Pearce's scheme is unusually compact, which is exactly what makes this question tractable. His system runs on a handful of primitives: valence as a real scalar he calls "the world's inbuilt metric of disvalue" ( Worthy Successor interview); motivation as gradient-climbing on that axis ( hedweb); suffering as a set-point parameter to be recalibrated; phenomenal binding as the on/off switch of consciousness; and "non-materialist physicalism," on which classical digital systems are zombies with "no more moral status than a rock" ( Sentient Futures 2026 slides). Because the axiom set is so small, the questions it cannot formulate sit precisely at each primitive. Here are the ones I'd locate.

1. Defeasibility: "suffering is bad" vs. "only suffering is bad." Pearce defends a lexical negative utilitarianism — no amount of bliss compensates for suffering ( William Joy interview). The missing distinction is between pro tanto and ultimate badness, and with it, the whole apparatus of moral epistemology: contrastive testing, moral uncertainty, defeasibility. The question he cannot ask is: "What conceivable observation would show that some non-valenced fact matters morally — and what would my scheme even count as evidence of that?" Since every candidate consideration must be cashed in experiential terms before it enters the ledger, the axiology is self-sealing. He treats the badness of suffering as perceptual, like seeing red — which forecloses asking what would make it revisable.

2. Translation vs. projection: the missing geometry of valence. His welfare talk is one-dimensional: hedonic ranges like "-10 to +10" becoming "+70 to +100" ( Sentience Research interview). Missing: any vocabulary for the dimensional structure of affect — arousal, savoring, psychological richness, calm vs. energetic bliss, awe as a distinct axis — and for incomparability between total states. The unaskable question: "When recalibration lifts everyone above hedonic zero, is that a translation into a richer region of the same space, or a projection of a high-dimensional affect manifold onto a line — and which goods have no coordinates once projected?" Two post-Darwinian minds with different bliss-textures might be neither better nor worse off than one another; his arithmetic has no place for that verdict.

3. Transformative experience and the authorship of consent. Pearce concedes his descendants will be "different creatures" whose "bad Darwinian memories will be gone forever," then dismisses the worry ( triple-S essay). What's missing is L.A. Paul's transformative-experience vocabulary: he himself insists superbliss is physiologically unimaginable to current neurochemistry — which means the choice to recalibrate cannot be made from a representation of what is chosen. The question he can't formulate: "When an agent commits an unrepresentable, irrevocable change to its own valence function, what makes it that agent's decision rather than a decision one organism makes about a stranger who will replace it?" His standard reassurance — that uplift "conserves our existing values and preference architecture" ( Worthy Successor interview) — presupposes precisely what's in doubt: that preference identity survives a global remapping of the currency in which preferences are denominated.

4. Moral status without a unitary subject. His scheme is binary: phenomenal binding yields a subject of experience; its absence yields a rock ( Sentient Futures slides). Yet his own monism makes micro-experience ubiquitous. The missing distinction is between phenomenal unity and moral patienthood. Unaskable questions follow: Can there be welfare without a unitary subject — partial binding, dissociated systems, hive-minds, dreamers under anesthesia? What do we owe systems that are neither persons nor rocks? His zombie verdict on digital AI forecloses an entire empirical research program (AI welfare science) as a category error rather than treating it as an open question with stakes.

5. Legitimacy, not just consent. Pearce says "the biggest obstacles to a cruelty-free world are ideological" ( 2011 interview) — opposition to abolition is a cognitive error awaiting correction. What's absent is the reciprocal, reflexive concept: that abolitionist policy could itself function as ideology. Missing vocabulary: power, biopolitics, second-order autonomy, non-domination. The blocked question: "What decision procedures could make the collective calibration of everyone's reward function legitimate rather than merely consensual or technically successful — and who audits the architects of bliss?" His toolkit addresses feasibility and justice; it has no register in which to raise legitimacy.

6. Homology vs. analogy of affect across phyla. "Abolishing suffering in all sentient life" treats suffering as a single natural kind with one currency across taxa — a "cross-species welfare state," predator reprogramming, additive moral arithmetic ( Wikipedia). Missing: the evolutionary distinction between homology and homoplasy. The unaskable question: "Are nociceptive-affective systems in cephalopods, insects, and vertebrates the same phenomenon at different intensities, or analogous implementations whose badness may be differently structured — and if the latter, what exactly is being summed when we compute total suffering?" His aggregation presupposes cross-species comparability while lacking any metric theory that could establish it.

7. The indexicality of abolition in a block universe. He predicts "the world's last unpleasant experience will be a precisely dateable event" ( hedweb), yet his physics flirtations (time-symmetric interpretations, an experiential block-world ontology) are tenseless. Missing: the A-theory/B-theory distinction, and with it any account of tensed value. The blocked question: "In a block universe, is suffering ever abolished, or merely confined to a coordinate region — and does 'abolished' name a fact about the world or an indexical remark made from inside a slice?" "Progress" may be a property of maps, not of the manifold.

8. Value without a valuer. In his scheme, value just is valence — by definition, not argument. So the question "Is anything lost when a beautiful but unperceived pattern is destroyed?" is not answered but unaskable: the term "value without experience" has no entry in his lexicon. He would bite the bullet if the question were forced on him from outside; the blind spot is that the scheme itself can't generate it.

Two closing observations. First, notice the shared signature: in every case a definition is doing the work of an argument (badness = negative valence, status = binding, value = experience, progress = later coordinate). The single most generative question for the scheme is the one it structurally cannot ask: of each primitive, what would you count as a counterexample? Second, to be fair, the scheme already creaks at its edges in Pearce's own hands — he muses about "the risks of premature amnesia" of suffering ( Sentience Research interview) and now asks "what could go wrong" ( Sentient Futures slides) — which suggests the boundary of the askable is already shifting. A tight conceptual scheme has this advantage: unlike most thinkers', its blind spots can be located to the millimetre. If you are the David Pearce in question, I'd happily stress-test any one of these — or try to build the missing vocabulary and see which questions survive translation into terms you'd accept.