all
directions, not just the desired one). The second is that it takes a
great deal of energy to maintain the flow once it is started. One of the
defining characteristics of behavior that is reinforced using negative
reinforcement is that it stops as soon as the undesired condition (or the
potential of one) is removed or reduced (the most that can be expected is
the least required to get by). If more of the behavior is desired, more
of the pushing force must be applied. If this pushing force is removed,
the behavior quickly returns to its prior equilibrium state (determined
most likely by the naturally occurring positive reinforcers for competing
behaviors - indeed, the lack of positive reinforcers for the desired
behavior will result in the EXTINCTION of the behavior over time, a kind
of "PASSIVE RESISTANCE", or friction that works against the emergence of a
new equilibrium state) and the desired behavior immerges.

Now that we have defined positive reinforcement as the attractor, or pull,
for behavior, negative reinforcement as a form of push for behavior, and
extinction as a form of passive resistance to behavior, we can consider
punishment from within the same framework. Punishment (which occurs when
the behavior is followed by an undesired outcome/consequence, resulting in
the decreased likelihood that the behavior will occur again in the future)
is an ACTIVE form of RESISTANCE to the flow of behavioral energy, also
generating friction that works against the flow of behavior and can
ultimately result in the immergence of the behavior.

>From this perspective, it is clear that postive reinforcement is
absolutely necessary for irreversible self-organization in a system of
behavioral energy. And it becomes clear why this is the case as well.

In LO20221, At helped us to understand the difference between a system
that is at a labile equilibrium and one that is at a stable equilibrium:

>They do acknowledge the "butterfly effect", but only in the sense of
>an amplification, causing an "avalanch" in a system which is at a
>labile equilibrium. A stationary car on the top of a mountain pass has a
>labile equilibrium. One push and it begins to rush to the valley below.
>On the other hand, a stationary car at the bottom of the valley is at
>a stable equilibrium. One push and it again grinds to a standstill.

We can also think of At's illustration of work and irreversibility in his
Primer on Entropy, part III B (LO20048), where he describes the difference
in free energy present in Jack when he is at the bottom of the hill vs. at
the top. There is much more free energy present when he is at the top,
and this free energy can be harnessed to do work (such as pulling Jill up
the hill. Clever Jack!). Jack's movement from the top of the hill to the
bottom of the hill is spontaneous (requires no work). However, moving Jack
up the hill when he is at the bottom requires the non-spontaneous
introduction of additional energy into the system to be used in the work
of getting Jack to the top.

This parallels the characteristics which distinguish negative
reinforcement and positive reinforcement. If negative reinforcement is
exclusively present as a means for pushing behavior, much work is required
and the resulting behavior is non-spontaneous (it will stop as soon as the
pushing force is removed, just as the car in the valley will grind to a
standstill --> immergence). When positive reinforcement is used to pull
behavior, much less effort is required ultimately to realize the same
behavioral movement (the change in behavior may even seem to be
spontaneous, as the car at the top of the mountain rushes towards the
valley - a new equilibrium state --> emergence).

As I first mentioned in 20211, the strength of a consequence to attract
behavior is dependent on several criteria. First, there must be a feedback
loop present to connect the individual with the particular consequence.
Next, the consequence must be desired by the individual (more so than the
other consequences that are also available for competing behaviors). The
other 2 dimensions that play a role in the ability of a consequence to
attract behavior are the immediacy of the consequence in relation to the
occurrence of the behavior, and the certainty that it will result. The
more immediate and certain the outcome, the more powerful the ability to
attract behavior.

In some behavioral systems, behavior is in a state of stable equilibrium,
that is, the positive reinforcers that are attracting behavior do not have
other competing positive consequences that are more desired. In such
cases, if we (as the baker of the process) want some other behavior to
occur, we must introduce additional energy into the system by providing
some socially mediated (preferably positive) reinforcer that will begin to
change behavior. In order for the behavior to continue, we will always
have to provide this social reinforcement. The instant that it is no
longer available, the behavior will begin to be extinguished. Without the
socially mediated reinforcement (be it positive or negative reinforcement
more prevalently used in these situations), the system will remain in or
return to its present state of [stable] equilibrium.

In many more situations, there are more desired outcomes present, but they
are unable to attract behavior in a head to head competition with other
available consequences due 1)to the fact that there may not be any means
for providing feedback to the individual regarding this outcome, or
2)other, more desired consequences may exist, but they may be sufficiently
delayed in time or uncertain to result so as to be ineffective in
attracting behavior. In these situations, we must "build the bridge"
between the behavior which occurs in the present and the future
consequences. In the first case, we will have to construct a feedback
loop to make the alternative consequence salient to the individual. In
the second case and in order to overcome the "decay of delay", we will
again have to provide some socially mediated positive reinforcement.
However, this intermediate reinforcement may not have to continue
indefinitely. If we are able to connect the individual with the future,
more desired consequences via our transition bridge, the realization of
these future consequences may become enough to begin to attract behavior
on their own, even when our bridge is withdrawn (in situations such as
this, consultants can play a powerful role in facilitating the emergence
of these new behaviors, and the engagement can be of limited duration).
This type of system is in a state of labile equilibrium with much free
energy. Our bridge intervention is the equivalent of pushing the system
over the edge (connecting with the future consequences) and allowing the
system to rush to its new attractor state of equilibrium.

At, the final question that I would like to address is one that you asked
in LO20232, although I don't know that I have an understanding of the
concept of the minimality of entropy production yet to give you an answer.
You wrote:

>What I never could ascertain clearly from the literature on behavioural
>psychology, is whether the A (antecedents) and C (consequents)
>"connect simultaneously". In other words, whether your scheme A=>B=>C
>is not actually A&C=>B. (the sign "=>" signifies a temporal separation
>while the sign "&" signifies the same "space-time" event. If that is
>the case, then we have much harmony in our tacit understanding. ...snip...

In my opinion, A&C=>B is the most accurate description of the relationship
of the influence of antecedents and consequences on behavior. The
simplest explanation (although I think it is inadequate to explain the
relationship) is that the consequences for one behavior become the
antecedents for future occurrences. In this case, the C's are actually
one of the A's. It is still necessary for us to draw the distinction
between the two because of the far greater influence that the C's have on
behavior than the "other A's".

Having said this, I think that the real answer (or at least a more
adequate one) to this question can be found in the "reverse arrow of time"
- in the minimality criteria for entropy production - but I haven't put
all of these pieces in place. In LO18655, you tried to explain this
concept in another context, and, although I have read it, I haven't been
able to make the translation to the present context. I have picked up on
the fact that it is the feedback loops for a system that determine this
minimality criteria (the upper bound and lower bound for the path that
entropy production must take), and that amplifying feedback and balancing
feedback are involved. If this is the case, then we can probably
translate it somehow into the push-pull & resistance characteristics of
consequences. After all, it is the consequences for a behavior the form
the basis of all possible feedback to a behavioral system.

Also playing into this is your description of equilibrium states as
attractor states from LO20163:

>It is even possible to think of the new equilibrium state itself as the
>ATTRACTOR state, something which lies somewhere in the future and acts
>backward into the past, pulling the disturbed system into a new order.
>This kind of thought, called teleology, is favoured by many scholars of
>the new science called complexity. But actually it is the "entropy
>production" in the present with its qualification "minimality" which does
>this guiding job!

Now we are back to the issue that Fred Nickols raised (particularly in
LO20098), that something which occurs in the future cannot possibly
effect, or change, something that has already occurred. If positive
reinforcers are the attractors that pull a system of behavioral energy
into an equilibrium state, then they must have a parallel in here

At, can you possibly make sense of all of this? Or, can you try to
explain the minimality qualification for entropy production again in the
present context? I think if you can, we will be able to revisit Leo's
comments and questions in LO20246 and your responses to his suggestions in
LO20274 with new insight.

Leo wrote:

>If these questions are of some valid, we have some clues for the
>behavioural sciences.
...snip that may be revisited in the future...

Many warm wishes in this new year!

Jon Krispin


"Jon Krispin" <jkrispin@prestolitewire.com>

Learning-org -- Hosted by Rick Karash <rkarash@karash.com> Public Dialog on Learning Organizations -- <http://www.learning-org.com>