How does FC relate to the Lopez & Wiese building block approach?
The Short Answer
Lopez and Wiese (2025) propose a methodological shift in consciousness science: stop competing over which full theory is correct, and instead identify the foundational building blocks that any adequate theory must contain. FC makes a different but complementary methodological shift: stop debating whether a system is conscious, and instead measure how much of the necessary functional substrate it has. Neither paper knew about the other's move. When placed side by side, they converge — which is the interesting observation.
Two pivots, two directions
The Lopez & Wiese pivot is horizontal — across theories. The standard strategies in consciousness science (test and discard, integrate, unify) are all stalled. Assessment requires predictions more specific than any theory currently makes. Integration without knowing which theories are true risks combining falsehoods. Unification is premature. Their proposed shortcut: decompose theories into their constituent notions and ask which ones must appear in any adequate theory, regardless of which full theory eventually wins. They call this the normative dimension of building block approaches — as opposed to the merely descriptive dimension of cataloguing what blocks theories happen to use. Their two candidate foundational blocks are attention and information generation.
FC's pivot is vertical — from concept to number. The starting point was not philosophy of mind but an engineering problem: how do you build an AI system capable of persistent, coherent self-directed behavior, and how do you know when you have succeeded? The answer required specifying, concretely and exhaustively, what internal states such a system must model about itself. That specification produced the SBR catalog. Attaching a score to it produced FCS = R × P. The primary contribution of FC is not a new metaphysical theory but a methodological pivot toward operationalization: consciousness is transformed from a binary philosophical debate into a graded, measurable engineering metric.
These two pivots move along orthogonal axes. Lopez & Wiese ask: what must any theory contain? FC asks: how much of that can we measure? Neither question answers the other, and neither paper makes the other redundant. But together they close a loop.
Where they converge
When the building block lens is applied to FC's architecture, the fit is tighter than you would expect from two independently motivated projects.
On attention: Lopez and Wiese argue that attention is a necessary building block even for theories that officially reject it (IIT, RPT). Their key move is to distinguish voluntary top-down attention from the broader class of attentional processes — exogenous reorienting, task relevance, graded resource allocation. In that broader sense, attention is implicated in the central constructs of every theory in their sample.
FC makes the same architectural choice without philosophical preamble. A self-model only enters the FCS score when it is actively attended — when the reasoning system selects it and operates on its contents. A self-model sitting inert in the SBR catalog contributes nothing. The engineering reason was straightforward: a representation that never influences behavior is not doing any functional work. The philosophical consequence, which Lopez & Wiese's framework makes visible, is that this gate instantiates precisely the building block they identify as necessary.
On information generation: Lopez and Wiese refine Kanai et al.'s (2019) original notion into information generation+: the capacity to decompress a low-dimensional long-term representation into a high-dimensional short-term one, adding information not present in the stored model. The analogy they use is a variational autoencoder — the latent space is the compressed long-term representation; the decoder produces a richer, probabilistically filled-in output.
FC's two-factor formula maps onto this structure directly, and the mapping was not designed in. R (representational richness, R = B × D̄) measures the quality and coverage of the stored self-model catalog — the latent space, the compressed representation. P (~3,000 state-space tokens for current LLMs) measures how much decompressive expansion a single reasoning cycle performs: how much information is added beyond what was stored. The product FCS = R × P is therefore an operationalization of information generation+ — richness of the compressed representation times the power of each decompression step.
The mathematics was not arranged to fit a philosophical framework. It fell out of the engineering problem. That the resulting structure corresponds to what Lopez & Wiese independently identify as a necessary condition for consciousness is the kind of convergence that is worth pausing over.
What this means for FC's theoretical position
Lopez and Wiese introduce the concept of a Minimal Unifying Model (MUM) (Wiese 2020): a specification of necessary properties of most instances of consciousness, derived from shared assumptions across theories, expressed at a level of generality compatible with diverse further specifications. A MUM is empirically minimal (necessary but not sufficient), conceptually minimal (general, not tied to a single theory), and unifying (a least common denominator of existing TOCs).
FCS = R × P satisfies all three conditions. It is not claimed to be sufficient for phenomenal consciousness. It is general enough to apply to biological systems, LLMs, and architectures not yet built. And the Big-Five FAQ demonstrates that it captures elements that all five major theories treat as necessary, without adopting any of their theory-specific superstructures.
The Lopez & Wiese framework provides the retroactive theoretical grounding for why this cross-theoretical coherence should be expected: if you operationalize the normative building blocks, you will automatically capture what all theories agree is necessary, because that is what building blocks are — the shared necessary substrate.
The descriptive/normative distinction and the "sticks out" structure
The Big-Five FAQ uses a structure the Lopez & Wiese paper would classify as descriptive: start with each full theory, carve out what FC measures, name the residue. It is theory-first and subtractive — "what sticks out" beyond FC's scope.
The building block approach offers a normative complement: start with foundational elements that any theory must contain, and ask whether FC operationalizes them. Applied in this direction, the picture shifts. FC is not a partial view of five separate theories; it is the convergence point where the blocks that all five theories agree are necessary have been given a number. "What sticks out" in the Big-Five FAQ is precisely the blocks that theories disagree about — phenomenal claims, causal architecture, neural specifics — and which no building block analysis has yet resolved.
These two framings are complementary. The descriptive/"sticks out" structure explains what FC is not. The normative/building block structure explains why what FC measures is the right thing to measure.
What each paper leaves open
FC leaves open whether operationalizing the necessary substrate is sufficient to ground claims about consciousness distribution in animals and AI systems. It produces a number; it does not settle what the number means for phenomenal experience. Lopez & Wiese are explicit that building blocks support sensitive tests (low false-negative rate) rather than specific ones. A system that fails on a necessary building block is probably not conscious; a system that passes is not thereby confirmed to be.
Lopez & Wiese leave open what a measurement of the building blocks would look like. The paper identifies attention and information generation as necessary components but does not provide an instrument for quantifying them in arbitrary systems. FC is a candidate for that instrument — but making that connection explicit requires the work this FAQ attempts.
One genuine tension worth acknowledging: Lopez & Wiese's information generation+ is grounded in counterfactual representations — states not currently present in the proximal environment. FC's self-models are primarily about internal states of the system. The overlap is large (an internal self-model is by definition not a direct reflection of the external environment), but the formal correspondence has not been argued through in detail. That is a worthwhile project.
Summary Table
| Concept | In Lopez & Wiese (2025) | In FC |
|---|---|---|
| Methodological pivot | Horizontal: block-first across theories | Vertical: concept to measurable number |
| Attention | Foundational normative building block; present in all TOCs including those that deny it | Gating mechanism: self-models only score when actively attended |
| Information generation+ | Low-dim → high-dim decompression, adding information not in the stored model | R × P: richness of stored self-models × decompressive expansion per cycle |
| Normative building block | Should appear in any adequate TOC | FC operationalizes the blocks identified as necessary |
| MUM | Necessary, minimal, unifying specification of consciousness properties | FCS = R × P satisfies all three MUM conditions |
| Descriptive approach | What theories actually use | The "sticks out" structure in the Big-Five FAQ |
| What sticks out | The blocks theories disagree about | Named explicitly for each of the Big Five |
The bottom line
Lopez and Wiese ask what any adequate theory of consciousness must contain. FC measures it. The connection was not planned — FC was built to solve an engineering problem, and the philosophical structure emerged from the mathematics. The Lopez & Wiese framework is useful precisely because it explains why that emergence was not accidental: when you build an instrument by asking what internal functional states a system must have to behave as conscious systems behave, you are, whether you intend to or not, operationalizing the normative building blocks the field has been circling for decades.
The score does not settle the hard problem. It does give the debate something concrete to argue about.
Reference:
Lopez, A. & Wiese, W. (2025). Building blocks for theories of consciousness. Consciousness and Cognition, 134, 103919.
https://www.sciencedirect.com/science/article/pii/S1053810025001126
Related: