Leiter on Originalism: A few weeks ago, I spoke on "The New Originalism" at a colloquium sponsored by the Philosophy Department at the University of Texas. Brian Leiter was there and he now has posted Justifying Originalism in which he thoughtfully and respectfully comments on my talk. After a nice summary of my oral presentation, he offers this critique:

[Barnett] thinks that the writtenness of the Constitution is a fact supporting the New Originalism, becaue the original public meaning as written is something that binds those who make laws that, in turn, purport to bind us as citizens. At this point, I confess I don't follow the arguments. If "constitutional legitimacy" in Barnett's terms (or "authority" in Razian terms) is the benchmark for the method of interpretation we should apply to a text, then I can not see how the writtenness of the text figures as an independent consideration.

Indeed, it seems to me a stronger conclusion is suggested: namely, that once we acknowledge (as Barnett rightly does) that a theory of constitutional interpretation must answer to a theory of constitutional legitimacy, then the constitution (as a written document, or as a document about which framers had some original intention, or as a document which had an original "public" meaning, and so on) drops almost entirely out of the picture: the theory of constitutional legitimacy tells judges how they should decide cases, and the written constitution is, at best, a proxy for what is constitutionally legitimate or is relevant because of its effect on the reasonable expectations of citizens (the latter being a factor bearing on constitutiona legitimacy). In other words, Barnett's theory of constitutional interpretation, because it (unlike most theories--Ackerman's, Amar's, etc.) has the correct moral structure is not really a defense of originalism, but a defense of whatever method of interpretation produces "legitimate" (i.e., morally authoritative) outcomes.

In my book, Restoring the Lost Constitution: The Presumption of Liberty, I present a normative argument for originalism. Here is a clearer breakdown of my argument for originalism that I believe shows why constitutional interpretation should not be reduced to whatever results accord with whatever background conception of justice accounts for a constitution's legitimacy. Note that each step in the following description requires further elaboration and defense. My goal here is simply to clarify the structure of the argument so as to advance the discussion.

(1) In Restoring, I begin by denying that the legitimacy of a constitution of the sort that governs a territory like the United States could ever be grounded on the "consent of the governed" because such consent is and must always be a fiction. If such a constitution is ever legitimate at all (and perhaps none can be), it must be because it is warranted in imposing its commands on nonconsenting members of the public.

(2) I then distinguish between the justice of a law and the legitimacy of a law-making process. A law is just and entails a duty of obedience if it does not violate the rights of the persons on whom it is imposed (with or without their consent) and is necessary to respect the rights of others. In contrast, a law-making process is legitimate if it offers procedural assurances that laws imposed on nonconsenting persons are likely enough to be just as to impart on these laws the benefit of the doubt. In other words, the fact that laws have been produced by these rights-protective procedures provides a prima facie reason why they ought to be obeyed. A just law could be illegitimate because not produced by such a process and a legitimate law could be unjust if the process has failed in a particular case to produce a just law.

(3) A written constitution, like such constitutional features as separation of powers, federalism, bicameralism, regular elections, varying terms of office, etc., should be considered just another structural feature of our constitutional order that defines a law-making process. All these features are means to the end of achieving (among other ends) a legitimate constitutional order in the absence of real consent. They either conduce to this end or they don't. Of course, it is easy to imagine constitutional orders without this structural feature--just look around the world--just as it is easy to imagine law-making systems without the other structural features of the U.S. Constitution. By the same token it is easy to imagine a system in which its written constitution is ignored or "updated" by the courts--just look at the U.S--but whatever enhancement of legitimacy is provided by writtenness, as described below, will be absent.

(4) The device of a written constitution was developed as means of imposing law on those who impose law on the people. In particular, it was a means both to authorize the use of lawful power and at the same time limit the scope of those powers to imposing laws that are both necessary and proper. It defines the process of lawmaking more particularly than an "unwritten constitution" does, and provides express limits on government power that are lacking in legal systems in which the governing parliament is considered sovereign.

(5) This normative defense of a written constitution is based on the claim that the people (either collectively or as individuals) retain their sovereignty or rights. They are the principals and the government as a whole, including Congress, are mere agents or "servants" of the people. In a principal-agent relationship, the principal retains some or all of her rights while delegating certain powers to the agent, who must exercise those powers on (a) on the principal's behalf and (b) subject to the principal's control. It is no coincidence that the Ninth Amendment refers to the "rights . . . retained by the people," while the Tenth Amendment refers to the "powers . . . delegated to the United States" by the Constitution.

(6) A written constitution defines the nature of this agency relationship and, by so doing in writing, helps police it. It is easier to see where the agent exceeds its proper powers when these powers are defined in writing. In Lon Fuller's terms, writings serve the evidentiary, cautionary and channeling functions of formality.

(7) The agents — the Congress, the President, the Judiciary — whose powers are defined by the written constitution, should not be able to change or expand the scope of their own powers, even in combination with each other. As Isaac Penington wrote in 1651: "They who are to govern by Laws should have little or no hand in making the Laws they are to govern by."

(8) So that they do not define or alter the scope of their own powers, some other body besides these government agents must approve any alteration or amendment, whether that other body be a "convention" or state governments. This principle is reflected in the procedures described in Article V. We could also imagine a referendum scheme to ratify amendments of the sort urged by Akhil Amar but the written U.S. Constitution does not provide for this process.

(9) Therefore, in the absence of a proper amendment, the meaning of the written Constitution should remain the same until it is properly changed. This is another way of describing original public meaning originalism.

To reiterate, each step in this argument merits additional explanation and defense, but this is a blog post that is overly long as it is. My objective here is simply to clarify the structure of the normative argument being made on behalf of originalism.

Notice that, while "justice" is the ultimate normative justification for originalism, the intermediate steps are crucial. If a written constitution is valuable for the reasons identified--to define and police the principal-agent relationship--then one cannot simply dispense with it in pursuit of greater justice. More precisely, agent-judges cannot on their own authorize agent-legislatures to exceed their proper powers as defined by the written Constitution in pursuit of greater "justice" than that document provides.

This implication is analogous to defenses of "rule" over "act" utilitarianism, or what Larry Alexander has called "indirect utilitarianism." Or if you think that rule utilitarianism always collapses into act utilitarianism, then consider another analogy: Placing the burden of proof on the prosecution as a means to the end of justice, especially the end of preventing the wrongful conviction and punishment of the innocent. Acknowledging that the presumption of innocence is ultimately justified as means the achievement of justice does not justify dispensing with the presumption in particular cases to convict someone we "know" is guilty. The presumption of innocence is a systemic way of dealing with a problem of knowledge--the problem of enforcement error--as well as a problem of interest--the problem of enforcement abuse. (I discuss both in my book, The Structure of Liberty: Justice and the Rule of Law.) By the same token, judges should not allow the legislature to exceed its powers by changing the meaning of the Constitution because doing so would better serve what they personally believe is justice. Once again, adhering to a written constitution is a way to address the problem of knowledge (judicial mistakes about what justice may be) as well as the problem of interest (judges who knowingly disregard justice). This normative defense of originalism is a special case of the normative defense of the rule of law I offered in The Structure of Liberty. Although the rule of law is ultimately dependent on a conception of justice, it functions independently and is normatively warranted as a means of solving pervasive social problems of knowledge, interest, and power.

Of course, one could contest the desirability of limiting the power to convict the guilty by means of a presumption of innocence or of limiting the powers of government by means of a written constitution. The latter is often the underlying objection being made by critics of originalism. But the nature of the objection is obscured by claims that the written constitution is merely being "interpreted" rather than frankly admitting that the Constitution is being amended for the better by judges. Perhaps this use of language is an homage that constitutional vice plays to constitutional virtue.

Two final points. First, my normative defense of originalism is not the only one out there. Keith Whittington defends originalism on popular sovereignty grounds that I reject. Second, Larry Solum makes the important distinction between descriptive originalism and normative originalism. According to this distinction, as a purely descriptive matter, original public meaning is simply what a text does mean according to the Gricean theory of language, which Larry thinks is the best available theory of meaning in this context. This may be the sort of assertion of meaning to which Brian reacts negatively, but needs to be hashed out as matter of the philosophy of language. Whether and to what extent judges or others ought to adhere to this meaning is what Larry calls "normative originalism" and here Brian might get back on the train. Larry has yet to write about this distinction, but I wanted to identify it as relevant perspective on this issue. I plan to address this matter in future work and am open to adopting it, but that is a separable argument than what I am presenting here today.

Update:Larry Solum has now posted on the distinction between descriptive or semantic originalism and normative originalism in Semantic and Normative Originalism: Comments on Brian Leiter's "Justifying Originalism"

Solum on Leiter (and Me) on Originalism: Larry Solum has an important new post on the New Originalism responding to Brian Leiter and my recent posts on the subject. Semantic and Normative Originalism: Comments on Brian Leiter's "Justifying Originalism" It is far too long and complex to summarize here, and I cannot even find a single excerpt that captures its entire analysis. Still, here is an excerpt that might get you to read it as anyone interested in originalism should do:

[snip]

The Case for Semantic Originalism

[snip] The meaning of an utterance is simply a fact about the natural world. For this reason, semantic theories might be called "descriptive" or "positive." Simplifying greatly, we might say that there are two leading theories of legal meaning in general, which we might call "textualism" and "intentionalism." These theories of natural meaning correspond, more or less, to the distinction made by philosopher Paul Grice between "sentence meaning" and "speakers meaning." In the context of constitutional theory, textualism is the view that the natural meaning of the constitutional text is its conventional semantic meaning—the meaning that the words and phrases would have conveyed to the public at the time the relevant portion of constitutional text was drafted, proposed, and adopted. Borrowing from Grice, we can call this sort of meaning "clause meaning." Similarly, the constitutional version of intentionalism is the view that the natural meaning of the constitutional text is meaning intended by the authors (framers and/or ratifiers) of the relevant constitutional provision. Following Grice we might refer to this kind of meaning as "framers meaning."

From what I have said so far, one might infer that we can choose between framers meaning and clause meaning, and that this choice would have to be justified on normative grounds. But this is not, in fact, the case. That is because the meaning of legal texts in general (and the Constitution in particular) cannot be understood as speakers meaning (or framers meaning). For speakers meaning to be possible, the audience of an utterance must have knowledge of the speaker's intentions and the speaker must know that the audience has this knowledge—call this the "reflexivity condition." But in the case of legal texts like judicial opinions, statues, and constitutions, the reflexivity condition is not met. This is most obviously true in the case of the United States Constitution, where the framers believed the evidence of their intentions (the proceedings of the Philadelphia Convention) would be kept secret. Recognizing that there are conditions under which speakers meaning is impossible, Grice recognized the existence of sentence meaning—the meaning that would be conveyed by the conventional semantic meaning of the expressions that constitute an utterance. In the context, of the Constitution, this is the "clause meaning" or the conventional semantic meaning of the constitutional text. In other words, "clause meaning" is the "original public meaning" of the New Originalism.

Of course, there might be other semantic theories of constitutional meaning. For example, someone might argue that the semantic meaning of the clauses is the conventional semantic meaning they have at the time of application—call this "contemporary public meaning." For example, the phrase "domestic violence" might have referred to "rebellion or insurrection" at the time the Constitution of 1789 was adopted, but it could refer to "spousal, child, and elder abuse" today. Or it might be argued that the meaning of the Constitution is whatever meaning a judge wishes to assign to it—on this theory, the Constitution is an empty vessel into which judges can pour any semantic content they wish. Or one might be a semantic skeptic, and deny that there is any semantic meaning at all. Although I won't make the arguments on this occasion, even a moment's reflection reveals that there are serious problems with contemporary public meaning or judge's meaning as semantic theories. Semantic skepticism faces even more serious problems, including the obvious one that it is self-effacing because it implies that it is not even meaningfully assertable.

How might one argue against semantic originalism? There are at least three strategies. First, one could argue for some particular alternative semantic theory. Second, one could argue for semantic pluralism—the view that the constitutional text has a "meta-ambiguity" between the meanings assigned to it by alternative semantic theories. Third, one might argue that the constitutional text lacks semantic content at all—that despite the superficial resemblance of the text to a meaningful utterance, the Constitution simply does not mean anything at all. Perhaps there are other strategies. But any argument that clashes with semantic originalism must assume a stance either about or from within a semantic theory. What won't work is to make a normative argument against semantic originalism. Normative arguments address a different question, and the attempt to argue against semantic originalism on the basis of a normative claim involves a category mistake.

Reframing the Normative Debate

The truth of semantic originalism, by itself, has no normative implications. Even if the semantic content of the constitution is the original public meaning, it could be the case that constitutional practice should deviate from that meaning. For example, one might argue that precedent should trump the original public meaning. Or one might argue that the difficulty of constitutional amendment so undermines the democratic legitimacy of the Constitution that the political branches are justified in "amending constructions" that give the text new and different semantic content. Or one might believe that the constitution is so undesirable or unjust that judges are justified in ignoring the semantic content of the constitution and deciding cases on the basis of their own beliefs about what the Constitution should have provided.

[snip] In a wide variety of legal contexts, the notion that officials are obligated by the semantic content of legal texts is simply not controversial. Quite the contrary, our usual assumption is that in a reasonably just society, the semantic content of valid legal texts creates obligations for judges, other officials, and citizens. The corresponding obligations of fidelity to the law attach to the semantic content and not the mere syntactic form of the texts.

Again, I should be very clear. I am not claiming that I have produced a normative argument for originalism. The aim of this discussion of the normative case for originalism has been very modest. My goal is to put the normative question in a new frame. Leiter is right to observe that many originalists—but not Balkin, Barnett, or Whittington—seem to beg the normative question. If I might be allowed to restate Leiter's observation in my own terminology. Many originalists seem to believe that by producing arguments for semantic originalism, they have provided a prima facie normative case. By reframing the debate, I think we can see how they made this mistake. These originalists believed that it was obvious that the correct theory of constitutional meaning entailed normative conclusions. I believe that was a mistake. "Obviousness" is in the eyes of the beholder. Once the normative claims of originalism are questioned, an argument must be produced.

But what is good for the goose is good for the gander. Nonoriginalists have their own burdens of persuasion and production. If nonoriginalists want to admit the truth of semantic originalism, but deny its normative force, they should do so explicitly. If they want to contest the truth of semantic originalism, they should lay their arguments on the table and make them explicit. If they want to deny the normative significance of the semantic content of the Constitution, but claim that the semantic content of other constitutional texts has normative authority and creates obligations of fidelity, they should produce the arguments that justify this distinction. And if they want to argue that some nonoriginalist theory of constitutional interpretation is justified on normative grounds, they need to produce the arguments.

Read the whole thing.