This is because operating a mouse-based user interface is a conscious process - one which cannot be fully relegated to "muscle memory." Unlike the keyboard, the mouse scarcely rewards learning at all.
Erik Naggum, in a Usenet post (circa 1997) explains this far better than I could:
"The clumsiness of people who have to engage their brain at every step is unbearably painful to watch, at least to me, and that's what the novice-friendly software makes people do, because there's no elegance in them, it's just a mass of features to be learned by rote... it's sadly obvious that we are moving into a way of working that is predominantly _conscious_, for which I believe the human brain was never prepared. we no longer have the time to let skills sink into the autonomous nervous system, as it were, and even if we try, the criminal in Redmond, WA, has a new, incompatible version out by the time we learned the last version... One of the joys of learning to ride a bicycle is to stop thinking about it -- the feeling that I had successfully programmed my body to master a bicycle at least thrilled me as a kid (except I didn't know the verb "to program")... we need to communicate to users that learning to use Emacs is like learning to ride a bicycle -- it does take some time and effort, it's a worth-while skill to have, and then you never forget. I firmly believe that the novice-friendly software is like giving people several sets of supporting wheels so they won't tilt, but could get moving right away, and then never taking them off, preferring that they keep using them and moving so slowly that they always need them. of course, if you argue that they should remove the supporting wheels to such users, they will, admittedly correctly, argue that they will fall _splat_ on the ground and ruin their three-piece suits. clearly a no-go."
(http://groups.google.com/group/comp.emacs/msg/821a0f04bab918...)