Psychologists have shown a continuing interest in the influence of knowledge on perception. The constructivist view of Helmholtz receives its modern expression in the work of Hochberg (1978) and Neisser (1967), who suggest that knowledge can affect the analysis of a complex stimulus. An expected stimulus, for example, is one whose occurrence is predicted by virtue of one’s knowledge of the world; it need be analyzed only enough to confirm this expectation. The same stimulus will be analyzed in greater detail in contexts where it is unexpected. Perception is seen as a form of hypothesis testing, in which hypotheses generated on the basis of knowledge and experience are tested against sensory information sampled from the world. (Seidenberg et al., 1982)

Вот если бы он сказал, что гипотезы не проверяются относительно информации внешнего мира, а защищаются от нее, было бы гораздо круче =)