The vacuity of the debate that preceded the referendum on EU membership is exceeded only by the emptiness of result that the referendum subsequently yielded. A slim majority of those who voted may have expressed a desire to “leave”, but what that means is such an open question as to render the referendum outcome largely meaningless. If ever the aphorism that “the devil is in the detail” were appropriate to a situation, then Brexit is it. One of the fundamental flaws in the UK’s referendum on EU membership is that while those who voted “remain” could be reasonably certain of what they were voting for — namely, the existing terms of EU membership as adjusted by David Cameron’s now-irrelevant “deal” — those who voted “leave” had the opportunity only to vote against the status quo. They could not have been expressing, and did not express, any clear view about what the UK’s future relationship with the EU should look like precisely because no vision of that relationship was on the table. (Nor is it now.) The use of a referendum in such circumstances was as rash as it was inappropriate.

Such considerations form part of the background against which the debate about Parliament’s role in relation to Brexit is currently playing out. I have argued elsewhere that there is no legal need for parliamentary involvement in relation to the triggering of Article 50 — a view that I continue to hold, notwithstanding the criticism that others have levelled at it. (Paul Bowen, for instance, is highly critical of my analysis, although in fact he appears to misunderstand my argument. My view is not, as he asserts, that “the [European Communities Act 1972] itself authorises the UK’s withdrawal from the Treaties”. Rather, my argument is that the royal prerogative authorises the triggering of Article 50, and that the Act does not displace the prerogative in this context.)

However, even if, as I contend, parliamentary involvement is legally unnecessary, it is arguably strongly politically desirable. Others — including Kenneth Armstrong, Colm O’Cinneide and Alison Young — have ably articulated normative arguments in favour of parliamentary involvement, and I will make no attempt here to summarise them. It seems to me, however, that one of the most persuasive arguments is that the referendum, for reasons indicated above, leaves in its wake not only the tatters of several political careers but a huge amount of unfinished — indeed, barely begun — business. In such circumstances, the case in favour of involving Parliament — our primary democratic institution — in shaping the direction of the UK’s EU policy is compelling.

There is, however, a further question — which is whether a second referendum should take place. I do not have in mind a re-run of the first referendum. Rather, I have in mind a referendum in which people are asked to vote on specific proposals concerning the UK’s future relationship with the EU, be that as a member of the European Economic Area or otherwise. There are good arguments against, one of which is the reverse side of the arguments in favour of parliamentary involvement. The UK is, at root, a parliamentary democracy in which members of Parliament make decisions on behalf of those who they represent. Referendums are a departure from, and an aberrant viewed through the lens, of that democratic model. On the other hand, for those who celebrated the first referendum as an opportunity for “the people” to have their say, the logic in favour of a second referendum is hard to resist. If the people are meaningfully to have their say, there has to be something, in the first place, for them to have a say about — and until there are specific proposals concerning the UK’s future relationship with the UK, that condition will be not satisfied.

Whether one favours a second referendum or parliamentary involvement alone turns in part on the extent to which one favours direct over representative democracy. For my own part, I consider the arguments in favour adhering to the UK’s traditional model of parliamentary democracy to be the more persuasive ones. But the question is not just whether it would be desirable to hold a second referendum. There is also the question whether there must, as a matter of law, be one. In an article on the In Facts website, Pavlos Eleftheriadis argues that a second referendum is “probably legally required”.

Central to Eleftheriadis’s argument is section 2 of the European Union Act 2011. Its effect is that a treaty that “amends or replaces” the Treaty on European Union or the Treaty on the Functioning of the European Union cannot be ratified unless one or other of two conditions is fulfilled. The first condition is that the treaty in question is “exempt” from the requirements of the Act. The second condition is that a referendum on the treaty is held and ratification of the treaty supported by a majority of those voting in the referendum.

The key passage in Eleftheriadis’s article is as follows:

Because the Act also covers what it refers to as ‘replacing’ treaties, it encompasses any treaties that may be concluded between the EU and the UK as a third party, if such a treaty will ‘replace’ the current EU treaties. It is obvious that any withdrawing agreement will replace those Treaties because by virtue of these treaties the UK will go from being a member to being a non-member. As ‘replacing’ treaties, the withdrawing treaty and the new trade agreement are likely to both fall under the scope of the European Union Act 2011.

For three reasons, my view is that the 2011 Act provides no legal guarantee that there must be a second referendum. The first point is as preliminary as it is obvious — namely, that, as part of the Brexit process, the 2011 Act might well be repealed. That might be politically difficult, particularly if it were to leave the Government open to the charge of moving the goalposts by dispensing with a referendum requirement that would otherwise have applied. However, in the light of my second and third reasons, I do not believe that the goalposts erected by the 2011 Act are legally relevant here.

The second reason, then, is that the central concern of the 2011 Act is the transfer of additional powers to the EU. As is clear from the Explanatory Notes that accompanied what was then the European Union Bill, the perceived mischief at which the Act was directed was the possibility of a Government, exercising its treaty-making prerogative, signing the UK up to new EU arrangements that would further tilt the balance of power in favour of the EU at the expense of Member States’ autonomy. In the light of that, very broadly speaking, treaties that do not have such effects are exempt from the referendum requirement in the 2011 Act. It is perfectly clear, therefore, that a treaty straightforwardly providing for the UK’s departure from the EU would be an exempt treaty, because far from involving the augmentation of the EU’s powers in respect of the UK, it would extinguish them.

The same may or may not be true of any further treaty addressing the UK’s future relationship with the EU. Eleftheriadis argues that the referendum requirement in the 2011 Act is “so broad” that “any trade agreement between the EU and the UK is likely to meet it”. He instances a trade agreement authorising the EU “to bring a claim against the UK before a WTO panel or an ad hoc arbitration tribunal”. It is true that such a treaty would likely fall outside the scope of the exemption to the referendum requirement because it would involve, as section 4(1)(j) of the 2011 Act puts it, “the conferring on an EU institution or body of new or extended power to impose sanctions on the United Kingdom”. That, however, leads onto the following, more fundamental, point, which forms my third reason for doubting that there is any legal guarantee or requirement of a second referendum.

That certain elements of a future EU-UK trade agreement might fall outside the exemptions to the referendum requirement under the 2011 Act is, in my view, irrelevant because, in the first place, the Act would not apply to them. A post-EU membership trade agreement entered into by the UK and EU would therefore trigger no requirement to hold a referendum — not because such an agreement would fall within one of the Act’s exemptions, but because such an agreement would not fall within the scope of Act to begin with. That is so because the Act only applies to a treaty which “amends or replaces” the two principal EU Treaties upon which the UK’s membership of the EU is based. The scope of the “amends or replaces” category of treaties upon which the Act bites falls to be understood by reference to the purpose of the  Act. That purpose is inextricably linked to the UK’s being a member of the EU. The Act exists in order to provide a direct democratic check upon the Government’s management — through the exercise of its treaty-making prerogative — of the terms of the UK’s membership of the EU, the idea being that there is a brake — in the form of the referendum requirement — upon any contemplated transfer of competence to the EU. In other words, the Act is concerned exclusively with treaties that shape the UK’s membership of the EU, and has no application to treaties that terminate that membership or which provide for an EU-UK relationship that amounts to something other than membership. A treaty providing for the UK’s departure from the EU is therefore outside the scope of the Act, as is a treaty staking out a relationship between the UK and the EU that does not entail British membership of the EU.

To conclude, the normative case in favour of parliamentary involvement in shaping the UK’s future relationship with the EU is compelling, albeit that such involvement legally need not extend to the enactment of legislation triggering Article 50. The normative case in favour of a second referendum is — given the parliamentary nature of our democracy — weaker, although those who seek to legitimise Brexit through reliance upon the first referendum face an uphill struggle in trying to explain why a second referendum, on a more specific and meaningful question, should not occur. Ultimately, however, this too is a matter that has to be resolved through normative rather than legal argument because, as a matter of law, there is no legal obligation to hold a second referendum.