That's the beauty of Skinnerian conditioning as a model for behaviour. The interpretation, the "why," doesn't matter so much That kind of reductionist thinking has it's limitations, of course, but when training animals, I usually find it more useful to look as objectively as I can at the variables that cause increase or decrease of behaviour. It also removes potentially inaccurate and useless negative associations from the equation, I.e. Saying a horse is "lazy." Maybe he is, maybe he isn't, but if you think of your horse as being lazy and needing a better work ethic, you end up approaching the horse with a more antagonistic disposition. If you think of your horse as not being well enough conditioned to move quickly off the leg aid (many are not) and he needs to learn this behaviour, you can think of this as a teaching problem and be less emotionally invested in the whole "Lazy sod. I'll show him" convictions you see at many barns.
Engaging in equine psychoanalysis, on the other hand, has far more pitfalls and is far more reflective of the person doing the analysing than perhaps the horse.