Explanation and inference in science

Why we should be very careful selecting hypotheses

So, we now come to the focal target this series of posts. As usual, some perfunctory comments to set the context.

Philosophy evolved (there’s no other term for it, I’m afraid) in classical Greece, although it goes without saying (or should) that philosophical, or at least philosophy-like, ideas and arguments arose independently in many other cultures, such as the Mesopotamian, Indus Valley, Semitic and others. One of my favourite philosophy-like texts is Koheleth, or as the Christians call it, Ecclesiastes. I still don’t know how it crept into the canon.

As a result of that Greek starting point, many of the problems, and their ancillary terms, have a long history; but nothing in culture stands truly still. Every debate is part of an ongoing table conversation, and words and concepts that have a certain set of contrasts when they are introduced will change these “foils” as van Fraassen called them (1977, 1981). This view, ascribed to my favourite nineteenth century logician, Archbishop Richard Whately (1787-1863), is that the field of contrasting answers is circumscribed; hence “contrastive explanation” is the name of this view. Something is explained when the answer given is the best “viable” option (my term). Those wishing to have a more non-philosopher-friendly presentation of this account should see, if they can find a copy, Alan Garfinkel’s 1981 book Forms of Explanation.[1]

Now this will be relevant later when we discuss how explanations are done in science a little more, but for now note that in debates, there are some constraints on what are suitable answers. If we are debating whether or not race is real[2], it is not an acceptable answer to say that the Ouija Board says it is (or at least not in science). So when a debate over, say, scientific method is conducted, what it meant in the nineteenth century to talk about hypotheses is quite different to what it means since the logical empiricists and their heirs and successors. So one needs to be aware of the context.