CAN CONSTITUTIONAL BORROWING BE JUSTIFIED?
A COMMENT ON TUSHNET
Matthew D. Adler*

In his illuminating contribution to this Symposium, Professor Tushnet raises some very general conceptual questions about the nature of constitutional borrowing.[1] He also provides reason for concern about the justifiability of constitutional borrowing, although his ultimate conclusion seems guardedly optimistic. I share the skepticism evinced by much, if not all, of Professor Tushnet's Article. In this brief Comment, I retrace some of the same conceptual ground that Professor Tushnet covers, and then suggest that a problem that he describes - what I shall term the problem of "modularity"[2] - does indeed constrain the scope of justified constitutional borrowing, even in a normative environment particularly hospitable for the practice.

What is a constitutional borrowing? Certainly it is something more robust than the following: A constitutional provision, doctrine, structure, norm, or institution (call it P) succeeds in constitutional regime C1; the very same provision, etc., is adopted in C2, and succeeds there. That alone does not constitute a borrowing, much less a justified one. Imagine that C1 and C2 have never had any contact. C1 is a constitutional regime on Mars and C2 is a constitutional regime on Earth. C2 adopts P and it turns out that, unbeknownst to C2, P has long been in use on Mars. We would not want to say, I think, that in this case a borrowing has occurred, because there is no causal connection between the adoption of P in C2 and its use in C1, let alone a causal connection of the right kind to constitute a borrowing. What kind of causal connection constitutes a borrowing? I would guess it's something like the following: C2 borrows a constitutional provision, doctrine, structure, norm, or institution P from C1 in virtue of its performance in C1. That is to say, constitutional actors in C2 take P's performance in C1 as a reason to adopt it in C2.[3]

In this Comment, for consistency, I'll use "C2" to refer to the borrower regime, "C1" to refer to the borrowee regime, and "P" to refer to what is borrowed. P could be any part, large or small, of the constitutional regime: a single sentence in the text of the constitution, a whole article in the constitution, a judicial doctrine interpreting some part of the constitution's text, a set of formal or informal understandings among legislators, the executive branch, or even among the population at large as to what the constitution requires. I'll call P, a bit awkwardly, a "segment," by which I simply mean that it is some part of the overall regime of constitutional law in C1 and C2.

I will freely admit that there are very large and interesting conceptual problems in figuring out what it means to talk about "the very same" segment in C2 and C1. If C2 and C1 speak the same language, and P is a short bit of text in the constitution in C1, then the very same segment is adopted in C2 if the very same string of words is put into C2' s constitution.[4] But what if P is not a bit of canonical text, but rather a judicial doctrine, or a norm among legislators? Or what if P is a bit of canonical text, but C2 and C1 don't speak the same language? In such cases, there will be something like P (call it P*) that is adopted in C2, but how much like P does P* have to be, and in what way, for this to constitute a borrowing, let alone a justified one? I will ignore this very fascinating question, however, because Professor Tushnet seems to suggest there are serious problems with explaining how constitutional borrowing can be justified even if we're satisfied that the very same P as was in force in C1 has been adopted in C2.

One more tangential point: we can, in theory, imagine negative, neutral, or positive borrowings. We can imagine that constitutional actors in C2 take P's negative performance in C1 as a reason to adopt P in C2. For example. P could be the Dred Scott decision in the U.S. constitutional regime.[5] Or, we can imagine that constitutional actors take P's neutral performance in C1 as a reason to adopt it in C2. For example, P could be the Third Amendment to the U.S. Constitution,[6] which has had virtually no impact, positive or negative, on the constitutional regime in which it has figured.[7] Dred Scott and the Third Amendment strike me as deviant cases, however, at least from the standpoint of a normative theory of borrowing.[8] From a normative standpoint, what we're interested in is whether P's positive performance in C1--equivalently, P's success in C1--justifies P's adoption in C2.

If a borrowing, paradigmatically, is the adoption of a constitutional provision, doctrine, structure, norm, institution or other constitutional segment, in one regime, in virtue of its success or positive performance in another, then what does "success" mean, here? Professor Tushnet's treatment of this issue is quite revealing.[9] If the constitutional elite in a developing country C2 adopts segment P, with the intention of attracting investment and new immigration, recognizing that this will exacerbate poverty and misery among the existing residents of C2, and P in fact works as planned, is it a success? Is it a success just because it has the effect intended by those who adopted it? Professor Tushnet is skeptical about this,[10] and so am I. At least from the standpoint of normative constitutional theory, a successful constitutional segment P is not one which succeeds for just any general reason, such as realizing the intentions of the constitutional elite that adopted it, or remaining in place over the long run (even a terribly oppressive P might become entrenched in the society), or--to note another kind of success about which Professor Tushnet seems skeptical[11] -- satisfying the international community. Rather, at least from the standpoint of normative constitutional theory, a constitutional segment P only succeeds if it is normatively successful: if it performs optimally or at least satisfactorily under the best theory of what a constitutional segment like it should do.

Given how I've defined "constitutional borrowing," and given how I've defined "success," what would it take to justify a constitutional borrowing? When would constitutional actors in C2 be justified in adopting P there, in virtue of its success (in the normative sense I've just elaborated) in C1? Or, equivalently, when would P's success in C1 provide a justifying reason for its adoption in C2? There are various kinds of justification, but it strikes me that the normal justification for a constitutional borrowing is evidentiary or epistemic--namely, P's success in C1 provides strong evidence that it will be successful in C2. A justified borrowing will be one where P's success in C1 is a reliable indicator of--provides strong evidence of--its success if adopted in C2.[12] And this, I think, is the core of the problem for constitutional borrowing, at least from the standpoint of normative theory, for how can that ever happen? How can the success of a constitutional segment, in one regime, serve as a reliable indicator that it will succeed in another?

The answer to this question depends, in part, on what kind of segment P is--an amendment to the constitution, a provision in the original constitution, a judicial doctrine--and, in part, on our normative theory of what a constitutional segment of that kind should do. Imagine, for example, that originalism turns out to be the correct normative theory of constitutional interpretation. Originalists argue that a judicial doctrine is normatively successful only if it reflects the original meaning of the text of the constitution, or the original intentions of the constitution's framers.[13] Now let's further imagine that P is a bit of judicial doctrine in C1. If indeed originalism is correct, then P is a normative success in C1 only if it reflects the original meaning of the constitution in C1, or the original intention of the constitution's framers in C1. But that kind of success will hardly be a reliable indicator of success in C2, at least in general. For P will be successful in C2 only if P reflects the original meaning of the constitution in C2, or the original intentions of its framers. Unless the texts of the constitutions in C1 and C2 are in relevant part identical, and unless these identical texts were adopted at virtually the same time by the same framers, what reason is there to think that the fact P matches the original meaning or intentions in C1 is strong evidence that P will match the original meaning or intentions in C2?

Even if one is an originalist, the justifiability of constitutional borrowing depends on what kind of constitutional segment P is. This is a point that Professor Tushnet notes.[14] Professor Tushnet also observes a distinction between borrowing for purposes of constitutional design, and borrowing for purposes of constitutional interpretation.[15] Originalism is primarily a theory about judicial doctrine, about how judges should interpret constitutions.[16] It is decisively not a theory about what the drafters and amenders of constitutions should do. The originalist does not say that the drafters of a bit of text for insertion in the constitution should conform to some prior original meaning or intentions. Rather, it is perfectly consistent with originalism to say that the drafters or amenders of the text of a constitution should insert the provision that best advances basic normative criteria such as equality, democracy, liberty, distributive justice, etc.[17] For short, I will call such criteria "straight moral criteria."

Generalizing a bit, it strikes me that the most hospitable environment for justified constitutional borrowing is one in which, on our best normative theory of constitutional institutions, the criteria for P's success in C2 and in C1 are indeed straight moral criteria. This might be the case because the P at hand is a bit of judicial doctrine, and the best normative theory is a non-originalist theory, such that straight moral criteria are the appropriate criteria by which to measure doctrine.[18] Or, it might be the case because the P at hand is a bit of canonical text to be adopted by the drafters or amenders of the constitutional text, such that even the originalist can agree that straight moral criteria are appropriate.

Can constitutional borrowing be justified in an environment where the criteria for constitutional success are straight moral criteria? Imagine that, on our best normative theory, P is a success in C1 if it satisfies the values of, for example, democracy, equality, and liberty, and P is a success in C2 if it satisfies the very same values. In this hospitable environment, will P's success in C1 (thus defined) be a reliable indicator that it will succeed, if adopted, in C2? Perhaps not. First, of course, there's the problem of moral conventionalism. The moral conventionalist says that what democracy, equality, liberty and other moral criteria require is necessarily relative to a given society.[19] Liberty in C2 is necessarily some function of the traditions and social conventions in C2, and likewise for liberty in C1. So, for example, the activity of abortion, or homosexual sex, will be a morally protected liberty in C2 only if (in some way, at the stipulated level of generality) that activity has been traditionally or conventionally recognized as morally protected, and likewise in C1. Clearly, I think, moral conventionalism poses problems for explaining how constitutional borrowing can be justified. Unless the liberty, equality, and democracy-defining conventions in the two societies are the same or quite similar, it's hard to see how success in C1 will be a reliable indicator of success in C2.

I should say that my own views are not moral-conventionalist. Thus, I certainly don't think that moral conventionalism is, in fact, a decisive objection to the justifiability of constitutional borrowing. My efforts here are rather to delineate what justifiability will depend upon, even in what I think is a highly hospitable environment, namely one in which the relevant criteria for the success of C1 and C2 are the straight moral criteria of liberty, equality, democracy, etc. Let me conclude by noting one more problem of this kind, which I believe Professor Tushnet raised in his paper, and which (I agree with him) is a very important problem indeed: the problem of modularity.[20]

The problem is this: Imagine that P is the very best provision, doctrine, structure, norm, etc., that exists in C1, in light of democracy, liberty, equality or other straight moral criteria. Further, assume that moral conventionalism is untrue, so that P succeeds in C1, and might succeed in C2, independent of conventions and traditions in either regime. Finally, let us assume--heroically--that constitutional actors in C2 know that P is a success in C1. I say heroically because, unless the constitutional actors have observed C1 without P, it will be difficult to rule out the possibility that C1 would be even better without it; and even if the constitutional actors have observed C1 without P and then with P, it will be difficult to rule out the possibility that democracy, liberty, and equality improved in C1 because of factors other than the adoption of P. But, in any event, let us assume that constitutional actors in C2 know that P succeeds in C1. The problem that I understand Professor Tushnet to raise is that P might be successful because, in juxtaposition with the rest of the constitutional regime and the societal context in C1, it succeeds there.[21] But if we move the segment to C2, where the rest of the constitutional regime is different, and societal characteristics differ as well, it might not succeed. For example, if P is the institution of legislator standing,[22] then P might generate an optimal pace of constitutional litigation in C1 (given norms of litigiousness, the number of constitutional courts, the existence, or not, of divided government, the extent to which frivolous and non-frivolous separation-of-power claims can be readily separated from each other at the pleading stage, the existence, or not, of a public interest bar, etc.) but perhaps not in C2, where these other factors are different.[23] Or, if P is the right to assisted suicide,[24] then recognizing that right might be morally optimal in C1 (given, for example, the existence of a constitutional right to health care, such that doctors and hospitals have less financial incentive to engage in euthanasia), but perhaps not in C2.

In short, P's success in C1 will be a reliable indicator of its success in C2, assuming some such differences of this kind between C1 and C2, only if P is a modular success: that is, only if P functions successfully more or less independently of social conditions and a fairly broad range of other constitutional segments with which it might be placed. The seeming upshot here is that constitutional borrowing is justified only where the borrowed provision is modular.[25] Perhaps that conclusion is too strong;[26] but at a minimum, I certainly agree with Professor Tushnet's point that the problem of modularity--whether the success of P is highly sensitive to, or broadly independent of, the constitutional and social context in which P sits--goes to the heart of whether P can be justifiably borrowed. Again, this is a central problem even in the hospitable environment where straight moral criteria are agreed to be determinative of success, and even if the anti-conventionalist view of morality is true.

Given these grounds to be skeptical about the justified scope of constitutional borrowing, I am puzzled as to the reasons (both explanatory and normative) for Professor Tushnet's ultimate optimism. He concludes his Article by suggesting that "even if the study of comparative constitutional law proves not to have the kind of reciprocal pay-off about constitutional policy that we might hope for, it still may be useful as part of a lawyer's liberal education."[27] Perhaps Tushnet means only to be optimistic about the pedagogic benefits of comparative constitutionalism (its role in "a lawyer's liberal education") as compared to the justifiability of constitutional borrowing ("the kind of reciprocal pay-off about constitutional policy that we might hope for").[28] I wonder whether that distinction works: How can the success of P in C1 enlighten the constitutional lawyers of C2, if the success of P does not justify its adoption in C2? One might say that what enlightens constitutional lawyers in C2 is not the success of P, but the success in C1 of the larger module in which P figures. But if that's the claim, then it's also true that the success of that larger module justifies its adoption in C1 (at least given the conditions, e.g., non-originalism and anti-conventionalism, that would enable the module's success in C1 to enlighten C2's lawyers), and we're left with no distinction.

Tushnet may be right that education and borrowing are separate enterprises, and that the conditions for the justified studying of C1 are weaker than the conditions for justified borrowing from C1. In any event, with respect to the latter problem, justified borrowing, I wholeheartedly agree with the implication of the first part of Tushnet's final sentence: The study of comparative constitutional law may not prove to have the kind of reciprocal pay-off that we might hope for.

--------------------------------------------------------------------------------

[*]

Assistant Professor, University of Pennsylvania Law School. Many thanks to Jacques de Lisle, Seth Kreimer, and Mark Tushnet for their comments.

[1] See Mark Tushnet, Returning with Interest: Observations on Some Putative Benefits of Studying Comparative Constitutional Law, 1 U. PA. J. CONST. L. 2 (1998).

[2] See id. at n.38 and accompanying text (distinguishing between "discrete constitutional provisions" and "modules or complexes of interconnected constitutional provisions").

[3] The problem of justified constitutional borrowing is whether the reason that causes P's adoption in C2, "that P succeeded in C1," also justifies its adoption in C2.

[4] Cf. FREDERICK SCHAUER, PLAYING BY THE RULES: A PHILOSOPHICAL EXAMINATION OF RULE-BASED DECISION-MAKING IN LAW AND IN LIFE 53-62 (1991) (discussing how linguistic criteria are partly free-standing from other criteria that are used as justification).

[5] See Scott v. Sandford, 60 U.S. 393, 451-52 (1857) (concluding that under the Constitution, the federal government has no right to interfere with the reserved power of states to regulate slaves except for the purpose of protecting the property rights of slave owners).

[6] See U.S. CONST. amend. III ("No Soldier shall, in time of peace be quartered in any house, without the consent of the Owner, nor in time of war, but in a manner to be prescribed by law.").

[7] Or at least virtually no direct impact. Cf. Griswold v. Connecticut, 381 U.S. 479, 484 (1965) (partially grounding right of privacy in the Third Amendment). For a rare case in which the Third Amendment was directly at issue, see Engblom v. Carey, 677 F.2d 957, 962 (2d Cir. 1982), which held that property-based privacy interests protected by the Third Amendment extend to those founded on lawful occupation or possession with a legal right to exclude others.

[8] The normative question of whether a constitutional borrowing may be justified is hardly the only interesting or important question to ask about constitutional borrowing. For example, one might examine, from a historical perspective, the major constitutional borrowings that have occurred. Or one might ask, from a social science perspective, why borrowing occurs? However, I take Tushnet's focus to be primarily, if not exclusively, normative, and it is that aspect of his article that is addressed here.

[9] See Tushnet, supra note 1, at nn.14-32, 50 and accompanying text.

[10] See id.

[11] See id. at nn.22-23.

[12] Cf. Heidi Hurd, Challenging Authority, 100 YALE L.J. 1611 (1991) (providing an epistemic account of legal authority).

[13] See Daniel A. Farber, The Originalism Debate: A Guide for the Perplexed, 49 OHIO ST. L.J. 1085, 1086 (1989).

[14] See Tushnet, supra note 1, at n.41 and accompanying text.

[15] See id. at n.9 and accompanying text.

[16] See Farber, supra note 13, at 1086.

[17] See id. at 1098 (describing majoritarianism as one of the fundamental underpinnings of originalism and asserting that democracy and the related belief of equality are "master norms" in our society).

[18] See Matthew D. Adler, Judicial Restraint in the Administrative State: Beyond the Countermajoritarian Difficulty, 145 U. PA. L. REV. 759, 781 n.69 (1997) (citing leading non-originalists).

[19] See id. at 803-04 (discussing conventionalism in the context of constitutional theory).

[20] The problem comes up throughout Professor Tushnet's paper, but particularly in his discussion of the appropriate "unit of comparison" between constitutional regimes. See Tushnet, supra note 1, at preceding text accompanying n. 33.

[21] See id. at n.29.

[22] See Raines v. Byrd, 117 S.CT. 2312, 2322 (1997) (denying legislators standing to challenge Line Item Veto Act).

[23] See Tushnet, supra note 1, at text following n.122 (noting that "the effects of legislator standing may differ in systems with centralized rather than dispersed judicial review, in those with and without a well-organized public interest bar, and in those with parliamentary rather than divided powers political systems").

[24] See Washington v. Glucksberg, 117 S.CT. 2258, 2271 (1997) (denying right to assisted suicide under Due Process Clause); Vacco v. Quill, 117 S.CT. 2293, 2302 (1997) (denying right to assisted suicide under Equal Protection Clause).

[25] Actually, there are two ways to make this claim. The more stringent version of the claim is that only the fact "that P is a modular success in C1" constitutes strong evidence of its likely success in C2. The less stringent version is that the fact "that P is a success in C1" constitutes strong evidence of its likely success in C2, where the success of P in C1 is modular. I'd reject the weaker version: if constitutionalists in C2 do not know that the success of P in C1 is indeed modular, then the mere success of P in C1 is not strong evidence that P will succeed in C2.

[26] It may be too strong because constitutionalists might know both that P succeeds in C1 and that social conditions in C1 and C2 are relevantly similar, without P being modular. However, it seems unlikely that constitutionalists without a high degree of expertise about both systems will know (or have strong evidence of) the similarity of relevant broader conditions.

[27] Tushnet, supra note 1, at n.126.

[28] Id.