Conclusion‐Oriented Typologies

The most commonly cited typology, Barbour's, is a conclusion‐oriented typology. Perhaps most famously enunciated in Religion and Science (Barbour 1997), Barbour's typology is constituted by four possible religion‐science relations: Conflict, Independence, Dialogue, and Integration. These four views are typically glossed in something like the following manner: Conflict means religion and science are opposed, and only one is legitimate; Independence means they deal with entirely different phenomena/aspects of human life; Dialogue means they pursue similar questions or have similar methodologies; Integration means they can be assimilated for a single purpose (see, e.g., Shin 2016, for a similar characterization).

Haught (1995) also offers a fourfold typology. As with Barbour, his system includes Conflict and Separation (relabeled “Contrast,” perhaps so that all the categories begin with C). But the other two categories differ because “I do not find a sufficiently crisp logical distinction between his third and fourth types, ‘dialogue’ and ‘integration”’ (p. 9, fn. 1). In their place, Haught provides “Contact” (science and religion have implications for one another and thus ought to adapt as either changes) and “Confirmation” (“religion supports and nourishes the entire scientific enterprise” (Haught 1995, 9).

What is common to both Barbour and Haught's typologies is their orientation toward the particular relationship between religion and science: the two relata are in Conflict or Contact, and so on. When applied to particular scholars, the typology focuses exclusively on their general conclusion. In a sense, the character of this kind of typology is holistic: religion and science are related in toto or all at once in one way or another.

Concept‐Oriented Typologies

In contrast to these conclusion‐oriented typologies are concept‐oriented ones. These focus on how particular aspects of religion and science are related. Thus, when approaching the question, “What is the RSR?” through a concept‐oriented typology, one must first ask: “What concepts, ‘religion’ and ‘science,’ are we talking about?” One might say that concept‐oriented typologies are more fine‐grained than conclusion‐oriented ones, but more importantly they have different starting points.

Drees offers a ninefold typology in this vein. He begins by canvasing three kinds of “challenges to religion” that have historically been generated by science: those related to new knowledge; ones concerning epistemology, or how we understand knowledge; and challenges regarding “our appreciation of the world” (Drees 1996, 39–41). This trio of challenges is accompanied by three ways of understanding the nature of religion: cognitive, experiential, and as traditions (Drees 1996, 42–43). By crossing the challenges and conceptions, we obtain a nine‐member matrix of “areas of discussion in science‐and‐religion.” Scholars who engage in a particular area of discussion will thus tend to focus on a particular kind of challenge posed by science to a particular conception of religion—though authors can, of course, engage in multiple areas of discussion at once.

Drees' typology cross‐cuts conclusion‐oriented typologies by slicing along the conceptions of religion (and of science) at play. In fact, Drees claims that Barbour's typology can be found distributed within particular areas of his nine‐fold typology (Drees 1996, 45). In that sense, it is more fine‐grained than Barbour's and Haught's: religion and science are not related wholesale but along particular dimensions.

An even more sophisticated typology which takes this dimensional approach further is developed by Stenmark (2004) (see also Stenmark 2010). The typology begins with three basic distinctions: religion and science might be entirely separate endeavors, overlap some, or be unified. But Stenmark points out that quite distinct views are wrapped up in the overlap and unity positions: one might think there is more or less overlap, or that science wins in the overlap (scientific expansionism) or that it loses (religious expansionism), or that science may come to totally encompass religion, or that science may instead eventually be just a subset of religion (Stenmark 2004, 251–59).

Furthermore, what are separated/overlapped/unified are not just single things, Science or Religion with capital S and R—they are not monolithic phenomena but complex social practices. To that end, Stenmark outlines a number of dimensions along which one might evaluate the RSR: the social, teleological (i.e., the goals of the practices), epistemological, and theoretical—though importantly this list is not meant to be exhaustive. Moreover, within each dimension are wrapped up a number of subdimensions (though Stenmark does not use that phrase). For instance, when thinking of the teleological dimensions of religion and science, one might think at the community level—what religious congregations or groups of scientists aim at achieving—or at the individual level—what particular religious practitioners or scientists seek. And the degrees of overlap may differ along different dimensions as well: one might be a teleological community‐level separatist (or “restrictionist”) but a methodological unitarian (an admittedly rather implausible—though logically possible—position).

In all, Stenmark allows for something on the order of 6⁴ possible characterizations of the RSR. Notice that just as with Drees', Stenmark's typology cross‐cuts conclusion‐oriented typologies: those who might have been labeled Conflict theorists (like biblical literalists), might be classed as scientific or religious expansionists; or historical interactions between Religion and Science which have appeared to represent Conflict or Harmony (e.g., the Galileo Affair), might instead both be categorized as instances of, say, theoretical overlap. Again, what separates Stenmark's typology from conclusion‐oriented ones is his focus on the particular conception(s) of religion and science at play.

Operating in a different cultural landscape, Malik also offers a concept‐oriented typology, albeit one constrained to the relationship between Islam and human evolution rather than between religion and science more generally. And rather than focusing on different aspects of Muslim/scientific practice, Malik zooms in on particular understandings of scripture and the theory of evolution—it is in fact explicitly unidimensional (Pear and Malik 2022, 632). Thus, Muslim perspectives on evolution are typed according to their understanding of that theory: as entailing that all animals were produced via evolution, that only nonhuman animals were so produced, or that at least Adam (the first human) was not generated by evolution. These groups all believe that Islam and evolution are compatible, in contrast to the “Creationists” who reject the evolutionary origins of any animals whatsoever (Malik 2021, p. 111; Pear and Malik 2022, 632).

This focus on conceptions of evolution is motivated, similar to Drees and Stenmark's, by a critique of conclusion‐oriented typologies as inadequately nuanced. However, rather than expanding to a multidimensional model, Malik specifically proposes a unidimensional system because it (1) “helps avoid confusing and mixing religious and scientific beliefs or attitudes” and (2) “can clearly demarcate between what individuals accept regarding evolution versus why they accept or reject evolution” (Pear and Malik 2022, 632; emphasis original). Thus, a concept‐oriented typology is thought to be more useful than a conclusion‐based one, an idea we will build on later.

Barbour's Aims

Recall that conclusion‐oriented typologies all refer to the same thing: the conclusions scholars draw about the RSR. But how is it helpful to classify the scholarship in this fashion, typing scholars and their work by their conclusions? One defense offered by Barbour is that such typologies serve a pedagogical function: “typologies might still be useful in introductory courses… Especially in dealing with contemporary thought students need to be aware of a wide range of alternative views that would be difficult to treat… in the time that is usually available” (Barbour 2002, 347–48). And indeed this is how many introductory textbooks (e.g., McGrath 2020) are in fact structured. Andrew Loke's new fourfold typology is also explicitly advertised as having significant pedagogical advantages (Loke 2023). But it is important to note that this kind of consciousness‐raising aim is more sensible for a typology in its first‐order, rather than second‐order, use. Yes, explaining to students that, say, Conflict, is not the only configuration of the RSR might be useful. But what about classifying scholars as proponents of Dialogue versus Integration? It is not clear to me that this contributes to raising awareness “of a wider range of alternative views” beyond what they can get from the first‐order classification. In any case, once we move past the introductory context, consciousness‐raising does not seem a proportionate justification for the widespread appearance of Barbour's typology in the scholarly literature.

Another, more second‐order‐focused reason cited by Barbour is that typologies offer maps of the religion‐science literature: “… a broad overview of a range of possible relationships can be helpful to readers new to this interdisciplinary field, even though an overview inevitably oversimplifies the complexities of the real world. A guidebook to any territory is not intended as a substitute for firsthand exploration but is intended to help people find their way around” (Barbour 2002, 2023, 348). Likewise Stenmark: “the aim of developing a typology is primarily to give a map which sorts out the main positions regarding how to relate science and religion” (Stenmark 2004, 262). True enough; the literature is vast and a guidebook/map would surely be useful. But just because one can trace the territory along certain contours does not mean the resultant sketch will provide relevant aid. What I find peculiarly missing is any explanation of how exactly slotting scholars into categories like “Conflict” or “Separation” is actually useful for the reader. In a sense, conclusion‐oriented typologies are almost trivially true: yes, Dawkins is indeed a Conflict theorist and Gould embraces Separation—we can get all that on the first page (or sooner). But what can a scholar (or a lay reader) do with that kind of information? It is not clear. (I will return to this cartographic aim later.)

Barbour draws a more sophisticated justification for typologies from the social scientific literature. Citing Weber and others, he explains that classification schemes are useful for highlighting the complexity of individual cases, for only very rarely will particular cases fall perfectly into the scholarly categories. Typologies thus help us compare individual cases by approximation to the idealized category (Barbour 2002, 348). I think this is an admirable aim, but unfortunately this is not how typologies, at least in the religion‐and‐science literature, are actually used. They are far more commonly used to eliminate nuanced differences—as in the case of scientistic atheists and biblical literalists in Barbour's own system. Be that as it may, I think Barbour is right that typologies should aim to clarify the unique contributions and views of particular authors and how they relate to others. Based on the critiques given of Barbour's typology (as with Malik's above), I would suspect that concept‐oriented typologists believe their typologies obtain this goal better.

The above aims are ones which I take any scholar would wish a typology to achieve. I must acknowledge, however, Barbour's overarching aim in presenting his typology: to advocate for Integration. The point of laying out Conflict, Separation, and Dialogue, and in that order, is didactic; it enables Barbour to highlight the issues facing the first characterizations of the RSR and thus build a case for Integration. Drees glosses the purport of the typology slightly differently, as a way of representing alternatives to Conflict in an increasingly secular world (Drees 2010, 1). Richard Olson likewise takes this to be an important use of Barbour's typology, and explicitly explains that a major reason for proposing typologies is their use in countering/defusing a simplistic form of Conflict which has great hold on popular imaginations (Olson 2011). I have chosen to leave out these kinds of aims, however, because they are partisan. Since typologies are analytic tools used by scholars across the spectrum of first‐order positions regarding the RSR, I limit discussion to those aims that can be recognized by scholars no matter their particular view of the RSR.

Three Other Aims

I now consider three other aims we might wish a typology to achieve in its second‐order usage: (1) illuminating ways in which contributions do/do not effectively engage with one another, (2) explaining why particular pieces of scholarship receive more public uptake, and (3) providing a useful guide for a public readership based on their values/reasons for interest in the RSR.

Conceptual Analysis

A common way of determining the relation between religion and science is via conceptual analysis, by which I mean the method proceeding as follows: first determine definitions of “religion” and “science” (or particular religions/sciences); then logically deduce their relationship on the basis of those definitions.

Stephen J. Gould, for example, uses this method: science and religion have very different “magisteria.” In fact, those magisteria do not overlap—and thus there can be no conflict, synthesis, or any other kind of interaction between them (Gould 1998). Reaching a very different conclusion using conceptual analysis is Tiddy Smith: religion and science do in fact overlap in their explanatory target (the world and its happenings), but they employ radically different epistemologies—religion makes use of highly individualistic evidence while science respects only intersubjective evidence. Given this, the two inevitably wind up in conflict (Smith 2019).

It should be pointed out that the particular way in which a conceptual analysis is conducted may vary greatly between scholars. One might, like Gould, simply intuit the notions of religion and science a priori. On the other hand, one could instead arrive at conceptions of religion and science more empirically. Frazer, for instance, does this in The Golden Bough: that religion makes appeal to Wills/agents while science appeals to regular Laws is a conclusion (supposedly) reached by induction over many cases (Frazer 1922, Ch. 4). But regardless of how they determine the definitions of “religion” and “science,” the above authors all arrive at their characterization of the RSR by comparing those definitions. This is the method of conceptual analysis.

Case Studies

Perhaps the most common method employed in the literature is what I'll call the method of case studies. Here, rather than comparing definitions, one performs a kind of induction over some number of historical episodes of religion‐science interaction. The goal is that such an induction will reveal the RSR.

Exemplars of this method go back to the early history of the discipline: the works of Draper and White. In their infamous histories, Draper and White enumerated dozens of historical episodes (some fabricated), on the basis of which they hoped to show that there is some kind of tension between religion and science. Working at a perhaps more modest scale, historians like David Hollinger and Marwa Elshakry have pushed for Harmony on the basis of their studies of twentieth‐century Jewish scientists (Hollinger 1996) and the reception of Darwin in Islamic cultures (Elshakry 2013).

These kinds of “positive,” or “constructive” inductive projects can be contrasted with more “negative” projects of a “debunking” nature. Indeed much of the historical work from the past five decades has focused on debunking the narratives of the classic Conflict theorists (Lightman 2019 calls it “myth‐busting”). Number's aptly named edited volume Galileo Goes to Jail and Other Myths about Science and Religion (2010) is representative, as are the revisionist accounts of the Galileo Affair by Heilbron (1999) and religious biographies of eminent scientists—like Iliffe (2017)'s Priest of Nature: The Religious Worlds of Isaac Newton. In all these cases, particular historical episodes or thinkers are consulted as cases in a broader induction to the RSR.

Relativizing

A related but quite distinct method can be called “relativizing.” While the method of case studies engages with historical actors and their actions, relativizing engages with the concepts used by contemporary scholars. It comes in (at least) two flavors: cultural and historical (i.e., “historicizing”).

The general idea is this: take the modern concepts “religion” and “science” in use today and show (or assert) that they either did/do not exist, or had/have radically different meanings in different times/places. On the basis of this, one argues that one cannot provide a universal and/or diachronically stable characterization of the RSR—any such attempt must either fail or be incredibly local (temporally and/or culturally).

I think the historicizing approach is best exemplified in the work of Peter Harrison, especially in Territories (2015), wherein he demonstrates how various sociohistorical contingencies from the sixteenth century till now set the parameters for how the West understands the RSR. Had things turned out differently (e.g., had the Protestant Reformation not happened), had our notions of “science” and “religion” taken slightly different forms, we may not have even been able to conceive of religion and science as being related in one of the four Barbourian ways. Ungureanu's recent work (Ungureanu 2019) likewise highlights how the notion of conflict between religion and science emerged from a very particular sociohistorical moment in nineteenth‐century Victorian England.

On the other hand, a good example of cultural‐relativizing is found in Jason Ãnanda Josephson's The Invention of Religion in Japan (2012). Josephson contends that prior to the Meiji Restoration and the United States' forceful “opening” of Japan's ports, there was no native Japanese conception of religion—or of science. Instead, this concept was invented (quite explicitly) by a number of scholars and political figures in order to appease the foreigners' demand for “religious freedom.” The moral of the story (if taken to heart) is that we ought not generalize our characterization of the RSR temporally or spatially; the things related are so radically different (perhaps even nonexistent) in different times and places that we cannot usefully provide a general account of the RSR.

Fieldwork

The final method I consider encompasses a range of methods drawn from the social sciences, and which I broadly call “fieldwork.” This embraces methods such as surveys, interviews, and ethnographies. What distinguishes these methods from the above is their explicit focus on the “everyday,” quotidian experiences of/encounters with “religion” and “science.” The essential idea behind fieldwork is that the proper characterization of the RSR is to be found reflected in the responses or actions of everyday, ordinary scientists and religious folk.

The classic example of this is Leuba (1916)'s survey of those listed in the directory American Men of Science. Observing that only 31.6% of the “greater men” among his sample indicated belief in a prayer‐granting God, Leuba claimed a basic incompatibility between religion and science. Leuba‐esque studies have been repeated several times in the intervening century‐plus, with varied interpretations (Larson and Witham 1997). A much more complex form of fieldwork can be found in the work of Elaine Howard Ecklund, sometimes in collaboration with Christopher Scheitle. In addition to surveying hundreds of academic scientists and everyday religious folk, Ecklund has also performed an extensive sleuth of interviews with such scientists and immersed herself in various religious communities across the United States (Ecklund 2010; Ecklund and Scheitle 2018).

Illuminate Effective Engagement

As with concept‐oriented typologies, a typology of methods can illuminate how particular works engage, or fail to engage, with one another—and can indicate how to most effectively engage with others. By calling explicit attention to the methods used by scholars, the typology encourages us to address the arguments rather than the conclusions found in the works to which we respond. The different methods outlined above clearly employ different kinds of evidence—for instance, the method of case studies does not rely on the first‐hand reporting of everyday laypersons, as does fieldwork. To try to use fieldwork‐based evidence against a scholar employing the method of case studies might thus be illegitimate. On the other hand, recognizing the diversity of methods in play could be immensely generative.

Unfortunately, discussion of scholars' methods is generally lacking in the literature. One place where the (possible) dangers of talking past one another due to different methodologies is acknowledged is in Smith (2019). Right at the start, he clarifies that he is speaking of an epistemic conflict between religion and science, not an historical one:

… I will argue in the course of this book that the conflict between science and religion is quite real, and further, that the conflict has a clear victor. The methods of science outcompete the methods of religion. I must emphasize from the outset that I do not dispute what has already been said by [historians]: the historical relationship between science and religion has been complicated… But this book is not about history. This book is about epistemology: the theory of knowledge. And the questions that this book seeks to answer are primarily about knowledge, not history. (Smith 2019, 1)

The message is clear: he believes it is simply not relevant to bring up historical case studies as objections to his account—religion and science are here understood as particular kinds of intellectual endeavors employing particular kinds of evidence to explain particular, overlapping classes of phenomena, and thus are by definition bound to conflict with one another, at least at some point (and in such a way that science will always come out on top).

By drawing our attention to the scholarly methods used in favor of a particular characterization of the RSR, a typology of methods can help determine what kinds of objections will be relevant to particular authors and their works. It can also point to potentially surprising places of disagreement. We see this perhaps most starkly in the case of historicizing and the method of case studies: if we fully embraced the historicizing method and all its implications, then we would not even permit the lumping‐together of distinct historical episodes to form the base for a case‐studies induction. Oddly, this tension between historicizing and the method of case studies has not, to my knowledge, been acknowledged in the field—and some are quick to endorse both simultaneously (in particular as ways of criticizing the Conflict Thesis; see, e.g., Lightman 2019, 5–6). On the other hand, focusing on methods can also highlight ways in which particular authors should alter their methods to better accommodate/acknowledge their critics.

Explain Public Uptake

We saw above that both conclusion‐ and concept‐oriented typologies provide some traction on the question, “Why do some pieces of scholarship receive more public uptake than others?” A typology of methods provides an another take on the issue. Parallel to the schema derived from concept‐oriented typologies above, the idea is simple: some methods are easier to understand/follow/digest than others. The historicizing method, in particular, is quite complex, and does not lend itself easily to public exposition or uptake. Other methods, however, are more liable to absorption by the public. At least some cases of conceptual analysis, for instance, are amenable to sloganization—“Science uses Reason, Religion uses Faith”—which can help their conclusions stick. Likewise, the narrative style employed by some instances of the method of case studies lends itself to public remembrance: who can forget the great struggle between Galileo and the Church or the burning of Bruno?

Notice that this is a distinct way of approaching the issue from that suggested by concept‐oriented typologies. Its explanatory power comes from focusing on the way various publics digest information rather than on their particular conceptions of religion and science.

Much work, of course, is still to be done in exploring exactly how this methodological strand of analysis can contribute to resolving the question of public uptake—and surely in the end it is not just method or rhetoric or politics, but a blend of all (and others) which do the explaining. But focusing on how the methods used appeal/fail to appeal to particular publics can offer fruitful insight into the issue.

Guide the Public

A typology of methods' greatest strength, I think, lies in its ability to provide a guide to the public in navigating the literature. We saw above that concept‐oriented typologies do this by asking what concepts of religion and/or science the subject has and directing them to literature which employ those same conceptions. In this way concept‐oriented typologies can help consumers (scholarly or not) find work that is actually relevant to them. A typology of methods can provide a similar though quite distinct guide.

A methods‐oriented guide builds off the idea that different methods are likely to appeal to different readers. Importantly, the consumers of the religion‐science literature are incredibly diverse. Readers have many different reasons for delving into the work on the RSR: some seek ways to defend their faith, others seek ways to attack others' faith; some have purely academic interests in the RSR, others a much more personal investment; some are embedded in a particular faith tradition, others are not. And the particular set of circumstances which lead readers (and researchers) to the literature contribute to the kinds of evidence they will find relevant (and convincing). Since different methods employ different kinds of evidence, it follows that the different methods will be more or less equipped to deal with different particular readers' concerns/interests in the RSR. By isolating what kinds of methods are best suited to which kinds of concerns, a typology of methods can thus help direct members of the public (and scholars!) to those works which would be most relevant to them. Schematically, the guidance would look like this: “Readers with concerns X should read works Y and Z because they use method A.”

This focus on values rather than concepts results in real, pragmatic differences. Recall, for example, the case of the freshman biology student who has grown up in a relatively conservative Evangelical Christian faith community. They are interested in evolutionary biology but have also heard rumors that such science is in deep conflict with their faith—and so they wonder whether they can flourish as both a biologist and a Christian. What kind of guidance could a typology of methods give? It might well be that what concerns this student most is whether they can fit themselves into a narrative of religious biologists. Given that, it would make sense to direct them to the case‐study literature, perhaps to the work on Darwin's reception in Victorian England and in the United States—rather than to historicizing work like Harrison's, or even conceptual analytic work like Plantinga's, which our student may find too abstract. Likewise, this kind of student might be interested in how they will be treated as an academic biologist who is also a Christian, in which case fieldwork‐esque studies will be the most relevant.

Now consider a case with broader social implications: a politician navigating her, say Muslim, constituents' opposition to stem cell research. The politician in this case wants to understand the root of the opposition, and thus find ways of defusing it or communicating it to her colleagues. Here, fieldwork studies, like the work done by the Pew foundation or studies of Muslim physicians (Everhart and Hameed 2013), will be more appropriate than historical case studies or conceptual analyses.

Note the difference in how this case is treated by a methods‐oriented guide versus a concept‐oriented guide. Using the latter, we would ask after the politician's conceptions of religion and science, or perhaps about their constituents' notions. But it is easy to see how this might not lead to a result that is actually useful for the politician. For suppose that both the politician and her constituents understand religion and science as competing forms of knowledge production about the natural world. We would then suggest that she read works from, say, Smith and Plantinga. But it is not clear how those works would lend themselves to actionable recommendations for the politician's actual situation: how does it help to know that indeed the methods of science outcompete the methods of religion or that true science is really compatible with true religion? Would this help our politician address or even make sense of her constituent's concerns? Instead, fieldwork studies which indicate how lay religious folk actually interact with science seem more likely useful—regardless of her own understanding of what religion and science are. Such studies can give the politician a better sense of what is “really” at issue in Muslim opposition to stem cell research because it builds on actual studies of on‐the‐ground individuals rather than abstract, idealized concepts.

Again, the question a method‐oriented guideline asks is “Why do you care about the RSR?” rather than “How do you understand ‘religion’ and ‘science’?” And in fact asking this latter question is likely irrelevant to the politician's purpose. It might be relevant to understand how their constituents conceive of religion and science. However, if the constituency is diverse, with many conceptions of religion and science present, then the recommendations from a concept‐oriented guide will quickly get out of hand: read everything!