1. Loving v. Virginia, 388 U.S. 1 (1967).

2. Bowers v. Hardwick, 106 S. Ct. 2841 (1986).

3. Barnes v. Glen Theatre, 111 S. Ct. 2456 (1991).

4. Lawrence v. Texas, 539 U.S. 558 (2003).

5. Christine Pierce, AIDS and Bowers v. Hardwick, 20 J. Soc. Phil. 21, 24 (1989).

6. Justice Scalia's dissent in Lawrence relies in part on this type of argument.

7. The most thorough discussion of liberty-limiting principles is Feinberg's four-volume work: Joel Feinberg, The Moral Limits of the Criminal Law (1984–1988). It includes Feinberg, I Harm to Others (1984), Feinberg, II Offense to Others (1985), Feinberg, III Harm to Self (1986), and Feinberg, IV Harmless Wrongdoing (1988). In Harm to Others, at 26–27, Feinberg introduces and provides a finely nuanced discussion of numerous such principles. Although I do not argue for the claim here, it seems that the issues of how we justify in general the institution of imposing legal sanctions and what justifies imposing legal sanctions on some person in some circumstances are separable. John Rawls's classic Two Concepts of Rules 64 Phil. Rev. 3–32 (1955), provides one example of how to separate these issues.

8. Mill's view is most clearly articulated in John Stuart Mill, On Liberty (1859), Hart's in H.L.A. Hart, Law, Liberty and Morality (1963), and Feinberg's in Feinberg, Moral Limits of the Criminal Law, supra note 7.

9. It should be emphasized that one crucial difference between Mill's concerns in On Liberty and the contemporary discussion of liberty-limiting principles concerns their domain of application: whereas contemporary discussions focus on the legitimate employment of legal sanctions, Mill was at least equally concerned with subtler forms of coercion, including the tyranny of the majority, which need not rely on legal sanctions for their efficacy. And even within the legal realm, some legal phenomena are, while not coercive in themselves, clearly coercive in effect (e.g., marriage restrictions, poll taxes, literacy tests, and photo identification requirements). I thank Bob Hoag for alerting me to this point.

10. Aquinas, Summa Theologica, at question 96, article 2.

11. Lord Patrick Devlin, The Enforcement of Morals (1968). Indeed, the only legal moralist who comes close to advocating the sufficiency view is James Fitzjames Stephen, Liberty, Equality and Fraternity (1967), and it is not clear that even he does so in a thoroughgoing and consistent manner; Stephen maintains that the costs of enforcement can be sufficient to outweigh reasons for enforcement owing to an act's wrongness. Stephen's view aside, other considerations that might outweigh a legal-moralist reason in favor of legal sanctions include pragmatic difficulties concerning detection, that a statute governing the conduct might be unenforceable, and that certain important rights might be sacrificed in the process. Thanks to an anonymous reviewer for alerting me to this point.

12. See text at note 21infra for further discussion of this point.

13. See Bayles, Michael, Legislating Morality, 27 Wayne L. Rev. 774 (1976)Google Scholar.

14. As Feinberg as argue, and as we shall see below, a retreat from the requirement of “true morality” to the weaker “critical morality,” which requires that the (conventional) standards to be enforced meet some test of plausibility, would not do either, since the critical morality to be enforced might be just as ignorant and mistaken as the standards enforced uncritically by conventional legal moralism.

15. Feinberg, Harmless Wrongdoing, supra note 7, at 9. The conservative thesis should be distinguished (at least in principle) from conventionalist legal moralism for the following reason: if it is possible that failure to enforce conventional morality will not in fact produce drastic, essential, or malignant change in society's way of life or valued institutions, then legal sanctions would not be justifiable according to moral conservatism but would be according to conventionalist legal moralism.

16. Id. at 133. He also states that pure, strict legal moralism represents the most difficult challenge for the liberal.

17. As will become clear, I owe debts to Ronald Dworkin for the general line of thought behind this position as well as to John Rawls for some important insights on the relationship between moral theory and political philosophy (which I co-opt and apply to legal philosophy) and to Carl Cranor for developing what he thought to be the strongest, but ultimately an unsuccessful, argument for legal moralism, and which he claimed to enable us to dispense with the view once and for all. See Cranor, Carl, Legal Moralism Revisited, 89 Ethics147–164 (1979)Google Scholar. If I am right, perhaps Cranor's fitting a casket for legal moralism is a bit premature. But if the version that I sketch here must also be rejected, then I am prepared to join Carl as a pallbearer at the funeral.

18. Railton, Peter, Moral Realism, 95 Phil. Rev.163–207 (1986)Google Scholar.

19. The term “Cornell Realism” is often used to refer to an especially robust version of moral realism, notably exemplified in the work of philosophers with ties to the Cornell University philosophy department such as Nicholas Sturgeon and David Brink, according to which the world contains objective, mind-independent moral facts.

20. I follow Chris Korsgaard in using the term “moral realism” to include both substantive realism and procedural realism. See Christine Korsgaard, The Sources of Normativity 36–37 (1996).

21. Cranor, supra note 17.

22. I am assuming that among ways of trying to discouraging conduct, legal sanctions would be regarded as among the most intrusive, hence demanding a higher level of justification, and thus would be reserved for the most serious offenses.

23. Cranor, supra note 17.

24. Cranor, supra note 17.

25. Dworkin, Ronald, Lord Devlin and the Enforcement of Morals, 75 Yale L. Rev., 986–1005 (1966)Google Scholar, reprinted in Dworkin, Taking Rights Seriously ch. 10 (1977).

26. The formal characterization structure is borrowed from Cranor, supra note 17.

27. As this view is stated, it could be interpreted as stipulating a necessary condition for applying legal sanctions, but that claim would probably be too strong. Understood in that manner, it would undermine the distinction between malum in se and malum prohibitum in the following respect: although not all conduct that is malum in se would be malum prohibitum (some instances of wrongdoing would not be legal wrongs), all that is malum prohibitum would be malum in se, since legal sanctions could justifiably be applied only to conduct wrongful in itself. Instead, it should be viewed as stipulating necessary conditions for having a good legal-moralist reason for employing legal sanctions, leaving it open that there may be other legitimate reasons that issue from other liberty-limiting principles (harm, offense, etc.), and leaving it open further that legal-moralist reasons could be overridden by other reasons.

28. Alternatively, this view maintains that legal-moralist arguments for imposing legal sanctions should be limited to conduct that falls within a Rawlsian overlapping consensus of reasonable moral positions: all reasonable positions would deem sanctions justifiable. See John Rawls, Political Liberalism (1993), on the idea of an overlapping consensus.

29. The unanimity version's formal characterization runs as follows:

1. (1)

   (1) A type of conduct C is immoral according to realist legal moralism (RLM-immoral) only if C is identified as morally wrong and, all things considered, ought not to be engaged in, according to every rationally defensible moral position.

2. (2)

   (2) A type of conduct C ought to be discouraged according to RLM only if (a) C is RLM-immoral, and (b) C ought to be discouraged according to every rationally defensible moral position.

3. (3)

   (3) The use of legal sanctions to prevent people from engaging in C is justifiable according to RLM only if (a) C is RLM-immoral, (b) C ought to be discouraged according to every rationally defensible moral position, and (c) C is a serious enough wrong to warrant the use of legal sanctions to prevent people from engaging in C according to every rationally defensible moral position.

4. (4)

   (4) The use of legal sanctions to prevent people from engaging in C is justifiable on legal-moralist grounds only if (1), (2), and (3) are satisfied.

Whereas the unanimity version requires that conduct be condemned as immoral and deserving of sanctions by “every rationally defensible moral position,” the consensus version requires instead that conduct be condemned and judged to be deserving of sanctions by “a clear consensus of rationally defensible moral positions.”

30. In his comments on an earlier version of this paper Bob Hoag points out that there is bound to be some overlap associated with a plurality of liberty-limiting principles, and I agree: some types of conduct maybe harmful, wrong, and profoundly offensive, in which case imposition of legal sanctions against that conduct might be overdetermined. Only Mill's “one simple principle,” as opposed to Feinberg's and Hart's endorsements of multiple principles, would have any promise of avoiding this consequence.

31. What appears to bother Feinberg is the possibility of some yet-undiscovered rationally defensible moral position that shows up in the future and demonstrates that however rationally defensible we might have viewed prior moral positions to be, they and the judgments that follow from them are ultimately indefensible. This theme runs throughout Feinberg, Harmless Wrongdoing, supra note 7.

32. Cranor, supra note 17.

33. Since most who reject legal moralism fear that it would allow an overly broad net of sanctions, I concentrate on the consensus version of realist legal moralism, since this “weaker” version of the view produces a wider net than the stronger version and since the difference between it and the unanimity version lies not so much in the kind of argument required to justify imposing legal sanctions (each requires appealing to the idea of a rationally defensible moral position) but whether every such position would judge sanctions as justified or whether a clear or overwhelming consensus of such positions would do so.

34. Palko v. Connecticut, 302 U.S. 319 (1937), cited in Bowers, supra note 2, at 191–192.

35. Moore v. East Cleveland, 431 U.S. 494 (1977), cited in Bowers, Id. at 192.

36. See discussion of Justice Kennedy's argument in Lawrence, supra note 4, discussed infra.

37. Bowers, supra note 2, at 196.

38. Id. brief for respondent, at 25, 29.

39. Id. at 199–200.

40. Id. at 204.

41. Id. at 206.

42. Id. at 210.

43. Id. at 211–212.

44. Id. at 212.

45. Some of the arguments in Justice Stevens's dissent also have legal-moralist overtones; and those arguments were noted and endorsed by Justice Kennedy in Lawrence, supra note 4.

46. Barnes, supra note 3, at 2461–2462.

47. Id. at 2463. There is also the following: to “reflect moral disapproval of people appearing in the nude among strangers in public places”; Id. at 2461.

48. Id. at 2468.

49. In his comments on a previous version of this paper Bob Hoag suggests, correctly I believe, that a position can be rationally defensible even if on some occasion a person (say, a justice) in some context (say, a judicial opinion) fails to offer a rational defense of it. And it is important to note that Dworkin's criticism of Devlin addresses this point: it is not that citizens need to be prepared to back up their moral convictions with reasons that satisfy the defeasibility conditions. Rather, as I understand it, Dworkin's position is that when considering legislation that is backed up by popular “moral” sentiment, the legislators are supposed to determine just what kinds of reason could be summoned in support of the conviction that underlies the legislation. Similarly, it seems that the whole enterprise of justices offering arguments for their conclusions commits them to offering a defense of (i.e., reasons for) their conclusions. A justice who is adopting the kind of position I have defended would, in appealing to that position as a premise in her argument, be committed to offering some reason for concluding that some type of conduct is judged seriously immoral enough to warrant legal sanctions according to a consensus of rationally defensible moral positions. Given that no such argument appears in these opinions, it is clear that the justices were not invoking anything like a realist version of legal moralism, and thus I conclude that they did not rely on a plausible version of legal moralism in making those arguments.

50. Barnes, supra note 3, at 2468.

51. Id. at 2469.

52. Some might insist that to refer to such phenomena as “effects” constitutes, at the very least, an abuse of language.

53. In discussing nude dancing and Barnes in his comments on a previous version of this paper, Bob Hoag (rightly) points out that one might view not just (or even) the nude dancing as the immorality to be combated, but the audience's watching the nude dancing or taking the pleasure that they do in observing it. But here are some thoughts on that line of argument: if the patrons are the ones behaving immorally, then legal moralism would seem to require that the sanctions be directed at the patrons, not the dancers. And if, as Hoag (giving voice to a certain kind of argument) suggests, this may be viewed a case of immorality without any immoral conduct, then the question would be: What kind of “immorality”—and what kind of legal moralism—are we dealing with? How “free-floating” can these “immoralities” be if they are not attached to any kind of immoral conduct? My guess is that when pressed, most who are likely to argue this way eventually reveal their concern with the secondary effects of such activities and rely on some version of moral conservatism in making their case.

54. Barnes, supra note 3, at 2465.

55. Id. at 2472.

56. Loving, supra note 1, at 3.

57. Id. at 12.

58. Lawrence, supra note 4, at 578.

59. Id. at 568 ff.

60. Id. at 582–583.

61. It might be argued that the rejection of conventionalist legal moralism by Justices Kennedy and O'Connor should be understood as rejecting that principle as providing a sufficient condition for imposing sanctions, leaving it open that they would accept it as providing a relevant reason. But Justice Kennedy states that “[t]he Texas statute promotes no legitimate state interest” (Id. at 578), which suggests that the presumed conventional judgment of the immorality of homosexual conduct carries no weight whatsoever. Regarding Justice O'Connor's position the evidence is mixed: she writes that “we have never held that moral disapproval, without any other asserted state interest, is a sufficient rationale” (Id. at 582) for imposing sanctions, which seems to support the “relevant reason” claim, but she argues that “[m]oral disapproval of a group cannot be a legitimate governmental interest under the Equal Protection clause” (Id. at 583), which supports the view that it is not even a relevant reason.

62. At times the claim is that only rights that are deeply rooted are fundamental rights, and fundamental rights demand higher scrutiny; at others it is that only those fundamental rights that are deeply rooted (leaving it open that some fundamental rights are not deeply rooted) get heightened scrutiny.

63. Id. at 593–594.

64. See Scalia's discussion of the “rational basis” test; Id. at 599.

65. An alternative reading of Justice Scalia's argument, suggested to me by Alan Fuchs, is that the Constitution delineates a set of fairly specific fundamental rights, which only a compelling interest can override and outside of which legislators are relatively free to do as they see fit. Interestingly, at a deep structural level, this position resembles that which Dworkin introduced in Taking Rights Seriously and has been refining since: principle overrides policy, but in the absence of some principle to constrain them, policy considerations may proceed relatively undeterred. I agree that this may even be the most charitable reading of Justice Scalia's argument, but of course Dworkin would disagree violently (and rightly so) with Justice Scalia's account of the nature and extent of those rights. And if Fuch's interpretation of Justice Scalia's position is correct, this would entail that legal-moralist considerations play almost no role at all in Justice Scalia's reasoning in this case, thus supporting my contention that the deficiencies in his argument are not due to his reliance on legal-moralist premises.

66. Perhaps this reliance on consistency might be viewed as evidence of Justice Scalia's adoption of at least one of Dworkin's defeasibility conditions, viz., that convictions be tested for consistency alongside other judgments. But a more natural way to understand this argument is as an application of stare decisis to the outcome of Lawrence: that all these other instances of “morals” legislation would now have to be deemed illegitimate. So the domains of Dworkin's consistency condition and stare decisis would seem to be fundamentally different. I would also add that accepting stare decisis is not sufficient for adopting a legal-moralist position. Thanks to Bob Hoag for bringing this point to my attention.