Database Reference
In-Depth Information
ing syntactic-semantic parsing in the hear mode (3.3.1). The
continuation
at-
tributes
fnc
(for functor) and
mdr
(for modifier) will characterize the relations
of the proplet to other proplets in terms of functor-argument, while
pc
and
nc
(for previous and next conjunct) will characterize relations of coordination.
The value of the
book-keeping
attribute
prn
(for proposition number), finally,
will be a number common to all proplets belonging to the same elementary
proposition and also indicate the temporal order relative to other propositions.
3.2 Representing Content
One requirement on proplets as a data structure is that they must provide for
a simple representation of propositional content. This applies specifically to
the compositional semantics of
functor-argument
and
coordination
, intra- and
extrapropositionally. Consider the content of an intrapropositional functor-
argument, represented as a set of linked proplets:
3.2.1 F
UNCTOR
-
ARGUMENT OF
Julia knows John.
⎡
⎣
⎤
⎦
⎡
⎣
⎤
⎦
⎡
⎣
⎤
⎦
noun: Julia
fnc: know
prn: 625
verb: know
arg: Julia John
prn: 625
noun: John
fnc: know
prn: 625
The simplified proplets are held together by a common
prn
value (here
625
).
The functor-argument is coded solely in terms of attribute values. For example,
the
Julia
and
John
proplets specify their functor as
know
, while the
know
proplet specifies
Julia
and
John
as its arguments.
5
A content like 3.2.1 may be turned into a schema by replacing each occur-
rence of a constant with a variable (simultaneous substitution):
3.2.2 T
URNING
3.2.1
INTO A SCHEMA
⎡
⎣
⎤
⎦
⎡
⎣
⎤
⎦
⎡
⎣
⎤
⎦
noun:
α
fnc:
β
prn: K
verb:
β
arg:
αγ
prn: K
noun:
γ
fnc:
β
prn: K
The schema 3.2.2 defines the same semantic relations between
pattern proplets
as does the content 3.2.1 between
content proplets
. A schema matches the
content from which it has been derived as well as an open number of similar
contents. A DBS schema is not just a
l'art pour l'art
linguistic generalization,
but allows using detailed syntactic and semantic properties for efficient high-
resolution retrieval (Sect. 6.5). The matching between a schema and a content
is illustrated below using the schema 3.2.2 and the content 3.2.1:
5
When we refer to a proplet by its core value, we use italics, e.g.,
John
, whereas for reference to an
attribute or a value within a proplet, we use helvetica, e.g.,
fnc
or
know
.
Search WWH ::
Custom Search