Introduction: How to Build a Talking Robot - Computational Linguistics and Talking Robots - page 2

The theoretical and computational modeling of this transfer mechanism is
based on the following properties, which are universal 3 among the natural lan-
guages of the world:

1.1.1 U NIVERSALS OF NATURAL LANGUAGE COMMUNICATION

1. The cycle of natural language communication is based on the hear ,the
think ,andthe speak modes of cognitive agents.
2. In communication, expressions of natural language are interpreted relative
to an agent-internal context of use.
3. All natural languages have a time-linear structure, i.e., linear like time and
in the direction of time.
4. All natural languages use the three kinds of sign symbol, index, and name ,
each with its own mechanism of reference.
5. All natural languages use coordination and functor-argument 4 to compose
content at the elementary ,the phrasal ,andthe clausal level.
6. All natural languages distinguish parts of speech, e.g., noun (object, argu-
ment), verb (relation, functor), and adjective (property, modifier). 5
7. All natural languages have the sentential moods declarative, interrogative,
and imperative .

The above universals characterize the agent-oriented approach of DBS, in con-
tradistinction to the sign-oriented approaches of today's linguistics 6 and phi-

losophy of language.
The language communication universals of DBS provide building blocks for
and constraints on the design of a talking robot. Once the natural language

3 The differences between natural languages are comparatively minor (cf. NLC'06, 4.6.1). In this re-
spect, we agree with Chomsky. However, while Nativism formulates the similarity between languages
as an innate Universal Grammar, DBS derives it from a common transfer mechanism.
Regarding Nativism from a history of science point of view, it is notable that this fragmented field
has been sworn again and again to the position that the different schools really all want the same,
e.g., Wasow, Postscript to Sells 1985. If they all want the same, why are there so many different
schools? Perhaps the gravitational pull of the central theory is not sufficient, resulting naturally in a
disintegration into all kinds of variants. This may be because the empirical base of a sign-oriented
approach is simply not broad enough. As an agent-oriented approach, DBS broadens the empirical
base substantially; it is also desigend from the outset to provide computational verification.
4 We use functor-argument throughout for what is commonly called functor-argument structure ,
simply because it is shorter and more proportionate to its sister, coordination .
5 Whether all natural languages have the same basic parts of speech is discussed controversially in
language typology and difficult to decide (Sect. 3.6).
6 See MacWhinney (2004) for an evaluation of the Nativist universals from the viewpoint of cognitive
psychology.