Memory in Perception, Not as Perception: A Reply to Barkasi

This paper critically examines Barkasi’s, (2021) claim that perceptual experience is “temporally thick,” shaped not only by current sensory input but also by the phenomenal contribution of long term memory. Barkasi defends two theses: that reactivated mnemonic traces partly constitute perceptual phenomenology (Past in Perception), and that memory should therefore be treated as a sensory modality (Memory as Sensory Modality). I argue that this dual proposal overextends the empirical evidence by conflating memory’s causal influence with phenomenal contribution, and by treating subpersonal mechanisms as phenomenally manifest. Moreover, elevating memory to the status of a sensory modality risks collapsing important distinctions between perception, imagination, and recollection. I develop a moderate alternative grounded in predictive processing models, according to which long term memory functions analogously to Bayesian priors: It shapes perceptual inference without supplying distinct phenomenal content. This framework preserves Barkasi’s empirical insight while maintaining the unity and present directedness of perceptual experience.