A causal explanation of physical systems, what we call PCN, is proposed based on the discussion of the contrast between causal and teleological viewpoints. PCN (physical causal network) is a kind of qualitative explanation of physical systems. Its nodes denote labels of physical quantities and its arcs denote physical causalities between nodes. PCNs are generated in a bottom up manner from fragmentary and general pieces of knowledge on physical causalities. The resultant PCN involves not only the essential phenomena for a certain teleological viewpoint but also unexpected (unintentional) phenomena which will be beneficial in the case where PCNs are utilized to synthesis processes. After discussing the temporal and logical aspects of causal relations, we propose a form for representing fragmentary pieces of knowledge on causal relations and a method for generating PCN based on CMS (Clause Management System).