Control Theory and Self-Reproduction

Abstract

Control theory is a branch of technology. A control task addresses an object to be controlled, some set of controls with at least two possible states, and some limitations on both the system to be controlled and the controller. The designer of the controller introduces his intentions through goal functions, or selection principles. Goal functions are usually expressed numerically, whereas selection principles can be simpler and qualitative, such as “on” or “off.” In either case, the designer intends to optimize the performance of the system under control.

Notions of optimality are very anthropomorphic. In applying control theory to biological processes, severe difficulties arise. Examples are given of the nonnumerical nature of skeletal muscle control, of self-reproduction, and of gene expression. In each case, the biological process has characteristics foreign to the domain of technological control theory. These characteristics are highlighted, and it is concluded that the images and concepts of technological control theory fail to take note of important features of biocontrol systems. The features that lie outside the range of modern control theory include the use of both positive and negative feedback relations between the information in a memory and the product of the readout operation at the genome; the increased possibility for richness of control that arises from mapping from a symmetrical memory structure (DNA) to an asymmetric catalytic structure (an enzyme); the merging of structure with function in the microscopic domain; and the acquisition of information by interaction with the environment. Furthermore, biocontrol is of the nonnumerical type. When all these features are considered together, one must conclude that modern control theory so far has had little to offer biologists. —The Editor