Introduction: How to Build a Talking Robot - Computational Linguistics and Talking Robots - page 4

3. from order-free content to ordered surfaces in the speak mode and back in
the hear mode (Chap. 3), and
4. from the STAR-0 to the STAR-1 perspective in the speak mode and from
the STAR-1 to the STAR-2 perspective in the hear mode (Chaps. 10, 11). 7

Given that each kind of conversion is realized in two directions, there are al-
together eight. Thereby, the conversions 2-4 are embedded into conversion 1
between the speak mode and the hear mode, as shown graphically below:

1.1.4 I NTERNAL STRUCTURE OF THE DBS C ONVERSION U NIVERSALS

speak mode hear mode

mod−free mod−dep mod−dep mod−free

order−free ordered ordered order−free

STAR−0 STAR−1 STAR−1 STAR−2

The Conversion Universals combine into a fundamental framework of struc-
ture and function, which forms the backbone of the DBS software machine.

1.2 Computational Verification

The continuous advancement of the overall system, by upscaling the lexicon
and the rules for language and nonlanguage cognition, is an important prac-
tical goal for many applications in human-machine communication. It is also
a method of computational verification for the overall DBS system and the
S LIM theory of language (FoCL'99, NLC'06) on which it is built.
The long history of natural and cognitive science provides many examples of
theories which sooner or later revealed some deep-seated flaw. In such a case,
the theory cannot be fixed without completely uprooting its conceptual and, if
present, mathematical framework, and in consequence much of its empirical
work, which is unfortunate.
Nevertheless, discovering such flaws is necessary for the progress of science
and constitutes a method of verification in the following sense: Theories shown
to have a fatal flaw 8

are verified to be unsuitable for further upscaling, while

7 A STAR represents the parameters of S pace, T ime, A gent, and R ecipient. Cf. FoCL'99, Sect. 5.3.
8 Sect. 2.6; FoCL'99, Sects. 8.5, 19.5.