In the Philosophy of Science, there is some debate over what extent we can ascribe existence to theoretical entities, like electrons (or neutrinos or other particles). Theoretical entities are ‘unobservable,’ meaning that we can never directly perceive them. Because we cannot directly perceive them, some philosophers, like Van Fraassen, maintain that theoretical entities are really just useful fictions — they are concepts that are only meaningful in the context of an empirically adequate theory. (An empirically adequate theory is one that accurately predicts and organizes relations between observable phenomena.)
I’ll argue that we can confirm the existence of theoretical entities. But this does not entail that we can ‘confirm’ the truth of theories themselves. Before exploring the views of any particular philosopher of this course, I’d like to provide a thought-experiment initially put forward by Dretske. The Martians are planning to invade Earth, but need to do recon first, so they decide to send Martian spies. There are two kinds of spies that they can send, (1) the Martian spy has an invisibility cloak and so can walk Earth unseen by humans, (2) the Martian spy disguises himself as a human (with appropriate makeup and costumes). There’s a wrinkle with (1); though the Martian is invisible, the cloak has certain magnetic properties that attract all metal objects (such that spoons, pans, and so on will attract to and stick to the invisible Martian). In light of this, the Martians decide to go with option (2); despite the fact the Martian is not invisible, there is a powerful intuition suggesting that he is, nevertheless, less detectable. Though the Martian is completely visible, he is seen as a human and does not arouse Earth’s suspicion. Had the Martian worn the invisibility cloak, the humans would be able to detect the Martian presence quite easily — they just have to recognize the strange behavior of pots and pans. The disguised Martian is observable, but less likely to be detected. The point of this example is to show that there are situations where ‘unobservables’ are more easily detected (and the presence of the corresponding object more easily confirmed) than ‘observables.’ The disguised Martian is observable, but less likely to be detected. Consequently, we shouldn’t put much stock in any rigid demarcation between observable and unobservable entities. Rather, it seems like there is no real, useful distinction between observable and unobservable entities.
The cloaked Martian is detectable because of the effect he produces, viz. he attracts magnetic objects. That is, we can infer some causal element that produces the observable effects. If we wish to maintain our notion of cause and effect, then we must accept the confirmation of theoretical entities. Van Fraassen provides arguments against one realist strategy, viz. the ‘inference to the best explanation.’ But inferring the presence of an electron from the streaks it leaves in a cloud chamber (or inferring the presence of a Martian from the unnatural behavior of spoons, pots, and pans) is not an inference to the best explanation. Rather, it is an inference to the best cause. And these two kinds of inferences are importantly different.
Here’s the reason Van Fraassen thinks that inference to the best explanation is illegitimate. An explanation of a phenomenon is not one of the ingredients of the universe. It is not as though the Author of Nature had written down that there should be various things like entities, qualities, and, additionally, more importantly, explanations. Explanations are always relative to human interests, they are not things that are true or false about the world. Our grounds for believing Einstein’s conclusions regarding his photoelectric effect was not because we thought that there was anything particularly explanatory about the photoelectric effect, but rather we believe in it because of its incredible predictive power. It consistently and accurately arranges observable phenomena in a rational way, and allows us to make predictions. It provides no veritable explanation of the phenomena in any way that might be said to be true or false.
But inference to the best cause is not a kind of inference to the best explanation. The two are quite distinct. From observing the existence an effect, it is not illegitimate to infer existence of a cause. If the existence of an effect did not entail the existence of a cause, then we cannot make sense of an ‘effect’ at all. Intuitively, the existence of an effect implies the existence of a cause, for if the cause did not exist, then what is it that brings about the existence of some effect? This is different from inferring to an explanation, for nothing about the explanation (in and of itself) implies that it should be true (which is analogous to existing). I can explain (that is, give an account, tell a story) about how Bob Dylan got to Memphis, if you ask, and it will be perfectly adequate, rational, and sensible. But you cannot infer the truth of my explanation (except on faith in the authority of my exlanations) from the fact that Bob Dylan is in Memphis and I can tell a good story about it.
So what have we said? From the existence of an effect we can infer the existence of a cause. But from the existence of an effect we cannot infer the truth of some empirically adequate explanation or theory. Fortunately these kinds of inferences are not the same. We have also shown that there is no real distinction between observables and unobservable. So observing the existence of an effect does not imply anything about the observability of the cause. We can infer the presence of the cloaked Martian spies the same way that we, by oberving streaks in a cloud chamber, infer the presence of an electron (or some other particle) which caused those observable streaks. Martian spies and electrons have observable, causal effects — and from this we must infer their existence.
But the most direct confirmation for the existence of theoretical entities comes from intervention in the physical world — experimentation. We can actually interact with these unobservable entities to produce new effects, new observable phenomena. We can ‘spray’ electrons to change the charge on a certain system of particles. If electrons were merely a convenient fiction, then it would be difficult to account for these results. If we can manipulate the (unobservable) object to produce real, observable effects, then how could we deny the existence of the object? Van Fraassen might ask, ‘Well whose electron are you spraying? Bohr’s? Stoney’s? Thomson’s?’ The answer is, of course, none of theirs. We are simply spraying the electron, the very same electron that Bohr, Stoney, and Thomson all took themselves to be referring to. The existence of an entity is independent from the truth or validity of the theory of which it is a part. Suppose that Newton is walking through the British countryside when we stumbles across a strange, alien device, an electron emitter. (Of course, he will not recognize it as an electron emitter. He has no conceptual apparatus for handling electrons.) He takes it back to his estate that has a beautiful orchard. Newton begins playing with the emitter (as Newton was notoriously fond of such tomfoolery1), points it his favorite apples (which happens to be made of niobium), and squeezes the trigger. Electrons are emitted from the device, strike the niobium apple, and consequently affect the apple’s charge. When Newton sees the apple drop, he observes that it drops at a rate slower than the rate at which the infamous apple that bonked him on the head had fallen. Newton could not have observed the cause, but he can observe the acceleration-change of the apple. (Of course, Newton will be mystified by this, because he does not know what he has done, nor has any of the relevant concepts like electron or ‘charge.’ It is unlikely that he would actually infer the existence of some electron(s) flowing from the emitter.) The point here is that the existence of entities does not presuppose any corresponding theory or background assumptions about the world. Entities and theories are separate. The electron is just the electron. Bohr’s electron is just the electron in its proper role/context within Bohr’s theory.
(And to really press the point, if an entity owed its existence to the truth of a corresponding theory, then we could never prove the existence of a theoretical entity in the same way that we can never prove the truth of a theory. For confirmation of a theory comes from continuous and diverse affirmations of the consequence of that theory. But the affirmation of the consequent can never entail the truth of the antecedent. [Consequently, theories are only ever disprovable, never provable.] If the truth of a theory cannot be proven, and the existence of a theoretical entity depends on the truth of its corresponding theory, then the existence of a theoretical entity can never be proven.)
The idea of experimentation or intervention is the best evidence we have for theoretical entities. It is the fact that we can manipulate these unobservables to create new, observable phenomena. We use electrons to manipulate our observable world in a way that we otherwise could not. The only way to really make sense of this is to suppose that electrons really are part of the world, and not just some useful fiction of our devising.
Here’s another way to think of it. Suppose there’s something like a duck-rabbit, and initially I think that it is a rabbit. Indeed, I say that there are rabbits and that this is a rabbit. But suppose on closer inspection I realize that actually what I thought were ears are clearly a bill (how could I be so blind before?). So I forswear my previous assertions: there are no rabbits, that is no rabbit. Now I know that it is a duck. What changed? Clearly there was a gestalt-switch, I once applied ‘rabbit’ to the duck-rabbit and now I apply ‘duck.’ My interpretation of the thing that I am looking at has changed. But it is not as thought the thing that I am looking at (the penstrokes on the page that constitute the duck-rabbit) has changed at all. The lines are still there, arranged in exactly the same way. Physically, it is the same as it ever was. The point is this: theory has no effect on what natural kinds of things are out there in the world. The things are simply out there, for us to name them, attribute properties to them, and categorize them as we wish. Our namings, attributions, and categorizations do not determine what is so. Electrons are constituents of the world like the inkstrikes that are constituents of the duck-rabbit. We determine the role that electrons play in our theories, but we do not determine the existential status of electrons themselves. (This is like how we determine the role that the arrangement of penstrokes play in our interpretation of the duck-rabbit, but we do not manipulate the strokes themselves.)
- Factually wrong. Newton hated tomfoolery. ↩