Abstract
Recently, it has been suggested that agent-based models (ABMs) can be construed as mechanistic models. It remains to be seen, however, whether they can also provide mechanistic explanations.
I narrow my focus on highly idealized and simple ABMs to see whether the notion of mechanistic explanation breaks down there. I argue that the highly influential account of mechanistic explanation, the model-to-mechanism mapping account (3M), together with the 'more details the better' assumption has to be replaced in the context of highly idealized and simple ABMs.
This is because it is often unclear how to de-idealize certain assumptions, or whether a certain idealization is crucial in the first place. I suggest to view these ABMs as providing hypothetical mechanistic explanations while distinguishing two notions of 'hypothetical.' A hypothetical mechanism might be similar to an actual mechanism, but the two differ in that the former is highly distorted.