Does FC operationalize Higher-Order Thought (HOT)?
The Short Answer
Yes — and it's the cleanest correspondence in the paper.
HOT says a state is conscious when it becomes the target of a higher-order representation available to the system's reasoning. FC's Definition 2 requires exactly that: a self-model variable mᵢ represents internal state sᵢ and must be available to global reasoning. Under mild assumptions, a state contributes to FCS > 0 if and only if it is HOT-conscious — making FC a quantitative formalization of HOT's binary criterion. It even naturally handles recursive "thoughts about thoughts" through meta-cognitive self-models, explaining why our introspection is finite rather than an infinite regress. HOT tells you which states are conscious; FC tells you how much.
Is it valid to say FC "implements" HOT?
Pretty much yes, and more cleanly than for the other four theories. HOT's core claim is that a mental state is conscious when it is the target of a suitable higher-order representation — i.e. there exists a meta-state that represents the first-order state. FC's Definition 2 requires self-models to make internal states available to global reasoning, which is structurally identical: the self-model is the higher-order representation, and "available to global reasoning" is the suitability condition. The mapping is direct enough that "implements" is defensible.
A sketch of a formal equivalence
Assumptions:
- A mental state s is HOT-conscious iff there exists a higher-order representation h(s) that represents s and is itself available to the cognitive system's reasoning processes (standard Rosenthal formulation).
- A self-model variable mᵢ in FC represents system state sᵢ with mutual information I(mᵢ; sᵢ) > 0.
- "Available to global reasoning" in FC (Definition 2) is equivalent to "available to the cognitive system's reasoning processes" in HOT assumption 1.
Claim: Under these assumptions, a state sᵢ contributes to FCS > 0 if and only if it is HOT-conscious.
Proof (sketch):
(→) If sᵢ contributes to FCS > 0, then by Definition 3 there exists a self-model mᵢ with I(mᵢ; sᵢ) > 0 and mᵢ is available to global reasoning (Definition 2). By assumption 2, mᵢ represents sᵢ. By assumption 3, mᵢ is available to the reasoning system. So mᵢ satisfies the role of h(s) in assumption 1 — sᵢ is HOT-conscious.
(←) If sᵢ is HOT-conscious, then by assumption 1 there exists h(sᵢ) representing sᵢ and available to reasoning. By assumption 3, h(sᵢ) qualifies as a self-model variable mᵢ under Definition 2. By assumption 2, I(mᵢ; sᵢ) > 0, so sᵢ contributes to R > 0. If P > 0, then FCS > 0.
What about recursive higher-order states?
HOT theorists like Rosenthal allow for recursive higher-order states — a thought about a thought about a thought — and FC handles this naturally through the meta-attention and meta-self-awareness self-models in the SBR catalog. A system that models its own self-modeling process has a second-order self-model with its own R and P, which feeds back into global reasoning and increases FCS. If that second-order model is itself modeled, you get a third-order contribution — and so on. In practice, each recursive layer adds diminishing returns (the mutual information at each level is bounded by the layer below it), which is why FC predicts that biological consciousness tops out at a small number of meaningful meta-cognitive levels rather than infinite regress — consistent with what we actually observe in human introspection.
Conclusion
FC operationalizes exactly the states HOT deems conscious, with R measuring the richness of the higher-order representation and P measuring the reasoning power that makes it consequential. HOT tells you which states are conscious; FC tells you how much.
The main caveat worth flagging: HOT is typically formulated about token mental states in biological systems, while FC is a graded metric over architectures. The equivalence holds at the structural level but you'd want to be careful claiming it extends to phenomenal consciousness — which HOT sometimes does and FC explicitly doesn't.