A more complex approach to the frequency hypothesis

An advocate for what would have to be a far more complex approach to the frequency hypothesis might argue that, in fact, a criterion that gen- erates multiple performance frequencies for the same ritual may not be so bad after all — precisely because it generates different values depending upon participants' roles. The amount of sensory pageantry to which par- ticipants are exposed in these rituals often does differ depending on their roles. Some rituals might be said to aim much of their sensory pageantry at specific participants. Special agent rituals unquestionably direct most of their sensory pageantry at their patients. For example, in a Christian wedding, although many people may enjoy the sensory pageantry (the finery, the flowers, the music, etc. ), it seems to be the bride, primarily, and the groom, secondarily, who are the manifest targets of most of the special effects. Similarly, in the initiations of the Baktaman, it is only the initiates, not the cult leaders or their assistants, who undergo the ordeals and days of torture, even if everyone does participate in the feasts with which some of these rituals culminate.

We will postpone for a while the question of whether this revised, “far more complex approach” results in anything legitimately describable as a version of the frequency hypothesis, and focus, instead, on its plausibility. The first thing to note is that it is not anything Whitehouse intends, since neither does he take up these problems when presenting either the frequency hypothesis or his account of the two religious modes nor does he discuss the possibility of assigning multiple frequencies to the same ritual depending upon different individuals' roles. But is this revised approach to the frequency hypothesis plausible, regardless of any specific theorist's intent?

If the point of this revised approach concerns the differences in the frequencies with which individuals serve in different ritual roles, then we need go no further than the Baktaman example to see an empirical prob- lem. The very reason the Baktaman initiations posed no problems for the simple, unqualified participation criterion was that the resulting frequen- cies were low for all of the participants, regardless of their roles. But that is just the reason why these initiations pose a decisive problem for this re- vised proposal. The levels of sensory stimulation among the different roles vary markedly (we suspect that torture probably qualifies as sensory pag- eantry of a fundamentally different order in just about everyone's book),

-130-

yet the frequencies for each of the ritual roles will differ little or not at all. It is the one and only time the initiates will go through this degree of initiation, but it is also the one and only time the assistants will assist, and it will certainly be no more than one of two or three times at most that the cult leader will preside.

It is, however, less plausible that we can make all of the necessary, fine- grained distinctions between levels of sensory pageantry that the evalua- tion of this “more complex approach” to the frequency hypothesis would require. Although neuropsychology does not provide the requisite in- sights about these matters now, it is not out of the question that it might someday. Drawing such distinctions is not currently practicable, but, ultimately, we do not wish for our negative argument to turn on this point. The critical deficiency of this more complex approach to the fre- quency hypothesis lies far deeper — as our discussion in the remainder of this section will demonstrate.

-131-

These initial considerations jointly suggest that, however promising on some counts this revised approach to the frequency hypothesis might initially appear, it is not very practicable. Establishing that it is not em- pirically plausible will take more work.

To repeat, testing the ritual form hypothesis directly does not depend upon obtaining a satisfactory criterion for rituals' performance frequen- cies. However, because testing the ritual frequency hypothesis does, a fair comparison of the two hypotheses' empirical merits will as well. Our aim is to be as fair as possible to the ritual frequency hypothesis. Ultimately, though, it will not matter which criterion of performance frequency its advocates adopt. The ritual form hypothesis makes the better predictions, regardless. That said, it should by now be clear that the requisite criterion must turn on some properly qualified account of participation.

Even-numbered rituals do not pose as many problems for the formula- tion of an adequate criterion for the relevance of performances as special agent rituals do. That is because, their different roles notwithstanding, ritual practitioners' and normal participants' opportunities to participate in special patient and special instrument rituals do not differ substantially. The overwhelming majority of Eucharists a priest performs are Eucharists of which a normal participant can partake, but even if they do not attend every Eucharist one priest performs, they may still attend other priests' performances.

It is special agent rituals that present the problems. Sometimes with these rituals (for example, with bar mitzvahs) ritual practitioners' oppor- tunities for participation may vastly exceed the opportunities available to those serving in other roles — but (for example, with Baktaman initiations) sometimes not. The critical point is that it is the frequencies with which the patients of special agent rituals participate in them that hardly varies. We stated before that in any religious system individuals take the patient's role only once — at least ideally — in these special agent rituals. (This is the pivotal sense in which these rituals are non-repeated. ) Even when the real world fails to conform to the ideal, normal participants' opportunities to participate in any special agent ritual (compared with those of ritual practitioners) are few. Even in Hollywood the number of times individu- als marry one another (or anyone else, for that matter) is small compared with the number of weddings typical members of the clergy perform.

These observations about real world failures to conform to the ideal occasion some reflections on one theoretically interesting complication

-132-

Things sometimes happen that necessitate the undoing of the super- permanent consequences of special agent rituals. Some couples must be divorced; a few participants must be excommunicated; an occasional practitioner must be defrocked; buildings or places must be desacralized. Any rituals by means of which such reversals would be accomplished would themselves have to be special agent rituals, since only CPS-agents have the power to undo what they have done. The reversals concern the first rituals' consequences.

We should highlight a distinction here. Throughout we shall distinguish between a ritual's (causal) effects and its (religious or logical) consequences. The former refers to the ritual's impact on physical, biological, psycholog- ical, and social arrangements. It concerns the ritual's place in the causal order scientists study. On the other hand, a ritual's consequences con- cern its objectives within the framework of the larger religious system and its impact logically on inferences and on the applications of cate- gories from the religious conceptual scheme. So, for example, one of the (psychological) effects of ritual scarification is to motivate participants and one of its (logical) consequences is the participant's eligibility to participate in communion rituals.

Although, in principle, the consequences of all special agent religious rituals are ritually reversible, in fact, there is little evidence that such rituals exist and even less that they are actually performed. Note at once that to say that is not to say that the consequences of these rituals cannot be reversed. They can. It is uncontroversial that juridical means often exist for reversing the consequences of special agent rituals. Sometimes the authority in question is secular, sometimes it is religious, and sometimes it is both.

This complicates the claim that odd-numbered, special agent ritu- als are non-repeated, because after a reversal the original special agent ritual can then be re-performed with the same patient(s) a second time. (Think about marriage and divorce in Hollywood!) Other than the rever- sals of marriages, though, such juridical reversals are comparatively rare. Other than remarrying after divorces, second performances of the orig- inal special agent rituals after a reversal are even rarer still. Yet, very occasionally such things do happen. For example, St. Michael's, which stands not too far from the gates of King's College in Cambridge, was

-133-

Special agent rituals incorporate comparatively high levels of sensory pageantry that they overwhelmingly steer in the direction of their pa- tients. The variation in the levels of sensory pageantry that different par- ticipants may experience, as they carry out their roles in these rituals, is typically a straightforward function of their physical proximity to the rituals' patients. ⁸ It is in this light that we propose what we shall call the “least direct connections” criterion of performances' relevance (LDC criterion, here- after) for the purposes of both making sense of the ritual frequency hy- pothesis and comparing it with the ritual form hypothesis. This criterion for ascertaining relevant performances directs us to look at the number of opportunities for participation in the ritual of those who are eligible to participate who have the least direct ritual connections with CPS-agents. Let us unpack each of the italicized terms in this statement of the LDC criterion.

Note that the ritual frequency hypothesis makes predictions about the levels of sensory pageantry in rituals on the basis of their performance frequencies. When we turn in the subsequent sections of this chapter to the comparison of the two hypotheses' predictions about the sensory pageantry associated with various sorts of rituals, the facts about those rituals' performance frequencies will be crucial. The LDC criterion pro- vides the grounds for deciding what the facts about rituals' performance frequencies are.

At the outset of this section we identified two problems of clarity in our preliminary comments about rituals' performance frequencies. The first concerned formulating an appropriate criterion for deciding the relevance of performances in calculating the rituals' performance frequencies. In re- sponse to that problem, we have argued at length for the appropriateness of the LDC criterion. The second problem we noted in our early com- ments about performance frequencies concerned how they are properly expressed. We turn to that problem here.

There are two obvious ways in which we might express performance frequencies. We might measure them by simply counting the number of times (according the LDC criterion) that some ritual has been per- formed. We shall refer to this approach to performance frequency as a “raw count” of the ritual's performances. However, a much better way

-135-

Sometimes raw counts will suffice for the comparison of two rituals' performance frequencies. But they can often be misleading. It depends upon the differences in the ages of the rituals and the lengths of the relevant time periods. Raw counts will deceive when we compare rituals' frequencies for time periods in which people performed one of the rituals throughout while they performed the other for some small fraction of the time period only, because, for example, it fell into disuse. If we were to express the performance frequencies of the Zulu naming rite and the Zulu circumcision rite for the past three hundred years for some clan in terms of raw counts, we would have a very skewed account of the data, since Shaka eliminated the circumcision rite over a hundred years ago. The same problem could arise in comparisons of long-lived rituals with those that have just emerged.

Expressing rituals' performance frequencies in terms of performance rates will avoid all of these problems. Both naming rituals and circumci- sion rites were performed once per lifetime. The Pomio Kivung perform the Cemetery Temple ritual daily whereas offerings to the ancestors at Bernard's Temple occur twice a week. That is the revealing comparison, regardless of which of the rituals came first.

Sometimes comparisons of rituals' performance rates can be no more helpful than comparisons of raw counts. Both can misdirect us, if the researcher uses too short a time period for the comparison. For example, an annual festival might include the performance of special rituals which during the critical week of the year would have the same raw count and the same performance rate as a ritual performed weekly throughout the entire year. Researchers must exercise some judgment in making such comparisons in order to avoid such unenlightening results.

Recall that our ultimate goal in this section is to make a case for the greater theoretical depth of the ritual form hypothesis. Augmenting the ritual frequency hypothesis with the LDC criterion for determining the rel- evance of ritual performances will not render it a robust version of that

-136-

hypothesis but, instead, a thin version of the ritual form hypothesis. What results is a “thin” version of the ritual form hypothesis for at least two reasons.

Second, the “more complex approach” to the ritual frequency hy- pothesis remains an exceedingly thin approximation of the ritual form hypothesis, because performance frequency remains its unexplained, in- dependent variable. By contrast, the formal machinery that the ritual form hypothesis recruits from the theory of religious ritual competence specifies the single most important factor influencing performance fre- quencies, viz., ritual form. Note this is not merely a point about how we choose to calculate performance frequencies. Rather it concerns the vari- ables that are the causes of rituals' performance frequencies. The ritual form hypothesis construes the space of possible ritual arrangements three dimensionally. (See figure 4.1. )

If, as we have argued, the LDC criterion is the only viable measure of performance frequency that advocates of the ritual frequency hypothesis can adopt, then the central question of that hypothesis, “is this ritual done more often than others?” largely reduces to the question “is this ritual a repeatable or a non-repeated ritual?” which is equivalent to the question “is this ritual a special patient or special instrument ritual or a special agent ritual?” Since, in effect, ritual form is a discrete variable, the ritual form hypothesis bifurcates the three-dimensional space of religious ritual possibilities into two regions. (See figure 4.2. ) Although the repeatability of a religious ritual does not provide the entire explanatory story for that ritual's performance frequency, it is probably the single most important factor to which that story appeals. As we demonstrated in chapter 3, the

-137-

Special agent religious rituals, which are odd-numbered in form, have high levels of sensory pageantry and the least directly connected par- ticipants participate as the patients of these rituals only once. For each patient these rituals need only be done once, because the CPS-agents in these rituals either act directly themselves or certify their intermediaries' actions indirectly. These rituals contain high levels of sensory pageantry and emotional arousal, both because participants should remember these unique rituals and because they should also emerge from them with the conviction not only that something profound has transpired but also that the actions of the gods are at least ultimately, if not proximately, respon- sible for these rituals' consequences.

Aspects of ritual form our theory of ritual competence pinpoints de- termine whether religious rituals are repeatable or not, which is the prin- cipal factor determining ritual frequencies (construed in accord with the LDC criterion). ¹⁰ The ritual form hypothesis possesses greater theoret- ical depth than the ritual frequency hypothesis, then, since it identifies

-138-