Category: ethics
Type: thought experiment
Associated thinker(s): James Rachels
Original publication or approximate period: Rachels, “Active and Passive Euthanasia,” New England Journal of Medicine 292(2) (1975), 78–80; DOI: 10.1056/NEJM197501092920206
Difficulty: Beginner
Themes: killing and letting die, omissions, moral equivalence
Related experiments: Drowning Child; Trolley Problem; Transplant Problem
Smith stands to inherit a large sum if his young cousin dies. He sneaks into the bathroom while the child is bathing and drowns him, making it look like an accident. Jones is in the same position, with the same motive. He too goes to the bathroom intending to drown the child, but as he enters, the child slips, hits his head and falls under the water. Jones stands by, prepared to do the deed himself if need be, but it proves unnecessary and the child drowns.
Smith kills the child; Jones lets him die. They share the inheritance motive and intended end, and both children die, but they reach the outcome differently.
Can a bare difference between killing and letting die explain a moral judgment when the agents share a motive and intended end, but the details of how the death occurs differ?
Rachels describes a conventional medical doctrine that treats actively causing death as worse than allowing death, and quotes an American Medical Association policy statement. His pair tests whether the bare difference between killing and letting die carries moral weight when other features are held close. The example does not settle any particular clinical or policy decision.
A further question is whether the argument affects duties of aid. That implication needs additional premises; it is not established by the Smith–Jones case alone.
Rachels argues that neither Smith nor Jones behaves better in the story and challenges the claim that killing is intrinsically worse than letting die. The SEP records distinct objections: some challenge whether the cases are properly matched, others dispute the apparent moral equivalence or the inference to a general principle. Iain Brassington, for example, discusses the role of intention and motive in further cases. Rights-based, causal, and responsibility accounts are different approaches, not a single “moderate” position. These disputes do not establish a consensus.
The case does not show that everyone has a duty to rescue everyone, nor that doctors should be indifferent between two courses of treatment. Rachels tests a particular inference, and practical conclusions depend on further premises. The scenario involves the death of a child; the discussion here stays with the reasoning rather than graphic detail.
One can compare cases in which the agent has no bad motive or where preventing death would be costly. These are teaching comparisons, not a summary of a single view held by Rachels’s critics. A pair of cases that differs in more than one respect cannot isolate the effect of acting versus omitting.
Rachels’s complete article was read in a university-hosted reprint that states it is reproduced by permission. The NEJM publisher record confirms the title, author, date, volume, issue, pages, and DOI, but its full text is subscriber-only. Relevant sections of the Fall 2025 SEP entry and Brassington’s peer-reviewed article were also read. The account attributes arguments to their authors and does not treat their disagreement as a consensus or as proof that any argument succeeds.
Research check: 2026-11-01. Claim-level evidence, access limits, and unresolved issues are recorded in the source-review ledger.