What Is the Higher-Order Theory of Consciousness?

Direct Answer

Higher-order theories of consciousness are a family of views unified by a single idea: a mental state is conscious when it is the object of another mental state — a higher-order state that represents it. On this view, the difference between a conscious and an unconscious mental state is not in the first-order state itself but in whether it is represented by a second-order state. You consciously see the red apple because you have a mental state that is about your visual experience of the apple — a mental state that represents you as seeing red.

The theory starts from a striking observation: most of our mental states are not conscious. Subliminal perceptions, implicit biases, habitual routines, and the countless computations of perception all occur without awareness. What, then, distinguishes the conscious ones? For the higher-order theorist, the answer is relational: a state is conscious if and only if a higher-order state represents it. The first-order state provides the content — redness, the apple — while the higher-order state provides the consciousness, the awareness of having that content.

Three versions of the theory must be distinguished. Higher-order perception (HOP) theories, associated with David Armstrong and William Lycan, model the higher-order state on perception: an "inner sense" scans our first-order mental states, and consciousness arises from this inner scanning. Higher-order thought (HOT) theories, the most influential version, defended by David Rosenthal, treat the higher-order state as a thought: a state is conscious when accompanied by the thought that one is in that state — roughly, when one has the thought "I am seeing red." Self-representational theories (Kriegel, Gennaro, and others) hold that a state is conscious when it represents its own occurrence: the state represents both the world and itself, and consciousness is self-representation.

The differences matter: HOP theories locate consciousness in a quasi-perceptual mechanism, HOT theories in concept possession and thought, and self-representational theories in the structure of the state itself. But all agree on the core claim — consciousness is a matter of representation, not of intrinsic phenomenal properties.

Historical Context

The idea that consciousness involves awareness of our own mental states has old roots. John Locke defined consciousness as "the perception of what passes in a man's own mind," explicitly conceiving of it as a kind of inner perception. Kant similarly held that the "I think" must be able to accompany all my representations — a higher-order unity of apperception. The early moderns thus already gestured at the relational structure that higher-order theories formalize.

In the twentieth century, the behaviorist and functionalist movements pushed consciousness analysis away from inner experience, and the higher-order approach emerged as a way to naturalize consciousness within a functionalist framework. David Rosenthal developed the HOT theory in the 1980s as an account of consciousness that is fully reductive: consciousness is analyzed in terms of mental states and the relations between them, with no irreducible phenomenal properties. The theory gained prominence because it promised to explain consciousness with ordinary cognitive resources — representation and thought — while avoiding the mysteries of qualia. It also aligned with the growing emphasis in cognitive science on metacognition — the capacity to think about one's own mental states.

Key Arguments & Debates

The central argument for higher-order theories is that they offer a reductive, naturalistic account of consciousness. They explain what it is for a state to be conscious in functional and representational terms, making consciousness tractable for cognitive science: consciousness becomes the availability of a state to higher-order representation, which can be studied and implemented. The theories also handle the explanatory gap in a deflationary way: if consciousness just is higher-order representation, there is no further "what it is like" to explain, because the phenomenology is constituted by the representation.

The empirical motivation comes from metacognition and from the phenomenon of "state consciousness" versus "creature consciousness": animals and systems that can represent their own states have the capacity for consciousness, and the theory predicts that consciousness varies with higher-order representational capacity. Some find support in "unconscious perception" phenomena: a stimulus that is represented at the first-order level without higher-order representation is not consciously experienced.

The objections are formidable. The regress problem: if a state is conscious when represented by a higher-order state, what makes the higher-order state conscious? If it needs its own higher-order state, we get an infinite regress; if not, the theory seems arbitrary about which level gets to be conscious. Rosenthal responds that the regress is harmless: consciousness requires only that the higher-order state be about the first-order state; the higher-order state itself need not be conscious. The targetless state problem: the theory seems to imply that a higher-order state could represent a nonexistent first-order state, producing an "experience" with no basis — which critics say misdescribes what experience is. The "empty" consciousness objection, pressed by Ned Block and others, argues that the theory gets the phenomenology wrong: raw sensory feels, like the quality of a toothache, do not seem to require a thought about themselves. And the problem of infant and animal consciousness: if consciousness requires higher-order thoughts with concepts, then creatures without such concepts — infants, dogs — would lack consciousness, which many find implausible. Defenders reply that the required concepts are minimal and that the empirical evidence on animal metacognition supports the theory.

Contemporary Relevance

Higher-order theories remain a major research program in the science of consciousness. They have direct implications for the neural correlates of consciousness: if consciousness is higher-order representation, then the correlates should include prefrontal or parietal regions supporting metacognitive monitoring, not merely sensory cortices — a prediction that current neuroimaging research tests. The theories also connect to the growing study of metacognition and confidence judgments in cognitive science, and they have influenced the design of "no-report" paradigms that aim to separate conscious experience from reportability.

For AI, the higher-order framework is attractive because it is implementation-friendly: if consciousness is a matter of a system representing its own states, then machine consciousness becomes a matter of building systems with higher-order self-monitoring — which is a tractable engineering problem. Whether such systems would thereby be genuinely conscious, or merely good at self-reporting, remains the live question. And the debate about whether higher-order theories explain away or fail to capture phenomenal feel continues to define one of the central battle lines in the philosophy of consciousness.

Related Concepts

• Phenomenal consciousness: what higher-order theories aim to reduce or explain.
• Access consciousness: the functional notion closest to higher-order representation.
• The hard problem vs easy problems: the question of whether higher-order representation solves the hard problem.
• Intentionality: the representational aboutness on which higher-order theories rely.
• Introspection: the first-person capacity that higher-order theories formalize.

Further Learning

• Stanford Encyclopedia of Philosophy: Higher-Order Theories of Consciousness
• Internet Encyclopedia of Philosophy: Higher-Order Theories of Consciousness
• David Rosenthal, Consciousness and Mind (2005).
• Peter Carruthers, Phenomenal Consciousness: A Naturalistic Theory (2000).