Sometimes, opinions change a lot even though the facts change only a little. So it seems to be with the case of MPs second jobs. Everybody knew for years that many MPs had them but most regarded it as acceptable. Until, suddenly, it is sleaze.

What's happening here is not so unusual. We often see it in financial markets. In the mid-90s, for example, several Asian countries ran large external deficits. And for a long time, these were seen as healthy because fast-growing countries should suck in foreign investment - until, suddenly, they weren't and we had a financial crisis. As Sushil Wadwani, a former MPC member said (pdf):

Countries with a current account deficit can have a currency that stays strong for a surprisingly long period, until, sometimes, there is an abrupt adjustment.

Similarly, in the late 90s high prices of tech stocks were seen as justified by the prospect of a new economy - until they suddenly were not. And the US housing market started to cool off in 2006, but for months nobody thought this was a problem - until suddenly it was.

A similar thing might be happening to Boris Johnson. Everybody has known for ages that he is a lightweight comic charlatan. He has not changed, nor done anything surprising. It's just that the same person is now provoking different feelings amongst his erstwhile supporters.

To understand what's going on here, we must remember that we rarely see things in isolation as bald facts. Instead, we select and interpret them. As MLJ Abercrombie wrote:

The information that a person gets...depends on the context, or total situation, and on his past experience, which is usefully thought of as being organized into schemata. Human relations play an important role in perception both in that they are often a significant part of the context, and in that they have contributed to the formation, testing and modifying of schemata (The Anatomy of Judgment, p18)

This is why we are so prone to mis-readings, typos and optical illusions: squares A and B in my picture are in fact the same shade.

Facts, then, come within frames (pdf). And as Daniel Kahneman wrote in Thinking Fast and Slow, "different ways of presenting the same information often evoke different emotions." He gives the example of how even trained doctors react differently to the statement "the odds of survival are 90%" than to "the mortality rate is 10%" even though they are different frames of the same facts. And, he adds, "your moral feelings are attached to frames, to descriptions of reality rather than to reality itself."

Even irrelevant changes - from a 90% survival rate to a 10% mortality rate - can elicit big changes in emotions. And if irrelevant changes can do so, so too can small ones. Owen Paterson's corruption, for example, might not be much different from that of other MPs in the past. But it has changed the frame. In the same way, small changes have tipped investors from not worrying about current account imbalances, valuations or housing markets to panicking about them.

But there's more. Frames are created by stories. And as Robert Shiller shows (pdf) in Narrative Economics, these are like viruses: some just die whilst others spread rapidly. And as they spread, so they change the frames in which we see things. Sir Keir Starmer is trying to achieve just such a change when he says of Johnson that "the joke isn't funny any more."

What I'm describing here is similar to but different from Timur Kuran's account of preference falsification. He shows how oppressive regimes can survive for years because everybody believes that everybody else supports the regime and so keeps quiet. But when, suddenly, people realize their error they rise up and topple the government. In this way tyrants such as Harvey Weinstein or Nicolae Ceausescu saw their power disappear overnight.

My account differs from his because people aren't misrepresenting their opinions. Sure, they are going along with the crowd - peer pressure being so powerful as Robert Frank has shown - but their expressed opinions are what they genuinely believe.

But it has something in common with it too. What we have in both cases is something described by Didier Sornette and Peter Cauwels; that social structures can seem stable only to suddenly change a lot.

And we cannot predict when or if this will happen or what will cause it; the triggers often seem innocuous. Which vindicates Jon Elster's crucial observation - that we can sometimes explain without being able to predict. All of which means we must be much more humble about our ability to foresee political change.