Overview: Forms of Collective Control in Social Life

An Example of Form #1 with Conflict: Tug of War

My earliest presentations and papers on collective control showed a classic situation of conflict between control systems, where two control agents, acting upon a commonly perceived environmental variable, both try to control their perception of the variable but with different references for its value.

As I’ve repeatedly shown in my publications, this simple form of collective control by agents with different references can result in an escalating conflict of output. The joint actions of the control agents, nevertheless, stabilize the contested variable at a virtual reference level intermediate between their preferred references. The effect on the contested variable is indistinguishable from the result of the actions of a more powerful control agent using this virtual reference level to control its perception of the contested variable. Martin and I have described this hypothetical controller as a giant virtual controller, or GVC.

My previous example of a giant virtual controller—an audience clapping—demonstrates that escalating conflict is not an inevitable feature of this simplest form collective control, even when participants have a range of different reference levels for the jointly produced environmental variables (in this case, the loudness and length of the sound of clapping). However, escalating conflict often results in the simple form of GVC when the participants cluster around two incompatible references for their perceptions of the same external variable, and a good example is in the game called tug of war.

According to Wikipedia (https://en.wikipedia.org/wiki/Tug_of_war), tug of war has ancient origins and has been played in many societies worldwide. In the modern version described in the Wikipedia article, two teams grab a long, sturdy rope and pull it in opposite directions, trying to drag the members of the other team along with it. The collectively controlled variable in this case is the physical location of a center point marked on the rope.

At the start of the game, the teams set up on either side of three parallel and equally spaced lines on the ground, and the rope is stretched out across the lines and perpendicular to them. Team members array themselves along either end of the rope and then pick it up and hold it with the center point positioned directly above the middle line. Both teams commence pulling, and the object of the game is for a team to pull the rope’s center point past the outside line on their side of the field in spite of the resistance of the other team.

After the signal to begin the game, the conflict between the two sides escalates quickly as teams try to pull as hard as they can to gain an advantage. If one team is considerably heavier and stronger than the other, the game ends quickly. When teams are evenly matched, however, the value of the contested variable—the position of the middle of the rope—may not change much at all, at least not immediately.

If both sides are pulling at the limits of their abilities and neither has a decisive edge, a high-conflict stalemate can often last for several minutes before one side is overpowered by the other. A YouTube video of the 2015 UK Tug of War Championships shows a match that lasted more than nine minutes, with the teams in a near standoff until the last couple of minutes (https://www.youtube.com/watch?v=gQoM8kKDn6Y). The team that finally succeeds in pulling the center of the rope over the line on their side of the field is declared the victor, after which the contest ends and everyone stops pulling.

Stalemated conflicts are a common feature of social life, from family feuds to the halls of Congress, and whenever the teams in conflict can recruit new members to replace older members as they drop out—thus bringing new energy to the fight—these conflicts can last far longer than the relatively brief struggles of a tug-of-war game. However, such long-running conflicts are generally much more complex than the simple form of collective control I just described.

It’s worth noting that even the tug of war game has some complexities. The teams collectively control several other perceptions in addition to the focal perception of the position of the rope on the field. Members of each team control the high-level perceptions of belonging to their own team, and all of the contestants control their perceptions of the rules of the game and personal adherence (we hope) to those rules.

Finally, some of the YouTube videos of tug of war contests show an extra team member, called a “driver,” whose role is something like that of the coxswain in a rowing competition. The driver doesn’t hold onto the rope but instead stands beside the team and calls out commands to help them coordinate the rhythm of their efforts, so that they all choose the same moment to make an extra pull and then take rests together by “hanging,” that is, planting their feet and leaning their weight back against the other team’s pull on the rope (https://en.wikipedia.org/wiki/Tug_of_war). Among the perceptions that team members collectively control, then, is the perception of following the references suggested by the driver. This coordination of references adds to the impact and efficiency of their collective control efforts, because team members avoid wasting their energy by inadvertently pulling against each other.

In my next post, I intend to offer some additional examples of the more complex form of collective control in which participants control different perceptions at lower levels of perception in order to collectively control a single perception of a higher level.

KM: … A round of applause nicely illustrates the properties of a giant virtual controller. Every member of a large audience can contribute to the collectively produced sound by joining into the applause, but no single individual has more than a negligible effect on the overall intensity of the sound. An individual member of the audience who feels that the applause is too loud can cease clapping. If the applause does not seem loud enough, a person can clap more vigorously or even whistle and stomp feet. But in a large audience, none of the actions of a single individual will have any great impact on the collectively produced sound. The larger the audience, the less that any individual contribution will matter.

RM: The problem with this “theory first” approach to understanding social behavior is that the theory can become a Procrustean Bed into which the the phenomenon must be fit. In this case you have a theory of “collective control” where multiple control systems are acting to control the same variable relative to the same or different reference levels. You have found, via computer simulations, that even when the systems all have a different reference for this variable, it will remain in a virtual reference state, as though it were being controlled by a “virtual control system”. So now you have applied this theory to “a round of applause” under the assumption that each individual in the audience is controlling for the same variable – the loudness level produced by their own clapping and that of anyone else in the audience who claps.

RM: But based on my experience in audiences this seems somewhat unlikely to me. The loudness level of audience applause has typically seemed like an uncontrolled side effect of the ambient level of appreciation of the performance by individuals in the audience. There have been times when I (or others) have tried to increase the overall level of applause for a performance or speech. But this has usually been done by other means than increasing my level of applause. It has been done by calling out things like “let’s hear it for…”, hopefully taking advantage of the possibility that members of the audience are controlling for following such an order.

RM: But your explanation may be right. What is needed is empirical evidence that would provide as test of this theory. Evidence such as measures of variations in overall loudness levels compared to variations in individually produced loudness levels. You need to find data that would provide a test of your hypothesis; data that could best be fit my you model (of overall loudness as a common controlled variable) and mine (of loudness as a side effect of the individual levels of loudness each individual wants to produce).

RM: I think it makes more sense to take a phenomena first than a theory first approach to understanding social phenomena. Some of these phenomena (such as the tug of war example you give later) may be best explained in terms of agents controlling the same variable with respect to the same and/or different references. But others (like applause level or the social behaviors explained by the CROWD program) may be explained as side effects that emerge for groups of agents mainly controlling different variables (the agents in the CROWD program are all controlling the same types of variables - they are all controlling for “proximity” to other agents, for example – but each agent is controlling a different environmental correlate of each type of variable – for example, each agent is controlling proximity to a different agent).

RM: I think there is much more to social control phenomena than groups of agents controlling the same variable.



Thank you for your reply to my post, Rick. It looks to me like we may be getting nearer to a meeting of minds on this topic than we have in the past. It seems like a big deal to me that you’ve conceded that my tug of war example may demonstrate the simplest form of collective control. For a long time, you’ve been telling me that you couldn’t think of any empirical examples at all. But there are also some other areas of possible agreement.

First of all, I agree completely with the final statement of your post:

I absolutely agree with you that most empirical examples of collective control are far more complex than simply groups of agents controlling the same variable. I plan to delve into some of those complexities in my subsequent posts on the various forms of collective control.

Second, while I think I understand your views on how science should work, and I agree with your views in some respects, I do not completely agree with the way you’ve outlined it in your post. Let me start by trying to paraphrase your views on science, and please correct me if I’ve got it wrong.

You argue that we should take a phenomena-first approach to science. That means, if I understand it, that we should look at a phenomenon that we would like to explain, make careful observations of it, gather data, devise a theory to explain the phenomenon, and finally demonstrate that the theory is correct by constructing a computational model that replicates the data. This all makes for a nice neat linear progression. Emphasis on the word linear .

The thing is, that I don’t think either a phenomena-first or a theory-first approach to science is even possible. I think instead that phenomena and theories in science are completely intertwined. In my view, science—as a form of knowing about the world—and thus as a form of perception—is a cyclical process, where it’s no more possible to separate phenomena from theory from techniques of observation than it is to separate perceptions from references from actions in a control-system loop.

You describe me as having a theory-first approach, but you may not appreciate the extent to which my theoretical ideas on collective control have been rooted in a phenomena-first approach. I have been studying, teaching, and doing research in sociology for more than 50 years now, and thus have spent my entire career observing and thinking about the widespread social and cultural phenomena that I have always hoped be able to explain scientifically. When I finally encountered perceptual control theory—which was not until almost halfway through my career—it opened a way for me to begin to make sense of phenomena that I had long been observing.

You yourself have argued in your quite brilliant article on “Control-Theory Glasses” that making observations and gathering data don’t do you much good until you find the correct theory to explain what you’re seeing. This insight provides a nice explanation for the apparent futility of a century-plus of social scientific research that has sought to find the “causes” of human behavior rather than the perceptions people are controlling.

When it finally dawned on me that widespread social and cultural phenomena were not the results of linear causal processes but instead were evidence of control—and that the control had to be happening collectively on a massive scale—I began seeing collective control everywhere in the social and cultural world. Everywhere. All the society-wide and world-wide patterns of cooperative human action, as well as the large degree of uniformity in the manufactured physical environments in which we live and work, could not possibly have resulted from the random actions of individual people controlling their own perceptions without any reference to the behavior of others. Large-scale cooperation just doesn’t happen randomly, any more than it could happen that all the molecules in a gas might suddenly line up and begin moving at the same speed in the same direction.

The widespread social and cultural phenomena that interest me may not be things that you think much about or even find very interesting. That’s OK. You’ve been trained as an experimental psychologist, and that training teaches you to distrust anything you can’t confirm in your own lab. Unfortunately, the macro phenomena that interest me aren’t easily amenable to being brought into a laboratory. For me to go out and do the kind of research that you recommend in your post, I would need to have assembled a sizeable research organization: a team of graduate students ready to go out into the field, as well as a substantial research grant to support them. As a teacher at an undergraduate liberal arts college for most of my career, I just haven’t had access to those kinds of resources.

To put it another way, doing research on widespread social and cultural phenomena, or indeed any serious program of research, requires a collective effort and a division of labor. Within that division of labor, I see a place for someone who works mainly as a specialist in theory and can help others in the team to understand what they see.

Bill Powers, when you think of it, was essentially a theorist. Although he was able to do some experiments, the fact that he never had an army of graduate students at his command or any big research grants was of course one of the reasons that his theory hasn’t gained more attention in the academic world.

My goal in doing what I can to advance PCT with the resources at my disposal is to provide some theoretical ideas that may hopefully someday be tested out by others, much in the way you would recommend. Until then, I don’t feel any need to apologize for playing my limited role in the collective scientific enterprise.

My next planned post, which I began working on before composing this reply to you, will be a commentary and critique on Tom Bourbon’s article, “Invitation to the Dance.” I actually regard Tom’s experiment as a good demonstration of collective control as I would define it. To give a little preview of my post, I will argue that he carried out an elegant experiment, but that Tom’s analysis of his results left something to be desired, because the theoretical framework that he applied was inadequate to the task.

Until I’ve got that post put together, here’s a question for further thought (or a “leading question” as a friend of ours used to say). Do you see any way at all that the CROWD demos actually do serve to demonstrate the simplest form of collective control, that is, groups of agents controlling the same variables?

What does it mean for control agents to control the “same” variable?

Rick made an argument in his response to my original post that he had evidently not thought through carefully, or he could have easily seen that the argument doesn’t quite compute.

Rick argues in the final sentence of this paragraph that when control agents are controlling the same types of variables, we can tell whether control agents are controlling the same perceptual variable or different variables by examining the environmental correlate[s] of the variables they are controlling.

We can see that this argument doesn’t actually make sense by applying that same argument to perceptions controlled by a single control agent over time and asking whether controlled variables change every time the environmental correlates of the controlled perceptions change.

Consider the CROWD program that he refers to. The individual agents in that program control two types of proximity variables, as I understand it. One variable is to stay within a prescribed distance of a designated target agent (call it the leader). The other is to maintain at least a minimal separation distance from the other agents.

A typical CROWD demo shows a group of control agents following the same leader through a field of other agents all moving in different directions. Because of the constantly shifting configurations of agents in the field of action, the environmental correlates of the proximity variables controlled by the group of follower agents are in constant flux. But does that mean that the perceptual variables they control are constantly changing?

Well, obviously not. Whatever the environmental correlates of their perceptions might momentarily be, they can only to control the same two proximity variables that they were programmed to control. And, in fact, they are all controlling the same two variables, because the agents are all programmed to be identical, and the programmer has assigned the same references for controlling these two variables to each one of them.

We can expect the environmental correlates of controlled perceptions to constantly fluctuate because of unpredictable environmental disturbances. Control of a perception, of course, results in some reduction of the variability of a perception’s environmental correlates, but never a complete elimination of the variability. We are talking about variables after all, and they have variable values by definition.

What makes a controlled perception the same from one moment to the next, or from one episode of control to the next, or even from one control agent controlling a perception to another controlling the same perception, is the organization of control loop and the similarity in the references used for the control.

If the perceptions controlled by two control agents (or one control agent at different points in time) are constructed in the same way (or highly similar ways), and the references used by tthose agents for controlling those perceptions are the same (or highly similar), we can say that they are controlling the same perception. Otherwise it makes no sense to say that any perception is the same as any other, and we must then resign ourselves to living in a world of perpetual and chaotic flux. Social life would not be possible in that kind of world.

Now, it’s true that no living control agent can have exactly the same transient perception as any other at the same moment, because no two living control agents can occupy the same point in physical space at the same moment, and no two living control agents (at least, those with complex internal organizations) have perceptual hierarchies that are organized in exactly the same way,

But for the purposes of my argument, variations in the organizations of perceptual circuits and the values of the references used by different control agents make no difference in determining whether collective control of the same variable is taking place. If agents use different references for controlling their perceptions of a shared environmental variable, there will indeed be some conflict between agents, and the agents may escalate their outputs to try to get their perceptions into control. However, my simulations have demonstrated that their joint control actions will continue to have the same cumulative effect on the environmental correlates of these perceptions, so long as the agents continue to control the same perceptions they started with and use the same references they started with.

What matters for collective control, then, is the combined impact of the control agents’ actions on the shared environmental reality affected by the agents’ actions, even when the agents do not all perceive that shared environment in quite the same way.

Getting back to the social world of humans, no one of us has perceptions that are exactly the same as any other, but we can have perceptions of our shared environment that are similar enough, and we can control those perceptions with references that are similar enough, to allow enough cooperation for the practical purposes of our survival as a human group (so far, at least). And by cooperating with each other, we have gained the immense advantages of collective power over our shared environmental circumstances. That is the payoff for collective control.

KM: What does it mean for control agents to control the “same” variable?

RM: This is a great question. Two controlled variables could be the “same” because they are the same type of variable or because they are both the same type of variable and a function of the same environmental variables.

RM: In the context of your model of collective control, where a virtual controlled variable emerges from two or more agents controlling the same variable relative to different references, the meaning of “same” must be the latter: the virtually controlled variable is the same type of variable and a function of the same environmental variables.

RM: The agents in a tug of war are controlling the same type of variable – the position of the flag relative to the dividing line between the two teams; call it a relationship type perception – that is also a function of the same environmental variables – the flag and line. Since it is physically impossible to have the flag in two different positions relative to the line at the same time, it is impossible for the two different teams to get their perceptions of the position of the flag relative to the line in two different states at the same time. So there is a conflict.

RM: The agents in the CROWD program are also controlling the same type of variable – their proximity to other agents and objects; also a relationship type perception – but each agent’s perception of proximity is a function of different environmental variables – the different agents and objects that they happen to be close to as they move through a scenario. Because the agents are controlling different perceptions that are functions of different environmental variables, there is no conflict.

RM: The variables controlled by the agents in the CROWD program are not virtually controlled; they are really controlled. The stable patterns that emerge in the various CROWD scenarios – such as the semi-circle of agents around a “guru” – are a side effect of each agent’s successful control of the variables it is controlling. The stable patterns are not a result of conflict between the agents acting like a “giant virtual controller” of these variables.

RM: This is why I think your “giant virtual controller” model of collective control doesn’t apply to social behavior like that seen in the CROWD program. In fact, I can’t think of any example of social behavior, except a tug of war and arm wrestling, where there is a stable result (even for a relatively short time) that results from conflict over the state of the same controlled variable – one that is the same type of variable and a function of the same environmental variables.

RM: When I refer to variables that are the same, I usually mean “same” in the second sense: same type and function of same environmental variables. When I am just referring to variables that are the same only in terms of type, then I will always try to say “same type” of variable. But I think what is meant by “same variable” is usually pretty clear from context.

I think Kent is mixing up the variable(s) being controlled with the value(s) of the variable(s) being controlled. The values are in flux, the environmental correlate(s) of the variable(s) is/are not. The value(s) of the environmental correlate(s) change. That’s what control is all about, keeping those values near their reference values. Control, collective or otherwise, isn’t about changing the variables being controlled. That’s something reorganization might do, but control does not.

In general, collective control does not mean that multiple controllers try to simultaneously control the same variable. It does mean that whatever variables the members of the Giant Virtual Controllers might be controlling, the environmental effects of controlling their own variables all influence a correlated common variable that might not be THE variable controlled by any of them. That’s perhaps a side-effect from the viewpoint of each individual controller, but it is not a side-effect for the Giant Virtual Controller.

A side-effect from the viewpoint of the GVC is something that happens unperceived by the controller as a consequence of the environmental effects of the controller’s output. A good example is the Tragedy of the Commons in which several controllers actions deplete the resource on which they all depend, but none perceives changing as a direct consequence of their own actions. Climate change, for example, where everybody says that their choice of car has no effect, when their choice does have an effect that is too small and too delayed for tem to perceive.

Seems to me that what makes it the same from one moment to the next, or from one episode of control to the next, is the perceptual input function. (Every PIF categorizes over a domain of effective input, and not all of its input need come from the environment.) And what makes it the same from one agent to another is a perception that another agent is controlling a particular variable. Although that variable is necessarily among one’s own perceptual variables, the agent takes it to be an actuality in the environment. This is the basis of intersubjective agreement about aspects of the environment.

Cooperation and competition are not the only phenomena that depend upon a perception that another agent is controlling a particular variable. Attribution of such perceptions (and control) to others is important for survival; I hope it’s not necessary to argue that point. Mutual perceptions of e.g. predator and prey, conspecific, and kin go back pretty far in the evolutionary tree. As Rick observed a few years ago, so-called ‘mirror cells’ are probably reference signals that fire, controlling in imagination “what I would be controlling if I were doing what I see that guy doing”. Of course we can’t just dismissively say “they’re only reference signals” without inquiring into why reference signals fire upon observing the activity of another agent.

Some aspects of the environment may serve in the environmental feedback path for control. More than one agent may use the same aspect of the environment in the environmental feedback path for control, even for control of different perceptual variables. Where environmental feedback paths intersect might be called a ‘crossroads’ or ‘intersection’ aspect of the environment. Under social conditions in which these collective phenomena occur, the stability of a common intersection of environmental feedback paths may itself become a controlled perception. This is how our built environment has come to supplant the natural environment. As Kent says, everything in the built environment is collectively controlled.


    February 17

BN: Seems to me that what makes it [the controlled variable] the same from one moment to the next is the perceptual input function.

RM: As I said in an earlier post that apparently no one read, that is one possible meaning of “same controlled variable”; when the perceptual input functions in two control systems are the same, the systems are controlling the same type of variable. But if the environmental inputs to these functions are not the same they are not controlling the same variable, just the same type of variable.

RM: For example, I could be controlling for opening a door and you could be controlling for closing a door. We are both controlling the same type of variable (angle of door relative to frame). But if we are dealing with two different doors then we are not controlling the same variable. So even though we have two different references for the state of the same type of variable (my reference being something like 80 degrees and yours being 0 degrees) there is no conflict and no “virtual” controlled variable. But if we are controlling the angle of the same door, then we are controlling both the same variable AND the same type of variable: the angle being the angle of that particular door relative to the door frame. In that case there will be conflict and, depending on our relative strength and stamina, the angle of the door will remain in a virtual reference state for a while, until one or both of us realizes that there are better things to do.

RM: Kent’s model of collective control as being the virtual reference states of controlled variables requires that the collective (group of control systems) be controlling not just the same type of variable but also the same variable. I think this limits the range of social phenomena to which the model applies to a very small number. I can think of many examples of social phenomena where a collectively produced result is a result of individuals controlling the same type of variables but not the same variables.



Comment on Tom Bourbon’s “Invitation to the Dance”

Just 30 years ago, Tom Bourbon published a paper called “Invitation to the Dance” [Bourbon, W. T. (1990). Invitation to the dance: Explaining the variance when control systems interact. American Behavioral Scientist , 34 (1), 95-105.]. This paper presented the first experimental evidence of collective control, as I would describe it, although Tom offered a much different interpretation of his results.

The paper begins with a concise description, well worth quoting, of how to tell from empirical evidence that control has occurred.

  • The most certain evidence of control is the elimination of variability in part of the world at the expense of increased variability in the actions of the person, or any other organism, that is in control. (p. 95)

Tom’s warning, that control efforts to create order in one place in our environment will inevitably produce side effects of disorder someplace else, sounds a bit like the Second Law of Thermodynamics applied to control theory. In the PCT world we may easily lose sight of that immutable fact.

Tom’s paper describes an elegant series of experiments designed to demonstrate how we can “coordinate the movements of parts of our own bodies, or our own movements and those of another person.” The paper thus offers a PCT-based theory of how people interact with each other, and it represents the earliest attempt I know of to apply PCT to social psychology. Unfortunately, in my view, his theoretical interpretation of his results was incomplete and thus misleading, and his arguments may actually have set back the acceptance of PCT among social scientists.

In the setup for his experiment, Tom attached two joystick handles to a personal computer (this was before the days of mice). He then programed the computer display to show a cursor in the middle of the computer screen that served as a fixed target for two cursors programed to wander vertically up and down the screen on either side of the target cursor in response to random disturbances.

He connected the handle of the joystick on the right to the cursor on the right, and the handle of the joystick on the left to the cursor on the left, which created an experimental setup that was a kind of double-barreled version of the classic tracking experiments done by Bill Powers. But there was a twist.

The movements of the handle on the right were programmed to affect not only the position of the cursor on the right, but also the position of the left-hand cursor, only just half as much. Similarly, the left-hand handle affected both cursors, but its effect on the right-hand cursor was only half as big as on the left.

In the first run of the experiment, one subject was given left-hand joystick and told to keep the left-hand cursor even with the target in the middle of the screen. The subject was easily able to perform this classic tracking task. By moving the left-hand handle to control the position of that cursor, the subject, of course, disturbed the position of the right-hand cursor, but it didn’t really matter, since the subject was only concerned with controlling (his perception of) the position of the left-hand cursor, and not the right.

In the second run of the experiment (with the same random disturbances), the subject was given both handles and asked to keep both cursors even with the target. The subject performed this more complex task successfully, in spite of the fact that the motions of the cursor on the left added to the disturbances on the cursor on the right, and vice versa.

In a third run with same set of random disturbances, two subjects participated, one using the handle on the left to control the cursor on the left, and the other using the handle on the right to control the cursor on the right. Again, the subjects had no difficulty performing these tasks.

Tom then presents graphs showing the handle positions, cursor positions, and disturbance vectors for these runs, which of course resemble the classic graphs for PCT tracking experiments. Tom points out in his writeup that the graphs for runs number two and three—where the task of keeping both cursors even with the target was performed first by a single person and then by two people— look almost identical. Of course, this result was not really surprising, since the handle motions that served to keep the cursors even with the target were the same in both cases.

Tom goes on in his paper to present a computational simulation of the results of third run of the experiment, where two subjects performed the task. Using control-system models (PCT bots, you might call them) with parameters tuned to simulate the performance of the two subjects on their previous tracking runs, Tom shows that the PCT models can do an excellent job of simulating the actual experimental results.

So far, so good. Tom’s evidence shows convincingly that, whether it’s the two hands of a single individual controlling the cursor positions, or two individuals, or two PCT bots, the control agents must resist disturbances in the exactly same way in order to keep the left and right cursors aligned with the target. Furthermore, he has shown that all of those combinations of control agents are equally able to accomplish the task.

In a concluding section of the paper, where Tom offers a “general discussion” of the results of the experiments, his interpretation goes off track. Tom begins in a reasonable way by discounting the prevalent psychological theory that the brain uses perceptual inputs to calculate in advance the movements necessary to keep the cursors even with the target, and then commands the hands and arms to move accordingly. As he argues, “commands computed in advance cannot anticipate the effects of unpredictable continuous disturbances like those used here” (p.103).

The problems with his explanation begin In the next two paragraphs, where he offers a critique of the theories of social scientists:

  • To explain coordination between the actions of two people, many social theorists invoke hypothetical goals that are allegedly “shared” by the people. Others hypothesize norms, standards, or guidelines that exist independently of people and subsume them in larger systems (families, armies, and the like), where each person functions as a component. But people are complete control systems, not isolated components such as input functions or comparators. Typically, theories that describe people as “components” in larger control systems cannot predict continuous coordinated actions.

  • In the simulations here, the only connections between the models were in the environment. When, as was the case here, two systems affect the same variables but to differing degrees, each system reduces the interaction to a net disturbance and then cancels the effect of that disturbance. The result? In a complex world, simple systems achieve coordinated control. (p. 103)

Tom’s theorizing is off base, not because his PCT-based arguments are incorrect, but because his PCT explanation is incomplete. He hasn’t thought through the experimental situation carefully enough to see all of the ways PCT might apply to it. Tom’s problem here is that his theorizing is strictly one-dimensional, but the experimental situation he set up involves two different levels of perception at the same time, and the social situations to which he wants to apply his results always involve perceptions at more than one level.

In the discussions 25 years ago on CSGnet about whether to use the acronym PCT for the theory that Bill Powers had presented in B:CP, Bill said his preferred label was HPCT, or hierarchical perceptual control theory, to emphasize the hierarchical nature of human perception and action, the fact that that humans control perceptions on many different perceptual levels simultaneously. In the interpretations he offers in this paper, however, Tom has completely ignored the hierarchical aspect of PCT.

It isn’t as if Tom is unaware of HPCT. In the concluding section of his paper, he makes this argument:

  • With its elegant simplicity and effectiveness, the control theory model stands in sharp contrast to thecomplexity and nonspecificity of most theories of coordination. Control theory offers the possibility of using the same few principles to explain coordination at every level, from movements of parts of our own bodies to interations [sic] like those in simple tracking tasks, in infant-parent dyads, in social gatherings, in marriages, and on the job. (p. 104)

Tom is arguing here that interactions occurring at any perceptual level can be adequately explained by the one-level PCT model that he offers in this paper. However, if we look carefully at his model, we can see how his model is even too simplistic for the relatively simple experimental situation he set up.

A one-level model of a control agent controlling its perception of a single environmental variable works fine for describing behavior in classic tracking experiments. But in Tom’s experimental setup there are three perceptual variables: the relationship of the left-hand cursor to the target, the relationship of the right-hand cursor to the target, and the simultaneous relationship of both cursors to the target.

This third perception is of higher order than the other two, because this perception is constructed hierarchically by combining the other two into a single more complex configuration. Tom completely ignores this third, higher-level variable in his diagrams of the experimental setup and in his simulation modelling. All the same, this phantom variable is what gives his paper its rhetorical punch.

The subject who uses two hands to do the experiment must be able to control this more complex variable, the perception of keeping both cursors even with the target simultaneously. Control of this variable is implicit in the task that Tom has presented to this subject, and the subject’s successful completion of the task is proof that the subject was able to perceive and control this higher-level perception, as well as the lower-level components of the perception, that each cursor separately must stay on target.

In the third run of the experiment, with two different subjects at the computer, both subjects presumably looked at the same computer screen, but as long as they focused their attention on the cursor that corresponded to their own handle, they could complete the assigned task without any problem. They didn’t have to pay attention to the third, more complex variable, and it might have confused them if they had.

In fact, another way for Tom to have set up this third phase of the experiment would have been to give each subject a separate computer with a monitor that showed only the cursor corresponding to their own handle and the target, as long as there was a connection between the computers that then transferred from one computer to the other the disturbances created by the movements of the other handles. The experimental results would have been exactly the same. It this case it was Tom who actually controlled the third variable, the one that he described as coordination of their movements, by the way he set up the experiment and his instructions to the subjects to do their tasks simultaneously.

In the computer simulation of the results, Tom was again the one who controlled the third variable, the one that combined the two lower-level perceptions of each cursor staying on target. The PCT bots in the simulation acted like Tom’s slaves, since he specified the perceptions they needed to control and also specified the references to use in controlling those perceptions. As one-dimensional control systems, the bots obviously couldn’t have comprehended the concept of keeping both cursors on target simultaneously, but it wasn’t any problem. Tom kept this higher-level variable in control, since he was the one who could monitor the coordinated movements of the cursors to make sure that the experiment worked as planned. If anything had gone wrong, Tom could have fixed the problem by reprogramming the computer.

But take note. For Tom’s article to be rhetorically effective—for readers of the article to buy his argument—his readers, too, must be able to control (in their imagination) this higher-level variable based on the combination of the other two variables. They have to be able to imagine the sight of the two cursors staying on target simultaneously. In fact, for his argument to be persuasive, Tom and his readers must be able to share this perception. In my terms, Tom and his readers have to be able to control this higher-level variable collectively.

My conclusions:

  1. In the second run of Tom’s experiment, where a single subject takes a handle in each hand, the subject is actually controlling three perceptions, not two: the perception of the right-hand cursor staying on target, the perception of the left-hand cursor staying on target, and the perception of both cursors staying on target at the same time, a higher-level perception constructed from the other two. Subjects are able to complete this complex task because the hierarchical organization of the perceptual control systems in their brain allows them to control different perceptions simultaneously at more than one level of perception. According to HPCT, the hierarchical organization of the brain implies that control of a perception at a high level depends on the simultaneous control of the cascade of lower perceptions from which it has been constructed.
  2. In the third run of the experiment, where two subjects were at the computer, each of the two subjects had to control two perceptions (at least): the perception of keeping their own cursor on the target, and the higher-level perception of cooperating with the experimenter by following Tom’s directions to complete the task. This is an example of my second general form of collective control: where two or more people cooperate by collectively controlling a single focal perception at a given perceptual level while independently controlling different perceptions at lower levels.
  3. For Tom’s argument to his readers to be persuasive, he and his readers must engage in my third general form of collective control: parallel independent control of “the same” perception or set of perceptions by different people who share a set of common references for their own independent control of those perceptions. In this case, Tom and his readers, who may be far separated in location and even across time, must be able to control similar perceptions what it means for two cursors to stay on target at the same time, in order for the readers to accept the argument that Tom’s model explains how “coordination of movements” takes place.

Some additional reflections: Tom argues, in effect, that coordination of movements might somehow spontaneously arise between two control systems controlling only their own independent perceptions without reference to each other. To me, this argument seems as silly as imagining that the molecules of a gas might all spontaneously start to go in the same direction at the same speed.

Social coordination does not, and cannot, happen by accident, and Tom’s argument that it could happen that way would seem absurd to other social scientists, too. No wonder that Tom’s paper was not taken as a great conceptual breakthrough by members of the social science community. Tom may have been arguing against the notion that society is a control system at the macro level and that people are simply components of that overall control system, but that theory was something advanced by some people interested in PCT in the 1990’s, not anything taken seriously by the social scientists of my acquaintance.

From the social scientific perspective, cooperation in social situations happens when people intend to cooperate. In PCT terms, we would say that the parties control a high-level perception of cooperation with the other parties involved, in addition to any lower-level perceptions they control to play their role in the overall cooperation. Which, in my book, represents the second general form of collective control.

Furthermore, our communication between each other hinges on the third main form of collective control. For communication to take place—for you as reader, for instance, to understand the arguments I am rendering symbolically by putting words on a computer screen, or for readers of Tom’s paper to have understood his arguments—the parties to the communication must be able to share similar perceptions, or in other words, to control those perceptions collectively. Now, of course, each person’s perceptions are unique, so the sharing cannot be perfect. Communication is never perfect, but in our semi-successful social life it has ordinarily proved good enough for practical purposes.

Thus, you and I communicate when your perceptions of these words as reader and my perceptions as author are similar enough that we don’t end up taking physical actions in control of our own perceptions that prevent the other person from continuing to control their independent but similar perceptions. As my simulations of collective control have shown, our perceptions need not be exactly the same for our joint actions to stabilize our shared physical environment more effectively than either of us could do on our own, and thus to enhance both party’s independent control of the relevant perceptions.

To sum up, my argument is that one-dimensional analyses of social life, like Tom’s analysis in his paper, are by definition insufficient and that one form of collective control or another is involved in everything that we do as social animals.

KM: It looks to me like we’re making some progress toward a meeting of minds here. As far as I can tell, the hangup that prevents you from fully understanding my theory of collective control is that the kinds of variables that you tend to think of (when you make your favorite argument, “I can only think of …”) are physical variables, or in PCT terms, variables at the lowest levels of the perceptual hierarchy. It seems to me that you, just like Tom Bourbon in his “Dance” article (see my comment below), have made the mistake of trying to analyze social situations in terms of one-dimensional models. But the theoretical analysis of social situations, in my view, always requires multilevel models involving the simultaneous control of low-level physical variables and high-level social variables based on combinations of those low-level variables.

KM: Collective control always takes place in a shared physical environment, when the actions of multiple people all affect the same set of physical variables. But the variables people are controlling may be social variables based upon those physical variables. If people collectively control the same high-level social variable, and their similar references for that social variable mean that their references for the physical variables contributing to this social variable are similar, their collective control of the social variable will have a stabilizing effect on the contributing physical variables, Alternatively, if they have different references for the social variable, which then implies different references for the contributing physical variables, the collective control will involve conflict. I know this may sound pretty complicated, but you need to think about these examples on at least two different levels of perception at the same time. The simplistic, one-dimensional theory that you have been apparently been using so far is inadequate to the task.

KM: Here’s an example of collective control that actually involves the physical variable you refer to in your post, the angle of door relative to frame, but involves at the same time a much higher-level social variable, and the variable being collectively controlled is the social one.

KM: My daughter’s husband is a German and they live in Germany. Sometimes, when my wife and I have visited our daughter’s family, we have stayed at the home of her husband’s father, who lives nearby. His house is a lovely house in the German style, with lots of dark wood paneling. But it tends to be dark inside, because all of the doors between rooms are kept closed all the time unless someone is going from one room to another.

KM: By keeping doors closed between rooms, my son-in-law’s father is keeping control, as I understand it, of his perception of a German cultural rule that the doors between rooms should always stay closed. I became aware of this rule because I am an American who does not have the same cultural reference for whether doors should be open or closed. As an American, I prefer houses to have open doors, lots of interior light, and a free flow from room to room. However, when I went from one room to another in his house and left the door open, my host would come in, give me a disapproving look, and ostentatiously close the door. When I asked my daughter why he was upset, she explained to me that this was just how the Germans did things. They like doors to be closed.

KM: My host and I were both controlling a social (cultural) perception about whether the interior doors of houses should be open or closed, and when my reference for controlling that perception did not match his, the result was (mild and transitory) conflict between us about the angle of a particular door. After I had stayed in his house once or twice and had become aware of the house rule about closing doors, I began trying to conform to the German rule by making an effort to remember to keep closing doors behind me. As a result of his and my collective control of this rule by means of similar references, the shared physical environment of the house was also stabilized to some extent, in that the doors remained closed for more of the time than when we were in conflict about the rule.

KM: Here is a similar and related example of cultural conflict involving physical variables and social variables simultaneously. At the house where we stayed, I was sometimes surprised on entering a room to find the windows thrown open as wide as possible. As an American, my cultural expectation is that when the weather is cold and damp (which it is most of the time in northern Germany, where my daughter and her family live), windows are to be kept closed, or only opened a little bit for ventilation. The German cultural rule, as my daughter again explained it to me, is that when a room starts to seem stuffy, all the windows should be opened wide to give the room a thorough airing (and then closed again before the people in the room start to shiver). Here, again, the control of physical variables about whether windows are open or closed depends on the collective control of a higher-level cultural variable about how interior environments of houses should best be maintained.

KM: I’m interested in what you have in mind when you say …

KM: I wonder if a more careful analysis of those examples would show that by the “same type” of variable you are really referring to a higher-level social variable that people collectively control, which then provides references for controlling their perceptions of the lower-level physical variables they are controlling independently. How about sharing a few specific examples to show whether or not that is the case?

My Best,


RM: I wish I could say the same.

KM: As far as I can tell, the hangup that prevents you from fully understanding my theory of collective control is that the kinds of variables that you tend to think of (when you make your favorite argument, “I can only think of …”) are physical variables, or in PCT terms, variables at the lowest levels of the perceptual hierarchy.

RM: Nope.

KM: It seems to me that you, just like Tom Bourbon in his “Dance” article (see my comment below), have made the mistake of trying to analyze social situations in terms of one-dimensional models.

RM: What’s a one-dimensional model?

KM: But the theoretical analysis of social situations, in my view, always requires multilevel models involving the simultaneous control of low-level physical variables and high-level social variables based on combinations of those low-level variables.

RM: Ah, you mean one-level, not one-dimensional, models. And I agree that complete models of social behavior will involve control of a hierarchy of perceptions in each participant. But the lack of inclusion of systems controlling higher level variables in Tom’s models of two-person interaction doesn’t invalidate those models any more than the lack of inclusion of systems controlling higher level variables in models of individual control invalidates those models. Tom’s models don’t purport to show how the participants in the control task ended up controlling for their cursors being aligned with the target; he just shows that when they do have this common goal they “automatically” produce this cooperative result by compensating for all disturbances to it, including those produced by each other.

RM: But why criticize Tom for having a one-level model of a social interaction when your is a one-level model as well. As I understand it, each agent in your “collective control” model is a one-level control system controlling the same variable as all the other agents. Each agent is controlling this variable relative to a different reference specification so the variable remains in a virtual reference state in the sense that it is protected from external disturbance. I have always understood that this virtual controlled variable can be “simple” , such as the temperature of the room, or complex, such as the official political position of all the agents in the group. Now that I think of it, I have just given examples of two situations where you model of “collective control” might actually apply! What would be nice now is if you could collect some data to show that virtual reference states actually exist in this situations.

RM: So I think I have discovered, finally, the “real world” situations to which your “collective control” model might actually apply. It applied to situations where groups have to get a variable that affects everyone in the group to a level that is “satisfactory” (minimal error) to each individual in the group.

RM: So do you know of any data that might be relevant to this!




The Coronavirus Pandemic as a Collectively Controlled Perception: Examples of the other two main forms of collective control

The current coronavirus crisis offers some striking examples of collective control on a macro scale. Let me start with the third main form of collective control, as I previously described it:

Last November, the word coronavirus would have meant nothing to most of world’s population. Of course, specialists in viral diseases would have known the term, and the word might also have been familiar to people directly affected by previous pandemic outbreaks caused by coronaviruses, like SARS or MERS. But the bulk of the world’s population, if asked what coronavirus is, would have drawn a blank. The first reports of the disease now known as COVID-19 began appearing last December, when the disease had not yet been given an official name by medical researchers. The concept of social distancing did not come into common parlance until March.

Today, these words are familiar to nearly everyone in many countries worldwide. In the language of PCT, these labels refer to perceptions that people around the world (as I write this in May) have begun actively controlling, perceptions that so loom large among their daily concerns that large numbers of people have given up the control of many of the other routine perceptions of their daily lives in order to focus on controlling their perceptions of the grave danger posed by this pandemic.

The rapid spread of coronavirus and COVID-19 around the world is a phenomenon too vast and multi-faceted for any individual to comprehend it in its entirety. But almost everyone shares the same reference for controlling their perceptions of the phenomenon: the agreed-upon reference for how much pandemic we want to perceive is zero. With so many people worldwide using similar references to control their perceptions of the “same” phenomenon, the coronavirus pandemic offers an excellent example of “parallel independent collective control,” “widely dispersed in time and location.”

The first thing to understand about our collective control of perceptions of the pandemic is that although no single individual’s efforts could possibly be sufficient to bring their perceptions of the pandemic into control, every individual’s efforts to control their own perceptions of it will have some slight impact on the development of the overall phenomenon. Every effort by an individual to protect against the virus (or not)—whether an individual avoids crowds, stays home, washes hands frequently, refrains from face-touching, practices social distancing in public, wears a mask or gloves or both in public, gets tested for the virus or its antibodies, seeks medical attention for possible symptoms of the disease, or does none of these things—will have some effect on the rapidity with which the virus spreads from person to person. Thus, the spread of the virus is under the control of a giant virtual controller emerging from the combined control actions of myriad individuals.

A second important point is that people’s perceptions of the pandemic, even though broadly similar, are not exactly the same. It’s a PCT truism that if you can’t perceive it, you can’t control it. But a phenomenon like a pandemic is so huge and complex that no single individual could hope to perceive every aspect of it. Our individual perceptions are necessarily limited, most obviously by the limitation that our physical bodies, with the body’s sense receptors for forming perceptions, can occupy only a single position in physical space at any one time. As a result of these differences in our micro-environments, each of us experiences the pandemic from a different perspective, and thus our perceptions are all different.

Of course, we can rely on other people’s accounts to form perceptions of things too complex for us to take in on our own, but these second-hand perceptions also have their limitations due to our unique locations in social space, that is, the limited range of our social contacts and the limited range of media sources we consult for stories and images about the pandemic. Although many people in many places have had firsthand experiences with one aspect or another of the overall phenomenon, many others (like me so far, I’m happy to say) have had no first-hand experience with the virus and COVID-19 (at least none that we’re aware of), and have formed our perceptions of the pandemic entirely from second-hand sources.

In short, the differences in our personal experiences of the same overall phenomenon— the worldwide spread of the viral disease—mean that each of us perceives it differently. How we then control our perceptions of it depends to a large extent on which aspects of the phenomenon we perceive to be the problem. People are limited in the range of their perceptual abilities—the unique repertoires of control systems available in their neural hierarchies— and these limitations allow them to perceive only certain aspects of this vast, complex, and many-stranded phenomenon. It is, of course, possible to enhance the range of one’s perceptions by making use of specialized perceptual tools to form a more detailed and comprehensive picture of the overall phenomenon, but many of us lack access to those tools, and the unequal distribution of perceptual tools is another limitation on what many individuals can perceive and thus try to control.

The specialized perceptual tools for understanding the spread of the virus tend to be more available to experts than to ordinary people. This pandemic results from the spread of a biological agent far too tiny to be detected by our unaided sensory organs. With scientific tools, like microscopes and gene-sequencing devices, experts can test for the presence or absence of the virus and its antibodies. But when tests for the coronavirus have been in short supply, as has been true in the United States and many other countries, it has hampered everyone’s ability to get a comprehensive picture of the virus’s spread, and has prevented government authorities from determining who is carrying the virus in order to isolate them and the other people they’ve been in contact with.

Statistical data and models provide another important set of perceptual tools for monitoring widely distributed phenomena, like the spread of a virus. However, many people have not had the education to understand these conceptual tools, and thus they lack access to perceptions that could enable more effective action against the pandemic. It’s particularly a problem when government officials, like the American President, for instance, don’t understand how to interpret statistical findings. Unless they prove more willing to rely on the advice of experts in statistical modeling than President Trump has been, their control efforts will miss the mark, as we have seen in the ineffectiveness of the American response compared to other countries. Thus, the ways in which our various individual perceptions focus on different aspects of the overall phenomenon will limit our collective ability to control it.

Whatever perceptions people form of this vast and complex phenomenon, their control efforts will also reflect the other perceptions they regard as important to control. If an extremely complex phenomenon offers many different possible options for trying to control it, no one of which is obviously the best choice, people are likely to select the control activities that are most consistent with other important perceptions that they try to control, even if those action options may not be the most effective for controlling the thing they’re worried about.

For example, President Trump’s go-to strategy for combatting the virus has been to institute travel bans and seal the borders against foreigners, in line with the anti-immigrant agenda he has pursued throughout his presidency. Conservative Republican governors of states in the US, including the governor of my own state of Iowa, have taken the coronavirus crisis as a convenient opportunity for shutting down abortion clinics. Autocrats around the world have taken advantage of the crisis to grab increased powers of surveillance and control of the public behavior of their subjects. And ordinary people for whom the whole thing is just too scary to think about have been controlling their perceptions of the pandemic by pretending that it doesn’t exist or is no more of a threat than seasonal flu.

When people have high gain (strong motivation) for getting something under control, but an imperfect perception of what the problem is, and whenever there is a significant time lag between the beginning of control efforts and when they actually take effect, people’s control efforts tend to be unstable and erratic. We see a pattern of overshoots followed by undershoots, and vice versa, as people first don’t do enough and then do too much and then pull back and don’t do enough again and can’t seem to find a response calibrated to be just enough to keep the thing in control. Again, the behavior of President Trump has taken this erratic course, as he first claimed that the virus was nothing to worry about and that everything was under control, then predicted hundreds of thousands of deaths and called on states to clamp down on social contact, then began urging governors to relax controls immediately and reopen the national economy.

With people perceiving the pandemic differently and then pursuing different strategies for controlling it, conflictive collective control is the inevitable result. Conflict is practically unavoidable when various people try to control their own differing perceptions of a phenomenon big and complex enough that nobody can control it on their own. Conflict arises as one person’s efforts to control their own perceptions of the problem interfere with and cancel out other people’s actions, and the result of such conflicting and ineffective control efforts is that the phenomenon gets stuck in some in-between state that isn’t what anybody wanted, as has happened in the US when the competition between states and countries for scarce resources like personal protective equipment has left some localities with surpluses and others with dire shortages, while the spread of the virus continues. In general, the more people involved in trying to control something, the less leverage any single individual has to bring things into line with their own preferences. And unless most people can get on the same page about what the problem is and what needs to be done, control efforts are often inefficient and counter-productive.

The only way this kind of mess is with cooperative collective control, with people combining their perceptions of a complex phenomenon in order to crowd-source a better picture of what is going on, and then optimizing and coordinating their control efforts to focus on the types of action that have proven to be most effective. Which brings me to the second major form of collective control outlined in my previous posts, “when people work together by doing different things toward the achievement of a shared goal.” Here is my more technical definition from my earlier post:

Governments, when functioning efficiently, also work on this principle, and the central function of governments is to bring stability to the living and working environments shared by the citizens of a locality, which is why the role of governments has been so important to our collective attempts to control the perceived effects of the massive disturbances caused by the spread of the virus. Other organizations, companies, hospitals, nonprofit service providers, etc., have also had important roles in combatting the pandemic, and while all of them exhibit the features of this type of collective control, none shares the governmental function of providing overall stability.

The hierarchical structure of organizations mirrors to some extent the hierarchical structure of the human perceptual hierarchy. With governments, for example, the politicians at the apex of the organizational hierarchy articulate the common goals of the organization. In governments, these goals take the form of laws and policies that serve, in effect, as high-level reference signals for the perceptual control actions of the lower layers of bureaucrats, aides, and front-line workers who carry out the policies and enforce the laws at the local level. Workers in the middle-managerial levels of a government bureaucracy translate the policies issued by their superiors into detailed regulations or plans for action by front-line workers. Workers in these middle levels of the bureaucracy also send reports back up the hierarchy that provide summaries of the perceived effectiveness of the actions of the front-line workers in meeting the overall goals. Thus, the downward flow of references for action and upward flow of perceptual feedback in these organizations somewhat resembles the perceptual control circuits of the human neural hierarchy.

Of course, the fact that the nodes in the layers of an organizational network are human beings, with all their internal complexity, instead of elementary control-system circuits, means that the analogy between organizational hierarchies and neural hierarchies is far from exact. In particular, it’s probably a mistake to think of organizations as social-level control systems, even if their objective is to carry out control functions collectively. All of the potential problems that I discussed above in regard to dispersed collective control also apply to this type of collective control, where people are coordinating their various control efforts to reach a common goal. The people in the various layers of the organization are likely to differ in their perceptions of what the goal is, as well as in their priorities and capabilities for taking action. Thus, conflict between members of the organization is an ever-present possibility in this type of collective control, as well.

I could write a great deal more about what can go wrong with coordinated collective control, and we are seeing object lessons in such failures here in the United States, as the country overall struggles to cope with the pandemic. But this has probably been enough to make the general points I wanted to make.

Good luck and good health to everyone, as we collectively seek to control the effects of the virus!

Rick is talking about variables controlled by different systems being “the same”, but in the bit that he quoted I was talking about the perceptual input to a single system being “the same from one moment to the next, or from one episode of control to the next”.

I went on to talk about “what makes it the same from one agent to another”, proposing that this “same variable” perception

I’m sorry you feel that no one reads what you post, Rick. The feeling is sometimes mutual. Does that make that feeling a similar perception, or the same perception? :slight_smile:

Kent, this is an excellent and important re-assessment of Tom’s pioneering work. The hierarchical context of control cannot be neglected. Bill often omitted to mention explicitly e.g. compliance with the experimenter among the controlled variables specified in an experiment, but I doubt he ever forgot it.

If alignment of cursor with target is a relationship perception, then this seems to be a perception of relationship between relationships. Rick and maybe Tom I think would argue that this is a perception in an observer, but that in the subject controlling both cursors simultaneously this is an emergent result from controlling each cursor independently–i.e. that the dependency is in the environmental feedback function, not in the subject. This would be exactly parallel to the emergence of rings and arcs and other phenomena in the CROWD demo. What data can show which analysis is correct? We know that the subject is instructed to control that perception of both cursors relative to the target, and the subject is demonstrably compliant, that is, has adopted the prescribed reference value for the described perception. We do not argue that taking a sip of tea is an epiphenomenon emergent from the lower-level perceptual control that brings it about, because we do not deny the higher-level perception and its reference. Is there something different about this case that authorizes us to deny the higher-level perception and reference?

Is the alignment of cursor to target a relationship? Then this higher-level perception appears to be a relationship between relationships.

The subject could as easily control the situation so that the left-hand cursor is above the mark and the right-hand cursor is below it by a certain distance. The subject could control a configuration such that cursor-l, mark, and cursor-r form a diagonal at an angle of 30 degrees. Reflecting on this awakens the observation that in the actual case the subject might be controlling a configuration such that cursor-l, mark, and cursor-r form a horizontal line. Whatever the facts for a given subject (and this might be testable), reviewing these possibilities may help a reader to appreciate with more immediacy what that higher-level perception might be, and maybe even open ways to demonstrate experimentally that it is not just the observer’s perception based on the experimental instruction, it is the subject’s perception based on the experimental instruction, and a perception ‘shared’ by subject, experimenter, observer, and reader.

Hi Kent

KM: The current coronavirus crisis offers some striking examples of collective control on a macro scale.

RM: I just saw this data in the NY Times that seems pertinent to your post:

RM: The graphs show changes in measures of economic behavior that are related to the social distancing that was eventually mandated by most states to deal with the pandemic. What is interesting about it to me is that people were beginning to social distance – as indicated by decreases in spending, time spent at work, number of small business open and hours worked in small businesses – 10 to 20 days before they were officially told to do so (to stay at home). Indeed, the overall data shows that people started social distancing around mid-march. The graphs show that whether states “closed” (required social distancing via stay at home orders) early or late or not at all, social distancing reached an asymptote at around April 1. So apparently people were social distancing as much as they were going to before it was required by the state. So why did the voluntary social distancing start in mid-March? I think it’s because it was around that time that the reported number of cases in the US was approaching 10,000 and deaths had exceeded 100. So my guess is that at mid-March people started to become afraid of the virus. So people started doing social distancing – and staying in as much as possible – to control for not getting the virus.

RM: Of course, people (or at least some) could also have been social distancing in order to control a perception of the pandemic, as you suggest. But to the extent that they were, they would have been very disappointed in their efforts. During the period covered by the graphs, the pandemic, in terms of number of cases and deaths, was getting exponentially worse. So if one were social distancing as a means of controlling the pandemic they would have been experiencing considerable error. This would be expected to result in reorganization. This reorganization would would be seen random increases and decreases in the amount of social distancing as the reorganization process looked for alternative ways to reduce the pandemic. And there is evidence that this occurs. If you look at the last graph, showing the measures of social distancing for states that never ordered closing, these measures start to vary considerably once they approach asymptote. This variability is large relative to that seen for the states where social distancing was mandated; with social distancing mandated, people who wanted to cooperate – even if they were experiencing error due to failure to see a reduction in the pandemic with social distancing) would continue social distancing anyway.

RM: Of course, there are other possible explanations for this variability. One likely alternative is conflict; doing social distancing for whatever reason – to avoid getting the virus and/or to control the pandemic – is likely to conflict with other goals, such as going out and being with other people. And there is evidence that people were in conflict – between wanting to cooperate by social distancing and wanting to make a living. But whatever the reason for this variability, the fact that it doesn’t show up in states that mandated social distancing suggests that people – most people – are controlling for doing what they are told by an authority.

RM: I/'m presenting this little analysis because there is going to be a ton of social behavior data available from this pandemic and I want to show that it should possible to use this data as the basis for testing models of how cooperation works (and doesn’t work) in human societies.



It’s a reasonable assumption that most people control positive values of these variables

  • Consumer spending is a way of controlling possession of needed goods, such as food.
  • Hours worked, time spent at work, and small businesses open (for those who operate a small business) are all ways of obtaining money as means of consumer spending.

Many variables could be identified here. Spending is also means of controlling social standing, public image, self image, which I expect that the business operator controls with higher gain than the employee/consumer, and employees control insurance coverage with higher gain, but set those variables aside for simplicity.

For the sake of a handy abbreviation, call these the public desiderata.

It’s also reasonable to assume that people avoid contracting a scary and highly contageous disease, and the main way of controlling this (the only way, absent testing and contact tracing, and absent a vaccine or cure) is to avoid other people who might unknowingly be carrying and spreading the virus. Control of this variable conflicts with control of the above desiderata. (The belligerent ‘open it up’ folks are controlling this by denying that it is real.)

We could look for changes in the environment coincident with the fluctuations in control of the identified desiderata. What made that big downward spike around April 12 or 13? And why did consumer spending spike upward soon after (but not the other desiderata)? The data appear to show that the start of a closure policy is not a variable that tips the balance in this tug of war. It would be important to know when the governor of each state announced the intention to close on the specified date. Relevant to this is the conflict between variables that the governor is controlling, commonly expressed metaphorically as ‘public pressure’ on the governor to issue a closing statement, in conflict with ‘pressure’ from business interests to defer or weaken closure. Communications that are identified as measures of this ‘pressure’ might be quantifiable as data. There may even be polling data.

For conservatives, measures to control avoidance of COVID-19 infection conflict with their preferred political policies. They reduce this conflict by denying the reality of the pandemic.

May collective control present subjectively as something like command hallucinations in Julian Jaynes’s bicameral mind theory?

Hi Gabriel,

It’s been years, decades really, since I read Julian Jaynes book. It struck me as interesting and possibly significant at the time, but this was before I became familiar with PCT, and I don’t remember much now about the “bicameral mind” theory.

Can you define “command hallucinations” for those of us who aren’t up to speed on the Jaynes book and describe in more detail what you see as the connection between his work and collective control?

Off the top of my head, it seems to me that in thinking about this it might be important to remember that the effects of collective control are entirely mediated by the physical environment. Collective control happens when two or more people try to control similar perceptions regarding the same piece of environmental real estate. If their references for controlling those perceptions are similar, the effects of their actions on the variable in the physical envinronment are additive. If their references are different, their actions interfere with each other. But in either case, the action all takes place in the physical environment.

That’s why I’m initially dubious when you talk about hallucinations. But maybe you’re talking instead about some kind of collective control that occurs within a person’s perceptual hierarchy (by the two halves of the bicameral mind, perhaps?) Please explain a little further. Thanks!


There’s a good summary of bicameralism in Wikipedia.

“a controversial hypothesis in psychology and neuroscience which argues that the human mind once operated in a state in which cognitive functions were divided between one part of the brain which appears to be “speaking”, and a second part which listens and obeys—a bicameral mind, and that the evolutionary breakdown of this division gave rise to consciousness in humans.”

Jaynes proposed that humans were this way as recently as the bronze age, ca. 3000 YBP, and that’s why Homeric characters heard the voices of the gods.

It’s difficult to square with Jill Bolte Taylor’s lived experience.