In debates over abortion, homosexual conduct, euthanasia, prostitution, drugs. Those who call themselves liberals often mount the same basic argument. A socially or morally permissive stance is necessitated towards such practices because people have a right to choose do what ever they like with their bodies. As Mill put it,
The only part of the conduct of anyone, for which he is amenable to society, is that which concerns others. In the part which merely concerns himself, his independence is, of right, absolute. Over himself, over his own body and mind, the individual is sovereign.  [Emphasis added]
Mill here draws a distinction between other-regarding actions, actions that harm other people and self-regarding actions, those that harm oneself. He argues that society, either by law or by social pressure, cannot justly regulate any action a person performs unless it is other-regarding; that is, it harms people other than the agent him/herself. As Mill’s position is typically interpreted harm is understood to be governed by the principle “volenti non fit injuria” (where there is consent there is no injury) and hence refers to things done to other people without their consent. On this interpretation, self-regarding actions are those that people consent to and that harm no non-consenting, third party. As Mill himself notes, a self-regarding action is that “which affects only himself, or affects others with their free and voluntary, and undeceived consent”.
Perhaps one of the more important counter-examples to this criterion is that of Gladiatorial Matches. Irving Kristol famously proposed this counter-example,
[T]he plain fact is that none of us is a complete civil libertarian. We all believe that there is some point at which the public authorities ought to step in to limit the “self-expression” of an individual or group even where this might be seriously intended as a form of artistic expression, and even where the artistic transaction is between consenting adults. A playwright or theatrical director might, in this crazy world of ours, find someone willing to commit suicide on stage, as called for by the script. We would not allow that-any more than we would permit scenes of real physical torture on the stage, even if the victim were a willing masochist. And I know of no one, no matter how free in spirit, who argues that we ought to permit gladiatorial contests in Yankee stadium, similar to those once performed in the Coliseum of Rome-even if only consenting adults were involved.
Kristol notes several cases of unjust homicide, homicide that any person of common sense would condemn but where the victims all consent to be killed.
In Harmless Wrongdoing Feinberg considers this as the most important and potentially-devastating counter-example and he spends considerable energy attempting to circumvent it. He offers three responses. Here I will argue that these responses fail.
The first response is to question whether, in fact, such cases are really cases of consensual homicide. Feinberg suggests that Kristol is too complacent about “the problem of determining genuine ‘willingness’ and voluntary ‘consent’”. He offers two lines of argument for this. Firstly, he notes that Kristol refers to such things happening in this “crazy world”. Feinberg states, “an agreement is hardly consensual if one of the parties is “crazy”. To exploit a crazy person in the way he describes is not distinguishable from murder”.
Feinberg is, of course, correct; people who are crazy in the sense of insane are not competent moral agents. However, it is doubtful that Kristol is, in fact, speaking of people who are crazy in the literal sense of the word. His reference is a colloquial comment on the kinds of depravity people will engage in if given the freedom to do so.
The second reason Feinberg offers is that if gladiatorial matches existed today one would need various mechanisms to ensure consent was actual; such licensing procedures would render the contests unobjectionable. Under the conditions necessary to discern consent, Feinberg suggests that both the contestants and the audience would agree to a set of rules that held that the contest must stop once one fighter has clearly gained his or her dominance.
This response is inadequate. It is unclear why contestants would agree to these provisions. Why, for example, could they not agree to a free for all or a fight to the death? Arneson notes, “Feinberg has projected a bit of his own nice character onto the world at large” and correctly adds “Feinberg’s liberal principles non-tendentiously applied would hold that we ought to let the free market decide how horrific or bloody the contests should become”.
Moreover, Arneson adds that even if the contestants and audience did agree to such rules, it is hard to see why the practice would then become unobjectionable.
Feinberg’s response contains the disturbing suggestion that a state-regulated version of Kristol’s gory spectacles would be acceptable to the liberal … so long as some analogue of the Marquis of Queensberry rules is enforced (no bludgeoning your opponent when he is down; no slashing below the belt; no disembowelling your opponent once he surrenders), commercial combat with lethal weapons should in principle be tolerated by a liberal society. Many would consider this a reductio of the liberal position.
At one point in the discussion, Feinberg appears aware that it is a stretch to suggest that such matches could not be consensual.
There could be a presumption that such a person doesn’t fully understand what he is doing, or is not fully free of neurotic influences on his choice, but these hypotheses are rebuttable in principle, and in some cases that we can easily imagine, only with minor difficulty and expense rebuttable in fact. The liberals second and third responses then are the more pertinent ones.
It is to these I now turn.
(b) Harm to Third Parties
Feinberg’s second attempt to circumvent Kristol’s counter-example is to argue that this case would inevitably involve harm to non-consenting, third parties. Reflecting upon the audience in a gladiatorial match Feinberg suggests,
We cannot hold an image of these wretches in our minds without recoiling, for each of them alone will seem threatening or dangerous, and thousands or millions of them together will be downright terrifying. … When the bloody maiming and slaughtering of a human being is considered so thrilling and enjoyable that thousands will pay dearly to witness it, it would seem to follow that thousands are already so brutalized that there is a clear and present danger that some innocent parties (identities now unknown) will suffer at their hands.[Emphasis original]
Like the former, this response is inadequate. First, the claim that such an activity will lead to the deaths of innocent, third parties is an empirical one. Feinberg simply asserts that it is correct but offers no evidence to prove this. In the absence of such evidence, the claim is speculative.
A second problem reinforces this. The people killed in gladiatorial matches consent to the fight; the innocent third parties Feinberg mentions do not. In order for enthusiasm for death and killing in a gladiatorial match to spill over into the killing of innocents, fans of such matches must be unable to distinguish between consensual and non-consensual killing so that they are unable to limit their enthusiasm for one without endorsing the other.
This assumption is implausible. For example, permitting people to engage in and enjoy consensual sodomy does not necessarily mean that those people will engage in sodomous rape. Allowing the consensual use of drugs will not mean that all drug users will then coerce others into drug use. Allowing people to watch pornography would not result in all people forcing others to have sex. Perhaps, even closer to the issue under discussion, allowing people to watch boxing matches does not appear to lead them to inevitably engage in assault. Why then should allowing people to watch consensual homicide lead them to engage in non-consensual homicide?
(c) Bloated Mouse
The final response Feinberg proposes is that these counter-examples are what he calls “a bloated mouse”. He concedes that in cases like these we do have a case of unjust homicide that is consensual. However, he suggests cases like this are hypothetical, “[t]here seems little likelihood that they will ever occur, at least in the foreseeable future”.
It is worth noting that Feinberg again relies upon an unsubstantiated, empirical claim. He asserts, without any reason, that gladiator matches are unlikely to occur at all in the foreseeable future but why think this? Dueling to the death was practiced for centuries. People clearly have been willing to engage in such activities and did so for many years. Moreover, it is also plausible that people would choose to watch such spectacles, even ordinary people. Public executions were popular as late as the nineteenth century, suggesting that people are willing, if allowed, to watch real death and violence. The practice of the circuses in a civilised culture such as Rome shows the appeal of watching such spectacles. There appears no reason for thinking that such an activity is unlikely to occur at all given the history of the human race.
Interestingly, Feinberg admits this elsewhere in his monograph. In a footnote he notes that the example of gladiatorial contests is no worse and in some respects better, than some forms of commercial brawling that exist in the U.S. He then cites disapprovingly the examples of “tough guy” contests where people fight bar-room brawls with no holds barred, no rules and often on racial lines. The sources he quotes note that such contests are popular, “the newest rage in spectator entertainment”. It is hard to understand then why Feinberg thinks that such things are unlikely to occur in the future.
It seems then the Feinberg fails to escape the counter example proposed by Kristol. Liberal Morality provides no basis for condemning blood sports and Gladiator matches and as such it is inadequate.
 John Stuart Mill, On Liberty (London: Penguin Classics, 1985), 69.
 Ibid. 71.
 Irving Kristol “Pornography, Obscenity and the Case for Censorship” New York Times Magazine March 28 1971.
 Feinberg Harmless Wrong Doing: Moral Limits of the Criminal Law. Vol. 4. (New York: Oxford University Press, 1987), 127
 Ibid. 130.
 This highlights another feature of Feinberg’s discussion. Feinberg appears to find it difficult to accept that a sane person would ever choose to engage in the depraved activities outlined in various counter examples. However, if one believes that human beings are capable of evil then much of what he says is implausible.
 Arneson “Liberalism, Freedom and Community” 373, 374.
 Ibid. 374.
 Feinberg Harmless Wrong Doing: 130.
 Ibid. 131.
 Ibid. 130.
 Ibid. 353.