Testing vs. Matching and the Attraction of the Aphorism

Don't ask me where it was, but I recently came across an internet comment that dismissed 19th century novels. These novels, the commenter said, traded in aphorisms, such as Tolstoy's "Happy families are all alike; every unhappy family is unhappy in its own way." An interesting question with regard to this statement is whether it's true, and if you want to know, the commenter argued, you should turn to the social and behavioural sciences, not old novels. To him (I think it was a he), 19th century novels were made redundant by the modern social and behavioural sciences like punchcards were made redundant by PCs.

My initial reaction (which I kept in my head) was that if you want to learn about family happiness, you should by all means turn to the social and behavioural sciences; read Tolstoy if you appreciate his language, characterizations, or plots.

But now I’m not so sure that gets to the heart of the matter. I am currently reading Proust’s À l’ombre des jeunes filles en fleurs (in German, mind), and it is full of apodictic statements about how the human mind works. Thing is, these are among the best bits of the novel. What’s going on?

In a different context, Karl Smith writes:

The world is extremely complicated and doesn’t make sense – at least not in the way we want it to. We don’t want to understand the world simply by following some complex routine of intellectual gymnastics. We want it to makes sense intuitively. We want it to bound up in a single completely digestible ball. The world, sadly, does not always comply.

So we build miniature models of the world in our minds – fictions that do make sense. When we run into a part of the world that doesn’t co-operate we either shoehorn our observations into that miniature model or tear through, blogs, articles and books until we find someone who can.

The statement “it makes sense” is, however, a statement about how pleased we are with our efforts to shoehorn observations into our miniature model. It is not a statement about our understanding of the world.

To check our understanding of the world we have to ask not “does it make sense”, but “how would I know if I was wrong?”

“How would I know if I was wrong?” is, of course, the Popperian ideal of Capital-S Science: You try to show that your guess was incorrect; if you cannot show this, you accept that your is correct – tentatively, for the time being. As a measure of caution, you’re biasing yourself against your guess.

The other mode of comparing data to guesses Smith describes is biased in favour of confirmation. This describes how people approach this problem most of the time. For example, there is a study (the citation to which escapes me right now) that looked at doctors diagnosing. The standard procedure seems to be that doctors listen to patients for a while, develop a guess about the problem and then ask follow-up questions in order to confirm that guess.

In other words, the mind uses a particular version of satisficing. If the data and the model match, that’s good enough, and if it’s good enough, that produces a pleasant sensation, like completing another row in Tetris.

I submit that the aphorism works by making sense of disparate data, thereby reducing the need to compute. A bit like factor analysis, if you don’t mind another simile. Reduction of the need to compute produces pleasure. It wouldn’t be hard to come up with a just-so evopsych story about why this should be so.

In contrast, when you’ve internalized Science, you’ve internalized uncertainty, which is unpleasant. This is one reason why reading novels is more fun than doing science.

Of course, testing is a much better way of getting at the truth than matching. But then, novels are not for truth, they're for entertainment.