A. “The Law” on Constitutional Rights is Constituted Only by Explicit Description of the Very Right in Question in the Text of the Constitution

This is the view that if there is a constitutional right to use birth control, to engage in nonheterosexual sexual intimacy, to have an abortion, to marry a person of a different race, to march on the streets of Skokie, Illinois, carrying a neo-Nazi placard, and so on, then the Constitution must itself specifically so say. Simply to state such a view, of course is pretty much to refute it. The rights provisions of the Constitution were written in no such codelike fashion. The Eighth Amendment, for example, was written generally to prohibit “cruel and unusual punishment”—it was not written so as to explicitly prohibit the rack and the screw, the public stocks, the ducking stools, the clean-up of latrines, or any other specific forms of punishments that are considered unusual and cruel even though not specifically so described.Footnote ¹⁵ The general phrases of the Constitution would have no application if this were what was required for there to be a Constitutional right. Such generally specified rights require interpretation of general language by the courts in order to have any meaningful existence.

Since every lawyer knows this,Footnote ¹⁶ the seeming assertion of this view of “the law” by Roberts and his conservative colleagues can only amount to the rhetorical hyperbole of partisan advocacy. It can thus be true that “the Constitution simply does not speak to the issue of same sex marriage,”Footnote ¹⁷ “the Constitution contains no ‘dignity’ clause,”Footnote ¹⁸ the “freedom to exercise religion is—unlike the right imagined by the majority—actually set out in the Constitution,”Footnote ¹⁹ “The Constitution itself says nothing about marriage,”Footnote ²⁰ the liberty to marry whom one chooses is one of those “liberties the Constitution and its Amendments neglect to mention,”Footnote ²¹ and yet also be true as a matter of sound constitutional law that there is a right of gay persons to marry as part of a general constitutional right to liberty.

These empty rhetorical flourishes of the dissenters in Obergefell remind one of Warren Burger's claim in Bowers v. Hardwick that at issue in that case was the right to engage in “homosexual sodomy,” a right nowhere mentioned in the Constitution.Footnote ²² Such rhetoric was roundly rejected by Blackmun and Stevens in the dissents that eventually became the Court's view in Lawrence v. Texas.Footnote ²³ One can have a right to practice nonconventional modes of sexual intimacy without the Constitution saying so, just as one can have the right to march in Skokie without the First Amendment saying so. One can hold the contrary view only on pain of none of the general rights-conferring clauses of our Constitution having any concrete application.

B. The “Law” on Constitutional Rights Is Constituted Only by the “Plain Meaning” of Constitutional Phrases such as “Liberty . . . without Due Process of Law”

This view concedes that “the law” includes more than the text of the Constitution; it also includes interpretation of that text and application of it to concrete situations. Yet, the argument is, when this application is done, the rights individuals have are limited to those plainly given by the accepted, conventional meaning of words such as “liberty” and “equal protection.”

Herbert Hart once opined that “law,” in some “central sense” of the word, referred to a “core of settled meaning” of the words appearing in authoritative legal texts.Footnote ²⁴ Often this counsel is taken to be a “plain meaning” approach to legal interpretation.Footnote ²⁵ So taken, such an approach runs into well-charted difficulties needing little rehearsal here.Footnote ²⁶ For whatever the merits of such an approach to the interpretation of statutory texts when those texts have the specificity of the rules of baseball,Footnote ²⁷ it seems hopeless as an approach to interpreting the grand clauses of the Constitution. For few would contend that the conventional usage of the word “liberty” is such that a “plain meaning” can be extracted from it sufficient to decide constitutional cases.Footnote ²⁸

C. The “Law” on Constitutional Rights Is Constituted by the Intended Meaning of the Framers

Linguists rightly distinguish what people say from what they mean (or “intend to say”).Footnote ²⁹ Thus the injunctions, “say what you mean” and “mean what you say,” are not empty tautologies. Likewise, linguists distinguish both what is said and what is meant, on the one hand, from what is understood by the audience of a statement, on the other.Footnote ³⁰ From these linguistic insights comes the “originalist” interpretive methodology. Thus Thomas construes the Fourteenth Amendment's use of “liberty” by what the Lockean-steeped framers meant and their audiences understood when the framers wrote and ratified the word.Footnote ³¹ Scalia finds determinative what the framers were understood to have intended by way of application of the words they chose to put into the Constitution:

Unlike both text and the plain meaning of text, it is controversial within constitutional theory whether the intent of the framers of constitutional language or the understanding of their original audience can be authoritative for later generations of those living under the Constitution.Footnote ³³ Advocates of there being a “living Constitution” often put aside originalist meanings in favor of Chief Justice Earl Warren's “standards of decency that mark the progress of a maturing society” kind of meaning.Footnote ³⁴ Even among so-called “originalist” theorists, however, many would disagree about the finality or determinativeness that Scalia attaches to the particular intentions or beliefs about particular practices of the framers and their audiences.

But suppose one gave all the credence to framer's intent and original understanding in determining constitutional interpretation that Justices Scalia and Thomas have wanted. Such an exclusively historical approach surely requires that one take the history as one finds it. And the history relevant here shows that crucially important framers of 1791 (for the Fifth Amendment's due-process clause) and of 1868 (for the Fourteenth's) were believers in natural rights in general and in a natural right to liberty in particular.Footnote ³⁵ Given such beliefs in the reality of the rights to which they intended their language to refer, such framers of the relevant constitutional texts used language the way all language users usually use it: to refer to entities whose actual nature gives the meaning and the extension of the words used.Footnote ³⁶ Such users may have their own ideas about the kinds of things to which their words refer. But if they use language the way the rest of us use it, they do not intend their ideas about acceptable applications to fix the meaning of the words they employ. They rather intend each audience to apply its own best understanding of the nature of the thing to which reference is made—even when that understanding differs from that of the original speaker and/or his or her original audience. That is the lesson of history that should go under the label “originalism,” not Scalia's and Thomas's conventionalist version, which is not faithful to the actual semantic intentions of the framers.Footnote ³⁷

D. The “Law” on Constitutional Rights is Constituted by the Distinctively Legal Meaning Imparted to Constitutional Phrases by Precedential Decisions of the U.S. Supreme Court

Charles Evans Hughes once famously opined that “the Constitution is what the judges say it is.”Footnote ³⁸ As is often pointed out, this confuses finality with correctness, power with obligation. But even when the needed corrections are made, it is indisputable that prior interpretations of the Court have some role in constraining its present interpretations.

There are two dimensions to precedent that have long been distinguished in jurisprudence,Footnote ³⁹ both of which have a bearing here. These are the dimensions of holding and of weight. As to holding, it is uncontroversial within the practice of horizontal precedent that no later court should be bound by a precedent court's characterization of the breadth of its own holding. Subsequent courts reformulate such holdings by narrowing, broadening, and distinguishing as they see fit. One upshot relevant here is that one cannot with a straight face dismiss the Griswold v. Connecticut line of cases as being irrelevant to construing “liberty” as used in the due-process clauses, just because some of the early opinions in that line of cases speak of a right to “privacy” rather than a right to “liberty.” When Roberts proclaims that “the privacy cases provide no support for the majority's position, because petitioners do not seek privacy,”Footnote ⁴⁰ he allows the deciding court's characterization of its holdingFootnote ⁴¹ actually to be its holding, contrary to this general wisdom about precedent. Every serious student of American constitutional law knows that she must penetrate the Lochner-inspired smokescreen (of privacy rhetoric) in this Griswold line of cases to see the true value being protected, which was liberty.Footnote ⁴² Treating alike cases that are truly alike, the essence of stare decisis, demands no less.

As to weight, it is also noncontroversial that precedent cases do not have absolute weight for subsequent decisions by the same court (so-called “horizontal,” as opposed to “vertical,” precedent). So even when (unlike Obergefell) there are precedent cases squarely on point, they cannot by themselves determine the correctness of an interpretation of the Constitution. Even Justice Scalia recognized the point when he acknowledged in Planned Parenthood of Southeastern Pennsylvania v. Casey that overruling a precedent such as Roe v. Wade involved a balancing of the degree of error the precedent represented against the conservative values that urge against change.Footnote ⁴³

E. The “Law” on Constitutional Rights is Constituted by “Tradition”

There is now something of a tradition for the Chief Justices that succeeded Earl Warren to extol reliance on “tradition” in interpreting the Constitution. Warren Burger extensively relied on tradition in Bowers to disavow there being any constitutional right to engage in “homosexual sodomy”;Footnote ⁴⁴ William Rehnquist formulated his defense of denying any right to physician-assisted suicide in Washington v. Glucksberg by relying on traditional condemnation of suicide;Footnote ⁴⁵ and Roberts relied most heavily on tradition as his basis for denying gay marriage rights in Obergefell.

One worry about relying on tradition in constitutional interpretation is the elasticity and indeterminateness of the concept. In many cases, indeed, there is room for doubt about the precision with which the content of any given tradition can be specified. “Tradition” as here used refers to a form of conventional belief. As such, it is subject to four indeterminacies of meaning and to a lot of evidentiary doubts, all of which contribute to the elasticity and indeterminateness of the concept.Footnote ⁴⁶ As to the indeterminacies in meaning: (1) how unanimous an agreement must there be in the beliefs of a people to constitute a tradition? (2) Who are the people whose beliefs count in constituting a constitutionally relevant tradition—U.S. citizens, U.S. residents, all “English-speaking people,”Footnote ⁴⁷ all peoples in the world, some educated elite within that world,Footnote ⁴⁸ and so on? (3) How long must the consensus exist through time in order to constitute a tradition? And (4) at what level of specificity in the description of the objects of traditional belief should one be operating when using tradition as a mainstay of constitutional interpretation? As to the evidentiary problem: How does a court (bereft as it is of the opinion-polling techniques of social science) ascertain what a group of people believes or has in the past believed?

These indeterminacies and evidential uncertainties are real enough in many cases but not the same-sex marriage cases. Roberts, Scalia, Thomas, and Alito are, by and large, right in their dissents in Obergefell: the clear, long-standing, and well-nigh universal tradition has been for marriage to be the union of a man and a woman. Even societies where gay relationships were without stigma did not recognize gay marriage.Footnote ⁴⁹ The pertinent question for Obergefell, thus, is not the determinacy and the evidential well-foundedness of tradition; it is rather, the question of the authority of tradition as a major determiner of the correctness of a constitutional interpretation. With the text of the Constitution, the plain meanings of the words used in that text, and the prior interpretations of that text by precedent cases, there is little doubt that such items have a place in constitutional interpretation. With framers’ intent and original understanding, there are at least plausible (if controversial) theories of the Constitution's authority that make these originalist items also legitimate ingredients in a good theory of constitutional interpretation. But of what force for judges is the fact that an institution such as marriage has been thought of—even universally thought of—in one way rather than another?

Roberts tells us that the authority of tradition in this context rests on “the need for ‘judicial restraint.’”Footnote ⁵⁰ It is, he intones, “the need for restraint in administering the strong medicine of substantive due process”Footnote ⁵¹ that requires the Justices to eschew their own best insights about what liberty requires in favor of the traditional view of the matter. Yet if the fear is one of judicial arbitrariness and judicial power, why use tradition as the restraint? Any standard not requiring judges to make evaluative judgments will serve as well to restrain judicial power. Indeed, a crisp and clear standard (such as the height, weight, or hair color of the litigants) would serve a restraining function even better. So it cannot be simply the restraining of judges that gives tradition authority in constitutional interpretation. There must be some more positive reason for tradition's authority in this context.

Four possible reasons come to mind. One is the kind of reason available only to the metaethical relativist. A relativist believes that natural rights exist when but only when some group of people believes that they exist. Rights, like all moral properties, entities, and relations, are relative to conventional belief—thus the name “relativism.” On the relativist view, it requires some significant consensus of belief before there can be a tradition and thus before talk of natural rights makes any sense. There being a tradition in favor of a liberty right of gays to marry, on this view, is essential for there being any such right at all. Insofar as constitutional rights protect natural rights, therefore, tradition has pride of place in determining what those constitutional rights are.

Yet relativism is not a view that is or can be accepted by decently educated people.Footnote ⁵² It is, as my late colleague Bernard Williams once wrote, a view pretty much limited to sophomore undergraduates—and even then, only before their first ethics class.Footnote ⁵³ It is fundamental to the moral experience of every one of us that we sense the (always potential, sometimes actual) divergence of what we believe is right from what we believe some group believes is right. The relativist's assertion that if Nazis alone had survived World War II, Nazism would be right, is anathema to almost everyone's sense of morality.

A second route to the authority of tradition is that of the utilitarian. On one well-entrenched version of utilitarianism, what most people most prefer constitutes what it is good or right to do.Footnote ⁵⁴ Or as Holmes, an avowed utilitarian put it, “I am so skeptical as to our knowledge about the goodness or badness of laws that I have no practical criticism except what the crowd wants.”Footnote ⁵⁵ A preference utilitarian might well regard tradition as authoritative because it is an index of utility. If most people have a traditional belief that marriage requires a man and a woman, that must be because they prefer that marriage be so limited; and no one can have a right contrary to such group preference because rights are constituted by such group preference.

Like the metaethical relativist, those who espouse utilitarianism as their exclusive kind of ethics are committed to taking the social facts as they find them. If most people prefer that Jews be exterminated, then it is right and just that that be done; if women prefer not to be raped less than men prefer to rape them, then it is right that rape be decriminalized so that such justice can be achieved.Footnote ⁵⁶ Again, almost everyone's experience of morality is such that their individual moral beliefs are not hostage “to what the crowd wants.”

A third route to defending the relevance of tradition in constitutional interpretation, lies in political philosophy rather than in metaethics (the first route) or in ethics (the second). This is the route of the kind of majoritarianism that Joel Feinberg dubbed the “conservative thesis.”Footnote ⁵⁷ On this view the consensus that constitutes a tradition has authority, not because it is right but rather because the majority are entitled to live by it even if it is not right.Footnote ⁵⁸ The majority, on this view, have the right to be wrong so that even their erroneous moral beliefs have authority.

Such a view defends the use of tradition on the basis of democracy. Tradition, on this view, represents a kind of informal democratic decision-making process. There are “belief-votes” such that a long-term consensus has the force—the imprimatur—of majoritarian decision. The lack of authority of the political philosophies of Herbert Spencer or of John Stuart Mill (Roberts's two examples)Footnote ⁵⁹ then does not lie in the incorrectness of their views; it rather lies in the fact that there have not been enough adherents to their views for long enough to constitute “a tradition.”

The conservative thesis does not run counter to our fundamental experience of morality in the way that relativism or preference-utilitarianism does, for it does not counterintuitively equate the moral fact as to what is right and just with the sociological fact as to what many have long believed to be right or preferred to be done. Still, to rely on tradition for this reason is surely unjustified. If one cannot claim for a practice that it is right or morally required but only that this is how most of us like to practice it, from whence comes our authority over others to tell them that they, too, must practice as we do? What, because we need their conformity to our way of doing things in order to sustain the value we place on our doing them that way? Thus, if we deeply value our marriage to someone who is of our own race but of the opposite sex, we need everyone to be so married (and thus we can justifiably prohibit marriages to persons of the same sex but of another race)? So grounding the authority of tradition is a pathetic admission of the insecurity one must feel about his own valuations of what he does. As such, it is a poor ground for taking away another's freedom of action.

Second, the conservative thesis looks particularly unappealing in the context of judicial review. Think about it. Legislation is the output of the most democratically constituted institutions we have: legislatures. Overturning that output surely wants some justification other than that some other institution—“tradition”—is supposedly even more democratic. As John Ely urged years ago, it just makes no sense to overturn the products of the most democratic institution we have in the name of some vaguely and only arguably democratic institution of informal consensus, however long-running it may be (and that is assuming that courts are any good at ascertaining the content of that vague and informal thing, “tradition”).Footnote ⁶⁰

Third, our tradition of judicial review enshrines the protection of minority rights. Demonstrably, our Constitution was fashioned in part out of a fear of the tyranny of the majority.Footnote ⁶¹ Minority rights are protected by judicial review in our tradition because of such fear. Now ask yourself this question: Do you protect the rights of a minority best by interpreting the content of those rights in accordance with the views of the majority, that is, the “tradition?” Indeed, is it any protection at all against popular views that those same views are used to determine the content of the rights to protection? Majoritarian justifications of protecting minority rights by judicial review make no sense, on the face of it.Footnote ⁶²

Fourth and finally, an unchecked conservative thesis would be an unacceptably dangerous grounding of tradition's authority. Even if such social facts (that as many long believed, slavery is just, that death is a permissible punishment, that race-segregated schools are permissible) were some reason to continue such institutions, whether such institutions should be continued would surely also be a function of just how wrong or unjust those institutions are.Footnote ⁶³ And this latter fact is not a matter of social fact or of traditional judgment—it is a moral fact that each judge or Justice must decide for him- or herself as he or she decides how much weight to give to traditions.

The fourth route to endowing traditional beliefs with constitutional authority is Burkean in nature.Footnote ⁶⁴ This is the view that traditions, although not constitutive of what is right and just, nonetheless are evidential of what is right and just. As Burke said (in answer to Locke and Rousseau specifically on the right to liberty): What intellectual power do we as individuals possess that can compare with the collective wisdom of the ages? Or as John Roberts echoed after reciting the universal tradition for millennia of marriage being limited to cross-gendered relationships, “Just who do we think we are” to go against that consensus?Footnote ⁶⁵

Unlike the previous three views, there is surely something to the Burkean argument. If there are right answers to questions of morality(such as the proper extent of our liberty), one might well be quite humble about one's own ability to find those right answers. For whatever one can be right about one can also be wrong about, and the disagreement of other well-meaning people with one's own views can often be a rational indicator of one's own errors. Yet just as surely such epistemic deference can never be complete.Footnote ⁶⁶ No one, judge or otherwise, can completely defer to others’ moral judgments and still be a person rather than, say, a mere parrot. That almost everyone in Athens thought slavery to be just, that most people in the pre-Brown American South thought that segregated schools did not violate equality, that a solid majority have long thought death to be a permissible punishment, cannot preempt one's own judgment about such matters. Evidence is just that, evidence, and as such, the evidence constituted by tradition can be outweighed by evidence pointing in the opposite direction.

I conclude that there is a limited case for some reliance on tradition in constitutional interpretation, based on Burkean considerations. But this limited case does not justify preclusion of individual judgments by judges and justices of what it is that liberty truly demands. Indeed, even the combination of all five ingredients that we have considered—text, plain meaning, precedent, tradition, and original intent/understanding—cannot preclude such individual judgment. Often this is for reasons of indeterminacy—for often these materials, even in combination, are insufficient to determine unique decisions in individual cases. But even when these materials do determine a result—as tradition in particular would do on the gay marriage issue—it flouts the obligation of judges in constitutional cases to rely on those materials alone to the exclusion of their own best judgments about the rights persons possess.

Those who framed the relevant clauses in 1791 believed in natural rights, as Thomas points out in his dissent.Footnote ⁶⁷ They believed that a general right to liberty was one of those natural rights and was, indeed, for those like James Wilson, the centerpiece of the entire natural-rights philosophy. Their successors also thought they were referring to such liberty when they drafted the Bill of Rights and the Civil War Amendments in the 1860s. Both sets of framers understood that people would differ in their views about the content of such rights, the right to liberty included. Such difference of opinion would exist despite the belief of these sons of the Enlightenment that (in the words of Locke) “one can demonstrate the truths of morality with the certainty that one can demonstrate the truths of mathematics.”Footnote ⁶⁸ Although moral truths about the content of natural rights may be this “self-evident,” as Jefferson put it in the Declaration of Independence, some things that are self-evident are nonetheless “evident only to the wise.”Footnote ⁶⁹ That was why it was important to get well-educated, right-thinking citizens to act as judges.Footnote ⁷⁰ Those who are neither wise nor well motivated need not apply.

John Roberts and his fellow social conservatives on the Court rightly see that their job in cases like Obergefell is not to apply some established, conventionally accepted political philosophy. Roberts quotes Holmes's off-quoted remark that the “Fourteenth Amendment does not enact Mr. Herbert Spencer's Social Statics.”Footnote ⁷¹ Roberts then paraphrases Judge Friendly: “the Fourteenth Amendment does not enact John Stuart Mill's On Liberty any more than it enacts Herbert Spencer's Social Statics.”Footnote ⁷² And of course Holmes (and Friendly and Roberts) are right: Their job is not to apply Spencer or Mill—or Aristotle, John Rawls, or St. Thomas Aquinas, for that matter—to resolve questions about the constitutional meaning of “liberty.” What such judges do not see is what their job really is instead. They are not to apply others’ views about what liberty demands, even if those others are numerous, famous, revered, or just plain old. They are to apply their own best judgments about the contents of the rights the Constitution names but does not create, rights that, as Justice Douglas noted in Griswold, “are older than our Bill of Rights.”Footnote ⁷³ Their job is to do the political philosophy themselves to decide what rights persons possess, not pass the job off to others. The job of judges, as Herbert Hart once put it, is—first, foremost, and always—to judge. Footnote ⁷⁴ That that judgment invokes political philosophy in no way renders such judgment extralegal, beyond the law, or in any other way improper. It is nothing other than what the law—here, the Constitution—demands.

Judicial conservatives (who, in America at least, tend to also be political conservatives) often protest that this conception of constitutional judging flies in the face of democracy. Roberts's earlier quoted “Just who do we think we are?”Footnote ⁷⁵ appeals to modesty of power as much as modesty of judgment. Democratic processes were at work on the gay marriage issue, on this view, and the Court had no business interrupting that democratic working by imposing the views of five unelected Justices.Footnote ⁷⁶

Now this is a question of political philosophy! The question of whether judicial review is a desirable institution in a democracy is a much debated question examined by political theorists in America and elsewhere for two hundred years. The question is a general one about the basic design of a constitutional system. For it is judicial review as such that arguably challenges democracy, not just judicial review about liberty. And in truth, there is much to be said against judicial review when it is exercised to overturn the products of democratic law-making. Indeed, my longtime colleague and sometime debate partner on this very topic, Jeremy Waldron, has said quite a lot on this side of the question, proclaiming (as Waldron puts it) that the “right of rights” is the right to participate in a society's decisions as to what rights we have.Footnote ⁷⁷ In a formal debate with Waldron,Footnote ⁷⁸ I put forward the other side of this question, urging that an increment of rights protection can reasonably be expected in legal systems with judicial review.Footnote ⁷⁹ Waldron countered that even if most rights are better protected with judicial review than without it, necessarily one right is flouted by the practice, namely, Waldron's so-called “right of rights” to participate in the decision about the content of rights.Footnote ⁸⁰ And there issue is joined as to which rights are most important.

This is a close debate on a question in political philosophy about the outcome of which reasonable minds could differ. But in America this political debate is closed to serious legal argument. For the more than two hundred years following Marbury vs. Madison,Footnote ⁸¹ whether U.S. courts should exercise the “great power” of judicial review has been answered in the affirmative. That judicial review costs us something in the way of democratic decision-making but that that cost is worth bearing for the greater rights-protection such review can plausibly be thought to achieve, is foundational for American constitutional law.

Of course, by the arguments given above with respect to the weight of precedent, Marbury might be overruled. But unless and until it is, that our courts’ actions overturning the decisions of elected representatives costs us something by way of democracy is true but irrelevant. The cost to democratic decision-making is a fixed cost of judicial review. We decided to pay that cost when we decided over two hundred years ago to have the practice, so the fact that it is a cost in a case such as Obergefell is no argument to decide the case one way or the other. Each and every time judicial review is exercised, democracy is set back a bit. But if a right is truly protected by the decision of the courts, that thwarting has been judged to be a cost worth paying.

So when the dissent in Obergefell complains about the thwarting of the democratic process by “stealing this issue from the people”Footnote ⁸² —that can only be a repetitive assertion of the conclusion that there was no right there to be protected. For again, if there is such a right, we long ago decided that its protection was worth the cost to democratic decision-making. There is thus no independent “who decides” question lurking in cases like Obergefell.Footnote ⁸³ There is no independent question as to whether the decision of whether a liberty right exists allowing gay persons to marry, “should rest with the people acting through their elected representatives, or with those lawyers who happen to hold commissions authorizing them to decide legal disputes according to law.”Footnote ⁸⁴ In our system, that decision is for the courts. Roberts is right when he says that “The Constitution leaves no doubt about the answer”;Footnote ⁸⁵ he is just wrong about what that answer is. As John Marshall intoned in Marbury, “it is emphatically the province and duty of the judicial department to say what the law is,”Footnote ⁸⁶ and that includes the constitutional law on the rights that we have. That business of “saying what the law is” requires courts to delve into questions of political philosophy, which I now propose to do.

A. Do Natural Rights Exist?

My main concern in this section is to make the case for the existence of a natural right to liberty as well as to chart its content. Necessarily, for such a right to exist there must be such entities as natural rights. I thus briefly address this preliminary but foundational issue. There are actually two issues here. One is about objective morality generally. The other is whether the shape of that objective morality includes natural rights. I address each separately below.

B. What Do We Mean by “Liberty” (What Is Liberty)?

“What is” questions are notoriously fuzzy. They also invite frivolous answers disguised as profound truths, as in, “What is love?” Answer: “Never having to say you're sorry.”Footnote ¹⁰¹ Or, “What is truth?” Answer: “The lonely voice of youth.”Footnote ¹⁰² They also look linguistic in nature, as if they could be answered by describing patterns of idiomatic usage of the word “liberty,” whereas, in fact, what such questions really ask is what (in the normative world of valuable things) is the valuable item(s) to which the word should be taken to refer? Is there some good or goods plausibly in the vicinity of what people mean when they speak of liberty, and if so, what is the nature of that item or items?Footnote ¹⁰³

That said, there still are some usage facts needed to get us started. Yet one has to pick carefully here. “Liberty,” “freedom,” and the like are honorific words, words of virtually universal commendation.Footnote ¹⁰⁴ Like all such words (“justice,” “equality,” “fairness,” etc.), “liberty” has been used to name all manner of things, the only common denominator of which is that someone somewhere at some time has valued them. Indeed, Isaiah Berlin purported to discover two hundred senses to the word “liberty,”Footnote ¹⁰⁵ and I suspect that his list is incomplete. Consider in this regard what the word means when used in film scripts such as Braveheart,Footnote ¹⁰⁶ where William Wallace (in 1307 no less!) is scripted to shout “freedom” at his death. Or ponder what George W. Bush meant with his various “Freedom” speeches given throughout Europe during his second term as president.Footnote ¹⁰⁷

As with many other useful concepts, we need to be somewhat discriminating in our selections of the facts of usage with which we begin. Consider the following usage. Some years ago, the Wall Street Journal reported on a group of motorcycle riders who have had great success in lobbying state legislatures to repeal the requirement that helmets be worn when riding a motorcycle.Footnote ¹⁰⁸ The group at one time called themselves A Brotherhood Against Totalitarian Enactments, or ABATE for short. They called themselves “freedom riders”; they styled states where the helmet legislation has been repealed, “free states.” Their leader, Sputnik (his legal name), had “Free” tattooed across his forehead.

ABATE's usages of “free” point to a familiar concept of liberty. Liberty in such usages is the absence of restraint. In particular, the restraint in question (the absence of which constitutes liberty) is the restraint given by the prohibitions of the criminal law when these prohibitions are backed by coercive sanctions such as fines or imprisonment. This (with some amending enlargements) is the sense of “liberty” that I use throughout this article, for (as Thomas points out in his dissent in Obergefell) this is pretty plainly the sense of liberty referred to by those framers of 1791 and 1868 when they spoke of a natural right to liberty protected by the Constitution.Footnote ¹⁰⁹

This sense of liberty is, of course, what is usually referred to as negative liberty—“negative,” because its essence is given by an absence of something, namely, restraint. The contrast is to positive liberty, a liberty whose essence lies in the presence of something rather than its absence. Just what positive liberty amounts to has been a considerably contested notion in political philosophy ever since Isaiah Berlin's revival of the notion (and its contrast to negative liberty) in the 1950s and 1960s.Footnote ¹¹⁰ In the midst of such controversy, I simply stipulate the usage of “positive liberty” that I intend, a usage I believe to be a common if not universal one: positive liberty is that part of power or ability that we call an opportunity. More specifically, positive liberty is the opportunity to achieve through one's actions what one most wants to achieve. Opportunities, so understood, are to be distinguished from capacities, which deal with the internal characteristics of an individual. Opportunities exist when no external impediments (such as natural circumstances, economic conditions, or legal restraints) stand in the way of achieving what one wants to achieve. Such opportunities may thus exist even when the individual who has them lacks the autonomy (i.e., internal capacities) to take advantage of them.

One might think that “positive” and “negative liberty” are but two ways of describing the same thing. The thought proceeds from the insight that restraints are the enemy of abilities, in the sense that the nature of a restraint is to reduce or eliminate the ability of one to do that which he is restrained from doing. If a positive liberty to do some action A exists only when there are no restraints on the actor doing A—no psychological restraints, such as inner compulsions, no economic restraints, such as poverty, no restraints imposed by a private party, such as the threat of a thief, as much as no restraints of coercive legal sanctions—and if negative liberty similarly exists only in the absence of all restraints, then negative and positive liberty are one and the same thing. Yet positive liberty (as I define it) is not to be equated with the absence of all restraints but only with the absence of external impediments to action. Moreover, negative liberty is not to be thought of in the holistic way either. Rather, such liberty is conceptualized in terms of particular kinds of restraints the absence of which is liberty-enhancing. Thus Hohfeldian liberties are negative in that they are constituted by the absence of moral rights and obligations. More specifically, in the justly celebrated Hohfeldian logic, x is at liberty as against y to do A if and only if: (1) it is not the case that x is under an obligation not to do A; and (2) it is not the case that y has a (claim) right that x not do A.Footnote ¹¹¹

The restraints that we are interested in here are not the restraints on normative powers imposed by moral obligation, as they were for Hohfeld. Rather, we are interested in actual, physical power to cause results in the real world. Even so limited, in this context we should not be interested in the internal constraints imposed by psychological compulsions such as are present in addiction and obsessive-compulsive disorders. Nor are we interested in the real-enough restraints on opportunities imposed by poverty and other economic disabilities. For political liberty, negative liberty is the absence of coercive legal sanctions. So defined, negative political liberty is in no danger of being confused with positive liberty.

Once one disambiguates “liberty” so as to make clear that one is referring, in this context, only to negative liberty, there remains some vagueness in the concept even so disambiguated. More specifically, the notion of (legal) restraint has residual vagueness in it due to unclarity on the line between a restraint (which diminishes opportunities) and an offer (which enlarges them). Justice Thomas in Obergefell exploits this vagueness in his narrow construal of negative liberty. Thomas urges that the state laws in question did not actively prohibit same-sex marriage with positive sanctions attached to such prohibition—that would have been a restraint, Thomas concedes. Rather, the laws at issue in that case merely omitted to offer the marriage status to same-sex couples—a violation of positive liberty, perhaps, but according to Thomas, no violation of negative liberty because of the absence of any legal restraint in those laws.Footnote ¹¹²

Justice Thomas's dissent raises a good question (even if, by my lights, he gets the wrong answer to that question). What is the distinction between a threatened detriment—a “restraint” which diminishes negative liberty—and an omitted offer of a benefit, which does not? A number of distinctions suggest themselves here, some more promising than others.

One immediately occurring thought might be that the state acts when it restrains whereas it only omits to act when it fails to confer a benefit. Yet a moment's reflection should show that state action is not what is at issue here. To begin with, there are some conceptual worries. When a state says to some of its citizens that they may marry, but this saying by implication means that others may not, has not the state acted? Consider this old chestnut from the act/omission literature: “the nurse gave the patient one-half of the required infusion.” Act (giving half) or omission (not giving all)?Footnote ¹¹³ Or suppose the state initially says that all may marry, but then amends that act by saying that same-sex couples may not. Is this now state action, whereas before, when the state initially offered only cross-gendered marriages, it was not? Second, and more important, there is a normative worry: the act/omission distinction usually matters morally because of the differential stringency of negative versus positive duties for individuals.Footnote ¹¹⁴ For states, that significance is significantly lessened, for states are not being punished in proportion to the degree of wrong done (which is the main normative implication of there being differential stringency of duties). If the state has a duty to its citizens, it matters little (in morality, whatever the state of the “state action” doctrine in constitutional law may be)Footnote ¹¹⁵ whether that duty is a positive duty to act or a negative duty to refrain from doing some act.

Another possibility for a distinction here is that threats and offers come as self-labeling items, in the sense that if one says, “I threaten you with” or “I offer you,” these speech acts are respectively those of threat and of offer. Yet this is as far from the truth as thinking that morally binding promises are always and only made by saying, “I promise.” If I “offer” to give your book back to you only if you comply with some request of mine, I have threatened you with deprivation of your book; if I “threaten” you with nonuse of my book unless you comply with some request of mine, I have offered you use of my book if you do the action requested. In the first case I have diminished your opportunities, while in the second case I have increased them.

This suggests that a second distinction is really at work here: threats promise detriments whereas offers promise benefits. The continued loss of your book is a detriment; the continued use of my book is a benefit. Yet this distinction wholly depends on there being already in place a baseline of entitlements.Footnote ¹¹⁶ It is because you are entitled to the use of your own book that my conditional “offer” is really a threat; it is because you are not entitled to the use of my book that my conditional “threat” is really an offer.

In its conditional spending cases the U.S. Supreme Court has seen this point with admirable clarity. Although in dicta, Chief Justice Rehnquist noted in South Dakota v. Dole Footnote ¹¹⁷ that the formal characteristic (of an offer of the benefits of federal highway funds) was not dispositive of the issue of whether the states were being coerced into compliance with a federal regulatory scheme that the federal government was without constitutional authority to impose on the states; similarly, in Sibelius,Footnote ¹¹⁸ Chief Justice Roberts more recently held that an “offer” to the states of the benefit of federal Medicaid funds was really a coercive order (“a gun to the head”)Footnote ¹¹⁹ to have certain federally mandated medical insurance, despite the federal legislation's use of the form of offer of benefit rather than the form of threat of detriment.

Now apply this baseline notion of a restraint to the marriage issue. Justice Thomas concedes that in the miscegenation cases,Footnote ¹²⁰ the Virginia statute prohibited mixed-race marriages and attached criminal sanctions to the prohibition, and thus impermissibly took away negative liberty.Footnote ¹²¹ Yet suppose Virginia had had no such criminal statute. Suppose Virginia simply “offered” the marital status to its citizens under certain conditions, one of which was that the citizens be of the same race. That, according to Thomas’ analysis, would not have violated negative liberty.

This is to exalt form over substance, a bad idea even in law but a preposterous view of morality. God, if she existed, would not reason like a tax lawyer. The relevant baseline question is whether citizens of states are entitled to marital status. If they are, an omission to offer it should be treated as being as much a denial of negative liberty as is an outright prohibition on such marriages backed by criminal sanctions.Footnote ¹²² Keeping someone else's property that happens to come into your possession through no action of your own (i.e., not “offering” to give it to them when they ask for it) is the tort of conversion as much as was the taking of the property to start with. The baseline of entitlements (property or otherwise) determines the rightness of the conduct, not its formal characterization as an omission to offer.

In ascertaining where the baseline of entitlements is on issues such as marriage, there are two ways one might think that citizens are entitled to a certain legal status such as the marital status. One is via a theory about the minimal laws all states owe to their citizens. One might think, for example, that all states owe their citizens an enforcement mechanism for promises such as is constituted in our legal system by the law of contract.Footnote ¹²³ Only some such form of law as contract law can give citizens the power to bind themselves by their own voluntary undertakings, to rely on others so binding themselves, and to reap the benefits of nonsimultaneous exchange that these powers make possible. As Lon Fuller once urged, only the state through public law can give individuals these valuable opportunities.Footnote ¹²⁴

In the Lochner era freedom of contract was a major ingredient in constitutionally protected (negative) liberty. As Justice Pitney once said, in one of Lochner’s progeny: “Included in the right of personal liberty . . . [is] the right to make contracts.”Footnote ¹²⁵ Thinking that, it surely did not matter whether the form taken by interference with this right was criminal prohibition (formally a threat) or one of holding such contracts void as a matter of public policy (formally a refusal to offer contract law protection to certain wage/hour or other contract terms to certain classes of contracting parties such as minors or women). Refusing validity to a contractual term (i.e., omitting to enforce it) can be just as effective a denial of liberty of contract as criminally prohibiting the making of such contracts, for in either case the citizen cannot enjoy the benefits of the contractual relation that it is the obligation of the state to provide. Indeed, conceptually there is very little room for separating a voiding from a prohibiting here: prohibiting a certain kind of contract from being made in effect voids such a contract, and voiding it in effect prohibits it from being made. The only difference would seem to be that in the prohibition of such a contract, the contract must exist for an instant so that the prohibition can be violated, whereas in voiding such a contract it need never exist even for an instant.

The second way a state may be obligated to provide certain public goods such as a law of contract is by monopolizing the production of such goods even when it was not obligated ab initio to provide them. Take public roads. Suppose one were of the view that no state owes its citizens the creation and maintenance of public highways; such provision could have been left, one might think, to private toll road operators. Yet in fact most modern governments have undertaken this function, and exclusively so. Given this state of affairs, suppose some state then conditioned the validity of licenses (for both motorcycles and for the driving of them) on the wearing of a helmet by the rider. Formally, this is an omission to offer a license to unhelmeted riders, not a sanction-backed prohibition on helmetless riding. Yet this diminishes the negative liberty of motorcycle riders no less than does outright prohibition by the state, given the monopoly the state has created for itself on roads on which one can ride.Footnote ¹²⁶

I conclude that negative political liberty is indeed best conceptualized as the absence of legal restraint—so long as “restraint” is understood in the baseline-related way above outlined and not in Justice Thomas’ formal way.Footnote ¹²⁷ The next question is why we should think such absence of legal restraint (or provision of obligatory legal opportunity) to be good—indeed, so good that it merits constitutional protection.

C. Why Is (Negative) Liberty Good?

There are a number of reasons for thinking that negative liberty is a value, a thing to be protected and furthered by legitimate state policy. Such reasons mostly have to do with the instrumental goodness of citizens’ being free of state coercion. I focus on the instrumental goodness of negative liberty because of the difficulties attendant on giving much content to the idea that negative liberty is an intrinsic good.Footnote ¹²⁸ Perhaps one could think that all coercion is intrinsically bad, state coercion included, so that its absence could thus be seen as intrinsically good.Footnote ¹²⁹ Yet it is more informative and less controversial to treat negative liberty as an instrumental good.

Here are five suggestions as to what the more basic goods might be to which negative liberty can be a means.Footnote ¹³⁰ First, there is positive liberty itself. Negative liberty is one of the things necessary for positive liberty to exist. One has the opportunity (positive liberty) to do some action A only if the state does not prohibit A by criminal sanction. Freedom from governmental coercion is, of course, only one of the items needed to have such opportunities—one needs physical nature and the economic situation also to be favorable, for example, as leftist critics of negative liberty are so fond of pointing out. But negative liberty is needed too, so that if positive liberty is of value, then so is one of its necessary means, negative liberty.

Positive liberty is itself something good. One might plausibly view it as an intrinsic good. This would be the view that enhancement of opportunity sets is always good, period, at least up to the point at which one gets paralyzed by “opportunity overload.” Alternatively, one might think that positive liberty is the kind of universal, instrumental good that Rawls called a primary good—something so useful to the attainment of a wide variety of ends that no matter what those ends might be, liberty would be good for their attainment.Footnote ¹³¹ In either way, positive liberty is valuable, and so thus is one of its indispensable means, negative liberty.

Second, there is what might be called Millian autonomy (so-called to distinguish it from Kantian autonomy, which is discussed shortly). Millian autonomy is the hard-to-pin-down idea that human choice gains value when freely made. Even bad choices have value in this sense, if they are free choices. The usual metaphor is that of authorship: the autonomous person is a being who is author of his or her actions, a self-determining agent in some sense. Mill's contrast case is a mere machine such as a steam engine: although it can do many things, it is not the author of any of them, so its actions possess no value as autonomously chosen.Footnote ¹³²

It is no small matter to flesh out the ideas of agency, autonomy, and self-determination implicit in Millian autonomy. Not wanted is some metaphysically libertarian idea of free will.Footnote ¹³³ Even if all choices are fully determined by factors themselves unchosen by an agent, that agent's choices can be self-determined in the relevant sense(s). David Hume teaches us as much when he urges that the freedom worth wanting is not freedom from causation in our choices; rather, what is valuable is the ability to choose and to act on such choices in a way that satisfies our desires.Footnote ¹³⁴ Being able to get what you want is valuable, Hume urges. Mill's addition is to find valuable the possession of such causal efficacy to one's choices when the choices are second-order choices about the kind of person one wants to become. Reflection about such matters and critical self-examination, as opposed to parroting views of others and unthinking adoption of a tradition, are part of what is needed to be self-determining in the relevant ways for Mill.

If Millian autonomy is valuable, it is easy to see how legal coercion prevents its attainment. A coerced choice is not an autonomous choice expressive of the agent's authorship. The threat of the law no less than the threat of a gunman can rob our actions of any of those decisional process one rightly values as autonomous choice. Duress is an excuse in the criminal law precisely because of this fact.

The third value, Kantian autonomy, is to be distinguished from Millian autonomy. For Kant, an autonomous action is a right act done for the right reason.Footnote ¹³⁵ Kant's thought is that there is value not just in doing the right act but also (and mostly) in doing it for the right reason. The emphasis is on the motives for action, not the processes of choice. Although choice may well be involved whenever an act is motivated by reasons, valuing correct reasons (Kant) is quite different from valuing free choice processes (Mill).

Note that in the Kantian sense of autonomy (in contrast to Mill's sense), there is no value in “autonomously” chosen wrong action, nor is there value in “autonomously” chosen right action when that action is done for the wrong reason. No matter how free one's decisional process may have been, Kantian autonomy is not achieved in such cases, because the right reason does not motivate the right action.Footnote ¹³⁶

Again, if there is value in acting for the right reasons, it is easy to see how legal coercion can prevent the attainment of that value, for the avoidance of legal sanctions can easily supplant the more virtuous motivations for an action that might otherwise have moved the agent to act. Consider, for example, gifts to the poor. If Kant is right about the locus of value residing in reasons and not just in actions,Footnote ¹³⁷ charitable giving is valuable not just because of the act of transferring wealth to those in need but also because of the benevolence motivating that act. Legally coerce such acts, and the fear of sanctions will often supplant that benevolent motivation. In such cases the virtue of benevolent giving has become the necessity of paying one's taxes.

The fourth consideration showing negative liberty to be instrumentally valuable is preference-utilitarian in character. A preference utilitarian must take seriously the simple psychological truth that people generally prefer to make their own decisions free of coercion. Apart from some largely only hypothetical happy slaves, people almost universally prefer their decision-making to be free rather than constrained. Therefore, from this fact alone, a committed preference utilitarian must give some positive evaluation to negative liberty.

Finally, there are well-known costs attendant upon the use of the criminal sanction. I refer not only to direct enforcement costs such as police, courts, and so on, although these are omnipresent costs of criminalizing behavior and punishing it. There are also less obvious costs of criminalizing certain behaviors that are: (1) typically carried on in private and so witnessed by no one other than the participants; (2) strongly motivated so that criminalization only minimally deters the conduct in question; and (3) typically harmful to no one other than those who willingly participate. Costs of criminalizing such behaviors are well known.Footnote ¹³⁸ They include invasions of privacy in enforcing prohibitions with these characteristics; costs in terms of disrespect for laws that predictably will be regularly ignored; opportunities for selective enforcement of laws that predictably will be underenforced; and the costs of funding organized crime by the “crime tariff,” namely, the artificial restriction of supply of the goods or services in question in the face of relatively inelastic demand. These costs can be significant for behaviors with the three characteristics in question and add weight to the value of nonprohibition of such activities.

These values add up to a reasonable case for presumptively leaving people free of state coercion. They raise what is usually called a presumption in favor of liberty and thus against criminal prohibition.Footnote ¹³⁹ As a mere presumption, their case in favor of staying the state's hand can be overcome.

D. Skeptics About Any Right to Liberty

The argument so far is insufficient to establish that there is any moral right to political liberty. For one can be a moral realist in one's metaethics, a believer in natural rights in one's substantive ethics, and a subscriber to the values behind the presumption of liberty—and yet still deny that there is any natural right to liberty.

Skeptics about there being a natural right to liberty appear from both sides of the political spectrum. Such a general right to liberty was famously conceptualized by Justice Brandeis to be: “as against the government, the right to be let alone—the most comprehensive of rights and the one most valued by civilized men.”Footnote ¹⁴⁰ From the right side of the political spectrum, those social conservatives of a Burkean persuasion accurately see that such a general right to liberty—which is, after all, a right to be free of state coercion in all of one's actions—would be a recipe for anarchy. For a natural rights’ normative force is such that it cannot be overcome by countervailing considerations as easily as can a mere presumption.Footnote ¹⁴¹ Seemingly a general right to liberty therefore gives each of us the right to rob, pillage, and murder free of state coercion.Footnote ¹⁴²

There is a limited number of strategies available with which to blunt this obvious but powerful Burkean challenge. One is to lessen the force of rights (the right to liberty included) to be only “prima facie,” “pro tanto,” or presumptive.Footnote ¹⁴³ Yet if this strategy is pursued to the extent needed here, this would be to eliminate what is distinctive about rights-based approaches to morality.Footnote ¹⁴⁴

A second strategy is that professed on the left side of the political spectrum by egalitarian liberals such as RawlsFootnote ¹⁴⁵ and Ronald Dworkin.Footnote ¹⁴⁶ On this view, there is no general right to liberty; there are only discrete rights (“liberties”) to engage in discrete activities, such as the right to speak freely, the right to exercise the religion of one's choosing, and so on. This “liberties, not liberty” strategy seems to be motivated by three concerns. One is to blunt the force of the Burkean reductio of liberty to anarchy; this is done by restricting the domain of actions to which the right to liberty applies to just some discrete liberties, and these need not include the anarchical liberty to murder, and so on. The second proceeds from the thought that there is no value to be found in being at liberty to do immoral actions.Footnote ¹⁴⁷ The thought is that liberty is valuable only when and to the extent the action (one is at liberty to do) is valuable. The third is the desire of egalitarian liberals to erode any basis from which to complain about the kind of big government needed to sustain egalitarian programs of redistribution.

Yet this second strategy, too, is flawed. The balkanization of liberty done by the discrete-liberties approach defies the intuition that there must be something common to free speech, free exercise of religion, and so on, that makes those activities meriting of special protection from state coercion. As to the second thought held by egalitarian liberals here, liberty as such is generally valuable; one can see that by adverting to the values standing behind the presumption in favor of liberty. Of those values, only Kantian autonomy is lessened by the badness of the actions one is at liberty to do. The other values uniformly urge the state to stay its hand irrespective of the moral status of the action contemplated. And last, such staying of the government's hand is presumptively a good thing, however much it increases the burden of argument for those seeking to justify various redistributive schemes.

I thus find both of these answers to the Burkean reductio inadequate. Yet Burke deserves some answer. The two answers I pursue below are as follows. The first is to characterize the right to liberty as protecting one against certain reasons for the state prohibiting some action; the doing of such actions themselves is not protected by such a right, so there is no anarchical worry about a liberty right to murder, and so on. The second strategy is (in a way distinct from the “liberties, not liberty” strategy of the egalitarian liberals) to restrict the domain of actions protected by the right to liberty. This is to seek Mill's (and the Supreme Court's) proverbial “sphere” or “realm” of actions immune to legitimate state regulation. As long as such a sphere does not include all actions, there is no Burkean reductio transforming liberty into anarchy.

E. The First (or “Derived”) Right to Liberty: The Right to Have Behavior Regulated for Only Some but Not Other Reasons

The opinions of the U.S. Supreme Court in its liberty decisions (whether of the Lochner or of the Griswold era) move seamlessly between discussions of the goodness/badness of the reasons for certain prohibitions and discussions of the importance/unimportance of the actions/choices being regulated.Footnote ¹⁴⁸ The same ambiguity of topic pervades political philosophy. Take the most influential Western text on this topic, Mill's On Liberty. Footnote ¹⁴⁹

As is well known,Footnote ¹⁵⁰ Mill famously distinguished four kinds of actions that legislation might seek to regulate: actions harmful to people other than the actor and his consenting coparticipants; actions harmful to the actor himself; actions harmful to no one but that are offensive to others; and actions that need be neither harmful nor offensive to anyone but which are immoral to do. Mill's famous “harm principle” is the twofold principle urging, first (and positively), that regulating the first of these kinds of actions is the state's business, and second (and negatively), that regulating any of the other three kinds of actions is not the state's business.

As articulated both by Mill and by later Millians, both parts of the harm principle are susceptible to two related but distinct readings. On the first it states a criterion for the proper subject of legitimate state regulation: the regulation must regulate behaviors harmful to others. On the second the principle states a criterion for proper legislative aim: legislators should aim in their legislation to prevent harms to others. On the first the harm principle delineates a sphere of action immune to state regulation no matter what the reason might be for that regulation; on the second, any action can be regulated but it must be regulated for a certain reason, namely, in order to prevent harm to others, and not for any other reason.

It is easy to see why those two different interpretations of a legislatively limiting principle like the harm principle are conflated. After all, when Mill's targeted audience was legislators, the difference disappeared: legislators are of course to aim at whatever is the proper justification for legislation, because that justification will define the sphere of actions protected by liberty. If the correct justification for state coercion is to prevent actions harmful to others, then legislators should aim to prohibit all and only those kinds of actions, because that is the boundary of legitimate regulation. Yet Mill's targeted audience for the harm principle was not always or only legislators. And when that principle is viewed from the point of view of those who are not legislators—such as citizens—the two interpretations come apart. It is one thing for citizens to complain that legislation regulates behavior that in fact is not harmful to others; it is another to complain that although the legislation in fact regulates behavior harmful to others, preventing such harm was not the reason motivating the regulation.

I urge below that the right to liberty encompasses both of these kinds of restrictions on legislation. Citizens are morally entitled to do certain things free of state interference however that interference is motivated, and they are entitled to be free of certain reasons for being coerced out of doing anything. I examine the “subjective” or reasons-oriented right here. In the next section, I address the “objective” or action-oriented right.

I have long called the right to properly motivated state coercion, the derived right to liberty.Footnote ¹⁵¹ The label is because this right of each citizen is derivative of its more basic correlative, a legislator's duty to aim at certain ends as she legislates. One should think of such legislative duty as part of the theory of legislative role. Every principled legislator must use such a theory of their role to do her job, much as a theory of proper judicial role is something every principled judge must operate with as he decides cases.

It is perhaps not obvious why there is value in legislators aiming at the right things independent of them in fact regulating the right things. The case for this starts with the Kantian insight that the reasons for an action are important to assessing that action's moral worth.Footnote ¹⁵² Properly motivated legislation is thus of greater value than improperly motivated legislation of identical content. That case is strengthened if we turn to the content of the legislator's duty in this regard. Needed is some more determinate source for this obligation of legislators in order to give any content to it.

Mill's influence here has been both fortunate and unfortunate. The good effect of Mill's political philosophy here was to turn the theory of legislative role away from the purely proceduralist theories of majoritarians like Mill's contemporary, Sir James Fitzjames Stephens. For Stephens denied that there could be any defensible substantive principles limiting legislation in a democracy.Footnote ¹⁵³ The idea was that the only proper item for legislators to aim at in their legislation was to reflect accurately the views of a majority of their constituents. Mill, of course, stood for the opposite view. His was the (in Teddy Roosevelt's language) “bully pulpit” view of democracy. On this view, democracy works best when elected leaders lead by their own substantive vision of how things should go, including the limits of their own role in passing laws; then, if the majority does not like those views, despite the use of the “bully pulpit” of elected office to promote them, that majority can throw such leaders out of office.

Mill's less beneficial and downright unfortunate influence here stemmed from his presentation of his principle of proper legislative motivation, the “harm principle,” as if it were a relatively freestanding principle of political philosophy—freestanding in the sense that its justification lay in its own intuitive plausibility and not because of its relation to other, more basic principles of political philosophy. Yet surely one's most general principles about the proper ends of all legislation in a democracy should govern the formulation of any principles about proper legislative motivation for the use of state coercion.

Mill, of course, had such a general theory of proper legislative ends—he famously was a utilitarian. Mill was thus formally committed to the view that coercive legislation, like all legislation, is proper when it is aimed to enhance utility. For a subjectivist, preference utilitarian like Mill, that meant that coercive legislation should aim at maximizing the net sum of satisfied preferences.

How did Mill get from his general utilitarianism to his harm principle as a limit on the proper aim of coercive legislation? Recall that the harm principle has two parts, one positive and one negative. The positive part states that aiming to prohibit behavior harmful to others is legitimate; the negative part states that any other aim is illegitimate (including specifically legislating prohibitions to prevent behaviors harmful only to the actor, to prevent behaviors offensive but not otherwise harmful to others, or to prevent behaviors that are immoral). One should seek the connection of the harm principle to utilitarianism separately with respect to each of these two parts of the harm principle.

As to the positive part of the harm principle, the connection is pretty straightforward. Both threatening and delivering on threatened state coercion (i.e., punishing offenders) are evils for a utilitarian; doing so can, for him, be justified only by preventing something sufficiently harmful that the good of its prevention overbalances the harms of coercion and punishment. Preventing harm to victims can fit the bill, so long as the harms prevented are greater than the harm done to prevent them.

As has long been recognized, connecting the negative part of the harm principle to utilitarianism is trickier for Mill.Footnote ¹⁵⁴ Consider, first, behaviors that do not cause physical, economic, or even reputational injury to others, but such behaviors do disgust and offend people when they witness them or even learn less directly of their existence. Cannot—do not—many people strongly prefer that such behaviors not take place, and cannot the strength and near-universality of their preferences outweigh the harm caused by punishing those who so offend? Perhaps Mill wanted us to assume that the “soft” harm of psychic offense (prevented by punishment of offensive actions) can never outweigh the “hard” harms imposed by the threatening and the inflicting of those punishments; yet some very careful dance through various ideas of welfare (pleasure, happiness, preference satisfaction) would have to be done to make this assumption at all plausible. The latter-day Millian, Joel Feinberg, refuses to attempt this dance. Feinberg invites his readers (all of us) to take a ride on his “bus of disgust”; we are then to ask ourselves whether the truly disgusting behaviors done on his bus may not be legally prohibited.Footnote ¹⁵⁵ Despite his generally Millian outlook, Feinberg urges that some behaviors offensive to others could be legitimately prohibited by law. This is particularly true if some liberty is preserved by prohibiting such behaviors only when done in public but allowing them to be done in private (where none need be offended except those with vivid imaginations).

Or consider a second aspect of the negative part of the harm principle: denying that prevention of harm to the actor can be a legitimate reason for state coercion. What justifies Mill in his assumption that the harms to self prevented by paternalistically motivated punishment will always be less than the harms of threatening and inflicting that punishment itself? Mills backs this assumption with the supposed epistemic superiority of each individual vis-à-vis the state in judging what is harmful to him,Footnote ¹⁵⁶ but surely (as Herbert Hart points out)Footnote ¹⁵⁷ that is a very optimistic assumption about the degree of each person's self-knowledge. As Gerald Dworkin points out,Footnote ¹⁵⁸ Mill really seems to have a nonutilitarian argument in mind with which to back the antipaternalist part of his harm principle.

I would construe Mill's nonutilitarian argument here in the following way. Millian autonomy (one of the goods giving instrumental value to negative liberty) requires that we be the authors of ourselves in the sense that we shape our first-order desires in light of our second-order desires about the kind of person we want to be. Paternalistically motivated state coercion goes directly against such Millian autonomy. When the state is paternalistically motivated, it substitutes its judgment of what desires a person should have for the ones he in fact has and thus gives no value to the person's own second-order desires in this regard. In such a way, the state says it knows the kind of person each of us should be. That we have not chosen to be that kind of person is given little or no credence. This cuts the heart out of Millian autonomy, which bases self-authorship on each person's own, genuine, second-order desires.

Now consider the third aspect to the negative part of Mill's harm principle, that which enjoins prohibiting behavior because of its immoral status. The first thing to get clear is just exactly what Mill's principle was here. Although my friends and Mill scholars, Richard FumertonFootnote ¹⁵⁹ and David Brink,Footnote ¹⁶⁰ each tell me that Mill never said what I am about to say, I remain convinced it is what he was committed to thinking here.Footnote ¹⁶¹

Recall that for utilitarians like Mill and Bentham, utilitarianism was not just a principle of legislation; it was also a (indeed, the) principle of ethics.Footnote ¹⁶² For them, utilitarianism fully determined what was morally right and wrong (and not just what could justifiably be legally coerced). So when Mill rails against prohibiting behavior “for no better reason than that it is immoral,”Footnote ¹⁶³ he cannot be railing against using the law to coerce the true, correct morality—for by Mill's own lights, that morality is exclusively utilitarian in character. Mill's actual target here thus must be conventional morality, or mores, not critical morality. Moreover, insofar as such conventional moral beliefs differ from the dictates of utilitarianism, it must be Mill's view that such beliefs are in error. Consider Mill's ridiculing of the Americans for their 1850s suggestion of sending an expedition to the wilds of Utah to stamp out the practice of polygamy by the Mormon settlers there.Footnote ¹⁶⁴ Mill's scorn is for the belief that polygamy is morally wrong, and it is that moral error that gives force to his animus against using state coercion for this reason.

So construed, it is easy to see how this part of Mill's harm principle follows from his utilitarianism. For promoting nonutilitarian goods (i.e., moral error) cannot be justified as what should be aimed at by a utilitarian's principle of legislation.

I am spending this time on Mill's theory of proper legislative aim because it is an historically important illustration of what kind of theory could provide content to the derived right to liberty. Yet Mill's theory has flaws debarring it as the theory of the restraints liberty places on coercive legislation in a liberal state.

One of the flaws Mill's theory (of proper legislative reasons for coercion) does not have is that first enunciated by Sir James Fitzjames StephensFootnote ¹⁶⁵ and repeated many times since.Footnote ¹⁶⁶ This is the objection that “no action is without some harmful effects on others besides the participants.” Whatever the force of this objection when it is directed against using the harm principle to define a sphere of action immune to state regulation for any reason,Footnote ¹⁶⁷ it is ill-suited as an objection to using the harm principle to state the proper and improper reasons for the use of state coercion. For even if all actions have some harmful effects on others, that does not make such harmful effects the reason for legislative prohibition. Riding a motorcycle without a helmet may increase medical insurance costs for nonriders, for example, but even if that is so, a legislature may well prohibit the practice not for that reason but for the paternalistic reason that such actions are not good for motorcycle riders.

The real worries about the harm principle lie elsewhere. If we focus on the preeminent form that coercive legislation takes—namely, that of criminal prohibition—the problems with the harm principle originate in the problems for utilitarianism as a theory of punishment.Footnote ¹⁶⁸ Harm is the wrong focus for a theory of criminal legislation; moral wrongdoing is the proper focus. One can see this intuitively by thinking of the criminalization of wrongs that harm no one—such as abuse of a corpse, cruelty to animals, destruction of a dead person's reputation—and by thinking of harms that it is not wrong to cause—such as destruction of a competitor's business by opening up a better store nearby.Footnote ¹⁶⁹ One can see this theoretically by adverting to the problems faced by utilitarianism, such as the problems surrounding omission liability: on one form of utilitarianism (so-called “negative utilitarianism”), we have no obligations of positive aid to others,Footnote ¹⁷⁰ while on another, we have as many obligations to aid others as we have not to harm others. What utilitarianism has difficulty with is our actual practice (which is also quite intuitive): we generally have fewer and less stringent positive obligations, but we do have some. (Mill himself attempts to finesse this problem by making an undefined “common social duty” separate the latter omissions from the former).Footnote ¹⁷¹

More broadly, utilitarianism as a theory of punishment has deep and unresolved problems with it, in that it seemingly licenses both punishment of the innocent and nonpunishment of the guilty.Footnote ¹⁷² Needed is a notion of moral wrongdoing as at least a necessary condition of punishment to make punishments just (thus alleviating the first problem); making such wrongdoing sufficient for a just punishment alleviates the second problem. Two theories of punishment alternative to utilitarianism alleviate these problems: the purely retributive theory alleviates both; the mixed theory of punishment alleviates the first only (although many regard the punishment of the innocent as the most serious of the two). I touch on each theory of punishment briefly to illustrate the limits on permissible reasons for coercive legislation that each implies.

Retributivism is the view that criminal punishment is justly imposed on all those who deserve it but only on those who deserve it.Footnote ¹⁷³ Such desert, in turn, is cashed out in terms of two moral properties, wrongdoing and culpability: one's degree of desert is determined by the seriousness of the wrong done and the degree of culpability with which that wrong is done.Footnote ¹⁷⁴

Retributivism so construed leads to legal moralism—the theory that all and only moral wrongs should be prohibited by the criminal law—as its theory of legislation. It does this via two other assumptions.Footnote ¹⁷⁵ First, one must assume that a principle of legality is good independently of the good of retributivism. Such a principle requires prospective, clear, noncontradictory legislative rules before punishments may be exacted. Such a principle arguably side-constrains the attainment of retributive justice, allowing punishments giving such justice to be given only if the morally wrong act was legally prohibited beforehand. Such an assumption thus requires that, prima facie, all moral wrongs be prohibited by the criminal law, for otherwise wrongdoers doing those wrongs could not (under the principle of legality) receive their just deserts.Footnote ¹⁷⁶

The second assumption is that law as such does not obligate citizen obedience, not even prima facie.Footnote ¹⁷⁷ This means that the legislative enactment of a legal prohibition cannot make (morally) wrong an act not morally wrong before. The upshot is that citizens do no wrong in violating the criminal law unless that law is reflective of an antecedently existing moral wrong. Since retributivism bars punishment of the innocent as much as it requires punishment of the guilty, and because one is innocent if he does nothing morally wrong, therefore retributivism bans the prohibition of acts that are not morally wrong.

Retributivism with these two assumptions yields what I shall call the “moral wrong principle”: criminal legislation must exclusively aim at preventing or punishing moral wrongs, and this it can do by prohibiting all and only those behaviors that are morally wrong.

It is fruitful to compare the moral-wrong principle of the retributivist with the harm principle of the utilitarian. They both purport to give a necessary and sufficient condition of the justified criminal legislation at which all legislators should aim, one in terms of harm to others, and the other in terms of moral wrongs. It could turn out that the two principles are extensionally equivalent in the laws that they justify; if all moral wrongs consisted of actions harming others, and if all actions harmful to others constituted moral wrongs, then the two principles would justify aiming at exactly the same legislation. In fact, however, there is only a large overlap in the two notions, not a perfect congruence. There are wrongless harms, such as competitive injuries in capitalist economies; and there are harmless wrongs, such as cruelty to animals, abuse of a corpse, extinction of a species.Footnote ¹⁷⁸ So while largely coextensive, the two principles diverge in these two classes of cases.

The principles also diverge in their modes of reasoning to Mill's three negative corollary principles. Whether paternalistically motivated legislation is permissible according to the moral-wrong principle depends on the question of whether we are ever morally obligated not to cause harm to ourselves. While there are certainly diverging views about this in ethics, one plausible view denies that there are deontic duties to selfFootnote ¹⁷⁹ and thus denies that there can be wrongdoing involved in actions harming the actor only. On this view of moral wrongdoing, the moral-wrong principle can be as condemning of paternalistically motivated legislation as can the harm principle.

Of course, if the Millian liberal has independent grounds for condemning paternalistically motivated legislation—grounds independent of the absence of a motivation of preventing harm to others—there is nothing preventing the legal moralist from adopting those independent grounds too. If paternalism independently insults liberty (because it insults one of the main values behind liberty, Millian autonomy), as I argue above is plausible, then a retributivist can condemn paternalism on this ground too, as well as can any Millian.

With regard to the prevention of offense as a reason for criminal legislation, how the moral-wrong principle applies depends on what one thinks of the moral status of causing offense to others. Surely sometimes intentionally causing serious psychic distress to others for no good reason is morally wrong, in which case it may be prohibited by a legal-moralist legislator. As noted above, latter-day Millians such as Joel Feinberg reach this same result by assuming that such serious psychic distress can be a greater harm than the punishment needed to prevent it.Footnote ¹⁸⁰ Here, too, the two principles need not diverge in their justification of (some) punishment/prevention-of-offense–motivated criminal legislation.

With regard to Mill's third corollary—prohibiting behavior because it is immoral—the two principles seem to diverge dramatically. What the moral-wrong principle requires—prohibiting all and only morally wrongful behavior because it is morally wrong—the harm principle seemingly condemns. Yet, as I argue above, all is not as it seems here. Above I speculate that Mill's animus against morals-motivated legislation seems largely to be directed at legislative reliance on erroneous if conventionally accepted moral shibboleths such as that abhorring polygamy.Footnote ¹⁸¹ Indeed, I argue that for Mill, that is all his third corollary can be directed against. As a dedicated utilitarian, Mill had to see his harm principle as an implication of his utilitarianism generally. Utilitarianism generally was his single, ultimate principle of morals. He must have thought that the touchstone of his harm principle was also the touchstone of morality. Behavior harmful to others was not only the only behavior prohibitable by the criminal law; it was also the only behavior that was morally wrong, according to his own view of the one true (utilitarian) ethics. Because of this, behaviors not harmful to others could not really be morally wrong—they could be thought to be morally wrong only by irrational people who did not see the truth of utilitarian ethics.

When Mill thus rails against behavior being criminalized because it is morally wrong, his real target is against criminalization of behavior that is incorrectly believed to be morally wrong by many or most people. His animus, in other words, is directed at moral error being crammed down people's throats for no better reason than that the majority subscribes to such error.

If erroneous moral belief is Mill's real target here, the retributivist/legal moralist can heartily join in its condemnation. Retributivist justice is achieved only by the punishment of morally wrongful behavior. If behavior is not really morally wrong but is only thought to be wrong by some sizable percentage of the population, only the appearance of retributive justice is achieved by punishment of such behavior. Retributivism, as I have long urged,Footnote ¹⁸² enforces real morality, not conventional moral beliefs. So Mill and the legal moralist can happily agree both in their rejection of conventionalist morality as the basis for criminalization and in their acceptance of real morality as that basis, while they significantly disagree as to the basic shape and the content of real morality.

The most widely held theory of punishment is neither purely utilitarian nor purely retributive in nature. Rather, it is a mixed theory that requires that both some net social gain and some retributive justice be produced by punishment in order for that punishment to be justified.Footnote ¹⁸³ For completeness, consider briefly what theory of legislation such a mixed theory of punishment implies. It implies what I am calling the “wrongful harm” principle. On this principle, criminal legislation is rightly motivated only insofar as it aims both to prevent future harms to others and to achieve punishment of the deserving. Actions may thus be prohibited only if they are both acts causing harm to others and acts that are morally wrong to do—“wrongful harms,” for short.

The wrongful-harm principle has many contemporary adherentsFootnote ¹⁸⁴ —no surprise, given the popularity of the mixed theory of punishment on which it depends. Notice that the principle shares with the moral-wrong principle the limiting of criminal prohibitions to those acts that are morally wrong. Thus acts harmful to others but not morally wrong to do are insulated from criminal prohibition by both principles, in contrast to the harm principle. Similarly, the wrongful-harm principle can also condemn paternalistically motivated legislation, on the Millian ground that there is no harm to others sought to be presented, on the legal-moralist ground that (since there are no duties to self) there are no moral wrongs to be punished, or on the ground that paternalistically motivated legislation independently insults Millian autonomy and thus liberty. Similarly, the wrongful-harm principle can follow the Millian and the legal moralist in (wholly or partly) condemning offense-motivated legislation and legislation motivated by erroneous but popular moral beliefs.

My own view on the content of the first, or “derived,” right to liberty is that of the retributivist/legal moralist. It is generally the right of each citizen to be free of improperly motivated criminal legislation, and what is more specifically prohibited is the use of the coercive apparatus of law when that use is motivated by anything except prevention and punishment of moral wrongdoing. Legislation aimed at social betterment in any other way—such as that aimed at preventing nonwrongful offense, nonwrongful self-harmings, or the satisfaction of popular but erroneous moral beliefs and preferences—is forbidden by such a principle.

Such a view of course requires a great deal of fleshing out to be complete. Two items in particular warrant detailed attention. The first concerns the nature of moral wrongdoing. Very generally, moral wrongdoing involves breach of moral obligations. It is not constituted by nonadherence to prudential reasonsFootnote ¹⁸⁵ nor is it concerned with aretaic failures of virtue.Footnote ¹⁸⁶ The moral obligations breach of which constitutes wrongdoing include both those that are most stringent (agent-relative, categorical, deontological) and those that are as a class less stringent (consequentialist duties, such as the duty of rescue).Footnote ¹⁸⁷

Even so limited, the criterion of moral wrongdoing has been criticized by some political theorists as being both too lenient and too restrictive in what it allows in the way of legislation. The too-lenient point is that some morally wrongful actions—the “nongrievance wrongs”Footnote ¹⁸⁸ or the “private wrongs”Footnote ¹⁸⁹ —may seem ineligible for criminalization because they are said to be “not the state's business.” The too-restrictive point is that many malum prohibitum crimes seem needed to solve coordination problems, prisoner's dilemmas, and other collective-action problems, yet they do not criminalize preexisting moral wrongs. I seek to defuse both of these criticisms elsewhere,Footnote ¹⁹⁰ and do not here repeat the exercise (necessary as it is to a complete working-out of the content of the derived right to liberty).

The second item crucial to any complete working-out of the content of the derived right to liberty concerns paternalism. Political philosophy since Mill has been rich with distinctions both as to what is and what is not paternalistic state motivation and as to what is and what is not permissible paternalistic state motivation. As I read that literature, seven such distinctions stand out.

1. (1) In the first place, there is the permissible, “soft” paternalism consisting of using state coercion to restrain people long enough to be sure that they are making the kind of choice that liberty rightly protects.Footnote ¹⁹¹ In the kind of examples made familiar by Mill,Footnote ¹⁹² if someone does not have relevant information (such as damage to a bridge over which he chooses to walk), or if he has mistaken beliefs (such as that he will float upward rather than fall to the earth if he exits a window), then interference for the actor's own good is only “softly” paternalistic and is permissible.Footnote ¹⁹³

2. (2) Likewise, when the actor's capacities to make rational decisions are impaired in ways more permanent than merely having a lack of information, as in the conditions of youth, retardation, and mental illness, then such disabled choices, too, are not fully worthy of protection, a point Mill himself also concedes in finding paternalistically motivated state coercion permissible for these traditional classes.Footnote ¹⁹⁴

3. (3) Relatedly, there are “weakly” paternalistic motivations, also often thought to be permissible.Footnote ¹⁹⁵ These are cases where the state does not replace a person's desires with ends the state finds better for them; rather, it is said, the state only compels use of means to satisfy the actor's own desires better than those means chosen by the actor herself. So if the state requires seat belts in automobiles for adult passengers, for example, not because the state rejudges the desirability of safety versus freedom of movement but rather because the state believes the subject herself most values safety, then the state is said to be only “weakly” paternalistic. For the state in such a situation is only furthering the subject's own desires with its coercive power.

4. (4) Consented-to paternalism is often thought to be a permissible kind of paternalistic motivation (or even to be nonpaternalistic). In the classic example of a Ulysses contract, the actor at t₁ “binds himself to the mast” by consenting to the use of state coercion later on if or when it proves necessary to get him to accept what he accepted earlier. When the state at t₂ does use its coercion for the reason that it is best for the actor if he performs the action to which he earlier consented, it is said not to act paternalistically, because it acts in the subject's own name.

5. (5) If Ulysses contracts are validly enforced by the state, why, then, cannot one sell himself into slavery? From Mill onward, choices to deprive oneself permanently of the capacities that make liberty valuable may justifiably be prohibited on paternalistic grounds, although some of us find the point contestable.Footnote ¹⁹⁶

6. (6) Much has been made in recent literature this past decade of the legitimacy of paternalistically motivated state influencing (rather than coercing) of human choice (such influencing going under the name of “nudging”).Footnote ¹⁹⁷ Such nudging approaches force one to draw lines between those permissible state interventions that aid individuals to make better decisions in the way that better information does, on the one hand, and those that so manipulate the decisional process (by using techniques of subliminal perception and the like) as to undermine the authorship of the subject so “nudged,” on the other.

7. (7) Last, one must distinguish permissible legal moralism from impermissible moralistic paternalismFootnote ¹⁹⁸ (or “moralistic legal paternalism”).Footnote ¹⁹⁹ Moralistic paternalism exists when coercive legislation is aimed at improving the moral welfare of citizens even when those citizens do not agree that it is an improvement. The idea is paternalistic because the state presumes to know what is good for someone's moral health, contrary to their own best judgment of the matter; the view is moralistic because it is not concerned with any subjective notion of welfare such as pleasure, happiness, or preference satisfaction but rather with the moral well-being of the actor. By contrast, the legal moralist aims to make the world better by preventing and requiting moral wrongs. The difference lies between making a person morally better and making the world (or “things” in general) morally better. These, in fact, are not the same motivations: the state of the world being morally better is not to be identified with the state of some person(s) being morally better persons—(or things being morally better for them).Footnote ²⁰⁰

With these (admittedly only lightly sketched) clarifications, the content of the derived right to liberty can be made complete. I thus turn to the basic right to liberty and its content. As a reminder, the second, or “basic,” right to liberty is not concerned with the subjectively motivating reasons for legislative prohibition of certain activities. Rather, it is concerned with the freedom to do those activities without state interference, no matter how that interference might be motivated. Still, there should be some relationship between the content of the two rights to liberty, as we see in what follows.

F. The Second (or “Basic”) Right to Do Certain Actions Free of State Coercion, However that Coercion Is Motivated

It will be helpful to divide the sphere of free action protected by the basic right to liberty into three parts, at least initially. One motive for doing so is to separate actions that should be free because of the absence of any good reason to coerce them (the first part) from actions that should be free because of the presence of reasons not to coerce them (the second and third parts).

G. The Limited Force of the Two Rights to Liberty: Threshold Deontology and Consequential Overrides

Like all moral rights, the two rights to liberty are not plausibly construed to be “absolute.” Although the Kantian injunction that “it is better that humanity perish from the earth than that a single right be violated”Footnote ²¹⁶ is the kind of stirring hyperbole that gets people to the barricades, it is surely false. Indeed, if the Kantian injunction were the only version of deontology, I would be a thoroughgoing consequentialist in my substantive ethics.

There are two ways in which the natural rights to liberty are not absolutes in this Kantian sense. One is within the theory of the right; the other is within the theory of the good.Footnote ²¹⁷ Starting with the first, a more plausible deontology is what I and others call “threshold deontology.”Footnote ²¹⁸ According to threshold deontology, rights (and their correlative duties on others) are immune to violation only up to a threshold of awful consequences; over such a threshold, “all bets are off,” in the sense that averting true moral horror justifies violating the rights of one or a few. If the only way to learn of the location of a terrorist nuke somewhere in New York is to torture the innocent child/spouse/parent of the terrorist who planted it there, such rights-violating torture is not only permissible but, in my view, even obligatory.Footnote ²¹⁹

Many have sensed the lumpy, ad hoc feel to threshold deontology and have therefore questioned its rationality.Footnote ²²⁰ My defense of it is not only Churchillian (it may be ugly but it is preferable to the alternatives) but is also (and more positively) in terms of the idea that obligations have different degrees of stringency such that the obligation to avert severely awful consequences can be sufficiently stringent as to override the less stringent even though categorical obligations of deontological morality. Such a defense, however, is for another occasion.Footnote ²²¹ Pertinent here is the fact that this route to defending the rationality of threshold deontology commits one to what I call “sliding scale threshold deontology.”Footnote ²²² This is the view that the location of the threshold (over which consequentialist justification holds sway) varies depending on the degree of stringency of the categorical obligation being violated. If, for example, the stringency of the obligations correlative to the two rights to liberty varies with the kinds of choices at issue in a given case—the liberty to make “fundamental” choices demanding greater stringency of obligation than the liberty to make “nonfundamental” choices, for example—then the threshold over which violation of such liberty rights can be violated will also vary.

With respect to the degree of rights protection within the theory of the good, consider Ronald Dworkin's fetching metaphor that rights “trump” utility.Footnote ²²³ Although it is not clear that Dworkin so understood the point, the best interpretation of his metaphor is not to be referring to the categorical nature of deontological moral norms. Rather, rights trump utility in the different sense that between two states of affairs, a right not being violated and net human welfare being enhanced, the former is significantly more valuable. A plausible consequentialism can think that both rights protection and welfare enhancement are intrinsically good states of affairs and think that the first is much more valuable than the second.

In my own view of deontic ethics, one reaches such consequentialist trade-offs once one is outside the agent-relating obligations and prerogatives of deontological morality.Footnote ²²⁴ Government officials are sometimes in just this position. There are no categorical obligations in the budget room, for example, for everyone in there is outside the scope of deontological obligation. There, governments must trade off enhancement of welfare against protection of various kinds of rights and justice. Surely even Kant, if he were sitting on the Königsberg City Council, would have funded services and facilities for enhancement of welfare, for protection of corrective justice, for prevention of rights-violating crime, along with funding services and facilities to achieve his favored retributive justice.

In making such trade-offs, rights protection (and justice generally) gets a “trumping” position vis-à-vis enhancement of welfare. Yet such trumping is hardly absolute; at some point alleviating human misery outweighs the enhancement of rights protection. When officials are outside the scope of deontological obligation and are considering how best to protect the natural rights to liberty, they thus are not obligated to eschew all trade-offs of utility as against rights.

A. Constitutionalizing the Derived Right to Liberty

B. Constitutionalizing the Basic Right to Liberty

C. Making Sense of Liberty's Being Overridden by a “Compelling State Interest” in Light of Threshold Deontology and/or the Limited Trumping by Rights over Utility

As set forth above, no natural right can plausibly be viewed as absolute in its force. Even when the correlative duty-holder is within the scope of a deontological (as opposed to a consequentialist) duty, that duty is justifiably overcome if the avoidance of truly horrendous consequences is in the offing. This is true of even the most stringent rights we have, such as the right not to be tortured or murdered; it is a fortiori true of the two natural rights to liberty that we all possess.

It is also true that state actors are sometimes outside the scope of deontological obligation, and thus the duties that they hold to respect the rights of their citizens take on a holistic, maximizing shape: they are to minimize rights violations and maximize the rights-correlative duties being kept in the policies they adopt in such situations. In such maximizing calculations, rights protection will be one good balanced against other goods, including welfare; in such balance, rights will “trump” utility, but only up to a point.

One might think that these two before-discussed points of moral and political theory would imply (and thus justify) the Supreme Court's current doctrine of overriding even fundamental liberty rights when there is a “compelling state interest.” The facts are otherwise, however, because the fit between doctrine and moral theory is not a clean one. Consider in this regard the abortion cases, where the Court has used its compelling-state-interest doctrine most extensively.

In Roe, Justice Blackmun for the Court identified three interests of the state in limiting a woman's liberty to choose to abort: “a state may properly assert important interests in safeguarding health [of the mother], in maintaining medical standards, and in promoting potential [i.e., the fetus's] life.”Footnote ³³⁸ These state interests ripened, Blackmun further opined, as the pregnancy wore on, the first becoming a compelling state interest at the end of the first trimester, and the third achieving that status at the end of the second trimester. Sandra Day O'Conner's opinion for the Court in Casey modified this analysis by focusing on the last of Blackmun's three state interests, finding that interest in potential life to be a “profound” one throughout pregnancy and one compelling enough to proscribe abortion once the fetus has achieved viability (which is roughly towards the end of the second trimester).

Whatever one thinks of the Court's characterization of the state's interests here, it is plain that they do not rise to the level of moral catastrophe measured by the thresholds of a threshold deontology. Deontology's kind of “compelling state interest” would perhaps be reached by items such as the population explosion—one might well think that global human society has reached the point in its population growth forecast by Malthusian predictions centuries ago and that therefore the liberty of persons to choose the number of their progeny is overridden by that looming catastrophe.Footnote ³³⁹ However, even if this is so, the woman's liberty to choose to abort would not be affected; rather, it would be her liberty to choose not to abort that such compelling state interest would override.

In any case, perhaps the Court's compelling-state-interest analysis can be likened to “rights trumping utility,” even if not to threshold deontology. A logically prior question to how that balance comes out is the question of how a court gets to balance the importance of rights against welfare here; for if women have a natural right to choose whether to carry a fetus to term, the state's correlative duty is seemingly categorical, precluding any consequentialist balance (short of the threshold of disastrous consequences just discussed).

The answer to this preliminary question has to be because the relevant state actors make their decisions outside the scope of categorical obligations. Consider legislators first. Except for the passage of private bills, legislatures are never in the position of overriding the liberty of some particular woman to choose whether to have an abortion. In passing legislation that is by its nature general, legislatures are always dealing with classes of persons, not individual persons. Because of this, legislatures are outside the scope of deontological (or categorical) obligations and may thus justify their decisions morally by the goodness of their consequences.Footnote ³⁴⁰

Appellate courts sometimes share this moral position of legislatures in being outside the scope of deontological obligations. The abortion cases give one instance of this. Because of a procedural anomaly about those cases—the shortness of pregnancy vis-à-vis the length of time it takes to prosecute an appeal—no particular woman's decision is ever being restrained by an appellate court order, because technically the named petitioners’ complaints are moot. In such situations as in Roe itself, the Court, too, is free to balance the consequences.Footnote ³⁴¹

Legislatures (and occasionally appellate courts) are thus free to balance consequences, and when they do so, they balance the goodness of not violating natural rights (such as the woman's right to choose to abort) against the utility gained by overriding that right. And when they make such balances, “rights trump utilities” in the sense that not violating the woman's right to make the decisions herself must be given great although not conclusive weight.

This moral balancing is closer to the Court's “compelling state interest” analysis. How close depends on one's assessment of the weight properly to be attached to the interest the Court finds “compelling” after viability, namely, the protection of fetal life and potential personhood. As the Court tells us in Roe, that interest is not to be seen as the protection of a right in the fetus to lifeFootnote ³⁴² —for then we would be balancing right against right.Footnote ³⁴³ Rather, the fetus constitutes only an interest the law should protect.

The weightiness of this interest is conveniently explored via Judith Thomson's famous violinist example. You go to the symphony as a member of the audience, fall asleep during the performance, and awake to find that the lead violinist has been hooked up to you with some life support system that only you can provide and that will kill him if it is withdrawn before nine months has passed. May you unhook the device anyway, even knowing that the violinist will die? Should the state allow you to do so? Thomson intends that your intuitive answers should be in the affirmative, in which case even the life of a full-fledged person is an interest less weighty than your right to determine whether your body will be used as a life-saving device. If, like me, you agree with Thomson here, that makes it difficult to think that fetal life (or potential personhood) can be morally weighty enough to become a “compelling” justification for overriding the woman's liberty and that makes it difficult to think that the Supreme Court's compelling-state-interest doctrine maps very closely onto morality's trumping of welfare by rights.

Yet abortion may be a special kind of example with which to gage the fit between legal doctrine and morality here, in possibly two ways. One is the opportunity for women to have exercised their liberty of choice with respect to child-rearing at a time prior to the time of abortion, except in cases of rape. This opportunity, of course, is when the decision to have sex was made, particularly if that decision involved unprotected sex. One might generally think that when we are at liberty to decide something fundamentally important to our lives, but we have two times when we can exercise that liberty—an earlier time when no other interest is involved and a later time when some other interest will be adversely affected—our liberty is of lesser weight at the second time than it would have been absent that first opportunity. The discount given the liberty to speak when one had other times, places, or manners of speaking that were less costly to the public is of this nature. (I do not myself think this about abortion because I do not think knowingly taking some risk of a consequence is the same as affirmatively choosing that consequence, but many others seem to find this feature of abortion to lessen the weightiness of the woman's right to abort.)

The second possible difference is that the right to decide to have an abortion is perhaps one of those fundamental rights that is “less fundamental” than other liberties on the earlier list of fundamental liberties. And if this is why the right to abort can be overridden by the interest in fetal life, one need only show that how compelling an interest need be varies with how fundamental is the liberty being protected. This would be to see, once again, the sliding scale on which our liberty judgments are made.