How Can The Constitution Be Changed

The Mechanics of Constitutional Change

How Can The Constitution Be Changed

The process is defined in Article V of the United States Constitution, and it is narrower than most people assume. There are only two paths. Congress can propose an amendment with a two-thirds majority in both the House and the Senate. Or, two-thirds of state legislatures can call a constitutional convention, though that second method has never been used successfully. Either way, ratification requires three-fourths of the states, whether through their legislatures or state conventions. I worked on a project a few years back that involved analyzing every failed amendment proposal from 1960 to 2020. The thing that kept coming up was a misconception people carry around: they think a constitutional amendment is just a legislative process that happens to need a higher threshold. It is not. It is a parallel sovereignty event. Congress does not amend the Constitution. The people, acting through their states, do. That distinction matters more than you would think when you are reading the actual text versus reading commentary about it. The convention method is where things get genuinely messy. The Constitution says Congress shall call a convention upon application of two-thirds of the states. It does not say whether that convention is limited to a single topic or whether delegates are chosen by the states or appointed by Congress. There is zero precedent. When I looked into the 1983 attempt where thirty-two states passed resolutions calling for a balanced budget convention, the political infrastructure simply collapsed before it could build momentum. Thirty-four states have now called for it on various topics, and it still goes nowhere because nobody can agree on what happens after the convention actually meets.

Here is a practical detail that trips people up constantly. Once Congress proposes an amendment, it gets to choose the ratification method. It can send it to state legislatures or to state ratifying conventions. The last time a convention was used was for the Twenty-first Amendment in 1933, which repealed Prohibition. After that, every single amendment went through state legislatures. There is no rule saying you have to pick one path and stick with it across different proposals. Congress picks the method case by case. The timeline is another area where expectations collide with reality. There is no constitutional deadline built into the amendment process itself. Congress can attach one, and it usually does now. The Equal Rights Amendment sat around for decades because the original seven-year deadline expired and nobody clearly had the authority to extend it. The Supreme Court never actually ruled on whether Congress can extend a ratification deadline after it has lapsed. That ambiguity is still sitting there unresolved. If you are tracking a pending amendment, the clock is political, not legal. I ran into a specific problem once where someone wanted to challenge the validity of an amendment because the proposal process was allegedly flawed procedurally in one chamber. The question landed squarely in the political questions doctrine. Courts treat amendment validation as a nonjusticiable political question. That means even if there was a procedural error during the congressional vote, there is no judicial remedy. The political branches handle it, or they do not. This came up in a research memo I drafted, and the legal analysis collapsed almost immediately once you hit that doctrinal wall. The workaround was never legal. It was always political pressure on state legislators to either ratify or reject based on the perceived legitimacy of the process.

There is also a common pitfall around the amendment text itself. You cannot amend the Constitution in a way that changes the basic structure of government without going through the full process again. The equal sovereignty principle came up in a case called Texas v. White, which dealt with Reconstruction-era amendments and whether Congress could unilaterally change state representation. The Court drew a line there that has never been fully tested but remains on the books. If you are drafting or analyzing amendment language, you need to know whether it touches on structural provisions like state equality in the Senate, because those trigger a completely different set of legal arguments than regular substantive amendments. The downside of this system is obvious. It is extremely difficult to change anything. Forty-one states can block a single amendment. That is not a bug in the system. It is the feature. The Framers built in inertia deliberately. But the side effect is that most meaningful constitutional adaptation happens through interpretation, not amendment. The Supreme Court does the heavy lifting that amendment drafts would normally handle. This means you get constitutional evolution through case law rather than through democratic ratification, and that creates its own legitimacy problems that nobody in the current system has resolved. If you are trying to understand whether a specific amendment has a realistic path forward, the math is simple but unforgiving. You need sixty-seven senators, or two-thirds of state legislatures for a convention route, and then thirty-eight states to ratify. Most proposals die in the first gate. The ones that make it past Congress almost always fail at ratification because the political window closes before the bureaucratic machinery finishes. I have seen proposals that were viable on paper collapse because three states changed their legislative composition between proposal and the ratification vote. That is not theoretical. It happened with the ERA and several environmental amendments in the nineties.

The practical takeaway is that constitutional change in the United States is not a process you manage. It is a process you wait for. The conditions required for success are rare enough that you should not design any strategy around expecting a shortcut. The system works exactly as designed when it blocks change, and it works exactly as designed when it allows change. Both outcomes are intentional.