It has created bitter division among both politicians and public between supporters of different possible resolutions to the question, including threats of violence against some MPs. It has caused deadlock in the House of Commons, as a result of which Britain’s worldwide reputation as a serious civilised country has been trashed. It has caused exasperation among our European partner countries as our government has twice had to beg for an extension to the deadline for leaving the EU, which was rashly fixed much too early by people who believed the irresponsible campaign claims that disentangling a 43-year process of economic and political integration would be quick and easy.

Threats of violence are not to be brushed off lightly after the murder of Jo Cox MP a week before the 2016 referendum, and the murder of Swedish MP Anna Lindh shortly before a Swedish referendum on joining the Euro in 2003.

So why do we have referendums at all?

On any given question, there are broadly three possibilities:

The public are mostly in favour;
The public are mostly against;
The public are divided, with no clear majority one way or the other.

The political class will already know which of these is the case. There are opinion polls, MPs receive letters from constituents, and politicians or media companies can hold debates in the form of focus groups (citizens’ juries or citizens’ panels).

If the public is generally for or generally against some policy, then that will be clear from the outset and there will be no need for a referendum to tell you what you already know.

There is only a case for a referendum therefore when the public is more finely divided on a question. This was indeed true of Britain’s EU membership in 2016.

But there is a paradox here. Referendums are only held on major constitutional questions, and represent a once-in-a-generation decision. (The previous EU referendum had been held in 1975.) If debate and polling before referendum day do not give a clear answer, then the vote will be a close one.

Can one actually decide such a major change on such a tightly contested result? Of all the claims made during the 2016 referendum campaign, the one that stands out in my memory is by Nigel Farage, when he said:

“In a 52–48 referendum this would be unfinished business by a long way. If the Remain campaign win two-thirds to one-third that ends it.”

Clearly, that remark was prophetic, even if not in quite the sense that Mr Farage originally intended.

The referendum paradox is therefore this: if there is a large majority one way or the other then that is already known beforehand; but if the majority is small and unpredictable then the precise referendum result obtained is indecisive.

That is, it is indecisive in a democracy governed by principles of fairness and hence of compromise wherever possible. If by “democracy” we mean a situation in which a narrow majority can trample roughshod over the wishes of a substantial minority, then a referendum won by a single vote (or by a single vote in 25, as in 2016) can be deemed to have decided the question.

Of course in a general election the choice of MP can be decided by an arbitrarily small majority, even a single vote. Unlike the referendum case, this is acceptable, because only one of 650 MPs is being chosen, and because the vote will be re-run in no more than five years time. The stakes are much less than in a nationwide and once-in-a-generation referendum.

I believe that where opinion on a referendum question is finely balanced between two options, an enlightened democracy must take account of the minority view as well as the majority one. But that situation will also be known in advance, and therefore there is still no need to hold a referendum.

Take the options for the UK as regards the EU: they are basically:

Hard membership (within Eurozone and Schengen area);
Soft membership (outside Eurozone and Schengen area);
Soft Brexit (within Customs Union and Single Market);
Hard Brexit (outside Customs Union and Single Market).

Since it was already known before the 2016 referendum that public opinion was divided close to half and half, it should therefore already have been clear that the only choices consistent with what the public as a whole would accept, ignoring extreme views either way, were Soft Membership (our actual position) and Soft Brexit. Was it really necessary to hold a referendum to decide between these two?

Of course what actually happened was that the government of Theresa May opted for Hard Brexit, thus totally ignoring the Remain vote, particularly the majority Remain votes in Scotland and Northern Ireland. In her recent statements one can hear her blithe assumption that she is speaking for the whole UK public, when clearly she is not.

Fortunately, her autocratic tendencies have been constrained by her political incompetence.

So where does the UK go from here?

My preference all along has been to accept that the 2016 referendum result constrains the UK to Soft Membership or Soft Brexit. Since we already have one of those, and since any form of Brexit is clearly both economically and politically damaging, if I had been prime minister in late 2016 I would not have initiated any kind of Brexit process.

I would, however, have been obliged to review Britain’s place in the EU and to strengthen our position against further integration, for example by ensuring that our place outside the Eurozone and the Schengen area was secure and sustainable, and that our collective defence would continue to be channelled through NATO rather than diverted to a hypothetical EU military as has sometimes been mooted.

But the mythology has grown up that the 2016 referendum produced a decisively large majority in favour of Brexit, that this constitutes a democratic imperative which must be obeyed (trumping the sovereignty of Parliament), that Soft Brexit is no Brexit at all, and that the government must therefore enact Hard Brexit, or even Ultra-Hard Brexit (leaving the EU without a proper agreement so as to cause maximum economic dislocation).

Parliament has nicely foiled this simplistic narrative by voting down Mrs May’s Withdrawal Agreement on three occasions, and by unprecedentedly large majorities, leading to the deadlock we’re in today.

In this situation, the only way out that preserves some shred of legitimacy is to have another referendum, this time offering the public a specific withdrawal deal (presumably the one that Mrs May has been unsuccessful in getting through Parliament) as well as a chance to reconsider the whole idea and maybe stay as an EU member after all.

Remain supporters will be anxious to turn a slim and indecisive Leave victory into a slim and indecisive Remain one. Leave supporters will be anxious to show that they can produce an even larger Leave vote after all that’s happened in the last three years. So it seems to me to be a win for both sides.

Especially as the likely result – a 50/50 split with the odd couple of percent producing a tiny majority one way or the other – will probably achieve little more than keeping the whole Brexit show on the road for another three years, in accordance with the Farage doctrine, until the EU gets so fed up with British stupidity and disruptiveness that they kick us out altogether.

That’s why I’m not in favour of referendums in our democracy.