Mystery Number Three - Computational Linguistics and Talking Robots - page 61

the semantic relations across token lines (3.3.2), and moving a query pattern
along a token line (4.2.2). Consider the following example:

4.4.3 C OREFERENTIAL NAVIGATION
⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

verb: sleep
arg: Julia
prn: 675

noun: Julia
fnc: sleep
prn: 675

noun: (Julia 675)
fnc: wake
prn: 702

verb: wake
arg: (Julia 675)
prn: 702

1
↔

2
↔

3
↔

Connections 1 and 3 are intrapropositional functor-argument relations between
(order-free) Julia and sleep and Julia and wake , respectively. Connection 2 is
extrapropositional and based on the coreference between the address proplet of
proposition 702 and the original Julia proplet of proposition 675. The content
of 4.4.3 may be realized in English as Julia was asleep. Now she is awake
(see Chap. 11 for the coreferential interpretation of third person pronouns).

4.5 Component Structure and Functional Flow

In DBS, pattern matching based on the type-token distinction 12 is used for the
following applications:

4.5.1 P ATTERN MATCHING BASED ON THE TYPE - TOKEN RELATION

a. Recognition :
matching between concept types and raw input (cf. NLC'06, Sect. 4.3)
b. Action :
matching between concept tokens and concept types (cf. NLC'06, Sect. 4.4)
c. Reference :
matching between language and context proplets (cf. 4.3.3; 6.4.3; NLC'06,
3.2.4; FoCL'99, Sect. 4.2)

Pattern matching based on restricted variables, in contrast, is used to estab-
lish the relation between an LA-grammar rule and a Word Bank content:

4.5.2 P ATTERN MATCHING BASED ON RESTRICTED VARIABLES

a. Natural Language Interpretation :
matching between LA-hear rules and language proplets (3.4.1)

12 The type-token distinction was introduced by C. S. Peirce (1933), Collected Papers , Vol.4:537. Steels
(1999) showed that new types may be derived automatically from similar data by abstracting from
what they take to be variable, and therefore accidental. See also NLC'06, Sect. 4.2, and FoCL'99,
Sect. 3.3.