There is no lack of academic works dedicated to religious vocabulary; however, there has never been an analysis of religious vocabulary in the empirical network of lexical links. An analyzed network arose as the result of an experiment taking advantage of a mechanism of verbal associations that was conducted cyclically on a group of 900 university students of Jagiellonian University (UJ) and of the Academy of Mining and Metallurgy (AGH) in Kraków, Poland. Religious vocabulary is the basis for an empirical subnet of a network of lexical associations. We also used words from the Kent-Rosanoff list as an incentive, as these words were used in an earlier test by Ida Kurcz. This list allowed us to compare the lexical links for such words as "Bible", "religion" and "priest" as used by 1000 people in the test conducted in 1964/1965 (Kurcz 1967) with our results obtained in 2011-2014. This comparison shows a stability of meaning obtained over the course of the past 50 years. At the same time, a comparison of the research results showed the appearance of various stereotypes and conviction systems linked with the referent of the words being researched.