This is a much more complicated question than it might seem. That the ban treaty should prohibit assembled nuclear warheads seems clear. But what about fissile material? If so, what is fissile material? And what about delivery platforms? Clearly, the ban treaty is not going to ban bombers, but what about ICBMs, SLBMs, or cruise missiles? What about missile defence? Also, is there a right to enrich? If so, to which level may uranium be enriched? In short: how far up the ladder of nuclearity would parties be allowed to climb

The best option, it seems to me, is to duplicate the NPT on this matter (which famously does not define nuclear weapons). Under the NPT, what is and is not OK? Is largely a political question settled by the IAEA board of governors and NPT review conferences. There is no doubt that this method is unfair. For example, there is (for good reasons) little doubt that Japan gets away with more than Iran. But the alternative of negotiating precise definitions and cut-off points could potentially open a can of worms (not to mention last for decades). A ban treaty should make clear that all states are expected to adhere to the same standards as non-nuclear-weapon states under the NPT (including, obviously, to accept IAEA safeguards).

The ban treaty should also provide for a treaty review process through which the treaty may be amended (e.g. by a positive vote of two thirds of the ban treaty membership). That way, more specific provisions or definitions could be added if needed.

MODEL A: DISARM THEN JOIN

The simplest solution would be to include a provision that any state wishing to accede must be able to demonstrate for example to the satisfaction of the IAEA or another suitable international body that it does not possess nuclear weapons. If this model were adopted, the ban treaty would be to today’s nuclear-armed states what the NPT was to South Africa before South Africa’s accession to the NPT in 1991. This model would allow the nuclear-armed states to complete the process of disarmament however they like unilateral, bilateral, or multilateral and under whichever verification mechanisms they prefer. Although opponents of disarmament have done their best to cloud the issue, a ban treaty is in no way incompatible with a step-by-step approach to disarmament. The whole point of the ban treaty is to separate the (normative) act of prohibition from the (material) process of disarmament.

MODEL B: JOIN THEN DISARM

An alternative model would be to allow nuclear-armed states to become parties on the condition that they committed themselves to verifiable disarmament within a specified period of time. But how would such a model work in practice? Would the nuclear-armed states negotiate stockpile destruction multilaterally with the entire ban treaty membership? If so, who would pay for these negotiations, the UN? Or would the nuclear-armed states negotiate stockpile destruction amongst themselves in a parallel process? What if this process ran aground? Could the nuclear-armed states be trusted to make the leap unilaterally and without verification mechanisms in place? If not, who should verify the disarmament process? What would happen if one or more nuclear-armed states failed to complete stockpile destruction within the specified period of time?

The least impractical way in which this model could work would be for the nuclear-armed states to join the ban treaty after having agreed to a plan for physical destruction of weapons (and probably weapon platforms) amongst themselves first.

To its credit, the ban treaty movement has shone a light on the role of a group of states often referred to euphemistically as umbrella states in perpetuating the nuclear status quo. These are states that, through membership in a nuclear alliance, rely on so-called extended nuclear deterrence in their security strategies. It is obvious to most that these states acquiescence and outright support for potential nuclear violence perpetuates existing policies and power structures. If it is to be effective, then, the ban treaty must force these states to change their ways. At the same time, the ban treaty cannot and should not prohibit alliances (that include nuclear-armed states). The challenge, then, is to allow states in nuclear alliances to join the treaty while at the same time forcing them to disavow potential nuclear violence completely.

But how, precisely, is this to be done? After all, the nature of the nuclear alliance states relationship to their nuclear patronvaries greatly. In the case of NATO, the nuclear aspect of the alliance is explicit, and all members bar France participate in the so-called Nuclear Planning Group. But in the case of Australia, for example, the nuclear aspect of that country’s alliance with the United States has not been formalised in any official document.

At a minimum, the ban treaty must prohibit nuclear alliance states from participating actively in nuclear war planning (e.g. the Nuclear Planning Group in NATO), military exercises that foresee the use of nuclear weapons by the allies in question, and from soliciting the use of nuclear weapons in their name. In addition, one might put in the preamble that the signatories unequivocally reject any use of nuclear weapons, including in their own defence.

North Korea’s withdrawal from the NPT prompted many to argue that the grounds for withdrawal should be sharpened. There is good reason to agree. A ban treaty that prohibited withdrawal would also seem to fit with the ethos of the humanitarian initiative and its unequivocal rejection of nuclear violence under any circumstances. However, a ban treaty that prohibits withdrawal entirely might struggle to gain the support of several influential non-nuclear-weapon states.

Most arms control and disarmament treaties stipulate that withdrawal shall not take effect until a certain period of time after notification of intention to withdraw. Certain treaties, such as the Cluster Munition Convention, further stipulate that if, on the expiry of the period in question, the withdrawing party is engaged in an armed conflict, withdrawal shall not take effect before the end of the armed conflict. This might not be a viable model for nuclear weapons, however, as it would seem to add to the incentive to wage war on the withdrawing party.