The Experimental Arche Continued: Von Foerster on Observing Systems

In the previous chapter we used Ashby’s cybernetic theory to discuss the “experimental arche” of organizations. This arche referred to a continuous and risky process of control, design and operational regulation with respect to organizational transformation processes. At the heart of our discussion of the experimental arche was Ashby’s regulatory logic, stating that, in order to regulate a particular concrete system, one has to: Select essential variables and desired values Identify parameters, disturbing the essential variables Design an infrastructure (a “mechanism”) by means of which: Disturbances are attenuated The system’s transformation processes can be realized Regulatory potential (regulatory parameters) becomes available And, given 1, 2, and 3: select values of regulatory parameters (= select regulatory actions) in the face of actual disturbances. Moreover, in this Ashby-based notion of regulation, one needs a model of the behavior of the concrete system: a transformation. According to Ashby (1958), a good (conditional, single-valued) transformation relates the selected variables and parameters in such a way that predictions can be made about the behavior of the concrete system. To arrive at such a transformation, the black-box method was introduced – a method enabling a regulator to derive a transformation based only on the values of the variables and parameters that are chosen to describe the concrete system that should be regulated. Ashby’s black box method seems to suggest that we can “objectively” select variables and parameters, and derive a transformation connecting them based on trial and error, without, as Ashby puts it, “reference to prior knowledge”. If this is what regulating systems is about, one might say that it does not contain much risk. It is “just” a matter of selecting variables/parameters; observation and deduction. The risk attached to it may have to do with the mistakes we make in selecting variables/parameters or in deducing a conditional transformation from empirical observations; or it may have to do with time-constraints we face while regulating; or with the probabilities appearing in a transformation and governing the behavior of the system.