The article addresses the issue of uncertainty in technological knowle
dge. ''Research by accident'' is seen as a central cognitive process t
o generate (unanticipated) knowledge about the characteristics of a te
chnology. In the example of nuclear weapons, scientific programs have
initially led to a sequency of unanticipated discoveries-a process of
research by accident that subsequently reveals properties and characte
ristics of a technology that were not expected initially. The second p
art of the article deals with cases in which uncertainty originates no
t in the absence of knowledge per se, but rather with whether knowledg
e is available in the right form and the right policy context. The exa
mple of energy studies of the 1970s illustrates that uncertainty also
includes cases where the problem at hand deals with the efficiency of
information distribution rather than information generation. There is
a difference between (scientific) information that exists somewhere an
d scientific information that is known in the right context to the rig
ht people (the persons with the capacity to take, or to resist, action
) at the right time.