A look at the Philosophy of Science and one theory of scientic inference; inference to the best explanation (IBE) and what it entails.
In the scientific world, scientists advance their theories and understanding of the world around us by way of making inferences. They often take what they observe and from this they will infer something else. Scientific inference has occurred for as long as scientists have been working. I plan to look at one account of scientific inference known as Inference to the Best Explanation and describe what it entails exactly. I will address its main problems and objections and then ascertain whether these problems can be solved or if they undermine this method of inference.
Inference to the Best Explanation [from this point referred to as ‘IBE’] is a view held by scientific realists which claims that the accepted hypothesis should be the one with the most explanatory power. Since it is “a vice of a theory that it explain nothing or very little”1, we are logically compelled to accept theory A if it explains more facts or explains them better than any other theory A*.
In accepting that one theory explains certain facts better than its rivals, then it is reasonable to believe that there is some truth contained in this theory. It is adequate to conclude that this theory has greater verisimilitude than its rivals.
In other words, if you are faced with a set of rival hypotheses and are seeking to find the truth, then you are obliged to choose the theory which is the best explainer.
Before exploring the merits and weaknesses of IBE, I think it would be helpful to define exactly what it means when we say we are looking for the ‘best’ explanation. Peter Lipton introduces two ideas; likely and lovely. The likeliest explanation is the most probable explanation. The loveliest explanation is, if correct, the most explanatory and provides the most understanding2. Of course, the best explanations will be both likely and lovely- they will be well warranted by evidence and conducive to understanding.
Since scientific realists are aiming for the truth we can see how it is tempting to accept that the idea of ‘inference to the likeliest explanation’ with likely pertaining to ‘most likely to be true.’ Lipton rejects this though, on two grounds. Firstly, he states that IBE should tell us on what basis we deem one hypothesis likelier than its rival. By stating we infer the likeliest hypothesis implies we already know the truth and this would belittle inference. Secondly, IBE should tell us how specifically explanatory considerations guide inference otherwise it loses its name and is in danger of falling into another account of induction. Thus, Lipton proposes ‘best’ should be identified with ‘loveliest.’3