Last year, UCSD psychology professor Christopher Bryan reported a fascinating finding about word forms and behavior. In an experiment exploring adults' responses to the verb use of "cheat" and the noun form of cheat referring to the person who cheats or "cheater," Dr. Bryan found that when people were directed "not to cheat" or advised "please don't cheat," they were more likely to cheat than if they were advised "not to be a cheater." Here is a link to the abstract of the research report.
More recently, in a collaborative study, Professor Bryan worked with a team focusing on word choice and effects on behavior in young children. Similarly, experimenters were able to get child subjects to help more often by asking them to be "helpers" rather than to "help." Findings such as this lead me to believe that we need to be more attentive than before to the word choices and more specifically, word forms, that we use to promote ethical behavior in our students. It would be interesting to know if the same effects apply to non-native speakers of a language as to native speakers.