-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update 40-3.cgel #29
base: main
Are you sure you want to change the base?
Update 40-3.cgel #29
Conversation
Added POS tags
Removed redundant lemma tags
"by the board of directors" |
"by the board of directors" Mod -> Comp
In this case we think it's reasonable to say that the PP "in an amount approved by the board of directors" attaches to payment rather than compensation. How would we treat this if we needed to attach it to compensation and NOT payment, considering the intervening indirect object PP "to an officer"? Are there cases in which attaching to "payment" would not work? How broadly attested is this? |
Discussed this with @nschneid yesterday: the PP would be coindexed with a gap that modifies "compensation"; "reasonable" modifies the phrase that consists of "compensation" and the gap. (I was mistaken: gaps in CGEL aren't limited to argument positions; other instances of In the present case: I think that if attachment @ "payment" (w/o gap) and attachment @ "compensation" (w/ gap) are both plausible and roughly meaning-equivalent (and if there aren't constituency tests that adjudicate between the two analyses?), we could lean on our default preference for low right attachment to favor attachment @ "compensation". Or we could introduce a new, "minimize nodes" heuristic that trumps the "attach low" heuristic.* [1] https://github.com/nert-nlp/cgel/blob/dec764379cf81f0915c8f81599d803ecc047ce89/datasets/twitter.cgel#L850 *as an aside: I think this would roughly correspond to the two-stage parsing model of Frazier & Fodor (1979) https://philpapers.org/rec/FRATSM-3 i.e. Minimal Attachment >> Right Association. Todo: revisit that paper |
No description provided.