"The creation of more therapeutic environments for dementia care may well require substantial change in the traditional roles of client, planner, designer, and researcher; such "shifting balances" may be essential if we are to reconceptualize the ways in which our society defines and provides places for care of the cognitively impaired. Many of these changes in roles, rules, and conceptualization of place were presaged more than 50 years ago by Kurt Lewin in his formulation of action research. Building upon Lewin's model, we identify six conditions necessary for effecting environmental and social change. Projects directed toward the creation of more therapeutic environments for dementia care _carried out by a range of researchers and care providers over the past two decade -- are then briefly reviewed to illustrate and reinforce these six preconditions for environmental change. Lewin argued that the implementation of social change __ whether this be change as ambitious as reducing racial prejudice or as seemingly prosaic as increasing consumption of beef hearts in support of the war effort __ first requires "unfreezing of the situation;" some "additional force" is required to overcome an "inner resistance" to change. In the context of environments for the elderly and infirm, the emergence of public and professional awareness of the enormous social and economic costs of Alzheimer's disease and related dementias has played a powerful role in unfreezing our understanding of what nursing homes are "supposed to be." Further impetus for such change is to be found in a number of "demonstration projects" implemented over the past two decades in which facilities have been consciously created as test beds for exploration and evaluation of philosophical, organizational, and architectural innovations. Widely disseminated in the professional as well as the academic and popular press, these facilities have served as powerful agents for change, presenting other care providers with an enhanced vision of what is possible as well as more specific substantive and procedural guidance. Often associated with such demonstration projects are efforts to effect change in those rides, reflected in both local and national codes and standards, which serve to shape environments for dementia care. Such demonstration projects may be seen as reflective of Lewin's call for a more "integrated" or "systemic" approach to social research. Lewin believed that social change requires consideration of multipIe levels of aggregation [i.e., family, group, social institutions, neighborhood, and community and thus requires cooperation between fields as traditionally autonomous as psychology, sociology, and cultural anthropology. Such an "integrated" or "systemic" approach can be clearly seen in the research of environmental gerontologists such as Lawton and Moos as well as in more recent work on "special care units" for the cognitively impaired, including the conceptual framework which guides the work of our Institute on Aging and Environment. Many dementia care demonstration projects have likewise endeavored to explore the relationships among, as well as the impacts of, change within organizational, social, and architectural subsystems. The next two themes distilled from Lewin's work -- "a spiral of steps" and "group decision making" are perhaps most clearly reflective of his initial conceptualization of "action research' and can be usefully considered together. The notion of iterative cycles of investigation_action_evaluation is clearly central to most models of "action research" or "reflective practice," as is a commitment to a more inclusive and participatory decision making process. Many dementia care demonstration projects have emerged from quite extended and highly participatory planning, programming, and design processes. Most have been the focus of systematic evaluation, in one instance involving feed forward to two subsequent new facilities. Unlike the more traditional positivist model of science prevalent in his era, Lewin recognized the need to link "general" principles emerging from research with "local" circumstances. Diagnosis of local conditions as well as general principles are necessary, Lewin argued, for formulation of an appropriate course of action. Our own efforts to provide design guidance for the creation of therapeutic environments for dementia care has endeavored to reflect and respect this distinction between the "general" and the "local." Beginning with very basic characteristics of people with dementia, we endeavor to formulate broadly applicable "therapeutic goals" [e.g., provision of privacy, continuity of the self]. These goals then serve as the foundation for more detailed "principles for design" with their application to specific local circumstances illustrated through "prototypical designs" and [more frequently in recent years] through actual projects. Finally, Lewin argued that we must add to our traditional focus on research and action a third element __ training. The initiation of meaningful social change, to be carried out by participatory teams in cyclical fashion, requires the training of professionals with skills in group building as well as research and action_taking. The National Alzheimer's Design Assistance Project [NADAP], conducted by our Institute on Aging and Environment and now in it's second 3 year cycle, may be seen as one such effort to link training, research and action. Teams of designers and care providers with a commitment to the creation of new and innovative dementia care facilities are brought together for two days of training, including both didactic and hands_on sessions. The NADAP project has as its goal the nurturance of a next generation of model facilities for dementia care. In the context of facilities for the cognitively impaired, change may be essential as well as inevitable. Perhaps only through a reconceptualiation of the nature of this place type. along with the roles those engaged in their implementation, operation, and evaluation, and the rules by which they are designed and administered, can we adequately shift the balance toward the creation of truly therapeutic environments for dementia care."