One of the most prominent theses which appear in recent writings on legal interpretation is the so-called anti-theoretical claim. This claim usually appears in two variants: a stronger and a weaker version. The strong form is advocated by Richard Posner, who holds that legal judgments must be grounded on a consequentialist reasoning that requires no value theory whatsoever. Legal decisions must be assessed on the basis of the political consequences that they ensue, rather than on any abstract theory about the legitimacy, morality or authority of the law (Posner 1990, 1998, 2003). The weak version, in turn, is held by Cass Sunstein and Adrian Vermeule. These authors believe that Posner is wrong to suppose that judges may reach any decision without a theoretical stance. If one is to evaluate a legal argument on the basis of its consequences, it is undeniable that one needs a value theory to determine which consequences are just, good or desirable, for this is the only way to ground any sort of consequentialist judgment. Sunstein and Vermeule uphold, however, that people with different background value theories may reach the same conclusions on the basis of an “incompletely theorized agreement” (Sunstein 2001). When this is the case, then any theoretical disagreement that participants may have should be bracketed and considered irrelevant for reaching the decision at hand (Vermeule 2006; Sunstein and Vermeule 2003). Although normative theories are necessary to support a legal judgment, they do not by themselves allow one to choose among different interpretations, and jurists should not spend much time and energy with them.

By the same token, I argue that Sunstein and Vermeule’s advocacy of incompletely theorized agreements is in no better position than Posner’s decisionism. From the practical point of view, there is little difference between adopting Posner’s strong version of the antitheoretical claim or bracketing all theoretical disagreements on the basis of an incompletely theorized agreement. Yet, though both versions of the anti-theoretical claim are untenable, there is a powerful argument underlying Vermeule’s assumptions about the significance of empirical considerations in debates concerning the choice of an interpretive methodology. Though I am persuaded by Dworkin’s objections to the Anti-Theoretical Claim, I believe that the advocates of its weaker version have successfully shown that no sensible theory of legal interpretation can be supported without a foundation provided by empirical evidence. To come to this conclusion, I proceed as follows. In Sect. 7.2, I introduce Posner’s “everyday” pragmatism, which is grounded on the strong version of the anti-theoretical claim. In Sect. 7.3, I expound Sunstein and Vermeule’s criticisms on Posner and their attempt to overcome the difficulties of the strong anti-theoretical claim by raising a “weaker” version of the same claim. Section 7.4, in turn, intends to dismantle the fallacy which underlies the anti-theoretical claim, while Sect. 7.5 acknowledges the strength of Vermeule’s normative argument about the role of empirical research in legal argumentation.

In his Holmes Lectures delivered at Harvard Law School on October 14 and 15, 1997, Richard Posner shocked the world of legal philosophy with two theses concerning the role of moral theory in legal argumentation.

The first thesis is an abstract sceptical claim that concerns moral philosophy in general, and holds that no moral theory can ever provide a solid basis for moral judgments (Posner 1998, p. 1639).

This ambitious theoretical claim is based on the following assumptions. First, morality is always and necessarily local, for “there are no interesting moral universals” and the few principles of social cooperation that might be valid across different cultures are “too abstract to serve as standards for moral judgment” (Posner 1998, pp. 1640–1641). Second, “many of the so-called moral phenomena can be explained without reference to moral categories,” as “most moral principles that claim universality are better understood as mere workaday social norms in fancy dress” (Posner 1998, p. 1641). The domain of moral theory is not composed of self-evident principles or moral truths, but conventional norms whose efficacy must be measured according to their ability to work as “means to a society’s ends” (Posner 1998, p. 1652).¹ Third, moral theory or, as Posner prefers to say, academic moralism, is incapable of improving moral behaviour, since “knowing the moral thing to do does not furnish a motivation for doing it,” as moral norms are “too feeble to override either narrow self-interest or moral intuitions.” The vivid disagreement amongst academic moralists makes it possible for the reader to find a rationalization for any course of conduct, regardless of its moral merits (Posner 1998, p. 1641). Fourth, “exposure to moral philosophy may actually lead people to behave less morally by making them more adept at rationalization.” And fifth, there is no uniform morality and it would be “a disaster” if academic moralists were successful in imposing their own moral view upon the majority (Posner 1998, p. 1642).

The second thesis, in turn, is specific about legal reasoning, and holds that even if moral theory could provide a solid ground for some moral judgments, it should not be used as a basis for legal judgments (Posner 1998, p. 1639).

The target here is limited to legal discourse. Yet in a certain sense the second claim is even more ambitious because it not only holds that moral philosophy is a mere tool for rationalizing one’s idiosyncratic moral views, but also argues that the separation between law and morality entails a sort of firewall that prevents moral considerations from influencing any relevant legal decision.

Here I am arguing that moral theory has nothing for law, but I am not limiting myself to academic moralism. The idea that racial discrimination is immoral owes very little to academic moralists; it owes a lot to non-academic moral entrepreneurs such as Abraham Lincoln and Martin Luther King, Jr. Yet we shall see in considering Brown v. Board of Education that the courts do not rely on these moralists, either, to support decisions in racial cases, and we shall see that there are good prudential reasons for this forbearance. I do not mean that moral entrepreneurs are never cited in judicial decisions, but they are cited as representatives of uncontested moral positions, rather than as authorities for taking one side or another of a moral issue (Posner 1998, p. 1698).

Judges and lawyers, therefore, should narrow down their ambitions and give up making abstract moral judgments to justify their legal decisions. Rather than philosophizing and rationalizing moral principles, judges should adopt a pragmatic approach to adjudication that evaluates a legal decision on the basis of its social consequences, and not of its moral worth.

To distinguish this “practical” approach to adjudication from the philosophical theories of pragmatism, Posner advocates a normative theory of adjudication called “everyday” pragmatism, which could be described thus:

Everyday pragmatism is the mindset denoted by the popular usage of the word ‘pragmatic,’ meaning practical and business-like, ‘no-nonsense,’ disdainful of abstract theory and intellectual pretension, contemptuous of moralizers and utopian dreamers. It long has been and remains the untheorized cultural outlook of most Americans, one rooted in the usages and attitudes of a brash, fast-moving, competitive, forward-looking, commercial, materalistic, philistine society, with its emphasis on working hard and getting ahead (Posner 2003, p. 50).

While adopting this “everyday” pragmatism, one bases one’s judgments “on consequences, rather than on deduction from premises in the manner of a syllogism,” without a commitment to any philosophical tradition on the basis of which these consequences will be evaluated (Posner 2008, p. 40).²

In his institutional theory of legal interpretation, Vermeule claims that no interpretive theory can be defended without careful empirical research about the interpretive capacities of legal institutions and the systemic effects of the allocation of decision-making power between or among institutions. This is what he calls his “minimal point” about interpretive theories (Vermeule 2006, p. 81).

Nevertheless, his institutional theory of interpretation is based on a second and more ambitious claim that in some cases “a second-best assessment of institutional issues might not only be necessary but indeed sufficient to resolve conflicts over interpretive theories,” inasmuch as people with different theoretical premises might agree on a particular interpretive strategy at the operational level (Vermeule 2006, p. 82).³

This argument is premised on the possibility of an “incompletely theorized agreement” in the sense defended by Cass Sunstein. Under this view, people who disagree about abstract moral principles might attempt a “conceptual descent,” i.e., a “descent to a lower level of abstraction” with a view to achieving a consensus about “concrete outcomes” (Sunstein 2001, p. 50–51). According to Sunstein,

The agreement on these points, more particular than their supporting grounds, is incompletely theorized in the sense that the relevant participants are clear on the practice or the result without agreeing on the most general theory that accounts for it. Often people can agree on a rationale offering low-level or midlevel principles. They may agree that a rule – protecting political dissenters, allowing workers to practice their religion – makes sense without entirely agreeing on the foundations for their belief (Sunstein 2001, p. 51).

The possibility of incompletely theorized agreements over the right interpretive theory for a given institution, therefore, allows meta-interpreters to put aside the theories on which they base their interpretive decisions at the operational level.

By relying on incompletely theorized agreements, judges should embrace what Sunstein has described as “decisional minimalism,” which advises judges to say “no more than necessary to justify an outcome,” and leave “as much as possible undecided” (Sunstein 1999, pp. 3–4). Decisional minimalism would have at least two important advantages which reduce the costs of judicial decision-making and contribute to the democratic justification of a legal ruling: first, it “reduces the burdens of judicial decision” and, more fundamentally, “is likely to make judicial errors less frequent and (above all) less damaging” (Sunstein 1999, p. 4).

This implies, according to Vermeule, that “institutional analysis might even enable interpreters to choose particular doctrines before, or in place of, choosing a value theory that specifies what counts as a good or bad consequence of interpretive practices” (Vermeule 2006, pp. 82–83). To give an example, Vermeule thinks that, “if, on certain empirical findings, it turned out that legislative history should be excluded on any high-level theory specifying what counts as a good or bad interpretation, then as far as the interpretive question goes, there would be no need to choose a fundamental theory” (Vermeule 2006, p. 83).

Vermeule’s account is, thus, admittedly anti-theoretical as he believes that most of the theoretical disagreements in meta-interpretive debates may be “bracketed as irrelevant to the operational problems and thus dispensed with altogether” (Vermeule 2006, p. 63).⁴

An adequate empirical analysis of the performance of a formalist (or any other) interpretive method for our institutions should, as Vermeule argued in an earlier essay co-authored by Sunstein, provide a reliable answer to at least the following three questions, which deal mostly with empirical issues: (1) The first question, as Sunstein and Vermeule argue, is “whether and when formalist decisions that produce clear mistakes will be corrected by the legislature and whether making the corrections will have low or high costs” (Sunstein and Vermeule 2003, p. 917). (2) The second question, in turn, is “whether a nonformalist judiciary will greatly increase the costs of decision for courts, litigants, and those seeking legal advice. A large issue here involves planning; if nonformalist approaches make planning difficult or impossible, there is a real problem” (Sunstein and Vermeule 2003, p. 918). (3) Finally, the third question is “whether a formalist or a nonformalist judiciary, in one or another domain, will produce mistakes and injustices” (Sunstein and Vermeule 2003, pp. 918–9).

These questions, for Vermeule, refer mostly to the “institutional capacities” and “systemic effects” of interpretive theories, which according to his account are the most important variables that should be balanced in order to support a theory of constitutional interpretation.

If this meta-interpretive strategy is consistently employed, then Vermeule thinks that interpreters will not struggle to conclude that judges should adopt a formalist strategy of legal interpretation, following the “clear and specific meaning of legal texts, where those texts have clear and specific meanings,” and deferring “to the interpretations offered by legislatures and agencies, where legal texts lack clear and specific meanings” (Vermeule 2006, p. 1). When interpreting the constitution, judges should “avoid high-level claims about constitutionalism, democracy, or the nature of law” and “enforce clear and specific constitutional texts according to the surface meaning,” because this procedure “will produce the best ground-level consequences for legal institutions” (Vermeule 2006, p. 33).

Although Vermeule offers other institutional considerations in support of this formalist method of constitutional interpretation, my impression is that the main argument for this view is the (empirically verifiable) “epistemic superiority” of legislatures over courts (Vermeule 2009