A case is made for the following claims. Higher animals have evolved the ability to make predictions because it increases their chance of survival. The basic steps in the prediction process are - registration of the internal spatial model with the world, initialization of salient objects (including actors) within the world model, projection of actor behaviors and movements through time, and assessment of results against drives and goals. Predictions are compared with situation estimates, and world models and previous estimates are revised based on this discrepancy. Better world models and situation estimates, in turn, lead to better initialization for the next prediction. This bootstrapping cycle improves both prediction and situation estimation. Thus, prediction can act as an evolutionary driver for more complex world models, higher-level behaviors, and greater sensory discrimination. Higher-level behaviors, with built-in contingencies, are time-consuming to build and resistant to change. However, a self-adaptive system must be able to change its higherlevel behaviors. When prediction is applied to one's own thinking a biofeedback loop is set up that can improve higher-level behaviors. It is observed that many meditation techniques are metacognitive processes that seek to setup this type of biofeedback loop.
展开▼
