Argument from marginal cases

The argument from marginal cases (AMC) is, in the literature on animal ethics, what logicians would call a dilemma. Take any cognitive capacity a defender of human exceptionalism proposes as the real basis of moral status — rationality, language, autonomy, moral agency, self-consciousness, the capacity for reciprocal contract. Now notice that some humans lack it: newborn infants, those with profound cognitive disabilities, adults in late-stage dementia, the permanently comatose. Either the capacity does not in fact draw the line the defender needs — because those humans still count and animals with comparable or greater capacities do not — or the defender must accept the conclusion that those humans may be treated as we treat pigs and chickens. Few are willing to do the second. The argument pushes the first.

The term marginal cases is a term of art, not a slur; it names humans who fall outside the cognitive profile philosophers have historically used to distinguish us from them. Daniel Dombrowski’s book-length treatment Babies and Beasts (Dombrowski, 1997) traces the argument’s lineage back through Porphyry and into contemporary analytic philosophy, and remains the standard reference.

Singer’s formulation

Peter Singer introduces the AMC in Practical Ethics (Singer, 1979) as a hammer against speciesism. If you say that humans matter more than non-humans because humans are rational, Singer replies: then what about the human who is not rational? Either rationality matters and those humans count no more than a pig — which is monstrous — or it does not matter, and the appeal was a smokescreen. Singer’s preferred conclusion is that the criterion worth keeping is sentience, which tracks interests; interests deserve equal consideration regardless of the species of the bearer.

The move is utilitarian in spirit but logically neutral. It does not tell you what grounds moral status; it tells you what cannot. Species membership, on its own, cannot, because every cognitive candidate for a morally relevant trait is either possessed by some non-humans or lacked by some humans — and often both.

Regan’s formulation

Tom Regan (Regan, 1983) builds the AMC into the architecture of The Case for Animal Rights. Regan’s criterion for inherent value is the subject-of-a-life: a being with beliefs, desires, memory, a sense of its own future, and welfare interests. This criterion is engineered to survive the AMC. It counts the profoundly cognitively disabled human and the year-old mammal on the same side of the line; both are subjects of lives, both have inherent value, and neither may be treated as a mere resource. Where Singer weakens the case for species-based discrimination, Regan constructs a rights framework that is species-blind from the start.

Cohen’s reply: species-typical capacity

Carl Cohen’s response, in a 1986 New England Journal of Medicine paper defending biomedical animal use (Cohen, 1986), is the most-cited counter. Cohen concedes that some humans lack the capacities relevant to moral agency but argues that moral status attaches to kinds, not to individuals. Humans belong to a species whose normal, mature members are moral agents; the profoundly disabled human shares in that status by virtue of kind membership even when she does not exhibit it. A chimpanzee does not, because chimpanzees as a kind are not moral agents.

Critics have pointed out that this move either collapses into naked speciesism — why should kind membership override individual capacity, except that we are drawing the line at our own species? — or borrows a framework (essentialism about biological kinds) that philosophers of biology have largely abandoned since the species-concept debates of the 1970s (Dombrowski, 1997). Cohen himself was unapologetic: he considered speciesism not a prejudice but a duty.

Carruthers: contractualism and the animals issue

Peter Carruthers’ The Animals Issue (Carruthers, 1992) takes a different route. Working within a broadly contractualist framework, Carruthers argues that morality is the set of rules rational agents would agree to under suitable conditions; non-rational beings are not parties to the contract and so fall outside direct moral protection. Animals matter, on this view, only derivatively — through the dispositions of human moral agents, whose callousness toward animals may bleed into callousness toward each other.

This invites the obvious AMC rejoinder: infants and severely cognitively disabled humans are also not parties to the contract. Carruthers’ answer is that they are protected by a social stability argument: a rule permitting their harm would destabilise the relationships of care that functioning societies require. Defenders of animals have found this thin. It makes the disabled human’s protection instrumental to everyone else’s comfort, and it leaves animals exposed when no such social-stability argument applies.

Scanlon’s trustees move

T. M. Scanlon, in What We Owe to Each Other (Scanlon, 1998), offers a subtler contractualist response. Scanlon’s core moral relation is justifiability — an act is wrong if it violates principles no one could reasonably reject. Non-human animals and marginal-case humans cannot, literally, reject principles. Scanlon handles this by extending the circle of those on whose behalf principles can be rejected: trustees stand in for those who cannot speak for themselves. A parent rejects on behalf of her infant; a carer rejects on behalf of the cognitively disabled adult.

The trustees move preserves the cognitive criterion for contract membership while saving the intuition that marginal-case humans are owed justice. But it also, some critics note, opens a door the defender may not want opened. If trustees can speak for non-contracting humans, why not for non-contracting animals? Scanlon’s own treatment of this is cautious; later contractualists (notably Mark Rowlands) have pushed the door further.

Dennett’s critique

Daniel Dennett, approaching the question from philosophy of mind, has argued that the AMC relies on a picture of cognitive capacities as sharp thresholds when the underlying reality is gradient and multidimensional (Dennett, 1996). The marginal-case human and the capacity-rich animal are, on Dennett’s view, not well described by the discrete cognitive tests analytic ethics likes to deploy. Dennett does not defend speciesism; his point is methodological. The argument survives his critique in its structural form — whatever the gradient looks like, some humans fall below some animals on any specified axis — but loses some of its sharpness.

Kasperbauer and the moral psychology turn

Recent work by T. J. Kasperbauer (Subhuman, 2018) examines why the AMC, despite its logical force, has made limited headway in popular moral intuition. Drawing on empirical moral psychology, Kasperbauer argues that the human-animal boundary is maintained by motivated cognition: we dehumanise animals as a routine cognitive operation, and evidence that complicates the boundary is discounted rather than integrated. The implication is that the AMC is philosophically watertight but psychologically uphill — an argument that wins on paper and loses in the gut, and that advocates need therefore to pair with narrative and acquaintance, not only syllogism.

What the argument actually does

The AMC is often mistaken for an argument that animals should be treated the same as humans, or that marginal-case humans should be treated the way animals currently are. It is neither. It is a consistency argument: whatever moral protection is owed to a human who lacks cognitive capacity X must be owed to a non-human who also lacks X, unless some further, non-speciesist reason is produced. It forces opponents to articulate that further reason explicitly. In the four decades since Singer and Regan, no such reason has emerged with broad philosophical acceptance — which is why the AMC remains, however uncomfortably, one of the most durable arguments in contemporary animal ethics.