This commentary argues for the need to revisit the foundations of attachment theory and its tradition of formulating testable mechanistic models of relationship development and change. Modeling and simulation may be useful to test novel theoretical propositions, such as the one stating that pleasure in parenting may be a determinant of secure father-infant attachment relationships (Brown & Cox, this issue). We discuss this proposition's plausibility, by relating parenting pleasure to the temporal patterning of parenting, a neglected property in parent-child interaction. Simulation work may not only offer first test runs of novel hypotheses, but may also guide empirical researchers to the most likely time-scale on which such hypotheses should be tested.