Sankara opens his commentary on the Brahma-sutras not with the sutras themselves but with a short preamble - the Adhyasa-bhashya - that stops to analyze the cognitive event on which the entire system will depend. Adhyasa: superimposition. The act of taking the properties of one thing to be the properties of another, and producing a cognition in which the two have fused.
The classical example is the rope and the snake. In dim light you see a rope on the ground and take it for a snake. The cognition has a real referent - the rope - and a wrongly-imposed form - the snake. This much it shares with viparyaya (Y3). Sankara goes further. The superimposition is a pre-reflective construction the cognitive apparatus performs and presents to consciousness as straightforwardly given. By the time you notice the snake, the superimposition has already happened.
The structure has three components. The adhisthana - the locus, the actually-present thing on which the superimposition lands (the rope). The aropita - the imposed form, the properties laid on the locus (the snake-shape, the threat, the coil). And the cognition itself, which presents the composite as a unified perception. Each component can be analyzed separately - and the analysis matters because the everyday vocabulary of “mistake” treats the whole event as a single undifferentiated thing.
The corrective is not effort or attention; it is the disciplined separation of the locus from what has been imposed on it. When the lamp is brought near, the rope shows itself as a rope; the snake-cognition doesn’t persist as an opinion to argue against - it dissolves, because the imposed form no longer has anything to rest on.
Where English Falls Short
“Projection” in English psychology is the nearest term and is too narrow. Freudian projection is the attribution of one’s disowned content to another - one specific motivated case. Adhyasa names the structural cognitive event of which projection is one application. The Sanskrit category ranges across cognitions that have no motivational structure - the rope-snake case is not motivated by anything; it is just how the perceptual apparatus assembled the cognition under those conditions.
“Cognitive bias” describes the surface effect - the systematic distortion - without naming the underlying structure. Adhyasa names the structure: properties being laid on a locus that does not intrinsically have them. Biases are surface taxonomies. Adhyasa is one level deeper, asking what cognitive event each bias is a specification of.
“Confabulation” in neuropsychology has structural overlap. A patient with certain memory deficits constructs a coherent account that is mostly false but presented as straightforward recall. The Vedanta frame would analyze the confabulation as adhyasa: the properties of a plausible biography have been imposed on the locus of the patient’s actual experience. The medical literature describes the phenomenon; the Vedanta analysis names its cognitive form.
Where it Shows Up
Anthropomorphism of AI systems. Users routinely impose the properties of a person - intention, care, understanding, agency - on the locus of a language model. The cognition presents the composite (“the AI is being helpful”) as straightforwardly given. The structural diagnosis is adhyasa. The corrective is the disciplined separation: what is the locus actually doing, and what properties have been imposed on it that do not belong to it?
Treating screens as people. A face on a screen, a voice on a phone, a chat window - each invites the superimposition of “person” onto a locus that is something else (an image, an audio stream, a text protocol). The composite cognition is operationally useful much of the time and structurally misleading when it is not.
Pareidolia. The face on Mars, the man in the moon, the cat in the cloud. A real locus (Martian surface, lunar topography, cloud-shape) and an imposed form (face, figure, animal). The neurological literature explains the pattern-completion machinery; adhyasa names the cognitive event the machinery produces.
Product evaluation by surface cues. A product is evaluated by qualities imposed on it from packaging, branding, or peripheral signals. The user is cognizing the composite - “this product is premium” - without being able to separate the locus (the actual product) from the aropita (the premium-signalling cues). Marketers exploit the structure; the corrective is the anvaya-vyatireka discipline (V3): ask what actually varies with what.
Cognition of one’s own emotional state. Anger about one thing is imposed on the locus of an unrelated current situation and presented as anger at that situation. The composite is invisible in the cognition. The contemplative version of the adhyasa analysis - developed in Yoga and Vedanta - is the cultivation of the capacity to see the locus separately from what has been imposed on it.
Model interpretation in ML. An interpretability tool produces a visualization; the researcher imposes the properties of “explanation” on the visualization and reads the model as having reasons. The structural question is whether the explanation-properties belong to the locus (the actual model computation) or have been imposed on it by the visualization framework. Many interpretability disputes are disputes about exactly where the adhyasa line falls.
Diagnostic Question
“What is the actual locus here, what properties have been imposed on it, and would I still see the composite if I separated the two?”
IKS Roots
The Sanskrit term is adhyasa (अध्यास), from adhi (upon) + as (to throw, to place) - literally “placing-upon.” Sankara (8th century CE) opens his Brahma-sutra-bhashya with the Adhyasa-bhashya, a short preamble that defines and analyzes the term before the commentary begins. His definition: smriti-rupah paratra purva-drshta-avabhasah - adhyasa is the appearance, in one place, of something previously seen elsewhere, in the form of memory. The structure: a thing previously cognized (the snake, the person, the premium product) appears in the cognition of a different locus (the rope, the screen, the actual product). The related term adhyaropa (imposing-upon) is used interchangeably in much of the Advaita literature. Sankara’s analysis builds on the Yoga treatment of viparyaya (Y3) and the Nyaya analyses of error. The Advaita position is the most systematic working-out of error-as-superimposition in classical Indian philosophy. The metaphysical extension - that the everyday self is adhyasa on atman - is the doctrinal use Sankara puts the structural analysis to; the structural analysis stands independent of the metaphysical conclusion.
See also Y3 (viparyaya - Yoga’s analysis of the same cognitive event in psychological register), Y4 (avidya - the master mis-taking that drives adhyasa), and V2 (sat / asat / anirvachaniya - the ontological status of objects cognized through adhyasa).
Further Reading
Sankara’s Adhyasa-bhashya, available in many translations of the Brahma-sutra-bhashya (Swami Gambhirananda’s translation from Advaita Ashrama is standard). Eliot Deutsch, Advaita Vedanta: A Philosophical Reconstruction, for an accessible philosophical entry. Bimal Krishna Matilal, Perception, for the wider Indian theory of error. Karl Potter, Advaita Vedanta up to Sankara and his Pupils, for the technical literature on the doctrine.