This paper critically reflects upon policy and research definitions of elder mistreatment in light of the findings of the United Kingdom Study of Abuse and Neglect of Older People that was commissioned by Comic Relief with co-funding from the Department of Health. The study uniquely comprised a national survey and follow-up qualitative research with survey respondents. This paper focuses on the findings of the qualitative component. One focus is the idea of ‘expectation of trust’, with an argument being made that the concept needs clarification for different types of relationships. It is particularly important to distinguish between trust in affective relationships and ‘positions of trust’ (as of paid carers), and to articulate the concept in terms that engage with older people's experiences and that are meaningful for different relationship categories. The qualitative research also found that ascriptions of neglect and abuse tend to be over-inclusive, in some instances to avoid identifying institutional and service failures. We also question the role and relevance of the use of chronological age in the notion of ‘elder abuse’. Given that ‘abuse’, ‘neglect’ and ‘expectation of trust’ are ill-defined and contested concepts, we recommend that although consistent definitions are important, especially for research into the epidemiology and aetiology of the syndrome and for informed policy discussion, they will unavoidably be provisional and pragmatic.