My friend and former professor Mike Dorf has a very interesting post on his blog about what he calls “the waning of constitutional tests.” Writing about Justice Thomas’s recent dissent from denial of cert in an Establishment Clause case, in which Thomas begs the Court to reconsider its endorsement test in this field, Dorf discusses the ways in which the famous Lemon test has been ignored but not overruled in the Court’s jurisprudence. Some key passages from his discussion:
The Supreme Court has become increasingly indifferent to the application of the multi-part, multi-stage “tests” that came to characterize its constitutional jurisprudence in the 1970s and 1980s. . . . [Examples omitted.] These are all important, high-profile cases of the sort that appeared to be governed by, or that seemed to call for the formulation of, a doctrinal test. Yet in each — and in satellite cases in both the Supreme Court and the lower courts — questions about what test applies seem not to play much of a role, if any. And note an oddity: Formal tests have come to play a decreasingly important role in the Court’s constitutional rights jurisprudence during the very same period that textualism has come to play an increasingly important role in statutory cases. Yet the reliance on rule-like tests and textualism are both versions of formalism. Why would formalism wane in constitutional rights cases just as it has been waxing in statutory cases? . . . . Another possibility is that the formalist Justices never thought that the old doctrine, with its multiple prongs and branches, was ever really rule-like. Justice Thomas says just this in his Utah Highway Patrol Ass’n dissent. He complains about the manipulability of the Lemon/endorsement test. But in the cases cited (including Heller), the Court is not replacing malleable tests with more determinate tests. If a somewhat malleable test is, well, somewhat malleable, then no test at all would seem to be more malleable still.
My own tentative view is more nearly the opposite of Justice Thomas’s. I think that the old tests were actually reasonably determinate, but that as the Justices found themselves increasingly unhappy with the results to which the tests led, they moved away from the tests. Seen in this light, the problem with the Lemon/endorsement test is not its flabbiness but that, if taken seriously, it would require much more separation of church and state than the Court’s conservatives can stomach. Ditto along different dimensions and in different directions for the other cases and tests.
I won’t engage the determinate/indeterminate question as such, and others might also ask whether it’s that tests have waned in constitutional law, or that they have been replaced by other, more categorical sorts of tests that work by up-front inclusions or exclusions (ie., Employment Division v. Smith) rather than a balancing of factors (ie., Sherbert v. Verner). I will add, though, that I’ve written about this a little in a Canadian context and will have more to say about it in an American context in my book First Amendment Institutions, which hopefully will be published next year. It strikes me as an interesting aspect of what we might think of as the lifecycle of constitutional adjudication, and perhaps adjudication in general:
1) A court approaches a new area and, to give some shape to the field and lay claim to adjudicative authority as against future courts, quickly lays down a seemingly clear set of rules or standards to guide adjudication in this area. 2) As time goes by, it becomes clear that these tests have not fully accounted for a variety of complex facts, different factual scenarios, and institutional and other contexts in which these tests don’t provide as much guidance as the judges initially thought. 3) The court is then caught in a tension between its institutional preference for clear, “lawyer-like” formal tests, whether rule- or standard-based, and its need to recognize the importance of various factual contexts. 4) For a while it tinkers with these tests, or recognizes various ad hoc exceptions and limitations, while maintaining allegiance to the fundamental soundness and applicability of those tests. 5) Eventually, something gives. It either replaces the test with a new one, or abandons tests altogether, or replaces one kind of categorical approach with a very different one; for instance, it replaces an interest-balancing standard with a formal, categorical rule. Of course, that test is subject to the same problems. So the cycle starts all over again, and the pendulum thus ends up swinging back and forth over time between a rule-based approach and a standard-based approach. 6) Either before or instead of step 5, the court will often simply leave the test in place while ignoring it and experimenting with various approaches, as it arguably has with Lemon. This may lead to sounder results in individual cases, but also results in a loss of guidance, coherence, and confidence. It is thus probably the case that step 5 can be delayed, but not avoided; it is inevitable at some point.
That’s my view, anyway. But read Mike’s post, which as usual says it better.
Posted by Paul Horwitz on November 10, 2011 at 10:13 AM
Comments
Hi Paul. Thanks for the shout. In response to my post, Mark Tushnet sent me a draft of an interesting paper of his on the First Amendment and risk in which he notes, inter alia, that the usual pattern is for rules to come late in the game. He channels an observation that Holmes made about how common law standards become rules over time, as courts see more and more cases, and so can be more precise about how the standards play out in particular cases. That point seems right but also in tension with what you and I both say. Maybe it’s possible for the evolution to go in either direction. It’s also worth noting, as Mark did in his email to me and as is implicit in what you and (the other) Marc say, that tests need not be rules, so abandonment of a test need not be abandonment of a rule. Anyway, thanks for the attention!
Posted by: Mike Dorf | Nov 12, 2011 11:49:45 PM
Paul, thanks for this, and for the link to Mike’s post. Extremely interesting. It may be that there never really was a time that purportedly hard-edged tests were applied strictly (at least in religion clause cases, sometimes I think this), or as strictly as we believe the fact of their hard-edgedness ought to suggest. The enunciation of a hard test in a judicial opinion is sometimes said to provide greater predictability than what a multi-factor approach can provide. But I wonder about this. Specifically, I wonder what we mean when we talk about the predictability that hard-edged rules provide. Is it more the simple (psychological) fact of comfort with having a written, enunciated rule, or instead the view that the fact of its enunciation gives litigants and the rest of us a greater sense for how courts will decide cases?
Or to put it another way: the vice of multi-factor approaches is sometimes said to be that one who follows them is only predictable to those who know how he will apply his multitude of factors. That is true, but it is no less true that one who follows a hard-edged rule is only predictable to those who know what his hard-edged rule is.
Posted by: Marc DeGirolami | Nov 10, 2011 12:02:57 PM