Beyond the Sixth Consciousness: A Yogācāra Diagnosis of Intelligence, Wisdom, and the Structural Limits of Artificial Cognition

This paper deploys the Yogācāra Buddhist eight-consciousness model (aṣṭa-vijñāna), as reconstructed from the Cheng weishi lun (T31, no. 1585) and Vasubandhu’s Triṃśikā (T31, no. 1586), to diagnose the cognitive-functional status of contemporary transformer-based AI systems, especially large language models. The analysis is structural-functional rather than phenomenological: Yogācāra categories are employed as a comparative philosophical vocabulary for distinguishing cognitive kinds, not as phenomenological attributions or ontological identity claims. The paper argues that current large language models achieve what it terms sixth-consciousness virtuosity — a highly developed but bounded extension of manovijñāna’s capacities in discrimination, inference, retrieval, and conceptual generation — while lacking the structural resources of the seventh consciousness (kliṣṭa-manas: pre-conceptual self-referential orientation) and the eighth (ālayavijñāna: dynamic dispositional substrate). Four converging arguments establish a structural gap between intelligence and wisdom (prajñā): the absence of pre-conceptual presentation, the absence of genuine metacognitive reflexivity (svasaṃvedana), the structural unavailability of liberatory āśraya-parāvṛtti, and the absence of existential grounding in duḥkha and karuṇā. The paper further distinguishes structural anātman (never having had a self) from realised anātman (the transformative dissolution of self-grasping), arguing that AI’s structural selflessness is not evidence of wisdom but precisely the absence of its preconditions.