Studies that involve getting information from people about other people, including pedigree studies, create obligations to those other people. Defining them as “human subjects,” however, does not solve the ethical problems and will, in some cases, make important lines of study impractical or even impossible. The more important task is to define what the ethical obligations are and how to ensure that they are carried out. Some heuristics based on the study of families with inherited Alzheimer's disease suggest that consultation with family members and attention to confidentiality should more directly address the ethical problems than try to treat families merely as collections of people with individual rights. The individual-based approach will necessarily fail when different family members differ in their judgments about the risks and benefits of disclosing information, which will, in turn, lead to the most restrictive individual controlling what others in the family can disclose. This solution is unlikely to be supported in practice by families or researchers or to be sanctioned by the courts. The current policy of the federal Office for Human Research Protections, based on an interpretation of the definition “human subject,” is incoherent and will need to be changed, preferably through a process that involves broad debate among all stakeholders but most particularly involving members of families being studied.