Computing Perspective in Dialogue - Computational Linguistics and Talking Robots - page 207

a noun. The result of this hear mode derivation equals the content 10.6.2 of the
requestor as speaker, except for the prn value and a new STAR-2:

10.6.5 R EQUEST STAR-2 CONTENT AS A SET OF PROPLETS
⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

⎡
⎣

⎤
⎦

verb: pass
cat: impv
sem: pres
arg: # ashtray
prn: 671

noun: ashtray
cat: snp
sem: def sg
fnc: pass
prn: 671

S: Paris
T: 1930-07-04
A: J.-P. Sartre
R: Simone de B.
prn: 671

Here the conversion from Simone's perspective as the requestor to J.-P.'s per-
spective as the requestee is limited to exchanging the A and R values in the
STAR-1 (= STAR-0) of 10.6.2, as compared to the STAR-2 of 10.6.5. This has
an effect similar to going from toi pass ashtray to moi pass ashtray , though
the imperative content has no explicit subject, and consequently no indexical.
At this point, J.-P. understands the request: the slight imbalance experienced
by Simone has been successfully passed on to him by means of natural lan-
guage. As step 3 of a time-linear request sequence, the requestee must take the
requested action in order to be cooperative. This may be a nonlanguage action,
as responding to Open the window! , or a language action, as responding to
Tell me more about your mother! (Weizenbaum 1965).
In our example, the requestee J.-P. has been asked to perform a nonlanguage
action. Therefore, he is looking around for the means to realize his blueprint
for action, which may be paraphrased as Moi pass ashtray . He recognizes
the ashtray on the restaurant table (we are talking about the year 1930) and
initiates a manipulation sequence which moves the ashtray over to Simone.
This results in J.-P. regaining his equilibrium.
The final step 4 is the requestor recognizing the fulfillment of the request by
the requestee. In our case, Simone puts out her cigarette in the ashtray. This
results in her regaining her balance and closes the sequence.
The main difference between the time-linear sequence of a question and a
request dialogue is that steps 3 and 4 in a question dialogue are normally re-
alized as language action, while the corresponding steps in a request dialogue
may be either language or nonlanguage action. However, because all essential
dialogue operations, i.e., adjustments of perspective and other inferencing, are
performed at the level of content, DBS can handle request dialogues just as
well as question dialogues.
In summary, this chapter has analyzed dialogue as a sequence of elementary
dialogues, using a fictional conversation between Jean-Paul Sartre and Simone
de Beauvoir in the Café de Flore on 1930-07-04 as our example: