NEWS: Chatbots.org survey on 3000 US and UK consumers shows it is time for chatbot integration in customer service!read more..
This book develops a quite idiosyncratic vision of the strange connection between a thinking man and a computing machine. Both agents can be said to be “rational”, but in different regards. Rationality, in people, relies upon their tendency to anticipate the results of their behaviour. Subsidiarily, they are also in the habit of evaluating the output of other people, or of machines, thus being able to question any computation. One may use a metaphor of “machine rationality”, inasmuch as computers, like most machines, follow a principle of causality. Moreover, software may be designed to reflect traditional inductive and abductive processes, by means of which users categorise any data they are presented with.
The first chapter of the book reminds us of the logical, linguistical and philosophical grounds of our (quite incomplete) knowledge of this realm. It particularly stresses the crucial role of learning processes and of managing imperfect knowledge. The second chapter is an attempt to model machine reasoning as an assembly of basic operations, which are clearly specified. This is a rediscovery of elementary computer science, but, contrary to the dominant paradigm, induction and abduction are viewed as primitives, in accordance with the author’s wish to adequately deal with categorisation processes. In the third chapter, we learn about “symbiotic reason” emerging from the conjunction, within a given context, of both reasoning processes described in the first two chapters. This joint activity is illustrated by the categorisation of formal entities, involving a subtle play of examples, counterexamples and non-examples. Various application domains are presented. A “language of aphorisms” is introduced as a means of communication between rational agents. It is also used as a counterpoint, all along this treatise, since its reader is also viewed as a kind of intelligent agent.