Must causal models distinguish default from deviant events? Yes, say Menzies (2004, 2007), Hitchcock (2007), Hall (2007), and Halpern (2008), inter alia. No, argues this chapter. It argues that adding defaults into causal models (1) generates complicating and under-constrained unclarities, (2) fails to solve the problems it has been claimed to solve, and (3) fails to fit the most psychologically plausible accounts of how norms influence cognition generally. Instead of adding defaults into causal models, it recommends clarifying the background constraints on what counts as an apt causal model, and attending to background cognitive biases about the availability of alternatives.