You may find some of the quantum optics texts dealing with the Second Quantisation of the Electromagnetic Field a gentler introduction, for example:
R. Loudon, "The Quantum Theory of Light"
Scully and Zubairy, "Quantum Optics"
Dietrich Marcuse (formerly of Bell Labs), "Engineering Quantum Electrodynamics"
These will deal of course with the nonrelativistic electromagnetic field. The justification for all this (at least for the nonrelativistic EM field) is not as lofty as any of your suggestions: one simply recognizes that solutions to Maxwell's equations are superpositions of harmonic oscillators (e.g. superpositions of plane waves with time-harmonic dependence), so one replaces each mode of the EM field with a quantum harmonic oscillator. Quantum harmonic oscillators are easy systems to do many body problems for; if we have a many body Hamiltonian comprising non-interacting QHOs:
$$\hat{H} = \hbar \sum_j \omega_j \left(\hat{a}_j^\dagger \hat{a}_j + \frac{1}{2}\right)\quad\quad\quad(1)$$
We can make the QHOs interact in intuitively clear and simple ways: we simply add a term of the form $\hbar\,\kappa_{\ell,m} \left(\hat{a}_\ell^\dagger \hat{a}_m + \hat{a}_m^\dagger \hat{a}_\ell\right)$ to model an interaction between the $\ell^{th}$ and $m^{th}$ oscillator. The term $\hat{a}_\ell^\dagger \hat{a}_m$ pulls one photon out of oscillator $m$ and put it into oscillator $\ell$, and the interaction terms always hang out in Hermitian conjugate pairs so as to keep the Hamiltonian Hermitian (thus the time evolution operator $\exp(i\,\hbar^{-1}\,\hat{H}\,t)$ unitary) - this is exactly the same as for classical harmonic oscillators where the time evolution operator must also be unitary to conserve energy. If we have a general coupled QHO Hamiltonian:
$$\hat{H} = \hbar \sum_j \omega_j \left(\hat{a}_j^\dagger \hat{a}_j + \frac{1}{2}\right) + \hbar \sum_{j > k} \omega_{j, k}\, \kappa_{j,k}\, \left(\hat{a}_j^\dagger \hat{a}_k + \hat{a}_k^\dagger \hat{a}_j\right)\quad\quad\quad(2)$$
then we can do a simple orthogonal transformation and diagonalise it to the form in Eq. (1), just as we can diagonalised a coupled system of classical harmonic oscillators and find the normal modes. Once we have done the analogous thing in the quantum case, we have an equivalent set of noninteracting oscillators.
Is the whole universe made of oscillators? As far as I understand QFT (and, beyond quantum optics, my understanding is patchy), well the universe is thought of in QFT as being made of fields very like the second quantized EM field that interact. We now don't think of them as nonrelativistic QHOs but more abstractly, where only ladder operators and their particles remain.
The applications part of the chapter is great again though.
– Wolpertinger May 09 '17 at 20:19