Everybody, it seems, wants a debate. Theresa May wants to debate her Brexit deal with Corbyn; Grace Blakeley invites her critics to “come debate me”; and Sarah O’Connor wants a “nuanced debate now about what we value in the economy”.

Include me out. Most debates – and especially televised ones between politicians – are worse than a waste of time. They are downright pernicious.

A debate, like many things, is a selection device. The question is: does it select for the truth, or against it?

There’s a presumption – best expressed by John Stuart Mill – that it selects for the truth. Honest and rational men, he thought, would present arguments and evidence to other rational men and so the truth would emerge:

There must be discussion, to show how experience is to be interpreted. Wrong opinions and practices gradually yield to fact and argument: but facts and arguments, to produce any effect on the mind, must be brought before it.

This is woefully over-optimistic. It’s very possible instead that debates are counter-selection devices. Far from selecting in favour of the truth, they select against it. I’m thinking of four mechanisms here:

 - Some facts are hard to present in debates. Numerical data can be cherry-picked or presented out of context, with the result that the debate degenerates into claim and counter-claim. The truth about numbers is often best shown in chart or tabular form, not by talking heads. In this way, “hard” data doesn’t get the weight it merits in debates. But not just hard data. As Michael Polanyi pointed out, there can be tacit knowledge – senses of comfort or disquiet that cannot be properly articulated. This too is underweighted in debate. Debates select for articulable arguments, which means selection against both hard numbers and tacit feelings.

 - Debates select not just for truth but for overconfidence. The confident assertion of a clear statement beats caution and caveats. Experiments tell us that people often mistake overconfidence for competence thereby selecting for it and against actual ability. Debates favour articulate overconfident posh folk who in fact know nothing – which is why we got into this mess.

 - Lies can win. Mill assumed that interlocutors would present facts. But this isn’t necessarily true. As Tim Harford has said, “a simple untruth can beat off a complicated set of facts simply by being easier to understand and remember.” As we saw in the Brexit debate, with claims that Turkey was about to join the EU and that leaving would give the NHS £350m a week, bare-faced lied can gain credence. Selection for dishonesty isn’t necessarily a problem if we are choosing our next Prime Minister: I’m happy to have a lying bastard deal with Putin and Trump. But it is if we want to ascertain the truth of an issue.

 - Debaters appeal not to reason but to cognitive biases. Mill assumed, implicitly, that audiences were good Bayesians who would yield in the correct fashion to “fact and argument”. We now know that they are not, that all of us (yes, even me) are a mass of cognitive biases. Debates can be won not by the side with the most rational arguments and strongest facts, but by those who best appeal to our irrationality. This, I suspect, was the genius of the Brexit campaign: it harnessed such biases as wishful thinking and prospect theory whilst remainers were mired in abstract facts about GDP. Cam_debate512

On top of these adverse selection mechanisms, debates have another baleful effect. They are polarizing. If we ask “who’s right?” we cannot build consensus. This problem is magnified by a phenomenon first pointed out in 1979 by Charles Lord, Lee Ross and Mark Lepper. They showed people mixed reports on the effects of the death penalty. They found (pdf) that, after reading these reports people who supported capital punishment became stronger in their support, whilst opponents of the death penalty also became more dogmatic. Even balanced and factual data thus produce polarization and extremism, through mechanisms we now know as asymmetric Bayesianism and the backfire effect.

As evidence for all this I need only point to Brexit. We’ve now had three years of debate about it. The result is ignorance and division. Or just watch Question Time, or any other of the endless "posh twats talk shit" shows that pollute the airwaves. 

I’m not saying that there is no point to political discussion. Instead, such discussions, if they are to be useful, must be carefully structured. There are strict rules on how court cases should be conducted and what evidence is admissible. The same should be true for political deliberation.

In its current form, political debate perpetuates the idea that the road to power lies in the ability to rationally persuade people. This is a lie.