The article deals with J. J. Gibson’s account of Paintings as arrested images, inadequate for the study of the principles of natural vision. Accepting most of Gibson's insights, the author discovers some incongruities of his account which lead Gibson to treat paintings as being less informative than an ambient optic array. Stressing the fact that a chirographic picture is an intentionally produced visual pattern, an extract of visusal information, an informational constant, which is invariant in every respect, the author reaches the conclusion that the complexity of pictorial codes may neither be considered as an evidence supporting the thesis about pictorial invariants being weaker and formless, nor does it prove the study of paintings to be irrelevant for the study of perception.
