In this paper, we raise issues about discovering and modelling purpose that, in our view, can often be missed within operational research practice. We suggest that, in problem solving, there is a danger of taking too little account of: the differences between espoused goals and goals-in-use; the potentially misleading nature of published goals; goals that express the need to avoid outcomes-'negative goals'; the meaning of goals in an action context rather than the semantics of goal statements; the dynamics and clarity implied by goal relationships; the potential that derives from multi-organisational settings where goals that express an outcome that can only be achieved collaboratively; stakeholder responses to expressed goalsthat good solutions can be sabotaged by others; the fact that some goals are contextually important but not a focus for problem solving because they are 'notour-core-goals'; and the need to design ambiguity of purpose in expressing goals systems. These issues are illustrated through a number of real case examples drawn from engineering, Police, NHS, a Research Institute, and a Utility company/Regulator setting.