Caveat:  the following is largely based on my recollections of a law review article I drafted over 30 years ago (which has long since vanished from my files), and not on any current legal research.   But at least its not nihilist black pill ranting, so I’ve got that going for me.

The Supreme Court decision in the abortion case (Dobbs v. Jackson Women’s Health Organization 19-1392 Dobbs v. Jackson Women’s Health Organization (06/24/2022) ( ) took the very unusual step of overturning a major previous Supreme Court decision.  This is extraordinarily rare, and even more so because Roe was itself a rare Supreme Court decision that limited the power of the government.  I have not yet closely read the Dobbs decision, but on a quick review the decision did not address the issues discussed here.

When the Court rules that a statute is unconstitutional, the statute is not repealed.  These statutes remain on the books until repealed by the legislature.  Courts can neither create nor repeal statutes.  Many states did not repeal their pre-Roe abortion laws.  Arizona’s pre-Roe law outlawing elective abortions, allowing them only to save the life of the pregnant woman, is still on the books.  The precise legal status of these statutes is not clear (more on that below), but at a minimum they were unenforceable after Roe and until Dobbs.  These laws are enforceable now, though.

And now, for the philosophical/jurisprudential musings.  I am going to use a couple of straw man jurisprudences – “legal positivism”, which is the theory that the law is whatever the courts (or the legislatures) say it is; there is no underlying order or natural law, merely the pronouncements of those in power, and “traditionalism”, which is the theory that there is an underlying order and the law is the recognition of what that order is.  Traditionalism is the philosophical foundation for common law, laws “made” by the courts, and recognizes that law can be illegitimate.  Positivism is a reaction to traditionalism, based on the belief that the underlying order is invisible and likely fictitious, and that all we can really say that the law is, is what is on the books.  Positivism is also the basis for critical legal theory and its bastard stepchild, critical race theory.

The traditional view of court decisions, including Roe and Dobbs is that they “discover” the law, they do not “create” the law.  Creating law is the province of the legislature, after all.  Courts applying statutes (or the Constitution) are merely explicating what the law already is.  Under this view, the Dobbs ruling means that abortion laws were never unconstitutional; the Court ruled that there was no Constitutional basis for Roe.  Wouldn’t that mean that those laws were always Constitutional?  The Constitution itself hasn’t changed, after all.  If those laws were always Constitutional, does that mean that abortion providers (and, depending on the statute, the women who had abortions) are now exposed to criminal prosecutions for abortions performed before Dobbs?

Not so fast.  This gets into the retroactive application of court decisions, which is a murky topic.

Retroactive application is messy.  Court decisions apply to the parties to the case.  However, the dispute that gave rise to the case occurred before the court made its decision.  When the court decides the case, it is applying its decision to something that happened years ago – the decision applies retroactively to whatever caused the court case in the first place.

Just to simplify things, let’s say the court adopted a new rule to decide the case – it used to be that the courts were silent on whether the terms of service for an app could authorize the software company to install surveillance software on your phone, or the courts even allowed it.  This court decides that, no, that would be a contract of adhesion and the software company has to give you specific notice and you have to opt in to the surveillance software.  At the time the software company installed the software, there was no rule against it, but now there is and they have to pay damages for something that wasn’t against the rules at the time.  The court’s decision applies retroactively to actions that took place before the rule was adopted.  This is an essential part of the courts’ core purpose of resolving disputes; there’s really no way around it.

Not only that, through the miracle of precedent and the common law, the new rule applies to other cases as well.  The new rule on surveillance software can be applied in pending cases, but not cases that are final.  For pending cases, the new rule will have retroactive effect, just as it did in the case where it was created.  Once a case is finally adjudicated (appeals are not taken or are exhausted), it is closed and can’t be reopened even if the new rule would have given the losing party good arguments, or even a lay-down win.  The loser is stuck with it, even though under the (current) law, they would have won.

Saying the rule didn’t exist until a court announces it is legal positivism (and the sharp-eyed reader will detect the odor of positivism in my references to a “new rule”).  To a positivist, it is a new rule, and it is applied retroactively, because the rule didn’t exist when the dispute arose, but is applied anyway.

Under a strict application of the traditional view that courts discover, not create, the law, the rule existed the whole time and was just now “discovered” by the court, so its hard to say that it is being applied retroactively.  There is definitely a tension between courts “discovering” the law and their decisions having only a limited retroactive effect.  If that’s what the law has always been, then why doesn’t the court’s decision apply retroactively across the board?  There’s a couple of reasons:

Part of the rule of law (and due process) is “finality” – once its over, its over, and everybody needs to move on.  Allowing every case to be reopened every time there is a new decision that might affect its outcome is destabilizing, to say the least.  That sets one limit on retroactive application of court decisions.  And, really, a pretty clean and reasonable one.

Another part of the rule of law (and due process) is “reliance”.  People are supposed to be able to rely on what the rules are when they take action, which is when legal disputes arise.  No ex post facto laws, right?  Statutes have prospective effect, and often even come with delayed effective dates to give people time to adjust.  However, statutes are laws created by a legislature, not rules “discovered” by the courts to be applied to resolve a live dispute.

Its messier trying to protect reliance on the law when you are talking about court cases.  Do you say whatever cases are on the books at the time the dispute arose are the law for that dispute?  Nobody would seriously argue that the decision in a court case shouldn’t apply to the parties, so that piece of retroactivity is pretty uncontroversial.  From there, it’s a short step, and one that is hard to argue with, to apply that decision in other cases not yet final – the law should be uniform, after all.  Applying new decisions retroactively to cases not yet final is often justified by an interest in “progress”, which seems an odd thing for a “traditionalist” approach, but there it is.  Fortunately, you can apply both approaches reasonably cleanly in the vast majority of cases, because any change in the rules is likely to be very marginal, if its detectable at all.

Cases like Roe and Dobbs are not part of this vast majority – they are both cases where the Court pretty explicitly changed the rules.  In Roe, all those state abortion laws became unenforceable.  In Dobbs, all those state abortion laws became enforceable again.  Those aren’t marginal changes, and make it harder to be a strict traditionalist.  You can’t really say both Roe and Dobbs “discovered” the meaning of the Constitution without doing violence to the words “discover” and “meaning”.  This tips the scales away from saying that those state abortion laws were always in effect and so states can prosecute people for abortions that happened before Dobbs (reliance on the “settled law” of Roe be damned).  On the flip side, I don’t think people who were convicted under abortion laws were freed when Roe was handed down.

As a practical matter, major changes like this tend to be applied prospectively with only limited retroactive effect only in live cases, and do not reopen closed cases.  If you were convicted before Roe, you stayed in jail.  If you performed abortions before Dobbs, you really don’t need to worry about being prosecuted.  And this seems intuitively correct, even if its not a pure application of either of positivism or traditionalism.

Allow me to complicate matters.  Arizona recently (before Dobbs) passed something much like the Missouri statute at issue in Dobbs, which allows elective abortions up to 15 weeks.  The bill had “Construction” language that says that it doesn’t repeal the old pre-Roe statute, and that it doesn’t “make lawful an abortion that is currently unlawful”.  At the time the new law was passed, were elective abortions legal, or were only the abortions allowed by the old statute, to save the life of the pregnant woman, legal?  What exactly was the status of the old, pre-Roe, statute, before the Dobbs decision?

If Roe made elective abortions legal, rather than merely barring enforcement of abortion laws, then the new statute would mean elective abortions up to 15 weeks are legal in spite of the old statute which is still on the books and is now enforceable.  If Roe means the old statute was simply unenforceable, then elective abortions were always illegal but the state was prohibited from enforcing the law.  Consider that healthcare providers, including hospitals, routinely make representations that they comply with applicable law, both in contracts and in representations to government agencies.  If elective abortions were illegal the whole time, what happens to those representations?  If they were legal while Roe was in effect, why was the old law still on the books?  Does it make any sense to say that something explicitly prohibited by law is actually legal?

One difficulty with the narrower reading of Roe (that it made abortion laws unenforceable) is that’s not really consistent with the way we understand and talk about our Constitution, which is more consistent with the traditionalist approach.  Government power is limited, and governments do not have the power to infringe on Constitutional rights.  Statutes that do so are ultra vires, beyond the state’s power, and as such should be void ab initio, because the government doesn’t have the power to pass such statutes in the first place.

The Roe Court did not issue an injunction against enforcement; it said the law was invalid in some sense.  Consider the holding in Roe v Wade:

Our conclusion that Art. 1196 is unconstitutional means, of course, that the Texas abortion statutes, as a unit, must fall. . . .

Although the District Court granted appellant Roe declaratory relief, it stopped short of issuing an injunction against enforcement of the Texas statutes. . . .

We find it unnecessary to decide whether the District Court erred in withholding injunctive relief, for we assume the Texas prosecutorial authorities will give full credence to this decision that the present criminal abortion statutes of that State are unconstitutional.

There are a couple of very interesting things here.  First, the District Court gave Roe, as a party to the lawsuit, protection against enforcement of the statute, but did not enjoin enforcement of the statute generally.  That kind of old-school judicial restraint has fallen by the wayside in recent years, with District Courts showing no hesitation to enjoin enforcement across the board, rather than only with respect to the parties to the case.  Enjoining enforcement across the board is more consistent with the traditionalist approach, and also with judicial activism.  Does your head hurt yet?

The Supreme Court also did not enjoin enforcement of the statute, but really left its status in suspense with its rather vague statement that Texas prosecutors will give credence to the Court’s finding that the abortion law was unconstitutional.  What did the Court mean when it said the abortion law “must fall”?  Why didn’t it issue an injunction against enforcement?  Both of these are consistent with the traditionalist approach and our general understanding of the Constitution, but the Court did not go on to say that the statute had always been unconstitutional, was ultra vires, and thus was void ab initio and to be struck from the books.  The Court carefully avoided the implications of its ruling and left its full effect indefinite.

This post has gone on long enough, so I’ll just wrap it up by saying that perfect philosophical consistency is not something that we should wish for in the law; the world is too messy.  Traditionalist retroactive application of new rules seems unfair, but how else will you apply the rule adopted by the court to the parties to the case where it was adopted?  What else is consistent with our view of the Constitution and the belief that it should be applied as written?  Do we think there is an underlying emergent order to society that the law should reflect, or not?  How fair is it to adopt a new rule and then, as a positivist, continue to enforce the old rule and the new rule side-by-side until older disputes clear the system?  Isn’t the positivist approach, that the law is what the courts say it is when they say it, essentially the same as having a “living Constitution”?  But don’t the positivists have a point when they say the law is what is on the books at any given time, and that the “brooding omnipresence” of natural law is, essentially a social construct?