Daniel Drezner has an analysis of the polling numbers on gay marriage that leads him to believe that an effort to amend the Constitution to ban gay marriage will fail. He may be right—if that is, indeed, what the actual amendment proposed does. Of course, there is no real movement on an amendment—and the necessity of an amendment, the proviso attached to the president’s support, has yet to be shown.

One thing to be seen is the Massachusetts legisature’s response to the court decision that legalized same-sex marriage—and whether the court will accept a civil unions provision as a substitute. If the state legislature implements Vermont-style civil unions, and the court accepts them, there’s no immediate federal issue—the only states that have to accept civil unions under “full faith and credit” are ones who already have them. Everyone breathes a sigh of relief until the next state court goes “off the reservation,” and no amendment is necessary.

But what if the Massachusetts legislature does implement full-fledged same-sex marriage? Then it becomes a “full faith and credit” question, which asks the Supreme Court whether Congress can decide what things are subject to full faith and credit—in other words, whether the existing federal Defense of Marriage Act is constitutional. If the Supremes say yes, again no amendment is immediately necessary—however, some states may still face the question of whether their state constitutions require them to recognize same-sex marriages, the same question faced by the Massachusetts court, which starts another iteration.

Now suppose instead the Supremes decide DOMA is unconstitutional. Then necessity becomes the mother of invention—and the question becomes, what form will the amendment take? An outright ban on same-sex marriage is one possibility; another is a narrower amendment that allows the states to decide what marriages performed outside their own states they will recognize.

The other question, somewhat lost in the debate, is how Congress will demand ratification. Dan’s analysis assumes that Congress will require ratification by state legislatures, as is normally the case; however, the procedure used for ratification of the 21st Amendment (the repeal of prohibition)—ratification by special state conventions in 38 states—is also available, which turns ratification into more of a referendum process. In such a scenario, same-sex marriage opponents have an advantage in elections to these state conventions: not only are the opponents more numerous in most states; they also are more passionate about the issue. And unlike other failed amendments—think of the Balanced Budget Amendment, the flag desecration amendment, or even the Equal Rights Amendment—more people genuinely care about the issue. That could be enough to tip one or two states that Dan enumerates into the “yes” column, particularly for a narrowly-tailored amendment that simply says “my state gets to decide whether or not to recognize a same-sex marriage performed elsewhere.”

That being said, Dan is essentially correct to point out that amending the Constitution is a long, drawn-out process with a lot of veto points along the way. But a relatively large, passionate group can succeed in doing so, and I think the chances of a narrow amendment succeeding are fairly strong.