Thickening description: towards an expanded conception of philosophy of religion

An increasingly common complaint about philosophy of religion—especially, though not exclusively, as it is pursued in the “analytic tradition”—is that its preoccupation with questions of rationality and justification in relation to “theism” has deflected attention from the diversity of forms that religious life takes. Among measures proposed for ameliorating this condition has been the deployment of “thick description” that facilitates more richly contextualized understandings of religious phenomena. Endorsing and elaborating this proposal, I provide an overview of different but related notions of thick description before turning to two specific examples, which illustrate the potential for engagement with ethnography to contribute to an expanded conception of philosophy of religion.

A common and thoroughly understandable complaint about much contemporary philosophy of religion is that its scope remains abysmally restricted. While purporting to investigate religion in general, most of what goes by the name of philosophy of religion continues to be preoccupied with a narrow range of questions concerning an ahistorical and decontextualized "theism." The restriction is twofold. First, the fixation on a homogeneous theism inhibits a fuller cognizance of the variety of forms that religions, both theistic and nontheistic, can and do take; that is, it inhibits due consideration being given both to the discongruities between different forms of theistic religion, whether within or beyond the Abrahamic traditions, and to nontheistic (or not straightforwardly theistic) traditions such as we see in certain versions of Buddhism, Jainism, Daoism, nondualist Vedānta and the indigenous traditions of many small-scale societies around the world. Second, the preoccupation with "beliefs" (commonly construed narrowly in terms of "propositional attitudes") and with the arguments supposed to support them diverts attention away from other aspects-what Ninian Smart (1996) termed the other "dimensions of the sacred," such as the ritual, mythic, experiential, ethical, legal, social, material and political dimensions. 1 Calls for more inclusive approaches are far from new. Not only Smart but also figures such as John Hick were pioneering in their advocacy of attention to religious ideas from outside the Abrahamic milieu; and over recent decades several philosophers of religion have made concerted efforts both to incorporate a wide array of religious traditions into the philosophical conversation 2 and in some cases to expand the discipline's purview to encompass philosophical analyses of aspects such as myth, ritual, physical location and prayer. 3 However, as perceptive innovators such as Timothy Knepper (cited in my first epigraph above) have argued, continued effort is needed if philosophy of religion is to genuinely reflect the diversity that obtains within the religious lives of humanity in general rather than only a small portion of it. This effort is important because it bears upon the extent to which philosophers of religion can legitimately claim to have got a firm grasp on the very phenomenon they profess to be studying. While it would be unreasonable to expect any philosopher to gain a thorough understanding of even a modest subset of the enormous range of religious traditions and phenomena that have characterized the cultural lives and histories of humankind, there remains an onus upon those who wish to speak about religion in general to explicitly recognize the diversity.
One way of conceptualizing the problem is in terms of abstraction. When philosophers of religion speak of "theism" or "bare theism" or "standard theism," etc., what they are often trying to do is enable their arguments to apply to more than one sect, denomination or religion. They may presume, for example, that Judaism, Christianity and Islam, notwithstanding the multiple differences between these 1 As Smart (1999) further notes-though this is not my main focus in this article-broadening one's analysis to encompass these dimensions also helps to highlight the complex interrelations and overlaps between "religious" and "secular" worldviews, thereby helping to avoid distortedly essentialist conceptions of religion. 2 See, e.g., Clayton (2006), Griffith-Dickson (2005), Kessler (1999) and the series of books by Arvind Sharma taking different perspectives on the philosophy of religion, such as Sharma (1995). 3 See, e.g., Schilbrack (2002Schilbrack ( , 2004 and Wynn (2009). For a feminist perspective on the importance of attention to religious practice in addition to "belief," see Hollywood (2004). For a study of prayer (albeit an older classic rather than a recent contribution), see Phillips (1965). religions and the sectarian divergences within them, are all "theistic." 4 On this account, to speak of "theism" rather than more specifically of Judaism, Christianity or Islam is a move in the direction of inclusivity. Some philosophers have taken this abstractive approach a stage further, by replacing talk of "theism" with talk of something common to all religions, be they "theistic" or "nontheistic." For John Hick (e.g., 2004), the unifying principle is "the Real"-a principle which no religion explicitly recognizes but which, according to Hick, they are all implicitly directed towards. Thus, although Hick was indeed a pioneer in promoting a more inclusive conception of religion, his conception could be regarded as unhelpfully homogenizing insofar as it presumes all religions to have a single essence. While calling his approach "pluralism," he ends up with a diminished appreciation of what D. Z. Phillips (2007) has termed the radical plurality of religious and nonreligious perspectives on the world.
A hazard of any abstractive move is the loss of contact with the real-life phenomena to which one's discussion was intended to have relevance in the first place. If the god of the philosophers-let alone some noumenal "Real"-is not a god to whom anyone wittingly prayed or bowed down in worship, then a problem arises for the whole enterprise of an abstractive philosophy of religion: to the extent that it loses touch with people's actual religious lives, it risks leading us away from rather than towards a deeper understanding of religion.
The issue of how to expand the scope of philosophy of religion is thus inextricably linked to matters of methodology. Without needing to deny that certain forms of abstraction can have a legitimate place in philosophical investigations of religion, it is worth exploring what other options are available. One option that a growing number of philosophers have begun to explore is that of increasing interdisciplinary engagement between philosophy and certain areas of the empirical study of religion, including the anthropology and sociology of religion along with the broad multidisciplinary area known as religious studies. A term that has gained prominence in these more empirically oriented disciplines since the 1970s is "thick description," which many scholars, following Clifford Geertz (1973), have adopted to characterize forms of description that endeavour to go beyond the mere surface of a given event or situation or state of affairs-to show the various layers of meaning or significance that constitute its place within the broader context of human lives. The term "thick description" in fact has its origins in philosophy, having first been coined in a pair of essays by Gilbert Ryle (1968aRyle ( , 2009Ryle ( [1968b), and hence recent suggestions that philosophers of religion would do well to deploy thick description themselves (Knepper 2013(Knepper , 2014Burley 2015, p. 236) are not calls simply to help oneself to methodological resources from empirical disciplines: they are in a sense also calls for philosophers to revive and develop philosophical methods that have lain largely dormant for several decades.
While the latter methodological suggestion is a fruitful one, there is still much work to be done in showing how it is to be realized in practice. The purpose of this article is to make a contribution to that task both by elucidating the concept of thick description, which, as we shall see, has itself become a highly variegated concept, and by exemplifying how thick description might promote a more nuanced philosophical appreciation of the ramified and multifarious possibilities of human religious life. This appreciation, in many ways, makes the project of philosophy of religion harder, as it brings to light the messiness or raggedness of religion, when many philosophers may prefer their object of inquiry to have smoother edges. But if our aim is to gain a deeper understanding, then smoothing over the rough edges is unlikely to serve that purpose. "What's ragged should be left ragged," as Wittgenstein concisely put it.

Varieties of thick description
The term "thick description" has had a long and complicated history since its first use by Ryle in two papers published in 1968. Ryle's purpose in introducing the term was to distinguish between different degrees of complexity and richness with which a description of someone's behaviour might be imbued. Although many commentators have tried to capture Ryle's distinction in terms of a general account of what "thin" and "thick" descriptions consist in respectively, it is doubtful whether any such general account could be successful, since the terms are relative both to each other and to the subject matter of the description. Thus it is advisable to think in terms of thicker and thinner-or more and less thick-descriptions, as opposed to descriptions that are, categorically, either thin or thick.
To illustrate the distinction, Ryle adduces several examples, the best known and most thoroughly worked out of which is that of the winking boys. Ryle begins the example by inviting us to suppose that two boys each do something that could, in the thinnest terms, be described as contracting their right eyelids. But one of the boys is merely twitching involuntarily whereas "the other is winking conspiratorially to an accomplice" (2009, p. 494). This conspiratorial winking, Ryle reminds us, is a single action comprising several elements. Hence to describe it fully-to describe it thickly-one should mention that it is a deliberate act intended to impart, without anyone else noticing, a particular message to a specific person in accordance with a previously agreed code (p. 495). Ryle subsequently imagines a third boy who wishes to parody the conspiratorial winker. In practising the parody, this third boy's action could be described in terms of multiple nested forms of trying: he is trying to prepare to try to entertain his friends by imitating someone trying to convey a secret message to an accomplice by trying to contract his eyelids (p. 496). As Ryle remarks, "The thinnest description of what the rehearsing parodist is doing" would be approximately equivalent to that of "the involuntary eyelid twitch; but its thick description is a many-layered sandwich, of which only the bottom slice is catered for by the thinnest description" (pp. 496-497).
We can see, then, that for Ryle there is no blanket answer to be given to the question of whether a thicker or a thinner description is appropriate, for appropriateness will depend, minimally, upon what is being described. Thus if someone has rapidly closed and opened her eye as a consequence of a habitual twitch, a thin, uncomplicated, description is likely to suffice. But in other instances there may be several descriptive layers required in order to capture what the action is that is being performed. Nor would it be unduly stretching Ryle's point to add that the appropriateness of a thicker or a thinner description will also hinge, in part, upon the purpose for which the description is being offered-or requested. For example, if one friend asks another what she was doing last night, and the second friend replies that she went out for a walk, the description might be adequately thick for the purpose, whereas if the same person is being interviewed by police as a crime suspect, the interviewing officers are apt to be hoping for more detail, more thickness.
Despite its having been coined by Ryle, the term "thick description" has become most closely associated with the anthropologist Clifford Geertz-the veritable "mahatma of 'thick description'" (Shweder 2005, p. 1)-who invokes it to define what he understands the research method known as ethnography to consist in. Notwithstanding the contrived nature of Ryle's example of the winking boys, Geertz (1973, p. 7) proposes that it "presents an image only too exact of the sort of piled-up structures of inference and implication through which an ethnographer is continually trying to pick his way." Several commentators have argued that there are considerable differences between Ryle's and Geertz's uses of the term "thick description." Indeed, it has been claimed that Geertz turns the term's meaning "on its head" (Descombes 2002, p. 439;Bazin 2003, p. 432, n. 22) by conflating description with interpretation. According to this contention, Ryle is concerned to explicate an order of complexity that is strictly logical: to describe someone as winking is, as it were, a first-level description, whereas to describe that person as winking conspiratorially brings in a second level-it combines verb with adverb in a logical structure that, in principle, could be added to indefinitely; Geertz, by contrast, is declaring that for any situation or occurrence in which more than one agent are involved, there will be multiple interpretations that the respective agents have of what is going on, and thick description consists in the ethnographer's attempt to build a determinate picture-to "construct a reading of" (Geertz 1973, p. 10)-the relevant situation out of the available data, including the agents' respective interpretations. There is thus a sense in which ethnography is a thoroughly interpretive endeavour: "explicating explications," as Geertz puts it -"Winks upon winks upon winks" (p. 9).
Against those who accuse Geertz of misunderstanding and hence of misappropriating Ryle's notion of thick description by conflating description with interpretation, it could be argued that the difference between presenting a contrived though culturally familiar and entirely plausible example on the one hand, and, on the other hand, describing real-life occurrences in cultures very different from one's own (and with which most of one's readers are unlikely to be familiar) is of considerable importance. Ryle and his readers can readily understand what it means to wink conspiratorially, and to imitate someone winking conspiratorially, and to practise imitating someone winking conspiratorially, and so on; and Ryle, as the author of the example, has no need to interpret for his readers what is going on: he simply describes it, as would the author of a novel. For this reason, plain description, albeit adverbially or logically "thick," accomplishes its task of conveying the requisite information relatively straightforwardly. In the ethnographic context, however, the researcher often encounters difficulties in simply understanding what the agents are doing. Figuring this out will commonly involve attending to an extensive range of contextual factors that include what the various agents say about what they are doing. While these latter accounts will not differ from one another in terms of logical order, a description that draws upon all or several of them will nevertheless comprise a certain layering or thickness, in something like the way in which the sound of a choir, or even of a cocktail party, exhibits a thickness that is lacking in a single voice. It is, then, possible to view Geertz's appropriation of "thick description" as an extension or adaptation-or recontextualization-of Ryle's original coinage without our needing to disparage it as a "misinterpretation" (pace Bazin 2003, p. 432, n. 22).
What Geertz's appropriation made possible was an explosion of further appropriations of thick description within the social disciplines and humanities more generally. Instrumental in this popularization of the term, and of the various descriptive methods to which it has been attached, is work by Norman Denzin, whose expositions of qualitative methods of data collection and analysis have influenced generations of qualitative researchers since the 1970s. 5 Among Denzin's innovations is the devising of a typology of eleven different forms that thick description can take (Denzin 2001, ch. 6). I do not have the space here even to merely summarize all eleven of these, but there are a few recurrent characteristics that are worth mentioning. These characteristics include the capturing and recording of "the voices of lived experience"-what Denzin, borrowing from Merleau-Ponty (1973), terms the "prose of the world"-thereby creating "verisimilitude, a space for the reader to imagine his or her way into the life experience of another" (Denzin 2001, p. 99). Thus not only voices in the literal sense, but also the "feelings, actions, and meanings of interacting individuals are heard, made visible" (p. 100).
Ironically, by Denzin's explicit criteria Ryle's original examples would count only as "typified, thinly veiled thick description" as opposed to "actual thick description," since the latter, according to Denzin (2001, p. 106), must describe experiences or actions of real-life individuals rather than of concocted characters. Denzin is, however, inconsistent on this point, as he has no trouble with invoking passages from literary fiction in order to illustrate what he means by "biographical thick description" (p. 108), which he appears to treat as genuine rather than as merely "thinly veiled." Given that descriptions of fictive scenes can be just as rich in such features as contextual detail and verisimilitude as can descriptions of real-life scenes, there seems little justification for regarding fictitiousness as precluding a description from being genuinely thick. But if we are thinking in typological terms, there are certainly good, and obvious, reasons for distinguishing between fictive and non-fictive descriptions, even if we acknowledge that in many instances the distinction is liable to become blurred.
For my purposes in this article, I am content to refrain from stipulating a tight definition of thick description. The impression one is likely to glean from a survey of the literature on this topic is that "thick description" is or has become, to use Wittgenstein's term, a "family resemblance" concept, with "a complicated network of similarities overlapping and criss-crossing" but no characteristics that are necessarily common to all uses (Wittgenstein 2009(Wittgenstein [1953, § §66, 67). 6 It is, nevertheless, a serviceable concept, convenient for characterizing certain modes of description in comparison with others. As I have suggested above, it is unlikely to be very informative to propose that a description is either thick or thin simpliciter. But it does not follow from this that it cannot be informative to describe one description of a given phenomenon as thicker than another, perhaps while also noting the particular respects in which its relative thickness resides. 7 Thickly descriptive philosophy of religion Notwithstanding its Rylean derivation, "thick description" is not a term that receives much recognition from philosophers these days, whether in the philosophy of religion or in any other branch of philosophical inquiry. There has, in recent years, been considerable attention devoted in metaethics to "thick concepts," and it has occasionally been noted that this latter term, originated by Bernard Williams (1985), was probably influenced by Ryle's notion of thick description (Kirchin 2013, p. 60;Väyrynen 2013, p. 1); but discussions directly about thick description are few and far between in the philosophical literature. Among the exceptions to this is recent work by Knepper (2013Knepper ( , 2014, who has called for thick description to be undertaken specifically by philosophers of religion. Knepper, in addressing the issue of how to promote approaches to philosophy of religion "that are more historically grounded and religiously diverse" (2013, p. x; 2014, p. 137) than has typically been the case, recommends a threefold methodology comprising "thick description, formal comparison, and multidimensional explanation and evaluation of religious reason-giving" (2013, p. 75). With regard to the first of these elements, Knepper argues for the claim that "thick description of religious reason-giving in all religions for the sake of critical understanding comes first" (p. 75, italics omitted), contending that in order to provide the sort of thick description that he has in mind, one would need to consider instances of reason-giving across a range of religions and in relation to the contexts in which the reasons are put forward. This would itself involve asking questions such as who "the speakers, audiences, and opponents" of the reasons are and under what conditions these reasons are ever "contested, modified, and abandoned" (pp. 37-38).
Although Knepper's proposals are innovative and potentially transformative of the subject, it is unclear why it should be "reason-giving" that is the primary object of thick description, whether in the philosophy of religion or anywhere else. The examples that Knepper offers to illustrate what he means by "religious reasongiving" are certain arguments presented by philosophers or theologians, such as the arguments advanced by Ś aṅkara in support of the "nondualist" view that, as Knepper puts it, "there is no distinction between Brahman and Ātman" (2013, p. 80) or, in Mahāyāna Buddhism, "Nāgārjuna's argument that saṃsāra is nirvāṇa" (p. 81). While these are no doubt worthwhile arguments to be investigating, by focusing as closely as Knepper does upon the ideas of reason-giving and argumentation there is a danger of overlooking the need for a more fundamental level of investigation, which would consist in the task simply of trying to understand what is meant when terms such as brahman, ātman, saṃsāra and nirvāṇa are used. What Knepper calls instances of reason-giving will indeed be among the discursive contexts in which these terms occur, but they are not the only such contexts. A fixation on reason-giving, especially where the paradigm cases are assumed to be specimens of philosophical or theological argumentation, is liable to delimit the benefits that thick description can bring to philosophy of religion.
Amid religious forms of life there will be instances of reason-giving, and these will occasionally take the form of arguments, but religious life amounts to much more than that. If philosophy of religion in general and the development of thickly descriptive approaches to philosophy of religion in particular were to be preoccupied exclusively with religious reason-giving, this would lead to a lopsided conception of what goes on in religion. Whether, or to what extent, reason-giving plays a part in religious activities ought not to be presumed in advance, but may be disclosed by investigations that are not hampered by an unduly narrow focus at the outset. What I propose is that thick description can enrich and broaden a philosophical appreciation of the diversity or radical plurality not merely of religious reason-giving but of varieties of religion more generally.

Radical pluralism: questioning lazy assumptions about religion
The term "radical plurality" or "radical pluralism" was formulated by the philosopher D. Z. Phillips (2007, pp. 203-211) in order to capture both the nature of religious phenomena in the contemporary world and an approach to the study of those phenomena that seeks to do "conceptual justice" to religious and to nonreligious perspectives in all their variety. Phillips contrasts radical pluralism with the kind of religious or theological pluralism advocated by Hick (e.g., Hick 2004, Part 4). While pluralism of the latter sort presumes that all religions share certain essential features, such as being directed towards a single divine reality or espousing an ethic of mutual love and compassion, radical pluralism remains open to the multifariousness of ontological and ethical commitments among people of different religious and nonreligious points of view.
An important feature of a radical pluralist approach is a deliberate unwillingness to neglect aspects of religion that one finds personally troubling or even repugnant. Far from deciding in advance of the investigation what kinds of ethical or other principles must be respected for any given perspective to count as properly religious, the radical pluralist actively seeks out counterexamples to any general thesis about what a religious perspective "must" consist in. It is by that means that lazy assumptions are challenged and a richer conception of the complexity of religious phenomena is developed; it is in that spirit that, in my view, thickly descriptive philosophy of religion is best carried out.
As we have seen, when Knepper offers examples of topics that call for thickly descriptive investigation, he foregrounds occasions of reason-giving that take the form of philosophical or theological arguments. This, we might think, is a reasonable approach for a philosopher to take, given that philosophers are trained specifically in methods of appraising arguments, normally in the form of written or spoken discourse, rather than in gathering descriptive information about the lives and daily interactions of complex sociocultural communities. Philosophers, after all, typically do their research at their desks and in the library rather than "in the field." However, the fact that philosophers are not usually well equipped to undertake fieldwork themselves need not preclude their adopting approaches to the study of religion that exceed the mere analysis of arguments. Nor need it entail that the only thick description in which a philosopher can engage is the devising of imagined scenarios of the sort offered by Ryle. Among the other options available to philosophers is that of drawing upon existing published material that provides thick description of religiously relevant human phenomena. Such material, as I have argued elsewhere (Burley forthcoming), could be in the form of insightful works of narrative fiction, including literature, films and plays, or in the form of biographical or autobiographical accounts; but so too, as I am proposing here, could it comprise ethnographic studies by professional anthropologists. Indeed, familiarizing oneself with a wide variety of ethnographic studies is one means of breaking out of the prevalent philosophical obsession with, precisely, the aspects of religion that involve reason-giving and argumentation, and thereby avoiding the fettered understanding of religion that commonly results from a restricted palette of examples.
In order to illustrate how a philosophical appreciation of the radical plurality of religious phenomena can be enhanced through engagement with thick description derived from ethnographic reports, I shall in the remainder of this article consider two poignant examples. Although further examples and more detail than I have space to provide here would be needed to make a thoroughly convincing case, I hope the following discussion demonstrates at least the potential of the approach at issue.

Example 1: Animal sacrifice at the Kāmākhyā temple
A focus on the textual and, more specifically, the doctrinal or philosophical sources of religious traditions has eventuated in a widespread neglect among philosophers of religion of religious practices in general and of ritual practices in particular. 8 In that light, it is thus of little surprise that ritual animal sacrifice is very rarely mentioned in the philosophical literature despite its prevalence in many religions throughout history and across the world. It is perhaps difficult for many contemporary philosophers of religion to know what to say about animal sacrifice, and this difficulty is accentuated when the philosopher's aim is to present religion as an overwhelmingly benign and morally edifying force in human life: the ritual killing of animals-let alone of human beings-does not fit neatly into the sanitized picture of religion that many wish to paint. If, however, our aim is to do justice to the radical plurality of religious phenomena, then, regardless of our personal moral attitudes, blood sacrifices should not be excluded from our philosophical purview.
Among the many places in which animal sacrifice features prominently in contemporary religion are various forms of goddess worship in India and Nepal. An especially graphic description of such ritual practices is offered by Patricia Dold (2011), who undertook collaborative fieldwork at the site of the Kāmākhyā temple in the north-eastern Indian state of Assam during 2008 and 2009. Kāmākhyā is the goddess of desire (kāma), and the Assamese temple complex in her honour dates back to the sixteenth century (von Glasenapp 1928, p. 121). One of the major annual festivals at the temple is dedicated to a snake goddess, Manasā, whose mythology emphasizes her benevolence to devotees combined with belligerence towards "those who refuse to worship her" (McDaniel 2004, p. 148). The festival, known as Debaddhanī or Deodhanī ["sound or echo of the deity" (Smith 2006, pp. 140-141)], takes place over a three-day period in August.
Animals are sacrificed daily at the Kāmākhyā temple all the year round, but this escalates during festival periods. Integral to the proceedings of the Debaddhanī festival is the ritual killing of dozens of animals, including goats, pigeons, cockerels and water buffaloes. At the centre of the devotional activities is a group of between nine and fifteen men who have temporarily become "possessed" by deities such as the god Ś iva and several forms of the goddess, including Kālī and Kāmākhyā. For the duration of the festival these men, known as Deodhās, become "sites of divinity" (Smith 2006, p. 141), the highpoint of the festivities being a ceremonial dance in which two of the Deodhās hold a scimitar blade-upwards while other Deodhās dance barefoot upon it to the accompaniment of vigorous drumming. It is this event that Dold describes in especially vivid terms. Dold (2011, p. 55) recounts how, as the Deodhās are taking turns to dance on the sword, there comes a moment when the mood of the observing crowd intensifies and the people push forward "as a priest appears carrying a live pigeon." The pigeon, with its wings flapping, is lifted by the priest in front of the mouth of the Deodhā who is now standing on the sword edge. "With a dramatic chomp of his teeth and shake of his head," the Deodhā bites off the bird's head, an act that is repeated two or three times by each of the Deodhās. As women in the surrounding crowd issue the "piercing, high-pitched, and triumphant cry 'ooolulululu'," each Deodhā in turn "chews, swallows, and then spits out clumps of flesh and feather." In emulation of the deities embodied in them, "the Deodhas are receiving animal offerings." Dold's description of these occurrences is "thick" in the sense that it evocatively captures something of the emotional atmosphere of the situation in addition to reporting the basic sequence of events. It constitutes what Denzin (2001, p. 99) terms "a form of performative writing" that enables "the willing reader to share vicariously in the experiences" being described. Although the details of the events at issue are liable to stir feelings of disgust in many readers, it is precisely the perspicuity of the description that facilitates this reaction. Framed by her exposition both of devotional songs recited during the festival and of traditional mythic stories involving deities relevant to the forms of worship that take place, Dold's first-hand description fosters a deepened understanding of the kind of meaning or significance that animal sacrifice has within the context of worship-albeit a meaning or significance that may not be straightforwardly articulable in terms other than the description itself. There may thus be a sense in which, as Wittgenstein proposes in response to James Frazer's attempts to explain sacrificial rites, "Here one can only describe and say: thus is human life" (1993, p. 121; translation amended).
Wittgenstein's point is not that our understanding of a ritual cannot be enhanced by further contextual information; on the contrary, adding contextual layers is exactly what thick description involves. The point, rather, is to draw a distinction between description and explanation: to urge caution about assuming that what understanding a ritual consists in must be the crafting of an explanatory account, perhaps couched in terms of what the ritual participants intend to achieve. At least in many instances, Wittgenstein is suggesting, such purported explanations may in fact deflect our attention from what is most significant in the ritual, leading us to characterize what is going on in unduly intellectualized terms, as seeking to fulfil an instrumental goal that is external to the ritual itself. Instead of immediately looking for an explanation, Wittgenstein is recommending that we dwell longer with the description, reflecting upon the kinds of feelings it stirs in us. In cases where the ritual involves a blood sacrifice, whether this be a human or an animal offering, one "answer to the question 'why is this happening?'" might be simply "Because it is terrible. In other words, what strikes us in this course of events as terrible, impressive, horrible, tragic, &c., anything but trivial and insignificant, that is what gave birth to them" (Wittgenstein 1979, p. 3e).
Dold herself does follow her description by briefly considering the meaning of the ritual. Responding to the question of whether "bloodlust" plays a part, she notes the regularity with which animals are sacrificed at the Kāmākhyā temple and the "serious solemnity" with which these sacrifices are observed. Indicative of the sanctity of the animals' deaths, Dold proposes, is the fact that priests forbid the taking of photographs at these moments. 9 Denying that sacrifice at Kāmākhyā is "a celebration of death or of the power to kill," Dold speculates that what may be being celebrated in the Deodhās' receiving of animal offerings is "the living presence of goddesses" (Dold 2011, p. 55).
A curious feature of Dold's speculation is why she thinks the sacrificial rites could not be both "a celebration of death or of the power to kill" and a celebration of "the living presence of goddesses." After all, Dold's description of the theatricality with which the pigeons' heads are bitten off and of the triumphal vocalizations of the assembled crowd is highly suggestive of an occasion that celebrates the power to kill while, in view of the ferocious depictions of goddesses in relevant Hindu mythology, also graphically displaying the presence of the goddesses in the rituals' protagonists. Indeed, it would seem highly plausible to propose that it is through the celebratory acts of killing that the presence of the goddesses is enacted. Thus, while it is entirely understandable to want to look for more than mere "bloodlust" in the ritual, there is no need to discount the possibility that celebrating death and the power to kill has something to do with it. We might also, following suggestions from Rush Rhees and Phillips, see in the ritual an expression of "wonder at what is terrible and what is evil … treating what is terrible as a sacrament" (Rhees 1994, p. 578;cf. Phillips 2007, p. 205). None of this gives us an explanation of the ritual, in the sense of attributing deliberate motivations or intentions to the ritual participants; rather, it offers "possibilities of sense"-ways of aiding those to whom the ritual is alien and remote to see how it might come to have a place in a form of religious life. Such observations of possibilities of sense are among the contributions that philosophers as well as anthropologists can bring to the discussion of descriptive accounts, though it is upon the anthropologists that the philosophers are dependent for furnishing the descriptive accounts in the first place.

Example 2: Buddhism and bereaved Sherpas in Nepal
The Sherpas are an ethnic group residing primarily in the Himalayan mountain range in northern Nepal. Male Sherpas are routinely employed by mountaineering parties from western and other countries both to act as guides and to help with the carrying of equipment and other tasks during climbing expeditions. Owing to the extreme dangers involved in such expeditions, many Sherpas have suffered death or injury as a consequence of this mode of employment, generating considerable emotional trauma among Sherpa communities. An anthropologist who has studied the Sherpas in depth is Sherry Ortner, in some of whose work particular attention is given to Sherpa responses to the deaths of fellow Sherpas in climbing accidents.
It is often assumed by foreign mountaineers that Sherpas, who are predominantly Buddhists, will be both fearless in the face of danger and emotionally detached when accidents occur. A stereotype has developed of the steadfast Sherpa who is hardy, sanguine and unselfish. Ortner documents how, though not without a grain of truth, this stereotype is extremely one-sided. In reality Sherpas do sometimes feel dispirited or even terrified by the circumstances they come up against on expeditions, but these reactions are rarely tolerated by the mountaineers who engage the Sherpas' services. Drawing upon Edward Said's (2003Said's ( [1978) analysis of "Orientalism"-where the latter term denotes a discourse that consolidates power imbalances by "essentializing" Asian subjects as inferior and subordinate-Ortner points out how Sherpas have been vulnerable to an Orientalist double bind. Displaying anxiety or other negative emotions when accidents occur is viewed contemptuously by foreign mountaineers, yet remaining impassive at such times is liable to be perceived by those same mountaineers "as a strange, and almost inhuman, form of 'Oriental fatalism'" (Ortner 1997, p. 141).
While acknowledging, then, that the kind of analysis of power relations that Said inherited and adapted from Michel Foucault has a legitimate role to play in the study of certain Asian communities, including that of the Sherpas, Ortner resists the facile tendency, characteristic of some academics, to view this style of analysis as precluding approaches that seek to expound the meanings of cultural phenomena, especially religious meanings. Approaches of the latter sort are typified by Geertz, who, as has been noted above, was instrumental in developing and popularizing the notion of thick description in ethnographic writing. What Ortner (1997, p. 157) argues is that "a rich and complex conception of culture and meaning in the Geertzian sense" is not only compatible with but is complementary to an examination of the discursive systems that instantiate and fortify "regimes of power/knowledge," in which the interests and the very lives of relatively disempowered individuals and groups are subordinated to those of others. Far from distracting us from issues "of how actors formulate needs and desires, plans and schemes, modes of working in and on the world," Geertz's work offers "a model for understanding the cultural construction of 'agency' in particular times and places" (p. 158).
Within the context of this analysis of agency, Ortner presents an especially poignant excerpt from her 1979 field notes. The excerpt describes an exchange between the family of a Sherpa named Ang Phu, who had recently been killed while assisting a mountaineering party on Everest, and a Buddhist Rinpoche (high lama) with whom some other Sherpas and Ortner herself had been having lunch. Ortner (1997, pp. 155-156;1999, p. 141) recounts how Ang Phu's father sobbed as he prostrated and pleaded for the Rinpoche to grant a blessing. In response, the Rinpoche remained "cool," telling the man that, as a religious practitioner, he ought to control his emotions, read prayerful texts and be a calming influence upon his family. Then, Ortner reports, the Rinpoche "closed his eyes and chanted quietly to himself for two or three minutes," after which he recommended some passages from his divination book that the father should read. Contemplating her own reaction to the situation, Ortner remarks that she had felt shocked by the Rinpoche's apparent "lack of nyingje"-a term that can be translated as "compassion, sympathy, pity"and by the jocularity of his tone. Later, however, Ortner is told by one of her Sherpa friends, Nyima Chotar, that this is simply how high lamas behave: even if it were to be the lama's own mother or father who had died, the lama's grief would dissipate after only five minutes' meditation. Responding to Ortner's contention "that the lama showed no nyingje," Nyima Chotar replied that what the lama had displayed "was nyingje" (Ortner 1999, p. 141).
As in the case of Dold's account of the sacrifice at Kāmākhyā, Ortner's description of the encounter between Ang Phu's bereaved family and the Rinpoche is "thick" inasmuch as it brings the situation to life for the reader, affording an insight into the emotional demeanour and mutual interaction of the participants.
Unlike contrived examples such as Ryle's delineation of the boy who contracts his eyelids in order to practise imitating another boy's wink, Ortner's description does not immediately disclose the intentions of the Rinpoche, for it is precisely these intentions that are initially opaque to Ortner. But there is sufficient detail in the description to enable the reader to make sense of Ortner's subsequent interpretations, arrived at by reflecting upon the Rinpoche's behaviour in the light of both Nyima Chotar's comments and Ortner's knowledge of Sherpa Buddhist culture more generally.
One of the key issues accentuated by the described incident along with Ortner's own reflections upon it is the extent to which ethical and religious principles, such as that of compassion, can be highly ramified and culturally inflected. When philosophers of religion such as Hick or popular historians of religion such as Karen Armstrong declare that, for example, love or compassion is an ethical ideal that is "common to all the great traditions" (Hick 2004, p. 14) or that "All the world faiths insist that true spirituality must be expressed consistently in practical compassion" (Armstrong 2010, p. 8), there are good reasons to be wary about such generalizations. One reason is that the very idea of "the great traditions" or "the world faiths" is modelled on a "world religions paradigm" that has come increasingly to be questioned by scholars who point out how it tends to obscure from view the heterogeneity that exists both within and outside the purportedly great traditions (see esp. Owen 2011). A second reason for wariness is that since notions construable in terms of "love" or "compassion" may in fact not be prevalent in all religious traditions, we risk unduly constraining our conception of religion if we assume a priori that they are. And third, even when traditions do apparently share a commitment to a compassionate ideal, it remains to be seen whether the ways in which that ideal is expressed and actualized bear any resemblance across and between distinct communities.
It is this latter point to which Ortner's example sharply draws our attention, as Ortner indicates that despite surface appearances, it is possible to understand the Rinpoche's "coldness" as an aspect of a "very specific higher Buddhist form of 'compassion,' which does not sympathize with grief in any ordinary sense, but rather 'shows the way' out of it" (1999, p. 142). This is an important lesson that thick description accompanied by culturally sensitive hermeneutical analysis can teach both the anthropologist and the philosopher of religion: it reminds us to look for the variegated possibilities of sense within a cultural or religious practice rather than assuming that we already know in advance of attention to particular examples what a certain concept, such as the concept of compassion, means-what senses it has-within any given form of life.

Concluding remarks
The examples I have summarized above provide glimpses of the rich potential that ethnographic thick description has to offer for expanding the range and depth of philosophy of religion. The first example, of animal sacrifice at the Kāmākhyā temple in Assam, shows how thick descriptions of religious practices offer opportunities for philosophical reflection on aspects of religion that are otherwise liable to be neglected. Taking my lead from Wittgenstein, and from certain philosophers influenced by him, I have suggested that the aspects reflected upon might include the possibilities of sense that a ritual can be seen to possess when we shift our gaze from an exclusive preoccupation with instrumental explanation and towards the ritual's emotional resonances. But other modes of reflection are no doubt also available. The second example, of the response displayed by a high Buddhist lama confronted by a bereaved Sherpa family, shows how the concept of a virtue such as compassion might be culturally or religiously variegated, thereby demanding caution on the part of philosophers who might be tempted to make broad-brushed claims about the centrality of certain virtues to religions in general. One way of developing the analysis of this example further would be to situate it within a broader comparative framework, drawing comparisons not only with other expressions of (what is commonly termed) compassion in Buddhist contexts but also with ostensible instances of compassion from other religious traditions. Such comparative analysis might reveal further variegation in the concept, or possibly that some instances of what has been called "compassion" are better conceptualized in other terms. My point here is not to prejudge the outcome of any such analysis but merely to indicate directions in which future research might go.
In more general terms, then, what I have argued is that thick descriptions of religious practices such as rituals, or of other interactions between religious practitioners, can usefully contribute towards the undermining of facile assumptions about what being religious "must" consist in and about what certain important ethical or religious concepts "must" mean. If, as Hick among others has remarked, philosophy of religion ought to be more than an inquiry into merely a very select portion of religion (Hick 2010, p. 13), then access is needed to "religions in all their variety and dimensions" (Schilbrack 2014, p. xi). Textual sources deriving from the traditions themselves-including doctrinal, liturgical, mythological and theological texts-afford valuable insights into the conceptual and doxastic dimensions of religions, but religion amounts to far more than what is discernible in religious texts: "'reading' religion must mean more than attending to the written word alone" (Collins and Arweck 2006, p. 4). Without philosophers having to become ethnographers themselves, existing ethnographies-whether these be written accounts or audio and visual material-constitute one among other fertile resources with which philosophers of religion can productively engage.
While attention to what Knepper (2013Knepper ( , 2014, following John Clayton (2006, pp. 5-6), terms "religious reason-giving" is and should be of interest and importance to philosophers of religion, there is no clear rationale for delimiting the use of thick description to the exposition and analysis of this area of religious life. Philosophical engagement with thick ethnographic description can go beyond the immediate and relatively uncommon contexts of explicit reason-giving to examine the broader repertoire of religious activities, many of which have been sorely neglected by philosophers. The engagement need not be unidirectional; that is, it need not consist exclusively in philosophers learning from ethnographic sources about the radical plurality of religious forms of life. It might also take the form of philosophers contesting assumptions that are present in anthropological analyses of the described phenomena. Critical engagement of this kind has a long pedigree, especially within the tradition of philosophy of religion influenced by Wittgenstein, and it is chiefly upon this tradition that I have been drawing in this article to illustrate fruitful points of intersection between philosophy and anthropology. But while Wittgensteinian approaches are eminently well suited for such interdisciplinary interaction, they are far from having a monopoly in this area.
Philosophy of religion is itself methodologically pluralistic, and has the potential to become more so. This strikes me as a strength rather than a weakness. What I have contended in this article is that thickening the descriptive material with which philosophers of religion operate is one important route towards an expanded conception of philosophical inquiry into religion in all its variety. 10