In general, questions such as “Is (subject X) a science?” are not very useful. After all, what does it matter what label you assign to something? In certain situations though, the label matters quite a lot. In the US, because of the Establishment Clause, the answer to the question of whether a theory is scientific or religious can determine if it can be taught in public schools. This is why religious people constantly seek to either label their religious beliefs (creationism and intelligent design) as science or seek to have those theories that threaten religion (like evolution) be classified as something other than science.

Apart from the purely legal and constitutional questions, to call a field a science is to give it a certain cachet and this is due to science’s well-deserved reputation for unearthing reliable knowledge. Thus academic disciplines sometimes squabble over whether their field is a science, and various disciplines have added the word science or a similar science-y sounding word to their titles, usually on the basis that their work is empirical and they use mathematical and statistical techniques as part of their analysis. The science label has become much in demand because it seems to add value to the discipline.

Recently, the old debate of whether philosophy is a science has resurfaced with philosopher Colin McGinn arguing in the affirmative and Julian Friedland arguing against.

McGinn says that the word ‘philosophy’ started out being a broad umbrella term that included the empirical and the non-empirical in the search for objective knowledge but as time went on the group split into moral philosophers, mathematical philosophers, and natural philosophers, with the last group later calling themselves ‘scientists’ and later further subdividing into physicists, chemists, biologists, and the like. Philosophy got left with just the moral philosophy part, essentially becoming a label that implies no empirical content, does not produce objective knowledge, and consists largely of telling people how they should live. He argues that being excluded from the new umbrella label of science has led to disrespect and that there is enough overlap between philosophy and science for the former to share in the label of the latter.

Our current name is harmful because it posits a big gap between the sciences and philosophy; we do something that is not a science. Thus we do not share in the intellectual prestige associated with that thoroughly modern word. We are accordingly not covered by the media that cover the sciences, and what we do remains a mystery to most people.
…
[M]ost of the marks of science as commonly understood are shared by academic philosophy: the subject is systematic, rigorous, replete with technical vocabulary, often in conflict with common sense, capable of refutation, produces hypotheses, uses symbolic notation, is about the natural world, is institutionalized, peer-reviewed, tenure-granting, etc. We may as well recognize that we are a science, even if not one that makes empirical observations or uses much mathematics. Once we do this officially, we can expect to be treated like scientists.

He suggests that philosophy change its name to ‘ontical science’ or ‘ontics’ to make it sound more like a science.

Friedland argues that using McGinn’s arguments, we could re-label every academic discipline as science. He says that calling philosophy a branch of science is misguided, because while “science and philosophy do at times overlap, they are fundamentally different approaches to understanding.”

Interestingly, he says that even though philosophy is not a science, it can produce objective knowledge and that its methods “can actually yield greater certainty than the scientific.” He makes his case by arguing that philosophy shares much in common with other fields that are believed to produce objective and quantitative knowledge like mathematics, theoretical physics, psychology and economics in that they “are predominately rational conceptual disciplines. That is, they are not chiefly reliant on empirical observation. For unlike science, they may be conducted while sitting in an armchair with eyes closed.” In other words, he argues that the lack of an empirical basis does not preclude the attainment of objective knowledge.

He gives examples from Socrates and Wittgenstein of how philosophy enables one to arrive at what he considers timeless truths and then concludes:

In sum, philosophy is not science. For it employs the rational tools of logical analysis and conceptual clarification in lieu of empirical measurement. And this approach, when carefully carried out, can yield knowledge at times more reliable and enduring than science, strictly speaking. For scientific measurement is in principle always subject to at least some degree of readjustment based on future observation. Yet sound philosophical argument achieves a measure of immortality.

While I agree with Friedland that philosophy is not science, I disagree with his characterization of theoretical physics as not chiefly reliant on empirical observation. Theoretical physics would be useless if it were not firmly embedded in the empirical world of physics as a whole. It is not true that it can be done ‘while sitting in an armchair with eyes closed’. While it can be done in an armchair, all theoretical physicists keep at least one eye on what is going on with experiments, to keep them from floating off into dreamland.

While I also agree with Friedland that philosophy’s tools of ‘logical analysis and conceptual clarification’ are invaluable in all areas of thought, I do not think that it arrives at truths that are immortal except in the case of tautologies. For that matter, neither does science arrive at immortal truths. All scientific knowledge is properly viewed as provisional.

A necessary condition for something to be science is that it have an empirical basis. I see philosophy as being that discipline that enables us to sharpen our use of language, clarify meaning, and bring questions into sharper focus. While extremely valuable in the study of any field, it is not empirical except in the trivial sense that it sometimes appeals to everyday observations. While it does generate hypotheses, it does not systematically investigate the consequences of those hypotheses by suggesting experiments that can be designed and constructed to see how well they compare with data, the way that theoretical physics does. Hence is not science.