A picture naming task and a semantic memory verbal questionnaire were given to normal subjects to assess the possible asymmetry between knowledge for non-living and living things. We first examined 60 elderly subjects with low education. Asymmetry between non-living and living things was found in the semantic knowledge questionnaire and living things fared worse. This difference was not explained by discrepancy in item frequency, familiarity or prototypicality. Using the same questionnaire, we analysed difficulty judgements given by younger, better-educated subjects: questions about living things were slightly, but significantly more difficult than questions regarding non-living things. In order to check for a possible sample bias, we submitted another verbal questionnaire with an analogous structure to different judges and replicated the previous results. These findings are discussed with regard to the selective semantic memory deficit for living things observed in patients. We suggest that the cognitive pattern presented by these cases may be linked to, but not fully explained by the greater difficulty living things present for normal subjects.