Monday, November 12, 2007

Why is so much social science rubbish?

Richard Feynman’s father used to tell him as a child that “it isn’t science unless you can write it down mathematically”. Some people think this is physics chauvinism, although biology, chemistry and economics also come with mathematical models.

I would even like to invert the point. Human groups are so prone to prejudice and the well-named ‘groupthink’ that it is a miracle that any opinions which defy a convenient orthodoxy ever get established. Truths which are inconvenient seem to me to be most likely to surface under the following conditions.

1. They are irrelevant to interest groups. Hard sciences such as astronomy used to be threatening, as Galileo discovered. But particle physics and cosmology today are so esoteric that they threaten few, so limitations on progress come mostly from other causes (see below).

2. It is possible to do well-defined experiments. Social science often falls at this hurdle, with positions being advanced with no thought of testing them through rigorous experimentation. Particle physics currently has an experimental problem as discussed by Peter Woit, Lee Smolin.

3. The inability to frame hypotheses in a precise, i.e. mathematical form. Without clear and unambiguous hypotheses to test, experiments cannot be properly designed and results are inconclusive. Feynman’s father was right.

Much social science doesn’t implement the dialogue between hypothesis and experiment, which drives convergence to truth. In this respect it is not ‘science’. In the absence of empirical correction via the scientific method, such work is prone to capture by activist interest groups. The results we know as political correctness, the erection of taboo areas and arguments from authority.

I think people who have been trained in a scientific tradition often don’t understand that their social science opponents are not actually operating in that paradigm. Unfortunately, the rhetorical styles of ‘appeals to common sense’, arguments from authority and even abuse are closer to everyday social discourse. They often have resonance with a lay audience, while the scientist flounders.