disclaimer: I’m just a humble master’s student in physics, so please take all of this with a grain of salt. It may sound like I’m making excuses, but please treat every concrete example in this post as nothing more than a way of coloring how I feel. Also, this post talks as if I knew what physicists collectively think, but some people may say “that’s not it at all…”
Debates over whether imaginary numbers exist seem to flare up on the Internet every so often. I don’t have much to say about the debate itself, but there were a few points in @mod_poppo’s blog post (in Japanese, and bearing the same title as this one) where I, as a physics student, wanted to offer a bit of a defense, or rather some additional context, so here goes.
First, I think the reason physicists cling to the reality of imaginary numbers, or rather complex numbers, is that quantum mechanics, the foundation of modern physics, is formulated in terms of the theory of linear operators on (complex) Hilbert spaces. People have apparently tried to see whether it could essentially be explained with real numbers alone instead of complex numbers, and it seems that doesn’t work1, so many physicists believe that complex numbers are deeply involved in the fundamental structure of the real world[^citation needed]. That’s why physicists can’t help reacting when someone calls imaginary numbers “fake numbers”. Moreover, countless physicists and mathematical physicists have worked hard on the formulation of quantum mechanics (with complex numbers) for many years, so I think it has been made rigorous enough that we wouldn’t be embarrassed to show it to a mathematician2. Maybe that’s why some people talk about it rather confidently on social media.
The delta function, on the other hand, is something we also use a lot, but I think we tend to regard it more as a handy tool, a convenient fiction. I don’t know how well Dirac understood it when he introduced it, but I think present-day students of theoretical physics are mostly aware that the delta function is a mathematically justified object (in the theory of distributions, i.e. generalized functions), and of course not a “function” that is infinite only at the origin. Still, I personally doubt whether the delta function ever truly represents a physical situation3.
If you give a delta function as the initial condition of the time-dependent diffusion equation, a Gaussian comes out as the solution. Physicists might take this as a model of, say, “the diffusion of sugar dropped onto the surface of water”, but that doesn’t mean any physicist seriously believes that “if you drop sugar into Tokyo Bay, the seawater in New York Bay instantly becomes a little sweeter” (I haven’t thought it through properly, but I suspect this problem goes away if you consider a discrete system). The point is, as long as it’s a good approximation of reality, that’s good enough. This sort of thing shows up beyond elementary physics, too. In high-energy physics, for instance, it’s not unusual for a to pop out when you compute some physical quantity. We then regard this delta function as a divergent quantity arising from an integral over an infinite range, and get a finite value by dividing or subtracting a constant (e.g. , where is the volume of the universe, and so on). Mathematicians may find this an utter mess, but what I just gave is probably a heavily abridged explanation; such things are handled seriously by a method called regularization4.
I meant to talk about imaginary numbers and the fire spread to the delta function, but to sum up: I think the way physics students understand and approach imaginary numbers is quite different from the way they approach the delta function. What physics students truly fear is a divergent integral, not imaginary numbers showing up in a theory. And we think fairly seriously about the infinities that appear in our theories. It’s not at the level of “the delta function is infinite at just one point, so this must be an elementary particle, huh.”
Conclusion
Once people learn that the residue theorem exists, I believe they will tremble at its usefulness and stop having conversations at the level of whether imaginary numbers exist or not.
Footnotes
-
Renou, MO., Trillo, D., Weilenmann, M. et al. Nature 600, 625–629 (2021). https://doi.org/10.1038/s41586-021-04160-4 ↩
-
I wrote that, but it may be a bit of a lie. Physicists carelessly use bra-kets and distributions together, but distributions are not elements of an ordinary Hilbert space… There do seem to be attempts to merge the two in a mathematically sound way: http://kir018304.kir.jp/nc/htdocs/?action=common_download_main&upload_id=204 ↩
-
Just to be clear, I’m not questioning the “reality” of the delta function or anything like that. As I wrote above, the delta function is an object explained by the theory of distributions, nothing more and nothing less. The same goes for how I think about imaginary numbers. ↩
-
You might think we only got into this shady business because we took the integration range and the volume of the universe to be infinite, but in fact it’s not that simple. Putting a naive cutoff into such integrals can break symmetries and invariances of the system that we want to hold physically. Problems that never arose in a system with infinitely many degrees of freedom surface once the degrees of freedom are finite. So regularization calls for all sorts of ingenuity. Research on regularization is still ongoing, and it’s an interesting problem. ↩
COMMENTS