-
Notifications
You must be signed in to change notification settings - Fork 4
ErgSemantics_IdentityCopulae
Abrams is Browne.
Browne is a manager.
The reason: the dog.
The theory is that Browne arrived.
The reason: Browne arrived.
The plan is to sleep more.
All Browne could do was arrive.
In contrast to uses of the copula which we analyze as semantically empty, there are cases where the copula verb itself introduces a predicate which (roughly speaking) denotes identity between its two arguments. The simplest cases are the ones where both arguments are expressed as noun phrases, but we also find uses of the copula with clausal complements (both open and closed).
In the simplest cases, such as Abrams is Browne, the two arguments are expressed as NPs and the predicate introduced by the copula describes identity between the referents of the NPs. We also use the same predicate in the analysis of predicative nominals, like Browne is a manager. In another class of cases, the second argument of the copula is clausal (open as in The plan is to sleep more. or closed as in The reason is that Browne arrived.). This is represented by having the label (local top handle) of the clause be the identity predicate's second argument. We sometimes find punctuation (namely `:') as the form of the copula, in both nominal and clausal complement cases as illustrated in the test suite examples.
Finally, we have grouped the do ... be construction (Flickinger and Wasow, 2013) under this category:
- All Browne could do was arrive.
In this construction, both do and be introduce special predicates (_do_v_be and _be_v_do, respectively). _do_v_be takes as its second argument an instance variable representing what was done. This instance is then also the ARG1 of _be_v_do, which like like _be_v_nv takes a handle as its second argument.
There are currently a wide variety of identity copula predications, so we do not have one unified set of fingerprints for this phenomenon. Rather, any of the following represent identity copulae:
_be_v_id[ARG1 x1, ARG2 x2]
[ARG0 x1]
[ARG0 x2]
_colon_v_id[ARG1 x1, ARG2 x2]
[ARG0 x1]
[ARG0 x2]
id[ARG1 x1, ARG2 x2]
[ARG0 x1]
[ARG0 x2]
cop_id[ARG1 x1, ARG2 x2]
[ARG0 x1]
[ARG0 x2]
_be_v_nv[ARG1 x1, ARG2 h2]
[ARG0 x1]
h3:[ARG0 e]
{ h2 =q h3 }
_be_v_do[ARG1 x1, ARG2 h2]
[ARG0 x1]
h3:[ARG0 e]
{ h2 =q h3 }
_colon_v_namely[ARG1 x1, ARG2 h2]
[ARG0 x1]
h3:[ARG0 e]
{ h2 =q h3 }
While the identity copula predications are mostly introduce by lexical items, the same EPs are in some cases introduced by syntactic constructions. In particular, in identity N-bar coordination, as illustrated by the following test suite example:
- My friend and colleague arrived.
the two conjuncts are interpreted as referring to the same entity. This is represented with an id predication taking them each as arguments.
Similarly, in absolutives with NP predicates, we posit cop_id as mediating between the two NPs inside the absolutive modifier:
- With Browne the manager, Abrams arrived.
-
No examples in esd.txt yet --- I put together a list based on the CCS sentences for this phenomenon, but with some modifications.
-
Is there good reason for the colon to introduce a different EP from forms of be?
-
Why no qeq between the ARG2 of _be_v_nv and its complement?
-
I don't yet fully understand what the analysis analysis of do-be. Should both of those verbs really be contentful? Is there a reason that _be_v_do isn't assimilated to _be_v_nv? Is there are reason that _be_v_do does have a qeq for its second argument while _be_v_nv does not?
-
Why is id in identity n-bar coordination not _be_v_id?
-
Why is cop_id distinct from both id and _be_v_id?
- 1214
- FIXME: Flickinger & Wasow 2013
-
ErgSemantics main page
-
Inventory of semantic phenomena (to be) documented
Home | Forum | Discussions | Events