The limits of logic and the lethal danger of the "self-evident."


Good logic is almost infinitely preferable to bad logic, but it has one serious problem: no matter how logical your analysis, it is only as good as the premises with which it starts.

Most computer users will be familiar with the acronym GIGO - Garbage In, Garbage Out.

This is the principle that if you feed a computer with rubbish data, the output will also be rubbish.

Many people do not pay as much attention as they should to the fact that this does not just apply to manufactured computers and devices: it is equally applicable to the organically grown computing device inside their heads.

Yes, the GIGO principle also applies to the human brain.

If you start considering the situation with inaccurate assumptions - even if on the information available to you those assumptions appear entirely reasonable - proceeding with the most perfect logic will usually lead you from false premises to false conclusions.

And of course the most dangerous assumptions are the unconscious ones you don't even realise you are making.

I have been following with interest the inquests into why the pollsters and most political parties did so badly - NOT out of schadenfreude (well, mostly not) but out of a wish to learn lessons.

Winners are even more liable than losers to make the mistake of re-fighting the previous war - although I will be posting a link later to an interesting article which suggests that one of Labour's problems was a debate among Ed Miliband's inner circle about how to move on to the issues of 2015 - and the author of the article quotes one insider who thought the party was effectively re-fighting 2010.

Whatever my involvement in politics may be in the 2020 General Election, I want to be fighting that election and not trying to rerun the tactics of 2015. So I'm following the debates about why we won in 2015 to see how our opponents are likely to change.

One of the posts on a debate on "Political Betting" in response to the Guardian's inquest on Labour's defeat referred back to an April 16th article by Alberto Nardelli suggesting that the odds were stacked against David Cameron.

My next post will be on the inquest itself, but this Guardian article is a classic example of how one really bad assumption, no matter how reasonable, can cause a completely rational mind to get things utterly and completely wrong:

http://www.theguardian.com/politics/2015/apr/16/in-the-election-numbers-game-the-odds-are-stacked-against-cameron

In this case the wrong assumption was that the voting intention surveys being published from all the main pollsters were an accurate predictor of how those electors who cast a vote were going to behave. Hardly a stupid thing to believe - the last time the polls had been as catastrophically wrong as they were to be in May 2015 had been nearly a quarter of a century before in the 1992 election.

Indeed the author of the article did refer to one or two of the items of evidence which were suggesting the possibility of a different outcome - saying, for example, that

"The Tories’ belief that the British public’s doubts over Miliband’s suitability for the premiership may still end up being prophetic, yet right now any such concerns aren’t putting a dent in the polling deadlock."

There were a very small handful of people who did suggest that the 2015 election was relatively new territory and the electoral conditions for a polling upset like 1992 might exist. Some of those people, including one or two who I spoke to or who often comment here, did predict a Tory win. In the event those people were right. But Alberto Nardelli, like the vast majority of people, assumed, perhaps without really thinking about it, that the pollsters had learned from 1992 and would not be so wrong again. The key indicator which gives away a lethal unconscious assumption are often words like "self-evident" and sure enough those words are there in the article:

"It is self-evident that the race to become the largest party is extremely close."

It may have appeared to be the case, and most of us including me believed it. I still think that the result was by no means predetermined, and it was right for me to say that there was everything to play for. However, with 20:20 hindsight the election was not as close as we all thought.

That phrase inspired me to write this blog piece because those words are one of the flags which should serve as a warning. When we are tempted to say something is self-evident, we should remember that this idea is one of the most common signs of an ill-considered assumption - a premise which has not been properly thought through properly precisely because it appeared to be so obvious.

The world we live in is often "counter intuitive" - which is a posh academic way of saying that things which appear to be obviously true are sometimes wrong. This is one of the most common causes of the kind of lethal wrong assumption which can wreck an analysis and lead people to follow a fatally flawed course of action. Whenever you come across this phrase, one of the first rules about dangerous assumptions should be this:

Things which appear self-evident are surprisingly often not the case.

Comments

Popular posts from this blog

Nick Herbert on his visit to flood hit areas of Cumbria

Quotes of the day 19th August 2020

Quote of the day 24th July 2020