A quantitative approach, to figural "goodness".

Simplicity: a unifying principle in cognitive science?

“…We argue that the best approach to (1) is to choose patterns that provide the simplest explanation of the data; and that this approach provides a powerful in-road to (2), in line with a long tradition of psychological research.…”

“…First, Shannon's information theory justified log 2 (1/p) as a code length for items with probability p. This is helpful for providing code lengths of highly repetitive data patterns, which can be assigned probabilities, such as low-level perceptual properties, phonemes, words and so on [2]. Second, the critical generalization to algorithmic information theory by Kolmogorov, Solomonoff and Chaitin defined the complexity KðxÞ of any object, x, by the length of the shortest program for x in any standard (universal) computer programming language [3].…”

“…The main theoretical alternative, the Bayesian approach to visual perception [12] is mathematically closely related to the simplicity principle [13]. [2,11]. † The simplicity of a code for a stimulus quantifies the amount of structure uncovered in that stimulus.…”

Competing global and local completions in visual occlusion.

“…Because global and local completions of the pattern diverge into different shapes, the pattern is referred to as a GLD pattern (global-localdivergent). (Hochberg & McAlister, 1953) conjoined with a descriptive system of visual patterns (cf. Hochberg & McAlister, 1953;Leeuwenberg, 1969Leeuwenberg, , 1971Restle, 1982) results in clear predictions of the preferred pattern interpretations.…”

Perceptual economy and the impression of visual depth

“…An alternative approach which might serve both as a theory of perception and of "misperception" is similar to that first explored by Hochberg and McAlister (1953) and Attneave (1954), and more recently by Garner (1962Garner ( , 1966Garner ( , 1970. This suggests that stimulus information is perceived as invariant because this is a more economical way of encoding the sensory data.…”