The first paragraph offers us the traditional explaining for earthquakes, that plate movement causes them, and then notes one fact that doesn't completely mesh with this model: that some areas with heavy plate movement which should have earthquakes have very little seismic activity. We can then expect the rest of the passage to explain this seemingly unexplained fact.
The second paragraph does indeed offer a possible explanation: that there are two kinds of subduction. In the kind of subduction that causes earthquakes the two plates move towards each other in opposite directions, which causes a relatively shallow angle of decline leading to more earthquakes. The kind that doesn't cause as many earthquakes involves two plates moving in the same direction, with the trailing plate moving faster than the leading plate. This causes a steep decline, and leads to fewer earthquakes.
The final paragraph explores a possible implication of this new theory.
