We find the reply by Kahana, Sederberg, and Howard helpful in clarifying the temporal-context model (TCM) function, in particular with regard to the elimination of the recency effect by a difficult distractor under parameters that still enable long-term contiguity effects to emerge. We agree with Kahana et al. that what matters most to the understanding of memory is the testing of models against actual data, while attempting to maintain the criterion of parsimony. We welcome, therefore, the challenge offered by this exchange, which has produced quite a number of novel predictions. Still, we are not convinced that TCM has been successful in offering a satisfactory account for memory dissociations between long- and short-term recency, that it is able to flexibly discriminate and recall items from different lists, or that it is more parsimonious than is our dual-store model. Our arguments have implications for the wider debate about short-term memory (STM) and long-term memory (LTM).