‘Not so much thought out as danced out’: expanding philosophy of religion in the light of Candomblé

Abstract

When the anthropologist R. R. Marett affirmed that certain forms of religion are ‘not so much thought out as danced out’ (1914, xxxi), he was, in effect, anticipating a criticism that has been levelled at philosophy of religion in recent decades – namely, the criticism that this branch of philosophy has frequently underplayed the extent to which religions often prioritize ritual activities (including dance) over intellectual matters. Taking Marett's observation as a point of departure, this article reflects philosophically on the Afro-Brazilian tradition of Candomblé as an exemplary case. Special attention is given to the themes of (a) dance as a mode of ‘gestural language’ (Wafer 1991, 178) and (b) ‘embodied knowledge’ (Daniel 2005). It is argued that these themes supply opportunities to enrich our understanding not only of a significant dimension of religion – that is, the dimension of dance – but also of what communication and knowledge can amount to in both religious and non-religious contexts.

The phrase ‘not so much thought out as danced out’ derives from Robert R. Marett's introduction to his book The Threshold of Religion, which was first published in 1909 and then reissued in a second edition in 1914 (see Marett 1914, xxxi). Although Marett (1866–1943) is generally regarded as an anthropologist or ethnologist, at Oxford he was initially employed to teach philosophy, which he did from the early 1890s until he succeeded Edward Tylor as Reader in Anthropology in 1910. Regardless of whether we call them philosophical or anthropological, Marett's ideas have significant implications for philosophy of religion. Reacting against what he perceived as a prevalent tendency to devise overly ‘intellectualistic’ theories of what religions are – a tendency, that is, to reduce religion to some particular ‘doctrine or system of ideas’ (Marett 1914, xxxi) – Marett accentuated the extent to which each religion is an ‘organic complex’, comprising affective and behavioural dimensions in addition to intellectual ones (1914, x).

An overemphasis on intellectual, doctrinal, or doxastic components of religion at the expense of other aspects has been notable in philosophy of religion. It is in an effort to counteract this overemphasis that some contemporary philosophers working in this area have sought to refocus attention on ‘lived religion’ (Clack 2018; Knepper 2021), ‘living religion’ (Hewitt and Scrutton 2018), or ‘the priority of practice’ in religious traditions (Cottingham 2005, 151; Burley 2018). Unwittingly echoing earlier figures such as Marett, Kevin Schilbrack identifies ‘intellectualism’ as one of three cardinal problems with what he calls ‘traditional philosophy of religion’ (2014, xi–xii), the other two problems being ‘narrowness’ (in the sense of a narrow preoccupation with only a small subset of religions) and ‘insularity’ (in the sense of a lack of engagement with other branches of philosophy and with other disciplines involved in the study of religion). Marett can thus plausibly be viewed as a prescient forerunner. By highlighting the risks of ending up with a distorted conception of religion if intellectual factors are fixated upon, he anticipated the very problem that has been a principal target of many critical voices in philosophy of religion in the twenty-first century.

When Marett describes religion as ‘something not so much thought out as danced out’, he refers to ‘savage religion’ in particular (1914, xxxi). Needless to say, the term ‘savage religion’ would these days be rejected by anthropologists and replaced by terms such as ‘small-scale religions’, ‘indigenous religions’, or, if more precision is called for, ‘indigenous religions of small-scale societies’ (see Winzeler 2012, 17–18). Marett's point, however, is likely to remain widely accepted among anthropologists of religion, provided his use of the term ‘dance’ is not understood too narrowly. As Birgit Meyer, a contemporary scholar of religion in Africa, has remarked, Marett's assertion may be read ‘as a provocative argument that introduces “dance” – in the broad sense of the moving body and, by implication, also the efforts to keep the body still – as a general feature of religion’ (Meyer 2016, 15). Although thinking of ‘dance’ as a synecdoche for (or exemplification of) bodily movement in general is a productive starting point for an expanded conception of religion, Marett's assertion also prompts us to consider the role of dance, in a more literal sense, as a form of religious practice. The foregrounding of dance in an investigation of religion brings to light the pervasiveness of this human activity both throughout history and across multiple religious and cultural milieus. From the Neolithic era onwards, and in virtually every known human society, it appears that dance has been integral to ritual practices, many of which can legitimately be thought of as religious in character (Gaston and Gaston 2014).

Owing to its pervasiveness, dance is a natural topic for cross-cultural comparative inquiry. Such an inquiry would, however, be an ambitious enterprise to even begin in a single journal article. To avoid spreading the discussion too thinly, the present article concentrates on a specific example, albeit one with a complex history and internally variegated structure, namely the Afro-Brazilian religion known as Candomblé. This is an especially pertinent example, not least because the very term candomblé is believed (by some) to derive from kandombele, a Kikongo term associated both with prayer (Henry 2008, 208; Thomas 2011, 224) and with dances or festivals (Johnson 2002, 202; de Castro 1976, 144). Translations of candomblé into English thus include ‘dance in honor of the gods’ (Hollander 2003, 8; Taliaferro and Marty 2018, 51), ‘dance in honor of the deities’ (Lartey 2016, 106), and ‘dance in honor of the spirits’ (Brill 2018, 233). In these phrases, ‘gods’, ‘deities’, and ‘spirits’ are approximations of Brazilian Portuguese terms such as orixás or inquices, these being Portuguese pluralized renderings of, respectively, the Yorùbá òrìṣà and the Bantu nkisi (da Silva and Brumana 2017, 172).

Following this introductory section, I first contextualize the inquiry into Candomblé by examining the relation between religion and dance. I aim to bring out the interplay between these components of human life as well as the extent to which dance can be viewed as a communicative and transformative activity. Attention is then given to Candomblé as a case in point. After introducing some key features of the Candomblé tradition, I explicate, with reference to this tradition, the themes of dance as a mode both of ‘gestural language’ (de Oliveira Cunha 1986, 2022; Wafer 1991, 178) and of ‘embodied knowledge’ (Daniel 2005). These themes, I argue, supply opportunities to expand and diversify our understanding not only of a significant dimension of religion – namely, the dimension of dance – but also of what communication and knowledge can amount to in religious, and indeed non-religious, settings.

Dance as a dimension of religion

The pioneering scholar of religion Ninian Smart provides a fruitful analysis of religion in terms of ‘dimensions of the sacred’ (Smart 1996). By doing so, he offers a conceptual framework with the potential to help philosophers of religion avoid the reductive tendency of those who overemphasize what Smart dubs the ‘doctrinal or philosophical dimension’ (1996, 10). In addition to the latter, Smart identifies as further dimensions the ‘ritual or practical’, ‘mythic or narrative’, ‘experiential or emotional’, ‘ethical or legal’, ‘organizational or social’, and ‘material or artistic’; as the most fundamental of all, Smart picks out the ‘political and economic dimensions of religion’ (1996, 10–11). Although dance does not feature in Smart's framework as a dimension in its own right, it is relevant to several of the dimensions that he specifies. In connection with ritual, Smart mentions in passing the use of dance and music among some groups of Ṣūfīs to engender purified states of consciousness (Smart 1996, 104). Additionally, it could be noted that dance is a common method of acting out myths or sacred narratives (Gianvittorio-Ungar and Schlapbach 2021); it is often associated with the expression or instigation of emotional experiences, both for the dancers themselves and for their audiences (Martin 1933, esp. 18); and dance clearly has an artistic or aesthetic aspect, which is why dance, whether spontaneous and improvised or deliberately choreographed, is generally regarded as one of the performing arts (Urmson 1976).

The phenomenologist of religion Gerardus van der Leeuw discusses dance at some length in a book on the relation between religion and art (van der Leeuw 1963, first published in Dutch in 1932). Following Marett, he defines dance simply as ‘ordered movement’ (van der Leeuw 1963, 73; Marett 1932, 43), though he identifies it with something primordial which brings human beings into reverberant connection with a cosmic pulse. Utilizing poetic vocabulary, van der Leeuw describes the rhythm of a person's dance as akin to ‘the distant sound of breakers which emanates from the beat of waves in the heart of the universe’ (van der Leeuw 1963, 68). What imbues dance with a religious quality, van der Leeuw suggests, is ‘the unification of one's own movements with the movement of the whole’ and the surrendering of one's will ‘to a stronger power’ (van der Leeuw 1963, 68). As other writers have commented, a collective dance amounts to more than the sum of its parts: it ceases to be a matter of individuals performing discrete movements, instead becoming a group phenomenon wherein the dance acquires a life of its own (Langer 1953, 175–176; Beaumont 1934, 167). These themes of surrender, merging with something greater, and transcending one's ordinary sense of individuation constitute a thread that runs through work on dance as a religious practice (e.g. Darling Khan 1976, 10; Watts 2006, 124).

Building upon van der Leeuw's ideas, a further point of reference for an analysis of the role of dance in religion is a fourfold typology formulated by the anthropologist Judith Hanna (1979). The four components of the typology are best viewed as overlapping or intersecting aspects of the ritual possibilities of dance, rather than as separate categories. The first aspect, ‘worship or honor’ (Hanna 1979, 106), consists in the utilization of dance in acts of worshipping, venerating, or otherwise treating with immense respect figures such as ‘deities, ancestors, and other supernatural entities’ (Hanna 1979, 106). The second aspect, ‘conducting supernatural beneficence’ (Hanna 1979, 110), relates to the role that dance can play in conveying or mediating, and concretely representing, divine or supernatural power. To illustrate the point, Hanna refers to a dance performed by the Sandawe of Tanzania. Carried out by moonlight, and featuring erotic movements that signify the fertility with which the moon is associated, the dance invokes that fertile potency, and thus ‘metaphorically conducts supernatural beneficence’ (Hanna 1979, 111; see also Ten Raa 1969).

The third aspect in Hanna's typology, ‘effecting change’ (Hanna 1979, 112), concerns the capacity of dance to engender and give expression to the transitions in status that occur through ‘rites of passage, death ceremonies, [and] curative and preventative rites’ (Hanna 1979, 112). An example, though not one of Hanna's, is the dance performed by Ẹlẹ́ṣin Ọba in Wole Soyinka's play Death and the King's Horseman (1975). As the chief horseman of a recently deceased Yorùbá king, it is Ẹlẹ́ṣin's responsibility to commit ritual suicide, ostensibly to lead the king's favourite dog and horse, which are to be ritually sacrificed, to be united with the king in the world of the ancestors (Gates 1981, 167). The traditional means of enacting this duty is to dance oneself to death. As described in Soyinka's stage directions, Ẹlẹ́ṣin's ‘dance is one of solemn, regal motions, each gesture of the body is made with a solemn finality’ (Soyinka 1975, 41). In the play, Ẹlẹ́ṣin is arrested by the British colonial authorities before his dance can be completed. This is perceived as a calamity by the local Yorùbá community. Ìyálọ́jà, the sagacious leader of the market women, declares that the deceased king will now be doomed ‘to wander in the void of evil beings who are enemies of life’ (Soyinka 1975, 71). I have analysed the significance of this play for philosophy of religion elsewhere (Burley 2020, 82–94); here I am citing it merely to illustrate one way in which dance may be perceived as transformative. Had Ẹlẹ́ṣin discharged his ceremonial duty, he would have been transformed from living horseman to revered ancestor.

The fourth aspect that Hanna identifies is that of ‘embodying the supernatural’ (1979, 114), which, as Hanna acknowledges, comes in several varieties (1979, 106). Central among the forms of embodiment that Hanna has in mind are two, one being what is commonly called possession (or ‘spirit possession’ or ‘divine possession’), the other being masquerade. Borrowing from work by Erika Bourguignon, Hanna defines possession as ‘an inner transformation’, involving a change in the very ‘essence’ of the person who is embodying – or, to use Bourguignon's term, ‘impersonating’ – a spirit or supernatural being; masquerade, by contrast, consists in ‘an external transformation, a change in the appearance of the actor’ (Bourguignon 1968, 13–14; Hanna 1979, 106; see also Bourguignon 1978, 504). By using the term ‘impersonating’ or ‘impersonation’, Bourguignon does not mean to beg the ontological question of whether someone who is described as being in a state of possession really is ‘possessed by a spirit’. In some places, she distinguishes between a ‘conscious imitation or impersonation’ – as carried out ‘in a theatrical performance or a masking ritual [i.e. a masquerade]’ – and a form of impersonation that involves ‘behavior that is culturally defined as being caused by the actual presence of these beings [i.e. ‘spirits’] in the actor’ (Bourguignon 1998, 185). Phrases such as ‘culturally defined’ enable anthropologists and other scholars to acknowledge how the phenomenon in question is described by the religious or cultural community being studied without requiring any ontological commitment on the part of the researcher. The emphasis of the inquiry, both for Bourguignon and for Hanna, is on disclosing the roles and significance of dance, whether in possession rituals or in other religious contexts, rather than on addressing metaphysical questions about the reality and agency of spirits or purportedly divine beings.

My emphasis in this article, too, will be on understanding the significance of dance in the lives of those who deploy it in religiously meaningful settings. The project is hermeneutically oriented rather than being an attempt to determine the soundness of the ontological or cosmological scheme that is intimately connected with the practices under investigation. But these purposes are not entirely separate, for coming to see the meaningfulness of a form of practice, which is internally related to a cosmological framework, is one way of coming to see the plausibility of the cosmological framework itself (although it does not follow that seeing its plausibility amounts to an argument for accepting the cosmological framework). The purpose of this section has been to introduce some key themes and concepts relating to dance as a dimension of religion. In the remainder of the article, we turn to a more focused examination of how that dimension is exemplified in the practices of Candomblé.

Candomblé and the dancing of the divinities

Attempting to define Candomblé has been described as ‘a dangerous undertaking’ on account of the heterogeneous nature of this and other Afro-Brazilian traditions (Capone 2010, ix, 8). There are different branches of Candomblé known as ‘nations’ (nações), some of the major ones being Nagô or Queto (Ketu), Angola or Congo-Angola, Jêje, and Caboclo (see e.g. Omari-Tunkara 2005, 13–17; Murrell 2010, 168–170). Limits of space prevent my expounding in detail the historical origins of these or other branches. It is worth noting, however, that certain terminological variations, such as the alternative terms for the ‘divinities’, are rooted in the different cultural and linguistic inheritances of the respective sub-traditions. For example, it is among practitioners of Candomblé Nagô that the term orixás is most prominent, whereas practitioners of Candomblé Angola are more likely to speak of inquices. For the purposes of this article, I shall stick with orixás, which is the more common of these terms.

Candomblé emerged from a blending of African conceptions of the spiritual beings or orixás, on the one hand, and ‘Iberian Catholic imagery’, on the other (Cole 1994, ix; see also Johnson and Palmié 2018, 447). In its modern form, Candomblé has been said to also incorporate practices and ideas from South American Indigenous culture and European Spiritualism (Goldman 2007, 103), although attempts are being made by some practitioners to re-emphasize the African elements of Candomblé and downplay the non-African ones (Selka 2007, esp. 73–96; Capone 2017). At the heart of Candomblé cosmology is a pervasive power or force known as axé. This force is held to underlie all existence; individual entities, including humans, other living beings, and orixás, are deemed to be modifications or ‘concretizations’ of it (Goldman 2007, 110). The practices of Candomblé centre upon ‘ritual dance, spiritual healing, divinatory science, spirit possession, sacrificial offerings, spiritual powers, and the celebration of living religious memories in Afro-Brazilian communities’ (Murrell 2010, 159). Music and dance are considered to be so essential to Candomblé that there could be no ritual without them (Taliaferro and Marty 2018, 51). Candomblé is thus regarded, along with certain other African-derived traditions such as Vodou of Haiti and Santería or Lucumí of Cuba, as one of the ‘dancing religions’ (Daniel 2005, 48).

An account of the role of dance in Candomblé ritual practice can be elaborated by reference to two salient themes, one of these being the notion of dance as a ‘gestural language’, the other being that of dance as a means of both acquiring and expressing ‘embodied knowledge’.

Gestural language

Part of my reason for invoking the phrase ‘not so much thought out as danced out’ is to draw attention to how the variegated phenomenon of religion can be distorted when philosophers concentrate exclusively on intellectual or cognitive aspects. There is, however, a risk of going too far in the opposite direction by implying that thought and dance have little or nothing to do with each other. This overcompensating tendency can be countered by considering the possibility that dance may in certain circumstances act as a vehicle for thought – or, at any rate, for the conveying of information. One way of reflecting upon this possibility has been to raise the question of whether dance is, or can be, a kind of language (Preston-Dunlop 1987; Bannerman 2014). Before applying it to Candomblé, let us dwell for a moment on this notion more generally.

A well-known proponent of the view that dance is not only a language but ‘the mother of all languages’ is the philosopher and historian R. G. Collingwood (1938, 244). What Collingwood means by the latter contention is that all forms of language are ultimately derived from ‘an original language of total bodily gesture’ and, of all human activities, dance comes closest to actualizing such an original language (1938, 246). One does not need to endorse Collingwood's theory in its entirety to appreciate the reminder that verbal language – what Collingwood calls ‘vocal language’ (1938, 242) – is not the only form that language can take, and that bodily movement and gesture can also play important roles in human systems of expression and communication.

The work of Judith Hanna is again instructive in this regard. She makes a credible case for dance being a form of sign language that utilizes the whole body rather than merely the hands or arms (Hanna 2001). As Hanna points out, there is a sense in which dance combines a vocabulary with grammar to convey meaning; the vocabulary comprises ‘steps and gestures’, and the grammar is made up of ‘rules for justifying how one dance movement can follow another’ (2001, 41). The meaning can take many forms, including the representation of someone or something through symbolic patterns of movement; these patterns might themselves utilize metaphor, such as when movements associated with a given animal are enacted to convey a person's mood, personality, or behavioural propensities (Hanna 2001, 43). Hanna also reminds us that just as verbal languages come in many varieties, so dance, rather than constituting one universal language, has its own kinetic languages and dialects (2001, 41).

The term ‘gestural language’ has been applied to Candomblé by the Australian anthropologist Jim Wafer, who borrows the term from work in Portuguese by Marlene de Oliveira Cunha (1986, subsequently published as a monograph in 2022). Wafer cites Cunha's analysis of how, in Candomblé ritual bodily movements, the earth is a prevalent reference point. The ground is pounded by the feet of the dancers, who frequently bow their heads and bend their backs ‘to keep the earth in view’; so too do they touch ‘the earth, then the forehead and the back of the neck, as a gesture of greeting to the ritual space itself and to individual spirits’ (Wafer 1991, 178, citing de Oliveira Cunha 1986, 144). Referring to the Concise Oxford Dictionary, the scholar of dance Valerie Preston-Dunlop, though not discussing Candomblé specifically, notes that ‘vocabulary’ can denote not just linguistic elements but also ‘a set of artistic or stylistic forms or techniques’, such as ‘a range of set movements in ballet’ (Preston-Dunlop 1995, 224). This sense of vocabulary, and of the importance of the order in which the components of a dance sequence are placed, is illustrated by Henrietta Bannerman, who observes that a sequence comprising ‘a run, jump and fall’ may convey elation or excitement followed by disappointment; if the sequence is reversed, however, what is conveyed may be unhappiness or despair followed by optimism (Bannerman 2014, 70). None of this shows that dance is a language in exactly the way that verbal language is, but it does contribute to an understanding of dance as incorporating methods of articulation and communication of meaning.

Some interpreters of Candomblé ritual dance construe the notion of communication broadly enough to encompass the idea of communicating across different ‘realms of existence’. Candomblé cosmology distinguishes between three main realms: the earthly realm, consisting of humans, animals, and plants; the realm of ancestors; and the realm of orixás or divinities, which Yvonne Daniel calls the ‘cosmic realm’ (Daniel 2005, 54). According to Daniel, the dances are performed both for the human community and for the divinities, with the aspiration that certain divinities should be encouraged to participate in the dance themselves. This participation occurs through the phenomenon commonly referred to in the anthropological literature as ‘possession’, in which a spiritual being – or, as Daniel puts it, ‘spiritual energy’ – ‘“mounts,” “manifests,” or “incorporates” within the dancing human body’ (2005, 85). The communication, or indeed ‘communion’ (Motta 2005, 295–297), thus comprises stylized movements on the part of the human dancers, both of the hands and of the body as a whole; these movements represent and serve to invoke specific orixás, who are held to respond in turn. Their response may consist in blessings granted to the practitioners or, in instances of possession, the temporary inhabitation of the bodies of particular dancers. Several authors have referred to the exchange between humans and orixás as involving an ‘embodied knowledge’ that is received by the ritual participants. It is thus worth examining this notion of embodied knowledge more thoroughly.

Embodied knowledge

A Dictionary of Human Geography defines embodied knowledge as ‘knowledge that is partial, situated, and developed through experience, contextualized with respect to the body, circumstances, life history, and locational context’ (Castree, Kitchin, and Rogers 2013, 125). With reference to Candomblé, Rachel Elizabeth Harding has used the term ‘embodied knowledge’ to denote ‘A way of experiencing a relationship to history, to divinity, to ancestry from within the movements of one's own body’ (2006, 17). Yvonne Daniel contrasts embodied with disembodied knowledge: the latter is ‘Intellectual knowledge without concomitant integration of somatic, intuitive understanding and the spiritual wisdom their combination yields’ (2005, 57). Embodied knowledge, for these authors, is thus more than what Gilbert Ryle famously identified as ‘knowledge-how’ (Ryle 1945–1946; 1949, ch. 2). For Ryle, to have knowledge-how (or know-how) is to know how to do something; having knowledge-how to play a musical instrument, for example, is simply being able to play that instrument (Ryle 1949, 28). This is to be contrasted with knowledge-that, which consists in ‘knowing that something is the case’ (Ryle 1945–1946, 4; 1949, 28). In the discussions of Candomblé to which I have referred, embodied knowledge is described as having a more prominent experiential quality than mere knowing-how.

Embodied knowledge is, in effect, a type of ‘knowledge by acquaintance’, a term coined by Bertrand Russell to denote the knowledge that is gained when one has ‘a direct cognitive relation’ or is ‘directly aware’ of some object (Russell 1910–1911, 108). But in the case of the dancing typical of Candomblé ritual practice, the acquaintance that is gained is deemed by the practitioners to be so intimate as to provoke changes in the knower's sensations and bodily comportment. As Yvonne Daniel delineates the experience, ‘the body is filled with praiseful movement’, the result being not merely a cognitive shift but a feeling of ‘spiritual awe or divine inspiration’ (2005, 161).

An illustrative example recounted by Daniel is her participation in a water ceremony dedicated to the orixá known as Oxalá at a temple in Casa Branca in the state of São Paulo. Oxalá is a contraction of Orixanlá, which is itself the Portuguese rendering of the Yorùbá Òrìṣàńlá (‘Great òrìṣà’). This is one of the principal names of the òrìṣà who is held to be responsible for the moulding of human beings out of clay (Olúpò̩nà 2011, 147–148; Prandi 2001, 23–24). Also known as Obatalá (=Ọbàtálá in Yorùbá, ‘Lord of the white cloth’), he is associated with the qualities of coolness, unhurriedness, and wetness, his sacred animal being the snail, which itself exhibits these qualities (Àjàyí 1998, 70; Burley 2022, 6, 10). In the Brazilian context, a distinction is made between the ‘elder’ Oxalá (known as Oxalufã) and his ‘younger’ and more energetic form (Oxaguiã) (da Silva and Brumana 2017, 180). It is specifically with the manifestation of the ‘elder’ form that the ceremony reported by Daniel is concerned. The ceremony begins with a collective exercise of carrying jugs of water up a flight of stairs in order to fill a large container. The experience of walking repeatedly up and down the stairs, carefully balancing a water jug on her head alongside dozens of fellow ritual participants, is described by Daniel as ‘mesmerizing, calming, taking me to a heightened sensitivity’ (2005, 200). The repetition and mental concentration induce a solemn atmosphere combined with a strong sense of connection between those engaged in or witnessing the activity. Already, then, we can see how the ceremonial performance is capable of facilitating a kind of embodied knowledge: not the acquisition of new propositional information, but the experiential or phenomenological knowledge of felt qualities, notably the placidity and serenity with which Oxalá is associated.

Daniel proceeds to describe the dance that is executed in honour of Oxalá, again from the perspective of an active participant. Both graceful and physically demanding, the dance moves resemble those that, with reference to work by Jim Wafer and Marlene Cunha, I mentioned briefly above – namely, moves that involve stooping towards the earth and taking steady purposive steps. These patterns of motion are, among other things, redolent of the elder Oxalá's characteristics of modesty, equanimity, resolve, and sagacity. Despite an aching back, Daniel perseveres. She imitates two elderly female dancers who display an impressive degree of endurance and fluidity in their movements. It turns out that one of these elderly women is ‘Oxalá himself’, in the sense that she, through her dance, is not merely mimicking the movements of Oxalá but embodying or manifesting him. To apply a distinction that we considered earlier, the woman is ‘possessed’ by Oxalá rather than only putting on an appearance, as in a masquerade. Attributing agency to Oxalá, Daniel describes him as having maintained the woman's ‘body in a very lowered position and at the same pace for hours’; playfully thinking of the dance as though it were a contest, Daniel adds that, faced with such a competitor, there was no chance that she herself could win in this ritual performance (2005, 203).

Subsequent to her participation in the ritual, Daniel experienced what she describes as a sense of ‘euphoria of both body and mind’, an increase in axé – the pervasive power that is central to Candomblé cosmology – and ‘a surging growth or development, a keen awareness of inner fortitude’ (Daniel 2005, 203). She perceives these benefits as blessings from Oxalá, ‘a wonderful prize for staying with the demanding dance for as long as I did’ (Daniel 2005, 203).

Lessons for philosophy of religion

It may not be immediately obvious what lessons can be drawn for the philosophy of religion from the above descriptions of gestural language and embodied knowledge in the context of Candomblé ritual activity. One notable lesson, however, comes in the form of the emphatic reminder that bodily movement – and dance in particular – can be an integral component of the kinds of practices that facilitate transformative religious experiences and the enhancement of religious conviction. Such a reminder may not necessitate an overhaul of the conception of religion with which philosophers of religion habitually operate; it does, though, provide a counterweight to the tendency among philosophers to overplay the intellectual and theoretical dimensions of religion at the expense of experiential, emotional, and imaginative ones.

As John Cottingham has insisted, far from being ‘simply a matter of intellectual assent’ or ‘even of moral orientation’, ‘the religious outlook . . . is something that is to be enacted and embodied in an intimate intermingling of belief and praxis that is in principle capable of infusing the entire tenor of a lived human life’ (Cottingham 2015, 66). Whether we should speak of ‘the religious outlook’ (my emphasis), as though there were some universally accepted way of being religious or of viewing the world from a religious vantage point, is doubtful. Nonetheless, Cottingham's spotlighting of the ‘enacted and embodied’ character of religious participation is a welcome corrective to the over-intellectualizing proclivity of much theorizing about religion, a proclivity to which Marett raised objections similar to those of Cottingham over a century earlier.

Even when discussing religious experience – the phenomenon that William Alston (1991) called ‘perceiving God’, for example – philosophers of religion have often described this kind of experience in ways that give little or no attention to the bodily comportment or movements of those who undergo it. As Grace Jantzen has observed, when arguing over issues such as the evidential value of religious or mystical experiences, philosophers have typically assumed something comparable to William James's well-known definition of mystical experience in terms of ineffability, noetic (i.e. knowledge-inducing) quality, transiency, and passivity (James 1902, 380–382; Jantzen 1994, 194). Although James himself recognized that the activities which are characteristically deployed to precipitate mystical states include ‘bodily performances’ (1902, 381), it is rarely these performances in which philosophers of religion have been interested, still less the often bodily nature of the experiences themselves. From feminist perspectives, Jantzen and others, such as Melissa Raphael, have argued that this lack of attentiveness to the body militates against the full acknowledgement of the experiences of women, whose ‘spiritual knowledge’ is commonly ‘body-mediated’, such as when it is arrived at through the experiences of pregnancy and childbirth (Raphael 1994, 521; see also Miller-McLemore 1992, 231). When feminists make such claims, they are generally careful to distance themselves from the dubious supposition that there is a blanket dichotomy between women's experiences, on the one hand, and men's experiences, on the other. Even so, there are good reasons to be cautious about accepting the insinuation that women's experiences are, or are likely to be, more ‘body-mediated’ than those of men. What can readily be acknowledged, however, is that there are indeed many forms of experience, undergone by both women and men, that are deeply rooted in the body, and the experience of dance is prominent among these. When religious or mystical experiences are assumed to involve an attenuation of somatic factors, scholarly attention is liable to be diverted away from thoroughly corporeal experiences that, in the absence of such an assumption, would be recognized as religious or mystical in character. This is something that feminist contributions to philosophy of religion have made more salient. As we have seen in this article, exploring the role of dance in ritual practices is one means of bringing to the fore the embodied experiential dimensions of religion, making more conspicuous the relevant experiences of both female and male practitioners.

Examining the practices of a specific religion such as Candomblé enables a more fine-grained analysis of the forms taken by (to borrow Cottingham's phrase again) the ‘enacted and embodied’ elements of religion. At the level of metaphysics, we might question what sense can be made of the contention that the bodies of ritual participants, such as the elderly woman described by Yvonne Daniel in her account of the Oxalá ritual, are temporarily appropriated by the divine beings whose characteristics their dancing figuratively emulates. At the level of phenomenological inquiry, however, it can be appreciated that these participants become so absorbed in the ritual proceedings – in the ecstatic momentum of the dance – that talk of ‘embodying the supernatural’ (Hanna 1979, 114) or of ‘manifesting’ some spiritual being or ‘energy’ (Daniel 2005, 85) may not seem out of place. Indeed, for those who are immersed in the imagery, mythology, and devotional practices surrounding a divine figure such as Oxalá, identifying oneself or another as having incorporated the spiritual energy of that figure is apt to come naturally. Attempting to articulate the experience in other terms may risk doing an injustice to its potency, failing to capture the extent to which the practitioner feels what we earlier saw Gerardus van der Leeuw describe as the ‘surrender of oneself to a stronger power’ (van der Leeuw 1963, 68).

When Daniel recounts her post-ritual experience of feeling elated and revitalized as a consequence of Oxalá's blessing, we need not read this as a mere imposition of religious vocabulary onto an essentially physiological or emotional experience. Again, describing the experience as one of having been blessed by Oxalá may be the most apposite form of words available. As Wittgenstein is said to have remarked when discussing another utterance of a religious nature, ‘It says what it says. Why should you be able to substitute anything else?’ (Wittgenstein 1966, 71). Wittgenstein's point need not be construed to be that religious discourse can never be paraphrased or explained in terms other than those that were originally used, but he is urging us not to be too hasty in assuming that, in every case, there must be a paraphrase or explanation capable of saying, at least as clearly and perhaps more clearly, what the original utterance was merely gesturing towards. In many instances, it may turn out that no other words can adequately replace what was originally said.

What Daniel's extensive description of her participation in the Oxalá ritual and of her invigorated state afterwards offers us is a thickly contextualized account of how things can appear a certain way – of how, given a particular ‘intimate intermingling of belief and praxis’ (to quote Cottingham again), it can make sense to understand one's experience in terms of receiving Oxalá's blessing. It is, of course, not only elaborate descriptions of rituals involving dance that can serve this purpose: descriptions of various religious activities and forms of life can be illuminating for philosophical inquiry into religion. But dance remains an important case in point. The interrelated concepts of gestural language and embodied knowledge are among those that open up the possibility of calling into question any supposedly sharp demarcation between bodily activity and cognitive engagement in the context of religion – and no doubt in other contexts as well. Ritual performances communicate as much through gestures, bodily attitude, facial expression, and the non-verbal interactions between the participants as they do through any overt verbalization. The rituals of Candomblé, in which distinct patterns of movement are associated with specific orixás, illustrate the point well. Through movement a certain atmosphere or mood is established, inaugurating changes in the state of mind of the practitioners. Gesture and demeanour embody and convey spiritual qualities, inviting the orixás themselves to join in the dance.

The type of knowledge that such ritual activity facilitates has been described as embodied because it is both acquired and expressed through bodily movement. Regardless of whether we think the woman described by Daniel ‘really was’ possessed by Oxalá, she was undoubtedly imbued with the characteristics resonant of that orixá. There is a sense in which not only she but also the other dancers become intimately acquainted with those characteristics through the motions both of their own bodies and of the bodies of others.

Concluding remarks

In May 1983 the philosopher of religion D. Z. Phillips delivered the prestigious Marett Memorial Lecture at Exeter College, Oxford. In that lecture, which was subsequently published in Religious Studies (Phillips 1986), Phillips draws a connection between Marett's prioritization of dance in the development of religion and the Wittgenstein-inspired idea that so-called ‘primitive reactions’ play a crucial role in the formation of concepts. Summarizing his contention in a later essay, Phillips writes that Marett ‘did not want to suggest that “the dance”, our primitive reactions, were logically or temporally prior to thought’; rather, ‘on the most charitable reading, Marett can be seen as saying that the dance is itself a condition of thought – that our primitive reactions and concept-formation go hand in hand’ (Phillips 2003, 189).

The present article, too, has adopted Marett's foregrounding of dance not as a means of driving a wedge between the ‘primitive’ or bodily dimensions of religion and the conceptually articulated cognitive dimensions, but as a starting point for reflecting upon ways in which thought and dance may indeed be intertwined. Unlike Phillips – and also unlike Birgit Meyer, whom I cited earlier – I have not immediately sought to enlarge the category of dance to encompass ‘the broad sense of the moving body and, by implication, also the efforts to keep the body still’ (Meyer 2016, 15). While recognizing that, in ritual contexts, dance is routinely accompanied by other forms of bodily movement, and hence that a call for philosophers of religion to be mindful of dance is at the same time a call for greater awareness of bodily factors more generally, my aim has been to keep dance in its primary sense firmly in view.

My objective has not been to pin down a definition of what dance is. But if we were to seek at least a working definition, there are various evocative accounts with which we might begin. The classicist George Thomson, for example, described dance as ‘gesture raised to a magical level of intensity’ (1941, 64) – a description that Wole Soyinka, among others, has found congenial (1976, 33). Comparably, the musicologist Curt Sachs remarked that ‘dance in its essence is simply life on a higher level’ (1937, 5). And the dancer and philosopher Kimerer LaMothe has affirmed that ‘to dance is to create and become patterns of sensation and response’ (2015, 4). Despite leaving unsaid much about dance, each of these three broad characterizations brings out an aspect of what dance can be while also hinting at why dance has featured significantly in many religious settings.

The Afro-Brazilian religion known as Candomblé exemplifies the intricate interweaving of dance and religiosity, the blending of rhythmic bodily movement with emotional and cognitive engagement. The focus on Candomblé has thus provided a case study not of how religion can be conceptualized as being danced out as opposed to thought out, but of how examining religion through the lens of dance discloses aspects that might otherwise be overlooked or underappreciated. Furthermore, by registering how dance can constitute a mode of language – ‘gestural language’ – our comprehension of the very concept of language, and of what communication can consist in, may be enlarged. So, too, by viewing dance as facilitating forms of embodied knowledge, we may come to develop an augmented epistemology, an appreciation of how religious understanding, and indeed non-religious varieties of understanding, may be cultivated through bodily experience. Given that traditions such as Candomblé have received relatively little interest from philosophers of religion hitherto, this focus also instantiates a means of expanding philosophy of religion beyond the confines of its standard preoccupations.