The Curious Case of the Twenty-Eighth Amendment, Part II: Did the ERA Expire?

Remember this? It's been a while.

Way way back in January 2025, a bunch of activists were lobbying outgoing President Joe Biden to take action to recognize the Equal Rights Amendment as having been adopted as part of the Constitution. At that time, I wrote the first of what I envisioned as a three-part series on the matter. Then, before I had a chance to publish another essay, Biden did declare the ERA to have been adopted.

Then, three days later, Donald Trump became president and we all had bigger things to worry about.

As a result, I never got around to publishing the rest of the series. But the topic came up the other day on Bluesky, which reminded me that I do want to finish this someday. And today is, apparently, that day. So here goes.

As you may recall from Part One – which you, uhhh, might want to go back and reread, since it's been a minute! – the question of the ERA is principally about whether a ratification deadline, imposed by Congress not in the text of the amendment itself but in the resolution proposing it, is valid. If so, then the ERA expired in 1979, or perhaps in 1982 (after a three-year extension). If not, then the ERA probably became law in January 2020, when Virginia became the thirty-eighth state to ratify the proposal. (There is a separate question concerning whether states can rescind ratification, for a handful of states purported to do this with the ERA. For now I am going to assume that they cannot.)

I should say at the outset that what I have to say in this essay largely tracks the work of three feminist legal scholars, Allison Held, Sheryl Herndon, and Danielle Stager, in their 1997 law review article making the case that the ERA was still pending before the states. My analysis does not exactly follow theirs, but the broad contours are the same, and I want to emphasize the primacy of their work.

The question concerns the validity of a certain clause of the resolution proposing the ERA; it helps, therefore, to begin with that resolution:

Resolved by the Senate and House of Representatives of the United States of America in Congress assembled (two-thirds of each House concurring therein), That the following article is proposed as an amendment to the Constitution of the United States, which shall be valid to all intents and purposes as part of the Constitution when ratified by the legislatures of three-fourths of the several States within seven years from the date of its submission by the Congress:

Emphasis added, for that portion of the resolution is what concerns us.

As mentioned in Part One, when Congress first began imposing a seven-year time limit for ratification of new amendments, it put that limit in the text of the amendment itself. The Eighteenth Amendment was the first to have this limit:

Section 3. This article shall be inoperative unless it shall have been ratified as an amendment to the Constitution by the legislatures of the several States, as provided in the Constitution, within seven years from the date of the submission hereof to the States by the Congress.

This is not, strictly speaking, a limit on the time available to ratify the amendment. If an amendment with this clause were ratified after the seven-year deadline, it would become a part of the Constitution, just like every other amendment! Just, it wouldn't do anything, by its own terms.

Of course this is kind of unsightly. Even where an amendment like this is adopted in timely fashion, you get the meaningless trinket text quoted above becoming part of the Constitution forever. In the worst case scenario, where the amendment was ratified after the deadline, you would have a whole amendment, formally part of the text of the Constitution, doing nothing. For this reason, beginning with the Twenty-Third Amendment they switched to putting the deadline in the proposing resolution, evidently thinking the change insignificant.

But that's wrong. Putting the deadline in the resolution, rather than in the text of the amendment, raises a whole new question of law: does Congress have the power to do this? The concern is that Article V prescribes a specific role for Congress in the amendment process, and nowhere gives it the power to set a deadline like this:

The Congress, whenever two thirds of both Houses shall deem it necessary, shall propose Amendments to this Constitution, or, on the Application of the Legislatures of two thirds of the several States, shall call a Convention for proposing Amendments, which, in either Case, shall be valid to all Intents and Purposes, as Part of this Constitution, when ratified by the Legislatures of three fourths of the several States, or by Conventions in three fourths thereof, as the one or the other Mode of Ratification may be proposed by the Congress

See? Congress does two things: it frames the text of the proposed amendment, and it decides which mode of ratification to employ. Nothing about setting a time limit. And, of course, Congress has only those powers given to it by the Constitution. So if Article V does not say that Congress can do this, it would seem that it cannot.

The question, then, is where this power would come from, if not the express text. The answer comes from a Supreme Court case called Dillon v. Gloss (1921). Now, the Court does not often have anything to say about the mechanics of Article V. For the most part, the amendment process is treated as a "political question," into which the courts ought not stick their noses. (And for good reason, as we will see in Part Three of this series!) But there are exceptions, and Dillon was one of these.

As mentioned above, the Eighteenth Amendment was the first to be proposed with a deadline. A certain Mr. Dillon had been convicted under the National Prohibition Act, passed under the authority of the Eighteenth Amendment. Dillon argued that the amendment was invalid – and hence the statue was invalid, and hence his conviction was invalid – because it contained a time limit for ratification.

Now you might notice that this is absurd. The Eighteenth Amendment's deadline was in the text of the amendment itself. There is no possible argument that this is improper. Though it would be passing strange, Congress has the power to propose an amendment reading only "This does nothing," if it wants. The Eighteenth Amendment was written so that it might read this way, under certain circumstances. There was, strictly speaking, no deadline on its ratification by the states. This all is an exercise only of Congress's power to craft the terms of proposed amendments. The Court should not have given Dillon's claim the time of day.

It did, though. Indeed, the Dillon Court does not seem to have drawn a clear distinction between the time limit in the Eighteenth Amendment and the style of deadline we saw with the Twenty-Third and thereafter. The question, as the Court saw it, was what the Constitution itself has to say about "the time within which ratification may be had." Of course the text itself says nothing about this, "neither that it shall be unlimited nor that it shall be fixed by Congress." But, the Court thought, this "is not, in itself, controlling, for, with the Constitution as with a statute or other written instrument, what is reasonably implied is as much a part of it as what is expressed." (Of course, this is music to my ears.)

The question, then, is what "is the reasonable inference or implication? Is it that ratification may be had at any time, as within a few years, a century, or even a longer period, or that it must be had within some reasonable period which Congress is left free to define?" And the Court concluded, without much hesitation, that it is the latter: that Article V itself presupposes that ratification must be done within a short time after proposal, and that, as the exact length of time is unspecified, it falls to Congress to specify it.

The Court gave three main reasons for this conclusion:

First, proposal and ratification are not treated as unrelated acts, but as succeeding steps in a single endeavor, the natural inference being that they are not to be widely separated in time.
Secondly, it is only when there is deemed to be a necessity therefor that amendments are to be proposed, the reasonable implication being that, when proposed, they are to be considered and disposed of presently.
Thirdly, as ratification is but the expression of the approbation of the people, and is to be effective when had in three-fourths of the states, there is a fair implication that it must be sufficiently contemporaneous in that number of states to reflect the will of the people in all sections at relatively the same period, which, of course, ratification scattered through a long series of years would not do.

Moreover, there were at the time four outstanding amendments that had been proposed but never ratified: two of the twelve proposed in 1789 along with what became the Bill of Rights; the Titles of Nobility Amendment, proposed in 1810; and the Corwin Amendment, proposed in 1861 in an attempt to avert the Civil War by protecting slavery against federal interference. If there were no limitation on the time for ratifying an amendment, then, these amendments would be

still pending and in a situation where their ratification in some of the states many years since by representatives of generations now largely forgotten may be effectively supplemented in enough more states to make three-fourths by representatives of the present or some future generation. To that view few would be able to subscribe, and, in our opinion, it is quite untenable.

Accordingly, the Court brushed aside Mr. Dillon's challenge. Of course none of this inquiry was actually necessary to do so! But, precisely because the Court did not rely on the point about how the limitation was placed in the text of the amendment itself, its analysis would seem also to apply to a deadline in the proposing instrument. Congress does have the power to impose a deadline that way, per Dillon, precisely because the Constitution already requires that ratification happen swiftly upon proposal. Congress is only filling in the details of this pre-existing constitutional requirement, as it so often does.

That is the end of the matter, yes? Dillon affirms in so many words Congress's power to specify the time limit for ratification. No wonder, then, that when Congress switched to doing this through the proposing resolution, it thought nothing of it! As of 1972, when the ERA was proposed, there was no reason to doubt the validity of the deadline. That deadline passed in 1979, at which point the proposal was no longer pending before the states. The supposed subsequent ratifications were, therefore, a nullity: there was nothing for these states to ratify.

Right?

Did you spot the problem?

The problem is that Dillon cannot be good law anymore. At least, not if the Twenty-Seventh Amendment is really a part of the Constitution. Because, you see, the Twenty-Seventh Amendment was proposed in 1789, and ostensibly ratified in 1992. It is, indeed, one of the four pending amendments that the Dillon Court expressly considered, and expressly said must no longer be regarded as live. Go back to that last block quote. Few would subscribe, the Court said, to the "untenable" view that the ratification of some states "by representatives of generations long large forgotten" could be "supplemented" by "representatives of the present or of some future generation" to make the requisite three-fourths majority.

But this is exactly what happened with the Twenty-Seventh Amendment. When it was promulgated in 1992, there were, supposedly, thirty-nine states that had ratified it. Seven of these had done so immediately after it was proposed in 1789, the last of these Kentucky in 1792. Then Ohio had ratified it, maybe, in 1873. The other thirty-one ratifications all happened from 1978 to 1992. This is, precisely, unambiguously, the thing the Dillon Court thought preposterous.

Around when the Twenty-Seventh Amendment was promulgated in 1992, the Justice Department's Office of Legal Counsel wrote an opinion about its validity. (This is what OLC exists to do: the Archivist, charged with promulgating amendments once ratified, had requested advice from OLC on the matter.) OLC concluded that the amendment was valid; its reasoning is almost entirely based on the text. The text of Article V lists certain conditions for ratification: all of those had been satisfied. Therefore, the amendment had been validly adopted. It says nothing about any kind of time limit, therefore there isn't one.

But what, then, of Dillon? Well, first of all OLC suggests that Dillon is not controlling authority, because its discussion of the pending amendments was mere dicta. That is to say, it was unrelated to the disposition of the matter before the Court. The question in Dillon was only whether Congress has the power to impose a time limit, not "a State's authority to ratify a long-dormant proposed amendment." Therefore Dillon was just not relevant to the Twenty-Seventh Amendment question.

For good measure, though, OLC explained why, in its view, Dillon had been wrong to say that Article V requires that ratification be contemporaneous with proposal. Each of the three reasons Dillon adduces is unpersuasive:

• The notion that proposal and ratification are framed, in Article V, as a single enterprise is circular, assuming its conclusion.
• The reference to necessity does not mean anything about timelines. Indeed, Article V has several features designed to guard against a too-hasty amendment process. It is not strange if the necessity of a proposal endures over time, as indeed that of the Congressional Pay Amendment seems to have.
• Finally, although it may be true that Article V looks to consensus, consensus need not be contemporaneous.

Moreover, OLC suggests that Article V is a peculiarly improper place to depart from a strict, literal construction of the text, because it must be crystal clear which amendments have or have not been ratified. The Dillon rule would put states in a position of having to guess whether a proposed amendment was still pending for their consideration or not.

Now, I am not mostly interested in fighting over who is right on the merits, Dillon or OLC. As it happens, I tend to think Dillon has the better read of things – hardly surprising, given my own attitude toward narrow textualism. In particular OLC's rejoinder about consensus is preposterous. Here, look at this nonsense:

The sort of lasting consensus that is particularly suitable for constitutional amendments may just as well be served by a process that allows for extended deliberation in the various states. There have been occasions when it has taken decades to build the consensus within Congress needed for a two-thirds vote on a proposed amendment. In the absence of a time limit in the original amendment proposal, it would appear to be equally true that it may legitimately take many decades to build the three-fourths consensus required for the states’ approbation.

True enough that it may take decades to form the consensus needed to propose an amendment. But that consensus expresses itself in a single moment! The ERA had been introduced in Congress repeatedly since 1920. But it is only the fact that two-thirds of the Congress in 1972 all agreed to support it that matters. You cannot add up the votes cast in the 68th, 69th, 70th, 71st, etc. Congresses to make up the two-thirds majority: that degree of support must exist at a single point in time.

But again, this is not really the point. The point is that OLC's argument that Dillon does not control the question is outright preposterous. This is not how the concept of dicta works. Yes, the question before the Court was Congress's power to impose a deadline, not whether a state can ratify an amendment proposed some two hundred years prior with no deadline. But the reason why the Court answered that first question a certain way bears upon the second question.

This is the extremely normal business of adjudication in a common law system. If it were not so, then the business of citing precedent would disappear altogether, for the question in one case is never precisely the same as that in another. Obiter dicta, things a court says that are not a part of the holding and therefore are not binding precedent, is defined as the converse of the ratio decidendi, the reason for the decision. This means everything logically necessary to answer the decision before the Court. In Dillon, the question of Congress's power to impose a limit is decided on the basis that Article V itself requires a measure of timeliness. That is absolutely part of the ratio decidendi; to say otherwise is absurd.

Even this, though, is not really our concern. The question is not whether Dillon was, in formal terms, controlling legal authority for OLC in considering the validity of the Congressional Pay Amendment. Rather, the question is whether the validity of the Twenty-Seventh Amendment can be squared with the idea that Congress has a power to impose a deadline on the ratification of constitutional amendments. Insofar as the foundation for such an authority is the idea that Article V itself requires a measure of timeliness – which is, to my knowledge, the only foundation anyone has ever asserted – I think it plainly cannot be. As OLC itself admits, if this is true then presumably after two hundred years, the window had passed.

What does all of this mean for the ERA? If we acknowledge the Twenty-Seventh Amendment as valid, then I think we must, in hindsight, consider the deadline for the ratification of the ERA as having been ultra vires and void. Thus, although everyone would have believed, for good reason, that the ERA had lapsed in 1979 and was no longer pending before the states, this cannot actually have been right, not under a Constitution that contains the Congressional Pay Amendment. Therefore, in fact, the amendment was still pending before the states in 2017, when Nevada claimed to ratify it. And therefore, when Virginia ratified in 2020, it met the three-fourths threshold for adoption and became the Twenty-Eighth Amendment.

Again, I am inclined to favor Dillon's reasoning, reasoning on a blank slate. But we are not working with a blank slate. Everyone's copy of the Constitution lists twenty-seven amendments. And that, I think, is the one number that is clearly wrong. Maybe there are twenty-six amendments, because Dillon is right. Maybe there are twenty-eight, because it is wrong. But there cannot be exactly twenty-seven.

For what it is worth, there is a second Supreme Court case from the early twentieth century that is relevant to the discussion. Coleman v. Miller (1939) concerned the Child Labor Amendment, proposed by Congress in 1924 but never adopted. In 1937, some thirteen years after the original proposal, the state of Kansas debated ratifying the proposal, and seemed to do so. There was some procedural chicanery, which is not our concern. But also, opponents charged that the amendment was no longer pending for ratification due to the passage of time. Congress had not specified any time limit for this amendment, but under Dillon, this would not necessarily matter, since Article V itself must set some outer limit on how long an amendment may remain pending.

The case made its way through the Kansas courts and eventually to the U.S. Supreme Court, which declined to invalidate Kansas's ratification. On the question of timeliness, the Court held that where Congress has not specified a length of time, courts should not do so. Instead, in these cases the task of providing specificity to the constitutional requirement of timeliness falls to Congress at the back end of the procedure, when "the time arrives for the promulgation of the adoption of the amendment." At that point, when confronted with what seem to be ratifications from three-fourths of the states, Congress can judge whether the proposal had actually lapsed before some of those supposed ratifications. Its power to make this judgment at the moment of promulgation, on this view, is the same as its power to specify a deadline in the act of proposing an amendment.

OLC, I think, rather mischaracterizes the way that Coleman distinguishes Dillon. The point, in Coleman, is not that Dillon has no bearing on the power of the states to ratify a dormant amendment; it certainly does. Rather, the point is that only Congress can judge whether an amendment is still timely. It was not for the Court to decide that Kansas's ratification of the Child Labor Amendment had been a nullity. That does not mean, however, that the proposal really was still pending before the states, or that Kansas really did ratify it. Coleman seems to contemplate that states might act without really being able to know whether their action was a nullity – something that OLC would doubtless abhor, as introducing an unacceptable uncertainty into the amendment process.

In any event. We might think that Coleman suggests a compromise view that can square both the deadline for the ERA and the adoption of the Twenty-Seventh Amendment. Article V presupposes a measure of timeliness in the ratification of constitutional amendments. And, per Dillon, Congress can specify the exact window for ratification when it proposes an amendment. But, per Coleman, where Congress has not done this at the outset, the amendment remains (in effect) pending before the states, and it is for Congress to judge timeliness at the moment of promulgation.

The Congressional Pay Amendment, of course, had no ratification deadline attached when it was proposed in 1789. This means that in 1992, confronted with the supposed ratifications of thirty-nine different states, Congress could have decided that the amendment had lapsed in the two hundred years since its proposing. But it didn't! Instead, Congress passed a concurrent resolution – by a vote of 414 to 3 in the House, and 99-0 in the Senate – affirming the validity of the amendment. Hence there are, in fact, exactly twenty-seven amendments – although it should be noted that, even on this theory, Congress would have the power now to recognize the ERA as valid, just as it did for the Twenty-Seventh Amendment.

I think all of this is wrong, though. The problem is that, if Dillon is basically right that Article V requires some degree of contemporaneous ratification, then a determination by Congress that the Congressional Pay Amendment was still live in 1992 is absurd. No one can reasonably think that the constitutional requirement of timeliness, should one exist, might be satisfied by a ratification process lasting two hundred years. Now, if Congress were expressly charged with making this determination, then this would not matter. Congress has an exclusive power to judge the qualifications of its members, for example, and if a majority of one House votes that one of its members is actually twelve years old, no one has the power to question this determination, even if it is patently absurd.

But Congress is not expressly charged with anything here. Its power to determine the appropriate length of the ratification process, if it exists at all, is an implied power: implicit in the requirement, a part of Article V itself, of a contemporaneous process. And if Article V requires that, then the Twenty-Seventh Amendment is not valid, cannot be said to be valid. I do not think the congressional resolutions endorsing the Twenty-Seventh Amendment can be understood as an exercise of the power Coleman describes. They can only be understood as denying altogether that there is any requirement that ratification be timely.

And if that is right, if our constitutional system is committed to that view, then I do not see how we can deny that the Equal Rights Amendment has also been validly adopted, and is this very day properly referred to as the Twenty-Eighth Amendment.