Compositional Semantics - Computational Linguistics and Talking Robots - page 126

7.3 Deriving DBS Graphs from Sets of Proplets

The leftmost graph in 7.2.5 is called a part of speech signature, or signature for
short, while the other graphs are called semantic relations graphs, or SRG sfor
short. 11 A signature resembles a phrase structure tree insofar as it uses letters
for nodes, while an SRG resembles a DG graph insofar as it uses English words
instead (7.1.1, 7.1.2). SRGs and signatures are conceptual superstructures on
top of sets of proplets serving as the full-fledged representation of content.
A signature is more general than an SRG because it abstracts away from
the content words by replacing them with letters representing their parts of
speech. 12 An SRG, in contrast, is more intuitive in that it illustrates a content
construction concretely with actual concepts represented by English words.
For the software procedures of DBS, however, the more differentiated proplet
representation of content must be used. This raises the following questions:

7.3.1 O N RELATING PROPLET SETS TO DBS GRAPHS

• What is the nature of the relation between the proplet representation

of a content and the corresponding SRG or signature?
•

Can an SRG or a signature be derived automatically from a set of
proplets representing a certain content?

Consider the content derived in 7.2.3 as an order-free set of proplets:

7.3.2 C ONTENT CORRESPONDING TO The little girl ate an apple.
⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

noun: girl
sem: def sg
fnc: eat
mdr: little
prn: 7

adj: little
cat: adn
sem: psv
mdd: girl
prn: 7

verb: eat
cat: decl
sem: past
arg: girl apple
prn: 7

noun: apple
cat: snp
sem: indef sg
fnc: eat
prn: 7

The first step of turning this set of proplets automatically into an SRG or a
signature is to form all possible unordered pairs of related proplets, i.e., { girl,
eat }, { apple, eat }, and { little, girl }, and to apply the following schemata to
each pair:

7.3.3 S CHEMATA INTERPRETING TRANSPARENT INTRAPROP . RELATIONS
(1) noun / verb

(2) noun \ verb (3) adjective | noun (4) adjective | verb

noun: α
fnc: β

verb: β
arg: α X

noun: α
fnc: β

verb: β
arg: γ X α

adj: α
mdd: β

noun: β
mdr: X α Y

adj: α
mdd: β

verb: β
mdr: X α Y

(5) noun − noun (6) verb − verb (7) adjective − adjective

noun: noun: verb: verb: adj: adj:

α β α β α β