Simultaneous Amalgamation - Computational Linguistics and Talking Robots - page 145

8. Simultaneous Amalgamation

A cognitive agent with language has the following sources of content: (i) cur-
rent nonlanguage recognition, (ii) nonlanguage recognition stored in memory,
(iii) inferences which derive new content from current or stored content, and
(iv) language interpretation. These contents, regardless of their source, may
be processed into blueprints which serve as input to the agent's action compo-
nents (Chap. 5), including language production.
Thereby, language interpretation as a mapping from ordered surfaces to
order-free content and language production as a mapping from order-free
content to ordered surfaces are based on a strictly time-linear derivation or-
der. However, no such order restriction needs to be assumed for nonlan-
guage recognition. Let us therefore define an LA-content grammar which ap-
plies the 16 elementary signatures defined in 7.6.4 and 7.6.5 as rules for the
simultaneous amalgamation of content.

8.1 Intuitive Outline of LA-Content

Simultaneous amalgamation in nonlanguage recognition is based on strictly
local and strictly binary relations between two elementary contents (nodes,
proplets), as shown by the following example.

8.1.1 S IMULTANEOUS AMALGAMATION OF CONTENT CORRESPONDING
TO E NGLISH The little black dog found the bone quickly.

3. dog 5. find

1.

dog

N/V

A|N

A−A

little black dog quickly bone

little

2. find 4. find little black

N\V

A|V

bone quickly bone