Reasoning, judgment and decision making are activities that often occur in a social context, and whose basic tools are connectives and quantifiers such as "if", "or", "some", "possibly", etc. When information is exchanged for the purpose of reasoning, the meaning that is ascribed to these terms is instrumental to the conclusions which are ultimately reached. All quantifiers and connectives have a rather consensual interpretation when used for reasoning, which is not necessarily their logical interpretation. Politeness, however, largely complicates their interpretation, as soon as the information that is being shared has the potential to offend or upset other people. This politeness-induced complication was investigated in a number of experiments, which are surveyed in this talk. The research line started with an experiment on the interpretation of the quantifier "possibly" in doctor-patient dialogues, and was later extended to the interpretation of the quantifier "some" in face-threatening contexts.  Other experiments investigated the interpretation of logical connectives, namely "or" in the context of announcing bad prospects, and "if" in the context of pointing out mistakes of others.  All experiments found evidence for the same phenomenon: (a) Quantifiers and connectives have a rather clear meaning in control contexts, but (b) confusion arises when they are used in face-threatening contexts, because (c) people start considering the possibility that they might be used for politeness purposes.