The British Social Attitudes survey reports a fall in the proportion of people saying they voted Leave in the 2016 referendum and a rise in the numbers saying they didn't vote. This is probably due to panel attrition. But it draws our attention to what is certainly a real phenomenon - a tendency to misremember our past preferences in a systematic way.

This was pointed out (pdf) by Gregory Markus back in 1986. He showed that when people were asked about their views on political issues such as gender equality or drug legalization nine years earlier their recall was poor. Their beliefs as they remembered them was much closer to their current ones than was in fact the case. They understated the extent to which they had changed their mind.

What was going on here is a desire for consistency. We want to believe that we are the captains of our soul rather than just a bundle of neurons responding erratically to stimuli and we redescribe our past to make it easier to believe this. There is a line in a lovely Kate Campbell song about a woman who "convinced herself she never needed a man". And when Orwell has the Ministry of Truth claiming that we have " always been at war with Eastasia" he is highlighting the public's willingness to change their perception of history to maintain their self-image of consistency. Giuliana Mazzoni says such false recall is common:

Picking and choosing memories is actually the norm, guided by self-enhancing biases that lead us to rewrite our past so it resembles what we feel and believe now. Inaccurate memories and narratives are necessary, resulting from the need to maintain a positive, up-to-date sense of self.

This echoes Adam Smith:

The opinion which we entertain of our own character depends entirely on our judgments concerning our past conduct. It is so disagreeable to think ill of ourselves, that we often purposely turn away our view from those circumstances which might render that judgment unfavourable.

But we don't just misremember the past to maintain a stable self-image. We also reinterpret the present to avoid telling ourselves we were wrong. Psychologists call this choice-supportive bias. If we buy a car that's slower than we wanted, for example, we congratulate ourselves on its fuel efficiency. Disappointed lovers or voters want to believe they were betrayed rather than that they made the wrong choice. And Brexiters downplay economics and focus on the benefits of sovereignty and need to respect the will of the people. As Robert Cialdini wrote:

Once we have made a choice or taken a stand, we will encounter personal or interpersonal pressures to behave consistently with that commitment. (Influence, p57)

This is one reason we are prone to the sunk cost fallacy: having made a bad choice, we stick with it.

Such self-deception devices mean, as Daniel Kahneman has noted (pdf), that our remembered utilities don't accord with what we experienced at the time. But it's memory rather than reality that shapes our choices. And this, he says, means we have "poor accuracy" in predicting what we'll want in future.

Here again, our attitudes are shaped by an urge for consistency. Just as we wrongly think that our past preferences are like our present ones, so too do we over-estimate the extent to which our future preferences will resemble our present ones. Matthew Rabin call this the projection bias (pdf): we project our current tastes into the future. This is why people pay too much for houses with swimming pools or convertible cars in the summer: they fail to anticipate these will be little use in the winter. It might also explain the well-known tendency for share prices to be too high in the spring and too low in the autumn: investors wrongly project their springtime optimism and their autumnal pessimism into the future. (And fund managers have always been loath to recognise this because they want to see themselves as consistent rational people rather than skittish mood-driven flibbertigibbets.)

All this suggests that we do not have the stable preferences we think we do. Stability and consistency are to some extent illusions which we impose onto our memories, perceptions and expectations. Which poses a question. If an individual's preferences are unstable to a greater degree than we think can we really speak of a stable "will of the people"? Maybe this is an emergent phenomenon - something that's true of groups but not of the individuals that compose the group. Or maybe it's just a fiction.