Elements of a Comparative Methodology in the Study of Religion

Abstract

While comparison has been the subject of much theoretical debate in the study of religion, it has rarely been discussed in methodological terms. A large number of comparative studies have been produced in the course of the discipline’s history, but the question of how comparison works as a method has rarely been addressed. This essay proposes, in the form of an outline, a methodological frame of comparison that addresses both the general configuration of a comparative study—its goal, mode, scale, and scope—and the comparative process, distinguishing operations of selection, description, juxtaposition, redescription, as well as rectification and theory formation. It argues that identifying and analyzing such elements of a comparative methodology helps, on the one hand, in evaluating existing comparative studies and, on the other, in producing new ones. While the article attempts to present the methodological frame in a concise form and thus offers limited illustrative material, the authors of the other essays in this collection discuss rich historical-empirical cases as they test the frame on their own comparative studies.

From the early days of the academic study of religion until today, comparison has been an important feature of the discipline. It was, and is, practiced in many different ways, while being reflected upon, praised, and scorned in heated theoretical debates. Interestingly, while scholars frequently speak of “the comparative method”,1 they are almost always more interested in the adjective (“comparative”) than in the noun (“method”). Normally “comparative method” is simply a synonym for “comparison” or for generic terms like “comparative approach” or “comparative perspective.” How comparison actually works as a method in the study of religion has not been discussed in greater detail so far. This raises the question: How seriously should we take the designation “method” when it comes to comparison? In his excellent survey chapter on comparison in the Routledge Handbook of Research Methods in the Study of Religion, Michael Stausberg addresses this point (Stausberg 2011, p. 34):

[C]omparison is most often not practised as a separate method, but as a research design, i.e., as a framework for the collection and analysis of data and the analysis of research problems. Comparative research designs use different kinds of techniques or tools for the collection of data (i.e., methods in a more narrow sense), for example discourse analysis, content analysis, document analysis, philology, hermeneutics, historiography, phenomenology, surveys, etc.

Conversely, while comparative research designs engage specific methods, many methods in turn operate comparatively. It bears pointing out that comparison is part of the working routine of most methods. On this more basic level, comparison works in the most unspectacular ways and is largely uncontroversial. To begin with, the formation of concepts and classifications and related forms of systematization rely on comparison, which therefore is enshrined in all research methods. Moreover, comparison of data is standard practice in all scholarly methods.

These are important observations. On the one hand, comparison often serves as a broader analytical framework—a research design rather than a research method. On the other hand, it is so fundamentally embedded in most research methods that it can hardly be recognized as a separate method. Consequently, Stausberg’s article on comparison was placed, in the Handbook, not under “methods” but under “methodology,” alongside articles on epistemology, feminist methodologies, research design, and research ethics.

While comparison can certainly be categorized this way, I would like to explore the—perhaps less frequently occurring—cases to which Stausberg’s first sentence alludes: “Comparison is most often not practised as a separate method” (my emphasis). How do we envision comparison in the study of religion when it is, intentionally and explicitly, practiced as a “separate method”? In other words, how are studies to be conducted whose primary research questions can only be answered by means of a comparative operation?

First, it is important to note that the points Stausberg makes in the first paragraph of the quote remain valid. Comparison cannot replace the listed techniques and tools, the “methods in a more narrow sense”—philology, content analysis, document analysis, etc. While the comparative research question informs the selection of the comparands, comparison itself is not a method that can or should be used in the first description and analysis of data. I suggest classifying it as a second-order method, which presupposes first-order-method research.2 It seems that comparing two or more items can be productive only if those items are being seriously studied.3 I wish to argue that like the methods used for the primary study of data, the second-order method of comparison features some elements that may permit us to view it as a method in the sense of an organized and controlled (or controllable) procedure.4

In the following I propose a methodological frame of comparison that consists of such elements. In their introduction to the above-mentioned Handbook, Michael Stausberg and Steven Engler define a scientific method as “the generally accepted mode of procedure in the sciences in a broader sense (including the humanities)” (Stausberg and Engler 2011, p. 4). Since comparison has rarely ever been discussed in methodical terms in the study of religion, it is hard to predict whether the following description will be generally accepted. But I believe that the elements discussed here are largely familiar to practicing comparativists, even if the terms may be partly new. My primary goal is to provide analytical categories, that is, a vocabulary that enables us to speak about the methodical components of comparison that most comparativists more or less intuitively exert in their scholarly practice. I argue that identifying and analyzing such elements of a comparative methodology may help in evaluating existing comparative studies and also in producing new ones.

The following description of the methodological frame has two sections. The first outlines the general configuration of a comparative study: Goals, modes, scales, and scopes. The second discusses various (potential) operations in the comparative process: selection (of comparands and the tertium comparationis), description (of data), juxtaposition, redescription, as well as rectification (of scholarly categories) and theory formation. For the purposes of this essay, I will discuss each element in a brief and concise manner; a more extensive discussion is in preparation. Since the other essays in the present journal issue provide deep analyses of particular case studies, I will also keep the discussion of examples brief.

1. The Configuration of a Comparative Study

Every comparative study is configured in a certain way with regard to its goal(s), mode(s), scale, and scope(s). However, the decisions that lead to its final configuration are rarely all made at the outset. More often, new insights emerging during the course of the research process yield reconsiderations and revisions of preliminary decisions. Thus the categories proposed here—which are also linked to one another in important ways—, are set up and defined merely for analytical purposes. I argue that they may be useful in evaluating (and also in conducting) comparative studies, but they do certainly not reflect a linear research process. Research flows are often complex and unpredictable.

1.1. Goals of Comparison

The most general question “Why compare?” is almost as philosophical as the question “Why do research?” and thus beyond the limits of this methodological discussion. Yet every comparative study has its own agenda and specific goals, and responsible scholars reveal and explain these goals in the introduction to their studies. For locating the goals more broadly it is useful to consider the disciplinary orientation and the intended audience of a study—the discourse out of which a study emerges and the readers to whom it is meant to speak.

The concept of academic disciplines, which ideally reflects a division of labor, has been questioned for some time, not least by university administrators and functionaries who believe that dissolving disciplinary boundaries will make scholarship more efficient. Time will tell if this is a productive approach, but at this point, and for our purposes, the disciplinary orientation is still valid. The scholarly discourse in which the author is operating, broadly revealed in a study’s bibliography, is mostly linked to a particular discipline such as religious studies, anthropology, sociology, history, theology, etc. Even more than the works that are cited (and especially those that are cited favorably), the subject at the very core of the inquiry determines this affiliation. Does the study strive to advance, first and foremost, the understanding of religion (religious studies), culture (anthropology), society (sociology), history (history), God or the sacred (theology)? Clearly this oversimplifies the matter outrageously. Identifying the abstract subject at the heart of the scholar’s interest (religion, culture, society, etc.) must certainly not be understood in a reductionist way. But it might be a helpful first step towards appreciating a study. Acknowledging the disciplinary orientation helps assess a study appropriately and fairly and can prevent unnecessary conflicts. For example, when a study is situated in a theological discourse and is interested, in the final analysis, in God or the sacred, it is quite pointless to criticize it for not advancing the general understanding of religion—and vice versa.5

Aside from the disciplinary orientation a study’s goal is also determined by the audience to whom it is meant to speak. Clues can be found in the list, in the publisher’s catalogue, of potentially interested reader groups or, for a journal article, in the journal’s profile. While many comparative studies are written primarily for readers in the author’s own and related disciplines, some studies are directed at a more general audience. Since comparative studies do normally not provide general introductions to specific fields of study but rather make a particular argument, it is useful to ask to which non-academic discourse the author intends to contribute or, in other words, which impact on public discourse s/he wishes to have. If the primary goal is non-academic, a close look at the comparative method is in order, for such a study’s benefit for scholarship may be low. Some studies of this sort place a one-sided focus either on similarities (in order to help rationalize a political conflict, e.g., between Catholics and Protestants in Northern Ireland) or on differences (in order to demonstrate the superiority of one religious group over the other).6

For the study of religion, pursuing two goals, description and classification, seems particularly productive.7 A comparative study whose primary goal is description aims at a better understanding of a particular historical-empirical item by means of comparison. Comparing that item with other items can serve a heuristic purpose by identifying aspects and facets that would otherwise be missed or neglected. It can produce insights by de-familiarizing the familiar. And it can be the method for testing hypotheses and causal analyses that aim at a more nuanced description of the respective item. This goal corresponds to the illuminative mode of comparison discussed below. Second, a study whose primary goal is classification uses comparison to form, apply, critically evaluate, and refine metalinguistic terminology in order to classify religious phenomena. This goal corresponds to the taxonomic mode of comparison. While it seems useful to separate these two general goals analytically, they can certainly complement each other, even within a single study.

1.2. Modes of Comparison

Jonathan Z. Smith coined the term “mode of comparison” in his article, “Adde Parvum Parvo Magnus Acervus Erit,” originally published in 1971 (Smith 1978), and discussed it further in his more widely quoted essay, “In Comparison a Magic Dwells” (Smith 1982).8 In accordance with Smith I understand modes as general styles of comparison that reflect both the spirit in which scholars compare and, to a certain degree, the goals of the individual study. Having reviewed previous scholarship Smith suggests that each and every study had been conducted in one of four modes of comparison—what he calls the ethnographic, encyclopaedic, morphological, and evolutionary modes. Smith’s model highlights certain methodical deficiencies: The ethnographic mode (which, I suggest, should better be called the spontaneous-associative mode) displays an undue use of intuition; the encyclopaedic mode yields superficial categorizations of little analytical value; the morphological mode decontextualizes phenomena in problematic ways; and the evolutionary mode posits analogies based on the broader theory of evolution. In Smith’s argument, the modes are meant to reveal the deficiencies of past scholarship (which he therefore sweepingly dismisses). While his conclusions raise further questions for discussion, the modes may be useful for the analysis of comparative studies—devoid of their polemical overtones and slightly modified. We should replace the misleading name of the first mode (see above) and refrain from sweepingly applying the morphological mode, the prime example of which is Eliade’s work, to all classificatory comparisons, because some avoid the described pitfalls. Further, it is important to note that the modes are not mutually exclusive but can appear, in some combinations, together in one and the same study. With these modifications, Smith’s four modes can be useful tools to test and evaluate comparative studies.

Another model, suggested by David Freidenreich, distinguishes four modes as well, but does so quite differently (Freidenreich 2004). Freidenreich’s first mode describes studies that display a strong focus on similarity and downplay or ignore differences between the comparands. The second mode, accordingly, refers to studies with a one-sided focus on difference. Freidenreich presents telling examples and argues that such approaches might serve certain political, social, or religious agendas well but produce few new conclusions for the study of religion. Studies in the third mode display a focus on genus-species relationship, constructing (or deconstructing) a genus (such as religious nationalism, scripture, or myth) by comparing various historical “species” and identifying differences and similarities between them. The fourth mode describes studies that use comparison to refocus, that is, to understand phenomenon a better in the light of phenomenon B, with a refocused lens.

Considering the mentioned weaknesses of some modes, two modes, which I propose to call the illuminative and the taxonomic mode, appear most promising for the study of religion. They correspond to the above-mentioned goals of description and classification, respectively, and echo the last two modes of Freidenreich’s model.9 The illuminative mode aims at illuminating a particular historical-empirical item, especially assumed blind spots, by drawing comparatively on other cases. This mode is asymmetric because the other cases are not studied in great detail in their own right; they simply help to illuminate the item at the center of attention. For example, cross-cultural comparison has been used to illuminate the depictions of sacrificing gods on ancient Greek vases (Patton 2009) or the meaning of giant stone footprints at an Iron Age temple in Syria (Thomas 2008). The taxonomic mode, on the other hand, aims at forming or modifying meta-linguistic typologies, taxonomies, classifications, or categorizations and thus at theory-formation. It is symmetric because every “species” of a “genus” receives equal attention.10 A recent example is the theorization of conceptions of the afterlife based on a comparative study of early civilizations (Shushan 2009).

Again, both modes may appear in one study. An example is my own comparison of discourses about asceticism (= the genus) in Brāhmaṇical texts from ancient India and Christian texts from late-antique Egypt (= two species). As a result of the comparison I was able to describe the structure of the asceticism discourse in theoretical terms, including a classification of its elements (Freiberger 2009, pp. 249–58). While the primary mode of this study was taxonomic, the illuminative mode was employed as well. For example, the frequent encounter with the ascetic ideal of sedentariness in the Christian Apophthegmata Patrum helped me identify and acknowledge the few passages in the Brāhmaṇical Samṇyāsa Upaniṣads that praise the ascetic practice of remaining in one place (as opposed to the otherwise omnipresent ideal of wandering about); here one source illuminated the other (see Freiberger 2009, pp. 235–38; Freiberger 2010).

Six of the modes of comparison discussed here (Smith’s four modes and Freidenreich’s first two) reflect comparative styles that seem problematic: An undue use of intuition, superficial categorization, inappropriate decontextualization, a positing of analogies that are rooted in an evolutionary model, a one-sided focus either on similarities or on differences. On the other hand, two modes, the illuminative and the taxonomic, seem promising. The modes thus constitute an analytical inventory that may be useful both for evaluating existing studies and for framing one’s own.

1.3. Scales of Comparison

The scale of a comparison marks the degree to which the study zooms in on the comparands. I appropriate the cartographic tool “scale” to indicate different levels of abstraction that maps have too. Maps can represent large territories like a country, a continent, or the whole world (small-scale maps), or smaller ones, like a city, a neighborhood, or a museum floor (large-scale maps).11 The scale is the degree of abstraction to which items are represented on the map. Choosing an item for comparison goes along with determining the scale and picking a map—that is, deciding on which level of abstraction one intends to compare. Since small-scale maps have other purposes than large-scale maps, comparing, say, an item that is visible on a state map of Arizona with an item visible on a floor map of a museum in New York City would need a creative explanation. Normally, productive comparative studies aim at balanced comparison, which means zooming in to the same degree for each comparand—for example, comparing items located on two state maps (or on two museum maps) of the same scale.

This applies, mutatis mutandis, to comparison in the study of religion.12 Existing studies compare at many different points on the scale, from comparisons of particular individual persons in their local settings to comparisons of entire religions. While the scale is continuous, it might make sense to broadly distinguish, related to how much the study “zooms in,” three levels: Micro, meso, and macro comparison. Micro-comparative studies zoom in on very specific items such as certain individuals or groups, certain texts, certain objects, certain practices, etc., and compare them. Macro-comparative studies compare entire religions, or several religions in view of one phenomenon (e.g., in studies entitled “Sacred Places in World Religions” or the like). Located in-between, on a mid-level scale, are meso-comparative studies, which cover more ground than micro-comparative ones but remain within clearly defined limits. It does not seem particularly useful to try and draw precisely defined lines between micro, meso, and macro comparison. They should rather be considered as rough dividers on a zoom scale that help categorize comparative studies.

It should be noted that today many comparativists in the study of religion have reservations towards comparison on the macro, or even meso, levels. It was macro-comparative studies, especially approaches in the phenomenology or religion (classical examples are (Leeuw 1967; Eliade 1958; Heiler 1961)) that were partly responsible for the bad reputation that comparison gained in the second half of the twentieth century. Those studies attempted to demonstrate, by presenting examples from many different religions, the true essence of religion or of certain religious phenomena. Aside from their religious or philosophical agenda, their main methodological weakness was, however, not the choice of scale but rather the fact that the (macro) scale at which the conclusions were drawn was not in accordance with the (highly eclectic) scales at which the actual study was conducted. Selecting certain items and declaring them as representative of a religious tradition essentialized that tradition; internal conflicts and discourses as well as historical change were largely ignored. When scholars declare to compare entire religions by juxtaposing such selected items, this only looks like macro-comparison. It is, in fact, a form of unacknowledged micro- and meso-comparisons conducted with a macro-comparative agenda. Such studies are methodologically problematic both as macro and as micro comparisons.

It is therefore crucial that the selected scale matches the question that a study seeks to answer. If that is the case, from a methodological standpoint all three levels are valid. Again, the map analogy might help clarifying this. Small-scale maps used for macro comparisons are indispensable for answering certain questions. Only these maps show, for example, that Texas is bigger in size than France, or that Austin, Texas (U.S.A.) and Cairo (Egypt) have almost the same latitude. Detailed maps of Texas and France or city maps of Austin and Cairo would be quite useless for answering such questions. I argue that macro comparison can be valid also in the study of religion if the research question is relevant and the conclusions drawn from it remain on the macro level. Such studies would often be based on quantitative and statistical research. One recent example is Norris and Inglehart’s study of degrees of secularization that argues, based on surveys from eighty societies that cover about 85% of the world’s population, that a society’s degree of secularization corresponds to its members’ experience of existential security (Norris and Inglehart 2004).13

1.4. Scopes of Comparison

The category “scope” reflects the distance between the items compared in a study. We may distinguish, very broadly, between contextual, cross-cultural, and trans-historical scopes. Studies with a contextual scope compare within one historical context or cultural milieu that can be delineated both spatially and temporally, for example the Mediterranean world in late antiquity, north-eastern India in the 5th century BCE, or contemporary Brazil. Studies with a cross-cultural scope, on the other hand, go beyond postulated cultural boundaries, like in a comparison of ancient Chinese and ancient Greek religion.

Clearly, these categories are not static and clear-cut. Spatial boundaries are often fluid, and cultural boundaries are constructions. The category “cross-cultural” is not meant to reinforce the notion that the respective cultures can be clearly delineated and are entirely separate from each other (cf. Welsch 1999; Juneja 2013). Rather, the categories have a specific analytical purpose. Scholar who compare in a contextual scope normally expect connections and interaction between the comparands, whereas scholars who conduct cross-cultural comparison normally expect unrelated developments.14

While comparison in a contextual scope is omnipresent in all historical-empirical scholarship on religion, scholars often hesitate to label their studies “comparative.” Yet comparisons are regularly drawn between various texts or manuscripts, ideas, rituals, objects, etc., especially for exploring and evaluating relations between the comparands. This relational approach is common practice, particularly in its genealogical variant, in which comparison is used to study potential borrowings and dependency. Genealogical comparison is indispensable for historical scholarship, but it can also come with an undue assumption of unidirectional flows and linear developments and the creation of reductionist pedigrees for certain religious phenomena (Smith 1990, pp. 46–53; Bornet 2016, pp. 73f). Recent work in “entangled history” or “connected histories” uses a broader relational approach that focuses on transregional flows and cross-fertilization. Despite the fact that these studies often transgress the conventionally set geographical and cultural boundaries, their scope, as defined above, is still contextual (not cross-cultural) because they study relations between the comparands. The “context” in such studies can become very large, even global (“global history”).15

Scholars commonly contrast genealogical16 (or more generally, relational) comparison with analogical comparison, which is based on the assumption that there is no significant historical link between the comparands. Rather, religious concepts, practices, objects, etc. are regarded as analogical due to observed similarities in their forms or functions. All studies conducted with a cross-cultural scope, as defined here, are analogical comparisons, and most studies with a contextual scope are relational. Yet analogical studies can well be done with a contextual scope, but such approaches are more complex, considering that potential links between the comparands need to be taken into account as well (see Mack 1996, p. 257).17

Studies with a trans-historical scope are comparisons across time and always appear in conjunction with one of the other two scopes. For example, a comparison of Hellenistic and medieval Judaism in the Eastern Mediterranean combines trans-historical and contextual scopes; a comparison of religious phenomena in medieval Europe and modern Japan combines trans-historical and cross-cultural scopes.

I argue that the general configuration of every comparative study is constituted by a certain combination of goals, modes, scales, and scopes. Analyzing existing studies with this differentiated model may not only help to identify and pinpoint potential problems but also stimulate new studies that highlight different aspects by modifying one or the other of this elements. For scholars preparing to conduct a comparative study the model provides a number of options to choose from when they configure their project. Which goal and which do I wish to pursue? Can I zoom in (or out) more, and how would that affect my project? What would happen if I modified the scope of my study in a particular way? Recognizing the respective risks and benefits of those options may lead to greater methodological awareness, and making decisions on the configuration with other possible options in mind may yield more sophisticated research.

2. The Comparative Process

In addition to the general configuration of a comparative study (by identifying goals, modes, scales, and scopes) the comparative process can be analyzed and categorized in methodical terms as well. Slightly revising and expanding Jonathan Z. Smith’s four-fold model of description, comparison, redescription and rectification (Smith 2000; described in greater detail by Burton Mack in (Mack 1996, pp. 256–59)) we may distinguish five operations that are potentially included in the comparative process: selection; description; juxtaposition; redescription; rectification and theory formation. While some activities must logically precede others (for example, an item cannot be redescribed before it has been described), most of them occur at various and often unexpected moments in the actual research process, and some are done repeatedly. For example, a redescription undertaken far into the study may cause the scholar to bring in (“select”) an additional, entirely new item and incorporate it in the comparison. Thus the order in which the five operations are presented here is analytical and pragmatic. It does not mean to suggest a neat linear, sequential procedure. It is also important to note that not every comparative study necessarily features all five operations; especially the last two, redescription as well as rectification and theory formation, are related to the respective goals of the particular study.

2.1. Selection

The selection of the sources and the tertium comparationis is arguably the most challenging operation for the comparativist.18 Put in general terms, every comparative act requires two (or more) items that are to be compared (the comparands) and a point or question with regard to which they are compared (the “third of comparison,” or tertium comparationis). For example, one may compare two religious texts with regard to certain aspects of their content, or with regard to their authorship, or to their religious significance in relation to other texts, or to their ritual function as religious objects, or with regard to any other identified feature.

All this seems fairly obvious, but a closer look reveals that the process of selecting both the comparands and the tertium comparationis is extremely complex. Multiple factors are at play in the selection process, from the researcher’s training and personal interests to cultural, academic, and disciplinary frameworks and paradigms. In addition, thorough reflection shows that the comparands and the tertium that eventually get chosen have been in a complex relationship—in the mind of the scholar and possibly also in academic discourse—long before they were put forward for comparison in an actual study. The selection of two comparands presupposes a prior act of comparison in which a productive comparability of the two was established. In other words, the assertion that two items deserve to be compared implies that they have already been compared.19 Furthermore, the comparands and the tertium may be modified in the course of the comparative process. Thus the selection process appears as all-encompassing and non-linear.20

This complex activity, labeled “selection” here, is also the least transparent of the five operations with regard to the researcher’s agency. In most scholarship in the humanities and social sciences, the reasons for why a researcher picks a certain subject for his or her study are manifold—being rooted not only in the academic discourse but also in very individual experiences, preferences, and agendas—and can rarely be traced to the full extent, even by the scholar him- or herself. This is even more relevant in comparative studies with its selection of not one, but two (or more) items and, most importantly, of the tertium comparationis which reflects the envisioned connection between the comparands (relational or analogical). The selection of the tertium in particular is closely linked to the goal of the study and thus also to the audience for which the study is intended. Since unstated agendas, unconsciously followed or intentionally concealed, can shape the research most effectively in the selection process, a high level of transparency is of paramount importance.

2.2. Description

Before juxtaposing the chosen items comparativists should provide a historical-empirical description that situates the items in their respective socio-historical and discursive contexts. In J. Z. Smith’s words, there is “[f]irst, the requirement that we locate a given example within the rich texture of its social, historical, and cultural environments that invest it with its local significance” (Smith 2000, p. 239).

Here a major issue for reflection is how an “item”—for lack of a better term I am using this generic and misleadingly reifying term as a placeholder—is to be delineated and thus separated from its “context.” Considering parallel sociological discussions about how to define a “case” (Ragin and Becker 1992), one may conclude that all items to be compared—the comparands—are, simultaneously, empirical units and theoretical constructs (see Freiberger 2018).21 The degree to which they can be identified as one or the other places them on a spectrum ranging from most theoretical to most empirical. At one end of this spectrum are highly abstract items such as fundamentalism, syncretism, or secularity. Fundamentalism in present-day America, for example, may be productively compared with fundamentalism in contemporary India. Here the comparands are primarily theoretical constructs, but they have to be studied empirically too; they need both theoretical and empirical properties. At the other end of the spectrum are items that, at first glance, may seem to be purely empirical, like a certain book or a specific piece of religious art. But these have theoretical properties too, because they have been conceptualized—both by the religious actors and by scholars—as items that can be circumscribed and delineated from their immediate context, that stand out in a certain way, and that have a specific religious value. Only this theorization makes them interesting for a comparative study. Smith’s remark about the “local significance”—which may also be called emic conceptualization—is thus closely connected to his second point: “The second task of description is that of reception history, a careful account of how our second-order scholarly tradition has intersected with the exemplum. That is to say, we need to describe how the datum has become accepted as significant for the purpose of argument. Only when such a double contextualization is completed does one move on to the description of a second example undertaken in the same double fashion” (Smith 2000, p. 239).

In practice, it must be noted, the description of the comparands in their contexts is also informed by the fact that the items will enter a comparative study. The description will highlight features that are most relevant for the subsequent comparison. The challenge lies in avoiding an overemphasis of those particular features—and in essentializing the item by reducing it to those features. The most productive studies aim at providing a comprehensive and rich description that takes the items’ historical-empirical context into consideration. The general rule is that other experts in the study of the respective context must altogether approve of the description.

2.3. Juxtaposition

The most essential operation of a comparative study is the act of juxtaposing the comparands. In the course of this juxtaposition the researcher observes and analyzes their similarities and differences with regard to the tertium comparationis. That both similarities and differences are equally important becomes apparent when we consider two seemingly contradictory perspectives on them. On the one hand, one could argue that since the comparands are separate items, that fact that they are different is obvious; the similarities need to be pointed out. On the other hand, the fact that they enter a comparison means that they have already been identified as members of the same class; what is now interesting is how they differ. Since both statements are valid, a careful and balanced approach is essential for conducting a productive comparative study. While juxtaposition is the most essential act of comparison, it is also the most individual act for each study. Where exactly the emphasis lies in the analysis of similarities or differences is determined by the goal of the particular study and the mode in which it is conducted.

2.4. Redescription

While the first three operations are inherent in all comparisons, the following two may or may not apply to a particular study. The first, redescription, is the act of describing a historical-empirical item once again in light of the insights gained from the juxtaposition with a different item. As Mack puts it, “[i]t may be that something will have been learned about factors that make the two situations similar, something about the difference another myth makes, something about the reasons for a people’s interest in or fascination with a particular notion, role, or activity, and so forth. These insights will change the way in which the examples under investigation are understood and thus require redescription. A redescription will register what has been learned in the study” (Mack 1996, p. 258).22

This act of redescription particularly applies to studies conducted in the illuminative mode. Studying an item through the lens of a different one, observing previously unnoticed features, discovering blind spots, etc. may result in a new description of the item that is more comprehensive or more refined. The new description of an historical-empirical phenomenon reflects the progress in scholarship that has been made as a result of the comparative study. Future studies of this item and its context will have to recognize and consider the revised description. In some studies illuminating phenomena by means of comparison happens in both (or more) ways. Arvind Sharma has called such a multidirectional process “reciprocal illumination” (Sharma 2005).

2.5. Rectification and Theory Formation

With the act of rectification “the academic categories in relation to which [the exempla] have been imagined” are rectified (Smith 2000, p. 239). Unlike redescription, rectification does not refer to the analysis of a particular historical-empirical item but to a revision of the definition and conceptualization of the (meta-linguistic) categories involved in the study. In contrast to Smith, who elsewhere blends the two terms when he says that redescription “expressed a central goal, the redescription of classical categories to the end that these be ‘rectified’” (Smith 2004, p. 29), I follow Mack’s understanding of the term that to rectify a category is to “rename the phenomenon of which our case studies are examples” (Mack 1996, p. 258). This separation seems analytically useful: We redescribe a concrete item in its historical, object-linguistic context, and we rectify a metalinguistic category.

Rectification is particularly relevant for studies that are conducted in the taxonomic mode. The comparison of “species” results in a better conceptualization of the “genus.” A cross-cultural and cross-tradition comparison of particular relic practices, for example, can result in identifying broader theoretical dynamics that enrich the scholarly conceptualization of the category “relic” (Trainor 2010). A thorough comparison of medieval Christian and Tibetan Buddhist texts can offer new insights on how “hagiography” works (Rondolino 2015, 2017).

Rectifying metalinguistic categories is an act of theorizing. According to the most general definition in the Oxford English Dictionary a theory is “[t]he conceptual basis of a subject or area of study” (Oxford English Dictionary Online 2015, s.v.). We may specify that theory can be regarded as a conceptual network of a certain area of study in which various metalinguistic categories are structurally interlinked. A comparative study may result in rectifying existing categories, but it may also lead to the suggestion of new ones. It may even help revise or create more complex theoretical formations. For example, a comparison of ancient Greek and early modern Indian texts can complicate the typologies of links between religion, gender, and violence (Pasche Guignard 2015).

This contributes, eventually, to a better theoretical understanding of religion. Again, in Mack’s words, “The point is nothing less than the construction of a theory of religion. A new designation for a recognizable phenomenon can become a building block for constructing a descriptive system. And the descriptions of phenomena in such a studied system can actually become mid-range axioms that might eventually be used to build a cultural (and in Smith’s case, cognitive) theory of religion” (Mack 1996, p. 259). In other words, a comparative study may result in the formation of a theory about a certain religious phenomenon, and this theory can be incorporated in a larger theory of religion. While this is one possible outcome, most comparative studies operate on lower—but equally relevant—levels of abstraction.

3. Conclusions

The methodological frame proposed here, addressing both the configuration of a comparative study and the comparative process, describes the comparative method in the study of religion. While many aspects could not be explored in greater detail, this outline already shows, I hope, that comparison is a rather complex operation. The proposed categories are meant to work as an analytical grid. The fact that some of them are flexible and that several of their subcategories may overlap acknowledges the dynamics of comparative research. Yet I argue that every comparativist makes decisions that eventually result in a certain configuration of the chosen goal(s), mode(s), scale, and scope(s). This configuration in combination with the way one chooses to proceed through selection, description, and juxtaposition methodically defines a particular comparative study. That implies that if some decisions were to be made differently—for example, choosing a meso-scale rather than a micro-scale or a cross-cultural rather than a contextual scope—, the study’s conclusions could be quite different. Thus, if it is possible to differentiate activities and identify particular choices, we might be able to envision comparison, much clearer than we used to, as an organized and controlled operation.

I argue that a refined comparative methodology provides useful tools for assessing the scholarly value of individual comparative studies. It can expose, among other things, sloppiness, imbalanced juxtapositions, or hidden agendas. In addition, a robust methodology of comparison also enables comparativists to confront, with factual analysis, certain sweeping criticisms of the comparative method.23 Much of what is regarded as problematic about comparison in the study of religion (decontextualization, essentialization, undue generalization) should be discussed, in my view, not only in theoretical but also in methodological terms. Identifying and isolating specific methodical problems in a study effectively confronts wholesale criticism and, at the same time, provides an opportunity to refine the methodology. At the same time, reflecting upon the various elements of this framework may help comparativists to configure and adjust the layout of their studies, justify their decisions, and possibly raise the study’s level of sophistication.

Certainly I do not claim that the proposed frame is the only possible way to describe the comparative method. An entirely different model of comparison might focus, for example, more on the intuition of the scholar and explore how exactly this intuition works, how it is developed, and how it distinguishes itself from the intuition that is present in other scholarship in the humanities and social sciences. If comparison is understood as a (second-order) method, there is no reason why it should not benefit from the exchange between, and competition of, different methodological approaches—just like we see it in the ongoing debates, in the respective fields, about the most appropriate philological, sociological, or anthropological method. The purpose of this essay was not to outline what I think should be an authoritative model but rather to propose a starting point for a serious, comprehensive, and productive debate about the methodology of comparison in the study of religion—a debate that, in my view, is long overdue.