Accurate Unlexicalized Parsing

8 Distance

1. Why is attachment height hard for a PCFG, and how did Klein and Manning approximate it?
   High and low attachments often use the same rules, and height or linear distance isn’t encoded in labels; marking yield sizes would explode the state space. Indirect indicators instead: BASE-NP marks NPs dominating only preterminals, DOMINATES-V marks nodes containing a verb, RIGHT-REC-NP marks NPs with an NP on their right periphery.
   • Klein & Manning, 2003, Accurate Unlexicalized Parsing (Proceedings of the 41st Annual Meeting of ACL), §8, pp. 429–430
2. How did Klein and Manning’s base-NP marking differ from Collins’s, and why did it matter?
   Collins (1999) also inserted a unary non-base NP above base NPs that lacked one; this erased a useful signal (base NPs are commoner in subject position) and cost 0.37 F1 in isolation. Marking base NPs without the unary gained 0.73.
   • Klein & Manning, 2003, Accurate Unlexicalized Parsing (Proceedings of the 41st Annual Meeting of ACL), §8, pp. 429–430
3. Why is it ironic that PCFGs can’t encode distance?
   Every chart item is a label over a span, e.g. NP[0,5], and a rule combines adjacent spans like NP[0,5] and VP[5,12] into S[0,12]; the span lengths are right there, yet only the labels are used in scoring.
   • Klein & Manning, 2003, Accurate Unlexicalized Parsing (Proceedings of the 41st Annual Meeting of ACL), §8, pp. 429–430