Agential value
To be presented as a symposium at the APA Pacific in San Francisco, April 6, 2023
Dogs, pigs, and humans, or most of them during much of their life, have moral status; cars, and computers don’t. Dogs, pigs, and humans are sentient and are welfare subjects; cars and computers are not. If only sentient beings can have well-being, and only welfare subjects can have moral status, then sentience is necessary for moral status. Sentientism is the view that sentience is both necessary and sufficient for moral status, that all and only sentient beings have moral status. This paper is a challenge to sentientism. I assume that individuals rather than holistic composites of natural entities have welfare and therefore moral status (pace ecocentrism), that many nonhumans have welfare and moral status (pace anthropocentrism), and that merely being alive is not sufficient for having welfare and moral status (pace biocentrism). The space I want to carve is for another sufficient and independent source of welfare and moral status: agency. To my knowledge, only Shelly Kagan, in his recent book, How To Count Animals: More or Less (2019: 16-36), has entertained the view in any detail. I take it seriously and motivate and unpack two important but understated elements of his discussion: one, how it is that agency contributes to welfare and grounds moral status; two, filling in the empirical details of the conjecture of agency without sentience.
Sentientism is widely accepted in the literature on moral status and beyond, though not always explicitly (see Lee, 2022, for a recent defense). It states that sentience, or the capacity for phenomenal consciousness, is necessary and sufficient for moral status. More specifically, sentience is usually interpreted as the capacity to experience positively or negatively valenced affective states (pain, pleasure, joy, frustration, warmth, hunger, anger, sadness, boredom, anxiety, distress, etc.). The fundamental insight of sentientism is three-fold. First, moral status protects morally significant interests, which requires the capacity for welfare. Second, the latter presupposes point of view; things can go well or poorly for a subject from their perspective. Finally, only sentient beings are subjects. There is something it’s like to be a cat, but not catnip or a cat food dispenser. The focus on sentience in animal ethics is such that the expansion of moral concern now seems exclusively fixated on which animals are sentient, whether it be fish, cephalopods, or arthropods, especially insects and decapod crustaceans (Browning and Birch, 2022). The question remains Bentham’s: Can they suffer? I take an alternative approach. Whether or not uncertain cases are sentient need not be resolved to make progress on the question of moral status. Instead, we tend to ignore another relevant question: Do they have agency?
On the view I defend, all sentient beings and all agents, regardless of their overlap, have welfare and moral status. Like sentientists, I assume that a fundamental link between welfare and moral status. What I deny is the claim that only sentient beings can be welfare subjects. I reject sentientism about moral status by rejecting sentientism about welfare. The implications of this rejection are notable. Some animals, robots, and AIs could have moral status even if they lack sentience. Another implication is that we need not take the moral status of robots and AIs to hinge on whether they are or could be sentient. Much of the conversation revolves around that question. But AIs could have moral status even if they are not sentient.
The central claim of the paper is that if some non-sentient beings are sites of valence, they have moral status. The crux of the issue then will be whether sentience is necessary for valence. I argue that it is not, that agency is also sufficient, hence that non-sentient agents are sites of valence, can be welfare subjects, and therefore have moral status. The thesis is:
Agency: Agency is a sufficient moral-status-conferring property
The argument will appeal to three central claims:
Welfare: All and only welfare subjects have moral status.
Agential Welfare: some morally significant interests are based on agency; it is basically and pro tanto good (bad) for an entity to have such agential interests promoted (frustrated, setback, thwarted).
Agential Value: some states and events are prudentially good or bad for a subject because they promote the subject’s aims or because they involve the exercise of agency (e.g., planning, decision-making, exploring, playing, etc.).
Agential Value supports Agential Welfare, which is key to Agency. Agents can be welfare subjects and therefore can have moral status.
Like sentientists, I assume that a fundamental link between welfare and moral status. What I deny is the claim that only sentient beings can be welfare subjects. I reject sentientism about moral status by rejecting sentientism about welfare. The implications of this rejection are notable. Some animals, robots, and AIs could have moral status even if they lack sentience. Another implication is that we need not take the moral status of robots and AIs to hinge on whether they are or could be sentient. Much of the conversation revolves around that question. But AIs could have moral status even if they are not sentient.
David Chalmers (2022: 335-337 and in this video; see Roelofs 2022 for discussion) has recently suggested that Vulcans, hypothetical emotionless intelligent beings inspired by the Star Trek series, would have moral status. It would be wrong to torture, kill, kidnap, imprison them. Vulcans are phenomenally conscious but lack affect. They can think, remember, plan, and desire, they have goals and projects, but none of these states are affectively experienced. Whether or not philosophical Vulcans are possible, we can imagine some creatures that are conscious in some sense yet lack the affective capacities associated with sentience in its narrow sense.
The crux of the challenge, as I see it, consists in two steps:
Step 1. We can conceive of beings who lack sentience but whose lives are sites of valence;
Step 2. If such beings exist, they have some moral status—there are ways it would be wrong to treat them for their own sake.
Informally, the next step would be:
Step 3. Moral status does not fundamentally depend on sentience.
This is not meant as a complete argument. If no such beings exist, can we move to step 3? Can moral status depend on the mere possibility of such beings? We need not resolve these questions. Together, steps 1 and 2 imply that if some non-sentient beings are sites of valence, they have moral status. The crux of the issue then will be whether sentience is necessary for valence. I argue that it is not…
Email me (ndelon@ncf.edu) for a full draft!
Cited References
Browning, H. and J. Birch (2022) “Animal Sentience.” Philosophy Compass 17 (5): e12822. https://doi.org/10.1111/phc3.12822.
Chalmers, David (2022) Reality+: Virtual Worlds and the Problems of Philosophy. New York: W. W. Norton.
Kagan, Shelly (2019) How to Count Animals, More or Less. Oxford University Press
Lee, Andrew Y. (2022). “Speciesism and Sentientism.” Journal of Consciousness Studies 29(3-4): 205-228.
Roelofs, L. (2022) Sentientism, Motivation, and Philosophical Vulcans. Pacific Philosophical Quarterly, https://doi.org/10.1111/papq.12420.
Very interesting work! What, in your view, constitutes agency? And do you discuss any ways we can use to determine who are agents and who are not, in your full paper? Finally, does your analysis restrict agency to individuals alone or is there scope for even collectives or groups to be agents?
This is really exciting work and would love to read a full draft soon! Out of curiosity, have you interacted with Jeff Sebo’s (2017) "Agency and Moral Status"? From memory, that paper may offer a lot of interesting/beneficial overlap with your own conception of moral status as it corresponds to agency.