Why is the apparent link between cognition and fitness so weak? Could it be that our proposed causal link between cognition, behaviour and fitness/selection is too simplistic and linear? We discuss possible alternative models