Mystery Number Five - Computational Linguistics and Talking Robots - page 102

In the proplets preceding the arrow, the semantic relation is coordination,
expressed by the attributes prev and next . In the proplets following the arrow,
the semantic relation is functor-argument in the form of modifier

modified, ex-
pressed by the attributes mdr and mdd . The content red circle is characterized
as a phrasal noun, and may be turned into a proposition by adding a suitable
verb.
In addition to the upscaling from coordination only to coordination plus
functor-argument there is the upscaling from nonlanguage content to language
content. The latter is treated as another refinement of the basic proplet struc-
ture. This is shown by the following Word Bank example containing the con-
tent red circles as coreferent language proplets and nonlanguage proplets:

|

6.4.3 C ODING NONLANGUAGE AND LANGUAGE CONTENT ALIKE
member proplets now front owner proplets

... ...

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

sur:
noun: circle
cat: pnp
sem: pl sel
fnc:
mdr: red
nc:
pc:
prn: 37

sur: circles
noun: (circle 37)
cat: pnp
sem: pl sel
fnc:
mdr: (red 37)
nc:
pc:
prn: 62

core: circle

... ...

...

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

sur:
adj: red
cat: adn
sem: psv
mdd: circle
nc:
pc:
prn: 37

sur: red
adj: (red 37)
cat: adn
sem: psv
mdd: (circle 37)
nc:
pc:
prn: 62

core: red

... ...

In contradistinction to the “vertical” matching between the language and the
context component shown in 4.3.2, the above example treats reference as a
“horizontal” matching between language and context proplets within token
lines, in accordance with the refined component structure of a cognitive agent
presented in 4.5.3 and 4.5.4.
The difference between the language and nonlanguage proplets concerns
their respective sur attributes. The language proplets with the prn value 62
have the nonempty sur values red and circles , whereas the corresponding sur
slots in the context proplets with the prn value 37 have the value NIL, repre-
sented by empty space.
The order of content and language proplets in a token line is determined
solely by the order of their arrival. In other words, they may be freely mixed.