Computing Perspective in Dialogue - Computational Linguistics and Talking Robots - page 195

As the result of a strictly surface compositional reconstruction of the speaker's
literal meaning, the content 10.3.1 represents the perspective of the speaker
J.-P. - except for the prn value 623 , which equals that of Simone's current
STAR.
In the hear mode, the following main perspectives 8 on incoming speak mode
content must be distinguished between:

10.3.2 M AIN HEAR MODE PERSPECTIVES ON LANGUAGE CONTENT

1. The perspective of the hearer as the partner in face-to-face communication.
2. The perspective of someone overhearing a conversation between others.
3. The reader's perspective onto the content of a written text (Chap. 11).

Given that Simone and J.-P. are partners in face-to-face communication, the
correct way for Simone to convert J.-P.'s speak mode perspective is from
I wrote you a letter yesterday. STAR − 1
to her own hear mode perspective as
You wrote me a letter yesterday. STAR − 2
This automatic conversion is based on the STAR-2.1 inference (see 10.3.3
below), which takes a content like 10.3.1 and a STAR-1 as input. The STAR-1
is attributed by the hearer to the speaker. In face-to-face communication, this
is easy because the speaker's STAR-1 and the hearer's STAR-2 correspond in
that their S and T values are the same, and their A and R values are reversed.
The DBS inference STAR-2.1 is defined as follows:

10.3.3 STAR-2.1 INFERENCE FOR DERIVING HEARER PERSPECTIVE

⎡
⎣

⎤
⎦ ⇒

S: L
T: D
A: N
R: O
prn: K

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

noun: moi
fnc: α
prn: K

verb: α
arg: {X moi toi}
prn: K

noun: toi
fnc: α
prn: K

⎡
⎣

⎤
⎦

S: L
T: D
A: O
R: N
prn: K

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

verb: α
arg: {X toi moi}
prn: K

noun: toi
fnc:

noun: moi
fnc:

α
prn: K

α
prn: K

In the output, the speaker's STAR-1 perspective is revised into the hearer's
current STAR-2 perspective by swapping the A and R values N and O ,and

8 Special cases are phone conversations which require the hearer to recompute the speaker's S (pace)
value; when talking across time zones, the speaker's T (ime) value must be recomputed as well.