Starting this topic is both thrilling and challenging, and I welcome the ensuing dialogue. As I was conducting my research on hurricanes and their impacts on the Gulf Coast of the United States, I was driven to (repeatedly) reconsider disasters: what they were, how they were experienced by affected communities, how they were portrayed in the media, and how they came to be understood within the academic community. What follows are some of the questions that emerged from that process.

Initially, I would like to consider what exactly a disaster is. This may seem an overly simplistic question, and I don't think consensus or an exact definition is nearly as useful as whatever dialogue might result from our attempts to answer it, especially if we can drill down to consider some other questions. How are disasters defined? Are these criteria useful? Or what is lost or missing in these designations? According to technical or bureaucratic criteria, the answers are presented as relatively clear. An earthquake above a certain magnitude within a populated area, or a flood event where the monetary damages to a community exceeds a certain dollar amount. Within the context of the U.S. governmental apparatus, disaster is a specific designation that facilitates assistance unavailable until such a declaration (of disaster) is made official. These designations seem bound to specific criteria rather than human experience, and may do well to regularize disaster from an administrative perspective, but they do little to help us better understand the experience of disaster. In this way, I wonder what might be gained with a more granular (read: ethnographic) understanding of disaster? How do we accomplish such a fine-grained view (and without fetishizing or sensationalizing the human tragedy linked to these disasters)?

Much of the way we define disasters, particularly the notion of "natural" disasters as opposed to, for example, industrial accidents, may be overly focused on the intrusion of the natural world into the normal operations of social systems or communities. I join Neil Smith and others in viewing the natural/human-caused dichotomy as a problematic distinction, given the level of technological intervention into our local ecosystems (as part of industry, residential development, and so on), the potential consequences of this environmental modification, and the long-term outlook for ecosystems heavily affected by industry. Looking more closely at technology and industry as it relates to disaster, we can look to any number of industrially linked accidents, including Bhopal, Texas City, or the Deepwater Horizon oil spill, for examples of disasters within a technological system. Given the inherent risks associated with industrial activity (both immediate, in terms of worker/community health, and in the long term, given ecological implications), the designation of disaster seems to be one for which the actual events exceed the predicted risks by some catastrophic margin, which means that these outcomes were in some ways anticipated: just not at the scale at which they were experienced.

In other cases, the anticipated risks of isolated systems might be understood, but the layered effects are not adequately considered. In the case of Hurricane Katrina in New Orleans, there was relatively widespread understanding of the poverty, institutional racism, and social inequality that afflicted the city, as well as the corruption and decay associated with the levee protection system. But it is less clear that anyone anticipated the layered consequences of the intersection of this social/environmental context with a large-scale hurricane event, although in hindsight, the likelihood of significant tragedy seems perfectly obvious.

Previously, I asked about characterizing extant disaster through an ethnographic lens. But as social scientists working on disaster, how much are we to focus on predicting the potential hazards of a given context? And in what forum would this information do anything to prevent or mitigate disaster potential in regions? Or are disasters, by their very nature, events that can be anticipated in the abstract, but whose specific impacts and effects cannot be predicted? (If they could be predicted, wouldn't we be able to effectively prevent most of them?)

Returning to a consideration of disaster as something more than the statistics that describe it or the thresholds it must reach to become known as such, I like Grant Otsuki's initial poking around into the counting and comparison aspects of disaster. Attempts at counting and comparison seem like distinctly human constructions, by which we attempt to regularize a disaster, to make sense of the event, to better understand a tragedy or crisis for which, especially for those directly affected, there is no sufficient explanation. Perhaps what is so destabilizing about disasters is that given our general knowledge of topics ranging from the potential risks of modern industrial society to weather patterns to plate tectonics, we are well aware of the potential for disaster and where they might occur, but despite our best predictive efforts, rarely can we know exactly when.

This leads to another question: what is the timeframe of disaster? Media coverage, federal disaster recovery efforts, and our general attention span often treat disasters as acute events, as specific moments in time with discrete boundaries. Yet communities may have long-term contexts that led up to the experience of the disaster, and the effects certainly have potential to be felt long after media coverage and national attention have waned. I think a challenge to social scientists studying the impacts of disaster is to consider how long a disaster (event) persists, and how we might determine the cutoff point for which outcomes and experiences are no longer considered part of the disaster event. This cluster of questions is easy to pose, but difficult to answer, so I hope this serves as further provocation within this context. In my own research on hurricanes, the timeframes are muddied and indiscrete, given the layered effects of prior storms affecting the region, coastal erosion and land loss increasing the risk of flooding during storms, and the layered impacts of planning and zoning changes that have drastically altered insurance regimes and rebuilding options over the past decades. These timelines do not fit neatly into disaster designation windows, and anthropology is well positioned to help document these experiences. But what we do with this documentation and how we best use it to better understand the past, present, and future is open to debate.