Abstract
In this paper, we offer an overviewof available category norms and of methodology of their creation. In the second part of the paper, category norms for 12 categories in Czech are presented (i.e. an alcoholic beverage, a colour, a crime, a four-legged animal, a fruit, a metal, a part of the human body, a relative, a sport, a type of vehicle, a toy, a weapon).
These norms are then analyzed in relation with linguistic frequency and token length. The problems of correlating linguistic frequency which is based on corpus data with associative frequency which is based on category norms are discussed.
Preliminarily, it seems that the members of more constrained categories are in a closer relation to each other and activate each other more strongly than members of more open categories. This can be explained based on the principles of the Spreading Activation Theory of Semantic Processing (Collins - Loftus, 1975).