Higher-Order Thought (HOT) Theories

Higher-order thought theories explain what makes a mental state conscious by appeal to higher-order mental states — thoughts or perceptions directed at the first-order state itself. A pain is conscious not because of anything intrinsic to the pain, but because one has a thought (or perception) about the pain. Unconscious mental states are states one is not in any way conscious of being in.

The view originates with David Rosenthal (City University of New York), who developed it as a way to explain the commonsense asymmetry between conscious and unconscious mental states while remaining within a broadly physicalist framework. HOT theories represent the most developed alternative to intentionalism for explaining phenomenal consciousness without invoking qualia as primitive inner properties.


The Transitivity Principle

Rosenthal’s foundational insight. When we speak of a mental state as “conscious,” we mean something close to what Descartes and Locke meant when they said one is conscious of one’s mental states. A state is conscious if, and only if, one is conscious of being in it in some suitable way.

The Transitivity Principle (TP): a mental state S is conscious iff one is transitively conscious of S.

This is not circular: the higher-order state that makes S conscious need not itself be conscious. A subliminal perception makes one conscious of an object without being conscious itself. Equally, a HOT that makes a pain conscious need not be a conscious thought — and usually isn’t.

TP receives support from the direction of explanation: no state seems conscious if one is in no way whatever aware of being in it. The question is only how we are conscious of our conscious states.


Two Versions

1. Inner Sense (Higher-Order Perception)

Armstrong and Lycan: a mental state is conscious when a quasi-perceptual inner monitoring mechanism tracks it. Consciousness involves a kind of inner sense directed at first-order states. The monitor relays information about ongoing psychological events, enhancing rationality and coordination.

Rosenthal’s objection: when we are conscious of our mental states, no mental quality belongs to the higher-order awareness itself — the awareness of pain has no intrinsic quality of its own. But perceiving always involves some qualitative character. Therefore our higher-order awareness of mental states cannot be perceptual. It must be cognitive — a thought about the state.

2. Higher-Order Thoughts (Rosenthal)

The canonical version. A mental state is conscious in virtue of being accompanied by a suitable HOT — an assertoric, non-inferential thought to the effect that one is in that state right now.

Key features:

  • Non-inferential: the HOT must not be mediated by inference; otherwise the resulting consciousness would feel mediated, which it doesn’t
  • Assertoric: the HOT asserts rather than doubts or wonders; doubting doesn’t make one conscious of anything
  • Occurrent, not merely dispositional: being disposed to have a HOT doesn’t make one conscious of a state; one must actually have the HOT (contra Carruthers’ dispositional variant)
  • The HOT itself rarely conscious: HOTs are conscious only when accompanied by third-order thoughts — which seldom happens. This is why we don’t notice being in two states at once when we have a conscious experience

Evidence Rosenthal offers:

  • Reportability: we determine whether a state is conscious by whether one can report it. Reports express thoughts about one’s states. So conscious states are reliably accompanied by thoughts about them.
  • Fine-grained variability: learning new words for qualitative properties sometimes makes one conscious of experiences in more fine-grained ways. This suggests HOTs (which employ concepts) govern the character of conscious experience.
  • Dental fear: patients with anaesthetized nerves sometimes feel pain due to fear and the sensation of the drill’s vibration. The HOT (about being in pain) determines what it’s like for them, not the first-order state.

Carruthers’ Dispositional Variant

Carruthers objects that Rosenthal’s view posits two states when experience seems to involve only one. He proposes: a state is conscious if a HOT is disposed to occur, not if one actually occurs.

Rosenthal’s reply: dispositions cannot implement TP. Being disposed to be conscious of something doesn’t make one conscious of it. And the dispositional theory cannot explain introspection as anything more than a disposition-to-a-disposition, losing explanatory traction.

Carruthers’ later move: build higher-order content intrinsically into the first-order state (via a disposition-to-cause-a-HOT that confers higher-order content). The conscious state itself has the content that it represents itself as a mental state. This eliminates the two-state worry. Kriegel (self-representationalism) develops a version of this independently.


The Core Objection: HOTs Without Targets

If a HOT can occur without a suitable first-order state, there should be a phenomenally conscious experience with nothing “underneath” it. Rosenthal bites this bullet: the HOT alone would suffice for what it’s like for one to have that experience. The dental fear case supports this — what it’s like is determined by how one is conscious of a state, not by the state’s intrinsic properties.

The mismatch problem: if the HOT misrepresents the first-order state (HOT says “I’m in pain” but the first-order state is fear), what is the experience like? On TP, it’s like being in pain — because what it’s like is determined by the HOT’s content, not the state’s intrinsic properties.

This generates a genuine puzzle for HOT theories about the relationship between phenomenal character and neural reality.


What HOT Theories Get Right

The view has genuine strengths:

  • The distinction between conscious and unconscious mental states: there clearly are unconscious mental states (subliminal perception, masked priming), and HOT theories offer a principled explanation of what makes the difference
  • The role of self-awareness in consciousness: there is something right about the idea that conscious experience involves a kind of awareness of one’s own states — the question is whether this is higher-order in Rosenthal’s sense or whether transparency provides a different explanation
  • The reportability connection: the link between consciousness and verbal report is real; HOT theories explain it neatly
  • The empirical orientation: Rosenthal takes cognitive science seriously and tries to make HOT theories testable — a genuine virtue

My View

HOT theories and intentionalism both reject qualia realism — both hold that phenomenal character doesn’t consist in intrinsic properties of inner objects. But they diverge crucially:

HOT: Consciousness requires self-representation

A state is conscious iff there is a thought about it. Phenomenal character is relational — it consists in being represented by a higher-order state. A first-order state with exactly the phenomenal character of a pain could exist unconsciously, with no phenomenal character, simply by lacking a HOT.

Intentionalism: Phenomenal character consists in world-presenting content

Among conscious states, phenomenal character consists in content under a perceptual or bodily-affective mode, with the relevant determinacy, field structure, and attentional organization. A separate state-consciousness question asks what makes such content figure in experience rather than remain unconscious processing; poising, recurrence, global availability, and related forms of integration remain competing threshold hypotheses. Intentionalism denies that a higher-order representation is required, not that a consciousness threshold must be explained.

Three structural advantages of intentionalism over HOT theories:

  1. No two-state problem: intentionalism doesn’t require a distinct higher-order state. The experience represents the world; its phenomenal character consists in that representational content. One state, one experience.
  2. Transparency explained, not merely accommodated: the Transparency Thesis falls out naturally from intentionalism — introspection returns us to the world because that’s what the experience represents. HOT theories must treat transparency as a datum to be explained after the fact: why, if the HOT is about the first-order state, does introspection seem to reveal the external world rather than an inner state?
  3. Externalism integrated: intentionalism connects naturally to semantic externalism and teleosemantics — what a state represents depends on its causal-historical relation to the environment. HOT theories are resolutely internalist: the HOT is about a first-order mental state, not about the world. They face the same problems as functionalism regarding content determination.

The dental fear case reconsidered: Rosenthal takes dental fear to show that what it is like is determined by the HOT rather than by the first-order state. The intentionalist need not grant that result. Fear, expectation, and vibration may alter the content and bodily-affective mode of the conscious first-order state so that a disturbance is presented as pain. That is misrepresentation within the experience itself. If no suitably organized conscious first-order state occurs, a higher-order thought that one is in pain does not by itself generate pain phenomenology.


  • Intentionalism — my preferred alternative; phenomenal character as world-directed first-order representational content rather than higher-order self-representation
  • Transparency — the transparency thesis is difficult to explain on HOT theories (the HOT is about an inner state, not the world); it falls out naturally from intentionalism
  • Qualia — HOT theories and intentionalism agree that qualia realism is wrong; they disagree about what replaces it — self-representation vs. world-representation
  • The Hard Problem of Consciousness — HOT theories attempt to dissolve the hard problem; Chalmers argues it survives even after HOT; my view agrees but takes a different dissolution route
  • Functionalism — HOT theories are a sophisticated functionalism about consciousness; they inherit functionalism’s problems about content determination and external grounding
  • Intentionality — HOT theories focus on self-directed intentionality (representations of representations); intentionalism focuses on world-directed intentionality as primary
  • Embodiment — HOT theories are resolutely internalist; they don’t explain how higher-order representations get their content grounded in the world