Any Other European State

Section 1The question before us

The North Atlantic Treaty was signed in 1949, and it says that members may invite the accession of "any other European state." Those five words are why the most successful military alliance in history cannot admit Japan, South Korea, Australia or Taiwan — and why an attack on Tokyo falls outside the guarantee that has kept Europe at peace for seventy-six years.

Nobody made that decision. It was a description of the world in 1949 that hardened into a jurisdictional limit, and the countries it excludes are precisely the ones the next century will be decided among.

We have an alliance that works and cannot expand, and a universal body that can admit anyone and requires nothing.

This paper proposes a league of free nations: a defensive alliance and trade bloc open to countries meeting a stated constitutional standard, inheriting NATO's command structure and personnel while leaving its 1949 treaty behind.

1.2   Our own share of the failure

For eighty years the United States told the world that elections are not the test — that what distinguishes a free country is constraint on power, an independent judiciary, protection of minorities against majorities, and a press the government cannot silence.

We are now being measured against that standard by the instruments we encouraged, and the reading is bad.

The Varieties of Democracy project, based at the University of Gothenburg, assesses 179 countries using 3,500 country experts. Its Liberal Democracy Index is built from the criteria Americans have always said were the important ones: legislative and judicial constraints on the executive, rule of law, individual liberties, freedom of expression.

In 2026 the United States fell from 20th to 51st on that index in a single year — from 0.75 to 0.57. It was the largest single-year decline in the history of the dataset, roughly three times the second-largest, and the only comparable drops in the modern era occurred in countries experiencing military coups. The indicators driving it were executive oversight, legislative investigation of the executive, government censorship effort, and academic freedom. The American score is now below every other member of the G7.4

A reader may dispute those measurements, and they are expert surveys rather than physical constants. We report them because this paper proposes a membership standard, and a standard that only ever judges other people is not a standard. It is a club with a flattering name.

We would rather propose a test that could exclude its own author than one designed to exclude only the people we already dislike. That is the entire difference between moral authority and self-congratulation, and it is the thing the United Nations lacks that we would be trying to supply.

Section 2Elections are not the test

The word "democracy" has stopped carrying information. Russia holds elections. Hungary holds elections. Venezuela holds elections. Citizens of countries excluded from Western alliances reasonably ask what disqualifies them when the ballot boxes are real and the votes are counted.

The answer exists and we have been too vague about it for thirty years, which is our fault rather than theirs.

Fareed Zakaria named it in Foreign Affairs in 1997: the rise of illiberal democracy, in which elected governments routinely ignore constitutional limits and deprive citizens of basic rights.5 Political scientists now measure the distinction directly. V-Dem separates electoral democracy — meaningful, free, multi-party elections — from liberal democracy, which requires in addition that citizens have individual and minority rights, are equal before the law, and that the executive is constrained by the legislature and the courts.4

An election decides who governs. A constitution decides what they may not do. The second is the one that protects you, and it is the one we stopped naming.

So the standard this paper proposes is not democracy. It is liberal constitutional democracy: free elections with broad participation, entrenched individual rights a majority cannot vote away, an independent judiciary, legislative constraint on the executive, protection of minorities, a free press, and a secular state that does not subordinate law to creed.

Stating it precisely is also the courteous thing to do. A country told it is "not democratic enough" has been insulted without being informed. A country shown a published index and the specific sub-scores on which it falls short has been given a map.

Section 3Why this particular club is worth joining

The empirical basis is unusually strong for this field.

Since the 1980s researchers have tested the proposition extensively against historical data. The results have been consistent: democracies very rarely go to war with each other, and the finding holds across different time periods, geographic regions, and varying definitions of both "democracy" and "war." Jack Levy called it "as close as anything we have to an empirical law in international relations."1

Michael Doyle, who has spent forty years on the question, put the precise version in 2024: liberal states may not act peaceably toward everyone, but they act peaceably toward one another. There are no clear-cut cases of one democracy going to war against another, nor do any seem forthcoming.1

That second clause matters and we will not hide it. This is not a claim that free countries are peaceful. They fight non-democracies frequently, including in wars that were mistakes. The claim is narrower and more useful: an alliance among liberal democracies is an alliance whose members will not fight each other, which is the foundation on which everything else — integrated command, open trade, shared intelligence — can be built without hedging.

The honest caveats. Critics argue democracies have not existed long enough to establish the pattern, and that results depend on how "democracy" is defined — which is exactly why Section 2 insists on a published, transparent measure.1

Section 4We built this once, and it worked

Nothing proposed here is novel. It is a restoration.

Between 1944 and 1949 the United States built an economic and military order among free countries: Bretton Woods, the General Agreement on Tariffs and Trade, and the North Atlantic Treaty. The result was the longest great-power peace in modern history and the largest reduction in poverty ever recorded.

What decayed was not the machinery but the standard. The institutions became universal — admitting everyone, requiring nothing — and the requirement quietly disappeared.

The United Nations is the clearest case, and we should be precise about the criticism because the usual version is wrong. The UN does have a declaration of rights: the Universal Declaration of Human Rights, adopted in 1948, drafted under Eleanor Roosevelt's chairmanship, which with the two 1966 covenants forms the International Bill of Human Rights.6

The problem is that it is a declaration rather than a treaty, that nothing in the voting structure privileges the countries which honor it, and that the Security Council veto belongs to whoever won a war in 1945. The General Assembly gives one vote to each state regardless of whether its government represents anybody.

The financial institutions have the opposite defect. Voting at the IMF and World Bank is weighted by financial contribution: the United States holds 16.49 percent at the Fund and 16.07 percent at the Bank, and major decisions require 85 percent — making fifteen percent a veto, which only America possesses.7 That is one dollar, one vote. It weights nothing toward freedom either.

And the G7 and G20 are not institutions at all. Neither has a charter, a secretariat, a permanent staff or a budget. The presidency rotates, the host pays, and the output is a communiqué that binds nobody. Their value is convening. They cannot be the vehicle for anything that requires an obligation.

Section 5Reform I: The standard, published and applied to everyone

Membership should require liberal constitutional democracy as defined in Section 2, assessed against a published, independent index with transparent methodology, applied identically to founders and applicants.

Outsourcing the judgment is the point. The question that has sunk every previous proposal is who decides, and the answer cannot be a foreign ministry. It can be an academic instrument with a published method, peer review, thousands of coders, and scores that anyone may check — with a threshold set in the charter rather than negotiated case by case.

Two consequences we accept in advance.

It will exclude countries we want. India is the world's largest democracy, the essential counterweight to China, and is not currently classified as a liberal democracy on these measures. Turkey and Hungary are NATO members today. A standard that bends for strategic convenience on its first application is not a standard, and everyone watching will understand that immediately.

And it may eventually exclude us. Section 1.2 gives the current reading. We would rather write that sentence ourselves than have it written for us.

Section 6Reform II: Four missions, and the fourth one narrowed

One. Collective defense of members. An armed attack against one is an attack against all — the Article 5 formula, unchanged, because it works.

Two. Free trade among members. Countries with comparable legal systems, enforceable contracts, independent courts and similar regulatory standards can trade more deeply and with less friction than countries that do not. This is the mission with the fastest returns and the fewest objections.

Three. Influence toward human rights, free markets and constitutional government — by demonstration, by the tangible benefits of membership, and by a published standard that tells any country exactly what qualification requires. Not by force.

And four, narrowly: the alliance does not guarantee world order.

This is the correction that matters most, because the opposite promise has been made before and it destroyed the organization that made it.

6.1   What happened to the last league

The League of Nations promised collective security against aggression anywhere. When Japan seized Manchuria in 1931 and Italy invaded Abyssinia in 1935, no member would fight for another member's interest. The United States never joined at all.

The failure was not the principle. It was that a commitment to prevent every invasion everywhere is a commitment to wars members will not fight — and an alliance discovered to be bluffing is worse than no alliance, because adversaries then discount the guarantees that were real.

NATO survived where the League died because Article 5 is narrow: an armed attack against a member. Seventy-six years of credibility rest on that limit, and it has been invoked exactly once, after September 11th.

6.2   And a member's own war is its own

The alliance defends members who are attacked. It does not join wars members choose to start.

This is already the law and already the practice, which is worth knowing. Article 5 triggers on an armed attack against a member, not on a member's offensive operation. In 2003 Article 5 was never invoked for the invasion of Iraq, and France and Germany declined to participate while remaining allies in good standing.

We would put it in the charter explicitly rather than leave it to practice, and add the corollary: a member that attacks another state without being attacked is subject to suspension. An alliance of free nations that cannot sanction aggression by its own members is not describing a principle. It is describing a faction.

Section 7Reform III: The voting rule, which is the whole thing

Decisions by qualified majority, not consensus. Expulsion or suspension by supermajority with the accused excluded — never by unanimity. And automatic suspension on bright-line triggers.

Two organizations have run this experiment and both results are instructive.

7.1   The European lesson: never set the bar at unanimity

The European Union's Article 7 permits suspension of a member's rights for serious breach of its founding values. It was triggered against Hungary in September 2018.

Eight hearings have been held, the most recent in May 2025. The Council has never moved to a vote.3

The reason is arithmetic. The decisive determination requires unanimity among the other members — so for years Hungary and Poland shielded one another and the threshold could never be reached. The Polish procedure ended in 2024 not because the mechanism worked but because Poland changed government. The EU eventually gave up and used a budget instrument instead, withholding funds under a rule-of-law conditionality regulation.3

Money worked where the constitutional provision did not, because the constitutional provision required the consent of the people it was written to restrain.

Two backsliders defeat unanimity permanently. The threshold must be a supermajority — two-thirds or three-quarters, with the accused state excluded from the count.

7.2   And NATO's consensus rule is the same trap

Every NATO decision is taken by consensus. As one legal summary puts it: "There is no vote; tacit agreement counts as consensus." Each member effectively holds a veto on everything.

The consequence was demonstrated recently. Finland and Sweden applied in May 2022 after the invasion of Ukraine. Turkey blocked ratification over grievances concerning Kurdish organizations and the Gülen movement; Finland acceded in April 2023 and Sweden not until March 2024, after Hungary also relented. And the legal observation that should govern our drafting: the treaty contains no procedure for a member state linking its approval to conditions lying outside the treaty entirely.2

7.3   Bright lines beat judgment calls

Assessing whether a country has become gradually less liberal is contestable forever. Assessing whether it has just had a coup, or just invaded a neighbor, is not.

The African Union suspends members after coups as a matter of routine, and it is the most functional suspension practice anywhere — because the trigger requires no judgment about degree. The Council of Europe expelled Russia in 2022, the first expulsion in its history, when the breach was unambiguous.

So the charter should carry both: automatic suspension on unambiguous triggers — a coup, the cancellation of an election, an unprovoked attack on another state — and a slower supermajority process for gradual backsliding, measured against the published index.

Section 8Reform IV: Inherit the machinery, not the treaty

This is the practical core, and it follows from a legal obstacle that cannot be negotiated away.

Article 10 of the North Atlantic Treaty permits accession only by "any other European state," and only by unanimous agreement. Japan, South Korea, Australia, New Zealand and Taiwan are ineligible by the text. Article 6 compounds it by defining the geographic scope of Article 5 itself — Europe, North America, Turkey, and islands north of the Tropic of Cancer. An attack on Tokyo falls outside the guarantee even if Japan were admitted.

Amending either requires the unanimous consent of all thirty-two members, including the members whose conduct a new standard would judge. That is the Article 7 trap in a harder form: the veto belongs to exactly the parties with reason to use it.

But the treaty is not what makes NATO work.

What makes NATO work is the integrated command structure, the interoperability standards, seventy-six years of joint exercise and planning, and a professional staff that knows how to move armies across borders. None of that lives in the 1949 text. A new instrument can designate the existing command as its own and carry the personnel across, while the North Atlantic Treaty continues to exist for its original members.

That is how Europe actually built itself — Coal and Steel Community, then the European Economic Community, then the Union, each a new treaty inheriting the institutions of the last.

8.1   And the nucleus already meets

NATO maintains a formal relationship with Japan, South Korea, Australia and New Zealand — its Indo-Pacific partners. They attend summits. They coordinate. What they lack is the guarantee.

That existing group is the natural starting membership: countries already at the table, already interoperable, already aligned, lacking only the instrument. Deepening that partnership requires nobody's permission, and it can proceed while the treaty work is done.

We propose the sequence deliberately. Deepen the partnership until it has substance. Give it its own treaty with better voting rules. Let it inherit the command. Slower than amending NATO, and it does not require the consent of Budapest or Ankara.

Section 10What we are not claiming

We are not claiming free countries are peaceful. They fight non-democracies frequently, sometimes wrongly. The claim is that they do not fight each other.

We are not proposing to spread liberty by force. Section 6 limits the alliance to defense of members and explicitly subjects a member who attacks another state to suspension.

We are not proposing to abolish the United Nations. Universal membership has value precisely because it is universal, and a body that excludes your adversaries cannot negotiate with them.

And we are not claiming this is easy or likely. It requires a new treaty, ratified by a Senate that has not ratified a major one in decades, among countries with their own politics, in an interval when the United States is not the natural author of a standard about constitutional constraint. We propose it because the alternative is drift, and because the thing being proposed is a restoration of what America built when it was at its most serious.

Section 11The argument you can carry

The paper compressed to what a person can remember and repeat.

And the one that indicts the argument's own side. For eighty years America told the world that elections are not the test — that what distinguishes a free country is constraint on power, an independent judiciary, and a press the government cannot silence. We are now being measured against that standard by the instruments we encouraged, and our score is below every other member of the G7.

Section 12Conclusion

In 1949 twelve countries signed a treaty saying that an attack on one was an attack on all. They wrote it about the North Atlantic because that was where the danger was, and they limited membership to European states because that was where the members were.

Seventy-six years later it has been invoked once, Europe has not had a great-power war, and the sentence is still true. It is also still about the North Atlantic, in a century whose decisive questions will be settled in the Pacific by countries the treaty cannot admit.

Meanwhile the universal institutions have arrived at the opposite failure. The United Nations will admit anyone and requires nothing. The financial bodies count dollars. The G7 and G20 have no charter, no staff and no budget, and issue press releases. None of them is capable of an obligation.

So we have an alliance that works and cannot grow, and a forum that includes everyone and asks for nothing. The answer is not to abolish either. It is to build the thing that is missing — using the command structure we already own, a standard published in advance, and a voting rule designed by people who have read what happened to the last two attempts.

One last thing, and it is the part that makes this serious rather than self-flattering. The standard has to be one we could fail. We are closer to failing it than we have been in fifty years, which is either a reason not to propose it or the best possible reason to — because a country that writes a rule it might be caught by is making a different kind of promise than a country that writes one for other people.