Conclusion - Computational Linguistics and Talking Robots - page 239

According to the consecutive hypothesis, the language ability evolved from
nonlanguage abilities. Theoretically, these could be divided into nonlanguage
communication and nonlanguage non-communication abilities. However,
because the consecutive hypothesis does not provide for a communication
component, language seems to evolve from non-communication abilities,
which is wrong.
The concurrent hypothesis, in contrast, assumes that the communication and
the non-communication abilities co-evolved from the outset. This provides the
two abilities with the same amount of time to evolve. Thus, when language
finally arrives as the last stage of the communication ability, it can rely on
many nonlanguage communication abilities evolved earlier. It can also rely
on a continuous, systematic interaction between the communication and the
non-communication (context) components.
Communication ability is necessary to all living beings at least to the de-
gree that reproduction is ensured. In evolution, reproduction is more impor-
tant than the survival of individual members because reproduction serves the
survival of the species as a whole. Communication is at the heart of evolution
because reproduction is the motor driving evolution, and communication with
conspecifics is at the heart of reproduction.
The concurrent hypothesis is in concord with modern work in ethology
(M. D. Hauser 1996). It suggests that the communication and the non-com-
munication abilities of an animal not only evolve in parallel, but also have
matching degrees of development. Accordingly, the cognitive abilities of a
squirrel to communicate with other squirrels, for example, are just as evolved
as its abilities for locomotion and behavior control. 11
Furthermore, because even nonlanguage communication refers, 12 the com-
munication and the non-communication abilities do not just grow in parallel,
but are continuously in close interaction. In DBS, this interaction is modeled
as a collaboration between the language and the context components. Natu-
ral language signs and contextual data interact in a software-mechanical way
during interpretation in the hear mode and production in the speak mode. In
short, designing the DBS robot in an evolutionary manner is another way of
broadening the empirical base.
The above considerations regarding the evolution of natural language have
a direct effect on the software design of DBS: if the communication and the
9 Sects. 6.1-6.3.
10 In addition to communication between conspecifics there is also the communication between different
species, for example, between a plant and an insect for pollination (symbiosis).
11 It would be thrilling to know what the little critters all have to communicate about.
12 For example, a warning call makes reference to a predator.