I am bothered by the constant talk of "virtual quanta." They are in a way abstractions as the top comment mentions, mainly because there are terms in the pertur

I am bothered by the constant talk of "virtual quanta." They are in a way abstractions as the top comment mentions, mainly because there are terms in the perturbation theory that "look like" photons and leptons that weren't there originally in the starting state. However, the "virtual" part of this is key, you can't actually observe them, so they might not really exist. They do predict changes to the interaction however. The important detail here that drives home how virtual they are is they only exist then when the regime is perturbative, for QED, at large spatial scales, and for QCD, at small spatial scales, hence why you can talk about them in hadrons like protons and neutrons as the author does. In the nonpertubative regimes, for example, strong QED, you cannot use perturbation theory and there are no "virtual photons" you can use. For example, the most basic strong field QED, the hydrogen atom, has no virtual photons in the coulomb field, you just solve the hamiltonian with the q^2/r potential given to you a priori. As the fourier transform of ~ 1/r shows you, you would need an infinite number of photons to make a couloumb field. This problem was already known from classical physics, in which the energy density of a couloumb field including the divergence at the point charge is infinite.