What is explanation in science all about? - Chapter 3

Science wants to explain the world around us, partly for practical ends, and partly to satisfy our curiosity. But what is scientific explanation? Our starting point to look at this question will be Carl Hempel who thought about this in the 1950s.

What is Hempel's covering law model of explanation?

Hempel figured that scientific explanations are usually given in response to "explanation-seeking why-questions", so questions that demand explanation. A scientific explanation provides an answer, and to know what a scientific explanation is, he thought, the essential features of such an answer should be determined.

He said scientific explanations usually have the logical structure of an argument: a set of premises, followed by a conclusion. The conclusion states that the phenomenon that needs explaining occurs and the premises then tell us why this is true. So the premises should explain the conclusion. Hempel's answer to this problem contained three things. First, the premises should entail the conclusion: thus, deduction. Second, the premises should all be true. And third, they should consist of at least one general law, sometimes also called laws of nature. 

The phenomenon to be explained is named the explanandum and the general laws and particular facts that explain it are the explanans. The explanandum may be either specific or general. 

So, this model says that the essence of explanation is to show that the phenomenon is "covered" by a certain law of nature. But not all scientific explanations fit this model. Hempel knew this, but said that if the explanation were spelled out completely and very detailed, there would still be some law of nature. Furthermore, Hempel said that every scientific explanation can also be a prediction. If the phenomenon was not yet known, it could have predicted it. And then also reliable prediction is also a possible explanation. So explanation and prediction are structurally symmetric. However, in some cases this definitely does not work, not everything that can be predicted can also be explained in the same way later.

Still, there are cases that fit Hempel's model but are not really scientific explanations. Two cases are:

• The problem of symmetry. Explanation is an asymmetric relation: that x explains y does not mean that y explains x in the same way. Hempel's model does not respect this asymmetric relation and therefore includes things that are not really scientific explanations. 

• The problem of irrelevance. A good explanation should contain the relevant information. "Explanations" that actually use the irrelevant information and can not be viewed as good explanations, but still have a correct  structure of explaining, can be included in Hempel's model which is not good.

What does causality mean in explanation?

When looking for a better model, some believe the key lies in the concept of causality: because often, what causes something is what explains it. In some of the problems with the covering law model, the causal models do succeed. It could also be that because causality is obviously asymmetrical, explanations are asymmetric too. 

Hempel was an empiricist, and empiricists are often suspicious of the causality concept. Empiricism states that all knowledge comes from experience. Like Hume, who thought: it's impossible to experience causal relations, so they don't actually exist; it's what we, humans, make of it. Empiricists treated causality with much caution, which is why Hempel did not take it into account. In recent years, empiricism has decreased in popularity.

Still, also causality-based accounts encounter less fitting cases. For example, theoretical identifications, like: water is H2O or temperature is molecular kinetic energy. These terms explain what these concepts are, but it has nothing to do with causality.

Can science explain everything?

There's many things science has explained, and there's many things left to uncover, probably possible due to always evolving science. However, some philosophers think science will never explain everything, because in order to explain something, something else needs to be invoked - and what explains the second thing then? Fundamental laws have to be used and probably at least some will themselves remain unexplained. Some philosophers think for example consciousness, and its subjective aspect, can never be fully explained. But it seems that only time can really tell.

What is multiple realization?

Most people think of physics as the most fundamental science, since all of the other sciences also study things that are made up of physical particles. Still, the other sciences seem largely autonomous and also important. But why can the other sciences not be reducible to physics? Some philosophers say that's because the objects studied by the higher-level sciences are multiply realized at the physical level. There is no limit on the range of different physical properties that a lot of things can have, because they are multiply realized at the physical level. For example, you can not say: "x is an ashtray if and only if x is...." as there are so many possibilities. The vocabulary of fundamental physics is not elusive enough for this.