Does it really take a Theory to beat a Theory?

Guest Blogger

For the Balkinization 20th Anniversary Symposium

Mark Tushnet 

For quite a while I’ve been irritated by the aphorism that “it takes a Theory to beat a Theory” in constitutional law and interpretation.[1] It strikes me as the sort of false profundity that gets thrown around in first-year college dormitories. And it also seems to have links to the silly historical tradition of attempting to treat law as a discipline akin to the physical sciences. This post could be treated as a promissory note for a longer law-review style essay were I interested, as I am not, in developing them into that form (where the arguments would admittedly be more qualified and have some nuances that the post will omit).

Start with the importation of the phrase into legal scholarship. My sense of the matter is that it came into law from the physical sciences via law-and-economics, a subdiscipline with its own pretensions (at least in its early years) to science. According to Larry Solum, whose account accords with my memory, its first use was by Richard Epstein.[2] 

It's not clear to me that even in the physical sciences it takes a Theory to beat a Theory.[3] Without trying to borrow trouble by “relying on” authorities in other fields, I think that one version of the current view is that theory-change occurs in the physical sciences through a combination of the identification of anomalies that the Theory in question has trouble dealing with, a crisp alternative narrative, association with prestigious authorities, and, importantly, longevity.[4] 

The aphorism has (to me) a strange subtext: “Sure, my Theory might be flawed, though my colleagues and I are trying to work out the kinks, but it takes a Theory to beat a Theory. We have one, you don’t, so we win.” I see a couple of problems with that. (1) Some Theories might not even pass a threshold of plausibility, which seems to me to be the state of play with respect to original-subjective-intent originalism and original-expected-applications originalism. (2) Some Theories might fall of their own weight, so to speak. As I understand it, that’s the standard pre-Kuhnian account of the displacement of Ptolemaic stories about the solar system by Copernican ones. Ptolemists (if that’s the right noun form) were confronted with observational anomalies and dealt with them by adding epicycles to their core theory. The epicycles “worked” in dealing with the observations but the theory just seemed too Rube-Goldberg-esque to be “right.” That’s especially true if one desideratum of a Theory is that it consist of a relatively small number of propositions—a characteristic I’ll discuss later.[5] 

But even if in the physical sciences it does take a Theory to beat a Theory, it’s not clear to me that it’s true in purely normative disciplines. Outside of academic novels like David Lodge’s, no one really asks whether Nozick’s normative theory beats Rawls’s, or whether Bentham’s beats Kant’s. 

So, what sort of discipline is (constitutional) law? Here’s my candidate characterization: Constitutional law is an exercise in practical governance guided by norms, written and unwritten, whose goal is the betterment of the material and moral well-being of those for whom it is a constitution.[6] Constitutional Theory A is better than (“beats”) constitutional Theory B if it does better at achieving that goal. 

Constitutions, for present purposes, involve institutional design, and especially the allocation of authority between legislatures and courts (or, if you want, the allocation of authority among the political branches and the courts). And Constitutional Theories of the sort referred to in the aphorism are theories about interpreting the Constitution.[7] 

Theories dealing with constitutional interpretation are necessarily second-order. That is, they deal not with the general question, “What’s the best way to interpret these words?,” but rather with the more focused question, “What way of interpreting these words will contribute best to a system in which courts and the political branches share responsibility for promoting human betterment?” So, for example, standard examples like “What’s best way of interpreting a grocery list you’ve been given?,” aren’t quite apt because (in the first instance) there’s no system into which that interpretation fits. Or, put in a slightly different way, in the first instance the standard example doesn’t take into account the possibility that the grocery list’s author will adjust her behavior in response to your interpretation. And, transferred to the context of constitutional interpretation we run into problems associated with entrenchment: Entrenchment makes it difficult (the degree depending upon the stringency of the amendment rule) to adjust the overall system to respond to judicial interpretations (that is, to the Theory the judges adopt). 

We can modify the standard example to insert a “system” into it. So, for example, one possible answer to the question about interpreting the grocery list is, “What interpretation will contribute best to maintaining harmony within the family on an on-going basis?” And here the second-order issues arise.[8] Maybe the best first-order Theory will do better than any other Theory here, but the simple fact that it is a Theory and indeed the best first-order Theory (without for the moment any competitors) doesn’t tell us anything about the question we should be interested in. A worse first-order Theory or, perhaps more interesting, no Theory at all but only “decision according to law” eclectically defined might be better than the best first-order Theory. 

How could we go about determining which of two candidate Theories is better (how do we know when one Theory beats another—or whether “we don’t need no stinking Theory,” which I’m going to characterize occasionally as a Theory, is better than a particular Theory)? Here are three interrelated possibilities. First: experiment. Use Theory A for a while, then switch to Theory B, and assess the human condition at the end of each period.[9] Second: comparison across systems. See how constitutional systems in which Theory A prevails do compared to systems in which Theory B prevails (subject to all the usual qualifications about the difficulties of comparisons across systems, particularly with respect to outcomes as general as “human well-being”). For a while we might have compared Canada under “living tree” interpretation to the United States with [pause] something else, but the increasing interest in originalism in Canada might make direct contemporary comparisons even more complicated. Finally: historical. Examine how things went when judges used Theory A, then how they went when they used Theory B. 

I pursue the historical approach in the remainder of this post, but have an important caveat (similar to the kinds of concerns associated with the comparative approach). I think it highly likely that the marginal contribution of interpretive Theory to human well-being is ordinarily quite small.[10] To me, this suggests, with some important qualifications, that the intelligence devoted to controversies about Theory might be more productively used in some other law-related enterprise. Sometimes I think that “Theorists” are like Reverend Casaubon in Middlemarch seeking “The Key to All Mythologies.” Perhaps that’s too harsh if we take ideas about comparative advantage into account: Theorists might be spinning their wheels to no real effect but maybe they’re better at doing Theory than they would be doing something else law-related like attempting to come up with statutory language whose enactment into law would make people better off. So, no real benefit from Theory as done by Theorists but no real harm either—and maybe harm would result if Theorists attempted to do something else. 

Now for a sketch of a historical inquiry, based upon my work on the Hughes Court. Not to hide the ball: The Hughes Court’s justices didn’t have anything we would recognize today as Theory.[11] We can retrospectively impose Theory on what they did, as my examples will show. But, it would be a mistake to do that and say, “Aha! There’s a Theory at work here (particularly if we clean up some things around the edges).” I suggest that the Hughes Court’s justices were doing ordinary lawyers’ work, not Theory (or proto-Theory, if you prefer). I offer a couple of examples, written by Chief Justice Hughes (who was no slouch as a lawyer), out of many that could be deployed.[12] 

Wood v. United States (1936) dealt with the question of juror disqualification for implied bias in criminal cases because the juror was an employee of one of the litigants. The Court had held in 1909 that people who were employed by or received pensions from the United States couldn’t serve as jurors in federal criminal cases. At the time this wasn’t a serious problem in the District of Columbia but with the expansion of government employment during and after World War I (and the extension of the federal pension system), it became one as a really large proportion of the District’s population was disqualified. Congress responded in 1935 with a statute specifically allowing jury service by federal employees and pension recipients. The question in Wood was whether this statute violated the Sixth Amendment’s guarantee of jury trial or principles of due process and fundamental fairness: What constituted a “jury” within the Sixth Amendment’s meaning? 

Hughes characterized the 1909 decision as relying on pre-constitutional and early constitutional experience to come up with a definition of “jury”—and disparaged the accuracy of the decision’s characterization of that experience, relying upon what he called the government’s “elaborate research” showing to his satisfaction that British common law practice allowed Crown employees to serve as criminal jurors and that early U.S. experience didn’t systematically depart from that practice. (This, in the face of statements by Coke and Blackstone that such employees were indeed disqualified—because they didn’t acknowledge that many “menial” employees were so “remote” from prosecution that it was unrealistic to believe that they would be biased in the government’s favor. Wood’s lawyers “pressed” their reliance on Blackstone “too far,” according to Hughes.) The Court was “unable to accept” the 1909 decision as “determinative” in light of the English precedents and other precedents. 

Beyond that was the 1935 statute. Congress had in essence offered its own definition of what a “jury” was. All the Sixth Amendment required was that the “essential elements” of jury trial “as recognized … when the Constitution was adopted” be preserved. A 1930 decision holding that a defendant could consent to trial by a jury of eleven listed three “essential elements,” and Hughes observed that the list didn’t include anything about juror disqualification. He then relied on a “principle of construction” drawn from Seventh Amendment cases for the rule that the ”essential elements” should be determined with reference to “the fundamental consideration that ‘New devices may be used to adapt the ancient institution to present needs and to make of it an efficient instrument in the administration of justice….’” 

What was at stake were questions about implied bias, for which “the Constitution lays down no particular tests and procedure is not chained to any ancient and artificial formula.” Hughes then explained in some detail why it was unreasonable to think that a person who was a mere government employee or received a government pension would be any more biased in favor of the prosecution (in a case involving an ordinary crime like shoplifting) than would any other person in the District of Columbia. He concluded by referring to “the spectacle of the exclusion en masse from that service of a body of citizens otherwise highly desirable in point of intelligence and character, solely by reason of their employment by the Government, and the imposition in consequence of a heavier burden upon other citizens….” 

I’ve gone through the Wood opinion in some detail for two reasons. It’s a completely ordinary case, and it might look like it’s an originalist opinion (or at least has enough originalism in it to be treated as an example of originalism as “our law”). The better reading of the opinion, I think, is that it’s “eclectic” (a term I’ve used before) or “pluralist” (the term seemingly in current favor). A couple of points: Hughes doesn’t treat the originalist materials as evidence of something like an original public understanding. This is clearest in his discussion of Blackstone: he doesn’t say that the framers didn’t rely on Blackstone but says instead that Blackstone was simply wrong or incomplete. Hughes looks for the reasons behind the practices he describes and, finding the reasons good ones on the merits, goes with them. His identification of the essentials of a jury—how he construed the term “jury,” if you like—was expressly purposive (“efficient instrument”) and present-oriented (“present needs”). For me, the ordinariness of all this shows that the “construction zone” jargon is unhelpful—if the term “jury” has to be construed, what doesn’t (other than the Constitution’s mathematical terms)?[13] And, of course, there’s the looming question of practicality that impelled Congress to enact the 1935 statute, a question more or less directly alluded to in Hughes’s concluding phrase referring to the “heavier burden” disqualification would impose. For me, then, Wood shows a very good lawyer working with the available materials—text, practice, precedents, “authorities” like Blackstone, practicalities—to come up with what he and his colleagues regarded as a sensible result—an ordinary and uncontroversial exercise in “reasoned judgment” rather than Theory (or perhaps an illustration of the proposition that  “reasoned judgment” is a Theory). 

Blaisdell, the Minnesota Mortgage Moratorium case, is my second example. The most obvious criticism of Blaisdell is that the Court there held constitutionally permissible exactly the kind of legislation that the Constitution’s drafters and ratifiers had in mind when they endorsed the Contracts Clause. That’s an original-expected-applications point, which these days shouldn’t count as an objection. And Hughes expressly rejected original-expected-applications interpretation. Another criticism is that Blaisdell recognized some sort of “emergency power” exception to constitutional limitations. Hughes went out of his way to reject that proposition: “While emergency does not create power, emergency may furnish the occasion for the exercise of power”—which has to be correct. A third criticism is that Blaisdell treated the Contracts Clause as outdated and for that reason alone inapplicable to the problem at hand. Justices Stone and Cardozo appear to have held that view and they pressed Hughes to say something along those lines in his opinion. Hughes incorporated some of their ideas into his opinion but not their observations about how the Contracts Clause didn’t fit well with a modern economy. 

What Hughes did was this: deploy the more or less ordinary techniques of common-law reasoning to show that, properly construed, the Contracts Clause didn’t stand in the way of what Minnesota did. The Court, Hughes wrote, had to construe the terms “impair” and “obligation”: “The inescapable problems of construction have been: What is a contract? What are the obligations of contracts? What constitutes impairment of these obligations?” His opinion addressed these questions by examining the Court’s prior cases dealing with claimed Contracts Clause violations—essentially every one of which had found a constitutional violation. For Hughes, though, “broad expressions contained in some of these opinions went beyond the requirements of the decision, and are not controlling”—dicta don’t count. Good lawyer that he was, Hughes was able to find in each one of them a negative pregnant: this one was unconstitutional because it was permanent (implying, Hughes said, that imposing a time limit would make it constitutionally OK), that one was unconstitutional because it completely deprived the obligee of any remedy (implying that giving the obligee something would salvage constitutionality), and so on. The precedents established that a properly qualified suspension of mortgage payments was constitutionally permissible.[14] 

I could go on; in a more complete version I’d discuss Jones & Laughlin, another Hughes-at-his-lawyerly-best opinion, and the eclecticism of the paragraphs of Footnote Four. That would be tedious in the blog format so I move to what I conclude from all this about whether it takes a Theory to beat a Theory. 

For generations—really, I think, until the end of the last century—constitutional law and the nation got along just fine without a Theory. Judicial practice of constitutional interpretation involved the application of the ordinary tools of legal reasoning (distinguishing, analogizing, and the like) to the materials of constitutional law (text, history, practice, precedent, policy). Sub-“methods” weren’t strongly ranked. Originalist materials would sometimes play a large role, as in Wood, sometimes almost no role at all; precedent would sometimes play a large role, as in Blaisdell, sometimes almost no role at all. Judges would deploy the materials in ways they believed “made sense” given the nature of the problem at hand—or, perhaps better, the best judges figured out how to present their arguments to make them as rhetorically persuasive as they could.[15] 

One might say, perhaps, that generations of practice show that the Court had no Theory at all—and so much the worse for Theory. Or, perhaps one might say that there was a Theory at work, and that Theory was: interpret the Constitution the way a good lawyer would.[16] 

Now, though, the definition of Theory matters. I have the sense that in current discussions of Theory, eclecticism or pluralism just doesn’t count as a Theory—indeed, I’ve picked up hints that pluralism is inconsistent with the Rule of Law.[17] 

What then is a Theory of the sort that could beat another Theory? Here’s a stab at a “definition”: A Theory is a relatively compact set of general principles whose application by judges of different backgrounds, ideological commitments, policy preferences (and more) will converge on outcomes that are roughly similar though there might be disagreements on (sometimes important) matters of detail and disagreements around the edges. 

Why define Theory this way? Perhaps because converging on outcomes (in the sense I’ve identified) is a desirable feature of adjudication because (perhaps) it provides some safeguard against idiosyncratic exercises of judicial whim/preference/judgment.[18] And, the definition has the advantage of at least getting in the ballpark of an explanation for the apparent view that Footnote Four/John Hart Ely isn’t a Theory, nor is a Dworkinian/Rawlsian interpretive approach: both seem to rest on the kinds of preferences or judgments that Theory aims to guard against. 

Yet, I think, Theory runs into trouble in seeking that goal. Perhaps an algorithmic (or, to use an older term “mechanical”) Theory would provide sufficient safeguards. And, though the official position appears to be that no Theory is really algorithmic, I do sense a hope on the part of Theory-proponents to make Theory as algorithmic as possible.[19] If Theory isn’t fully algorithmic, though, the risk of judicial whim/preference/judgment sneaks in at the places where something other than an algorithm does the work.[20] 

That doesn’t mean that the desired convergence won’t occur. Convergence can occur if almost all judges share the same whims (though then they probably aren’t best described as “whims”)/preferences/judgments or if idiosyncratic judges can’t figure out how to exploit the interstices of the algorithm (in terms I’ve come to use, if the judges aren’t talented enough). So, Theory will do what it’s designed to do either if the selection process produces generally like-minded judges or if produces a pool of generally less talented judges.[21] 

Now, though, we can return to “the past.” One reason for thinking that past judicial practice should count as Theory is that the former condition—selection produces generally like-minded judges—was satisfied. Not that they were like-minded with respect to politics or the like, but they were like-minded with respect to what they would describe as the qualities associated with being a good lawyer. This isn’t to say, of course, that their like-mindedness is freed from the possibility of normative criticism, but only that their practice satisfied what I take to be the requirements of Theory. 

They might describe what judicial selection looked for was the capacity to display (good) “reasoned judgment.”  As I’ve noted, that’s disparaged today (mostly, I think, by people who lack that capacity), but it used to be praised as the “artificial reason of the law,” to use a phrase chosen by Charles Fried to describe what judges do (even when they’re doing constitutional law). 

One concluding thought: Perhaps the selection process used to create a pool of judicial candidates who had the capacity for good reasoned judgment but no longer does so. Instead, perhaps that process generates a pool of generally less-than-highly-talented candidates. If that’s the case, maybe it really should take a Theory (other than the Theory of reasoned judgment) to beat a Theory. I suspect, though, that that’s not a conclusion with which contemporary proponents of Theory would be comfortable. 

Mark Tushnet is William Nelson Cromwell Professor of Law emeritus at Harvard Law School. You can reach him by e-mail at mtushnet@law.harvard.edu.