Bill, I read through the Feynman lecture in you link. It uses
wave theory, not quantum mechanics or photons. It uses math that works with electric fields of electromagnetic waves and oscillating electrons. I don't think the word "photon" appears in the whole piece.
The behaviors of electrons assumed in the math is "assume that the atoms are little oscillators, that is, that the electrons are fastened elastically to the atoms, which means that if a force is applied to an electron its displacement from its normal position will be proportional to the force." That is distinctly different from assuming that electrons are limited to specific energy levels by quantum mechanics. And, the electromagnetic waves are modeled as if they have a uniform electric field across the surface of the material, not as discrete packets (phtons) of light that can individually interact with photons.
So, there is no photon absorption by single electrons and reemission of the same energy/direction of photons by electrons in this article.
The problem with light is that we really don't know how it can be a wave that travels through space without some medium to travel in, but we can't find a medium in space. So, we try to think of light as quasi particles, but can't explain how those particles can behave like waves in experiments where they seem to do that.
There are a lot of mathematical models that treat light in different ways (waves vs particles) which work well in some situations but fail in other situations. So, we just accept a "duality" and use experience to tell us which models work for which situations.
That works OK until we try to use those models to understand more things that we don't currently understand. If we extrapolate different models for the same situation, we get different answers that conflict. We really don't have a basis to know which, if any, is right, because we don't have a way to test the models in those particular situations. And that leads to a lot of unverifiable speculation, with people "shooting down" each other's theories using tenants of models that may or may not really apply to those situations.
I think that is what we have going on here. And it is going on elsewhere, too. See
https://www.physicsforums.com/threads/how-does-refraction-work-on-the-quantum-level.499240/ for example. That is just one of many such discussions that come up when you search "quantum mechanics refraction theory".
The problem is that nobody I have found can actually
explain what physically happens and
prove that it happens that way with an experiment. It is
all conjecture, and mostly by people who (like me) are not recognized experts on the subject. Even recognized experts like Feynman can only
assume that light travels along many many paths simultaneously and is only visible due to "constructive interference" at the places we see it. So, that at best get us back to the double slit paradox where somehow what we want to think about as "one photon"
sometimes acts like one particle, and sometimes like a distributed wave,
depending on how we choose to detect it.
So, let's drop this discussion here. There just does not seem to be sufficient understanding or agreement to help with understanding the original question for this thread.