duker, are you implying that no extra work is done when the treadmill is at an incline? If that is the case, then logically you must not receive any benefit from it set to a decline (in which case I must assume that you always place a few cinder blocks under the back end of your treadmill ;-). Of course there's extra work. Who says you have to increase PE for there to be work (again, by your body, not via classical physics)? Here's a little thought experiment: Suppose you are running up a mile long hill that's at a 10% grade at about 10 min pace. That's almost a 530 foot rise in 10 mins, a good bit of extra work. Now, suppose the second time you're running up, this entire hill slowly sinks into a city-sized mass of quicksand at the rate of about 53 feet per minute. Given the huge mass involved, there won't be direct measurable interaction between you running and the hill sinking (i.e., your footplants will add/subtract an infintesimal force to the sinking hill). At the end of 10 mins the hill will have sunk about 530 feet leaving you in exactly the same vertical position, and thus, no net PE gain. Are you saying this would feel no different than running at 10 min pace over flat land?
You seem to be implying the same argument as those who say running on a treadmill is easy because the belt "moves under you". No, you still have to accelerate your body to stay in the same spot relative to the room. With an incline, we're just looking at a vector so many degrees above horizontal, so now you need to react against a vertical as well as horizontal component. You could take this to the extreme and imagine someone climbing up a ladder while the ladder is slowly lowered over the side of a building. If the vertical velocities match, the climber does not change PE, but they certainly are doing work! Look at it this way: relative to the bottom of the ladder, they've increased PE by equal amounts either way.