Mystery Number Four - Computational Linguistics and Talking Robots - page 83

The (apple 46) 48 proplet contains the feature [eval: attract] . Assuming that
the corresponding subactivation for the second intersection in 5.5.1 happens
to evaluate the (müsli 80) 82 proplet as [eval: avoid] 16 (not shown), the agent
will use only the first, and not the second, subactivation as a trigger situation
to activate a consequent in the inference chain 5.2.1:

5.5.3 S TORED CONTENT MATCHING CONSEQUENT IN INFERENCE CHAIN

β be hungry K cm β eat food K+1

rule level: pre

Word Bank # #

β get food K+2

down β get α K+3 exec

# #

β locate α

at γ K+4 exec β take α

K+5 exec

# #

β eat α K+6 up β eat food K+7

(Mary 25) eat (apple 46) 48 #

The trigger content (Mary 25) eat (apple 46) 48 matches the consequent
K+6 of the penultimate inference. All other inference parts in the chain do not
have matching contents at this point, indicated by #.
Next, the matching content (Mary 25) eat (apple 46) 48 is used for a sec-
ondary (i.e., extrapropositional) subactivation. It includes propositions which
precede at the agent's content level, and which contain nouns coreferent with
(Mary 25) and with (apple 46) . Pattern matching selects subactivated propo-
sitions which fit antecedents and consequents preceding

K+6 in the
inference chain 5.2.1, resulting, for example, in the following correlation of
inferences and contents:

β eat α

5.5.4 E XTENDING MATCHING CONTENT BY SECONDARY SUBACTIVATION

rule level: β be hungry K cm β eat food K+1 pre

Word Bank: # #

β get food K+2 down β get α K+3 exec

# #

β locate α at γ K+4 exec

(Mary 25) locate apple at cupboard 46

β take α K+5 exec

(Mary 25) take (apple 46) 47

β eat α K+6 up β eat food K+7

(Mary 25) eat (apple 46) 48 #

16 The assumed evaluations reflect the agent's preference for eating apples over eating müsli.