Equipping Symbolic Frameworks with Soft Computing Features

This paper proposes to have a fresh look on the neural-symbolic distinction by focusing on the strengths and weaknesses of the two antagonistic approaches. We claim that in both worlds, the symbolic and the subsymbolic world, there is a tendency to embrace new methods borrowed from the respective other methodology. Whereas, this seems to be quite obvious from the neural perspective we focus on sketching ideas where soft computing methods are used in classical symbolic, logic-based frameworks. We exemplify this claim by some remarks concerning certain soft computing features of Heuristic-Driven Theory Projection (HDTP), a symbolic framework for analogy-making and concept blending.