Collective control can result either in conflict or cooperation between the participating control agents, depending on the degree of alignment in the reference values they use in controlling their perceptions of an environmental variable.
When the control agents use identical reference values to control their perceptions of a shared environmental variable, the result is cooperative. My computational simulations of collective control have shown that the environmental variable is stabilized to the same extent as if some super-powerful control agent were tightly controlling its perception of the variable by applying to the control process an amount of loop gain equal to the sum of loop gains of the all the participating parties.
If, however, the participants in a collective control process disagree in their choice of reference values for controlling their perceptions of the environmental variable, conflict is the inevitable result. As Bill Powers often explained (see, for instance, the chapter on conflict and control in Behavior: The Control of Perception ), conflict occurs whenever two or more control systems attempt to control their perceptions of the same environmental variable by using different reference values.
The visible sign of conflict is that the participants begin working against each other, pulling in opposite directions, as it were, each one trying to bring the value of their own perception of the environmental variable to their preferred reference point. Although it’s easiest to envision conflict as occurring between two controllers, each trying to control their perception of a single environmental variable, conflict occurs whenever more than N controllers try to control an environmental variable of N dimensions. See below for additional discussion of these complex forms of collective control.
An important but paradoxical finding from my computational simulations is that collective control with conflict can stabilize an environmental variable just as effectively as happens with cooperative collective control, in which the reference values of the participants are perfectly aligned. Thus, the joint actions of control agents in conflict over the perceived state of an environmental variable can produce environmental stability that seems as firmly set as if a super-agent were tightly controlling its perception of the variable with a loop gain equal to the sum of the contributions of all the participating control agents.
In practical situations in which humans are involved, the stability created by collective control with conflict may be less long lasting than is true of cooperative collective control, because the people caught in the conflict may not like the experience, and so may try to escape it. Nevertheless, I could offer many examples of long-lasting “frozen” conflicts, where people have continued to engage in deadlocked conflicts over years or even decades.
Although collective control with conflict will often produce a relatively stable outcome, the outcome is not necessarily what any of the participants would prefer. Powers described the value at which the environmental variable is stabilized as a “virtual reference level” that represents a compromise among all the participants’ differing preferences. Mathematically, the virtual reference level will be equal to the weighted average of all the participants’ reference values, weighted by the loop gain that each control agent contributes to the joint control effort.