We show how to use model classes of partial logic to define semantics of general knowledge-based reasoning. Its essential benefit is that partial logics allow us to distinguish two sorts of negative information: the absence of information and the explicit rejection or falsification of information. Another general advantage of partial logic, which we discuss in the first part, is that its meta-theory is very close to the meta-theory of classical logic. In the second part notions of minimal, paraminimal and stable models are presented in terms of partial logic and we show how the resulting definitions can be used to define the semantics of knowledge bases such as relational and deductive databases, and extended logic programs.

Logic Programming (acm D.1.6), General (acm F.3.0), Specifying and Verifying and Reasoning about Programs (acm F.3.1), General (acm F.4.0), Mathematical Logic (acm F.4.1)
Many-valued logic (msc 03B50), Logic in computer science (msc 03B70), Problems with incomplete information (msc 49N30), Logic programming (msc 68N17), Logic in artificial intelligence (msc 68T27), Knowledge representation (msc 68T30), Problems with incomplete information (msc 93C41)
Department of Computer Science [CS]

Herre, H, Jaspars, J.O.M, & Wagner, G. (1995). Partial logics with two kinds of negation as a foundation for knowledge-based reasoning. Department of Computer Science [CS]. CWI.