Different knowledge disciplines mean different things by the verb “to understand”. For economists and physicists, a domain or a problem is not understood unless and until it is modeled, and often only by a particular type of model. For most economists, for instance, agent-based models do not provide understanding, because they only show sufficient and not necessary conclusions. For mechanical engineers, understanding usually only comes from a physical prototype. For computer programmers, understanding happens through and with the writing of a software programme for the problem. For legal scholars, it arises with and from the writing of a narrative text reflecting on the problem and its issues.
Here is economist and game theorist Ariel Rubinstein on models in economics: