Database Reference
In-Depth Information
fore be processed by the same language-dependent LA-speak rules for lan-
guage production. As an example, consider the inference
eat
food (discussed in Sects. 5.1, 5.3, and 5.6). The connective cm stands for coun-
termeasure:
β
hungry cm
β
4.6.3 I NFERENCE PRODUCING OUTGOING SURFACES
antecedent
consequent
noun:
fnc: hungry
K
β
verb: hungry
arg:
noun: ( K)
β
verb: eat
noun: food
fnc: eat
rule
level
β
cm
fnc: eat
arg:
prn:
( K)
β
food
K+M
prn:
prn:
K
prn:
K+M
prn:
K+M
matching
new content (output)
8
7
and binding
noun: moi
verb: hungry
noun: moi
verb: eat
noun: food
fnc: eat
content
level
fnc: hungry
arg: moi
prn: 211
trigger situation (input)
fnc: eat
prn: 220
arg: moi food
prn: 211
prn: 220
prn: 220
6
synthesis
3, 5, 7
I
would like to eat
some food
surfaces:
The impulse activating this inference is a sensation of hunger provided by the
agent's I/O component (antecedents 3, 5, 7, 8). One way to use the content de-
rived by the consequent (7) would be as a blueprint for nonlanguage action. In
4.6.3, however, the newly derived content is mapped into language, utilizing
the core values for the synthesis of surfaces (6). This kind of language pro-
duction is used especially in dialogue (Chap. 10), which requires and triggers
the agent's reasoning for real-time reactions to the actions of the partner in
communication.
 
Search WWH ::




Custom Search