In the literature there can be found a widely applied distinction between entity realism and structural realism. On its face value, structural realism does not involve entity realism—the only problem is whether there is a minimum requirement as for the construction of the theoretical entities through assumptions in our theories. In other words, is it possible to complete a structurally realist program through models built up in an instrumentalist way? Can our assumptions be simply cooked-up?
I got the most important impulse for scrutinizing this problem from Kevin Hoover’s Milton Friedman’s Stance: The Methodology of Causal Realism (2009). On this present occasion it is not possible to look into the alleged instrumentalism of Friedman, so here I can only drop a hint to my surmise. According to this, the most problematic part of F53 is the long paragraph where Friedman writes about the assumption of leaves rationally maximizing their utility. For Hoover, the way the assumptions regarding entities are constructed is subordinated to the purpose of causal realism. The assumptions underlying a model must be designed in order that the resulting theory could be capable of highlighting the relevant causal mechanisms. Elsewhere (in Galbács 2017, forthcoming in Acta Oeconomica, 67(2)) I called attention to the fact that presumptions generally sorted under the label of “unrealistic assumptions” may actually come from two distinct strategies, from which the one is purely instrumentalist and the other is the type that Max Weber assigned to neoclassicals. The latter are those ideal-types that are instrumental in highlighting some lines of the real causal structure and that bring to the fore those causal mechanisms the analysis of which is traditionally covered by the neoclassical orthodoxy. However, it is doubtful whether both are compatible with the idea of causal/structural realism, or successful causal realism needs a certain degree of entity realism as a prerequisite. Richard Boyd in his On the Current Status of Scientific Realism supports a naturalistic theory of reference in a similar vein. Its primary purpose is to underlie our strategies of building assumptions as to entities. Accordingly, the contents of our concepts are to reflect the real features of actual entities—which is still not the case of direct descriptive relevance. Even though one may set as a prerequisite the idea that our concepts must refer to the actual attributes of real-world entities, this is not an argument for entity realism involved by structural realism. Is it possible to map the real causal structure even by neglecting the real attributes of the related entities?
The intuitive answer to the question is negative. Distinguishing entity realism and structural realism as a first step and arguing that structural realism requires entity realism as a second step is a remarkable way of highlighting Friedman’s instrumentalism and Lucas’ realism. Doing so, Lucas can be introduced as a structural realist, interested in revealing the causal structure underlying macro-economic phenomena. Dropping adaptive expectations hypothesis and using rational expectations hypothesis instead, Lucas, this is my surmise, opted for entity-realism-based structural realism, since structural realism does not seem to be achievable by an instrumentalist strategy as for entities.