I’m reading Jim Manzi’s Uncontrolled: The Surprising Payoff of Trial-and-Error for Business, Politics, and Society right now. No complaints, though that’s no surprise, as I’m familiar with the broad outline’s of Manzi’s work, and have found much to agree with him on in the past (though there are issues where we differ, never fear). That being said, I did ponder one aspect of Manzi’s characterization of science: that it makes non-obvious predictions. This is not controversial, and I don’t want to really quibble with it too much. But in the context of social science in particular I think one of the gains of ‘science’ is the clarification of obvious predictions.
To illustrate what I’m talking about, the inverse-square law defines the decay of the intensity of light from a radiation source. Is this non-obvious? The precise decay function isn’t obvious, but the general trend is clearly obvious. Intensity decreases with distance. We know this intuitively. But it is obviously a gain to quantitize and formalize this phenomenon, as it can then be integrated algebraically into a broader system.
And so it is with social science phenomena. For example, I can say that most of personality variation within a population is accounted for by genes and non-shared environment. But what does this mean? It could mean that 20% is accounted for by variation in genes and 60% by non-shared environment, with 40% shared environment. Or, it could mean that 40% is accounted for by variation in genes and 40% by non-shared environment, and 20% shared environment. These are not distinctions without difference. The ‘science’ here is less a counter-intuitive framework of prediction and projection than a clarification and precision of broad trends which might be unsurprising, with the ‘non-obvious’ aspect only coming to the fore when quantities are integrate into larger frameworks. So, for example the prediction that heritable inequality may actually increase when you maximize environmental inputs.