In this paper, we consider the problem of. answering queries using material
ized views in the presence of negative goals. The solution is carried out b
y "inverting" views and deriving both positive and negative knowledge. In o
rder to derive negative knowledge, we invert conjunctive views with negatio
n into a set of (extended) views which may also have, in addition to negati
on-as-failure, a different form of negation called classical (or strong) ne
gation. We also consider the case of disjunctive views and present. a techn
ique which permits us to infer both positive and negative knowledge. Furthe
rmore, we extend previous techniques for inferring knowledge from views bas
ed on relations with functional dependencies. Finally, we present a prototy
pe of a system developed at the University of Calabria.