Disaster relief work is often extremely challenging for those involved. One of the most obvious problems is coordination: coordination with headquarters in the home country, local organizations and authorities, and other teams in the field. Disaster relief is often about having to deal with unanticipated situations – the potential for difficult surprises inheres in the complexity and confusion of disaster-stricken areas of the world. The central dilemma in managing these surprises, though, appears to be this: disaster relief workers either have the knowledge to know what to do (because they are out there in the field), or the authority to do it (because they are in a position in the hierarchy that gives them that authority). Seldom, however, are the local knowledge of what to do and the authority to do it located in the same person. This mismatch between having either local knowledge or authority creates instability. This instability generates pressure for change. Such change occurs through what I describe as renegotiations of authority - where people or teams who are not officially in charge take authority to act, because they know what to do and they know how urgent it is to do it.

This dissertation investigates the renegotiation of authority by interpreting and re-interpreting a large body of qualitative data (hundreds of hours of interviews, observations and participant observations) through a number of cyclical encounters between empirical material, theory and analysis. While the fundamental reasons for renegotiating authority might always be similar (having local knowledge but not authority), not all renegotiations take place between the same parties or along identical trajectories. This research has identified multiple pathways along which authority can travel in order to accommodate local demands for action. Not even the distribution of renegotiations over time appears to be constant: as experience (team leaders recognizing typical problems and knowing what to do) and common ground (team members and leaders knowing each other) accrues, disaster relief workers rely less on formal procedures and hierarchical structures for knowledge of what to do or who to ask. While this is true, not all renegotiations leave the people involved equally happy: further analysis revealed two major, basically exclusive dimensions:renegotiations can be mutually agreed or asserted, and they can be graceful or ungraceful. Individual moments of renegotiations can vary largely independently on these two dimensions, depending on the starting conditions (e.g. are there contingency plans or not). In all renegotiation, there appears no substitute for experience or common ground: these are critical mitigating variables that help renegotiation (and thus disaster relief work) become successful. Directions for training are identified that pull such experiences forward in people's exposure of disaster relief work: both legitimizing the drift away from procedures and structure, and systematizing the conditions under which renegotiation (and thus coordination) will be experienced as problematic or not. Finally, new technology does not necessarily solve coordination problems. In fact, technology can make coordination more difficult (for example by producing data overload), and coordinating via technological artifacts can exacerbate problems of establishing trust between people. Coordination in disaster relief work is not just about getting in touch with each other. It involves issues of hierarchy, experience, history, local knowledge and affective aspects including trust that may be insensitive to technological interventions alone.