Question Dear Philosophers, When philosophers write about scientific method, are they proposing a description of the actual practices of scientists or are they attempting to produce a normative theory of what science should be like? If it's the former, then shouldn't this be answered by historical study and not philosophy? If the latter, why do philosophers talking about scientific method bother to look at the history of science at all if one cannot gurantee an 'ought' from an 'is'? BMW

Accepted:
August 30, 2006

Comments

Generally, when philosophers write about scientific reasoning, they are interested in how scientists (or, more broadly, how anyone) ought to reason. For example, they might be interested in specifying what it takes for a piece of evidence to count in favor of a given hypothesis, and why certain pieces of evidence should count for more than others. They are generally not interested in explaining why scientists in fact regarded a given piece of evidence as counting for more (that's for historians to figure out) except insofar as this explanation goes via some account of why that piece of evidence *should* have counted for more. It is sometimes said that philosophers are trying to give a "rational reconstruction" of scientific reasoning.

But the history of science is not at all irrelevant to this task. For example, any account of scientific reasoning that regards as unjustified some renowned episode of scientific reasoning has a great (though not impossible) hurdle to climb, just as any epistemological theory according to which (for example) I am currently unjustified in believing that my name is Marc Lange faces a high (though, some would say, perhaps not absolutely impossible) burden. As another example, a "rational reconstruction" of what is meant by an "ad hoc" theory would gain credibility insofar as it deems "ad hoc" many of the candidates from the history of science that were at the time widely so labelled (and does not deem "ad hoc" many of the theories that were widely considered justified). Of course, scientists might have been mistaken in which theories they considered "ad hoc". But just as we test a moral theory by seeing how well it fits our moral intuitions, always bearing in mind that some of those intuitions might be mistaken, so likewise we can test a theory of good scientific reasoning by seeing how well it fits our historical exemplars of scientific reasoning, always bearing in mind that some of those exemplars might in fact not be as exemplary as we had thought.

Philosophers often think of the philosophy of science as being less of a descriptive enterprise than is either the history or the sociology of science. The philosophy of science, it is often said, concerns itself in part with an evaluation of scientific practice. For instance, a philosopher of science does not just want to know what scientists have, as a matter of fact, accepted as good explanations; scientists might, after all, have jointly succumbed to some widespread error. Rather, the philosopher of science wants to know what would really constitute a good explanation—where it is assumed that scientists might on some occasions have taken something to be an explanation which was not. The philosophical project is thus in some sense a normative one, namely to determine what the scientist should take an explanation to be. Likewise, to consider another example, the philosopher of science is not particularly interested in whether scientists do believe that evidence about the past provides one with grounds for believing anything about the future. The point, the philosopher of science, will say, is not to conduct a poll, but rather to determine whether past evidence ought to convince us about the future.

In this way, the philosophy of science often positions itself as independent of scientific practice. But at the same time, there are limits regarding this independence. For should the philosophy of science cut itself off entirely from actual practice, it will lose its claim to provide an understanding of that practice. For instance, if a philosopher of science should conclude that, contrary to what one might have thought, no object is actually observable using our five senses—not just objects like atoms, but trees and dogs as well—then one might well wonder whether that philosopher has ceased to talk about the activity of observation as it occurs in scientific practice. Likewise, should a philosopher conclude that, strictly speaking, scientists have yet to offer a single explanation, then it seems a good question whether that philosopher could possibly be shedding light on the concept of explanation as scientists employ it.

It seems, then, as if a philosopher must be beholden in some way to the facts of scientific practice—but not so beholden as to render the philosophical enterprise purely descriptive. Philosophical reflection is thus pulled both away from actual practice and towards it.

This tension is at the root of what can make philosophy so difficult (in a certain sense of "difficult"). It is often a vexing matter to work out, loosely put, where the philosopher stands when he or she "does philosophy." How can a perspective that is intimately tied to the practice that initiates a philosophical inquiry provide a stance from which rational appraisal of the practice is possible? And yet, how can a perspective that is substantially removed from the practice be understood as providing an appraisal of that practice? Philosophical inquiry, unlike research in the natural sciences, relies largely on judgments (or "intuitions," as they are sometimes called) for its data. A fundamental question that must exercise anyone who is engaged in philosophical reflection is how a philosopher can take a critical vantage point on such judgments without thereby sawing off the branch on which he sits.

Here is another way in which the normative and factual mix in the philosophy of science. One of the central normative issues is whether we are justified in saying that our best scientific theories are (at least approximately) true. The best known argument for saying that our best theories are true is the miracle argument, according to which the truth of our best theories is the only account that avoids making their remarkable predictive successes miraculous. The best known argument for saying that our best theories are not true is the pessimistic induction, according to which the fact that so many of the best theories in the history of science have turned out to be false (even when they were predictively successful at the time) makes it very likely that our current best theories will turn out to be false too. The miracle argument and the pessimistic induction address the normative question of whether we are justified in believing our best theories to be true, and they depend on factual claims about the success of our best current theories and about the failure of best theories of the past.