The paper investigates wether humans can give reliable information abo
ut themselves, at least in some respects. In a first step the presuppo
sition is discussed that certain concepts of ordinary language which r
efer to mental phenomena do not become obsolete and cannot be reduced
to observable or physiological phenomena as a result of scientific pro
gress. in a second step, the possible inference is critizised that sel
f-observation could be a way to assess psychological processes or stat
es. These abjections, however, do not diminish the validity and the re
levance of knowing and speaking about oneself. rt is argued that a per
son's knowledge about mental states and phenomena as well as the possi
bility to comprehend these phenomena (e. g. to express them verbally)
are constitutive for them. A comprehensive psychology, however, cannot
succeed without these mental aspects of human life.