I was guessing at what Bad Wiggins meant by talking about "stress." I interpreted it to mean approaching the limits of aerobic energy production where the anaerobic contribution ceases to be negligible. Blood lactate starts to get elevated within about 20 seconds of MP. You can measure this with a cheap testing kit if you think I'm talking out of my .... as you put it. It doesn't follow, however, that aerobic adaptations only happen when you approach the limits of aerobic energy production.
p.s. I pretty much agree with everything else you said regarding energy systems being on a continuum and about other training tools also being effective.
It certainly DOESN'T need to be long and slow. Though you have to be careful not to make it too hard relative to its importance in your total training plan. For marathoners, the long run is so hard (in the specific phase) that other workouts often necessarily end up being extremely easy. For track runners, it may make more sense to do progressions. Usually these progressions will be shorter and sharper than a marathoner's progression. For example, you might do 10 miles at 7 minute pace followed by 3 miles at 5:30 average, whereas a marathoner might gradually progress from 6:30 to 5:45.
I don't think so. Typically there's only one type of quality training that will be present every week for the duration of a mesocycle (whatever it is that you're focusing on). Everything else will usually be less frequent.