When people discuss the risks associated with low doses of ionizing radiation, central to the discussion is the definition of a low dose and the nature of harm. Standard answers such as "doses below 0.1 Gy are low" or "cancer is the most sensitive measure of harm" obscure the complexity within these seemingly simple questions. This paper will discuss some of the complex issues involved in determining risks to human and nonhuman species from low-dose exposures. Central to this discussion will be the role of communicable responses to all stressors (often referred to as bystander responses), which include recently discovered epigenetic and nontargeted mechanisms. There is a growing consensus that low-dose exposure to radiation is but one of many stressors to impact populations. Many of these stressors trigger responses that are generic and not unique to radiation. The lack of a unique radiation signature makes absolute definition of radiation risk difficult. This paper examines a possible new way of defining low dose based on the systemic response to the radiation. Many factors will influence this systemic response and, because it is inherently variable, it is difficult to predict and so makes low-dose responses very uncertain. Rather than seeking to reduce uncertainty, it might be valuable to accept the variability in outcomes, which arise from the complexity and multifactorial nature of responses to stressors.