When faced with a potential threat, the immune system has to answer two questions. The first is "shall I respond?". If the answer is "yes", then the second question is "what kind of response should I make?" Burnet's self-non-self model suggested that the immune system answered "yes" to the first question if the potential threat included foreign molecules. Janeway's Pattern recognition model suggests that the immune system responds if the potential threat includes conserved molecules from bacteria, and the Danger model suggests that immune responses occur when damaged tissues send alarm signals. None of these models cover the second question, namely how the immune system determines the effector class of a response (for example, does it generate killer cells? Or IgE? Or IgG? Or activated neutrophils vs eosinophils?) However, 15 years of working with the Danger model led me to the possibility that this decision is not made by the immune system, but instead by the tissues in which the response is needed. In fact, this newly expanded Danger model suggests that most immunological decisions are not made by the immune system itself, but by the tissues that the immune system was designed to protect. It is a damaged tissue that sends alarm signals that initiate immune responses. It is a healthy tissue, expressing its antigens in the absence of alarm, that induces peripheral tolerance, and it is a tissue that determines the effector class of a local immune response.