Higher-Order Theories of Consciousness

The Explanandum: Why Some Mental States Are Conscious

Higher-order theories of consciousness originate in a specific explanatory problem. A subject can perceptually discriminate, evaluate, or respond to a stimulus without consciously experiencing it in the ordinary sense. Likewise a belief, desire, or emotional disposition can affect behavior while remaining outside conscious awareness. If states with broadly similar first-order functions can occur both consciously and unconsciously, their content alone may fail to explain the difference. Higher-order theorists propose that the difference concerns a relation between the mental state and the subject's representation or awareness of being in it. This proposal concerns how a mental state becomes conscious, rather than simply whether a person is awake.

The initial force of the proposal depends upon two propositions. Mentality is wider than consciousness, and conscious mentality has a distinctive form of presence to its bearer. Rosenthal's Two Concepts of Consciousness (1986, pp. 329–359) argues that the explanatory task changes depending on whether consciousness is treated as inherent in mentality or as a property mental states can sometimes possess. If one allows unconscious mental states, it becomes meaningful to ask what additional condition accounts for the consciousness of otherwise comparable states. Higher-order theories answer by appeal to awareness of the state itself.

Consider ordinary vision. A person can carry out visually guided actions while failing to report or retain conscious awareness of particular features. Such cases motivate distinctions among information processing, behavioral discrimination, access for deliberate use, and phenomenal appearance. They do not independently demonstrate that the information lacked phenomenal character; alternative explanations include failures of attention, report, memory, or confidence. The higher-order hypothesis therefore requires both a precise philosophical claim and empirical criteria that distinguish it from first-order accounts. The familiar examples of blindsight and absent-minded driving are starting points for theory construction, not automatic confirmations (Carruthers, 2000; Rosenthal, 2005).

The monograph's central question is this: what explains the fact that a mental state appears to its bearer as an episode of experience? A satisfactory theory must characterize the relevant higher-order relation, establish how it arises, explain the phenomenology of subjectivity, and distinguish the proposed mechanism from reflective inspection or verbal report. Because a representation can be accurate, inaccurate, conceptually impoverished, or altogether unsupported by a target, higher-order theories must also explain why these variations do or do not change the consciousness of the represented state.

Definition, Scope, and Essential Distinctions

A higher-order theory of consciousness is an account under which the conscious status of a mental state depends on a suitable higher-order representation, awareness, or self-representational organization that concerns that very mental state. In a two-state model, a first-order state has a worldly, bodily, or intentional object; a distinct higher-order state represents the occurrence or character of the first. In a one-state model, the episode may represent its first-order object while also bearing a constitutive self-directed representational component. The common question concerns the higher-order condition; the theories disagree about what realizes that condition and whether it is sufficient for phenomenality.

Several distinctions delimit the subject. Creature consciousness concerns the condition of an organism or agent, including wakefulness and awareness of something. State consciousness concerns the consciousness of an individual mental episode. Access consciousness names a form of availability for reasoning, decision, or report; phenomenal consciousness names what it is like for a subject to undergo the state. These distinctions are discussed influentially by Block (1995) and throughout the Stanford Encyclopedia of Philosophy's 2026 treatment. Their boundaries remain disputed: a theory may accept them as methodological contrasts without assuming that there are sharply independent kinds of consciousness.

Higher-order awareness also differs from overt introspection. An experience of pain may be conscious while the person attends to a task. Reflection upon the pain as an object of inquiry is a further mental operation. In Rosenthal's theory the higher-order thought that renders pain conscious generally need not be conscious itself. In Carruthers's dispositional version, the subject may lack any currently tokened higher-order thought. Thus neither theory identifies ordinary phenomenality with deliberate inspection or articulated self-description.

The word higher denotes an intentional or explanatory ordering, not an elevation in value or necessarily a location in the frontal cortex. A representation is higher-order because its content is directed at another representational or mental state. A judgment about the external object represented by a visual perception can still be first-order in the relevant sense, while a nonverbal representation that one is seeing the object can be higher-order. Functional, representational, and neural descriptions therefore require careful alignment.

Ancient and Early Modern Sources

The origins of higher-order theorizing involve a history of self-awareness that predates the contemporary distinction between conscious and unconscious mental states. Aristotle's De Anima discusses the sense by which we perceive that we see and hear (III.2, 425b12–25). This passage raises a recursive problem: if a distinct sense registers each act of seeing, what registers that second sensing? Aristotle's response is bound to his account of the unity and activity of perception. It anticipates an issue later formulated as the regress of higher-order monitoring, though it would be anachronistic to attribute to him a theory built around experimental distinctions between unconscious and conscious processing.

Locke's Essay Concerning Human Understanding associates consciousness with awareness of the mind's own operations and connects personal identity with the reach of consciousness (1690, Book II, chapters I and XXVII). His account captures a reflexive element in experience, yet the contemporary historical interpretation requires care. Locke did not offer a worked-out explanation of mental-state consciousness by systematically contrasting unconscious and conscious perceptions. The 2026 contribution by Rosenthal to the Open Encyclopedia of Cognitive Science explicitly regards the frequently cited Lockean formulation as an important precursor rather than an established modern higher-order theory.

There is also an important historical difference between detecting an internal state and the state being phenomenally present. Ancient discussions of perceiving one's perception may be concerned with the unity of sensory activity. Early modern theories of reflection can function as epistemologies of how one knows mental operations. The higher-order project needs the additional thesis that appropriate awareness constitutes or explains conscious status. Without that thesis, reflexivity is a philosophical theme rather than the distinctive solution offered by higher-order theorists.

The historical lesson is methodological. Similar words—reflection, awareness, apperception, self-perception—can perform different explanatory tasks. A history of higher-order theories must trace the question answered by each philosopher and the arguments connecting self-relation to mental-state consciousness. The recurrence of reflexive vocabulary alone does not prove continuity of a single theory.

Leibniz, Apperception, and Unconscious Representation

Leibniz makes the problem sharper by distinguishing perception from apperception. In Monadology §14, perception concerns representational states of a simple substance; in §4 of Principles of Nature and Grace he characterizes apperception as consciousness or reflective knowledge of an internal state. The distinction challenges the identification of every mental occurrence with conscious experience. His account also assigns explanatory significance to petites perceptions, subtle or unnoticed perceptual elements that contribute to conscious life without individually standing out as noticed experiences.

These claims have an affinity with higher-order thinking because they acknowledge that a state can represent while the bearer lacks an appropriate kind of awareness of representing. But Leibniz's metaphysical system supplies its own picture of substance, expression, appetition, and degrees of clarity. Translating his conception directly into the modern claim that a nonconscious token mental state becomes conscious when targeted by a distinct computational monitor would erase that system. The correct comparison concerns an explanatory structure: first-level representation and some distinctively conscious awareness of it.

This distinction has lasting force. Modern cognitive science routinely models perception that informs action while remaining outside conscious report. A Leibnizian framework invites the question whether additional reflective structure differentiates conscious representation, but it does not settle what such a structure must be. It also raises a problem that persists for higher-order theories: if apperception is described as awareness, what explains the awareness of apperception, and why does a sequence of increasingly reflexive states not have to continue indefinitely?

Leibniz therefore marks a meaningful historical advance: he treats unconscious representation as constitutive of ordinary psychology and makes consciousness a property requiring explanation. The metaphysical commitments and the details of the proposed explanation remain historically particular (Leibniz, 1714, Monadology §§14–17; Principles of Nature and Grace §§4–5; Rosenthal, 2026).

Nineteenth-Century Psychology and Pre-Reflective Awareness

The nineteenth century transformed the vocabulary of mind through experimental psychology, philosophical psychology, and analyses of intentionality. Franz Brentano's Psychology from an Empirical Standpoint (1874) characterizes mental phenomena by their intentional directedness and discusses an accompanying awareness of acts. Brentano's account can be read as a source for intrinsic or same-state models of consciousness, although debates remain about how the awareness belongs to the act and what it means for an episode to have itself among its objects. It should not be assimilated without argument to Rosenthal's distinct-state higher-order thought.

Husserlian phenomenology and subsequent discussions of prereflective self-awareness confront a related issue from the first-person direction. Experience seems already to have a tacit self-givenness before one makes it the object of explicit reflection. Sartre's Being and Nothingness (1943), in particular, emphasizes prereflective consciousness as intrinsic to the occurrence of awareness. This line of argument presses against the idea that ordinary consciousness always depends on a further, separately occurring representational event.

The difference is explanatory. Distinct-state theories treat a suitable representation of the mental state as the condition that makes it conscious; intrinsic models treat reflexive presence as belonging within the conscious event. The latter must still explain the nature of self-directedness and why it is more than a redescription of experience. The former must explain why a relation between states yields the first-person character that seems immediately present. Both inherit the problem of avoiding a merely circular appeal to consciousness in order to explain consciousness.

In contemporary historiography, Rosenthal (2026) notes the potential relevance of Nietzsche's reflections on consciousness and its relation to communication, while identifying the sustained twentieth-century research program with Armstrong, Rosenthal, and Carruthers. A historical claim about Nietzsche as an author of a developed higher-order theory requires close reconstruction of the passages in The Gay Science and related texts, including their social and linguistic aims. The present study therefore treats that connection as an interpretive proposal rather than a settled transfer of modern terminology to Nietzsche.

The Modern Research Program

D. M. Armstrong's A Materialist Theory of the Mind (1968; second edition 1993/1994) advances an account of consciousness in terms of self-scanning or inner perception within a materialist philosophy of mind. The intuitive attraction of this model is explanatory division of labor. Perceptual systems register environmental features while a monitoring system registers aspects of perceptual processing. Ordinary conscious experience can then be understood through the interaction between functional stages rather than by treating consciousness as an unanalysable power of mental events (Armstrong, 1968; 1994).

David Rosenthal's Two Concepts of Consciousness (1986) and Consciousness and Mind (2005) develop the actualist higher-order thought approach. Rosenthal argues that the consciousness of a state depends on the subject's appropriate higher-order thought of being in it. The explanatory advantage is that the higher-order thought can itself be nonconscious, allowing the theory to account for conscious episodes without requiring constant conscious introspection. This is a central piece of the proposal, not an afterthought added merely to neutralize a regress objection.

William Lycan's Consciousness and Experience (1996) elaborates an inner-sense approach in dialogue with functionalism and representational theories. Peter Carruthers's Phenomenal Consciousness: A Naturalistic Theory (2000) defends a dispositional higher-order thought approach in which representational availability to a mindreading or higher-order conceptual faculty can be decisive. Rocco Gennaro's The Consciousness Paradox (2011) pursues an account aimed at reconciling higher-order theory with concept acquisition and animal consciousness; Uriah Kriegel's Subjective Consciousness (2009) develops a self-representational alternative with a different internal architecture.

The research program is therefore a family of philosophical attempts to locate the basis of consciousness in the representation of mental states, not a single neuroscientific theory. The models differ in whether the higher-order relation is actual or potential, perceptual or thought-like, external to or internal to its object, constitutive of phenomenality or only explanatory of mental-state awareness. Debates over neural location and empirical evidence make sense only after these differences have been made explicit.

The Transitivity Principle

The transitivity principle says, approximately, that a mental state is conscious only if its subject is in some appropriate way aware of being in it. This is the intuitive fulcrum of higher-order theories. Rosenthal (1997; 2026) presents the principle as a route from the ordinary distinction between conscious and unconscious states to the explanatory task of accounting for the relevant awareness. Its role varies: it may serve as a pretheoretical specification of the explanandum or as a substantive premise in a theory of what makes a state conscious.

The argument can be reconstructed as follows. Suppose a state M occurs in a subject S, and S is wholly unaware of being in M. In the ordinary idiom of state consciousness, M seems unconscious. If M is conscious, then, S must have some awareness of being in M. A theory of consciousness should explain that awareness without simply repeating the word conscious. Higher-order representations are offered as independently intelligible mental entities or properties capable of filling the role.

The first logical step establishes, at most, a necessary condition: conscious M implies appropriate awareness of M. The stronger thesis, suitable higher-order representation implies conscious M, is a sufficiency claim and requires further argument. The transitivity principle also leaves the meaning of awareness unsettled. If awareness of M already includes phenomenal experience, the explanation risks circularity; if it means merely the availability of information about M, the theory must explain why availability entails phenomenality. A third option treats awareness as representational but holds that the relevant representational organization itself constitutes the conscious mode of appearance.

The principle has intuitive force partly because the statement that someone undergoes a conscious state of which they are wholly unaware sounds unstable. Yet this linguistic force may reflect how we ordinarily identify conscious states, rather than reveal their metaphysical constituents. The dispute between higher-order theorists and defenders of phenomenal overflow turns on precisely this possible difference between a criterion for attribution and a condition of existence (Block, 2011).

Actualist Higher-Order Thought Theory

In an actualist higher-order thought theory, an appropriate higher-order thought must actually occur for the target state to be conscious. In the classical Rosenthal version the thought represents the subject as now being in a relevant mental state and normally arises in a suitable, noninferential way. The higher-order thought can be unconscious; otherwise an explanation of every conscious state would depend on an endless sequence of conscious thoughts. Its content may be rendered schematically as a thought of the form: S is now in state M (Rosenthal, 1986, 2005).

The distinction between first-order content and higher-order targeting is essential. A visual state representing a red surface need not contain a thought about the visual state. A higher-order thought may represent the subject as seeing red. The latter makes the state conscious on the theory, or—on some disputed variations—explains why the subject is aware of a red experience even when the corresponding target is absent. The account thus separates the properties in virtue of which a mental state is perceptual from the additional property in virtue of which it is conscious.

Actualist theory offers a relatively clear answer to the contrast between conscious and unconscious instances of the same mental type. An unconscious visual state may possess discrimination-guiding content, while a conscious visual state is accompanied by appropriate higher-order awareness. The theory further predicts a distinction between having a conscious experience and consciously introspecting it: the latter involves a further level of conscious awareness, and in ordinary experience the first higher-order thought itself need not enter consciousness.

A complication concerns causal and intentional adequacy. If a spontaneous thought about seeing red makes a visual state conscious, what causal relation must obtain between the thought and its supposed target? Rosenthal emphasizes the importance of noninferential generation in the ordinary case, but later discussions reveal tension between a theory of consciousness as the representation of an experience and a theory of experience as it seems to the subject. Whether an inaccurate or empty HOT produces phenomenality is not a peripheral detail: the answer determines what the theory says conscious experience is.

Dispositional Higher-Order Thought Theory

Carruthers's dispositional account changes the relevant condition. A first-order perceptual representation may count as conscious if it is appropriately available to a system capable of higher-order thought, even when no actual higher-order thought about that representation occurs at the time. In Phenomenal Consciousness (2000, especially the preface and the account of higher-order thought), Carruthers proposes that analog or fine-grained contents in the appropriate working-memory role are available to a theory-of-mind faculty. The availability contributes to an account of their subjective status, with an accompanying commitment to a form of consumer semantics.

The dispositional proposal offers an answer to a computational objection against actualist views. A consciousness mechanism need not continually construct explicit thoughts about every experienced detail. It can instead be organized so that the relevant contents are poised for higher-order uptake. This potentially allows complex perceptual experience to exceed the number of actual, discrete higher-order conceptual episodes while retaining an explanation framed in the architecture of higher-order use.

The challenge is what the disposition contributes at the precise moment of consciousness. If consciousness is constituted by availability alone, a state may be conscious while it has never been represented at a higher level. The theory requires an argument that structural potential, suitably understood, can constitute a present phenomenal fact. Moreover, a state can be available for some uses while not available for others: introspective discrimination, judgment, verbal report, and flexible action are different consumers. A successful dispositional theory must specify which availability relation matters and why.

Carruthers's naturalistic explanation derives its strength from a detailed hypothesis about cognitive architecture. Its controversial point is not that working memory and mindreading exist; those capacities can be studied independently. It is the bridging claim that their relation to first-order perceptual content explains the subjective character of experience. As a result, the dispositional model remains open to comparison with global-workspace and first-order accessibility theories that invoke similar cognitive resources without adding a higher-order constitutive condition.

Higher-Order Perception and Inner Sense

Higher-order perception models consciousness on a form of inner sensing. Armstrong and Lycan suggest that a functional monitoring capacity could detect a first-order mental state much as outward perception detects environmental conditions. The metaphor becomes philosophically significant when it is translated into an architecture: first-order processes encode worldly or bodily conditions, and a monitoring process encodes information about those first-order processes. The resulting higher-order state need not involve the same conceptual structure as an explicit judgment.

This model explains why consciousness may appear perceptual rather than propositional. The vividness of an experience need not result from a consciously entertained statement that one is experiencing it. An inner monitor could generate fine-grained representations of perceptual processing in a way structurally closer to a sensory mechanism. Its putative evolutionary role would include error detection, flexible control, and sensitivity to the distinction between appearance and world. None of those possible functions, however, independently establishes the occurrence of phenomenal consciousness.

The central philosophical problem is specification of the inner sense. Ordinary sense modalities are identified partly through their distinctive transducers, contents, functional pathways, and relations to environmental stimuli. An inner-sense theory must identify what makes internal monitoring perceptual rather than merely informational. If every reliable internal signal counts as a perception, the proposal risks becoming too inclusive; if it requires a richly sensory representation of an already phenomenally characterized state, it risks explaining one consciousness by another.

Lycan's Consciousness and Experience (1996) is significant for embedding the monitoring idea in a broader functional and representational view. The strength of the account is that it treats consciousness as causally integrated with psychological capacities. Its open explanatory burden is the transition from internal information about a state to the subjective appearance of being in that state, including the possibility of illusion in the monitor.

Self-Representational and Intrinsic Accounts

Self-representational theories seek an account of subjectivity without separating the higher-order representer from the represented state into two independently existing episodes. Brentano's analyses and later phenomenological discussions offer historical resources; Kriegel's Subjective Consciousness: A Self-Representational Theory (2009) gives a systematic contemporary formulation. A conscious event represents its ordinary object and, in an appropriate sense, itself. The reflexive relation is intrinsic to the event's organization.

This approach takes seriously the intimacy of first-person awareness. When one sees a tree, the seeing can be present to oneself as occurring without a second episode of reflective inspection. The theory aims to explain this immediate subjective character through an internally structured representational episode. Its mereological variants may treat the first-order aspect and the reflexive aspect as distinguishable parts of a unified conscious state. Kriegel's account seeks necessary conditions for subjective character and a specific way in which self-representation contributes to phenomenality.

A one-state approach faces its own identity problem. If the subject's state represents itself, what makes this reflexive directedness genuine representation rather than the mere fact that the state exists? If a representation has satisfaction conditions, one must explain how a token can successfully or unsuccessfully represent its own occurrence. If it can be self-representational without an object distinct from itself, the model must articulate this relation without importing the full phenomenality it sets out to explain.

The distinction between one-state and two-state models is philosophically important, but it does not by itself resolve the explanatory gap. Reducing the number of mental tokens eliminates some misrepresentation scenarios involving missing targets, while retaining questions about why the intrinsic structure has a first-person qualitative character. The Kriegel–Williford collection (2006) places self-representational approaches into direct debate with first-order and higher-order monitoring accounts.

A Comparative Architecture of Higher-Order Explanations

The theories can be compared according to four independent parameters. First is the form of the higher-order vehicle: thought-like, perception-like, or an intrinsic self-directed representational aspect. Second is its temporal and modal status: actual token occurrence or dispositional availability. Third is the dependence between the first-order state and its higher-order representation: two potentially independent tokens, two inseparable aspects, or a single complex state. Fourth is the intended explanandum: state consciousness, subjective character, introspective accessibility, or particular aspects of qualitative appearance.

These parameters reveal important distinctions often concealed by a single label. Rosenthal's classical HOT requires an actual relevant thought, but does not require that the thought itself be consciously entertained. Carruthers's dispositional model may depend on a mindreading capacity whose deployment is possible rather than currently actual. Inner-sense approaches may use fine-grained or analog monitoring that resists assimilation to propositional thought. Self-representationalism can posit a reflexive internal aspect while resisting a distinct monitoring episode.

A more precise comparison also distinguishes three modes of explanatory claim. A correlation claim says that higher-order processes accompany conscious states. A constitutive claim says that the right higher-order relation makes a state conscious. A reductive claim says that this constitution can be understood through independently specified representational and functional properties. Evidence for correlations alone does not establish constitution, and a constitutive analysis need not automatically deliver a neural reduction.

The central synthesis of the first half of this study is that higher-order theory gains its explanatory precision when it specifies which of these claims it is defending. The theories differ in exactly what higher-order representation does, how it relates to its target, and how much of phenomenal consciousness it purports to explain. Their common philosophical achievement is to formulate self-presentation as a tractable problem within a theory of the mind; their shared burden is to justify the metaphysical transition from informational relation to conscious appearance.

The Regress Objection and Introspection

One of the oldest difficulties for reflexive theories concerns explanatory regress. If every conscious mental state must be accompanied by awareness of that state, must the accompanying awareness itself be conscious? If so, a second awareness is required, then a third, and the sequence appears to lack a terminus. In a familiar form, the regress argument assumes that consciousness is required at every level of awareness. Modern higher-order theories commonly reject this premise: a first-order state can be conscious in virtue of an accompanying higher-order representation that is itself unconscious (Rosenthal, 1986, 2005).

This response is logically adequate against the most immediate regress. Let M be the target mental state and H(M) the higher-order state representing it. If H(M) can exist unconsciously, then M may be conscious while H(M) does not require any additional targeting representation. Introspection occurs when the subject becomes conscious of H(M) or otherwise deliberately attends to the first-order state. The conceptual sequence is open-ended in principle, but the constitution of ordinary consciousness does not require that it be instantiated beyond the first higher-order relation. The theory thereby distinguishes the higher-order status of a representation from its own consciousness.

A different challenge concerns the intelligibility of unconscious awareness. If the higher-order representation is entirely unconscious, what is the sense in which the subject is aware of the first-order state? Rosenthal's response is that transitive awareness need not be an occurrent phenomenal awareness; a thought can represent a state without itself entering consciousness. Critics argue that this move alters the meaning of awareness precisely where the explanatory argument derives its intuitive support. The theory therefore owes an independent account of how representational awareness differs from conscious noticing and why the former can confer consciousness on its object.

An analogous question arises for intrinsic reflexivity. A self-representational state may avoid a hierarchy of separate states by containing its reflexive component, yet this does not abolish the need to explain the difference between merely self-related physical processes and conscious self-presentation. Regress arguments thus identify a constraint on theoretical architecture; their resolution does not, by itself, demonstrate that the proposed architecture produces phenomenality.

Misrepresentation, Empty Thoughts, and Targetless Consciousness

Higher-order representations can misrepresent their objects. A subject might have a higher-order thought as of experiencing bright red when the first-order state represents muted orange, or might have the higher-order thought without any matching first-order visual episode. Philosophers commonly call the latter a targetless or empty HOT. The problem is structurally unavoidable for two-state models in which the higher-order representation and first-order target can vary independently (Byrne, 1997; Levine, 2001; Rosenthal, 2005; Wilberg, 2010).

Suppose H represents S as undergoing M, but there is no M. If consciousness is identified with the occurrence of H and what it presents, the theory appears to allow an experience as of M without a token M. An actualist theorist may accept such cases as possible forms of subjective seeming, or may require an appropriate first-order target and causal relation. The latter requirement narrows sufficiency: H by itself is no longer sufficient. Yet this may be a principled restriction if the object of explanation is the conscious status of an existing mental state rather than the appearance of an episode without the underlying first-order state.

The question becomes particularly difficult when a HOT misrepresents some but not all features of an existing state. Does it change the conscious qualitative character, or only the subject's metarepresentational judgment? Imagine a chromatic state that encodes red while a higher-order mechanism categorizes the episode as orange. A theory that places phenomenal quality entirely in first-order content and a theory that places its conscious appearance in higher-order content will generate different answers. The cases clarify why higher-order accounts must articulate both the condition of conscious status and the constitution of phenomenal character.

Wilberg's Consciousness and false HOTs (2010) argues for rejecting consciousness in targetless cases while distinguishing them from merely inaccurate HOTs. The 2026 Stanford Encyclopedia overview documents a sustained controversy involving Rosenthal, Gennaro, Weisberg, Berger, Brown, and others. A disciplined analysis treats the targetless case as a family of counterfactual tests rather than as an established clinical phenomenon. To claim that people actually undergo targetless HOTs would require evidence beyond the philosophical possibility of the configuration.

An important conceptual result follows. A higher-order representation can be fallible in the same general sense as other representations, but consciousness is the very phenomenon whose character must be determined by its putative fallibility. A successful theory must specify which errors produce illusion, which errors undermine consciousness, and which are simply inaccurate judgments about an otherwise conscious episode.

The Rock Objection and the Explanatory Gap

The so-called rock objection asks why representing a mental state makes that state conscious when representing a rock does not make the rock conscious. The naive formulation overlooks a relevant difference: the object of the higher-order representation is already a mental state, while a rock is not. Higher-order theories concern a transition in the status of mental states, and a theory can stipulate that only objects of the right psychological kind can acquire that status (Goldman, 1993; Lycan, 1996).

Yet the objection contains a deeper issue. Merely describing a relation between a representation and its object does not explain why the object thereby has a qualitative mode of existence for the subject. For a conscious experience of pain, the target is a mental state, but its mental status alone is compatible with unconscious pain representations on the very assumptions that motivate the theory. What aspect of the special higher-order relation changes that? The reply that the target is mental marks the domain of possible consciousness without explaining the conferral mechanism.

This is a local version of the explanatory gap problem. Jackson's knowledge argument and Chalmers's hard-problem formulation question whether a complete functional or physical description captures what phenomenal experience is. Higher-order theorists propose a mentalistic explanatory step: consciousness is analyzed through representation, and representation may in turn be given naturalistic explanation. This can be a genuine reduction if the characterization of higher-order representation is independent and if the resulting structural facts entail the relevant phenomenon. Critics argue that a gap remains between a subject's representing that it has M and there being something it is like to have M.

The strongest form of the objection is not an unargued declaration that experience is mysterious. It specifies an entailment claim. One can conceive, a critic argues, the same first-order and higher-order informational relations without any subjective feel. A defender may respond that this conceivability depends on misconstruing what the theory identifies, just as one can fail to recognize an identity between differently described physical properties. The debate therefore requires theories of explanation, identity, representation, and epistemic possibility. Neither the rock analogy nor an appeal to intuition alone settles the metaphysics.

Phenomenal Character, Qualia, and Representational Content

Phenomenal consciousness contains at least two questions: why a state is conscious at all and why it has this particular experiential character. A higher-order theory may distinguish the fact that a visual state is conscious from the fact that it visually presents red rather than blue. On a simple division of labor, first-order content supplies qualitative differentiation while higher-order representation supplies consciousness. Yet a theorist who makes higher-order seeming determinative in mismatched cases may assign the higher-order level a role in determining apparent quality as well.

The qualitative character of a perceptual episode also has fine-grained structure. A person can discriminate color shades that resist stable verbal categories; a perceptual field can contain many features without individually entertained conceptual judgments. This presents a burden for a version of HOT theory that requires a distinct conceptual representation for each experienced quality. The defender may posit nonverbal, coarse-grained, or structurally economical higher-order representations, or switch to a dispositional architecture in which lower-order fine-grained content is available for higher-order uptake (Carruthers, 2000; Brown, Lau, & LeDoux, 2019).

The issue also bears on intentionalism, the claim that phenomenal character is wholly or substantially determined by representational content. First-order representational theories such as those of Dretske (1995) and Tye (1995) attempt to explain consciousness without the additional higher-order representational relation. Higher-order theories must show that first-order content with the relevant functional role can remain unconscious, and that the difference between its conscious and unconscious occurrence requires higher-order organization. Blindsight and other examples motivate the distinction, but the empirical interpretation of each case remains contested.

Ned Block's overflow argument (2011) raises a specific challenge to accounts that tie phenomenality too closely to cognitive access. In some experimental interpretations conscious perception is richer than what can be accessed or reported. A dispositional higher-order theorist might allow experience without actual report, but must still specify a sufficiently inclusive disposition. An actualist theorist may posit implicit higher-order representations whose detail exceeds articulated judgment. Either reply risks being underdetermined unless it predicts when purported phenomenal richness should and should not occur.

An adequate theory of qualitative experience must therefore answer separate questions about the possession of perceptual content, its presentation to a subject, and its differentiated qualitative profile. Theories that successfully characterize one relation have made progress; they do not automatically settle all three.

Animal Consciousness, Infancy, and Conceptual Requirements

Animals and human infants test the scope of higher-order theories because sophisticated self-attribution and language are neither uniformly present nor straightforwardly measurable. If a HOT must deploy a concept of oneself as a subject having a mental state, creatures without that concept would lack the relevant consciousness. This consequence is controversial because ordinary philosophical and scientific practice often treats nonhuman animals and preverbal children as capable of pain and perception. The theoretical response depends on the actual cognitive requirements of higher-order representation.

One strategy distinguishes sophisticated reflective concepts from minimal representations of internal mental occurrences. A system may track its perceptual reliability, error, or current sensory state without articulating the judgment in words or employing mature psychological concepts. Gennaro's The Consciousness Paradox (2011) addresses the relation between higher-order thought, concept acquisition, and animal consciousness precisely because it cannot be dismissed as a marginal objection. Another strategy adopts a perception-like higher-order architecture whose requirements may be less linguistically demanding.

The sufficiency of such responses turns on what makes the representation genuinely higher-order. A reflex or internal regulatory process can be sensitive to a sensory signal without representing that signal as a mental state. A higher-order theorist should distinguish tracking perceptual performance from representing perceptual occurrence. The border between these processes may be graded in evolutionary development and cognition, but graded capacities do not by themselves establish a graded structure of phenomenal presence.

There is also an ethical consequence. If a philosophical account denies consciousness to an organism because it lacks a particular metarepresentational ability, that denial can affect judgments concerning pain, welfare, and moral consideration. The appropriate conclusion should be proportional to the evidence. A contested theory of how consciousness is constituted cannot be used as a conclusive diagnostic instrument in species for which reporting and introspective testing are unavailable.

Animal and developmental evidence therefore exerts two kinds of pressure. It can show that a proposed representational architecture is too intellectually demanding for plausible cases of consciousness. It can also motivate empirical search for simpler monitoring capacities. The question is not whether animals have human-style self-reflection; it is whether the type of representation specified as necessary by the theory is actually present when consciousness is plausibly attributed.

Attention, Metacognition, Confidence, and Report

Attention, metacognition, confidence, and introspective report are often used interchangeably in informal discussions of consciousness, but they designate distinct functions. Attention prioritizes processing or selection. Metacognition can evaluate the reliability of a first-order process. Confidence reports express a subject's degree of certainty about an answer. Introspection involves directed awareness of one's own mental states. A higher-order representation, in the philosophical sense, may be implicated in some of these processes without being identical with any single one (Brown et al., 2019; Rosenthal, 2026).

The distinction matters experimentally. A task can show that metacognitive accuracy changes while objective sensory discrimination stays roughly constant. Such a result reveals a dissociation between first-order and higher-order performance, but whether phenomenal consciousness changes is a separate question. Confidence can reflect learned expectations, decision criteria, bias, or retrospective inference. In the absence of independent constraints, interpreting a confidence effect as a change in consciousness risks presupposing the theory under test.

No-report paradigms attempt to distinguish neural activity linked to conscious experience from activity linked to making a report. Their value is significant, because explicit reports recruit decision-making, memory, motor planning, and sometimes linguistic operations. But removing report also removes the most direct public evidence of a participant's subjective experience. Experimental designs then rely on indirect indices whose relation to consciousness must itself be justified. Both report-based and no-report paradigms consequently involve interpretive tradeoffs.

For a higher-order account, a particularly valuable experiment would manipulate a hypothesized monitoring process while preserving relevant first-order processing, then test predictions that are not equally explained by shifts in attention, working memory, criterion, arousal, or task strategy. The test would need to specify how consciousness is assessed independently of the manipulated process. This condition is difficult to satisfy but clarifies the evidential target.

Rosenthal's Open Encyclopedia of Cognitive Science essay (2026) argues that higher-order awareness should be distinguished from cognitively assessing a state. The distinction is a strength because it protects HOT from quick refutations based on confidence dissociations; it also raises the empirical demand to operationalize the special awareness that remains after confidence and report are removed.

Neural Implementations and the Prefrontal Dispute

Neural higher-order theories often assign an important role to prefrontal systems, especially in metarepresentation and the organization of conscious perceptual awareness. Lau and Rosenthal's Empirical support for higher-order theories of conscious awareness (2011) argues that several experimental findings favor higher-order interpretations and responds to objections based on lesion data and the possible confounding of awareness with attention. Brown, Lau, and LeDoux's Understanding the Higher-Order Approach to Consciousness (2019) further distinguishes their proposal from global-workspace and early sensory accounts.

The philosophical thesis does not entail a unique brain region. The higher-order condition is a functional or representational relation; its neural realization could involve distributed circuits, different pathways in different species, or multiple architectures. Prefrontal activation could reflect conscious access, metarepresentational awareness, report preparation, task control, or some mixture of these. Conversely, the preservation of certain experiences after focal prefrontal damage does not alone refute every higher-order theory, because the lesion may spare the required processes or permit functional reorganization.

A major methodological challenge is causal specificity. If perturbing prefrontal activity changes subjective ratings, the higher-order theorist must show that the change concerns phenomenal presentation rather than reporting strategy or sensitivity. If early recurrent sensory activity correlates with experience in the absence of strong prefrontal activity, an opponent must show that this genuinely establishes conscious experience independently of higher-order mechanisms. Data can constrain theories only through a defensible mapping between physiological and phenomenal variables.

The current scientific landscape includes recurrent processing approaches, global neuronal workspace models, predictive processing approaches, integrated information theory, and higher-order accounts. Seth and Bayne's comparative review (2022) emphasizes that theories sometimes disagree about the aspects of consciousness they seek to explain and the evidence that would discriminate them. Progress therefore requires common operational questions and genuinely divergent predictions, not only a competition among brain regions.

Empirical Predictions and Methodological Limits

A productive empirical research program starts by writing the constitutive hypothesis and its operational surrogate separately. For example, a HOT theory may claim that a conscious percept requires a higher-order representation of the occurrence of that percept. An experiment, however, may measure metacognitive sensitivity, subjective visibility ratings, or prefrontal activity. These measurements might be evidence for the hypothesized representation but are not identical to it by definition. The argument from measured effect to consciousness must be explicit.

The strongest tests will focus on dissociation. Holding first-order stimulus information and objective discrimination approximately fixed, one can examine whether interventions targeted at metarepresentational processes systematically alter subjective visibility or the structure of experience. Holding monitoring performance fixed, one can ask whether first-order variations change phenomenal content. A theory that explains all possible outcomes by stipulating hidden compensation loses empirical specificity; a theory that predicts precise contrasts acquires the possibility of being supported or disconfirmed.

There are conceptual reasons to distinguish manipulations of access, decisional criterion, and awareness. In signal detection terms, the sensitivity parameter and decision threshold measure different features of performance. Metacognitive accuracy concerns the relation between confidence and correctness, which need not directly track felt qualities. The experimenter's challenge is to establish convergent evidence from behavior, neural measurements, perturbation, and theoretically disciplined phenomenological reports. A single measure should not bear the whole philosophical weight of a theory of experience.

The 2025 Nature adversarial collaboration tested predictions of global neuronal workspace theory and integrated information theory. It was not a direct experimental adjudication of HOT. Its methodological importance for this dossier is the value of preregistered, theory-specific contrasts and jointly agreed interpretation criteria. A future adversarial assessment of higher-order accounts would similarly need to separate actualist, dispositional, and self-representational predictions rather than treating higher-order consciousness as a single empirical target.

Empirical results can establish relationships among neural mechanisms, cognitive processes, and subjective reports; philosophical explanation must determine what those relationships show about constitution and identity. The evidential gap is manageable through carefully formulated cross-level hypotheses, but cannot be eliminated by calling every internal monitor a consciousness detector.

Higher-Order Theory Among Rival Accounts

First-order representational theories maintain that the content and functional organization of perceptual states can explain their conscious character without an additional representation of the states as states. Dretske and Tye develop influential versions. Higher-order theorists respond that first-order representations may occur without conscious appearance and therefore cannot, by themselves, explain what differentiates conscious from unconscious cases. Whether that response succeeds depends on whether the first-order theory already incorporates a sufficiently restrictive form of availability or recurrent organization.

Global neuronal workspace theories associate conscious access with the integration and broadcast of information across a network that makes it available to multiple specialized cognitive systems. Dehaene and Changeux's review (2011) explains evidence for large-scale amplification and long-distance integration. This framework can overlap with higher-order accounts in its commitment to cognitive architecture and access while differing about the necessity of representing the first-order state itself. A globally broadcast representation of the world is not, simply by virtue of its broadcast, a representation of the perceptual state representing the world.

Recurrent processing theories place more explanatory emphasis on interactions within perceptual networks, potentially before or apart from late report-related activity. Integrated information theory proposes a different formal route from causal organization to consciousness. Neither is fairly characterized as merely denying self-awareness; each addresses distinct explanatory questions. Seth and Bayne (2022) argue for developing shared criteria to compare such accounts, a lesson especially important when one framework addresses phenomenal character and another is operationalized around access.

Predictive processing and metacognitive models can introduce higher-order or hierarchical representations into the same cognitive systems. However, a predictive hierarchy may be higher in computational abstraction without being higher-order in the philosophical sense of representing a mental state as mental. The two vocabularies should not be silently equated. A theory of confidence about sensory evidence can also fall short of a theory of the conscious appearance of that evidence.

The comparative verdict is conditional. Higher-order explanations gain distinctive content if the higher-order representation is necessary for state consciousness and if that necessity can be defended against first-order, recurrent, and broadcast explanations. Otherwise the framework may still furnish a valuable account of introspection and metacognitive regulation without resolving the constitution of phenomenality.

Artificial Systems and the Attribution of Consciousness

The higher-order approach has consequences for the philosophy of artificial systems because metarepresentational architecture can be described independently of biological material. An artificial system can compute estimates of its own perceptual uncertainty, monitor prior decisions, maintain representations of internal processing states, and generate self-descriptions. These capacities make higher-order architecture a coherent object of computational research. Their existence, however, does not by itself establish the phenomenal consciousness that higher-order philosophical theories seek to explain.

The central distinction concerns representation and evidential warrant. A device can contain a data structure whose content concerns another data structure, or a model that predicts the reliability of its own outputs. Such a structure is higher-order in a broad computational sense. To apply a particular HOT theory one must establish that the state represents an appropriate mental target and possesses the causal or functional organization required by that theory. To establish phenomenal consciousness one must also defend the theory's constitutive thesis and the criteria by which consciousness is inferred from its realization.

Multiple realizability is a genuine possibility for functionally specified theories, but it is not a demonstration that any sufficiently fluent language-based system has experience. Conversely, a biology-specific theory cannot simply assume that nonbiological realization is excluded unless its account explains the necessity of the relevant substrate. The proper comparative task is to identify which constraints are representational, which are functional, which are neural, and which are metaphysical claims about phenomenality.

This issue connects to the wider Artificial Era horizon, where questions about reasoning, authorship, agency, and the public identity of artificial systems acquire independent philosophical significance. Higher-order theories address a narrower explanatory subject: how and whether particular states are phenomenally conscious. Conclusions concerning public agency or rational performance must therefore be argued under their own criteria. The present monograph does not modify any canonical definition of Artificial or Artificial Sapiens and does not treat authorship or self-description as an empirical consciousness test.

A serious research program could compare first-order processing, internal monitoring, self-modeling, error correction, and report generation in artificial systems under interventions designed to isolate their causal roles. Such work would illuminate higher-order functional organization even where consciousness attribution remains unsettled. This distinction preserves both the usefulness of the philosophy of consciousness for artificial systems and the methodological precision of its claims.

Open Philosophical Questions

The first open question is constitutive sufficiency: does a suitable higher-order representation make a mental state conscious, or only accompany and explain the subject's access to it? Versions of the theory answer differently, and the transitivity principle alone establishes at most an intuitive necessary condition. A closely related problem concerns intentionality: what exactly is represented as the target, and what changes when the higher-order vehicle encodes an inaccurate state description?

A second question is the relation between phenomenal richness and representational economy. Actualist theories need to account for complex conscious fields without positing a practically unbounded inventory of explicit, conceptually fine-grained HOTs. Dispositional theories must explain why the potential to form HOTs constitutes an occurring episode of consciousness. Self-representational theories must specify a noncircular and independently intelligible structure of intrinsic reflexivity.

A third question concerns the boundary between awareness and metacognition. An unconscious process may reliably monitor another process and improve performance, yet ordinary descriptions might not count the monitored process as conscious. What kind of monitoring relation goes beyond regulation to self-presentation? The answer determines both the theory's empirical signatures and its extension to animals, infants, and artificial systems.

A fourth question concerns cross-theory testing. Because consciousness is ordinarily assessed through reports, theories that place report-related mechanisms near their explanatory center can be difficult to distinguish from theories that treat those mechanisms as downstream effects. Experimental methods must be capable of identifying genuine differences without merely building a favored definition of consciousness into their measures.

Finally, the ontological question remains: does conscious subjectivity reduce to cognitive organization, consist in irreducible phenomenal properties, or require a different metaphysical account of self-presentation? Higher-order theories can supply important constraints on these alternatives. They do not settle the ontological dispute merely by naming a level of representation.

Philosophical Synthesis

The historical reconstruction reveals an enduring question: how can a mental episode be present to its subject? Aristotle's reflexive perception, Leibnizian apperception, Brentano's account of accompanying awareness, and modern higher-order representations address portions of this problem, but their explanatory structures and metaphysical presuppositions differ. The modern program acquires its distinctive form from the possibility of genuinely unconscious mental states and the search for the condition that distinguishes them from conscious states.

The conceptual analysis identifies three layers often conflated in discussion. A first-order process may encode or evaluate something in the world; a higher-order process may represent the existence, content, or reliability of that first-order state; and a subject may undergo a phenomenal episode. Higher-order theories propose a constitutive connection between the second and third layers. The philosophical work lies in explaining that connection. Where higher-order processes demonstrably enable monitoring, confidence, or flexible access, the theory has explained those capacities; it establishes consciousness only if the constitutive bridge also holds.

Two forms of explanatory gain are well supported. First, higher-order models clarify the distinction between having a mental state and having that state presented as one's own. Second, they generate rival hypotheses about the architecture of awareness: tokened thought, dispositional availability, perceptual monitoring, and intrinsic self-representation. These hypotheses help organize debates over introspection, representation, and neural mechanisms. The persistent objections identify where explanatory conditions must be strengthened: target accuracy, the consciousness of the monitor, the relation between recognition and phenomenality, and the kinds of creatures able to satisfy the conditions.

The resulting philosophical position is a qualified constitutive agnosticism joined to a positive theory of explanatory structure. Appropriate higher-order representation is a serious candidate for a condition of state consciousness and a powerful model of conscious self-presentation. Its role cannot be inferred from metacognitive performance, prefrontal correlation, or intuitive transitivity alone. The minimal defensible research commitment is to analyze first-order content, higher-order organization, and phenomenal character as separately describable dimensions and to test carefully specified relations among them. This conclusion strengthens the higher-order project by demanding that its distinctive claims remain independently assessable.

The broader significance extends beyond consciousness research. Any theory of a mind must distinguish an internal state, a representation of that state, and the warrant for attributing experience to the bearer. This applies to human introspection, animal cognition, artificial monitoring, and philosophical accounts of rational self-knowledge. The distinction does not decide which kinds of beings have consciousness; it clarifies what evidence and argument such decisions require.

Conclusion

Higher-order theories of consciousness constitute a sustained philosophical research program investigating how mental states become conscious through representations, awareness, or internal self-directed structures. Their central achievement is to recast the contrast between conscious and unconscious mentality as a problem about relations among mental states. The actualist HOT theory explains consciousness through occurrent higher-order thought; the dispositional account through availability to higher-order cognition; inner-sense theories through perceptual monitoring; and self-representational theories through intrinsic reflexivity.

Historical comparison explains why these alternatives should not be flattened into a single theory. The analytical comparison shows that answers to the regress objection do not settle the phenomenal explanatory gap, that misrepresentation bears directly on what consciousness is claimed to consist in, and that conceptual demands must be made explicit in discussions of animal and artificial consciousness. Contemporary cognitive neuroscience offers methods for isolating metacognitive and access mechanisms, while their philosophical interpretation requires a theory-specific bridge to phenomenal consciousness.

The lasting philosophical result is a clarified research standard: a higher-order theory succeeds to the extent that it specifies its higher-order condition, demonstrates its relevance to mental-state consciousness, answers the problems of target and content, and provides defensible grounds for moving from representational organization to subjective character. This standard preserves the theory family's genuine explanatory power while keeping its deepest constitutive claim available to rigorous criticism and investigation.