Questions

1. Make a sample parse for "I sent him the letter", similar to the table in Figure 1 (you don't need to draw the final dependency tree, just describe each step of the parsing process: Transition, Stack, Buffer, A). Relations you can use are {nsubj, root, obj, iobj, det}, ignore punctuation. Here obj means direct object and iobj indirect object.

2. Why is it important for all RIGHT-ARCs associated with an element to occur before a LEFT-ARC?

3. Why can't the arc-standard algorithm be used to parse non-projective sentences? Do you know (or even better: can you come up with) a way how to allow parsing non-projectivities?

4. The Introduction mentions features of four different properties: higher-order, higher-support, dense and sparse. Can you explain each of the four properties. Which properties pairs are mutually exclusive?

5. What do you like and dislike about the paper? Is there anything unclear?