Because the screen is so tiny, you could use the crown to scroll the screen much farther in one smooth motion than you could do in one swipe. The crown keeps you from having to do a bazillion tiny little swipes to scroll down any list larger than one screen. They could theoretically let you scroll faster with a swipe than the actual distance your finger moves on the screen to achieve the same affect, but as you said, we've been trained to directly manipulate on-screen objects, so making the screen scroll faster than your finger moves would just feel strange.Understood. But iPhone/Pad has trained us all to directly manipulate on-screen objects. Heck, there's even times I'll try to pinch-zoom a paper magazine.
So, IMHO, the crown seems redundant if it simply lets you achieve the same end-result should you chose not to use the gestures you already know well. Why would anyone do that?
Like others, I do see a future where they eventually make the side of the watch case something you can swipe to simulate a digital crown.