. Previously, scholars have generally ignored any notion of time. Now, we need to make explicit our use of time in understanding disaster. Such an application, I believe, will give us a much deeper understanding on defining disaster, how and why such events unfold, and how various social entities attempt to return to normal after the event. Finally, the use of social time in disaster can provide sociologists a deeper look into understanding key theoretical issues related to social order, social change and social emergence, along with voluntaristic versus deterministic patterns of behavior among various units of analysis.

the field. Such a fresh approach possibly improves a wide range of conceptual issues in disasters and hazards. In addition, such an approach would give us insights on how disaster managers, emergency responders, and disaster victims (recognizing that these “roles” may overlap in some cases) see, use and experience time. This, in turn, could assist with a number of applied issues (e.g., warning, effective “response,” priorities in “recovery”) throughout the process of disaster.

Second, I believe that the concept of entrainment could open new doors for understanding post-impact behavior, or the transition from post-impact to pre-impact (or everyday) behavi

The use of social time may also give us a unique view on understanding slow moving disasters (e.g., environmental events, famines and droughts) when compared to sudden impact events. Social time and social disruption provides tools to define disaster without making assessments (i.e., good/bad) of the event.

In this paper I have shown that the application of social time and social disruption to disaster settings (i.e., pre-impact, impact, post-impact) along can provide a unique way to understand disasters. First, by integrating the ideas of event time, social time, and social disruption, we can develop a foundation for creating an empirically based continuum of everyday life/emergency, disaster, and catastrophe. An implication of this approach is that we use specific social criteria (based upon social disruption via the various time concepts) to define the event. As a result, drawing upon social time and social disruption casts a rather wide net in understanding events what we call disa

It should be no surprise recognizing the problematic-theoretical simi­larities between disaster research and its "half-sibling" (e.g., see Wenger 1987), collective behavior. Both disaster research and collective behavior are at a theoretical standstill (e.g., see Aguirre and Quarantelli 1983), and both still rely upon every-day language rather than a broader scheme for describing events (e.g., see Weller and Quarantelli 1973, McPhail 1992). Works by Kreps and Bosworth (1994) and Dombrowsky (1981; 1987), (and perhaps Barton's 1970 classic work) reflect the theoretical attention disaster research needs to build.

The use of disaster periods provides a useful heuristic device for disaster researchers and disaster managers. These various approaches of disaster Phases give researchers an important means to develop research, organize dates, and generate research findings. Similarly, the use of disaster phases benefits disaster managers in attempting to improve their capabilities. Yet, the current use of disaster phases creates broad definitional problems of the field. I show that the current attempts to describe disaster phases are good heuristic devices, but not effective scientific concepts. Yet, scientific, empirical, and theoretical conclusions are drawn from the use of these Phases.

The next logical step to aid analysis would be to cross-tabulate the temporal periods of "before, during, and after" with functional activities. This type of analysis and consideration could be further extended by including both the unit of analysis and various social categories such as social class or ethnicity. This type of three-dimensional approach would also strongly highlight the idea that disaster phases are multilayered. Overall, not only do different groups and units of analysis experience the phases at different times, but that multiple aspects of time (i.e., objective and subjective; before, during, after a disaster) intermesh with specific activities.

We must differentiate whether the use of any phrase refers to temporal or functional aspects of disaster. They should not have multiple meanings

Both researchers and practitioners have seemed to have socially defined some notion of disaster phases for their use. The researchers' job is to create theoretical clarification.

Yet, as Alexander (1982) cautions, good theory must have components of both deterministic and voluntary behavior at all levels of analysis. We do not live in a strictly deterministic world. Thus, as family cycle researchers have learned, we must move with theoretical caution when embracing phases as part of a theoretical model.

Also, with disaster research having strong theoretical ties with the study of collective behavior(Wenger 1987), and with the field of collective behavior often looking at issues related to social change {e.g., riots, social move­ments), another link between disasters and social change has implicitly

The issue of disaster, and phases and social change is an important consid­eration since the "what is a disaster?" debate draws directly upon whether disasters are a "social problem," {e.g., Drabek 1989; Kreps and Drabek 1996), part of a "reconstructionist" approach {e.g., Stallings 1991), or part of social change {e.g., Quarantelli 1987a). I do not see these important aspects of these approaches as contradictory. Interestingly, these three approaches reflect the three main theoretical issues of sociology (see Hinkle 1980): social statics {generally associated with functionalism), social change {generally associated with conflict theory), and social emergence{generally associated with symbolic interactionism).

Cross-cultural disaster research may also provide further insights regard­ing disaster phases.

As the field of collective behavior highlights, individuals in social settings have different perceptions of reality-social settings are not homogeneous (e.g., Turner and Killian I 987).' Thus, to tap further the mutually inclusive, multidimen­sional and social-time aspects of disaster phases, researchers should draw upon multiple publics and their definition of disaster phases.

Consid� e�g the redefinition of disaster phases based on social time may help us WJtb the broader and more important struggle of defining disaster.

Consid� e�g the redefinition of disaster phases based on social time may help us WJtb the broader and more important struggle of defining disaster

D!saster and hazard researchers have recognized the social time aspect of disasters. Dynes_ (1970) alludes to social time regarding the social consequences of a disaster. Dynes observes that social time: is important because the activities of every community vary over a period of time duri�� �e day, the week, the month, and the year. S�c� patterned acuv1nes have implications for potential damage within thecommurnty, for preventative activity within the commu­�ty, for the inventory of the meaning of the disaster, for the rmm�?1ate tasks necessary within the community, and for the mobilizanon of community effort. (Dynes 1970, p. 63)

The Phases Should Reflect Social Rather Than Objective Time Giddens (I 987), although not the first, makes an important theoretical distinction between social and objective time. Giddens defines clock time as the use of quantified units. Clock time represents "day-to-day" structured activities. Typically, studies refer to disaster phases with hours, days, weeks, or years. Social time, however, is contingent upon the needs or opportunities of a society.

Disaster Phases Are Multidimensional Another component of the mutually inclusive nature of disaster phases is that they are multidimensional

Disaster Phases Are Mutually Inclusive As previously indicated, disaster phases overlap. From a theoretical and applied viewpoint, researchers and practitioners must first recognize that disaster phases are not discrete units.

Second, the manner the field handles the issue of disaster phases actually reflects a larger problem in the field. Specifically, how do we define disaster? Kreps (1984, p. 324) comes closest to recognizing the relationship among disaster phases, the theoretical components of disaster (i.e., social order and social action), and the definition of disaster (primarily in a heading in his paper). Unfortunately, he does not elaborate upon the connection of defining disaster and disaster phases. Thus, recognizing and recasting our notion of disaster phases may actually help the field more precisely understand or define "disaster."

Researchers have at times treated the disaster phases as scientific constructs to order data and for scientific analysis. However, as the organizational and family literature show, assumptions based on life-cycle approaches and assumptions often fall outside the realm of appropriate scientific analysis. Here, the phases within the disaster life cycle fall outside the scientific necessity of well-de­fined, mutually exclusive concepts.

One high-level manager involved with the Federal Response Plan made the following reflection about two weeks after Hurricane Andrew occurred: My feeling is that recovery needs to start day one, or even prior to a disaster. It would be wise to set up a group or task force, or a committee. They get together to gather information as the disaster begins. The potential for fragmentation is enormous. It actually goes back to intelligence, damage information. It is difficult to plan for recovery when you do 001 have a sense for how long it could take. You know, recovery has already begun. FEMA has already issued over one million dollars worth of checks .... Anyway, why not have a recovery unit? That would be cool. They should deal the long term recovery within hazard mitigation. In any event that needs to be happening ftom day one.

Finally, the notation of the disaster phases affected emergency-respond­ers' decisions. The lexicon of the four phases appeared to force disaster managers and responders to think and respond in a linear, separate-category fashion. Thus, this paradigm in the end can hurt effective response

Addinonal �g­recovery research by Phillips (1991) shows that different_ categones of disaster victims exit and enter disaster-housing phases at different nmes. She finds that some special-population groups ( e.g., elderly, Hispanics) take a much longer time to transition from temporary to pennanent sheltenng, and from sheltering to temporary and permanent housing than other popu· lation segments.

Of equal importance, both structural-and non­structural-mitigation techniques would lessen dramatically response needs. In essence, effective mitigation and preparation would lessen response time. Logically extended, effective mitigation and preparation when coupled with an effective response could decrease the time for both short-and long-tenn recovery. Tlus analysis further convinced me of the interconnec­tiveness of the disaster periods.

In summary, the ECGs study from DRC showed me that the use of disaster periods created analytical problems. The categories often over­lapped, different groups perceived and experienced the disaster phases differently, and individuals or groups defined differently the actual or potential event

emergent citizen groups (ECGs) in disasters (e.g., Neal I 984, Quarantelli 1985)

Emergency response and recovery is not a linear process; decisions that are made during the emergency phase will impact the recovery process. In practice, however, recovery often takes place in an ad hoc fashion because key decisions are not part of a strategic program to restore services and rebuild communities. (Dumam et al. 1993, p. 30)

Therefore, if disaster researchers wish to improve the theoretical devel­opment of the field dramatically, I argue that we should reanalyze the current heuristic related to the phases.

In fact, the Functions and Effects Study generated the notion that the relationship between mitigation, preparedness, response, and recovery is not even linear. Rather, some preparedness activities (like educating government officials) could really have mitigation effects; and some recovery activities mitigate against future disasters (like using housing Joans to relocate residences out of a flood plain). The Functions and Effects experts hypothesized at least a cyclical relationship among !hese four phases of disaster activity. (National Governor's Association 1979:108)

Others also allude to the fact that these categories are not mutually exclusive. Haas, Kates, and Bowden (1977) make the following important observation:

Despite a positive approach regarding time models, Stoddard concludes the discussion by saying, a simple or complex time model is not comprehensive enough by itself to integrate completely disaster research. Additional con­structs are required for methodological and theoretical compari­sons and liaisons between findings and the various disaster studies. (Stoddard 1968, p. 12)

Since the beginning of the field, disaster researchers have observed various types of disaster periods. Specifically, different events seem to occur at different times related to a disaster. Also, both academics and practitioners assume that these phases exist, and act as if they do exist Yet, in the last 30 years or so, disaster researchers or practitioners have accom­plished little in defining or refining the use of disaster phases. Yet, as I show in the next two sections, both researchers and practitioners have questioned the use of disaster phases since their initial use.

The National Governor's Association Report

Other empirical studies show that the recovery process is not a simple, linear, or cyclical process. Different units or groups may experience, or perceive that they experience, the different stages of recovery I) at different times and 2) at different rates of time.

Overall, recovery studies suggest that subcategories of the recovery process exist. However, different units of analysis (e.g., individual versus group) or different types of groups (e.g., based on ethnicity or social class) may experience the phases of recovery at differing rates. Thus, patterns, phrases or cycles of recovery are not linear.

Phillips' (1991) analysis of housing following the Loma Prieta Earth­quake confirms these different phases. Also, her study shows that different groups of people, often based upon such factors as social class or ethnicity, go through the phases of housing recovery at different times.

The edited work by Haas, Kates, and Bowden ( 1977) illustrates the complexity of the recovery process. Unlike most other overall codification efforts, the above authors explicitly recog­nize that recovery reflects a complex process. They note that people use several subcategories (e.g., restoration, recovery, rehabilitation, redevelop­ment, reconstruction) to describe aspects of the recovery period.

In Drabek's (1986) more recent codification effort, he modifies the disaster phases. His revision reflects the language of the National Governor's Association's 1979 recommendations (i.e., preparedness, response, recovery, mitigation) of disaster phases.

Stoddard argues that the use of time-and-space models in disaster research

By the 1960s, researchers had studied many disasters to allow codifica­tion efforts (Quarantelli and Dynes 1977). For some (e.g., Dynes 1970) disaster periods refer to a temporal category (e.g., before a disaster strikes, while a disaster strikes, after a disaster strikes). In other cases, the use of the phases may refer to functional activities that may or may not also be embedded with temporal considerations (stocking supplies, search and rescue, responding while the disaster strikes, attempting to recovery from the impact). For example, Barton (1970) combines both functional and temporal considerations of disaster. Yet, these and other writers never fully explored the theoretical implications of using the phases in their research

He adds that this phase could be important with sudden impacts, but not as important with slow-moving impacts.

Disasters, Dynes argues, follow a general temporal sequence despite the agent. Dynes em­ploys these phases to argue successfully for an "all hazards" approach to disaster

Mileti, Drabek, and Haas developed their six categories of: 1) preparedness/adjust­ment; 2) warning; 3) pre-impact, early actions; 4) post-impact, short-term actions; 5) relief or restoration, and 6) reconstruction. They justify these categories by noting that "Numerous researchers have documented how activities and nonnative definitions appear to vary across time and vary greatly among events" (Mileti, Drabek, Haas 1975, p. 9). The six phases serve as a central component of the authors' codification effort (it organizes the book chapters). Yet, the authors do not provide a more specific defini­tion for each category. Other theoretical underpinnings in the book receive much more detailed justification (e.g., collective stress, social nature of disaster).