Thanks for your thoughtful comment Dunja. I'll forward this to the authors. My initial thought is that these are different types of ambiguity. With the verbal model, you are not certain about either the assumptions or the deductive impliications. With a formal model you are. However, with just one formal model, you are still uncertain about the range of conditions that would lead to a similar result, so in some sense, there's ambiguity in the extend to which you can inductively infer a broader conclusion from the one formal model. As Karthik and I have discussed, there's also the ambiguity (especially in complex Agent-Based Models) where you don't really know how all of the interactions are leading to the model output.