Do Physicists Assume that Electrons Exist?

Scientific realism was the main topic of my PhD dissertation, so I feel that I should say something on the recent debates on Twitter on the existence of electrons, between Philip Goff and Sabine Hossenfelder. Here is a quick comment.

An important aspect for this debate that I wish to mention, which is notably emphasised by Van Fraassen (an anti-realist!), is that accepting a theory (using it to explain, inquire, etc) involves talking as if the theory was true, and this often means picturing situations inhabited with entities such as electrons occupying energy states and the like.

See, as an illustration, the two first paragraphs of an influential article from Cooper (1956) on superconduction, published in Physical Review Letters.

It has been proposed that a metal would display superconducting properties at low temperatures if the one-electron energy spectrum had a volume-independent energy gap of order 6 kT between the ground state and the first excited state. We should like to point out how, primarily as a result of the exclusion principle, such a situation could arise.

Consider a pair of electrons which interact above a quiescent Fermi sphere with an interaction of the kind that might be expected due to the phonon and the screened Coulomb field. If there is a net attraction between the electrons, it turns out that they can form a bound state, though their total energy is larger than zero. The properties of a noninteracting system of such bound pairs are very suggestive of those which could produce a superconducting state. To what extent the actual many-body system can be represented by such noninteracting pairs will be discussed in a forthcoming paper.

The paper talks about metals, and describes a kind of situation that could arise in them. The situation involves electrons and phonons, among other entities, that enter into various interactions. Of course, mathematical calculations are involved a bit later in the same paper, but they are introduced from a “picture” of some phenomena, and the results of the calculations are always interpreted back into this “picture”, in terms of physical properties of the objects involved. For example, the energy of electrons. It is very clear from the text that an electron is not a mathematical structure to be found in a platonic realm, but a particle to be found in a metal (the math is only used to represent the electron). And yes, this “picture” can have measurable consequences (assuming we embed our model into a model of experiment to derive them), but importantly, it does not itself correspond to anything directly observable nor necessarily measurable. The point of the paper is not to predict that we could observe paired electrons in the microscope. It is rather to explain some more directly accessible phenomenon associated with superconduction that would derive from this picture.

Some tweets from Sabine Hossenfelder in the recent Twitter discussion can be read as denying that such a “realist” picture is involved in scientific explanations at all. Maybe this is a misinterpretation on my part, but in any case, this idea is not really up for discussion. It should be uncontroversial. Even anti-realists such as van Fraassen accept this: scientists explain observable stuff by describing unobservable stuff all the time. And what is implausible, in light of this, is a particular kind of instrumentalism that would say: doing physics is only making mathematical calculations in order to predict measurement outcomes. Well no, maybe that’s the core part of experimental research papers, but take any theoretical paper, and you will see that physicists barely do this, they do something different (they propose a "picture"), or at least, interpreting them as doing only this seems far-fetched.

An interesting question is the following: could physicists dispense with “pictorial” representations (or realist talk) of this kind in principle and only rely on math instead?

I personally doubt it, or it would be very cumbersome at best. You could imagine theoretical physicists just doing math, and experimenters just taking the math and interpreting it in terms of measurement outcomes. But how would experimenters coordinate each other in their interpretation? A purely mathematical structure can apply to anything. And how would theorists justify the relevance of their maths if it’s not clear to what it applies? So perhaps theorists should also mention measurement outcomes in their papers? But interpreting the structure of an abstract model in terms of some measurement outcome or other would strongly limit the its scope to very specific experimental situations. That would impair the development of science, the look for new applications for an existing model for instance or its ability to explain anything that happens outside of a laboratory (similar and other devastating remarks against attempts to reduce physics to a mere discourse about observations were famously made in philosophy of science a long, long time ago, but the idea keeps coming up again and again outside of philosophical circles. You all should know this by now! *sigh*).

One way of viewing things is that the seemingly realist description, the physical "picture" (the model that is) mediates between an abstract mathematical structure and some potentially concrete observable consequences (yet foreseeable or not) associated with a class of phenomenon, here superconductivity, without strictly reducing to any of the two. This is a very useful, not clearly dispensable functional role in practice, and arguably, this is where most of the interesting action takes place in theorical science.

A second question that ensues is whether the fact that this “realist” way of talking is the usual way in physics, and that perhaps it is even indispensable, implies that we should all take this “pictorial” representation seriously (if we take scientists seriously) as an actual description of a mind-independent reality, or, to say it differently, that believing in the actual existence of electrons is required in order to explain most physical phenomena. Maybe not when the explanation is tentative, as was Cooper’s theory above at the time it was published, but at least when it becomes well established. This is what the debate on scientific realism is about, and the question here is whether scientific realism is forced upon us by the indispensability or ubiquity of seemingly realist descriptions in science.

Some tweets from Philip Goff could be read as answering yes, and implying that this was a majority view among philosophers, that denying it requires very strong arguments, etc. and again, it might be a misinterpretation on my part, but here, the answer is clearly no, no one is obliged to take this view. There are reasons to be a realist, but not all physicists are realists, not all philosophers either, and there are good reasons to resist being a realist too. Oceanographers still explain tides with models using Newtonian forces of gravitation today, but according to relativity theory, there are no such things, right? Or not fundamentally at least: forces of gravitation are what the effect of deformations of space-time approximately look like in certain regimes. So, we can explain some stuff very successfully by describing stuff that does not really exist. Or take phonons for example: are they real particles? I think most physicists, even the realist ones, would say that it is just a convenient way to talk about mechanical energy transfer in some materials (or something along these lines), but that they are not really particles. Do these things exist? “kind of” is perhaps the best answer. They can be taken to exist for all practical purposes, but that could be description-dependent, or a matter of convenience rather than a robust natural category of stuff. And yet, in scientific papers, physicist do talk as if phonons really existed: the paragraph cited above is one example.

So, do electrons exist in a different (more robust or more fundamental) way than Newtonian forces of gravitation and phonons? Not necessarily, would say a non-realist. That could be a description-dependent or convenient way of describing reality. Perhaps the scientific realist puts too much metaphysical weight on scientific discourse, and this weight is not really there. Perhaps the aim of science is only to find very convenient ways of describing reality in order to make reliable predictions in all circumstances. Or perhaps this aim is to produce a fiction that can explain everything that is observable. Is it possible to actually do science while holding such anti-realist views? Well of course it is. It is even possible to do science without ever bothering about these questions.

I don’t want to rehearse here the details of the philosophical debate on scientific realism, it can be found in many places (you can read my book Modal Empiricism where I develop a non-realist position). I only wanted to make the following points, which I think should be acceptable by everyone in this debate: (1) physicists talk as if they were realists and (2) this does not mean that they are or that they (or we) must be.