This article demonstrates that theory-laden perception is a pure fiction of some philosophers of science and does not in fact exist. It shows by examples from L. Fleck that non-neutral or person-bound observation is an important source of scientific knowledge and suggests that we can explain those changes in scientific knowledge that are caused by divergent perceptions of different observers by means of differences in the repertoires of visual concepts of the respective observers. Visual concept is introduced by means of (...) a performance model of perception. It is contrasted with definable or language-generated concept and compared to Wittgenstein's family resemblance of concepts. (shrink)