It's natural, once one has collected workouts and divided them into groups - as I described in the last two posts - to switch from asking how the workouts differ to asking why they differ, to ask what does this particular workout get me that this other one does not? Once you formulate an answer to that question, several problems arise.
The simplest theory devised is to divide workouts into cardiopulmonary and neuromuscular. One piece of advice from the 1970's that's based on this still holds, namely "Once per week, run so fast that your lungs burn; once per week run so far that your legs burn." The most popular proponent of this theory today is Brad Hudson, whose book "Run Faster from the 5K to the Marathon" takes 250 pages to explain that simple idea. Because both short sprints and long slow runs don't cause one to breathe hard, he folds them both into "neuromuscular," but does anyone believe that those two are equivalent?
The most popular theory today is any variation of the "energy systems" theory, that started by separating workouts into "aerobic" and "anaerobic." The most popular proponent of this theory is Jack Daniels, who in "Daniels' Running Formula" divided workouts into "repetition," "interval," "threshold" and "easy." He then added "strides," then "marathon pace running" and "fast intervals." In the second edition, in table 2.2, he adds a "10K zone" because of another gap in his theory. Every time there's a workout to add that doesn't fit the theory, the theory gets altered; there's nothing special about runs at marathon pace, unless one's training for a marathon, certainly no physiological rationale for it.
When one compares competing versions of the energy systems theory, one sees the flaw in thinking that something you've given a name has a real existence. Take "anaerobic threshold" for example. Daniels defines a "tempo run" - what others call a "threshold run" - as a 20 minute run done at that one precise velocity where the accumulation of lactic acid in the muscles begins to exceed the ability to remove it, at a point called the "lactate turnpoint," which has never been proven to exist. He then admitted that that velocity was subject to change with terrain, weather, general health and so on. Then he decided that going a little further a little slower was also good, coming up with complicated charts of times and paces.
Daniels' protege' Pete Pfitzinger took a more practical approach. Since one cannot measure lactate level when running (I've heard of attempts to do just that), he says that "threshold" is "15K to 1/2 marathon pace." Daniels has said that it's 5K pace plus 24-30 seconds per mile, 83-88% VO2max, about 88-92% maximal heart rate. That's four different ranges in my case. Both refer to it as "comfortably hard."
Brad Hudson says he uses three different threshold paces (so much for Daniels' one specific pace!), the paces one could race for an hour, for 90 minutes and for 2.5 hours. Other coaches have different definitions, all of which conflict. When you've run for a number of years, you get a feel for what they mean, but you should remember that their definitions are not for anything real.
The inductive switch
A complication of the theorist approach comes with experimentation. Once one is convinced that a theory "makes sense," one forces things to fit the theory. Workouts others do that don't fit the theory are declared a useless waste of time and energy or at best inferior to workouts that fit the theory. The coach thinks in the shorthand of the theory and prescribes a "threshold run," whatever he or she happens to believe that is, because it's a requirement of the theory, not because it's in the best interest of the athlete.
When a theory becomes firmly entrenched, one stops learning. One of the reasons I started running ultramarathons was because no theory could explain how to train for them and I was forced to see things from a new perspective. Of course, now there are ultrarunning theories, too.