“TO ask the hard question is simple”, according to WH Auden. Whether you agree may depend on your view of Auden: Joseph Brodsky called him “the greatest mind of the 20th century”, while Hugh MacDiarmid thought him “a complete wash-out”. But is it really simple to ask the hard question?

It may be less contentious to skip to a later bit in the poem, where he claims: “But the answer / Is hard and hard to remember”.

All the same, I dimly recall that the answer to the question of whether the UK should leave or remain in the EU was “Leave”, something we haven’t yet done. It might then seem surprising that a lot of people (50 per cent, to be precise) would now, according to a new Sky Data poll, like a referendum on “the deal the Government suggests, no deal or staying in the EU”.

But it’s not the most surprising finding. Nor is the very slight majority (51 per cent) who now think Brexit will be bad for the country overall.

The astonishing finding is that 10 per cent of respondents to the poll think that the Government is doing a good job in the negotiations and a mere 78 per cent think that Theresa May is making the biggest meal of it in the history of dogs or dinners.

The fact that the latter figure is not 100 per cent is almost enough to make anyone, Leaver or Remainer, doubt the wisdom of ever asking the people anything in any kind of poll or referendum.

There are other reasons, however, for not wanting what its advocates like to call a “People’s Vote” – which raises the question of who cast 33,577,342 votes just 25 months ago if it wasn’t people.

The first reason is just that. Why heed anyone who wants a second vote if, by the very fact of calling for one, they dismiss the answer the public has already given? It’s not as if the referendum has been implemented and, some way down the line, having changed their minds, people want another one. It took the UK more than 40 years to do that the last time, after all.

This is more like the tactic the EU has previously used when it doesn’t like the result of a vote and – as happened in Denmark over the Maastricht Treaty and (twice) in Ireland, over the Treaties of Nice and then Lisbon – has another to get the correct answer. I had thought it had given that up; in 2011, it just imposed a new government in Italy, and in Greece in 2015 it ignored the vote altogether. But apparently people such as Lord Mandelson, a former EU commissioner, think it’s worth reviving the technique.

He’s wrong not just because of the principle, but the practicalities of attempting a blatant wrecking tactic. The Remain camp seem not to have considered that a rerun of the vote cannot, in any circumstances, produce a good outcome.

A binary Leave/Remain question might – indeed, such would be the resentment among many voters at another referendum, probably would – produce the same result. Which would improve matters not one bit.

Or it might, almost certainly by an incredibly narrow margin, reverse the original vote. You’d then have two incompatible results of equal democratic value, and a recipe for fury amongst half the population who would feel (entirely justifiably) that they had been betrayed.

Of course, lots of Remainers are furious, but thems the breaks when you lose. No doubt, had the vote gone the other way and the EU was now planning some scheme which the Remain campaign assured us it wasn’t planning, such as the EU defence integration Jean-Claude Juncker is in fact planning, Nigel Farage would be jumping up and down asking for another referendum. But most sane people wouldn’t.

The impossibility of getting a better answer is one practical objection to a straight rerun. The only binary option that makes any sense at all is the Government deal versus World Trade Organisation terms.

There are even greater practical difficulties with the suggestion, originally from the former Conservative minister and Remain campaigner Justine Greening, of a vote on three options: whatever plan the Government and the EU eventually agree (if they do), a no- deal Brexit on WTO terms, or revoking Article 50 and remaining after all.

This is popular with Remainers, because they think it’s the one where they would get a majority, or at least the largest chunk of the electorate, by splitting the Leave vote. But it’s unworkable, because a referendum more or less has to have a binary question.

What if you think Mrs May’s plan is better than Remain, but Remain is better than WTO? Or – which, despite the scaremongering over “no deal”, seems to be a more popular position – that WTO is better than Remain, but Remain better than anything Mrs May comes up with (a reasonable calculation, based on her performance so far)?

You may remember the devolution referendum presented a similar potential problem. It was in practice impossible to vote “No, Yes”, though it was (as a very few people, such as Tam Dalyell, argued) perfectly logical to vote against a Scottish assembly, but to think that, if there were one, it ought to have tax-raising powers.

Perhaps in that instance it didn’t much matter, but our departure from the EU is different, because it’s not confined to those three options, anyway. If we could get rid of the Prime Minister and get someone halfway smart, there might be innumerable others.

The EEA and Efta remain theoretical, and quite practical, options. The EU has already offered a variant of the so-called Canada ++ deal. Given that previous predictions of huge unemployment, wholesale departure of foreign firms and collapse of growth failed to materialise, some people think WTO terms (on which most of the world trades already) might not be as scary as Remainers claim.

I’m inclined to think they are, but only because the Government hasn’t planned for them. Any moderately competent PM could make them attractive; for example, simply by unilaterally abolishing tariffs and non-tariff barriers on imports, which is all that matters to ensure cheap goods and secure supplies of medicine, just-in-time manufacturing lines, and food.

Alas, we’ve got simple Theresa May instead. Yet, wrong though all her proposed answers have been, she’s still right that asking a hard question again would be not only pointless but damaging.