Alliance by footnote

In 1963, Charles de Gaulle asked a question that Washington never really answered: would an American president actually trade New York for Paris? Would the United States be willing to risk its own cities in the defence of Europe's? He did not wait to find out. France built its own nuclear deterrent, the force de frappe, and three years later pulled out of NATO's integrated military command altogether. The alliance did not collapse. But it caved in, quietly, from the inside, for one member no longer trusted the guarantee but had begun hedging.

That is the part of alliance politics that people tend to overlook. Treaties do not fail in a single dramatic breach. They cannot stand the accumulation of small doubts that each member state develops, which each turn leads to the creation of a plan B, until the plan B is the actual policy and the treaty is just a piece of paper.

A similar thing is stirring right now. President Trump has publicly hedged on Article 5, NATO's founding promise that an attack on one member is an attack on all. It was written to be automatic. No debate, no calculation, no place for the president to make real-time cost considerations. Trump's public doubts about it don't change the treaty. It does something quieter and more corrosive: it turns an automatic promise into a conditional one, and conditional promises are not really guarantees at all.

Robert Jervis, the late Columbia scholar of international relations, argued that complex systems do not absorb shocks but metabolise them. Statesmen are not chess players moving isolated pieces. They are nodes in a system that reads their signals and reorganises around them, often in ways they did not intend. Trump's hedge is just this type of signal. Washington is already reducing its NATO force contributions, including a reduction in the number of fighter, refuelling and maritime reconnaissance aircraft, as part of a defence strategy that pledges "more limited support" to European defence. The message that European capitals are getting isn't subtle: budget for yourselves.

And Europe is following France's path in the 1960s. Once politically taboo, plans for rearmament are now commonplace in Berlin, Warsaw and elsewhere. Joint procurement schemes are picking up the pace. None of this is framed as an exit from NATO. It looks instead like insurance bought quietly against the possibility that the guarantee will not hold when it is tested.

This is the trap of ambiguity in deterrence. In some respects, a guarantee that is not honoured is worse than none, since it is expecting two failures to occur at once. It doesn't fully reassure allies, who already start hedging. And it fails to fully deter adversaries, who read the hedge as an opening. Moscow does not need Trump to withdraw from Article 5. It requires only a bit of doubt to be planted in Brussels and Warsaw, enough to make the math of testing NATO's eastern flank work out.

None of this means NATO collapses tomorrow. Alliances rarely die of a single wound. De Gaulle's France remained a member of NATO until 1966 and participated in NATO summits and deployed troops as necessary. In 2009, however, Paris re-entered the alliance, but it was not the same one that it had previously left. It had discovered the Washington doctrine, which held that the president's words were subject to conditions, and it had followed the doctrine for more than four decades.

Trump may yet reaffirm Article 5 without qualification. But the doubt, once spoken aloud by a sitting American president, does not go away entirely. Allies do not un-hear it. They plan around it. Europe's current rearmament drive is not really about Russia's army. It is about America's ambiguity, and it is the surest sign that the hedge has already been factored in.

The lesson from 1963 is not that alliances end in fire. It is that they end, slowly, in footnotes, the moment members stop assuming the guarantee will hold and start building for the day it does not.