Conclusion - Computational Linguistics and Talking Robots - page 240

noncommunication components are in such close functional interaction and
at such equal levels of evolutionary development, the simplest approach is to
program the two components essentially alike while paying close attention to
their interaction with each other and with their external interfaces. Designing,
using, and reusing software constructs which are as uniform and simple as
possible is also regarded as good programming practice.

12.4 Semantics

Is DBS really a semantics? Let us answer this question by a point by point
comparison with the reigning queen of semantics, Symbolic Logic.
The most important property common to Symbolic Logic and Database Se-
mantics is that they are both Aristotelian. At the core of the Aristotelian ap-
proach is the use of only two basic semantic relations of structure, namely (i)
coordination and (ii) functor-argument.
In Symbolic Logic, this fundamental insight is realized by (i) proposi-
tional calculus for extrapropositional coordination and (ii) predicate calculus
for intrapropositional functor-argument. These may be extended into (iii) ex-
trapropositional functor-argument 13 and (iv) intrapropositional coordination. 14
Extrapropositional coordination is represented in Symbolic Log by expres-
sions like ((p

r) . They are composed by the syntactic-semantic rules of
propositional calculus (cf. FoCL'99, 19.3.2).
In DBS, the logical connectives are introduced by function words, lexically
analyzed by automatic word form recognition. Called conjunctions, they carry
their semantic contribution as a value inside their proplet representation. In
DBS, the extrapropositional coordination corresponding to f(a)

∧ q) ∨

∧ f'(a') ∧ f”(a”)

is shown by example in 3.2.1 and abstractly as the schema 3.2.6.
Intrapropositional functor-arguments are represented in Symbolic Logic by
expressions like f(a,b) ,where f is a functor and a and b are arguments. These

13 However, the Donkey sentence 11.5.5 shows that the extension to subclauses is not always possible.
Though a lower level defect, it has been regarded as serious enough to spawn a massive body of
literature trying to repair it - within, or almost within, the tradition of predicate calculus.
14 Extending propositional calculus from extrapropositional to intrapropositional coordination also cre-
ates a problem for Symbolic Logic. It arises with examples like All the students gathered; John
and Mary are a happy couple; Suzy mixed the flower, the sugar, and the eggs ; etc., which
intuitively suggest a collective, and not a distributive, reading (see Zweig 2008 for a overview; see
also Hausser 1974; Kempson et al. 1981).
The “mix” problem does not arise in extrapropositional coordination, just as the “donkey” prob-
lem does not occur in intrapropositional functor-argument. Within Symbolic Logic, the “mix prob-
lem” can be solved by proposing new quantifiers and/or connectives, while the “donkey” problem is
caused by failing scope and cannot be solved without uprooting the quantifier-based syntax and the
bracketing structure of predicate calculus. The DBS solution is based on the use of addresses (11.5.7).