Self-Government: Part II - The Loophole in Plain Sight

A safeguard switched off in broad daylight, without a single step announcing itself as lawless.

In Part One I made you a promise: that the loophole was not a theory but a thing already done — with the result sitting in the Oval Office — and that I would show you exactly how it was managed. I also drew a line I need you to keep in view, because the entire trick depends on your losing it. The loophole is not Section 3. Section 3 is one of the Constitution's plainest self-defense clauses. The loophole is what was done to it: the Constitution's own doctrines turned against one of its own defenses, by no one in particular, with no one's fingerprints on the result.

So let me take the machine apart and show you how it runs. I'll warn you in advance — it is more elegant than any crime has a right to be.

Three hands, no fingerprints

Here is what happened, stripped to the bone.

A court held a trial. It heard the evidence — days of it, witnesses and exhibits and cross-examination — and made a finding of fact: that the man who is now President engaged in insurrection against the United States, after swearing an oath to defend it.¹ Sit with how extraordinary that is. Not a tweet, not a pundit's verdict, not a party-line resolution — a court of law, on the evidence, found the precise thing Section 3 turns on.

Then the case went up, and the Supreme Court of the United States did something stranger than anyone had braced for. It did not say the finding was wrong. It did not say he hadn't done it. It took the gravest factual finding a court can make about a candidate for the highest office in the land, set it gently to one side, and never laid a finger on it. What it said instead was: even if he did — granting every word of the finding — Colorado may not act on it. Enforcing this clause against a federal candidate, the Court announced, is not for the states — not Colorado, nor any state proceeding like it. It is the work of Congress.

And Congress — where his own party held at least one chamber the whole time, and both by the time it counted — did nothing. No hearing. No legislation. No vote. It let the clock run, and the calendar did the rest.

Now look at the shape of it, because the shape is the entire point. Three moves, three different hands, and not one of them will own the result. The trial court found he was an insurrectionist — and was told it had no power to act on its own finding. The Supreme Court was handed an adjudicated insurrectionist — and said, in effect, not our department; take it to Congress. Congress was handed the matter — and answered with silence, the one answer that never has to sign its name. At the end of the chain, a man the Constitution bars from the office holds the office; and if you go hunting for the official who switched the safeguard off, you will not find one. Each points down the line to the next. No one's fingerprints are on the erasure.

That is the trick. And I want to be exact about what is, and isn't, the scandal — because the easy version (corrupt judges, gutless congressmen) is true enough and still misses everything that matters. If this were merely venality, it would be an ordinary disgrace, with an ordinary remedy: throw the bums out. It is worse than venality. It is structural. Every actor in that chain could have been a person of spotless integrity and the same result would have been sitting there for the taking — because the chain was built so that no single honest choice anywhere along it was enough to stop the outcome. That is not corruption. That is the loophole — the missing wall from Part One — doing precisely what Part One warned it would do.

Which is why this was never really about one man. Strike his name out entirely. What the maneuver proves is that the Constitution's one safeguard against insurrectionists holding power can be switched off — quietly, lawfully, deniably — by the very faction the safeguard exists to stop. Build the machine once and it doesn't retire; it waits, reusable, for the next one. And see what the machine is really for, because it is bigger than any single seizure: it is how a faction defeats the one thing self-government cannot survive without — the power to reverse it — and the man in the chair is only the case that made the machine visible. So before we say another word about the man, we have to take the machine apart and watch it run. It runs in three moves. The first is the clause itself — and the clause, read honestly, does not say what the Court needed it to say.

"No person shall"

Read the clause. Not a paraphrase of it — the words:

No person shall … hold any office, civil or military, under the United States … who, having previously taken an oath … to support the Constitution of the United States, shall have engaged in insurrection or rebellion against the same, or given aid or comfort to the enemies thereof. But Congress may, by a vote of two-thirds of each House, remove such disability.²

Notice what kind of sentence that is. It is not a procedure. It does not say a person may be removed from office if a tribunal finds, after due process, that he engaged in insurrection. It says no person shall hold office who did. It is a flat bar — a prescription, in the same grammar the Constitution uses for no law shall abridge the freedom of speech or no person except a natural born citizen shall be eligible to the office of President. The disability is not a penalty some body decides to impose. It is a status that attaches the instant the facts are true, exactly the way ineligibility attaches to a thirty-four-year-old who wants the presidency. No court makes her ineligible. She simply is.

And then — hold onto this — the clause grants exactly one power over that status, to exactly one body, in exactly one direction. But Congress may, by a vote of two-thirds of each House, remove such disability. That is the entirety of Congress's role in Section 3's operative sentence: not to impose the bar, not to find the facts, not to decide whether it applies — only to lift it, and only by a supermajority. The Constitution hands Congress an eraser, and a demanding one. It hands Congress no pen.

Now the honest objection, and I'll give it more than its due: a status is one thing, and giving it effect in the world is another. Someone has to determine the facts are true; someone has to keep the barred man off the ballot or out of the chair. A flat bar with no one assigned to enforce it can sit unenforced for years. Grant all of it — concede, freely, that Section 3 does not enforce itself, any more than a speed limit pulls you over. The concession costs nothing, because it quietly swaps two different questions: whether the disability exists, and whether anyone has acted on it. The command is antecedent. It is true the moment the facts are true, written into the Constitution, owing nothing to anyone's later permission. Leave it unenforced and you have not repealed it; you have only let it lie dormant. And dormant is not dead. A dormant command is still law — it is simply law that no one has yet picked up and used.

Which leaves the only question that matters, the one the whole trick turns on. A dormant command has to be activated to bite: someone must find the facts, and someone must give the bar its effect. So — who? Hold that question — because activating the command, it turns out, is not one job but two, and the Constitution hands the two out very differently.

Finding is not enforcing

Picking up the dormant command is not one job but two, and prying the two apart is the move that unlocks everything. There is finding — determining, as a matter of fact, that this person did what the clause names: took an oath, and then engaged in insurrection. And there is enforcing — giving that finding its effect, keeping the barred man off the ballot and out of the chair. Two different jobs. The Constitution assigns them to different hands, and Anderson's root error was to mash them into one lump and hand the lump to Congress.

Before we sort out which hand gets which job, fix the one number that quietly decides every question that follows. The text sets a single price, and sets it deliberately high: it takes two-thirds of each House to lift the disability — to put an insurrectionist back into eligibility. That is the constitutional cost of his return, paid out loud, by an overwhelming and near-bipartisan supermajority. Hold that number against everything else, because it hands us a test sharp enough to settle the rest almost by itself: any reading under which some faction smaller than two-thirds of both Houses can secure the insurrectionist's eligibility — whether by blocking the finding, blocking the enforcement, or simply sitting on its hands — is void by construction. Hold onto that phrase — it carries the whole weight of Part One, and I will lean on it from here on. To be void by construction is to be null by the architecture of self-government itself: Part One's rule, that a reading turning the Constitution's own doctrines against its own defense is a child raised against its parent, and a child that kills its parent inherits nothing. The voidness is built into the structure; no one has to vote it into being. It has to be. A reading that lets a minority buy, for nothing, the precise outcome the Constitution prices at a supermajority is not interpreting the clause; it is quietly tearing off the price tag. Run each candidate for the job through that test, and the answers fall out on their own.

Who finds the fact?

Start with the worst answer — the one Anderson drifted toward: Congress. Suppose finding the fact, the determination that the man engaged in insurrection, is Congress's job, something Congress must convene and vote to establish. Run it through the test. An insurrectionist with his eye on the office now needs only enough of his own party to keep the finding from ever being made. Not two-thirds of each House to clear himself — merely enough to jam the machinery. Under the Senate's own rules, even forty-one senators — a filibuster, not a majority — could see to it that the fact is never found, the disability never bites, and he keeps the office the Constitution says he forfeited. And the filibuster is only the sharpest illustration; the defect is broader than any one procedure. A bare majority that votes no, a committee chair who never schedules the hearing, a Speaker who never calls the vote — any blocking faction short of two-thirds would do. Forty-one, or fewer, to buy what the text prices at sixty-seven in the Senate and two hundred and ninety in the House. That is not a reading of Section 3; it is the loophole wearing Section 3 as a mask — a minority seizing by obstruction exactly what the clause reserves for a supermajority. Void by construction.

And we should have known it would be, from the other direction, because we have seen this machine before and watched the drafters refuse to build it. A body that convenes, takes evidence, and votes on whether to disqualify is no new invention — it is impeachment, which had sat in the Constitution for eighty years by the time the Fourteenth Amendment was written.³ The drafters had the template in hand. Had they wanted Congress to try the fact of insurrection, they had only to say so. They said the opposite, in silence: no trial, no Senate sitting as a court, no presiding officer, not a syllable. Expressio unius — to build the political tribunal in one place and pointedly leave it out of another is to mean the omission. Congress was never meant to be the finder; it was kept out of the role on purpose. And the test tells us why that instinct was sound: hand the faction the finding, and you have handed it the loophole.

So where does the finding belong? Where the finding of facts has always belonged in this country — to a competent court, in a real case, on the evidence, under the discipline of due process, where party arithmetic cannot reach in and switch the answer off. A court finds the fact whether the political winds blow for the man or against him; it finds because finding is its work. Which is exactly what occurred: a Colorado courtroom, witnesses, exhibits, cross-examination, a finding. The finder the Constitution implies is the finder the country actually used.

Can a court even reach it?

Here the legal formalist plants a flag, and it is the sharpest objection in this whole stretch, so I won't skate past it. Federal courts do not rule into the air. They decide cases — real disputes, real parties, a real injury a ruling can redress — and they are barred from issuing advisory opinions, free-floating pronouncements that this man is an insurrectionist with nothing turning on the answer. So the challenge lands: if no court may simply declare the fact, isn't "let a court find it" an advisory opinion slipped in through the back door?

No — and the answer is already sitting in the clause we read. The disability is not a remedy a court hands out; it is a status the Constitution attaches the instant the facts are true, the way ineligibility attaches to the thirty-four-year-old or the foreign-born candidate. A court asked is this person eligible to hold this office? is not volunteering an opinion or inventing a punishment — it is finding a fact the Constitution has already wired to a consequence. And that is no abstraction floating free of any case: the question reaches the bench the ordinary way, riding a live dispute with parties and stakes on both sides — a ballot challenge, a fight over a candidate's qualifications, a quo warranto contesting his very title to the office. The insurrection fact is simply the thing those cases turn on. Concede the honest residue — that a court asked to declare in the abstract, with nothing concrete riding on it, that some citizen is an insurrectionist would be handing down forbidden advice — and it costs nothing, because that is not how the fact ever has to come. It comes the way every other eligibility question comes: fastened to a real contest over a real office.

There is a deeper form of the worry — that the finding redresses nothing if whoever must act on it stays free to ignore it. But that one answers itself a few steps from here, the moment we see that enforcement is a duty, not a favor: a finding wired to a consequence the Constitution itself compels is the opposite of advisory. The point is simple, and it stands on its own — the courthouse door is open. A court may find this fact in the course of deciding a real case, and nothing in the case-or-controversy rule shuts it.

Now for the objection that did the real work in Anderson — the very one that visibly frightened the Supreme Court: if any court can make this finding, don't we get a patchwork, Colorado disqualifying where Texas would not, fifty courtrooms and fifty answers? It is a real worry, and it deserves a real answer rather than a wave of the hand. So here is the answer — and it does more than dispose of the patchwork. It exposes the deepest dishonesty in the whole maneuver.

Many doors, one roof

Begin with what a patchwork actually is, because the cure is hiding inside the complaint. A patchwork is too many answers — and the remedy for too many answers has never been no answer; it is one answer. You do not stop fifty courtrooms from disagreeing by locking forty-nine of them and trusting the fiftieth to a faction. You let the courtrooms do what courtrooms do, and then you put a single roof over all of them.

Take the many doors first, because they are not a flaw to be tolerated but the very source of the design's strength. If only one body on earth were permitted to find this fact, you would need to capture only that one body to bury the fact forever — which is exactly what "only Congress" delivers: a single chokepoint, handed to the faction with the most to gain from jamming it. Many doors is the reverse. Let the fact be raised and found in any competent court — a ballot challenge here, a quo warranto there, a qualification dispute somewhere else — and there is no one lock to pick, no single room to seize. The truth has too many ways in. Decentralizing the finding is not the threat to uniformity; it is what makes suppression impossible.

Then raise the roof. Over all those doors sits one forum whose task is to take the question, once it has been found below, and settle it for the whole nation — the Supreme Court, doing the single thing it exists to do. Many doors in; one roof over. The fact may be found in fifty places; the answer is fixed in one. There is your uniformity — and you bought it without surrendering a thing: no chokepoint, no eligibility left to anything short of the two-thirds the text names, no courtroom door padlocked.

And now see the cruelty of what actually happened, because this is the hinge of the entire case. The Supreme Court is the roof. It already occupies the exact seat the design requires — the one tribunal that can take a contested national question and answer it once, for everyone. So the patchwork fear was never an argument against letting courts find the fact. It was an argument for the Court doing its job: taking the Colorado case, reviewing it, and announcing one answer for the whole country — disqualified, or not. Anderson seized that fear — a genuine fear, one that called for exactly one thing, the Court's own judgment — and used it as the excuse to do the opposite. It declined the merits, refused the single national answer only it could give, and handed the entire question to Congress instead. The Court did not lack a cure for a patchwork. It was the cure, and it climbed down off the roof.

And here someone will object, rightly: the roof is not safe either. The Supreme Court can be captured like anything else; a partisan majority could take the Colorado finding, review it, and throw it out. True — and we conceded exactly this in the first essay, where we proved that no design closes itself and that the people remain the last word. The roof is not bulletproof. But notice that the difference it makes was never between safe and unsafe. It is between laundered and un-launderable — and that difference is everything.

Force the Court onto the roof — force it to reach the merits, to take the Colorado finding and rule, in the open, thumbs up or thumbs down, on whether that finding was sound: whether the man got his due process, whether the evidence met the standard, whether the law was applied honestly. Now a captured Court that wants him seated has exactly one move left, and it must be made in broad daylight: it has to take a fully litigated finding of insurrection and throw it out anyway — naked partisanship, with the Court's own name signed to the page, impossible to mistake for anything but what it is. That is sunshine. That is an abuse no robe can launder.

Set that beside what Anderson actually did. It did not review the finding and reverse it in the open — that would have cost the justices their deniability. It refused to reach the finding at all. It reached instead for a threshold — this isn't ours to decide; it belongs to Congress — and let the question drop into a Congress that would do nothing, so that silence could accomplish what no justice had to sign for. That is the laundering: the hot-potato move let the Court strangle a self-defense clause and call the strangling modesty — humble, jurisdictional restraint. The same dead safeguard, dressed as legitimate jurisprudence instead of the partisan kill it was. A Court forced onto the roof has to own the body; a Court allowed to toss the question to Congress gets to walk away whistling. So no — making the Court the roof does not make the safeguard safe. Nothing does; we said so and meant it. What it does is take from every captured court the one thing it needs more than a friendly ruling: a way to do the deed without anyone being able to prove it did. The merits in the sunshine, or the dodge in the dark — that choice is the whole ballgame.

And there is a deeper futility in the dodge still. The abuser's hope is that by keeping the Court from ever affirming the finding on the merits, he keeps the disqualification from ever becoming truly national — leaves it a local verdict the country never ratified. But that mistakes what the roof does with a finding like this. It does not confer national reach on a fact; a fact takes its truth from the evidence, not from the altitude of the last court to touch it. A man's age is not thirty-five in Colorado and twenty in Texas, and an insurrection found on the record is not an insurrection only inside one county line. What the roof supplies is not reach but an audit — the single national check on whether the finding was sound: fair process, sufficient evidence, honest law. And an audit has the one feature a conferral lacks: to decline to fault a finding is to leave it standing. A roof that refuses to look has not withheld a blessing the finding needed; it has waived the objection it might have raised. So the dodge buys the abuser nothing but deniability — the predicate stands, carried and un-struck, as national as the day it was found. There is no move that unmakes a sound finding by refusing to look at it.

The remedy may fail, but facts must stand

There is a subtler defense of the seating than any we have met — let me give it its full strength. It grants almost everything: the disability is real, only two-thirds can lift it, a court may find the fact. But — it says — before that finding can bind the nation in a presidential race, the predicate must be established through some nationally competent process, and a lone state's judgment, reversed once it reached the Supreme Court, never ripened into that. So no faction smaller than two-thirds made him eligible; he was simply never nationally disabled in the operative sense. No nullification, no laundering — only a fact that never reached the form the office demands. It is the most serious objection in this whole essay, and the cleanest way to see what is wrong with it is to run the very same sequence on a fact no one on earth could dispute.

Suppose a twenty-year-old runs for President. Someone sues in a state court; the court holds a real adversarial case, finds the obvious fact — he is twenty — and, because Article II requires thirty-five, orders him off the ballot. The case climbs to the Supreme Court, which reverses: not because he is secretly thirty-five, and not because Article II tolerates a boy in the Oval Office, but on precisely Anderson's ground — that a single state may not enforce presidential qualifications in a way that splinters the national ballot into a patchwork. Then Congress does nothing. The voters, who like him, elect him anyway. His electors cast their votes.

What follows? The remedy has fallen; the fact has not. He is still twenty. Ballot access does not make him older. Votes do not make him older. Electoral votes do not make him older. Congressional silence does not make him older. A Supreme Court that looked away does not make him older. Sworn in, he is not the President — he is the beneficiary of a procedural failure wearing the President's seal. And the reason is airtight: a ruling about who may enforce a qualification cannot, by simple logic, reach back and change the fact the qualification turns on. The Court can knock a state's hand off the ballot. It cannot make twenty into thirty-five.

Now the objection this invites — and the example that disarms it. Age is mechanical, someone will say; a birth certificate settles it, where insurrection takes judgment, so the two are not alike. Grant the premise and watch it fail to do the work. Take, instead of the boy, the President already twice elected who wants a third term. The bare fact — twice elected — is as plain as a birthday; but whether the Twenty-Second Amendment bars this particular run is not plain at all. His own allies have floated more than one theory for slipping it — run as the running mate and ascend by succession, or read the Amendment to forbid being elected a third time but not serving one — each of which a court would have to weigh and decide. That is judgment, not a birth certificate. And still no one imagines that a friendly Court reversing on a threshold, a Congress sitting on its hands, and a winning vote count, together, hand him a lawful third term.

Run the whole ladder and watch what moves and what holds still. From age, to natural-born citizenship, to the third term, to insurrection, the amount of adjudication a qualification demands climbs steadily — a date, a definition, a contested reading, a five-day trial — while the one rule that never budges is that once the disqualifying conclusion has been properly reached and left undisturbed, nothing downstream can undo it. The difference between age and insurrection was never whether the bar can be switched off by institutional failure. It is only how much must be adjudicated before the bar stands established. Establish it — by whatever proof the case demands — and the constitutional operation is identical: the enforcing hand does not create the disability; it recognizes one the Constitution had already imposed.

And weigh what establish means here, because the safety of the whole doctrine lives in it. A finding is not a label a sympathetic court affixes; it is a positive burden carried — evidence marshaled, an adversary answered, the conclusion reached under the discipline of proof and left undisturbed on review. That is why the operation runs one way and cannot be reversed. You can carry the burden of proving a man engaged in insurrection; there is no burden one can carry to prove he did not. A court that hears the case and falls short has established nothing affirmative — its silence is the absence of a proof, not the presence of an innocence, just as an acquittal establishes no one's virtue. So a hundred forums that never reached the question, or reached it and came up short, do not outvote the one that carried it: proof is not a poll. And the feature that makes one finding enough is the same that makes it safe — though not, let me be exact, by making a false finding impossible. Nothing here is; a captured enough system can counterfeit a trial as surely as it can dodge one, and this series has promised no safeguard that closes itself. The safety is narrower and sturdier: the counterfeit cannot be worked in the dark. To fake the predicate, a faction cannot merely declare an innocent an insurrectionist — it must stage the whole ordeal, a trial court manufacturing an adversarial record and the roof affirming that record on the merits, its name signed to the page in full view. That is the very sunlight the dodge was built to avoid: the real Anderson worked by silence, the move that never has to sign itself — and a fabrication is the one thing that cannot. The predicate is hard to make and impossible to launder; and that, not any promise of impossibility, is the safeguard.

And here the comparison stops being a hypothetical, because we are standing inside it. The very opinion that let him keep the chair recites the Colorado courts' finding that he engaged in insurrection, and overturns it on nothing but the states-cannot-enforce ground — never reaching, never disturbing, the fact itself.¹ Strip off the robes and we stand exactly where the twenty-year-old's country stood: the enforcement door swung shut, and the constitutional fact did not move an inch.

One lawyer's escape still waits, and it should be shut before it can open. The Colorado judgment was reversed, the argument runs, so its finding binds no later proceeding — no preclusion, and therefore no predicate. But this mistakes a rule of litigation housekeeping for a constitutional cure. Issue preclusion governs exactly one thing: when a fact found in one case must be accepted, without relitigation, in the next.⁴ It is not an eligibility clause, and not a power to grant amnesty. Put it to the fork and it fails both ways. If preclusion recognizes the finding, the disability is recognized. If preclusion declines to bind a later court because the remedy was vacated, that refusal is not a ruling that the man is innocent of insurrection; it decides only that a later forum is not mechanically bound by the earlier one. Non-preclusion is not negation. A reversed remedy is not a merits reversal. A vacated order is not an amnesty.

Press one step further, because preclusion is itself a creature of the system — a doctrine derived to manage litigation, a child like every other. To let that child be wielded to erase a parent self-defense clause's predicate, for less than the two-thirds the text demands, is the loophole again in fresh costume: the very sub-two-thirds bypass we already named void by construction, now gowned in civil procedure. So be exact about what non-preclusion could ever buy. At its absolute most, it means a later forum is not automatically bound and might be asked to find the fact afresh — never that the fact is false, and never that the disability is gone. Here even that ceiling goes untouched: the finding has already run the full course — adversarial trial, appellate review, undisturbed on the merits — so the predicate stands established unless a competent tribunal rejects it on the merits, or Congress lifts it with the only key the Constitution cut for that lock. An established disability has exactly one off-ramp, and it is two-thirds of each House. A failed enforcement route is not one. A failed preclusive effect is not one. Anything offered as an off-ramp is only the price tag torn off again.

So set the sophisticated defense back down and weigh it honestly. To say Anderson "established his eligibility" — that the predicate never became operative — is to say that a court finding the fact, a higher court reversing only the remedy, a Congress doing nothing, and a sympathetic electorate together amount to a constitutional cure. But that is precisely the machine that would seat the twenty-year-old and hand the twice-elected man his third term, and no one will follow the argument there. A theory that dissolves every presidential qualification the instant its enforcement is blocked has not defended this outcome. It has proved far too much — and forfeited itself in the proving. And the last turn cuts deepest: age, foreign birth, the term limit all come with no designated enforcer at all — no Congress charged with policing them, no statute standing ready to keep the ineligible off the ballot — and still the twenty-year-old never becomes President. If the bar holds even where the Constitution named no one to enforce it, it holds all the more where a court has done exactly the finding the Constitution asks for. Of the whole family, Section 3 is not the weak case for the predicate surviving. It is the strongest.

Who carries it out?

Now grant Anderson the whole of its holding — grant, for argument's sake, that enforcing Section 3 against a federal candidate belongs to Congress and to no one else. Watch how little that actually concedes, once finding and enforcing are kept apart. Even at its most generous, the power Anderson describes is a power to enforce — to carry out the disqualification the found fact already commands, by whatever procedure Congress prefers. It is not a power to find. The finding is done, in court, upstream; nothing in the word "enforce" reaches back to undo it. So the most the Court can hand Congress is the role of executor — handed a verdict it did not render, and ordered by the Constitution to carry it out.

And even that runs straight onto the test. If Congress's enforcement is discretionary — if it may look at an adjudicated insurrectionist and simply decline to lift a finger — then the eraser has quietly become a veto, and we are back at the wall we already hit with the finding: the body does nothing, the disability never bites, and the supermajority the text demands is supplied by no one at all. Void, by the same construction, for the same reason. The only version of "Congress enforces" that survives the test is one in which enforcement is ministerial — a duty Congress must discharge once the fact is found, not a favor it may withhold. Grant Congress the executor's role if you like; you cannot grant it a veto. A veto is only amnesty by another name, and amnesty has a price — two-thirds, paid out loud — that the mere executor was never handed the power to waive.

The highest office cannot be the escape hatch

There is one door left, and the careful reader has been holding it in reserve this whole time. Everything so far has assumed that Section 3 reaches the presidency at all — but the clause never says the word. It bars the oath-breaker from "any office, civil or military, under the United States," and a man fighting to keep his chair can plant himself in that silence and ask: is the presidency an office under the United States? Am I an officer of it? If the answer is no, none of the machinery we just took apart ever starts. No finding, no two-thirds, no finding-versus-enforcing — because the clause was never aimed at this seat.

That is not a clever hypothetical. It is the dodge that was actually run, and it nearly held. The same Colorado trial court that sat through the evidence and found, as a fact, that the man engaged in insurrection then turned and let him off — on precisely this ground. It ruled that the presidency is not an "office under the United States" and the President is not an "officer" of it, so Section 3 does not touch him. The insurrection finding stood. The disqualification did not, because the court decided the clause had a hole shaped exactly like the one office that matters most. The Colorado Supreme Court reversed that part and got it right — resting on the plain sweep of any office and the absurdity of reading the chief executive out of it: for purposes of Section 3, it held, the presidency is an office under the United States and the President is an officer of it.¹ Then the case went up, and the Supreme Court disposed of it on enforcement and never reached the question at all — so it sits there still, a door a hostile reader can keep rattling. It deserves an answer on the merits, not a wave of the hand.

Here is the answer, and it is the one the whole series has been walking toward. The presidency question looks like a different species — not did he do it but does the clause reach him, a question of legal meaning rather than fact. Grant that. Then run it through the same engine anyway, because the engine does not care whether the threat arrives dressed as a fact or as an interpretation. Part One's principle was blunt: a derived reading that disables the Constitution's own self-defense is void, because the defense is the parent and the reading is only its child — and no child is allowed to kill its parent. An interpretation that carves the presidency out of Section 3 is exactly that kind of child.

Look at what the reading would actually do. It would bar an oath-breaking insurrectionist from serving as a state officer, a state legislator, a single member of the House — every lesser office in the land — while leaving open the one office that commands the armed forces, directs federal law enforcement, appoints the officers who execute every other law, and holds the pardon power. The clause would stand guard at every door in the house except the one that opens on the war room. It would stop the insurrectionist from becoming a deputy and wave him through to commander in chief. If Section 3 reaches any office, it must reach this one first — because this is the office from which an oath-breaker can do the very thing the clause exists to stop: not merely hold power, but turn the machinery of the state to shielding the insurrection itself. He can point federal prosecutors away from his confederates and toward his critics. He can pardon the foot soldiers, the organizers, the financiers — the whole ecosystem of the act he was found to have committed. The pardon will not lift his own disability; only that two-thirds vote can do that. But it lets him cover everyone around him — and his return covers more than the pardon ever could: every hand that knowingly held the door open can now hide behind the man it admitted. For the project as a whole, that is the next best thing to lifting his own bar — and sometimes the better one.

So the exemption is not a narrow reading of Section 3. It is a self-canceling one. A clause written to keep dangerous men out of power, construed to usher the most dangerous man into the most powerful seat, is not being interpreted — it is being switched off, as surely as if a faction had blocked the finding, and at the one spot where switching it off is fatal. If the clause reaches any office at all, it reaches the presidency first; and a reading that says otherwise is void by construction — the loophole's last door, picked with the same key as the rest.

The test a rival must pass

Put it all back together and the reading that seated him is not merely mistaken; it is void at every joint. Make Congress the finder the drafters deliberately refused to make it — void. Treat enforcement as a discretion when only a duty can survive — void. Carve the presidency out of the clause that must reach the presidency first if it is to defend anything at all — void. Three doors, three dodges, and one mechanism behind them all: take some job the Constitution split, or some office it covered, and use it to deliver — for less than two-thirds — the very eligibility the text prices at two-thirds. And the subtler defenses — that the predicate never truly vested, that a reversed remedy unmade the fact — we have already run down and found void by the same construction. Read honestly, against the single price the text actually sets, Section 3 barred him the instant the facts were true; the court's finding did not create that bar — it made one already standing legally cognizable. Everything that came after — the hand-off, the silence, the swearing-in — was not the Constitution working. It was the loophole working. The man holds the office; the Constitution says he cannot; and nothing that put him there survives the test the text itself supplies.

And that last phrase is not a flourish; it is an invitation. The test the text supplies is also the test any rival reading must pass — so let me set it down in the open, not as a dare but as the plain standard any serious account has to clear. I am not asking you to take this reading because it is the comfortable one. Constitutional self-defense is hard, and every theory of it — mine first — deserves to be pressed. I ask only this: that a rival carry the same burdens this one carries, and explain the same facts, without breaking the Constitution somewhere else.

So try to build it. Construct a reading under which an established Section 3 disability still leaves the candidate free to hold the office, though Congress never lifted the bar by the two-thirds the text demands. Then run it through three questions, and watch where it goes.

Does the reading let some faction smaller than two-thirds accomplish — by silence, by obstruction, by ordinary majority vote — what the Constitution assigns to two-thirds alone? Then it fails the text.

Does it also imply that a twenty-year-old, a foreign-born citizen, or a President already twice elected could hold the office lawfully — if only the right official declined to act, or the Court reversed the remedy and spared the fact, or enough voters chose him anyway? Then it proves too much, and forfeits itself in the proving.

Does it concede that once the right derived machinery is captured — the courts, the count, enforcement discretion, the doctrines of finality — the capturing faction may turn that machinery against the parent Constitution and keep the office the Constitution bars him from? Then it has not preserved self-government. It has abandoned it.

A rival may take any road it likes — finality, federalism, due process, national uniformity, the political-question doctrine, the legitimacy of the ballot, the competence of one institution over another — and it may be built with more care than I have managed here. But it has to survive all three questions at once: the two-thirds the text names, the qualifications no one would read away, and the rule that nothing derived from the Constitution may be turned against it. Fail a single one, and the reading has not displaced this account. It has only mapped one more route by which a constitutional order is argued into becoming complicit in its own defeat.

Another door in the same wall

And Anderson was not the last move; it was not even the only one that year. A few months later, in Trump v. United States, the Court reached for a second doctrine found nowhere in the text — a sweeping presidential immunity — and aimed it at the same wound from the other side.⁵ Where Anderson made Section 3 harder to enforce against an insurrectionist candidate, immunity made the underlying conduct harder to prosecute, and harder even to lay before a jury as evidence. Different door. Same missing wall.

Consider the absurdity through one of the oldest self-defense clauses of all — the writ of habeas corpus, the guarantee that the government cannot seize a person and hold him beyond the reach of any court. The Constitution treats even suspending it as nearly unthinkable, permitted only in cases of rebellion or invasion. Yet ask the question the immunity doctrine forces on us: what would-be tyrant needs to suspend habeas at all, if he can order the man killed as an official act, wrap the killer in federal authority, seal off any inquiry into the command that sent him, and pardon away whatever federal charge survives? Habeas can drag a prisoner back before a judge; it cannot resurrect a body. Reach for the state-law answer and it dissolves in your hand — a state may still have a murder statute on its books, but the proof runs through presidential orders, chain-of-command deliberations, and official-act communications the immunity doctrine risks placing beyond any prosecutor's ordinary reach. And even that assumes a prosecutor willing to bring the case. A remedy that asks a local official to indict the very President whose theory of power treats violence against his political obstacles as an official act is not accountability — it is a dare. It is the same trick that runs through this whole essay: split the deed across enough locked rooms and no single forum can ever assemble the entire crime. That is not a refinement at the edge of executive power. It is the hole beneath the whole floor. It is Article II made sovereign over the Constitution that created it.

And this is no horror conjured after the fact by dissenting justices. The assassination scenario was put directly to Trump's own lawyer — twice. In the D.C. Circuit, Judge Florence Pan asked whether a President could order SEAL Team Six to kill a political rival; his lawyer answered that such an order could not be prosecuted unless the President were first impeached and convicted by the Senate. At the Supreme Court, Justice Sotomayor pressed the same hypothetical, and his counsel allowed that it "could well be an official act."⁶ When the decision came down, Sotomayor traced where the doctrine pointed — assassinate a rival, stage a coup, take a bribe for a pardon: Immune, immune, immune — and the majority waved it away as fearmongering over extreme hypotheticals. But the hypothetical was never the dissent's invention. It had been spoken aloud, in open court, to Trump's own attorney, who declined to rule it out. The abyss was not hidden from the Court. It was named — and the Court wrote its doctrine broad enough to leave the door standing.

I won't try the full immunity case here — it has its own machinery, and Part Four gives it the reckoning it has earned. Notice only the shape, because by now you know it cold: a doctrine derived from the Constitution, turned to shield the one man accused of striking at the constitutional order itself. It takes nothing from the Section 3 verdict, which comes from Section 3 and stands on its own. What it adds is the pattern. Twice in one year, when the Constitution's self-defense reached toward the presidency, a doctrine drawn from the Constitution reached back and made the defense stop short of the chair. Once is an outrage. Twice is a design.

A verdict is not a remedy

So the Section 3 proof is finished, and it lands where Part One warned it would. Read against the one price the text actually sets, the man holds an office the Constitution denies him, and every doctrine, handoff, and refusal that put him there is void by construction.

Not my construction. Not partisan construction. By the very premise — and promise — of self-government itself. Grant only that this Constitution exists so that a free people can govern itself, and the nullity follows on its own. A reading that lets the constitutional order be destroyed from within — its own defenses bypassed, turned against the very thing they were built to guard — cannot be, is not, compatible with an order whose whole purpose is to keep a people in charge of itself. It is the greatest betrayal that premise can suffer. This is what self-government means, taken at its word.

But look hard at what that verdict does not do. It does not move him out of the chair. He is still there as you read this — sworn in, wielding the office, every day since. A safeguard proven dead on the page is still dead in the world. We said it in the first essay and it comes due now: the parchment has no army. A void is only as good as a people willing to refuse to honor it.

Which leaves the question this whole series has been driving toward, and it is no longer abstract. If the machine was built so that no honest actor inside it ever had to switch the safeguard off — and so that proving it void switches nothing back on — then who does? Not the court that climbed down off the roof. Not the Congress that ran the clock. The cure was never going to come from inside the machine; Part One spent five fields of self-reference establishing exactly that. It has to come from the one place the system cannot capture: the people, acting through the instruments built to carry their judgment. How they do it is the work of the parts ahead: first the guardrails — what a free people may lawfully reclaim, and the lines that keep the reclaiming from becoming a tyranny of its own — and then the machinery, by which a disqualification void by construction is actually recognized and given effect when every door is itself capturable.

And there is a second thread I left deliberately untied, because it is too dangerous to tug at without its own set of locks. We watched a chain of hands hold the door open — and I told you that every hand that knowingly did so can now hide behind the man it admitted. But the Constitution has something to say about hands like those. If seating an adjudicated insurrectionist is itself a way of giving aid to the insurrection, then the people who knowingly did the seating are not bystanders to the betrayal — they are part of it. That is a claim with a blade on both edges: get it wrong, and "aid" curdles into a purge, a license to brand every opponent a traitor. So it cannot be tossed off in a closing line. It needs its own set of locks — and it gets them when we build the machinery, in Part Four.

That is where this ends, and where it turns: not on the machine, but on the only thing that was ever going to stop it. A constitution cannot correct itself any more than it can defend itself — we proved that on the road to Trenton. So something else must. The self-correction is the people. Parts Three and Four are how they do it.

If you want it the moment it lands — the repair, and the reckoning — subscribe. This is the fight we don't get to lose.

Read next → Part Three: The Guardrails (Coming July 4 - 2PM)

---

Article Sources:

1. "Anderson v. Griswold, No. 23SA300 (Colo. 2023)", Colorado Supreme Court; and "Trump v. Anderson, 601 U.S. 100 (2024)", U.S. Supreme Court (per curiam).

The exact posture matters, because the trick depends on blurring it. A Colorado district court held a five-day trial and found, on the evidence and by clear and convincing proof, that the man now President engaged in insurrection on January 6, 2021 — but ruled the presidency was not an "office under the United States" and let him stay on the ballot. The Colorado Supreme Court reversed that second part (the presidency is an office; the President an officer) and left the insurrection finding intact. The U.S. Supreme Court then reversed on a different ground entirely — that a state may not enforce Section 3 against a federal candidate, which it said is Congress's charge under §5 — and pointedly never reached, much less disturbed, the finding of insurrection. So "he was never really found to have done it" and "but the Supreme Court reversed it" are both false: the finding was made after an adversarial trial, survived state appellate review on the merits, and was left standing. What the high court changed was who may act on it, not whether it was true.

2. "U.S. Const. amend. XIV, § 3", Constitution Annotated (congress.gov).

The disqualification clause, quoted in full from the official Constitution Annotated text. Two features of its grammar carry the section's whole argument. First, it is written as a flat bar — "no person shall ... hold any office" who, having sworn an oath, then engaged in insurrection — the same prescriptive form as the natural-born-citizen and age requirements: a status that attaches when the facts are true, not a penalty a tribunal elects to impose. Second, it grants Congress exactly one power over that status, in one direction — "by a vote of two-thirds of each House, [to] remove such disability." The text hands Congress an eraser and no pen, the single textual fact from which the rest of the section's reasoning unspools.

3. "U.S. Const. art. I, §§ 2–3; art. II, § 4 — the impeachment power", Constitution Annotated (congress.gov).

The section's expressio unius argument turns on a contrast with these clauses. The Framers already had, and had had for eighty years, a full apparatus for a political body to take evidence and vote to disqualify an officeholder: impeachment by the House, trial by the Senate, with disqualification from future office among the available judgments. When they wrote Section 3, they reproduced none of it — no trial, no Senate sitting as a court, no presiding officer, no political vote to impose the bar. Cited here not as the authority that settles the question but to show the drafters knew exactly how to build a congressional fact-finding tribunal when they wanted one, and pointedly did not; the silence is a design choice, and the choice was to keep the finding out of Congress's hands.

4. "Collateral Estoppel (Issue Preclusion)", Legal Information Institute, Cornell Law School; see also Restatement (Second) of Judgments § 27 (1982).

The black-letter rule of issue preclusion (collateral estoppel): an issue "actually litigated and determined by a valid and final judgment" is conclusive in later litigation (Restatement (Second) of Judgments § 27). The essay invokes the doctrine only to subordinate it. Two features matter. First, preclusion governs when one court's finding binds a later court — a rule of litigation economy, not a measure of constitutional truth or eligibility. Second, its predicate is a valid and final judgment, so a judgment reversed or vacated may lose preclusive force. That cuts the opponent's way only on the surface: if Anderson's reversal stripped the Colorado finding of preclusive effect, the most that follows is that a later tribunal is not mechanically bound to accept the insurrection finding — not that the finding was false, and not that the constitutional disability was lifted. Non-preclusion is not negation. The disability is a constitutional status the text removes by one route only — two-thirds of each House — which no doctrine of litigation finality can supply.

5. "Trump v. United States, 603 U.S. 593 (2024)", U.S. Supreme Court.

The immunity decision, handed down months after Anderson. Chief Justice Roberts, for a six-justice majority, held that a former President has absolute immunity for the exercise of his "core" constitutional powers, at least presumptive immunity for all other official acts, and no immunity for unofficial acts — and barred prosecutors from using official-act conduct even as evidence. Justice Sotomayor's dissent traced where the doctrine points — that a President could, as an official act, order an assassination, stage a coup, or take a bribe for a pardon, and be immune ("Immune, immune, immune") — while the majority dismissed those scenarios as "fearmongering on the basis of extreme hypotheticals." Cited not to relitigate the case but to mark the second instance of the pattern: a doctrine drawn from the Constitution, turned to shield the very conduct the Constitution's own defenses exist to reach.

6. "Oral Argument, United States v. Trump, No. 23-3228 (D.C. Cir. Jan. 9, 2024)", U.S. Court of Appeals for the D.C. Circuit; and "Oral Argument, Trump v. United States, No. 23-939 (U.S. Apr. 25, 2024)", U.S. Supreme Court.

The assassination scenario the majority called fearmongering was not invented by the dissent; it was put to the President's own counsel, in open court, twice. At the D.C. Circuit, Judge Florence Pan asked whether a President could order SEAL Team Six to assassinate a political rival and escape prosecution; counsel answered that he could, unless first impeached and convicted by the Senate. At the Supreme Court, Justice Sotomayor pressed the same hypothetical, and counsel allowed that ordering such a killing "could well be an official act." Cited to establish one narrow but decisive point: the abyss the doctrine leaves open was named aloud, to the people defending the doctrine, before the Court wrote it — so the breadth of the ruling cannot be waved off as a dissent's paranoid hypothetical.