This article assesses the historical foundations of how linearity at low dose became accepted by the scientific/regulatory communities. While the threshold model was used in the 1920s/1930s in establishing radiation health standards, its foundations were challenged by the genetics community who argued that radiation induced mutations in reproductive cells followed a linear response, were cumulative and deleterious. Scientific foundations of linearity for gonadal mutations were based on non-conclusive evidence as well as not being conducted at low doses. Following years of debate, leaders in the genetics community participated in the U.S. National Academy of Sciences (NAS) (1956) Biological Effects of Atomic Radiation (BEAR) BEAR I Committee, getting their perspectives accepted, incorporating linearity for radiation-induced mutational effects in risk assessment. Overtime the concept of linearity was generalized to include somatic effects induced by radiation based on a protectionist philosophy. This affected the course of radiation-induced and later chemically-induced carcinogen risk assessment. Acceptance of linearity at low dose from chemical carcinogens was strongly influenced by the NAS Safe Drinking Water Committee report of 1977 which provided the critical guidance to the U.S. EPA to adopt linear at low dose modeling for risk assessment for chemical carcinogens with little supportive data, much of which has been either discredited or seriously weakened over the past 3 decades. Nonetheless, there has been little practical change of regulatory policy concerning carcinogen risk assessment. These observations suggest that while scientific disciplines are self correcting, that regulatory 'science' fails to display the same self-correcting mechanism despite contradictory data.