On deviant causal chains – no need for a general criterion