This paper outlines two of the dominant ecosystem-management paradigms, one based on the traditional paradigm of ecological risk assessment and the other based on the health paradigm of ecological risk assessment. After showing the central ethical strengths and weaknesses of the traditional paradigm and the health paradigm, the essay argues that the uncertainties surrounding these two paradigms require ecosystem management to focus on default assumptions. The argument is that, just as most of the major battles of risk management are being fought on the ground of stakeholder participation and default assumptions, so most of the important battles of ecosystem management need to be fought on the battleground of stakeholder participation and default assumptions. Otherwise massive scientific uncertainties will compromise the credibility of both ecology and environmental management. Coping with problems from dioxin to nuclear power, the contemporary risk-management community is witnessing a dramatic battle between 'environmental hypochondriacs' and 'industrial cannibals' [Shrader-Frechette, K., 1994. Risk and ethics. In: Lindell, B. (Ed.), Comprehending Radiation Risks. Swedish Risk Academy, Stockholm, pp. 167'182]. The environmental hypochondriacs often argue that only zero risk is ethically and environmentally acceptable. They forget that virtually nothing has zero risk, provided analysis can detect the smallest threats. The industrial cannibals frequently maintain that almost any level of risk is justifiable, provided that the economic benefits are substantial enough. They forget that not everything has a price. In more sophisticated and realistic risk-management circles, a similar battle pits the wholists against the reductionists. The wholists believe that risk is a multiattribute concept. For them, health and environmental risks are complex and extra-scientific, involving not only threats to human and planetary safety but also threats to values to trust, autonomy, sustainability, cultural integrity, equal protection, due process, future generations, and free informed consent [Slovic, P., Flynn, J., Layman, M., 1991. Perceived risk, trust, and the politics of nuclear waste. Science, 254, pp. 1604'1605; Shrader-Frechette, K., 1991. Risk and Rationality. Univ. of California Press, Berkeley]. The wholists believe that we manage health and environmental risks by reducing them and by enlisting the active participation of stakeholders in risk management. The reductionists, following part of the classic National Research Council/National Academy of Sciences 'Redbook' account of risk, maintain that risk is a scientific concept. For them, health and environmental risks must be assessed scientifically, typically as an average annual probability of fatality [National Research Council, 1983. Risk Assessment in the Federal Government: Managing the Process. National Academy Press, Washington, DC]. Following this definition, many reductionists believe that we handle health and environmental risks by educating members of the public who supposedly are irrational, scientifically ignorant, and paranoid [Weinberg, A., 1988. Risk assessment, regulation, and the limits. In: Woodhead, A., Bender, M., Leonard, R. (Eds.), Phenotypic Variation in Populations. Plenum, New York, pp. 126-127] and by having technical experts control risk management.