Maxwell "Theoretical entities"

Maxwell's central thesis is that the theoretical entities of science are (insofar as the science is correct) just as real as all the other ("observable") entities.

The first issue is, "Who ever thought otherwise?"

One possible attitude toward theoretical entities is that they are "convenient fictions."

A full-blown philosophical offspring of that attitude is instrumentalism: Mathematical theories are not true, they are instruments (tools, like hammers) useful for predicting what will be observed. The theoretical statements in such theories are no more true or false than is a hammer. They are (merely) useful or not.

There are examples of theories we use instrumentally: most people use the algorithms of arithmetic in that way, and sailors used to use a geocentric astronomy for navigation.

The instrumentalist thinks that every scientific theory is like that.

Another offspring is the received view: according to it, every sentence in which there are theoretical terms is translatable, using bridge laws, to one without. The theoretical terms are nothing more than convenient shorthand. (The translations, however, might be infinitely long.)

All agree that theories, and the theoretical terms in them, are useful. The point of instrumentalism and the rest is to eliminate "pseudoproblems" like "Are electrons real?"

There is another version, similar to a view advocated by van Fraassen, which says that it is perfectly possible that whatever the theoretical terms purport to refer to is real, but it should be of no concern to scientists, since the theories provide no reason for thinking that they are real.

All of these views take observation to play a key role in giving evidence of existence. That is, the only reason we can have, according to such theories, for thinking some kind of thing exists, is that they have been (or at least can be) observed.

What is observable? There are, broadly speaking, two schools of thought on that.

1. Sense data.
2. Medium-sized dry goods.

Maxwell has a very particular problem in mind: he wants to show that electrons are real. He can therefore dismiss phenomenalism (the first view) out of hand: According to phenomenalists, even medium-sized dry goods are theoretical entities, nothing distinguishes them from, say, electrons in the eye of the phenomenalist. Thus, the phenomenalist thinks that electrons are just as real as tables and trees. Maxwell is happy with that.

Maxwell's problem is, surprisingly, with the second view. He asks, "What is observable?"

Actually, he starts by asking, "What is, in principle, observable?" After all, anything that is in principle observable might turn out to be real: think of microbes and microscopes. Things that are in principle observable are named by observational terms, and, even if we haven't made the observations yet, they may or may not be real (as opposed to being mere instrumental terms for which the question of reality does not arise)>

Atoms were for a long time taken to be in principle unobservable, and many, perhaps most, scientists had doubts about whether they were real into the 1920s. It took Einstein's theory of Brownian motion, which makes atoms observable, to convince the holdouts.

There is a lesson here: we have no idea what is "in principle" unobservable. Anything that appears in a theory might someday become observable.

The standard way of pushing back on Maxwell's arguments is that Brownian motion is not an observation of atoms and we do not observe microbes in a microscope. What we see in a microscope, defenders of instrumentalism say is an image, not the thing.

Maxwell notes, and this is a distinctive contribution, that that is a slippery slope: if we don't observe with microscopes, what about magnifying glasses? If not magnifying glasses, what about eyeglasses? If not eyeglasses, how about when looking through a window? If seeing through a window is theoretical, why isn't seeing through air?

Two points follow:

1. Nothing is unobservable in principle.
2. There is no clean, sharp division between what is actually observable and what is not.

Of course, for various particular purposes, we may make such divisions. But that is not good enough for a view according to which the question whether the things theoretical terms purport to describe are real doesn't even make sense or even for a view according to which we can have no evidence for or against the existence of such things.

The slippery slope argument I just rehearsed concerned the means of observation. Maxwell gives another slippery slope argument about the things themselves: molecules range in size from atoms to large macrosopic objects.

Notice that, while Maxwell's argument is not specifically targeted at the observational/theoretical distinction as drawn by the logical positivists, it does seem to show

1. The notion of observable cannot carry the epistemological weight the logical positivists wanted it to hold.
2. There is no fundamental distinction between observation sentences and theoretical sentences.

There is a more ordinary consequence of all this: People frequently think of experimental work and theoretical work as being entirely distinct, in principle, aspects of science. In fact, it take theory to make some things observations (theory of lenses, microscope) and the observations verify other theory, which produces new methods of observation,... . The two are intertwined.

I should have mentioned the past.

-- ShaughanLavine - 27 Jan 2009