Phenomenology, Representations and Complexity

Abstract. The paper refutes the general phenomenological argument that knowledge cannot be completely represented by symbols and, hence, symbolic AI does not work. Moreover, the viewpoint of algorithmic information theory is introduced for considering the limitations of arti cial intelligence. Based on the formal considerations of the notion of algorithm we will come to new conclusions from the phenomenological critic at the possibility of (symbolic) AI.

• Type: Conference
• Year: 1992
• Where: ECAI
• Authors: Achim G. Hoffmann
• Keywords: Arti Cial Intelligence, Artificial Intelligence, ECAI 1992, General Phenomenological Argument, Symbolic Ai