Angus Jenkinson (2018-01-02)
Kent
Thank you for these remarks. I am responding to the specific passage about modelling, and how and why you have turned to it.
I do not have reservations about modelling as such .
It is an important and valid way of approaching understanding. This also very much includes mathematical modelling. Any reservations that I have relate practically to the nature of the model itself. Following good cybernetic practice, a model is only capable
of giving a true representation of the variety that it models and the problem is when people extrapolate beyond the level of explanation that the model is able to give from its own design. Worse is when they discount other modes of knowing. But modelling is
an excellent tool and I have used it myself many times in different ways over many years. Moreover I have designed many thinking tools, methods of analysis and design, that have been widely used by practitioners, each of which is a form of model. My aim has
been to develop what I call archetypal models, i.e. models that embrace a wide variety of concrete differentiated instances.
What I have also found personally helpful as another route for understanding PCT is a method that I have been involved in for some decades. It could
be called personal experience, but the specific aspect of it is the ability to enter consciously into personal activities of thinking, feeling and will. With respect to the last, and the bodily activities of movement, sensing, and so forth, I met a dance
practice and movement practice back in the 1970s called Eurythmy when I was a teacher and at the time it was part of the curriculum for the pupils, so I had to take part with the Eurythmy teacher. This means that a purposefully controlled awareness directed
behaviour has been part of my world for decades: what PCT brought was a precision of operational explanation. That’s really why the activity was part of the curriculum: it brought conscious awareness – as well as practical self-conttrol – to the processes
whereby the imaginnatively conceived goal is realised in movement.
Over time I also came to appreciate the way in which the normally entirely unconscious realm of our physical activity can be rendered more conscious.
I mean that in the normal case we walk along the road with very little awareness of the actual activity of walking. We are not aware of what our feet are actually doing, nor the whole movement and flow through the limbs. We know it may not even be particularly
conscious of the processes of observing the pavement that we are navigating. It is easy to see in others on the street. This whole realm of what is called Will, the conative, and which is expressed in behaviour, is normally one in which we more or less go
to sleep, it is rather unconscious. Anyway, the upshot is that as a supplement to other forms of thinking and analysis, I can to some extent directly observe PCT in my own behaviour. I also find such experiments as the rubber band exercise compelling.
···
…â………………………………………………………………………….
Angus Jenkinson
On 27/12/2017, 01:39, “McClelland, Kent” MCCLEL@Grinnell.EDU wrote:
From Kent McClelland (2017.12.26 16:30)
KM: My reason for getting involved in mathematical modeling was that the dynamics of control systems in action and interaction weren’t obvious to me when I first encountered PCT, and so it took a lot of playing around with models for me
to get a clear sense of how control systems work. Maybe a person can gain a deep understanding of PCT without trying out any mathematical models, but some hands-on tinkering with models really helps, in my experience.
If you’ve made an intellectual commitment, as I have, to the proposition that the dynamics of control systems are our best conceptual model for understanding the dynamics of human behavior, then you need to have a thorough knowledge of
how control systems actually work in order to know what to look for in examining empirical instances of human behavior. I don’t pretend that I’ll ever be able to model a strip of human interaction in all its real-time complexity, but some basic modeling seems
really helpful for cutting through some of that complexity.
I would say that the reason for using a model of any kind—mathematical or conceptual—is to highlight significant pattterns that might be possible to pull out of the messy complexity of an empirical phenomenon. Of course, models can never
be more than abstractions and simplifications of the reality to which they are applied, but for scientists (and folks like me who try to think scientifically) mathematical models are particularly useful because they provide such a compact and rigorous way
of communicating scientific descriptions of empirical phenomena. In comparison, ordinary language tends to be fuzzy and imprecise.
Does this speak to your reservations about modeling, Angus?
Kent
…………………………………………………………….
Angus Jenkinson
On 10/10/2017, 23:42, “McClelland, Kent” MCCLEL@Grinnell.EDU wrote:
From Kent McClelland (2017.10.09.1500 EDT)
Re: Angus Jenkinson (2017-09-26. 23.57), Martin Taylor 2017.09.27.10.40
HI Angus and Martin,
At long last, I’m back in Grinnell and have had some time to think about the questions you asked, Angus. I’ll start with a response to that post
and treat Martin’s comments in another.
AJ: Thank you Kent, I am very grateful to you for sharing this.
I have started studying your paper and I can see the elegance of having a mathematical model for the purposes. But also analytical thinking. I regard it
ass a major contribution. More of which later…
KM: Thanks!
AJ: Although I have worked with conflict, it is not a speciality of mine. So what follows is an enquiry.
KM: My focus on conflict has led me to see it as a ubiquitous feature of social interaction, just as cooperation. collaboration, or coordination of one kind or another occur in almost every interaction. I think
of conflict and cooperation, then, as two varieties of the more general phenomenon of collective control. So, I’d say that anytime you study social interaction you’re implicitly dealing with conflict, and cooperation, too, of course.
KM: Granted that most everyday interactions between people don’t take the form of a conflict escalation spiral, like those described by Collins and analyzed in my paper. The reason that most conflicts don’t get
out of hand, as I see it, is that everyday interactions nearly always require collective control at a variety of perceptual levels simultaneously, and while conflict may be occurring at one perceptual level, cooperation more typically occurs other levels of
control. Thus, for us to do a comprehensive analysis of any given social interaction, we have to think in terms of the participants’ collective control of perceptions on a hierarchy from low-level perceptions of the shared physical environment to the middle-level
perceptions involved in language and communication, to high-level perceptions, such as norms or social identities.
KM: Some examples: Two sports teams are in conflict over which team will win the game, but they cooperate at low levels of perceptual control by playing together on the same field or pitch, using the same ball,
etc., and at higher levels of control by adhering to the perceptual principles defining the rules of the game. In most contests, the conflict between the teams doesn’t ever escalate into unbridled conflict. However, sometimes a member of one team commits what
appears to be a deliberate foul that injures a player on the other team (or just humiliates him). A series of tit-for-tat physical aggressions may then occur as the teams seek to even the moral score, and we see an example of the familiar conflict-escalation
pattern.
KM: Two people argue with each other. Cooperative low-level collective control nevertheless occurs as they use a common language and orient their bodies to face each other (although at these low perceptual levels
some of their physical gestures may symbolize threats of violence to the other). While everyday arguments commonly result in patterns of conflict escalation, as the participants trade increasingly cutting insults, for instance, most arguments don’t degenerate
into fisticuffs. A typical reason for this restraint may be that the participants are cooperatively controlling some higher-level perception, perhaps a mutual friendship or relationship that they want to preserve. Another possibility is that both of them recognize
and adhere to higher-level perceptions of norms that prohibit doing violence to another person or prescribe penalties for physical aggression.
KM: In general, the conflict-resolution literature suggests that one good way to resolve a conflict is to focus the combatants’ attention on some goal that they both share, or in other words to get them to cooperatively
control some higher-level perception, rather than conflictively controlling the low-level perceptions that they are fighting about. Interestingly, the MOL method of therapy operates on a similar principle of directing the attention of a client caught in an
internal conflict to some higher-level perception connected to the two lower-level perceptions in conflict, which then enables the client to prioritize which of the incompatible lower-level perceptions to control, or do something else entirely.
KM: Another example: A work team or a committee work together to complete a project or reach some other common goal. The overall tenor of the interaction is cooperative, but along the way lots of little conflicts
pop up about the best means toward reaching their goals. If the group is working together effectively, these conflicts are handled by talking these conflicts out with each other or else deferring to a boss whose word is taken as authoritative and who tells
them what to do. The basic purpose of verbal communication, in my view, is to enable people to align the reference values to be used for their collective control of perceptions. What communication communicates, I would say, are reference values for the perceptions
that the speaker wants the listener to control. People who want to work together cooperatively need well aligned reference values for their common perceptions, and lack of alignment is tantamount to conflict.
- AJ: You say: “Of course, every approach to modeling relies on simplifications and abstractions, and degrees of realism exhibited by agent-based models will vary with the level of detail built into the
modeled agents…â€?? You also say that it does not reproduce the detail of Collins’ explanation. Understood, but one of the cybernetic principles was that the best model was reality itself. Moreover, you comment that Collins does not give precise enough definitions.
But part of the problem of “definition� is that it tends to reduce the variety of the phenomenon to a boundary defined by the definition. So I am interested in the degree to which you think that real life (sic) can be reproduced in a mathematical model? I
return to this below…
KM: The highly simplified mathematical models that I used in my paper pertain to only a single level of perception. When multiple levels of perception are in play, as in the examples I’ve suggested above, these
simple mathematical models obviously can’t begin to reproduce the real-life complexity of cooperation on one level of perception and conflict on another. More sophisticated multi-level mathematical models might be able to do a little better job, but it would
take an enormously complex mathematical model to simulate the multi-level complexity of the real-life interaction. What mathematical models are good for, in my view, is to alert us to the basic patterns of dynamic interaction we might expect to see in these
real-life instances and also to provide a theoretically grounded explanation of why such patterns are likely to emerge.
- Building on that, as an illustration, you say:
- The other core variables in Collins’s model also lack of the kind of conceptual precision needed to support mathematical reformulation. Collins explicitly defines the variable atrocities as a matter
of the perceptions of people on one side of the conflict interaction: “Atrocities are opponents’ actions that we perceive as especially hurtful and evil . . .� (2012:2). However, he does not offer any operational definition * that
would enable observers to judge whether a particular action committed by one side is likely to be defined as an atrocity by the other*. My emphasis
- One of the conditions of conflict I understand is that the supposedly rational judgement is lost and people tend to find unacceptable, and perhaps extremely so, based on, let’s call it, a biased reaction.
So what the person deems to be an atrocity might not be seen as such by an observer. That’s easily understood.
KM: Nevertheless, many of the violent atrocities that Collins describes would strike almost any observer as atrocities. When you kill scores of people, for example, that’s an atrocity in anybody’s book. My definition
of violence, from a PCT viewpoint, is the destruction of environmental feedback paths that someone else needs to keep their highly valued perceptions under control. When you kill someone, for instance, you render their physical body nonoperational, and the
physical body always forms two essential links in the circuit of pathways used by individuals to control their perceptions: on the output side, the muscular movements that enable individuals to manipulate their physical environments, and on the input side,
the sensory organs that report the effects of those manipulations. But violence can also consist of destroying physical objects that individuals rely on to control perceptions, like their houses or tools. It’s when you get to higher levels of perception, what
Pierre Bourdieu referred to as symbolic violence, that definitions of what an atrocity might consist of get more subjective. Is destroying someone’s good name, for instance, an atrocity? It’s certainly an essential link in the feedback path that people use
to maintain a positive sense of self.
- One of the rediscoveries of neuroscience is the role of emotion in decision making. I am interested in how the strength and roles of those appear in the model?
- You refer to reference values as memories (aka mental pictures). So when you say you “will take polarization to refer to the extent of differences in reference values�, I take that to mean that the
(actively filtered and probably distorted) perception is compared by each party to something conjured from their own memories (forgive the unmathematical description).
- You also say, “PCT models show the output of the two interacting systems diverging, with one system, in effect, pulling in one direction to bring the environmental variable into line with its preferences,
while the other system pulls in the other (see McClelland 2004, 2006). Hence, the variable used in my simulations is the intensity of the conflict, measured by the degree of divergence in system outputs.�Outputs are to the control variable of the other as
I understand it.
KM: As Martin notes in his reply, one’s outputs affect the shared environmental reality, and the opponent controls a separate perception based on that environmental reality, but that perception isn’t necessarily
based on the opponent’s perception of the individual’s output. What happens in conflict is that a positive feedback loop emerges that links the two opponents’ outputs, the portions of the environment on which both outputs are acting, and their respective perceptual
control loops.
- So how do you process emotion in the model?
KM: Martin’s reply, directing you to Ch 17 in the second edition of Behavior: The Control of Perception, is a good start for answering the question about the place of emotion in this model. What I’d add is that
Powers argues that the experience of prolonged inability to bring a valued perception under control inevitably produces negative emotions like frustration and anger. In conflict situations, the actions of the opponent by definition prevent one from bringing
the contested perceptual variable into control at one’s own reference level, so the PCT definition of conflict implies that these situations always produce negative emotions. Emotion is just an expected part of conflict. The other thing I would say is that
because people are operating on several different levels of perception at the same time, emotional reactions do not preclude rational behavior. The two processes can go on simultaneously.
- How do you see the model contributing to practitioner work? And what other variables would need to be included? Is it more than offering a conceptual framework to think differently? Meaning, how would
a version be used in Trump v Kim JU?
KM: The mud-slinging match between Trump and Kim sure looks like a classic conflict-escalation cycle to me. What my model does not do is predict the outcome of this escalating conflict, because it doesn’t take into
account the other presumably higher-level perceptions that these two actors may be controlling. Let’s hope that there are some perceptions that restrain them before they start pushing nuclear buttons!
- It seems that being able to prove similar results to positive feedbacks through negative feedback is very useful in rethinking what happens.
- It also seems to me it replicates what practitioners in conflict resolution do. They ask each party to see how the other party sees things. Negotiators teach the same thing.
KM: Good point.
- Might however the +ve and –ve bee mirrors of each other according to how one analysis the situation? One is dealing with a series of escalating outcomes that propel each other and the other a series
of negative (internal) cancellations that also effectively propel each other? Your strength is to show the interaction, maybe?
KM: I’m not quite sure what you mean by the ±ve in the model. The value of the environmental variable [v] times what [e]?
- The paper also introduces some clarity in thinking about issues. This is a typical case in all model building and as long as one does not allow a spurious impression of detail to get in the way, can
help. I like for example the parsimonious description: “From the PCT perspective, conflict occurs whenever the two sides use different reference standards in their attempts to control an environmental variable, and victory or defeat in the contest is represented
by the extent to which side or the other succeeds in bringing the variable into line with its own preferred reference conditions�.
- I wonder whether it does not verbally – as a result of yyour thinking through the issues with the help of modelling – give the practitioner as goood an insight as a mathematical model?
KM: The model goes beyond the verbal description to show the dynamics of the interaction, something the verbal description can’t really do.
BTW: You mention: ““reflex arc� as conventionally understood�; which version of this do you mean? Presumably not the one that Dewey critiqued?
KM: No, I am talking about Dewey’s reflex arc, as well as the set of internal neural connections that provides the focus of most current neurological analyses.
KM: Thanks again for your good questions and your interest in my work!
Kent
—&nbssp;
A
On 24/09/2017, 20:20, “McClelland, Kent” MCCLEL@Grinnell.EDU wrote:
[from Kent McClelland 2017.09.24.1735]
Hi Angus (and Christopher),
I’ve done a lot of thinking and some PCT-based modeling that relates to questions of collective action, organizations, and, more broadly, how society is possible.
The modeling I’ve done may strike you as pretty simplistic for your purposes, but the best summary of my work along that line can be found in a paper published in 2014 in the journal Sociological Theory called “Cycles
of Conflict: A Computational Alternative to Collins’s Theory of Conflict Escalation.” You can download a copy of this paper, as well as my other published papers on PCT, from my page in ResearchGate.
A much longer and more detailed, but strictly theoretical, exposition of my ideas about “collective control”, as I’ve called it, is contained in my chapter for the Living Control Systems IV volume being published
by Bill Powers’s sister, Alice. The chapter is called “Social Structure and Control: Perceptual Control Theory and the Science of Sociology.”
Unfortunately, the LCS IV volume has been slow to appear and still remains unpublished. I don’t know when it will be out, but the publisher has requested that authors not distribute any draft copies of the chapters
prior to publication, so I can’t share my chapter with you at this point. I’ve posted a summary of the chapter on ResearchGate, but it’s probably too condensed to provide any very good sense of my argument.
Martin Taylor’s contribution to the LCS IV volume may also be of interest to you when that book appears. I’ve seen some early drafts of his chapters, and many of his ideas were immediately influential to me as I put my own chapter together. He comes at the
question of how to apply PCT to social interaction from a somewhat different angle than I do, but his work is unlike anything else I’ve seen, and I think it will prove foundational for lots of future applications of PCT in this area.
Please send me some examples of your work if there are things you’d like to share. I’m definitely interested.
And thanks to all for the birthday greetings! I’ve had a fun birthday going with two of my young grandsons to the chocolate museum in Hamburg, Germany.
Kent
On Sep 24, 2017, at 3:08 PM, Christopher Paul Geddes geddes.chris@gmail.com wrote:
Angus,
Thanks for writing. I hope all is well. I have done some modeling in the past however I was kind of limited and now I am more focused on Systems Dynamics which is making me take more of an interest in modeling now
as well.
I would like to know more about what you are looking for if I can help. I like to think of organizations as entities that I can influence with technology so that things so the way the team wants them to go.
Christopher
On Sun, Sep 24, 2017 at 5:48 AM, Angus Jenkinson angus@angusjenkinson.com wrote:
[Angus Jenkinson, 2017-09-23.11.46]
Kent, Bruce Nevin has suggested that I should reach out to you in relationship to some work on modelling that you have been doing on collective behaviour.
I am interested, and working with, the behaviour of organisations and social groups and the relationship between the individual and the group. I understand you have done some computer modelling of this. Is there something you can share?
And happy birthday!
………………………â€â€¦â€¦â€¦â€¦â€¦â€¦………………………….
Angus Jenkinson
–
Regards,
Christopher Paul Geddes
Email: geddes.chris@gmail.com
Mobile: +1-876-393-1366