ABSTRACT
Most psychophysical models are empirically testable only when expressed as a probability space or class of probability spaces, that is, as elementary events and their associated probabilities or probability densities. Within a given experimental milieu, the model attempts to reproduce the probability spaces defined by observable events (or some aspects of the space) by postulating certain unobservable events that “explain” or “give rise to” the behavior. To the extent that the fit between data and model spaces is deemed acceptable, the model is judged to be worthy of further interest and research.
