Introduction

How can we characterize the way scientific modelling works? In Chapter 1, I suggested economists use models as a means of investigation or enquiry: they both enquire into the small worlds in their models, and use those enquiries as a way to interrogate the nature of the world. In this chapter I explore how these kinds of enquiries work by treating model-based reasoning as akin to experimental investigations.

Treating model reasoning as a form of experiment inevitably raises questions about the nature of such experiments. In Chapter 6, I portrayed model usage as a process of asking questions about the circumscribed and limited world in the model and using the model to derive answers about that small world. This is a process in which scientist and model are jointly active participants: neither is passive – the scientist experiments by manipulating the model, that is, he or she uses the model’s resources (both its subject specific and deductive resources) to demonstrate answers to questions of interest to the scientist. In this chapter I show how the heart of the experimental action lies first in the ways a model’s resources are used to demonstrate answers to questions about that model world, and second in using these experimental demonstrations to make inferences from that model world to the real one.