Theory Construction as Disciplined Imagination

Critical Analysis

In this text, the author describes the qualities that make up a strong theory, focusing on the process of producing the theory itself.

Trivial theories often result from priority placed on validation instead of usefulness, and this has a negative impact on the entire idea of theorizing: it minimizes the value of “imagination, representation, and selection” that are fundamental to the theorizing process. More specifically, “theorizing consists of disciplined imagination that unfolds in a manner analogous to artificial selections”—it is made up of problem statements, generated thought trials, and then narrowed through selection criteria.

Theory is a dimension, not a category (however, many poor theories are categorizations and descriptions). Theory itself is a set of assertions about a general behavior that seems to hold true against many specific instances. It is ordered, generic, and holds against a range. A more general theory holding against a larger range is “more deserving of the label theory.” Good theories are also plausible, and that is judged to be true if it is interesting, novel, unexpected, narrative, aesthetically pleasing, or corresponding with things already believed to be true.

There are many different views of theory construction. One is when a relationship is proposed between at least two properties, and can be determined by experience (Homans)—it must emphasize connections and interdependencies. Another is through inductive abstraction, or by developing predictions in a hypothetical system (White, MacKenzie, Bourgeois).

Bourgeois presented a seven-step process towards building middle-range theories, including “partitioning the topic under investigation, method of theory construction, review of literature, construction of theory-induction from empirical base, extension of theory-deduction into propositions, metaphysical elaboration, and conclusion.” The unique part of this approach is that literature review, construction of theory-induction, and extension of that theory deduction into propositions occurs at once. Bourgeois concluded with ways of working that include “read some of the old masters,” “ground your theory in data,” and “take advantage of serendipity.”

Another approach is to theorize through trial (conjectures) and error (refutations) of ideation (Campbell). Learning is the sum of that process, and theorizing is then a selection of a few constructions that “refer more competently to their presumed ontological referents.”

These approaches are mechanistic, and leave little room for intuition, serendipity, and creative quality; they also do not provide detail on the selection process of a good or bad theoretical thread to pursue. Additionally, many emphasize sequential thinking at the expense of parallel processing, and all view theorizing as equated to problem solving, while other reasons to theorize are ignored (like elegance, usefulness, and simplicity). An alternative view is theorizing as sensemaking (Astley, Dubin). But this is problematic for theorists because “the correspondence between concepts and observables is so loose, the system being studied is open rather than closed, and because the dissemination of earlier sensemaking alters the relationships that theorists are currently trying to order.”

Theory building is a process of building and interpreting imaginary experiments; it is “artificial selection,” as compared to natural selection. It begins with a context, often formed through ethnographic research, and the context suggests a problem. Different conjectures try to simulate scenarios to explain the problem, and selection criteria for any given conjecture is based on if it is interesting, plausible, consistent, or appropriate. The process is guided by imaginary representations, which are selected, accumulated, and rejected. Problems of theory are “more likely to be solved when the problem is represented more accurately and in greater detail with assumptions made more explicit, as a greater number of heterogenous variations are generated, and as more selection criteria, of greater diversity, are applied more consistently to the variations that are generated.” Middle range theories solve problems that have only a limited number of assumptions, and the problem is well-structured and detailed.

The conjectures that are developed can be viewed as thought trials. Some of these are then selected to advance, and the author states that “self-conscious manipulation of the selection process is the hallmark of theory construction.” Quantity in exploration is important, as the more criteria (and the more diverse the criteria) that are pushed against the conjecture, the higher the possibility that a strong and defensible theory will result. Validation is not important for selection criteria, and that means that a contribution of theory is one that offers relationships and connections that were not previously noticed. Plausibility is one of the most common ways of assessing a theory, and this typically emerges through the ideas of being interesting (as compared to absurd, irrelevant, or obvious). Interesting implies that a conjecture has been compared to a previous understanding, and the understanding has been challenged or found lacking. The idea that something is obvious is just as important, as it allowed a theorist to reject a theory: it adds to selection criteria and the selection process.

In addition to being interesting, theories can also be connected, or can imply connection between events. Believability signals that a conjecture has narrative appeal: a theory is more likely to be retained if it strengthens an element of a story, or supplies something missing.

Theory construction requires imagination, “disciplined” by artificial selection. There are moments to create, modify, and improve theory across the entire process of creation: when the problem is stated, when thought trials are imagined, and when criteria is applied. Theorists depend on metaphors, pictures, maps, and diagrams, as these are “compact descriptions of complex phenomena.” Conjecture emerges, “preserved in well-crafted sentences,” that are then tested by people who have stake in the outcome. The author concludes by naming this process—it is a “simulated evolutionary system” that allows for artificial creation and artificial selection.

Research Value

This is a great text, and both valuable and immediately applicable to my work.

First, his theory development description makes sense, and is so far the only time I’ve read something that made it both understandable and concise. Theory comes from identifying a problem, imagining many solutions, and selecting a solution. The simple path unlocks lots of questions, like—what is a problem? What is a solution? How do you select?—but that’s the work, really.

He also introduces some new words to explain theorizing, which are words more familiar to me and more helpful. A conjecture makes it clear that an initial theory (and even a refined one) is a thoughtful guess, stated with some assertion. Imaginary makes it clear that theorizing is happening in and a result of the imagination, and that’s a lot more understandable to me than the arguments around if the “answer is in the data” or not, although I think it’s getting at the same point. Thought trials is in the same space: designing things (in your head), and trying them (in your head). Selection criteria references the strawman of boundaries. Artificial selection, as compared to natural selection, is a cute rhetorical device that makes the validation-is-not-important argument clear.

Most obvious to me is the almost perfectly-aligned relationship between this and the process of design strategy development as I understand it. Of course, Weick was already sort of there, in the space of business and organizational theory; I cited him extensively when I was down my sensemaking rabbit hole.

Anyway, this seems like a starting point for something... interesting: