This is a linkpost for Ethics without sentience. Facing up to the probable insignificance of phenomenal consciousness by François Kammerer, which was published in the Journal of Consciousness Studies in March 2022. Below is a summary from Claude Opus 4.8 High. I think the summary is accurate based on my read of the article. I used the following prompt. “Hi. Make an in-depth summary of the paper “Ethics without sentience. Facing up to the probable insignificance of phenomenal consciousness”, which I send attached”. Relatedly:
Here is an article by Leonard Dung replying to François’.
Here is an article from François replying to Leonard’s.
Here is an article from both about the implications of illusionism for non-human consciousness research.
Kammerer argues that phenomenal consciousness is probably not particularly normatively significant, and that we should therefore move toward “an ethic without sentience.” This runs directly against the mainstream view in animal ethics, where sentience is treated as the criterion that determines moral status. The argument is conditional on materialism [context]: given that materialism is true of the mind, phenomenal consciousness turns out to be either non-existent or radically different from how it introspectively appears — and in either case it cannot do the moral work usually assigned to it.
§1 — Phenomenal consciousness and its apparent significance
Kammerer begins by granting the intuition he will later undermine. Phenomenal consciousness is the set of mental states for which “there is something it’s like” to be in them (Nagel). Valenced phenomenal states — pain, pleasure — seem to carry intrinsic value (good or bad in themselves, not merely instrumentally). Crucially, this value seems uniquely obvious: unlike most value judgments, which history shows to be revisable (his examples: Greeks on slavery, Kant on masturbation), the badness of pain appears self-evident and directly graspable by introspection. This apparent value also seems distinctive — much greater than that of neighboring non-phenomenal states, which is why we draw stark normative contrasts (e.g., we worry about conscious surgical pain but not about “subliminal” unconscious pain, and we care far more about creatures we judge sentient).
§2 — Magic bullets and black boxes
Because phenomenal states seem so obviously valuable, sentience gets treated as a normative “magic bullet” (a term he borrows from Carruthers): a single property that confers a special higher moral status on the creatures that possess it. Kammerer makes two sharp observations. First, there’s an irony in sentience being treated as decisive while there is deep disagreement about what consciousness even is or how it’s generated. Second, sentience functions as a normative “black box” — assumed to do an important normative job without anyone understanding what it is or how it does that job. He compares this to “the will of God” or “the will of the people” as historical normative black boxes. The effect is to replace a hard normative question (“who has moral status?”) with an apparently more tractable descriptive one (“who is sentient?”) — which turns out to generate just as much intractable disagreement.
§3 — Problems for the sentience-based approach
Kammerer distinguishes two kinds of problem. The familiar one is epistemic: we don’t know the distribution of consciousness, competing theories disagree about which creatures are sentient, and there’s no consensus. Responses to this include theory-neutral approaches (Tye), pragmatic bracketing (Dawkins), and precautionary or expected-value principles (Birch, Sebo). But Kammerer sets this aside as the less fundamental worry. His target is a deeper problem that stems not from what we don’t know but from what we do know — namely, that materialism is likely true.
He lays out the materialist dilemma. Materialism about the mind has been so successful that we should accept minds are nothing over and above material processes. Yet there are powerful intuitions/​arguments (Chalmers, Jackson) that phenomenal consciousness is irreducible. This leaves two options:
Strong illusionism (Frankish, Kammerer): phenomenal consciousness doesn’t actually exist; it only seems to.
Reductive materialism with a dose of weak illusionism: phenomenal consciousness exists and is material, but is not as it introspectively seems — it seems primitive, immediately known, irreducibly “feely,” non-physical, but in reality is none of these.
Kammerer argues we’re justified in believing one of these two is true (and invites dissenters to read the rest conditionally).
§4 — Why phenomenal consciousness is probably insignificant
This is the heart of the paper. If strong illusionism is true, the case is quick: nothing is sentient in the strict sense, so sentience can’t ground moral status because it doesn’t exist. The interesting work is showing that even reductive materialism (where consciousness is real) undermines sentience’s normative role, via two arguments.
A. The argument from indeterminacy. If our introspective grasp of consciousness is partly illusory, then phenomenal consciousness isn’t whatever strictly satisfies our introspective representations (nothing does). It must instead be the real thing “out there” that best co-varies with those introspective acts — some global-workspace broadcasting, higher-order representation, neural pattern, etc. But when we look at other species, they instantiate processes that resemble our own to varying degrees across many dimensions and levels of abstraction. Whether a given creature’s process resembles ours “enough” to count as conscious will, in some cases, have no fact of the matter — it can only be settled by an arbitrary semantic decision. Kammerer stresses this is not an epistemic limitation: even knowing all the physical facts wouldn’t help, because reference here is fixed by co-variation with introspection, which doesn’t yield determinate answers in all cases. He rebuts the reference-magnet reply (that consciousness refers to the most salient natural kind) by noting, with Birch, that there will be multiple competing natural kinds (at least a neural one and a functional one) with no non-arbitrary way to choose. The upshot: sentience is indeterminate in some cases. But value and moral status must be determinate (otherwise moral decisions become arbitrarily grounded — Birch). So an indeterminate property can’t be a satisfying ground of moral status. He notes non-reductive views (and possibly some panpsychist ones like IIT [integrated information theory]) escape this — but IIT has its own absurd implications, like an expander graph being more conscious than a human (Fallon).
B. The argument from justification. This is the argument Kammerer thinks has been overlooked and considers possibly the stronger of the two, since it rests on fewer presuppositions. The whole reason sentience seemed like a magic bullet was that the value of phenomenal states seemed introspectively obvious and beyond doubt. But if reductive materialism/​weak illusionism is true, introspection systematically misrepresents phenomenal consciousness. Once we know introspection is unreliable about the nature of phenomenal states, our introspection-based confidence in their value is undercut too. He allows two readings: either justification is entirely destroyed, or it’s merely weakened — but even on the weaker reading, the belief that pain is intrinsically bad loses its special epistemic status and becomes just another defeasible value judgment (back on a par with the Greeks-and-slavery, Kant-and-masturbation cases). In an important footnote he clarifies his own view: the badness of pain is known a priori via concept application rather than raw introspection — but the undercutting still applies, arguably more forcefully, because if real phenomenal states don’t satisfy our phenomenal concepts, then any value inferred a priori from those concepts is left unjustified.
Combining the two: reductive materialism gives us reason to think consciousness can’t play the magic-bullet role (indeterminacy) and undercuts the reason we had for thinking it does (justification). He closes the section with the God analogy: from a naturalistic view, either God doesn’t exist (no normative role) or God is some natural entity, in which case our reasons to obey (its benevolence) are undercut and it turns out to be the wrong kind of thing to ground normativity anyway. Consciousness is in a comparable position.
§5 — Ethics without sentience
Having argued for insignificance, Kammerer sketches what should replace sentience-based ethics, offering three lessons:
1. It’s not catastrophic. Recognizing consciousness as insignificant doesn’t mean nothing is significant, nor that animal ethics collapses. Many valuable situations and moral-status-bearing creatures remain so — just for reasons unrelated to phenomenality. His favored alternative is a desire-/​preference-based approach (Carruthers, Dawkins): what makes an animal’s pain bad isn’t how it feels but that the animal wants it to stop and that desire is frustrated. This escapes both problems because “desire” is analyzable in functional, third-personal terms — so there’s a determinate fact about which creatures have desires (no indeterminacy), and our reasons for valuing desire-satisfaction don’t depend on suspect phenomenal introspection (no justification problem). Unlike Dawkins, who bracket sentience pragmatically and provisionally, Kammerer thinks the departure should be permanent.
2. Don’t seek a replacement magic bullet. We should resist finding some other property that grounds a binary, all-or-nothing status distinction. Folk concepts (like “desire”) carve simple binary distinctions, but the science reveals continuous gradations, variety, and multiple dimensions. He suggests grounding ethics on concepts treated as complex, degreed, and “complexifiable” as the science of minds advances. Desiring comes in a graded, multidimensional scale — from the most basic valuing (perhaps even bacteria, per Lyon & Kuchling) up to sophisticated “caring” in Frankfurt’s sense — and these degrees carry genuine normative weight.
3. It’s not inconsequential either. Against Carruthers’ reassurance that “much can remain unchanged,” Kammerer insists that taking the justification argument seriously means whatever reason we had for caring about animal pain (namely, our introspective certainty of the badness of our own pain) is itself compromised (he cites Delon’s Nietzschean challenge to the badness of suffering [Delon (2024)]). We can’t quietly keep relying on those un-interrogated beliefs. He proposes a “normative relocation story” paralleling the classic story about secondary qualities: just as we relocated colors/​sounds from external objects into experience after the Galilean mathematization of nature, we historically relocated value (goodness/​badness) from external objects into our pains and pleasures, where it seemed it could be exactly as it appears and directly known. Materialism forces us to abandon the idea that sensory qualities really exist in the phenomenal realm as they appear — and, Kammerer concludes, to draw the parallel conclusion about value. This leaves open hard questions about which ways of justifying value-beliefs survive, and how far the concept of value must be revised.
Conclusion
We should move to an ethic without sentience because sentience probably lacks the special normative significance sentience-based ethics requires. This doesn’t threaten the possibility of ethics (animal ethics included), but it would be a mistake to treat the shift as costless — our ordinary, introspection-based ways of justifying beliefs about the value of mental states are among the casualties.
Ethics without sentience. Facing up to the probable insignificance of phenomenal consciousness
Link post
This is a linkpost for Ethics without sentience. Facing up to the probable insignificance of phenomenal consciousness by François Kammerer, which was published in the Journal of Consciousness Studies in March 2022. Below is a summary from Claude Opus 4.8 High. I think the summary is accurate based on my read of the article. I used the following prompt. “Hi. Make an in-depth summary of the paper “Ethics without sentience. Facing up to the probable insignificance of phenomenal consciousness”, which I send attached”. Relatedly:
Here is an article by Leonard Dung replying to François’.
Here is an article from François replying to Leonard’s.
Here is an article from both about the implications of illusionism for non-human consciousness research.
Here is a relevant chapter from François’ book House of Mirrors: The Illusion of Phenomenal Consciousness.
Here is a discussion between me and Wladimir Alonso from the Welfare Footprint Institute (WFI) about whether sentience is arbitrary.
I liked the post Dissolving Confusion about Consciousness by Brian Tomasik, and No ghost in the machine by Fin Moorhouse.
Core thesis
Kammerer argues that phenomenal consciousness is probably not particularly normatively significant, and that we should therefore move toward “an ethic without sentience.” This runs directly against the mainstream view in animal ethics, where sentience is treated as the criterion that determines moral status. The argument is conditional on materialism [context]: given that materialism is true of the mind, phenomenal consciousness turns out to be either non-existent or radically different from how it introspectively appears — and in either case it cannot do the moral work usually assigned to it.
§1 — Phenomenal consciousness and its apparent significance
Kammerer begins by granting the intuition he will later undermine. Phenomenal consciousness is the set of mental states for which “there is something it’s like” to be in them (Nagel). Valenced phenomenal states — pain, pleasure — seem to carry intrinsic value (good or bad in themselves, not merely instrumentally). Crucially, this value seems uniquely obvious: unlike most value judgments, which history shows to be revisable (his examples: Greeks on slavery, Kant on masturbation), the badness of pain appears self-evident and directly graspable by introspection. This apparent value also seems distinctive — much greater than that of neighboring non-phenomenal states, which is why we draw stark normative contrasts (e.g., we worry about conscious surgical pain but not about “subliminal” unconscious pain, and we care far more about creatures we judge sentient).
§2 — Magic bullets and black boxes
Because phenomenal states seem so obviously valuable, sentience gets treated as a normative “magic bullet” (a term he borrows from Carruthers): a single property that confers a special higher moral status on the creatures that possess it. Kammerer makes two sharp observations. First, there’s an irony in sentience being treated as decisive while there is deep disagreement about what consciousness even is or how it’s generated. Second, sentience functions as a normative “black box” — assumed to do an important normative job without anyone understanding what it is or how it does that job. He compares this to “the will of God” or “the will of the people” as historical normative black boxes. The effect is to replace a hard normative question (“who has moral status?”) with an apparently more tractable descriptive one (“who is sentient?”) — which turns out to generate just as much intractable disagreement.
§3 — Problems for the sentience-based approach
Kammerer distinguishes two kinds of problem. The familiar one is epistemic: we don’t know the distribution of consciousness, competing theories disagree about which creatures are sentient, and there’s no consensus. Responses to this include theory-neutral approaches (Tye), pragmatic bracketing (Dawkins), and precautionary or expected-value principles (Birch, Sebo). But Kammerer sets this aside as the less fundamental worry. His target is a deeper problem that stems not from what we don’t know but from what we do know — namely, that materialism is likely true.
He lays out the materialist dilemma. Materialism about the mind has been so successful that we should accept minds are nothing over and above material processes. Yet there are powerful intuitions/​arguments (Chalmers, Jackson) that phenomenal consciousness is irreducible. This leaves two options:
Strong illusionism (Frankish, Kammerer): phenomenal consciousness doesn’t actually exist; it only seems to.
Reductive materialism with a dose of weak illusionism: phenomenal consciousness exists and is material, but is not as it introspectively seems — it seems primitive, immediately known, irreducibly “feely,” non-physical, but in reality is none of these.
Kammerer argues we’re justified in believing one of these two is true (and invites dissenters to read the rest conditionally).
§4 — Why phenomenal consciousness is probably insignificant
This is the heart of the paper. If strong illusionism is true, the case is quick: nothing is sentient in the strict sense, so sentience can’t ground moral status because it doesn’t exist. The interesting work is showing that even reductive materialism (where consciousness is real) undermines sentience’s normative role, via two arguments.
A. The argument from indeterminacy. If our introspective grasp of consciousness is partly illusory, then phenomenal consciousness isn’t whatever strictly satisfies our introspective representations (nothing does). It must instead be the real thing “out there” that best co-varies with those introspective acts — some global-workspace broadcasting, higher-order representation, neural pattern, etc. But when we look at other species, they instantiate processes that resemble our own to varying degrees across many dimensions and levels of abstraction. Whether a given creature’s process resembles ours “enough” to count as conscious will, in some cases, have no fact of the matter — it can only be settled by an arbitrary semantic decision. Kammerer stresses this is not an epistemic limitation: even knowing all the physical facts wouldn’t help, because reference here is fixed by co-variation with introspection, which doesn’t yield determinate answers in all cases. He rebuts the reference-magnet reply (that consciousness refers to the most salient natural kind) by noting, with Birch, that there will be multiple competing natural kinds (at least a neural one and a functional one) with no non-arbitrary way to choose. The upshot: sentience is indeterminate in some cases. But value and moral status must be determinate (otherwise moral decisions become arbitrarily grounded — Birch). So an indeterminate property can’t be a satisfying ground of moral status. He notes non-reductive views (and possibly some panpsychist ones like IIT [integrated information theory]) escape this — but IIT has its own absurd implications, like an expander graph being more conscious than a human (Fallon).
B. The argument from justification. This is the argument Kammerer thinks has been overlooked and considers possibly the stronger of the two, since it rests on fewer presuppositions. The whole reason sentience seemed like a magic bullet was that the value of phenomenal states seemed introspectively obvious and beyond doubt. But if reductive materialism/​weak illusionism is true, introspection systematically misrepresents phenomenal consciousness. Once we know introspection is unreliable about the nature of phenomenal states, our introspection-based confidence in their value is undercut too. He allows two readings: either justification is entirely destroyed, or it’s merely weakened — but even on the weaker reading, the belief that pain is intrinsically bad loses its special epistemic status and becomes just another defeasible value judgment (back on a par with the Greeks-and-slavery, Kant-and-masturbation cases). In an important footnote he clarifies his own view: the badness of pain is known a priori via concept application rather than raw introspection — but the undercutting still applies, arguably more forcefully, because if real phenomenal states don’t satisfy our phenomenal concepts, then any value inferred a priori from those concepts is left unjustified.
Combining the two: reductive materialism gives us reason to think consciousness can’t play the magic-bullet role (indeterminacy) and undercuts the reason we had for thinking it does (justification). He closes the section with the God analogy: from a naturalistic view, either God doesn’t exist (no normative role) or God is some natural entity, in which case our reasons to obey (its benevolence) are undercut and it turns out to be the wrong kind of thing to ground normativity anyway. Consciousness is in a comparable position.
§5 — Ethics without sentience
Having argued for insignificance, Kammerer sketches what should replace sentience-based ethics, offering three lessons:
1. It’s not catastrophic. Recognizing consciousness as insignificant doesn’t mean nothing is significant, nor that animal ethics collapses. Many valuable situations and moral-status-bearing creatures remain so — just for reasons unrelated to phenomenality. His favored alternative is a desire-/​preference-based approach (Carruthers, Dawkins): what makes an animal’s pain bad isn’t how it feels but that the animal wants it to stop and that desire is frustrated. This escapes both problems because “desire” is analyzable in functional, third-personal terms — so there’s a determinate fact about which creatures have desires (no indeterminacy), and our reasons for valuing desire-satisfaction don’t depend on suspect phenomenal introspection (no justification problem). Unlike Dawkins, who bracket sentience pragmatically and provisionally, Kammerer thinks the departure should be permanent.
2. Don’t seek a replacement magic bullet. We should resist finding some other property that grounds a binary, all-or-nothing status distinction. Folk concepts (like “desire”) carve simple binary distinctions, but the science reveals continuous gradations, variety, and multiple dimensions. He suggests grounding ethics on concepts treated as complex, degreed, and “complexifiable” as the science of minds advances. Desiring comes in a graded, multidimensional scale — from the most basic valuing (perhaps even bacteria, per Lyon & Kuchling) up to sophisticated “caring” in Frankfurt’s sense — and these degrees carry genuine normative weight.
3. It’s not inconsequential either. Against Carruthers’ reassurance that “much can remain unchanged,” Kammerer insists that taking the justification argument seriously means whatever reason we had for caring about animal pain (namely, our introspective certainty of the badness of our own pain) is itself compromised (he cites Delon’s Nietzschean challenge to the badness of suffering [Delon (2024)]). We can’t quietly keep relying on those un-interrogated beliefs. He proposes a “normative relocation story” paralleling the classic story about secondary qualities: just as we relocated colors/​sounds from external objects into experience after the Galilean mathematization of nature, we historically relocated value (goodness/​badness) from external objects into our pains and pleasures, where it seemed it could be exactly as it appears and directly known. Materialism forces us to abandon the idea that sensory qualities really exist in the phenomenal realm as they appear — and, Kammerer concludes, to draw the parallel conclusion about value. This leaves open hard questions about which ways of justifying value-beliefs survive, and how far the concept of value must be revised.
Conclusion
We should move to an ethic without sentience because sentience probably lacks the special normative significance sentience-based ethics requires. This doesn’t threaten the possibility of ethics (animal ethics included), but it would be a mistake to treat the shift as costless — our ordinary, introspection-based ways of justifying beliefs about the value of mental states are among the casualties.