Category: consciousness
Type: thought experiment
Associated thinker(s): David Lewis
Original publication or approximate period: Lewis, “Mad Pain and Martian Pain” (presented in 1978; published in Ned Block, ed., Readings in the Philosophy of Psychology, vol. I, 1980, pp. 216–222)
Difficulty: Intermediate
Themes: functionalism, identity theory, multiple realizability
Related experiments: Inverted Spectrum; Philosophical Zombies; China Brain
In his essay, David Lewis asks us to imagine two unusual beings. The first is a madman who sometimes feels pain just as we do, but whose pain has an unusual causal pattern: moderate exercise on an empty stomach brings it on; it helps him concentrate on mathematics while distracting him from other things; and intense pain makes him cross his legs and snap his fingers rather than groan or writhe. He is not motivated to prevent or end the pain.
The second is a Martian with a very different physical constitution. When pinched, hydraulic cavities in his feet inflate; that state disrupts his thought and activity, makes him groan and writhe, and motivates him to stop the pinching. Lewis stipulates that this state feels like pain, although it is realized by bodily machinery unlike ours.
Can a materialist theory of mind accommodate both cases: pain whose causal role is atypical for an exceptional human, and pain realized by a different physical kind in another population?
Lewis sets two simple theories against each other. A simple type-identity theory, which identifies human pain with a particular human physical state, can count the madman as in pain if he has that state, but seems to exclude the Martian, whose pain has a different physical realization. A simple role-only functionalism can count the Martian as in pain if his state plays the pain role, but excludes the madman because his state does not play that role for him.
Lewis argues that these simple versions are too restrictive. His proposal combines causal role with physical realization relative to a population: a state is pain for a population when, with exceptions, it occupies that population’s pain role; an individual is in pain when in the state that occupies the role for the appropriate population. The madman can be an exception within humanity, while a different physical state can realize pain for Martians. This resembles multiple realization across populations, although Lewis allows that different population choices may yield different senses of “pain”; the scenario does not establish that such realizations actually exist.
Lewis’s population-relative account is a hybrid, not simply the claim that pain is whatever role a creature happens to occupy. It treats role occupancy as population-relative and allows a state to realize pain for a population even when an exceptional member does not have the usual causes and effects. Lewis also acknowledges that the criteria for selecting an “appropriate population” can conflict, leaving some cases unsettled.
A later objection concerns a mad Martian, who is atypical both in causal role and within his species. Peter Alward argues that Lewis’s criteria do not clearly select among the human, Martian, and mad-Martian populations in that case. This is a published challenge to how the account handles the case, not a settled verdict on Lewis’s view.
The cases do not show that Martians exist, prove that either imagined case is possible, or establish functionalism as true. Lewis explicitly says he does not know how to prove the possibility of mad pain or Martian pain; he presents his judgments about those cases as intuitions a credible theory should accommodate.
Nor does Lewis treat phenomenal feeling as irrelevant. He says pain is a feeling and argues that a theory of what makes a state pain is also a theory of how it feels. That is Lewis’s defended position, not a result demonstrated by the thought experiment. The broader question—whether causal-role descriptions can capture the qualitative character of pain—remains disputed in philosophy of mind.
As editorial teaching prompts, one can ask whether a robot or other artificial system could feel pain, or compare these cases with the China Brain and fading-qualia discussions. Those comparisons raise related questions about functional organization and experience; they are not additional cases presented by Lewis, nor conclusions established by his argument.
The cases and the population-relative proposal are David Lewis’s, not Ned Block’s; Block edited the anthology in which Lewis’s essay appeared. Lewis’s text says the paper was presented at a Rice University conference in April 1978; the cited anthology printing is from 1980. The consulted passages cover the two scenarios, the contrast between simple identity and role-only theories, Lewis’s proposed account and population criteria, and his discussion of phenomenal character. Janet Levin’s Stanford Encyclopedia entry supplies context on multiple realization, identity theory, and objections about qualitative experience. Peter Alward’s 2004 article presents a specific later criticism of Lewis’s population criteria. Further comparisons and questions in this entry are editorial teaching framing, not historical attributions.
Research check: 2026-11-01. This is an educational explanation, not an expert review or a substitute for the cited works.