ABSTRACT
This chapter proposes a model aimed to detect and to typify semantic relations in each text. According to Semiotics, each text sets up its peculiar semantic relations. Considered as a process, the text progressively determines the type of semantic relation between terms and their strength. A. J. Greimas first noticed an analogy between the process of stabilization of coherent semantic layers and the probabilistic models in use in Information Theory. Since the co-occurrence of two words in a context is a matter of probability, a second goal of our model is to weigh semantic relations. G. Deleuze and F. Guattari first proposed the notion of abstract machine, distinct from its physical implementation, and identified it with a probabilistic Markov chain. The chapter discusses information on semantic relations in the context of each word, where the semantic relation between terms should be stronger.
