One way is the "closing the scissors" movement, which implies pressing down on the stern-most blade so that it sweeps over the bow of the kayak underwater.
Another way is to think of the paddle shaft as a sort of "parallel bar" (think gymnastics). You grab it with both hands and use it as a reference point to shift your weight and hold your torso fixed while you rotate your hips and legs and, thus, the kayak down into the water.
Which "vision" is more accurate? Or is there a better one?