What high-frequency, high-voltage techniques accomplished, long before theory could explain it, was a more selective jostling of electrons. Trial and error, of course, played a large part. It was a search for gases and phosphors that behave mostly as light emitters, and not heaters. Two approaches are possible, however: either delivering a narrow range, only certain frequencies, of excitation energy, trying to couple with fewer electron states; or starting out with an emitting substance that possesses a more limited and desirable range of transitions. The latter choice is often more practicable. In neon, for example, pulsed high-voltage excitation results in a restricted range of re-emitted energy.
Thus, in familiar lighting typesfluorescent, neon, sodium vapor, etc.the object is to create mostly those excited states which drop to lower-energy states in permitted ways that are richer in useful photons (visible light). White-hot filaments, by contrast, are a broader mix of excited states, with a broader range of output. Of course, the efficient technology of "tuning" light can occasionally succeed too well, providing illumination that is deemed harsh or unnatural, precisely because so many frequencies are absent from the mix.