Psychological and Cognitive Science Approaches to Consciousness
Third Lecture (3/5)
Part One and Part Two
For most of the twentieth century, academic psychology treated consciousness as an embarrassing ghost left over from introspective philosophy. Behaviourism banished it, cognitivism replaced it with information-processing metaphors, and consciousness survived only as a synonym for verbal reportability. The cognitive revolution of the 1980s and 1990s changed everything. A series of simple, repeatable experiments demonstrated that attention and consciousness are neither identical nor independent, that perception without attention is rich and structured, that conscious access is late, serial, and capacity-limited, and that the intuitive transparency of our own experience is largely illusory. Psychology was forced to re-admit consciousness, not as a mystical residue but as a natural phenomenon amenable to the full arsenal of experimental methods: masking, binocular rivalry, inattentional blindness, change blindness, the attentional blink, continuous flash suppression, partial report, dual-task interference, and, increasingly, first-person training protocols and neurophenomenological correlation.
The foundational discovery was that attention and consciousness dissociate in both directions. Koch and Tsuchiya (2007, 2012) showed that a wide variety of features—colour, motion, orientation, and semantic category—can be processed to high levels outside focal attention and outside awareness. Inattentional blindness (Mack & Rock, 1998; Simons & Chabris, 1999) and change blindness (Rensink, O’Regan, Clark, 1997–2023) revealed that observers routinely fail to notice large changes or salient objects when attention is elsewhere, yet eye-tracking and priming studies prove that the unnoticed stimuli were registered unconsciously. The partial-report paradigm (Sperling, 1960; revisited with modern masking by Bronfman et al., 2019) demonstrated iconic memory: subjects briefly exposed to a 4×3 letter array report seeing all twelve letters yet can only name three or four; something it is like to see the entire array even though most items never become reportable. This “overflow” phenomenon remains the central battleground between global workspace theories (which identify consciousness with access) and recurrent-processing or higher-order theories (which allow phenomenal consciousness without broadcast).
Global workspace theory in its psychological guise (Baars, 1988; Dehaene, 2014; Dehaene et al., 2017, 2023) has become the dominant functional model. Consciousness serves to mobilise otherwise encapsulated modules—face recognition, language, motor planning, emotion—by broadcasting a selected content to a coalition of prefrontal, parietal, and temporal areas. Reportability is the gold standard because verbal report is the most flexible and cross-modal output system. The theory predicts, and experiments confirm, that conscious contents are those that survive the ignition threshold, remain stable for hundreds of milliseconds, and dominate working memory. Unconscious contents are fast, parallel, and automatically decay unless attention stabilises them.
Yet the overflow argument (Block, 2007, 2011, 2023) refuses to die. If subjects genuinely experience more specificity in the Sperling array than they can report, then phenomenal consciousness exceeds access consciousness. Block’s case rests on phenomenology plus indirect measures (ensemble perception, subjective inflation paradigms) and has been strengthened by recent machine-learning decoding studies showing that unseen items in crowded displays leave decodable traces in early visual cortex even when subjects deny seeing anything beyond a vague gist. The debate has become extraordinarily technical, turning on the interpretation of confidence ratings, the validity of gist perception, and the possibility of intermediate states between full consciousness and complete unconsciousness.
Higher-order theories (Rosenthal, 2005; Lau, 2008; Brown et al., 2019, 2024) offer a different solution: a mental state is conscious only when the subject has a higher-order representation (not necessarily linguistic) that they are in that state. This elegantly explains why some thoughts feel conscious and others do not, why dreaming feels conscious despite reduced metacognition, and why transcranial magnetic stimulation of prefrontal cortex can abolish consciousness of a stimulus while leaving first-order processing intact. The price is the familiar regress worry: does the higher-order state require a yet-higher-order state to be conscious? Rosenthal bites the bullet and claims higher-order thoughts can be unconscious yet consciousness-making, a move that many find counter-intuitive but logically coherent.
Recurrent processing theory (Lamme, 2006, 2010, 2023; Block & Lamme, 2022) locates consciousness in local recurrence within sensory areas, especially visual cortex. Feedforward sweeps are unconscious; sustained recurrent activity between higher and lower areas (V4 ↔ V1, MT ↔ V1) is sufficient for phenomenal consciousness even without prefrontal involvement. The theory predicts conscious overflow and receives support from decapitated visual cortex preparations in monkeys and from human intracranial recordings showing sustained local gamma in posterior electrodes long before the P3b wave reaches frontal sites. The most recent versions integrate predictive-coding ideas: recurrent processing is the settling of a perceptual inference that minimises prediction error at intermediate hierarchical levels.
Illusionism (Dennett, 1991; Frankish, 2016; Pereboom, 2022; Kammerer, 2024) is the most radical psychological stance: phenomenal consciousness as ordinarily conceived does not exist. Introspection systematically misrepresents functional, reportable states as private, ineffable, intrinsically qualitative qualia. The hard problem is not hard; it is an artefact of a broken user-interface. Empirical support comes from the success of predicting reportability without positing private phenomenal properties, from change-blindness and masking studies showing that we are routinely wrong about what we experience, and from the apparent continuity between conscious and unconscious processing. Strong illusionists predict that once the neural mechanisms of judgment and report are fully mapped, the residual mystery will evaporate. Critics reply that illusionism denies the one datum no theory can coherently deny: that there is something it is like to be us right now.
Micro-phenomenology and neurophenomenology (Varela, 1996; Petitmengin, 2006; Depraz, 2021) have re-legitimised trained first-person investigation. Experienced meditators can reliably report pre-reflective shifts—changes in the felt location of awareness, the onset of a thought, the micro-structure of an emotion—that untrained subjects miss entirely. These reports correlate with distinct neural signatures: increased gamma synchrony, reduced alpha suppression, enhanced phase-amplitude coupling between theta and gamma. Long-term practitioners show decreased default-mode activity during present-moment awareness and increased insula activation during interoception. The field has moved from suspicion to mainstream respectability, with major laboratories now incorporating first-person training into consciousness protocols.
The psychedelic renaissance since 2018 has been nothing short of revolutionary. Psilocybin, LSD, DMT, and 5-MeO-DMT dramatically increase brain entropy, disrupt default-mode integrity, weaken top-down priors, and produce states of ego dissolution, synesthesia, and ontological rewiring. The REBUS model (Carhart-Harris & Friston, 2019; Singleton et al., 2024) explains these effects as a temporary relaxation of high-level beliefs, allowing bottom-up information to flood consciousness. Perturbational complexity (PCI) and repertoire size increase beyond normal waking levels, and the subjective intensity is often described as “more real than real.” Crucially, these states are accompanied by lasting reductions in depression, anxiety, and addiction, suggesting that consciousness is not only malleable but therapeutically malleable.
Hypnotic suggestion provides the cleanest psychological dissociation. Highly hypnotisable individuals (top 10–15 %) can abolish pain, create vivid hallucinations, erase memory, fracture identity, and reverse agency with nothing more than words. Neural correlates are now well mapped: suggested analgesia suppresses anterior cingulate and insula activation despite intact nociceptive input; suggested hallucinations recruit the same sensory areas as veridical perception; post-hypnotic amnesia reversibly disconnects prefrontal-hippocampal loops. Hypnosis demonstrates that the final contents of consciousness are under direct linguistic control, confirming predictive-coding models and undermining any theory that treats phenomenal properties as intrinsic and inaccessible.
The enactive and sensorimotor approaches (O’Regan & Noë, 2001; Engel et al., 2016; Di Paolo et al., 2023) reject representationalism entirely. Consciousness is not a picture in the head but the structure of the organism–environment coupling. Mastery of sensorimotor contingencies—the lawful ways that sensory input changes with movement—is what it is like to see red or feel a surface. Deafferentation, prism adaptation, and sensory substitution devices provide striking support: when contingencies are disrupted, qualia change or vanish. The approach dissolves the hard problem by redefining experience as action rather than passive contemplation.
Finally, metacognition and confidence research (Fleming, 2023) have revealed that our sense of conscious clarity is itself a constructed judgment, often poorly calibrated. Subjects can be highly confident of having seen nothing when decoding shows rich unconscious processing, or uncertain about stimuli that were consciously perceived. The machinery of confidence appears to rely on prefrontal and parietal regions distinct from the machinery of first-order perception, explaining why disorders of consciousness, blindsight, and hypnotic suggestion can selectively impair metacognitive accuracy while leaving phenomenal experience intact (or vice versa).
After four decades of rigorous experimentation, cognitive science has established beyond doubt that consciousness is late, limited, editable, and deeply intertwined with attention, memory, self-modeling, and language. It has also shown that our folk-psychological intuitions about consciousness are systematically mistaken: we overestimate its capacity, its stability, its transparency, and its necessity for sophisticated cognition. Yet the central mystery remains: why does any of this feel like anything from the inside? Psychology can map the functional architecture with ever-increasing precision, but the leap from function to phenomenology continues to elude third-person science. The field has not solved the hard problem; it has made it sharper, more embarrassing, and, paradoxically, more fascinating than ever.
Part Two: Consciousness Across Waking and Dreaming
Tanmoy Bhattacharyya
