-
Notifications
You must be signed in to change notification settings - Fork 4
TheAbbey_Chrysalis2014
This page is here to keep track of topics discussed at the small group mid-year meeting hosted by TheAbbey and held at The Chrysalis in February of 2014.
Though we need not stick to any particular schedule, tentatively we have 9 sessions (in PST):
| Day | Time | Notes |
| Saturday | 10:00 - 11:00 | Morning session |
| 11:00 - 12:15 | Early lunch | |
| 12:15 - 14:45 | (Possibly) intercontinental session | |
| 14:45 - 15:30 | Coffee break / walk | |
| 15:30 - 17:30 | Afternoon session | |
| Sunday | 9:00 - 11:00 | Morning session |
| 11:00 - 12:15 | Early lunch | |
| 12:15 - 14:45 | (Possibly) intercontinental session | |
| 14:45 - 15:30 | Coffee break / walk | |
| 15:30 - 17:30 | Afternoon session | |
| Monday | 9:00 - 11:00 | Morning session |
| 11:00 - 12:15 | Early lunch | |
| 12:15 - 14:45 | (Possibly) intercontinental session | |
| 14:45 - 15:30 | Coffee break / walk | |
| 15:30 - 17:30 | Afternoon session |
Our initial agenda is to investigate the following question in as much detail as time and interest dictates.
How can MRS (and its implementation in the ERG, for concreteness) represent some of the following types of distinctions in an underspecificied manner, where disambiguation is frequently partially but not fully not constrained by syntax?
- Complement vs. adjunct PPs:
- a PP that appears in complement position can usually also be an adjunct. I threw the chocolates on the table.
- maybe: PP attachment more generally:
- syntax puts incomplete constraints on where a PP can attach
- proper name vs. definite noun:
- is "Wall Street" a single proper noun, a compound proper noun, a compound simple noun, something else? does the MRS need to say which?
- roles played by constituents in gapping constructions:
- "John gave me a red present and Mary a yellow one." Is Mary a giver or a recipient?
- binding theory:
- syntax can put constraints on anaphora, but doesn't always
- Francis says: Another phenomena that may be of interest is topics in Japanese, where we may know (or think we know) it is an argument of a verb, but not be sure which argument.
A longstanding general design principle for MRS has been:
The MRS should express all the distinctions that are grammaticalized (i.e. implied by the syntax), but no more.
This is the grounds, for instance, for not including a fine grained word sense inventory in the PRED values. Cases where the syntax partially but not fully constrains some aspect of the meaning that seems atomic from the perspective of current MRS design (such as the above list) make it challenging to apply that principle consistently, however. Two possible explanations are (a) the principle is not clear enough about cases like this, or (b) the current shape of the MRS in these situations is incompatible with following the principle.
We need not discuss every item on that list, and other topics are welcome.
Home | Forum | Discussions | Events