Relying on von Foerster's distinction between trivial and non-trivial
machines, this paper discusses some crucial arguments for a constructi
vist concept of 'empiricity'. The concept 'empirical' is oriented towa
rds (social) knowledge and knowledge-constructing operations instead o
f objects or reality, thus following the strategy of systems theory to
replace the subject/object difference by the difference system/enviro
nment. Consequently, scientific empirical research is located at the l
evel of second-order observation in the social system science.