Centering on dementia, this paper provides linguistic perspectives on causes of inherent difficulty with terminological accuracy. Attention is paid to the interface between the positivist imperatives of clinical evaluation and the relativist interpretations that help make sense of uses of terms across contexts. Two types of reason are explored for why the language generated by people with dementia is sometimes hard to feature and forecast: the theoretical challenges inherent in exploring the language of dementia and the social variables that impact how that language is manifested.