-
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 & Wasow, 2014) 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]
h1:[ARG0 x1]
h2:[ARG0 x2]
{ h3 =q h1, h4 =q h2 }
_colon_v_id[ARG1 x1, ARG2 x2]
h1:[ARG0 x1]
h2:[ARG0 x2]
{ h3 =q h1, h4 =q h2 }
id[ARG1 x1, ARG2 x2]
h1:[ARG0 x1]
h2:[ARG0 x2]
{ h3 =q h1, h4 =q h2 }
cop_id[ARG1 x1, ARG2 x2]
h1:[ARG0 x1]
h2:[ARG0 x2]
{ h3 =q h1, h4 =q h2 }
_be_v_nv[ARG1 x1, ARG2 h2]
h3:[ARG0 x1]
h2:[ARG0 e]
{ h4 =q h3 }
_be_v_do[ARG1 x1, ARG2 h2]
h3:[ARG0 x1]
h4:[ARG0 e]
{ h2 =q h4, h5 =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.
Huddleston & Pullum (2002, p.266) distinguish ascriptive and specifying uses of the copula, where the specifying use (illustrated by the most salient reading of the first example below) ‘defines a variable and specifies its value.’ The ascriptive use in their terminology, on the other hand, relates a property (the complement) to a theme (the subject). They thus group together ascriptive uses of the copula with NP complements and the copula that takes adjectival or PP complements, as in the second example below.
-
The chief culprit was Kim.
-
His daughter is very bright / a highly intelligent woman.
Huddleston & Pullum (2002, p.267) also note that while NP-complement uses of the copula are in general ambiguous between the ascriptive and specifying readings, it-clefts are unambiguously specifying:
-
His first proposal was a joke.
-
It was a joke that he proposed first.
-
What he proposed first was a joke.
Similarly, they note on p.270 that interrogative who can only be used to question specifying uses: Who are they? and on p.271 that personal pronouns as predicative complements of be are necessarily specifying and certain NPs with singular (count) noun heads but no determiner are necessarily ascriptive (She is secretary of the bushwalking club.). The fact that syntax disambiguates in these cases suggests that we might want to create finer-grained representations of the possible interpretation of the copula with an NP complement. On the other hand, assimilating ascriptive NP-complement copula uses to the AP/PP complement cases (where we treat the copula as semantically empty) seems unnecessary.
In a later chapter, Huddleston & Pullum (2002, p.402, fn 35) note that the use of be in examples like Max was Macbeth is a special use ‘that cannot be subsumed under either the specifying or the ascriptive use.’
-
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
Huddleston, R., & Pullum, G. K. (eds) (2002). The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press.
Flickinger, Dan and Wasow, Thomas. (2014). A Corpus-driven Analysis of the Do-Be Construction. In Hofmeister, P. and Norcliffe, E. (eds). The Core and the Periphery Data-Driven Perspectives on Syntax inspired by Ivan A. Sag. Stanford: CSLI Publications.
-
ErgSemantics main page
-
Inventory of semantic phenomena (to be) documented
Home | Forum | Discussions | Events