What does "expect" mean, particularly in vision? For Minsky, expectation is the content of filled-in nodes of frame-arrays (the defaults) -- but qualifies this as an oversimplification.
NB: What does expect mean -- a neat question! One wonders the impact of this question on the debate of the cognitive impenetrability of vision. To what extent might top-down visual effects be due to wired in mechanisms, and thus only seem "cognitive"?