7 - Information in Internal States

Summary

The final, critical step in developing the information-and-evolution framework is to understand how selection drives informational tracking of world-states by the internal states of neurological control systems such as human learning. The bacterial navigation model allowed us to understand the essential fitness feedback loop – the way it results in acquired traits tracking environmental changes and the way it effectively directs open-ended creativity toward the solution of evolutionary problems. The shortcoming of that model was that the acquired traits involved were implausible analogs for human beliefs, in that nothing the least bit like memory was involved. If the upper level is interpreted as consisting of beliefs, then the all-important interaction with the world via resulting behavior was left out of the model. In this chapter, we add a third, intermediate level for preference formation to the model, which will bias variation in a random walk behavior pattern. We again model a system much simpler than humans – bumblebees – but the important thing here is the principle of information transfer involved. Again, the less we assume about complex structures, the better off we are in terms of the epistemological challenge. If information transfer depends on certain structures, then more complex systems which have those structures as well should be able to exploit information in the same manner.

THE MODEL: REAL'S BUMBLEBEES

For a number of years, behavioral ecologist Leslie A. Real (1991, 1992) has been studying the foraging behavior of bumblebees in enclosed environments.