Database Reference
In-Depth Information
Which language proplet refers to which content proplet is coded by addresses
defined as core values. An original may have many coreferent addresses. It
holds that an address always follows the original in a token line.
For example, the coreference between the language and the context proplets
is coded by the address values
(red 37)
and
(circle 37)
in the language pro-
plets. The core value of the language proplets is a concept type corresponding
to the owner proplet, while the core value of the context proplets is usually a
concept token (4.3.3).
The determiner properties of the quantifiers
x(
for all
)
of predicate calculus are recoded in DBS as values of the
sem
attribute, i.e.,
[sem: pl sel]
for
some
and
[sem: pl exh]
for
all
.Thevalues
pl, sel,
and
exh
stand for plural, selective, and exhaustive, respectively. For a set-theoretical
characterization of these values as well as the definition of other determiners
like
a(n), every,
and
the
, see NLC'06, 6.2.9.
In the hear mode, the constellation of coreferential nonlanguage and lan-
guage proplets shown in 6.4.3 comes about by (i) parsing the surface
some
red circles
into connected
red
and
circle
proplets, (ii) turning these proplets
temporarily into patterns by using a variable as their
prn
value, (iii) using the
patterns to search their token lines for matching referents, (iv) replacing their
core values by suitable addresses, and (v) storing them at the
now front
of the
Word Bank. In the speak mode, the context proplets are (i) copied to the
now
front
, (ii) provided with a new (current)
prn
value and addresses for the core
values, and (iii) are passed to LA-speak for language production.
The uniform coding of language and nonlanguage proplets is in concord with
the Humboldt-Sapir-Whorf hypothesis (linguistic relativism, cf. Sect. 3.6), ac-
cording to which a natural language influences the thought of its speaker-
hearers. DBS adopts also the reverse direction by treating natural language
expressions as a direct reflection of thought. This has the following impact on
the computational treatment of reference in DBS:
∃
x(
for some
)and
∀
6.4.4 T
HE
F
IFTH
M
ECHANISM OF
C
OMMUNICATION
(M
O
C-5)
Exporting coordination and functor-argument from language content
to nonlanguage (context) content, coded uniformly as sets of proplets,
allows us to model
reference
as a pattern matching between language
content and context content in the hear mode, and between context
content and language content in the speak mode.
In DBS, linguistic relativism is counterbalanced by the universals of natural
language listed in 1.1.1-1.1.4.
Search WWH ::
Custom Search