Theory building from the points of view of a native,an anthropologist,and a philosopher: Commentary on Suzanne Rivard’s “Theory building is neither an art nor a science. It is a craft”

What is meant by “art,” “science,” and “craft”?

In the image of theory building as an art, according to Rivard, “the theorist has sparks of inspiration, and the theory emerges during trancelike periods of inspired writing. Once the theory is in writing, this object of art should not be ‘touched up’.” We expand this with the following definition of art from the Oxford English Dictionary (OED):

Any of various pursuits or occupations in which creative or imaginative skill is applied according to aesthetic principles . . .; (in plural with the, sometimes personified) the various branches of creative activity, as painting, sculpture, music, literature, dance, drama, oratory, etc.

When Rivard says that theory building is not an art, I interpret her as saying that theory building is not just the same as applying creative or imaginative skills according to aesthetic principles. This is not to deny that theory building does not involve applying creative or imaginative skills; this is only to say that theory building and art are not exactly the same.

In the image of theory building as a science, Rivard says,

there exists a series of activities—such as reviewing the literature to circumscribe the phenomenon of interest, analyzing the constructs and relationships from extant literature to identify gaps, and developing new theoretical propositions—that if dutifully followed will result in good theory.

Consistent with this image is the following definition of science:

A branch of study that deals with a connected body of demonstrated truths or with observed facts systematically classified and more or less comprehended by general laws, and incorporating trustworthy methods (now esp. those involving the scientific method and which incorporate falsifiable hypotheses) for the discovery of new truth in its own domain. (OED)

In the context of this definition, I interpret Rivard as saying that theory building is not just the same as applying, among other things, trustworthy methods (or, in her words, “a series of activities”). This is not to deny that theory building does not involve applying trustworthy methods; this is only to say that theory building and science are not exactly the same.

As for theory building as a craft, Rivard notes the following: “in addition to possessing these qualities [of art and science], the theorist as an artisan must also demonstrate care and ingenuity and work with patience and perseverance, being engaged in the work in and for itself (Sennett, 2008).” I looked up Sennett’s book; I could not find his definition of craft, but he describes craftsmanship in the following way:

Craftsmanship names an enduring, basic human impulse, the desire to do a job well for its own sake. Craftsmanship cuts a far wider swath than skilled manual labor; it serves the computer programmer, the doctor, and the artist; parenting improves when it is practiced as a skilled craft, as does citizenship. In all these domains, craftsmanship focuses on objective standards, on the thing in itself. . . (p. 9).

Of the OED’s many definitions of craft, perhaps the one most salient to Rivard’s usage is its definition as a verb: “To make or construct skilfully.” And in addition to Sennett’s commentary and the OED’s definition, I would like to make explicit a key notion implicit in Rivard’s conception of theory building as a craft: It is the well-known notion of tacit knowledge (Polanyi, 1969), which Smith (2005) describes in the following way:

Thinkers [including researchers who are building theory] often perform complex rule governed tasks even though they have no conscious or explicit knowledge of the rules involved. To explain these capacities, philosophers and psychologists have supposed that thinkers have unconscious or tacit knowledge of the rules for executing the tasks. A prominent example is speakers’ ability to produce and understand indefinitely many grammatical sentences of their language even though they cannot state the rules of grammar they conform to.

It is through experiences in theory building which take place over the course of a researcher’s career that the tacit knowledge, as a needed input to theory building, is itself built up. Central to theory building as a craft is its reliance on and use of tacit knowledge.

In summary, I interpret Rivard as saying that an information systems researcher, as a theory builder, calls upon not only the approach of art (involving, among other things, creativity, imagination, and aesthetic principles) and the approach of science (involving, among other things, trustworthy methods) but also the approach of craft, which has to do with the researcher’s explicit and tacit knowledge, applied in his or her own hands-on activity of carefully, ingeniously, patiently, perseveringly, and skillfully making or constructing a theory while focusing on objective standards.

Given these conceptions of “art,” “science,” and “craft,” what concerns might we add to Rivard’s thoughts about theory building being not quite an art or a science, but a craft?

Concerns from the perspective of a “native” of the world of information systems researchers

As a “native” myself in the world of information systems researchers who are busy working on theorizing, I have three nuts-and-bolts concerns about Rivard’s image of theory building. They might also be labeled practical day-to-day concerns faced by the individual researcher when engaging in theory building.

First, given Rivard’s “spiral model of theory building” and Rivard’s advice that theory building may properly seek to deliver only “approximations of theory” rather than complete “full-blown theories,” I wonder about exactly what forms an approximation of theory may take so that it can be considered ready for inclusion in a manuscript to be submitted for possible publication in a journal. How much would a theory still in development need to approximate a complete, full-blown theory for it to be considered publishable?

Second, I wonder and worry about what roles reviewers and editors might play in my efforts of theory building. Rivard’s spiral model of theory building includes no explicit entry points for input from reviewers and editors, but clearly they have a say in the acceptability of the theory that I build. What form might or should the spiral model of theory building take over two or more review cycles, especially where some of the input from reviewers and editors is not understanding and constructive but misinformed and disruptive?

Third, it is one thing for just an individual researcher to follow the steps of the spiral model of theory building; however, when a researcher is working with co-authors, it is quite another thing for the researcher to resolve differences in or diversity of opinion. In the way that Rivard explains her spiral model of theory building, it is an activity for an individual to perform, not a group. Perhaps a group decision support system would be needed!

The conception of theory building as a craft, rather than just an art or a science, can help to address (but not completely resolve) the above three concerns. The tacit knowledge about theory building that a researcher builds up over many experiences of theory building can help (1) to identify when an approximation of theory is ready to be written up for and submitted to a journal, (2) to anticipate the reactions of reviewers and editors, and (3) to manage the diversity of opinions among co-authors. However, the same perspective also tells us that younger scholars, such as recently minted PhDs, will be disadvantaged in addressing the three concerns because they do not yet possess the needed tacit knowledge—a disadvantage that can be mitigated only by commencing with the work of engaging in the craft of theory building itself.

Concerns from the perspective of an “anthropologist” studying the “natives” of the world of information systems researchers

There is more to the world of information systems researchers than just individual researchers or even groups of individual researchers. There is also their culture and social structure, which successive generations of individual researchers enter into, pass through, and leave from and which have an existence that changes much more slowly than the turnover of people in it. In this regard, Rivard’s portrayal of theory building as a craft is a purely individual model, not yet accounting for the influence of the culture and social structure of the community of information systems scholars.

One concern I have is that the community’s culture and social structure can constrain or even throttle the building of certain kinds of theory, even if the activity of theory building itself by individuals has been, from an objective perspective, exemplary. For examples in information systems research, one only has to think of the exclusion or oppression of interpretive theory until the 1990s and design theory until the 2000s. Interventions to alter normal operations can help to mitigate this and is addressed in the next concern.

The enabling of Rivard’s spiral model of theory building will require not only enlightened individuals to do it but also well-placed power brokers—including editors-in-chief, senior editors, special issue editors, conference track organizers, and dissertation chairs—to intervene with such an effort. In other words, the community should take advantage of whatever already existing roles in its social structure may be enlisted to promote theory building as a craft; this suggestion reinforces a point already made by Rivard.

An overall concern is how the information systems research community’s culture and social structure could be developed so that theory building as a craft not only is a practice that individuals can choose to perform but also is institutionalized. One step that could be taken is the creation of journals, or sections within journals, and conference tracks that welcome the products of theory building as a craft. Another step could be to undertake actions to deliberately promote the practice of theory building as a craft, rather than just to let the practice evolve naturally; such promotion can involve, for doctoral students, coursework and research projects that highlight theory building as a craft and, for established scholars and doctoral students, case histories of successful and unsuccessful instances of theory building as a craft.

Concerns from the perspective of philosophy

In so far as “philosophy is thinking about thinking” (Quinton, 2005) and given that Rivard’s model is an example of scientific thinking, what scientific thinking can be offered on Rivard’s scientific thinking? One thought is how Rivard’s spiral model of the theory-building process is “empirically testable” or otherwise assessable.

Rivard’s model is not a scientific theory, and it is scientific theories which are subject to what is called “empirical testing,” which assesses a scientific theory through hypothetico-deductive logic using predictions drawn from the theory. Just the same, though not a scientific theory, Rivard’s model needs to be assessed. But how? That would depend on what Rivard’s model is, if not a scientific theory.

In my view, Rivard’s spiral model of the theory-building process is a design, which can be defined as a “plan or scheme conceived in the mind and intended for subsequent execution; the preliminary conception of an idea that is to be carried into effect by action; a project” (OED). And in so far as a design is a plan, the latter can be defined as an “organized (and usually detailed) proposal according to which something is to be done; a scheme of action; a strategy; a programme, schedule” (OED). I would like to add that, in the process of designing and planning, alternative means for achieving a preset goal are considered, where the best alternative is selected (and where the selection of the “best” alternative can involve considerations such as efficiency, effectiveness, and the extent to which the goal is achieved). As a design for building theory, how may Rivard’s model be assessed?

First, apart from Rivard’s own experiences on which she draws for building her model of theory building and with which, therefore, her model can be deemed consistent, what additional, new experiences are needed to test or assess the efficiency and effectiveness of the model? It would be helpful to have, in the future, accounts of individual experiences (case studies, perhaps) of attempts to use the spiral model of theory building, including reflections on problems encountered and successes experienced.

Second, is Rivard’s model itself necessarily put together or “designed,” as it were, in the best possible way? Does Rivard’s way of combining activities, which include “read,” “reflect,” and “write,” with “ions” constitute the only or the best way of putting these elements together for the purpose of theory building? What feasible, alternative ways are worthy of consideration and, of these alternatives, which one would be most efficient and effective in theory building? Again, individual experiences or case studies would be helpful in such an assessment effort.

Third, how well can Rivard’s model be adopted as the practice of a craft? Rivard’s spiral model of the theory-building process needs to be tested, by others in addition to Rivard, for how well it allows the researcher to (in Rivard’s own words on what a craft is) “demonstrate care and ingenuity and work with patience and perseverance, being engaged in the work in and for itself.” Also, regarding what I consider to be the most important element of a craft—which is an individual researcher’s tacit knowledge—there are the considerations of how much tacit knowledge is needed simply to apply Rivard’s model in the first place and how conducive Rivard’s model is to building up the tacit knowledge of the individual researcher who is using the model.

A final concern worth mentioning from the perspective of philosophy is the matter of theory building as distinct from theory testing. The difference between formative validity and summative validity is salient here; Lee and Hubona (2009: 246) explain,

We define formative validity as an attribute of the process by which a theory is formed or built . . . We define summative validity as an attribute of the sum result or product of the process, namely, the theory. A theory achieves formative validity by following one or another accepted procedure in the process of its being formed [such as Rivard’s spiral model of the theory building process]. A theory, once formed, achieves summative validity by surviving an empirical test that uses the logic of modus tollens. Theory testing, we will show, involves comparing the theory’s observational consequences with the observed evidence. Theory testing can lead to either of two outcomes: if the evidence is consistent with the theory, then the theory has summative validity; if not, then the theory lacks summative validity.

In other words, no theory, no matter how well it is formed or built according to whatever method of theory building, may be accepted until and unless it subsequently survives empirical testing. Even if we assume that Rivard’s spiral model itself is flawless as a way of building a theory and even if a researcher were to apply the model perfectly, the resulting theory would only possess formative validity, not summative validity. For a positivist theory, the latter would involve deriving predictions from the theory as it applies to a specific setting and then comparing what the theory predicts against what is actually observed in that setting. It would not be until a theory survives an attempt at empirical testing that it could be accepted. Theory building, by whatever method or model, is but a step along the way to theory testing.

Conclusion

Rivard’s spiral model of the theory-building process involves elements of art, science, and craft, whether we consider them according to Rivard’s original definitions or this commentary’s expanded definitions. As an art, theory building involves sparks of inspiration, inspired writing, creativity, imagination, and even aesthetic principles. As a science, theory building involves a series of certain activities and trustworthy methods. And as a craft, theory building involves care, ingenuity, patience, perseverance, and explicit and tacit knowledge, all of which the researcher calls upon when engaging in the work in and for itself and focusing on objective standards. Rivard’s spiral model puts together the elements of art, science, and craft in a coherent and richly detailed way, such that the resulting theory which is built would be, in my view, more likely to survive the rigors of subsequent empirical testing.

Still, as examined in this commentary, it is important to keep the following in mind. There are practical day-to-day concerns that we have as individual researchers ourselves engaging in the practice of theory building. There are higher level cultural and social structural concerns that enable and constrain what individual researchers, as members of the community of information systems researchers, can accomplish in theory building. And there are philosophical concerns that arise in reflective thinking about the thinking involved in theory building.

Rivard’s spiral model of the theory-building process is provocative, and its success will ultimately depend on the extent to which the practical day-to-day concerns, the higher level cultural and social structural concerns, and the philosophical concerns are all addressed.