What is a scientific explanation?

An important issue concerning emergent phenomena is whether part of their definition should be that they "cannot be explained/predicted" solely in terms of the properties of the components of the system.  As I have discussed before, there are several alternatives such as difficult to explain, extremely difficult to explain, impossible to explain, and possible to explain in principle, but not in practice...

To further consider this issue, it is helpful to back up and consider the general question, "What is a scientific explanation of a specific natural phenomena?" This has received a lot of attention over the past century from philosophers of science. A nice accessible introduction is found in Philosophy of Science: A Very Short Introduction by Samir Okasha. In chapter 3, he discusses at length the Covering Law model of explanation, developed by Carl Hempel.

Here are the basic elements of the model. Hempel observed that "scientific explanations typically have the logical structure of an argument, i.e., a set of premises followed by a conclusion." More specifically, 

1. The argument should be a deductive one. The conclusion is the phenomena to be explained.

2. All the premises must be true.

3. One of the premises should be a scientific law.

Here is an example:

Premise a. Astronomical data about planetary motion.

Premise b. Kepler's laws of planetary motion can describe the data.

Premise c. Newton's theory of gravity.

Conclusion. The motion of planets can be described by Newton's theory of gravity.

Hempel considered that explanation and prediction were two sides of the same coin. 

Okasha also discusses some weaknesses of the covering law, namely that it is too broad allowing "explanations" that are not really scientific. Specifically, there is the problem of symmetry and the problem of irrelevance.

In the next post, I will explore how this model for scientific explanation might help frame discussions about whether emergent phenomena can be predicted.