Sign Concept

In the context of a language-learning agent one has to introduce the more general concept of a sign as a framework for all the special sign systems, which one wants to use. The discipline which is at stake here, is called semiotics^5.1 as the general science of signs compared to more specialized disciplines like linguistics, phonetics, psychology of language, etc.

Although the basic ideas of semiotics are as old as philosophy the more modern concepts are connected with names like Ch.S.Peirce (1839 - 1914), F. de Saussure (1857 - 1913), and Ch.S. Morris (1901 - 1979)^5.2. Since the times of these pioneers of semiotics the field of semiotics has grown a lot. A straightforward extension of semiotics is computational semiotics combining semiotics with computation. Starting in the 1960s in the realm of control theory (cf. overview [116]) the applications are widening their scope in the direction of human computer interaction (HCI)[3], text linguistics [116], organizational Semiotics [116], as well as semiotic agents[66]. It was especially Gudwin working with Queiroz who investigated different possibilities to use the sign concept of Peirce within computational semiotics (cf. [115], [117], [118], [119], [120]). Another subject within computational semiotics is labeled language games. Strongly influenced by the talking heads experiments from Steels 1995^5.3 [305], [303], this topic grows very fast (cf. [306], [307], [308], [196], [23], [20]. [231]). Doeben-Henisch has shown, how it is possible to formalize the concept of grounding - which is part of the concept of a language game - within the concept of semiosis of Peirce [67]:124ff.

Stimulated by the mentioned authors we will introduce here a basic characterization of the concept of a sign, transform this concept in a small sign theory and show then how this theory can be implemented in the the concept of a conscious learning semiotics system as described in the preceding chapters.