What I'll say to the Russians

Our department is in the middle of a two-day video conference with philosophers at Moscow State University. The whole thing is pretty freewheeling, with people presenting on the possibility of progress in philosophy and on what they think about the last several decades in their specialty.

I present tomorrow, with the rather grandiose aim of discussing the last fifty years in philosophy of science. I have only been attentive to philosophy of science for the last decade and a half, so I feel like some of this might be confabulation. My usual preference is to work from notes rather than to read a paper. As an aid to the translator, however, I've written up my remarks in advance. I'm posting them here, below the fold.

Philosophy of science in the 21st century

Philosophy of science in the United States, at the middle of the last century, was primarily logical empiricism. From an internal point of view, this was a response to two dramatic developments: Modern logic and revolutionary physics.

The first (modern logic) had dramatic effects in every area of philosophy. Anything important was translated into expressions in a formal language.

The second (modern physics) was important primarily for philosophy of science. Newtonian mechanics, which had been the paradigm example of established science, was overthrown by relativity and by quantum mechanics.

Enough time has passed that philosophers writing about Carnap and Quine now count as writing the history of philosophy. I am going to tell an imprecise and idealized story: The last half-century in philosophy of science can be seen as a reaction against these two earlier influences.

Logic as the language for philosophy

For the logical empiricists, theories were sets of sentences in a formal language. This Received View of theories succumbed to serious criticism in the 1960s. Opponents developed the semantic conception of theories, according to which a theory is a set of models. In early formulations of the semantic conception, this meant formal models: abstract semantic structures. Subsequent philosophers have been more permissive about what may count as a model, allowing in drawings and physical scale models.

As it stands now, I do not think that there is a consensus about what theories are or how they may best be represented. In my own work, I have argued for theory concept pluralism. This is the view that there is no single correct conception of `theory.' Instead, there are several equally legitimate conceptions of theory that are useful for different purposes.

Physics as the paradigm science

It is no longer the case that philosophy of science and philosophy of physics neatly coincide. Philosophers take the special sciences seriously.

This is reflected in the structure of the profession. Many departments look to hire specialists. There are specialist journals for philosophy of biology, philosophy of psychology, philosophy of chemistry, and so on.

The difference is also substantive. Some of the questions asked by philosophers of a special science may have little relevance for philosophers of other sciences. For example, much has been written about the question of what `fitness' means in the theory of natural selection. This is a foundational question for evolutionary biology, and philosophers have had helpful things to say. Yet our understanding of other sciences, or of science generally, has nothing to do with what we say about `fitness.'

At the same time, philosophy of physics has become its own autonomous specialty. Technical questions about how to reconcile relativity with quantum mechanics, for example, have no more relevance for the philosopher of biology than `fitness' has for the philosopher of physics.

So are there any questions left for general philosophy of science, rather than the philosophies of the special sciences? In the next section, I consider the obvious candidate: Scientific realism.

Scientific realism

Scientific realism, roughly, is the view that our best scientific theories accurately represent both the observable and unobservable world. More precisely, it is the view that our mature theories are approximately true.

By the 1970s, philosophers of science had come to accept that there was no principled difference between vocabulary that identifies observable entities (like `tree' and `chair') and vocabulary that identifies theoretical entities (like `atom' and `electron'.) One might express this lesson, as Quine did, by saying that trees and atoms are all posits introduced to make sense of our experience--- the words `tree' and `atom' are both, in some sense, theoretical. At the same time, he said that we should look at our best science and believe in whatever entities are posited in its theories. So atoms are just as real as trees. As such, after the demise of anti-realist logical positivism, many philosophers were scientific realists.

In the last few decades, even most anti-realists have accepted that words like `atom' refer to atoms just as `tree' refers to trees. They have insisted, nevertheless, that we should believe in trees but not atoms; that is, we should believe in the observable parts of our theories but not the unobservable parts.

For Bas van Fraassen, whose book The Scientific Image was published in 1980, anti-realism is closely connected with the shift to the semantic conception of theories. The distinction between the observable and unobservable is not a distinction to made between two kinds of vocabulary, according to van Fraassen, but instead between two kinds of thing; that is, the models of a theory are divided into observable and unobservable parts.

Debates about scientific realism have generated a massive literature. Since it is about scientific theories generally, rather than theories in physics or biology especially, one might conclude that general philosophy of science is alive and well.

This would be too quick, I think. Despite realists' claim that the theories in our mature sciences are approximately true, realism is simply more appealing for some sciences than for others. Take two examples:

1. The theory of plate tectonics would be perverse if there were not actually tectonic plates moving very slowly across one another. Even though we cannot apprehend this motion with our unaided senses, it is hard to imagine accepting plate tectonics without believing in continental drift.

2. The entities posited in particle physics are more exotic. One can imagine accepting particle physics and using it for practical purposes without actually believing in, for example, the top quark.

So the question of scientific realism is not one that can be decided for science simpliciter. Perhaps we should believe in some of the unobservable entities posited by our best scientific theories, but not in others.

The opposition between scientific realism and anti-realism cannot be settled for all or most of science by a master argument. It can only be settled for particular theories by particular arguments. Determining whether or not we should be realists about black holes, for example, does not settle whether we should be realists about genes. One might worry that these determinations just become a matter of doing science, rather than doing philosophy at all.

The disunity of science

In the late 1980s and 1990s, some philosophers of science argued that there is no such thing as science simpliciter--- there are just many different sciences. This `disunity of science' position was partly a reaction against the attitude, discussed earlier, that physics is the paradigmatic science.

When physics was taken to be the exemplary science, it was common to think along these lines: `The correct account of the world is the one given by physics. Biology, then, must just be a simplified summary of the physics that goes on in organisms. Psychology must just be a simplified summary of what is happening at the biological and so ultimately physical level. And so on. Everything must reduce to physics, at least in principle.'

Yet philosophers of biology found that reduction doesn't seem to be possible, even in principle. Does this make biology unscientific? Only if science must provide a unified, hierarchical picture of the world. If we take as a premise that biology is scientific, then the failure of reduction means that science does not provide a unified picture of that kind.

The claim of disunity suggests that there might not be any room for general philosophy of science, as opposed to philosophies of special sciences.

Is there any philosophy of science left?

Of course, I think general philosophy of science is possible. Most of the articles I have published are about science generally, rather than just being about physics, biology, or information technology--- although I have used examples from all these areas.

Consider a physicist describing the projectile motion of a cat as it's thrown through the air and a biologist describing the metabolism of the cat. They are, of course, describing the same cat. The various sciences, although not perfectly unified, are not entirely separate. They give accounts of the same world. So it would be weird if there were nothing general that could be said about science.

So we strive to make our arguments as general as possible, while acknowledging that no single argument can tell us what must be the case for all of science. The challenge for philosophy of science is resisting the appeal of naïve abstraction while working at a greater level of generality that the specific sciences.

Thu 26 Feb 2009 01:31 PM


from: Im curious

"If we take as a premise that biology is scientific, then the failure of reduction means that science does not provide a unified picture of that kind." Thus, the picture is still unified but of a different kind, right?

Fri 27 Feb 2009 06:41 AM

from: Administrator (P.D. Magnus)

My wording was deliberate, so as to allow for the possibility you suggest. It is still integrated, and you can call that "unified" if you want.

Fri 27 Feb 2009 08:20 PM

from: Greg

For what it's worth, that weaker sort of unity is all that the folks we'd consider 'classical' logical empiricists were committed to. They (Carnap and Neurath in particular) explicitly say that their notion of 'Unity of Science' is a unity of LANGUAGE, NOT a unity of LAWS. (By 'unity of laws,' I mean that biological laws could be derived in some sense form physical laws.) Neurath, the biggest booster for the Unity of Science movement, even argues against unity of laws.

Putnam and Oppenheim's "Unity of Science as a Working Hypothesis" is the first piece (I know of) written by a major figure working in the logical empiricist tradition that comes out in favor of the unity of laws.

Sat 28 Feb 2009 08:23 AM

from: Administrator (P.D. Magnus)

Greg: As I say, mine is an "imprecise and idealized story." Although I know that the original Vienna Circle guys were more subtle than cartoon history like this suggests, I don't know as much of the details as I probably should. Thanks for the correction.

As a further note: I ended up skipping the section on the disunity of science when I presented the talk, just for reasons of time.

Sat 28 Feb 2009 09:14 AM