How true is a fact?
Of what value are the language games of philosophy?
Absolute truths and ‘by and large’.
Truth and facts are not as closely related as most people seem to imagine. Truth is related to a word-game system known as philosophy. Facts are related to checkable experience. The practical-minded who are not much concerned with such matters can skip to the next chapter.
If every swan we happen to see is white, can we make the bold statement that ‘all swans are white’? We can and we do. For the moment that statement is a true summary of our experience. In this sense it is also a fact.
The first black swan that we see makes the statement untrue. So we have switched from true to untrue with remarkable abruptness. Yet if we are looking at facts, one hundred white swan experiences are still set against one black swan experience. So as a matter of experienced fact we can say: ‘most swans are white’; ‘by and large swans are white’; ‘slightly more than ninety-nine per cent of swans are white’.
This ‘by and large’ stuff is immensely practical (by and large children like ice cream; by and large women use cosmetics) but of no use at all to logicians. The ‘all’ is essential in the statement ‘all swans are white’. This is because logic has to move from one absolute truth to another. ‘If this is true … then this follows …’
When we come across the first black swan, the statement ‘all swans are white’ becomes untrue. Unless we choose to call the black swan something else. Now it becomes a matter of words and definitions. If we choose to keep whiteness as an essential part of the definition of a swan, then the black swan is something else. If we drop whiteness as an essential part of the definition, then we can include the black swan and we base the definition of a swan on other features. It is the design and manipulation of such definitions that is the essence of philosophy.
White hat thinking is concerned with usable information. So the ‘by and large’ and ‘on the whole’ idioms are perfectly acceptable. It is the purpose of statistics to give these rather vague idioms some specificity. It is not always possible to collect such statistics, so we often have to use the two-tier system (belief, checked fact).
… By and large corporations that base their spending on extrapolated future sales run into trouble. (It is possible to point to a few companies that have done this and been successful.)
… Sales will tend to rise if prices are lowered. (When house prices rise there may actually be increased sales for reasons of speculation, fear of inflation and fear of being left behind.)
… If you work hard, you will be successful in life. (A lot of hard-working people are not particularly successful.)
The spectrum of likelihood might be expressed as follows:
Always true
Usually true
Generally true
By and large
More often than not
About half the time
Often
Sometimes true
Occasionally true
Been known to happen
Never true
Cannot be true (contradictory)
How far along this spectrum is it permissible to go with the white hat role? As before, the answer to that question lies in the framing of the information. For example, it can be useful to know things that happen only very occasionally.
… Measles is usually harmless, but it can sometimes be followed by secondary infections, such as ear infections.
… In very rare cases inoculation can be followed by encephalitis.
… When irritated this breed of dog has been known to snap at children.
Obviously there is a value in being aware of this sort of information. There is also a dilemma. In the second example given, people’s perception of the danger of encephalitis following inoculation may be thousands of times greater than the actual statistical danger. So it can be important to give actual figures in order to avoid inadvertent misinformation.
Are anecdotes acceptable under white hat thinking?
… There was a man who fell out of an aeroplane without a parachute and survived.
… Ford is said to have designed the Edsel on the basis of market research and it was a total disaster.
These may indeed be statements of fact and as such the white hat thinker has the right to put them forward. They must be framed as ‘anecdotes’ or ‘instances’.
… Designs based on market research can often fail. Take, for instance, the Edsel car, the design of which is said to have been based on market research. It was a total failure.
The above statement is not legitimate white hat thinking – unless there is much more support for the claim that designs based on market research fail. Cats can fall off roofs but that is not normal behaviour.
Exceptions stand out simply because they are exceptions. We notice black swans because they are usually in a tiny minority. We notice the man who survives a fall from an aircraft without a parachute because it is somewhat unusual. The Edsel is always referred to for the same reason.
The purpose of white hat thinking is to be practical. So we must be able to put forward all sorts of information. The key point is to frame it properly.
… All the experts predict that the interest rate will fall by the end of the year.
… I talked to four experts and each of them predicted that the interest rate will fall by the end of the year.
… I talked to Mr Flint, Mr Ziegler, Ms Cagliatto and Mr Suarez and all of them predicted that the interest rate will fall by the end of the year.
Here we see three levels of precision. Even the third level may not be good enough. I may want to know when you talked to these experts.
There is nothing absolute about white hat thinking. It is a direction in which we strive to get better.