Euthanasia is legal only in the Netherlands and Belgium, but it is on occasion performed by physicians elsewhere. We recruited in France two convenience samples of 221 lay people and of 189 professionals (36 physicians, 92 nurses, 48 nurse's aides, and 13 psychologists) and asked them how acceptable it would be for a patient's physician to perform euthanasia in each of 72 scenarios. The scenarios were all combinations of three levels of the patient's life expectancy (3 days, 10 days, or 1 month), four levels of the patient's request for euthanasia (no request, unable to formulate a request because in a coma, some form of request, repeated formal requests), three of the family's attitude (do not uselessly prolong care, no opinion, try to keep the patient alive to the very end), and two of the patient's willingness to undergo organ donation (willing or not willing). We found that most lay people and health care professionals structure the factors in the patient scenarios in the same way: they assign most importance to the extent of requests for euthanasia by the patient and least importance (the lay people) or none (the health professionals) to the patient's willingness to donate organs. They also integrate the information from the different factors in the same way: the factors of patient request, patient life expectancy, and (for the lay people) organ donation are combined additively, and the family's attitude toward prolonging care interacts with patient request (playing a larger role when the patient can make no request). Thus we demonstrate a common cognitive foundation for future discussions, at the levels of both clinical care and public policy, of the conditions under which physician-performed euthanasia might be acceptable.