A more simple approach could also yield good results for something like this.
j 下一段next speechk 上一段previous speech
