1 See, e.g., Mill, John Stuart On Liberty (1859)Google Scholar Ch. V; Feinberg, Joel ‘Legal Paternalism,’ Canadian Journal of Philosophy l (1971), 105–24CrossRefGoogle Scholar; Hodson, John D. ‘The Principle of Paternalism,’ American Philosophical Quarterly l4 (1977), 61–9Google Scholar; and the diseussion of some other current views on fustified paternalistic interference with adults in Thompson, Dennis ‘Paternalism in Medicine, Law, and Public Policy,’ in Callahan, Daniel and Bok, Sissela eds., Ethics Teaching in Higher Education (NY: Plenum 1980), 245–75.CrossRefGoogle Scholar In many of these discussions, nonvoluntariness and incompetence are treated as being equivalent (see, for example, Feinberg, who uses the terms interchangeably). But part of my purpose in this paper is to show that this is a mistake.

2 I use these cases because of the current concerns regarding admitting the poor to paid, non-therapeutic research programs, and admitting the involuntarily in-carcerated to any sort of biomedical research study. Case II comes from KAIMOWITZ v. MICHIGAN DEPARTMENT OF MENTAL HEALTH, Civil Action 73-19434 (Wayne County, Mich. Civ. Ct. 1973), where this decision was actually made. There are two concerns of equal moral interest in such cases, viz., the exploitation of vulnerable populations and the question of acceptably voluntary consent of the vulnerable. I deal with the latter issue here, but the former merits detailed consideration as well.

3 Frankfurt, Harry for example, claims that ‘Only another person can coerce us or interfere with our social or political freedom …,’ (‘Coercion and Moral Responsibility,’ in Honerich, Ted ed., Essays in Freedom of Action [Boston: Routledge and Kegan Paul 1973]. 83–4Google Scholar, emphases in original).

4 It will not do to protest here that refusing Margaret's admission is not to force her to stay in her circumstances, since she could earn the money she needs in some other, acceptable way. For if this is the case, then she has other live options, and her choice of the toxicity study cannot be considered substantially unfree – at least not because of her external circumstances (i.e., her ‘predicament’). But this anticipates the discussion in 11.2.

5 Thus, the cases I consider here for purposes of illustration are cases of human experimentation which I assume would be available to those whose choice to participate in them is considered acceptably voluntary. The question of what criteria should be adopted for the acceptability of research on human subjects in general is a different (and important) question, but one I do not take up here except peripherally.

6 See Greenspan, P.S. ‘Behavior Control and Freedom of Action,’ The Philosophical Review 87 (1978), 225–40CrossRefGoogle ScholarPubMed; and Bergmann, Frithjof On Being Free (Notre Dame: University of Notre Dame Press 1977)CrossRefGoogle Scholar, Ch. 4. In putting the matter this way, I follow people like Greenspan who want to detach the question of acting freely from the traditional question of freedom versus determinism.

7 See Annas, George J. Glantz, Leonard H. and Barbara F., Katz Informed Consent to Human Experimentation: The Subject's Dilemma (Cambridge, MA: Ballinger 1977)Google Scholar, Ch. 4. These authors make the same point, although I suspect that they would make a general presumption of vitiated voluntariness much more easily than I would. Also, they stop with this point, concluding that research ought not to be permitted in such places. But, as I have argued earlier and am about to emphasize again, when conditions are so abysmal that they justify a general presumption of vitiated voluntariness, prohibition of research participation available to persons at liberty (i.e., for experiments which meet currently accepted societal standards for responsible research on human subjects) is a paternalistic policy which is unjustifiable.

8 See my ‘Liberty, Beneficence, and Involuntary Confinement,’ The Journal of Medicine and Philosophy 9 (1984), 261-93.

9 Ibid.

10 Interestingly, Boswell noticed signs of this in Johnson's, ‘ … anxious care to go out or in at a door or passage by a certain number of steps from a certain point, or at least so as that either his right or his left foot (I am not certain which), should constantly make the first actual movement when he came close to the door or passage. Thus I conjecture: for I have, upon innumerable occasions, observed him suddenly stop, and then seem to count his steps with deep earnestness; and when he had neglected or gone wrong in this sort of magical movement, I have seen him go back again, put himself in a proper posture to begin the ceremony, and, having gone through it, break from his abstraction, walk briskly on, and join his companion.’ (Life of Johnson Vol. I [1791] [NY: Oxford University Press 1933], 232.)Google Scholar

11 Feinberg, Joel ‘What is So Special About Mental Illness?’ in his Doing and Deserving (Princeton: Princeton University Press 1970), 282Google Scholar

12 Greenspan; also Dworkin, Gerald ‘Acting Freely,’ Nous 4 (1970), 367–83.CrossRefGoogle Scholar

13 Cf. Dworkin, 378.

14 Dworkin, 378-9, my emphases.

15 For interesting testimony on this, see some of the experiential accounts of mental illness in Kaplan, Bert (ed.), The Inner World of Mental Illness (NY: Harper and Row 1964), e.g., 235–6.Google Scholar

16 For an illuminating discussion of weakness of will, see Watson, Gary ‘Skepticism About Weakness of Will,’ The Philosophical Review 86 (1977) 316–39.CrossRefGoogle Scholar Watson accepts the irresistible impulse doctrine, and he allows that for any given action, the weak person, because of his weakness, may not be able, like the compelled person, to overcome an impulse. But, according to Watson, the weak person is responsible for his weakness in a way the compelled person is not responsible for his compulsion. One could reject the irresistible impulse doctrine, yet accept the other central points of Watson's view. For example, one might accept the coercion model of psychological compulsion and allow that, given someone's weakness, it would be unreasonable to expect him to do A rather than B (where A is, generally, reasonable to expect of people), yet hold him responsible for being someone who is so weak, just as we might hold the inebriated driver who causes an accident responsible for his inebriation. Watson's view of weakness rests on considerations like those introduced in Section II – weakness is, for Watson, decided (in part) on the basis of what, in general, can be reasonably expected of people in the way of self-control.

17 Pennock, J.R. Liberal Democracy (NY: Rinehart 1950), quoted in Dworkin, 380.Google Scholar

18 Dworkin, 380-1; his reference is to Oppenheim, F. Dimensions of Freedom (NY: St. Martin's Press 1961).CrossRefGoogle Scholar

19 Dworkin, 381

20 Dworkin, Gerald ‘Autonomy and Behavior Control,’ Hastings Center Report 6 (February 1976), 23–8CrossRefGoogle ScholarPubMed

21 See note 16. If one takes a view like Watson's, then the weak-willed (sometimes, at least) may act nonvoluntarily. But one need not take this kind of position.

22 Feinberg, ‘What is So Special About Mental Illness?’ 286; emphases in original.

23 For an articulation of what I take these side constraints to include, see ‘Liberty, Beneficence, and Involuntary Confinement,’ 271-3.

24 The final version of this paper was written with the support of the Louisiana State University Council on Research. I am grateful to Raymond Martin, Tziporah Kasachkoff, and referees for this Journal for their helpful comments on earlier drafts, and to Nancy P. Oraftik, R.N., M.S., C.S. for her professional advice on cases like Ms. Jay's and Mr. Dee's.