Hi Brandon, thanks for your support!
(1) Muscular failure here refers to the point where the lifter cannot complete the concentric (lifting) phase of the exercise through a full range of motion with correct technique. This is the point at which all of the muscle fibers that can be activated (subject to the level of voluntary activation that can be achieved) are activated. I haven’t discussed the use of forced or cheat reps, but I will write another article about these soon.
(2) When we say that the level of neural drive (voluntary activation) is higher in trained individuals, this means that the number of motor units that they can activate voluntarily is larger. So when untrained individuals fail in a set, they fail before they have recruited all motor units (and therefore some muscle fibers do not get stimulated). When trained individuals fail, they are recruiting most motor units, and most muscle fibers are getting stimulated. Over time, beginners gradually increase their ability to recruit more motor units, so they can activate more muscle fibers. In the meantime, this doesn’t matter that much, since they have far greater capacity to grow each stimulated muscle fiber anyway.
(3) The problem here is identifying the nature of the fatigue. We often talk generally about lacking motivation, being tired, or being fatigued, but ultimately, the only thing that matters to the muscle is whether the muscle fibers are (a) activated, and (b) shorten slowly. If the type of fatigue is local to the muscle itself, then this does not affect whether the muscle fibers can be activated. In reality, this is how training with lighter loads “works” for hypertrophy, because it causes a reduction in muscle shortening velocity while simultaneously increasing the number of active muscle fibers. If the type of fatigue is in the central nervous system, then this will lead to muscle fibers not being activated, and therefore not being stimulated. We are still working out what factors are connected with central nervous system fatigue, and the research will probably advance quite quickly in this area over the next few years.