Marshall, Thomas, Text, and Preemption: Take III

John Marshall’s opinion in Gibbons v. Ogden, I argued in an earlier post, is manifestly inconsistent with Justice Thomas’s textualist-originalist concurrence in Wyeth v. Levine. Only one of the decisions and opinions can be right on the question of federal preemption; and the conflict arises not from some wrinkle in the statutory scheme but from differing jurisprudential commitments. Mike Ramsey’s eloquent Originalism Blog post notwithstanding, I still think that I’m right, that Marshall was right, and that Justice Thomas was wrong. I harp on the disagreement because I fear that textualist originalism won’t let the Constitution do what it is supposed to do.

Gibbons Revisited

Recall the basic question in Gibbons: did New York’s monopoly grant over steamship traffic into its port conflict with a federal statute under which a would-be competitor had obtained a license for the coastal trade? Mike Ramsey helpfully excerpts the portion of Marshall’s opinion that purports to demonstrate the federal statute’s preemptive effect. He acknowledges that the license may be read to “signif[y] only federal approval, subject to additional state requirements as applicable.” But he calls that reading “strained,” “as the statute says that federally licensed ships ‘shall be … entitled to the privileges’ of coastal trading.”

Beware of textualists bearing ellipses: that is not what the statute said. It said that certain ships (identified in an earlier act) “and no others, shall be deemed ships or vessels of the United States, entitled to the privileges of ships or vessels employed in the coasting trade.” A natural reading of the text—“shall” refers to “deem” and says nothing about the privileges—strongly suggests that this was a mercantilist statute to exclude foreign vessels, not a free-trade statute to break down domestic barriers. Chancellor Kent treated the statute that way in the lower-court opinion in Gibbons, and adhered to it afterwards. Justice Johnson’s concurrence in Gibbons took the same position. Felix Frankfurter deemed Marshall’s statutory construction “exotic.”

My own preferred word is “question-begging.” A “license,” Marshall wrote, “transfers to the grantee the right to do whatever it purports to authorize.” And this textualist tautology is helpful—how? What did the grant purport to authorize—a privilege to ply a coastal trade in, say, diseased cattle or in convicts, regardless of any contravening state requirements? Surely not. Protection against any physical interference with traffic even on navigable but minor in-state waterways, as opposed to the majestic lower Hudson? Nope: see Willson v. Black Bird Creek Marsh Co. (1829), where the plaintiffs held a very similar federal license but lost. (Marshall conveniently neglected to discuss the license issue in that case.) In short: to make a go of this argument, you need a theory that tells you what the grantor was authorized to grant in the first place; what is or is not “authorized” by a license that is completely silent on the question; and what state requirements are preempted and why.  You cannot possibly get all that from the word “license,” either in isolation or even in statutory context (especially not when the context, as here, cuts against you).

The theory is famously explicated in the remainder of Gibbons. You can agree or disagree with it. For what it’s worth Justice Thomas and many other textualist originalists disagree because it strongly supports, and stops just short of affirmatively relying on, what later came to be known as the dormant Commerce Clause. But you cannot reasonably doubt that it’s the background theory, not the elliptical license argument, that does the work in Gibbons.


Under the Federal  Food Drug and Cosmetics Act (FDCA), drugs sold in the United States require an FDA-approved label—the elaborate, incomprehensible (to laymen) sheets you find inside every package. Every sentence is dictated by FDA requirements, down to the font and letter size. Violations of these requirements, and the sale of drugs without the label or a different label, are subject to very severe penalties. The statutory scheme operates to the explicit exclusion of any state regulatory (administrative) scheme. What Wyeth asks us to believe is that state juries may nonetheless hold drug manufacturers liable, for accidents caused by use in direct contravention of the federal label, on the grounds that the federally required label was inadequate. Meticulous compliance with federal requirements doesn’t preempt “failure to warn” liability under state common law. The federal scheme doesn’t preempt the state suits explicitly or by “necessary implication,” said Justice Thomas, concurring with the majority. And any appeal to the “purposes and objectives” of the federal scheme is verboten.

While  I believe this outcome to be irresponsible and not even minimally rational, let’s stick with the point at issue—the compatibility with Gibbons. Neither statute contained any explicit preemption provision with respect to the state law at issue. There, the similarities end—and not in a way that would allow one to harmonize the decisions. The license in Gibbons was readily available to all comers, provided they were American. Permission to sell a “safe and effective” drug is subject to a comprehensive federal statute;  a licensing regime that fills an entire part of the CFR, (pt 201) with seven subparts and 70 subsections; and to an approval process that, with respect to the product at issue in Wyeth, consumed fifty-four years. Yet we are supposed to believe that a perfunctory license conveys a right to trade in derogation of (wholly unspecified) state requirements, whereas the requirements for drug sales are merely a ”minimum”: even after the FDA has told Wyeth and the country that the drug is safe, any jury in the country is free to rule that it isn’t. To deny a preemption conflict by “necessary implication” under these circumstances is like saying in Gibbons that ships are free to sail and trade unless a state says otherwise.

Contrary to Mike Ramsey, it is not the case that the FDA could easily “clarify” that its requirements are preemptive: the agency said so in Wyeth, but got blown away. In any event, it is absurd to insist that the agency, or for that matter Congress, should have to say what is obvious in any event: if safety determinations are in the hands of juries, why have an FDA? But there, I suppose, is th’ole extra-textualist mistake against: to ask “why” is to inquire into purpose, and we can’t have that.

If you try to hang the difference on the word “license” (versus requirement), in complete abstraction from the statutory scheme, this really is (contra Mike Ramsey) a jurisprudence of magic words. I don’t think that this is what’s going on here, though. In Wyeth, as in Gibbons, what passes for statutory interpretation reflects deep background assumptions about how the Constitution is supposed to work. Justice Thomas, like Chief Justice Marshall, explicates those assumptions very forcefully.  And they are not the same assumptions. Marshall’s deep theory makes him take license with “license”; Thomas’s theory makes him look for reasons to deny preemptive force to any federal statute and scheme that doesn’t declare preemptive intent in haec verba.

John Marshall was not infallible. (Mike Ramsey is entirely right that M’Culloch, which poses similar problems, is very difficult.) On the other side, much in Justice Thomas’s preemption riff is right and commendable. In short, this is a debate worth having. But the first step has to be an acknowledgment of the conflict.