My purpose here is not to disagree with Steve Heilig’s conclusion concerning physician participation in torture. My purpose is to provide a moral foundation for his argument. That is necessary, I think, because Heilig’s argument is unusual in a discussion of bioethics (including academic medical ethics). There is, in particular, little said about the “principles of bioethics” (autonomy, beneficence, nonmaleficence, and justice) or about the nature of medicine, the physician-patient relationship, the physician’s “social role,” or similar arguments common in bioethics. Instead, Heilig rests his argument primarily on the AMA’s Code of Ethics (and international equivalents)—what most bioethicists tend to treat as mere custom, etiquette, law, or statements of opinion (when they do not ignore the codes altogether). Although I believe that Heilig is right to rely on the code, he is, I also believe, wrong to have done so without some response to common objections—for example, those that Fritz Allhoff made in this journal only a few years ago.^{Footnote 1}

If I have any disagreement with Heilig, it is with his choice of one term. I shall substitute “hostile interrogation” for his “torture.” Of course, the term “hostile interrogation” is a euphemism, like “take off the gloves” or “give him the third degree.” In war or peace, interrogation is always hostile; that is, it is an attempt to get information from someone who has good reason to withhold it. The term’s friendlier counterparts include “questioning,” “debriefing,” and “interviewing.” Insofar as “hostile” is added to “interrogation,” what is added, at a minimum, is cruel, inhuman, or degrading treatment of the person interrogated—and, often, torture strictly speaking.^{Footnote 2} The same international conventions that forbid “torture” forbid such mistreatment. Anyone who must argue that her conduct is not torture but merely hostile interrogation is almost certainly engaged in conduct both legally and morally well beyond the permissible. I nonetheless adopt the term so that those Heilig and I mean to criticize understand that we mean to criticize them.

My thesis is, then, perhaps a bit broader than Heilig’s but entirely consistent with his: physician participation in hostile interrogation is both unethical and unprofessional. Of course, “ethics” has at least three senses: ethics as ordinary morality, ethics as moral theory, and ethics as special, morally permissible standards. I will reserve “ethics” for the special-standards sense—with “morality” and “moral theory” serving (respectively) for the other two. In this preferred sense of ethics, a code of ethics is a formal statement (in large part at least) of special, morally permissible standards—that is, it is a statement of standards that apply to members of a specific group because they are members of that group, not a statement of ordinary moral standards (which apply to everyone) or of moral theory (systematic reflection on morality). Though most codes of ethics result in part from reflection, none is itself a reflective statement (in the sense, say, that Aristotle’s Ethics is). A code of ethics sets standards rather than offering arguments for them.

Because our subject is physician participation in hostile interrogation, we should ask: who counts as a “physician”? A physician is not merely a person trained in medicine. An unlicensed physician is not a physician, except in the way an unsigned check is still money. The potential is there, but not the actuality. If an unlicensed physician identifies himself as unlicensed at a hospital or clinic, he will not be allowed to serve as a physician, whatever his knowledge, skill, experience, and other credentials. If he omits mentioning that he is unlicensed, he will be guilty of a crime (practicing medicine without a license).

Why is the license so important? It establishes that the person in question is a physician in good standing, not merely someone with medical training, skill, and experience. In the United States, part of maintaining that good standing is following the appropriate code of ethics (the AMA’s or, if different, the code adopted by the licensing jurisdiction). In the nineteenth century, before physicians were licensed, it was harder to tell the difference between a physician in good standing and an unethical, incompetent, or merely unconfirmed physician, but the distinction nonetheless existed. To be a physician, strictly speaking—not a mere counterfeit—was to be someone on whom patients, other physicians, and the public were entitled to rely to maintain medical standards, ethical as well as technical. A physician was someone who truthfully professed those standards. In the same way, a physician today is someone who truthfully professes today’s version of those standards.^{Footnote 3} Licensure (the requirement of a license meeting certain conditions) is one of those standards.

My point here is not that the knowledge, skill, and experience of an unlicensed physician may not give her obligations beyond those of an ordinary person. Knowledge, skill, and experience are always relevant to assessing, for example, our duty to render aid in an emergency. A swimmer will have a reason to rescue a drowning child that a nonswimmer will not. My point is that any such obligation belongs to ordinary morality, not to professional ethics. Beyond even these obligations from special knowledge or skill is another sort of obligation, one that only licensed physicians have: the obligation to follow the medical profession’s standards.

The U.S. military seems to understand this. To be a military physician, one must be a licensed physician (i.e., licensed in some U.S. civilian jurisdiction). The military does not have its own physicians (medically trained but only certified by the military), though it certainly could have. Even physicians the military trains in its own facilities must obtain an ordinary medical license in a civilian jurisdiction before being allowed to practice in the military. The military wants to be able to say that the care its healers provide meets civilian standards.

The interrogators (or, at least, their superiors) also understand the importance of the profession’s standards. The interrogators want physicians present, not merely someone with medical training, skill, and experience, so that they can say, “There is always a physician present.” They do not want to have to say that the person present was “medically trained” but not a licensed physician. The licensed physician guarantees something that mere medically trained personnel, however knowledgeable, skilled, and experienced, cannot—maintenance of medical standards, ethical as well as technical.

The term “physician” functions much as does a trademark. Those who claim to be physicians without being entitled to make that claim are engaged in a deceptive practice. They are seeking to have the benefits of the label without satisfying the requirements the label presupposes. They are seeking to pass for what they are not—members of the medical profession.

A profession is not a mere collection of experts in some field. It is a cooperative undertaking. Each member of a profession contributes to the profession by following its standards, even when doing so is burdensome. The overall reputation of the profession depends (in large part at least) on individual members following those standards. Each member benefits from that general reputation, provided that the reputation is for meeting a higher standard, that is, for conduct beyond what the law, market, morality, and public opinion would otherwise require.

A profession is voluntary in at least three respects. First, no occupation need organize as a profession. Members of a possible profession may instead remain an honest occupation in which the members generally do as the law, market, morality, and public opinion require. Medicine was once such an occupation and could be again, as herbalists and plumbers still are. Second, even when an occupation is organized as a profession, membership in the profession is voluntary. So, for example, no one has to be a physician. Indeed, joining a profession is generally difficult. One must study hard, pass various tests, and otherwise win admission. Third, even after winning admission, no one has to remain in the profession. Physicians are free to give up the license whenever they choose. They need only see current patients into safe hands.

Professions are connected with morality in at least three ways. First, what professions claim to do is morally permissible. Morality provides a baseline. There can be no profession of thieves or assassins. Second, each profession is defined in part by a moral ideal, that is, an end (or set of ends) that all rational persons would (at their rational best) recognize as desirable enough, all else equal, to support, at least in minor ways (by providing resources for its pursuit, praising those who follow it, or at least not interfering with its pursuit). The medical profession seeks, insofar as possible, a state of affairs in which disease and disability are absent. That is a state of affairs all rational persons recognize as good (all else equal). That is part of medicine’s moral ideal.

By professing to be a (licensed) physician, a physician claims whatever benefits come from the profession’s reputation for maintaining its special standards. Some of those standards are technical: for example, the way a surgeon is supposed to make a certain incision or an internist is expected to record medical tests. Some standards are ethical: for example, the obligation to “report physicians deficient in character or competence.”^{Footnote 4} The distinction between technical and ethical standards is unimportant here. All such standards should (and generally do) share a feature that gives professional standards a third connection with morality, indeed, that makes them a part of morality in much the way that a valid promise is; professional standards add to a professional’s moral requirements.^{Footnote 5}

Medical standards, whether technical or ethical, are morally permissible rules that each physician (at least at her rational best) wants every other to follow, even if that means having to do the same. Each physician has good reason to want the rest to follow those standards: Those standards encourage patients to prefer physicians to other “healers”—aroma therapists, Christian Science practitioners, homeopaths, chiropractors, and so on—insofar as they do.

Because (as just explained) a profession is a voluntary, morally permissible cooperative practice in which each doing her share contributes to the benefits the others receive, those members of the profession who violate its standards are engaged in cheating (i.e., violating the rules of a voluntary, morally permissible cooperative practice). Because cheating is (all else equal) morally wrong, violating the standards of one’s profession is (all else equal) morally wrong. To violate the standards of one’s profession is unprofessional.

The AMA’s Code of Ethics—the combined Principles of Medical Ethics and Opinions—is among the profession’s standards. Whether or not they are members of the AMA, physicians have a moral obligation to follow those standards, an obligation only licensed physicians have. So, the problem that Allhoff takes to be so difficult—that is, explaining why a licensed U.S. physician participating in hostile interrogation is, all else equal, acting immorally—is not difficult at all. The conduct is morally wrong, all else equal, because (as Heilig in effect says) it violates a (morally permissible) AMA standard (as well as several international medical standards). Opinion 2.067 says, “Physicians . . . must not participate in torture for any reason.”^{Footnote 6} Because that opinion defines “torture” as including any “deliberate, systematic, or wanton administration of cruel, inhumane, and degrading treatments,” participating in hostile interrogation of any sort is unprofessional conduct for a physician. What is unprofessional (a kind of cheating) is, all else equal, morally wrong (because, all else equal, cheating is morally wrong).^{Footnote 7}

The proviso “all else equal” may seem to give someone like Allhoff a way around this professional condemnation of hostile interrogation. It does not. The proviso is simply a recognition that moral rules often have exceptions. For example, though killing in general is morally wrong (a violation of the rule “Don’t kill”), there are exceptions—self-defense being the least controversial. The moral rule “Don’t cheat” may well have exceptions too. The question is whether any of those exceptions apply in the context of hostile interrogation. I can think of no such exception. (The principle of nonmaleficence is not, contrary to Allhoff’s claim, enough to provide an exception—as I will soon explain).

This argument for professional obligation is general across all professions. A profession may, in addition, have a special argument for the moral authority of its professional code. For example, in many jurisdictions, lawyers must be sworn in before they are licensed. Because all such special sources of obligation are both parochial and redundant (doubling the moral obligation in a particular jurisdiction without changing its content), I shall say no more about them here.

How might a defender of physician participation in hostile interrogation like Allhoff respond to this argument? There are at least six possibilities. First, he might argue that opinion 2.067 does not represent an actual standard, that it is in practice “dead.” There are several reasons to reject this response. One is that (as far as we know) few physicians violate the standard (even when they have the opportunity). Another is that most physicians will, if asked, support the standard (especially after giving the question careful thought). Yet another reason to reject this response is that physicians have good reason to support the standard insofar as they, especially physicians working for the police or military, want their patients to trust them, not to look on them as potential assistants in their mistreatment. A last reason worth mention here is that physicians do seem not to want to be known as colleagues of those who mistreat people. The few physicians involved in hostile interrogation seem to avoid public notice if they can.

I list these reasons not to prove that the rule in question is right but to provide evidence against supposing it dead. (The existence of good reasons for a rule is, all else equal, a good reason to believe that reasonable people follow the rule.) I am now replying to an objection to the argument from profession, not making a positive argument.

A second response someone like Allhoff might make to the argument from profession is that the standard in question, though not dead, is nonetheless inappropriate for inclusion in a code of ethics. The problem with this response is that it is hard to see how to defend it. If we accept the analogy with promises, the only standards that can be ruled out in principle are the morally impermissible ones, and hardly anyone would claim that, all else equal, prohibiting physician participation in hostile interrogation is morally wrong. Any argument for physician participation in hostile interrogation must depend entirely on special cases, in which the balance of consequences favors allowing a physician to participate. That the standard in question might be ineffective in serving the moral ideal that medicine is supposed to serve, or not be the best means of serving it, or be damaging to some other ideal, would be a good reason to change the standard but would not automatically render the standard morally invalid. If a foolish promise can nonetheless be a valid promise, until withdrawn or amended, so a foolish professional standard can be binding until repealed or amended.

The third response to the professional argument is that physicians who are helping with hostile interrogation are not acting as physicians. They have put down one role (that of physician) to take up another (that of interrogator’s assistant).^{Footnote 8} This response would be devastating if the argument from profession relied on the patient-physician relationship or another role-based premise. The physician helping with hostile interrogation may well have before her someone who does not consider her to be his physician, and she may well agree that he is not, because she thinks of herself as having put aside the role of physician (as indeed she may have). She is, she may think, like a physician assisting a butcher or veterinarian. That is something a physician may properly do, but not as a physician.

Nonetheless, that response is irrelevant here. The argument from profession does not depend on the existence of any role or relationship but only on the person in question being a physician subject to opinion 2.067. Because that opinion applies to a physician in whatever role she is playing, the only way for a physician to avoid violating it while assisting at a hostile interrogation is to leave the profession, that is, to resign her license to practice. She could then do what she likes with her medical training without breach of medical ethics. What she cannot do without violating medical ethics is participate in hostile interrogation while a physician (even while acting in another capacity).

Fourth, a defender of physician participation in hostile interrogation might respond that nothing in the argument from profession shows that medically trained persons, whatever called, could not (in a world not too different from this one) morally participate in hostile interrogation. The argument from profession is contingent on what U.S. physicians happen to profess. Opinion 2.067 was only adopted in 1999. Before then, physicians could (it might be argued) participate in hostile interrogation.

To this, I reply that medical ethics, though contingent, is not that contingent. Medicine has standards in addition to opinion 2.067. For example, the AMA’s Principle I requires physicians to “be dedicated to providing competent medical care, with compassion and respect for human dignity and rights.”^{Footnote 9} That standard was in place well before 1999. Phrases like “compassion and respect for human dignity and rights,”^{Footnote 10} though quite general, still seem to rule out assisting with hostile interrogation. It would be hard to be so dedicated and not object to hostile interrogation or withdraw rather than assist in any way. In addition to AMA standards, various international medical standards, decades old, also seem (as Heilig notes) to rule out participation in hostile interrogation.

But, assuming none of these standards were in place, then a physician participating in hostile interrogation would, I agree, not be acting unprofessionally—that is, not be violating medical ethics. In a world not very different from this one, medical ethics might allow physicians to participate in hostile interrogation (including torture). Medicine does not have a Platonic nature that forever rules out such participation. Medical ethics is contingent on what physicians decide their standards should be (provided, of course, that the standards are morally permissible). That there are good—or even decisive—arguments for a standard does not, as such, prove the existence of a morally binding standard. A code of professional ethics is a work of history, not of metaphysics or abstract logic.

That admission should, however, not much comfort a defender of physician participation in hostile interrogations. On the one hand, it leaves open the question of the moral status of participation in hostile interrogation, that is, participation whether or not the participant is a physician. So, only if the defender of physician participation could show that morality allowed participation in hostile interrogation would there be no moral grounds for objecting to physician participation. On the other hand, there is, contrary to what we have just been assuming, an ethical bar to participation. The professional standards do in fact prohibit participation in hostile interrogation. The argument from what is possible is not relevant to practice in this world (but only to practice in a world somewhat different from this one). In this respect at least, the argument from possibility is simply beside the point—as long as the point is what requirements actually apply to physicians.

Fifth, a defender of physician participation in hostile interrogation might try to get around the argument from profession by appeal to “principles of medical ethics.” They are (he might say) a morally better guide for physicians than the profession’s code of ethics. So, for example, Allhoff says: “At least in some cases, torture would not be worse than death. Therefore, resuscitating someone merely so she or he can face more torture does not necessarily violate the principle of nonmalfeasance [nonmaleficence] if, absent resuscitation, she or he would have been even worse off (e.g., dead).”^{Footnote 11} The physician should, Allhoff concludes, participate in torture (and, we should add, other hostile interrogation) whenever participating would help the individual “patient” overall (e.g., when the “patient” is subject to torture that will, as a matter of fact, end in a few hours or days, and the alternative is dying now). Participation in hostile interrogation would then not be “doing evil.” Medical ethics cannot (Allhoff concludes) rule out physician participation in hostile interrogation.

This conclusion is no surprise to anyone who understands that the principle of nonmaleficence—and its twin, the principle of benevolence—are essentially act consequentialist. The act preferred depends on the exact facts assumed, because these determine the balance of good and bad consequences, actual or probable, that would result from so acting, which is all that is relevant according to those two principles taken alone. The other two principles commonly taught beside them—justice and autonomy—tend to compensate for this consequentialism—as Allhoff’s substitutes (confidentiality, honor, and loyalty) do not. But even without Allhoff’s substitution, there would be a problem of how to weigh one principle against another. Do considerations of autonomy or justice outweigh nonmaleficence here? How are we to tell? Medical ethicists generally provide little help with such questions.

That is no small failing. The point of having more specific standards, such as the AMA’s code, is (at least in part) to rule out certain otherwise reasonable conclusions that some might draw from ordinary moral principles, when others might draw different conclusions. The point of having more specific standards is to coordinate the profession’s conduct whenever such coordination seems likely to produce better results overall than the uncoordinated conduct of individuals.^{Footnote 12} Even if the nonmaleficence argument were sound in some situations (as perhaps it is), it would not follow that patients in general are better off overall if physicians reason in the way that it requires. The systematic absence of physicians from hostile interrogation might substantially reduce hostile interrogation and thereby reduce every patient’s risk of mistreatment—a better result than letting each physician decide for herself. The systematic absence of physicians might achieve that by, for example, denying those engaged in hostile interrogation whatever legitimacy they derive from being able to say that physicians are present. Denying interrogators that legitimacy may thereby eliminate the very hostile interrogations that the physician’s presence is supposed to ameliorate.

Sixth, and last, a defender of physician participation in hostile interrogations might respond that the argument from profession establishes only that a profession creates a moral obligation of some sort. The argument does not show that the obligation so established carries much moral weight. Professional obligations might still be morally trivial, easily overridden by an important moral principle such as nonmaleficence—say, when a physician is asked to look after the medical needs of someone undergoing hostile interrogation.^{Footnote 13} To this response, there are at least two replies, one simple and the other involving larger issues of moral theory.

The simple reply admits that some conventions, for example, those relating to the proper placement of tableware, may indeed carry little (or no) moral weight—except perhaps in special circumstances. But (the reply continues) that admission is not relevant to professions, such as medicine, in which a breach even of a convention that might seem analogous to the placement of tableware—such as the use of certain terms for certain instruments or acts—can easily be the cause of serious injury or death. The sixth response is therefore beside the point unless the convention in question, the medical profession, is of little more moral importance than the placement of tableware, something that seems quite unlikely.

That is the simple reply. The other reply accepts an interpretation of the distinction between rules and principles that Ronald Dworkin made popular.^{Footnote 14} A rule either applies to an act or does not. If it applies, it decides the act’s rightness or wrongness. Rules have exceptions (as already noted). When an exception exempts an act from the rule, it makes permissible what would otherwise be impermissible (under that rule). An exception provides the basis for justification of an act otherwise unjustifiable (under the rule). The moral weight of a rule matters only if the act is wrong (according to the rule). The rule’s moral weight (or, rather, the weight of the principles giving reasons to have the rule) then has a part in evaluating excuses and determining what response is appropriate. The less weighty the rule, the easier it is to offer a good excuse for breaking it or a decisive reason to impose a penalty at the lower end of the penalties allowed. But principles cannot, as such, weigh against a rule (though they may provide reasons to add an exception or limit application of the rule in some other way). The sixth objection thus seems to rely on a mistake about rules, that a principle can overrule a rule (justifying what the rule, even with all reasonable exceptions, forbids). Of course, that is a mistake only if (something like) Dworkin’s interpretation of the distinction between rules and principles is correct, an assumption the defense of which is well beyond the proper scope of this article. For now, it is enough to see that the sixth response is far from making an obvious point. The argument from profession seems safe from it.

We may, I think, draw both a practical and a theoretical conclusion from the forgoing. The practical conclusion is that physicians do (as Heilig claimed) have a professional obligation to avoid participation in hostile interrogation, that the obligation is morally binding, and that no principle of medical ethics overrides it. The theoretical conclusion is that bioethics, including academic medical ethics, should pay more attention to the argument from profession than it has until now. The argument might provide important insight into other contested areas of bioethics, just as it has done here.