Defamiliarization operates through friction. The work resists the perceiver's attempts to assimilate it into existing categories. The resistance slows perception, extends attention, and forces the mind to construct rather than retrieve. The transformation that results is not informational — the perceiver has not acquired new facts — but structural: her capacity for seeing has been altered, and the alteration operates in every domain of subsequent experience.
Greene treated defamiliarization as the essential educational operation, more important than the transmission of content. A student who has been genuinely provoked by Morrison's Beloved has undergone a small unsettling of her assumptive world. She has seen, if briefly, that her own perspective is one position among many, that the reality she inhabits is constructed rather than natural, that alternatives exist which her habitual categories had rendered invisible.
The AI tools pose a specific challenge to defamiliarization. Large language models generate outputs that regress toward the statistical center of gravity of their training data — toward the already-familiar forms that define the corpus. The outputs can be technically accomplished, structurally conventional, and fluently readable. What they rarely are is strange. They do not resist the perceiver's categories; they confirm them. The encounter with the AI-generated artifact is more likely to be an efficient retrieval than a defamiliarizing transformation.
This is not a claim that AI cannot contribute to defamiliarization. A skilled user can direct the tool toward unfamiliar connections, toward outputs that cross domains in ways that disrupt habitual perception. Koestler's bisociation describes the operation. But the disruption depends on the user's wide-awakeness, not on the tool's automatic operation. The default of the tool is the default of the corpus — and the default of the corpus is the already-familiar.
Shklovsky introduced the term in the 1917 essay 'Art as Technique,' published in Theory of Prose. Greene absorbed it through her study of Russian formalism and made it the mechanism behind her theory of aesthetic education, developed most fully in her work with the Lincoln Center Institute.
Perception as end. Art exists to prolong and intensify perception, not to convey information efficiently.
Resistance as method. The work defamiliarizes by refusing to be easily assimilated — its difficulty is its gift.
Educational core. Greene treated defamiliarization as the essential educational operation, more important than content transmission.
AI challenge. Generative models regress toward statistical familiarity; their default mode is the opposite of defamiliarizing.
Transferable capacity. The perceptual shift produced by aesthetic encounter operates beyond art, reshaping how the perceiver meets every subsequent experience.