Inference to any old explanation

We’ve seen that scientists often proceed by inference to the best explanation: they infer that a hypothesis is probably true from the fact it would, if correct, best explain their observations and evidence.

Darwin inferred the truth of the natural selection hypothesis from the fact that it explained his observations better than any alternative.

We use inference to the best explanation in everyday life too. I’m looking out my window on a winter’s day. The grass is white. I infer that it’s cold outside. A truck carrying a load of flour could have overturned upwind sprinkling its load on my lawn, but on a winter’s day where I live that is a less likely explanation than ‘it’s cold outside. I’m looking at frost’.

The conclusions in these examples do not follow deductively from the observations or evidence. The flour truck explanation could be true; Darwin’s observations could be explained by a mischievous deity out to fool us (see the Philip Henry Gosse example in the Science and falsification video). Rather the conclusions follow from the fact that they provide the best explanation of the observations.

When making inferences to the best explanation, then, we need to show that the explanations to which we infer really are the best available. Inferences to ‘any old explanation’ do not give us grounds to be confident about the truth of those explanations.

For reasons we have already encountered, however, we are often tempted to infer explanations for our observations which are not the best available. Some reasons include:

• Most generally, people – probably all of us some of the time at least – simply fail to think very carefully about which explanation is the best. We are attracted to an explanation because of who it is offered by rather than because it makes the events it describes unsurprising and fits with the rest of best theories about the world, or because it’s striking rather than a bit tedious.

• Confirmation bias leads us to prefer explanations which confirm our pre-existing beliefs. If we’re already sceptical about the claims of mainstream medicine, we’re more likely to accept an explanation of falling disease rates which gives little or no credit to vaccination.

• There is a tendency, to seek ‘grand’ or complex explanations for grand events: people simply find it unsatisfying to think that President Kennedy could have been shot by a lone gunman: his death was too important to be explained by the actions of a disaffected misfit working on his own.

It is sometimes suggested that these and related reasoning vices explain the popularity of conspiracy theories, theories which explain events such as the assassination of President Kennedy, the attacks in the US on September 11, 2001, and the moon landings, as the result of plotting by covert groups or organisations. Of course, as recent writers on conspiracy theories have pointed out, there are genuine conspiracies. Someone conspired to attack the US on 9/11: the question is who?

Understanding inference to the best explanation, however, makes clear that we have to ask whether an explanation which involves a conspiracy (and, if so, a conspiracy between who,) is the best explanation for what we see. When we seek explanations from this perspective, we need to bear in mind the reasons we might be tempted to prefer ‘any old explanation’ to the best explanation.