My epistemological papers discuss the demise of positivism and the antiscience inclinations of postmodernism. I draw on recent developments in philosophy of science for an epistemological program that accepts the reality of coevolutionary, nonlinear organizational ontology while at the same time pursuing a modern normal science-based justification logic aimed at improving the truth claims of beliefs about how to improve organizational functioning that are asserted by organization and management researchers.
A number of arguments suggest that organization science has lost its legitimacy with two external institutions, the philosophy of science community and various user communities. Philosophical institutional legitimacy is missing for three reasons:
- Organization science never followed the reconstructed logic of the Received View, whether logical positivism—which no practicing scientists could follow—or logical empiricism;
- Whatever partial legitimacy organization science might have gained from the Received View, or Hstorical Relativism, disappeared when these two epistemological projects were abandoned by philosophers in the 1970s; and
- Organization science seems largely ignorant of the normal science postpositivisms emerging after the collapse of the Received View. Further, a rather active subgroup seems bent on setting postmodernism and other relativist postpositivist epistemologies in place as “guides” for organization studies—this instead of worrying about the lack of legitimacy among external user communities.
Instead of sliding down the anti-science path outlined by the postmodernists, my program rests on bringing organization science up-to-speed in terms of the four postpositivisms that have the attention of current philosophers of science: The Legacy tenets remaining from the Received View; Scientific Realism and Selectionist Evolutionary Epistemology as interpreted for organization science via Campbellian Realism; and the Semantic Conception of Theories. Besides identifying twelve realist tenets organization science should aspire to follow, I add the model-centered definition of effective science promulgated by the Semantic Conception epistemologists. In essence scientific activities become divided into those focusing on:
- Coevolutionary development of the theory–model link and truth-testing for experimental adequacy, that is, testing the ability of the model to test the predictive nuances of the theory, given various conditions; and
- Coevolutionary development of the model–phenomena link and truth-testing for ontological adequacy, that is, testing the ability of the model to refer to or represent real-world phenomena defined as within the scope of the theory.
Empirical tests in organization science typically are defined in terms of a direct “theory?phenomena” corroboration, with the result that:
- We do not have the bifurcated separation of theory-model experimental and model-phenomena ontological tests;
- The strong counterfactual type of confirmation of theories is seldom achieved because the attempt is to predict real world behavior rather than model behavior;
- Model structures are considered invalid because their inherent idealizations usually fail to represent real world complexity—instrumental reliability is low; and
- Our models are not formalized—though this latter criterion may be optional. In organization science there are some formalized models, such as game theoretic, agency, and decision making mathematical or computer models. But most theories are not formalized. If they are, they have little ontological adequacy, and if the testing of counterfactual conditionals is any indication, most have little experimental adequacy either.
Organization science could move to a stronger epistemological footing if it followed the Semantic Conception. Bifurcating activity into theory-model predictions and model-phenomena comparisons would enhance both experimental and ontological adequacy—it would actually make the task of producing more effective science easier. Presupposing that model structures representing a complex real world can be developed, then: (1) Theoreticians could work on developing formalized computational models, both activities of which require technical skills outside the range of many organization scientists; (2) The organization science equivalent of laboratory scientists could work on enhancing model-phenomena adequacy by testing counterfactual conditionals by making and testing predictions; (3) Empiricists could make comparison tests between model and phenomena “within the scope” of the theory and work on generating findings comparing model structures with functionally equivalent structures appearing in the real world without having to worry about testing counterfactual conditionals and making predictions of behavior—somewhat akin to Kaplan’s pattern model.
The package of Campbellian Realism combined with the model-centered Semantic Conception does make effective science a more realistic objective for organization science for a number of reasons:
- A falliblist realist epistemology lowers the standard of truth-seeking from unequivocal Truth with a capital T, to a more approachable human-scale definition of verisimilitude, that is, more truthlike theories remain after the more fallible ideas have been selectively winnowed away.
- A model-centered epistemology that separates the theory–model link from the model–phenomena link makes each activity more manageable, sets up differentiated standards for truth-testing, and allows scholars to become more specialized in one or another side of science, if they wish.
- The new normal science postpositivisms are actually closer to the logic-in-use in organization science than reconstructions following from the Received View, though the standards required for an effective science are still far from being achieved—see the Guttman scale in paper #16.
- An organization science that is more legitimate in terms of the current normal science postpositivisms should produce results that in fact will also increase legitimacy in terms of criteria held dear by user constituencies.
The best way to fend off anti-science attacks by the postmodernists is to develop an organization science that works better because it better meets the institutional legitimacy requirements of both academic and user external communities. This calls for a “marriage” between the postmodernists’ view of organizational ontology and the modern “normal” science developed by complexity scientists, including heterogeneous agent modeling. These themes are elaborated in my papers. A good starting point is paper # 17 because it comes with a 62 word Glossary. An overview of my program—as it extends from logical positivism to agent modeling—appears in paper # 20.