Mary Douglas: Purity and Danger (1966) Summary. The aim here is to summarise the work generally, highlighting ideas of particular interest.

Chapter 4: Ritual re-assessed

Hits since 28-01-12:    17568

Links to Other Chapters:
Purity and Dangerch 1ch 2 |  3 |  ch 4ch 5 |  6 |  7 |  8 |  9 |  ch 10

Extended Comments Pages:
Ndembu Doctor (ch 4)Zande Poison Oracle (ch 5)

In her fourth chapter, Douglas challenges earlier anthroplogical views on ritual, views which she sees as having arisen from a fundamental prejudice of our culture to the effect that foreigners know no true spiritual religion [p 59].

Frazer and Malinowski

In Douglas's opinion, it was on the basis of this prejudice that Frazer erected his theory of primitive magic, according to which primitive tribes expected their ritual words and gestures to work automatically, the way magic happens in the Arabian Nights. From this conception of primitive magic came a mistaken distinction between primitive culture and modern culture.

Frazer's theory was accepted uncritically by Malinowski [1] and elaborated by him. Malinowski saw ritualised words and gestures as arising from originally almost spontaneous verbal and physical expressions of intense desire for some particular outcome, such as revenge.


Douglas suggests we might consider ritual in terms of miracle rather than of magic. Thus, in the miracle-believing ages of Christianity [p 60] Christians did not expect Church rituals to produce results magically, i.e. automatically. What they actually believed was that there was always the possibility that a miracle might happen. [1a]

With different nuances, this kind of attitude was general among pre-literate believers: the idea that some kind of miraculous intervention into human affairs is always possible, but that there is no sure way of turning the possibility into reality. Such an idea was involved, for instance, in the Teutonic concept of luck.

The Judeo-Christian tradition

Douglas sees our culture's prejudice against ritual as inherited from the Judeo-Christian tradition. In that tradition, there has always been a conflict between interior religious life and exterior religious observance. The Old Testament prophets and the New Testament apostles, as in the case of St Paul, all emphasised the importance of interiority against legalistic exterior performance. That Biblical conflict has continued on throughout the history of Christianity, as in Protestant anti-ritualism. [2]

Douglas acknowledges that empty ritual has to be guarded against. However, she finds a danger in the tendency to suppose that any ritual must be empty ritual, that exterior observance must necessarily mean an absence of genuine interiority. [3]

This tendency can all too easily lead to an unwarranted misconception with regard to pre-literate religions: the assumption that if they have rituals, then they must lack an interior dimension.

Social rituals

Indeed, for Douglas there can be no religion at all without externals: this is the case for our lives more generally, as she now proceeds to argue at some length. The author points out that for all the value we place on spontaneity in our personal relationships, these relationships could not exist without social rituals to give them reality: we need letter writing, for instance. [4]

For, according to Douglas, we need in our lives what she refers to as framing, what this site would term articulation. We need rituals of one kind and another to give focus to our experience.

What is more, without rituals certain types of experience would not be possible. Douglas points out how the days of the week create an experience of time that would not be possible if we did not have them. Not only do the days have their own particular characteristics for us, but also their meaning for us is partly a matter of their position in the sequence. Thus Sunday is the day of rest for us, but it also has a significance relative to Monday. Douglas sees regular rituals generally as helping shape our experience of life.


Going back to religious rituals specifically, Douglas looks at how thinking about them in the British context was influenced by Radcliffe-Brown. The latter had himself been influenced by Durkheim, who had aimed to show how religious ritual creates society by making individuals aware of themselves as social beings. Consequently, for Radcliffe-Brown, as opposed to Frazer and Malinowski, pre-literate religious ritual was not a matter of mere magic.

However, there is for Douglas a serious problem with Radcliffe-Brown's approach. He disregarded the truly religious dimension in Durkheim's discussion, making ritual a merely social phenomenon. Losing sight of the importance of ritual in relation to issues of ultimate human concern, Radcliffe-Brown reduced it to matter of expressing particular social values.

For Douglas, this left the connection between particular rituals and particular social values quite arbitrary. Thus, in Radcliffe-Brown, particular food taboos among the Andaman Islanders were about inculcating respect for those older than oneself. But there was no explanation as to why some foods rather than others should figure in these taboos. Douglas makes the same sort of point about the Radcliffe-Brown influenced discussion of M N Srinivas on pollution among the Coorgs of South India.


Contrastingly, Douglas apparently approves of Lienhardt's [5] understanding of the religion of the Dinka people of southern Sudan; she now launches into a relatively lengthy account of it. [5a] According to Douglas, Lienhardt's main point is that Dinka religion ritual is about creating and controlling experience.

It is in this light that Lienhardt sees the Dinka ritual in which the Master of the Fishing Spear meets his death by being buried alive. The ceremony is about controlling the way in which the Spear Master comes to die. A Spear Master cannot be seen to die a normal death, because that would contradict his religious role as guardian of the life of the Dinka people. [6]

Our own rituals

Douglas claims that we can apply this sort of argument to the rituals of our own society as long as we allow for one important difference. The rituals of peoples like the Dinka are unified, each one a part of a religion that penetrates almost every area of life. But our rituals are not unified, because our lives are compartmentalised. [7]

Allowing for this difference, Douglas sees a comparison between the pollution concerns of tribal cultures and some of the allegedly hygienic concerns of our own culture. [8] Thus, among the Bushman people of South West Africa there are female and male sides of the campfire, while amongst ourselves the men are supposed to use the downstairs toilet and the women the upstairs. In both cases the effect is to impose pattern on domestic space, turning it into home.


Douglas now turns to the effectiveness of ritual. [9] For her, there is an illuminating comparison to be made between ritual and money. This seems worth detailing:

  • Money is a sign, standing for operations that would otherwise be hard to pin down.
  • Money mediates transactions; ritual mediates experience [p 70].
  • Money gives us a standard by which we are able to measure value; ritual creates standardised experiences, enabling us to assign a value to them.
  • Money links the future to the present; ritual likewise.

Douglas remarks that money is actually a very particular type of ritual. [10]

Pursuing her analogy further, Douglas notes that, as with money, ritual depends for its efficacy on public confidence. She then offers two examples from recent anthroplogical research that illustrate the power of ritual when it is backed by belief. [11]

Turner and Lévi-Strauss

The first example is a case of shamanistic ritual reported by Turner. The shaman is able to cure a subject because the ritual he performs involves what is effectively group therapy directed at bringing out into the open the social problem with the community that lies behind the subject's psycho-somatic symptoms. [Ndembu Doctor]

The second example is drawn from Lévi-Strauss. This shows how a shamanistic ritual is able to help a woman through a difficult childbirth, without the shaman ever touching her. Lévi-Strauss demonstrates how the shaman performs a chant which enables the patient to feel she understands exactly what is happening to her. This relieves her panic and makes her feel that the situation is under control. [12]

Douglas claims to have demonstrated that pre-literate ritual is not about Arabian Nights style magic, but about creating and controlling human experience.


1 British anthropologists

In the course of this chapter, Douglas refers to the majority of important British anthropologists to date. Apart from Frazer, we find:

  • Bronislaw Malinowski [1884 - 1942]: he came to Britain from Poland in 1910 to study anthropology, his interest having been fired by his reading of Frazer's The Golden Bough. He remained here, to become the leading anthropologist of the interwar period.
  • A R Radcliffe-Brown [1881 - 1955]: he succeeded Malinowski as Britain's leading anthropologist.
  • Godrey Lienhardt [1921 - 1993]: Douglas's Obituary of Godfrey Lienhardt appears in her Implicit Meanings [1975/1999] as Chapter 14, pp 188-192].
  • Victor Turner [1920 - 1983]: his important work, The Ritual Process [1969] was obviously still in the future when Douglas published Purity and Danger.

Douglas's own important teacher, Evans-Pritchard, and her leading contemporary, Edmund Leach, are not mentioned in this particular chapter, but we do find the pre-eminent French anthropologist, Claude Lévi-Strauss [1908 - ].     [Back to Article]

1a The Catholic sacraments

October 2006: The main Catholic rituals are those associated with the Seven Sacraments of baptism, penance, eucharist, confirmation, matrimony, holy orders, extreme unction. These are believed to work ex opere operato, that is to say automatically as long as the relevant formal conditions in terms of intention, words, actions, materials and people are met. Thus in the case of the eucharist, bread and wine automatically become the Body and Blood of Christ when the conditions are met. It is difficult to suppose other than that this belief dates from Douglas's miracle-believing ages of Christianity.

Presumably, when Douglas claimed that Christians hoped rituals would produce miraculous results rather than believed they would, she was thinking about special rituals associated with pilgrimages to the shrines of saints, as in the modern example of the pilgrimage to Lourdes. But how she could omit consideration of the sacraments in the context of a discussion of Christian ritual is inexplicable.     [Back to Article]

2 A Catholic perspective

Mary Douglas [1921 - 2007] does not specifically acknowledge any religious allegiance in the book, but there are indications of her Catholic background.     [Back to Article]

3 The real issue

Douglas does not appear to be addressing the real issue here. She apparently fails to appreciate that it is not just empty ritual that Protestantism has a problem with, but any ritual: anything that seems to put a distance between the believer and God.     [Back to Article]

4 Religious rituals, interpersonal mediations

It just will not do for Douglas to argue about religious rituals on the basis of the sort of interpersonal mediations she chooses as examples here: she is not comparing like with like. For instance, if Protestants object to the Catholic Mass as putting a distance between God and the believer, you cannot blithely counter that with an analogy about letter writing: today's equivalent would be using the mobile phone. The difference is not just about quantity, it is also about quality: as in the old saying, two's company, three's a crowd.

For the essential point about a religious ritual like the Mass is surely that it seems to place a third party, in the person of the celebrant priest as mediator, between the believer and God. That third party could seem to undermine the personal relationship between the believer and God, the way a parent-in-law might come between a husband and wife. Surely that is the kind of thing the old anthropologists were referring to when they theorised on what they called primitive ritual: tribal ceremonies involving priests and the like.

In fact, all Douglas has addressed is the question of the believer's relationship with the divine needing personal scale devotional practices, such as kneeling down to pray or using some particular technique to facilitate meditation. These might be seen as the religious equivalents of letter writing and the like. To counter Protestant assumptions, Douglas needed to demonstrate the need for rituals with priests.     [Back to Article]

5 Lienhardt

Douglas does not actually say she approves of Lienhardt's analysis: we have to work that out for ourselves.     [Back to Article]

5a Douglas, Lienhardt and the Dinka

Jan 2007: Issues of detail with Douglas's accounts of Turner and Evans-Pritchard (see the Ndembu and Zande links at the top of this page) have prompted me to check her references to Godfrey Lienhardt: Divinity and Experience [1961], the classic Lienhardt work on Dinka religion. Douglas does not provide page references, but I have so far found these problems in the first two references looked for.

Malaria ceremony

In the first paragraph of chapter 4, Douglas refers to a Dinka annual ceremony aimed at curing malaria. She tells us that this was timed [p 59] to coincide with the time of the year when malaria diminishes.

Douglas then says:

A European observer who witnessed it remarked dryly that the officiant ended by urging everyone to attend the [European] clinic regularly if they hoped to get well. (Lienhardt 1961). [p 59]

I have traced Lienhardt's discussion of a Dinka malaria ceremony to op. cit. pp 279 - 280 and found that there is no basis there for the Douglas sentence just given. What the missionary in question actually says in the quotation from him provided by Lienhardt is:

I may add that people on the way home the same day seemed to feel no inconsistency in coming to us for medicine just as they had done the day before, and I suppose there is no-one in the whole community who turns to us more quickly for help in case of sickness or accident than the old rain-maker himself who acted as high-priest on that day.[op. cit. p 280]

(I must note in passing that a couple of days ago I watched a TV programme about the work of a present-day English neo-shaman. This featured a seriously ill person who was being treated simultaneously by the neo-shaman and by conventional hospital therapy. Clearly, the desire for relief from suffering overrides considerations of logical consistency not only for simple early c20th Dinka but for sophisticated c21st Europeans as well. But to get back to Douglas.)

Elsewhere in his book, Lienhardt describes a Dinka spear master recommending to a patient of his with an ulcer that he should consult European medicine as well. Perhaps Douglas was confusing the two things.

Douglas's error is particularly strange because a few pages later in her work, p 67, she gives a quotation drawn from later in the very same paragraph on p 280 of Lienhardt as the missionary quotation I have given above. So she manages to misrepresent one part of a Lienhardt paragraph and quote another part of it.

Note also that you might suppose from Douglas's reference that the malaria ceremony was important among the Dinka. It turns out not to have been observed by Lienhardt in his fieldwork or by any other major source of information on the Dinka, but apparently merely by the one missionary maybe twenty or thirty years previously in just two villages near his station. It is only the once that Lienhardt refers either to the missionary or to the ceremony.

Grass knotting

Douglas's second reference in chapter 4 to the Dinka is likewise problematic. She says:

The Dinka herdsman hurrying home to supper, knots a bundle of grass at the wayside, a symbol of delay. [p 64]

What Lienhardt actually says in his discussion of [t]he practice called thuic [op. cit. p 282] is:

when Dinka are making a journey they often tie knots in the grass growing beside the path with the intention that the preparation of food at the end of the journey may be delayed until their arrival.[op. cit. p 282]     [Back to Article]

6 Mediation and ritual

Lienhardt's point that ritual creates experience can be extended to mediation generally: we may say that mediation creates and sustains experience. This insight puts us in a position to understand how Douglas goes astray in this chapter. There are two considerations we need to help us.

Interpersonal mediation

Firstly, living relationships between people cannot exist in the absence of mediation. This applies even to the primal relationship between mother and baby. That relationship is created and sustained in a humanly meaningful sense only through mediations: the physical contacts between the two, the exchange of smiles, the mother's baby talk and the infant's gurgles, the play, the feeding process and so on.

This consideration certainly applies to the kind of relationship Douglas referred to, friendship sustained by exchanges of letters. Of course, only in the exceptional case of penpals did written correspondence create friendship in the first place. But any friendship sustained by letter-writing must have been created initially by some other mediation: repeated encounters and conversations, shared experience of work or leisure etc.

Assuming religious belief, it is possible to reflect on how this consideration applies to religion, on whether a meaningful relationship between God and the individual believer can exist in the absence of some form of mediation. The answer is surely not: the relationship is created and sustained by a specifically religious mediation, prayer.

Of course, it could be argued that the kind of exceptional encounter between the exalted individual and God discussed by James and Otto does not involve mediation, that it is a matter of mediationless immediacy. But this sort of encounter may not in itself constitute an actual relationship between the exalted individual and God.

Interpersonal ritual

Our second consideration is that most interpersonal relationships involve rituals of some kind. Thus, in our culture, friends send each other Christmas and birthday cards, meet up regularly for some shared activity, such as participating in some sport or having a drink.

Similarly, we can surely detect a ritual element in most ongoing relationships between individuals and God. For don't most people routinise, indeed ritualise their prayer, the way they do their friendships. Don't they set aside a particular time each day for prayer, don't they adopt a particular posture, such as kneeling, don't they use aids to prayer, such as traditional set prayers.

Again, don't people have in their relationship with God the religious equivalents of the secular calendar rituals of birthdays and Christmases. In Christianity, don't they go to church on Sundays and so on.

Clearly, there are varying degrees of ritualisation in religion. Within Christianity we can see a continuum running between Catholic ritual maximalism and Protestant ritual minimalism, applying at both the social and the personal levels of religion. Catholics have their Mass and their rosary beads, Protestants don't.

In fact, many Protestants would probably reject any idea that their religious behaviour involves ritual at all. For them, the term ritual would probably refer essentially to the stylised ceremonies with priests wearing robes.

But such a narrow understanding is not going to throw any light on how Douglas goes astray in this chapter. For that, we need to understand ritual broadly, as implied above, as a matter of significant patterns of behaviour prescribed by some source of authority: tradition, custom, routine, habit, the Church, the Army and so on. See note 10 below for more on the definition of ritual.

Douglas's error

Douglas's error is that she confuses ritual with mediation. In the light of our two considerations this becomes quite obvious. We have seen that relationships are always a matter of mediation and that relationship mediation may involve ritual. What Douglas does is miss out the may: she writes as if relationships always involve ritual.

Stated another way, this is an example of the elementary all and some error, the second in the catalogue of erroneous arguments in Robert Thouless: Straight and Crooked Thinking and allocated a whole chapter there: Making a statement in which all is implied but some is true [p 193]. Douglas writes as if all mediations are rituals, when only some are.     [Back to Article]

7 Primitive and modern

At the start of the chapter, Douglas referred to a mistaken distinction between primitive culture and modern culture. We now find that she has her own distinction between primitive and modern, expressed in terms of her unified experience ~ compartmentalised experience dichotomy.

This site rejects the Douglas formulation of the distinction along with all the others, however expressed in cultural, intellectual or psychological terms. In particular, there seems to be little to choose between the Douglas dichotomy and Eliade's hokum about religious man and modern man.     [Back to Article]

8 Pollution as ritual

This is not an error of the summary. Douglas actually does write to the effect that the pollution concerns of tribal cultures and our own culture's concerns with hygiene are themselves rituals rather than concerns that may give rise to rituals.     [Back to Article]

9 No signposting

In fact, Douglas says she is now returning (rather than turning) to the issue of the effectiveness of ritual. However, it is not at all clear at what stage previously she had been discussing the matter. This is because she does not signpost an argument in this chapter. She makes various points, but what significance of a given point may have in the course of some overall argument may be far from apparent. Indeed, in so far as this chapter summary seems to have a certain coherence, then to that extent it is not an accurate representation of the original.

Your summariser's best guess is that Douglas is referring back to her point about miracles not magic: that believers hope for results from their rituals, rather than expect them. This does not seem to link up with the immediately upcoming comparison between ritual and money, but it does have something of a connection with the subsequent point about rituals working because people have faith in them.     [Back to Article]

10 Money as ritual

Remarks like this reinforce the notion that Douglas's understanding of ritual in this chapter is unacceptably flexible.

In general usage a ritual is surely a pre-determined sequence of actions that carries a particular significance. We probably apply the term basically to religious ceremonies, in which a prescribed sequence of actions has in context an important symbolic meaning. An obvious example would be a church wedding ceremony.

However, we extend the term to apply to certain sequences of actions observed in everyday life, such as a person's routine for getting undressed and ready for bed. The implication in such instances may be that the person's behaviour is neurotic, not merely determined by habit but under the control of irrational forces within.

Again, we extend the term to animal behaviour, such as the sequences of posturings we observe in the courtship of birds. Here too, there is the idea of pre-determined sequences of behaviour, in these cases programmed in as instinct.

We saw earlier on in this chapter how Douglas assumed that letter writing between friends was a type of ritual. This is already extremely debatable, because there is no prescribed sequence of behaviours involved in merely exchanging letters.

Then we saw how Douglas discussed pollution concerns as if they were rituals. Clearly, rituals can refer to pollution concerns, as in rituals that remove pollution: uncleanness, defilement. But that does not mean that the phenomenon of religious pollution is a ritual, nor concerns caused by belief in pollution.

Now we have Douglas claiming money as a specialised type of ritual. But in spite of the list of features that precedes the claim, she is mistaken. Money may be involved in all sorts of rituals, but it is emphatically not itself a ritual.

This is because money fails the critical test. However we extend the meaning of ritual, we are always talking about a pre-determined sequence of actions or behaviours, prescribed by some controlling source of authority: tradition, the Church, habit, neurotic personality, animal instinct etc.     [Back to Article]

11 The power of ritual

These examples of the power of ritual are interesting and persuasive in themselves, but what they might signify in terms of the chapter as a whole, Douglas does not make clear. She had started the chapter by evoking a Bushman rain ritual made fun of by a European observer who was viewing it from the perspective of ritual as magic. So we might have expected these final examples to point to a preferable understanding of ritual. But merely to suggest that rituals addressed specifically at humans are not to be laughed at gives us a rhetorical finish to the chapter and no more.     [Back to Article]

12 A chronological plan

By the time we reach the end of this chapter, it turns out that Douglas has touched chronologically on the views on ritual of the majority of the main British anthropologists and has contextualised these views in terms of the Judeo-Christian tradition in general and of Protestantism in particular.

Consequently, it would have easily been possible for the author to present her views in terms of an explicitly chronological chapter plan instead of the disjoined meanderings we actually get. One wonders why she didn't do that.     [Back to Article]

(c) John C Durham, 2004-2006

Links to Other Chapters:
Purity and Dangerch 1ch 2 |  3 |  ch 4ch 5 |  6 |  7 |  8 |  9 |  ch 10

Extended Comments Pages:
Ndembu Doctor (ch 4)Zande Poison Oracle (ch 5)