05-022 Ch 07.qxd 2/10/05 4:55 PM Page 127 7 Memory and Analogical Thinking in High-Arousal Rituals Rebekah A. Richert, Harvey Whitehouse, and Emma Stewart Imagine that you are a teenage boy calmly sleeping one night. You are sud- denly awakened and dragged from your bed to a dark location on the pe- riphery of the settlement, along with other terrified boys. Your head is cov- ered with a hood and you are brutally attacked by the senior males of the village and herded onto a ceremonial platform. You are whipped with nettles and forced over a fire so that your skin is licked by the dancing flames before strange concoctions of disgusting substances are forced into your mouth and daubed over your skin. You are given no indication of the purpose of these experiences; you are simply required to endure whatever comes next. Such behaviors are quite typical of male initiation rituals in Papua, New Guinea (see Whitehouse 2000). Moreover, rituals like these are found all around the world. This chapter sets out to explore how participants in highly emotionally arousing rituals make sense of their unique experiences. In addressing this issue, we need to appreciate at the outset that ritualac- tions differ from nonritual actions by virtue of being irreducible to means- end technical motivations or to the intentional states of the actor. Reaching out to catch a ball is a qualitatively different kind of action from making the sign of the cross. It is different in part because there is a direct means-end connection between stretching out one’s hand and the stopping and grasp- ing of a ball, whereas self-crossing could be a means to almost any end, but never in a way that conforms to intuitive ideas about mechanical causation. People might cross themselves as an expression of relief or fear, or to show respect or piety. Sometimes they might have no idea why they do it. Of course, they might belong to a religious tradition that specifies exactly what this action means as well as when and why it should be carried out. But whatever ritual actors tell you about the meanings of their behavior, it will 127 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 128 128 Richert, Whitehouse, and Stewart not be based on the same kinds of cognitive operations that lead us to ex- pect a ball to stop when a hand gets in the way. Further, rituals involve alterations to our normal intuitive inferences about the intentions that lie behind actions. When somebody catches a ball, we naturally assume that the behavior is motivated by intentions located inside the catcher. By contrast, ritual actors are not the authors of their actions in any usual sense (Humphrey and Laidlaw 1993; Bloch 2004). They are not the ones who decided that the procedures should take this particular form rather than that, or what clothes should be worn, or any of the myriad other fea- tures of the ritual that are prescribed (as well as the things that are forbid- den). Rituals are constructed out of clusters of actions that are stipulated in advance, rather than expressing the intentionalities of those who carry them out. It is not intuitively obvious why the particular stipulations take the form that they do, rather than some other form. Who was it who first came up with these rules for performing the ritual, and (even more puzzlingly) why? The question of who made up the rules for a given ritual practice might elicit standard answers: it was God, or a messianic leader, or the ancestors, or “the tradition” (conveniently construed as an agent, for this purpose), or some such entity. But the why-type question can present more of a challenge to ritual participants. Whitehouse (2004) argues that the challenge is only taken on in rather special circumstances. If you perform the same ritual many times over to the extent that you can carry it out competently without having to reflect on how it is done, you are less likely to worry about why it has to be done in a certain stipulated way rather than in some other fashion. If, on the other hand, the ritual is performed in such a way that causes you to think a great deal at an explicit level about the nature of the procedures involved, then you are likely to be more exercised by the problem of what it means. Both hypotheses are potentially testable. The second, however, is the main focus of this chapter. Whitehouse (e.g., 1992, 1995, 2000, 2004) has argued that the lower the frequency of a ritual, the more arousing it is likely to be, because if it were not arousing it would fail to elicit the right kinds of memories and motiva- tions and would either die out or come to be more frequently performed. In response to this argument, much debate has focused on the question of what constitutes the “right kinds” of memories and motivations. Most researchers in this area agree that low-frequency, high-arousal rituals give rise to endur- ing episodic memories. In some of his early work, Whitehouse argued that the survival of rarely performed rituals depended on episodic memory for the procedures involved. More recently, McCauley and Lawson (2002) have developed a highly instructive theory of religious ritual transmission that is based partly on the same premise. Others, however, have criticized this line of argument on the grounds that low-frequency rituals are seldom recon- structed wholly (or even largely) with reference to episodic memory, and 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 129 Memory and Analogical Thinking 129 more commonly depend on the guidance of elders and experienced experts who have developed general scripts for the rituals through repeated in- volvement (see, for instance, Barth 2002 and Houseman 2002). In light of these arguments, Whitehouse has started to focus his attention instead on the consequences of episodic recall for the way people think about the mean- ingsof high-arousal rituals (e.g., Whitehouse 2002a, 2002b, 2002c). According to Whitehouse, a major effect of vivid, enduring episodic mem- ory for rituals is that it encourages long-term rumination on what it all means. We may assume that there is intelligent agency behind it all, but that is not what really commands our attention. What preoccupies participants in high- arousal rituals is the question of why the originator of the rituals insisted on certain prescriptions and proscriptions with regard to conduct, demeanor, dress, speech, and so on, as opposed to all the other possible ways of doing the ritual that could easily be imagined. Answers to this question constitute what anthropologists call “ritual exegesis.” Exegesis presents a considerable challenge for human memory. Whitehouse argues that it may be quite easy to remember ritual procedures (what you must and mustn’t do), but people have great difficulty remembering the official meanings of the acts unless these meanings are frequently repeated. In the case of high-frequency rituals, there are many opportunities to rehearse official exegesis, and this fact (com- bined with low rates of spontaneous reflection on the matter) explains why standardized authoritative exegesis is mainly found in routinized religious tra- ditions. It simply could not survive in conditions of very low-frequency trans- mission. In the case of rarely performed rituals, exegesis is based not on ver- bal testimony but on independent reflection. What typically drives this reflection is episodic memory for intrinsically puzzling actions, that is to say, for ritual actions. According to Whitehouse, elaborate ritual exegesis gener- ated in this way takes on the character of a body of revelations—that is, of ini- tially surprising insights into esoteric mysteries that, over time, develop great motivational force. What drives the elders and ritual experts to reproduce the tradition is their deeply held conviction that the knowledge they have gained is valuable and necessary to preserve. In his contribution to this book, Pascal Boyer expresses doubts about the capacity of low-frequency, high-arousal rituals to stimulate elaborate ex- egetical thinking and to motivate subsequent transmission. He argues that ritual ordeals may well “focus the mind,” in the sense of causing people to pay attention to what is happening during the rituals. They may also give rise to vivid and haunting episodic memories. But what evidence is there, he asks, that these experiences give rise to revelatory religious knowledge based on personal rumination about the meanings of the rituals? Whitehouse has based his case mainly on ethnographic evidence. Many social and cultural anthropologists have tried to show that participants in rare, traumatic rituals (such as violent initiation rites) gradually develop elaborate 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 130 130 Richert, Whitehouse, and Stewart esoteric insights into their meaning, culminating in cosmological expertise on the part of cult leaders (see Whitehouse and Laidlaw 2004). This view has more recently been reinforced to some extent by reports from historians, ar- chaeologists, and classicists (see Whitehouse and Martin 2004). Still, evidence of this kind is not as precise and comprehensive as any of us would like. Rit- ual participants may be very reluctant to talk or write about their private ru- minations on matters of ritual exegesis, particularly in traditions that regard rituals and their meanings as secret. If mystery cults are founded on personal revelations rather than publicly transmitted teachings, as Whitehouse main- tains, then how are we to access those “private moments” of reflection? Ethnographic and historiographical evidence does this only indirectly, by pointing to cryptic hints from those involved, their insistence that they are driven by “deeper” forms of knowledge and insight, the often astonishing profusion and systematicity of cultic iconography, and patterns of behavior that seem to imply “hidden” generative schemes shared only by experts but never explicitly divulged. But all of this remains less than convincing if we cannot demonstrate the exact psychological mechanisms involved in the pro- duction of this kind of exegesis and the way in which it unfolds over time. Whitehouse has argued that one of the most prominent psychological mechanisms involved in the production of exegesis is analogical thinking (see Whitehouse 2002a, 2002b, 2004). Psychological research on analogical thinking traditionally draws a distinction between source and target analogs. Analogies postulate links between a particular object, event, or state of af- fairs (the “target” analog) and some previously independent object, event, or state of affairs (the “source” analog). Analogical reasoning has been defined as “the process of understanding a novel situation in terms of one that is al- ready familiar” (Gentner and Holyoak 1997, 32), suggesting that source analogs consist of familiar situations and target analogs consist of novel ones. This definition has serious drawbacks, however. Some instances of analogi- cal thinking are based on the recognition of connections between two situa- tions that are both novel or that are both familiar. We maintain that what is characteristic of spontaneous analogical thinking is that the connections it proposes are novel, insofar as the person is drawing parallels that he or she has not recognized before (even if others have already done so at least partly by independent means). As is outlined below, research suggests that the production of analogies is a natural cognitive process and seems to be used often to impart information that is less easily conveyed in other ways. Although until very recently no at- tempt has been made to examine the role of analogical thinking in relation to emotionally salient rituals, we suspect that analogical reasoning plays a significant role ritual participants’ attempts to make sense of their experi- ences, at least in highly arousing rituals. Whitehouse refers to this process as “spontaneous exegetical reflection” (or SER). 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 131 Memory and Analogical Thinking 131 SER AND ANALOGICAL THINKING A key defining feature of SER is that of “analogic depth.” According to Holyoak and Thagard’s (1997) multiconstraint theory, there are three gen- eral constraints that guide the selection of analogical comparisons. These constraints are similarity, structure, and purpose. The constraint of similarity requires that analogies are guided by perceptual commonalities between the source and target. For instance, a spider web and a fishing net are similar in appearance insofar as both are comprised of interconnected strands of ma- terial. In terms of the ritual context, a superficial analogy might connect red paint used to cover a ritual instrument to blood—as both are red. Analogies based exclusively on similarities of this sort may be described as “superfi- cial,” in the sense that they have common outward appearances or visceral connotations rather than because they are formally or functionally similar. Source-target pairings based on structure and purpose, however, are capa- ble of delivering much greater depthto analogical thought. Deep analogies capitalize on parallels in the relational structure of ele- ments in the source and target domains (Gentner 1983; Gentner and Mark- man 1997; Holyoak and Thagard 1997). Although there may be many surface similarities between two analogs, the most meaningful and informative (“deep”) analogies are those in which the source and target analogs share connected systems of relations, resulting in more elaborate parallels. A spi- der web and a fishing net may have superficial (perceptual) similarities, but deeper connections might be made between their trapping functions, allow- ing further analogical connections between spiders and fisherman, or be- tween flies and fish, air and sea, and so on. In the case of the web–net anal- ogy, superficial similarities might draw attention to “deeper” functional parallels. But since ritual behavior is not functional in the same way as spi- ders’ webs and fishing nets (i.e., is not reducible to technical motivations), deep analogical thinking in relation to problems of ritual meaning cannot be triggered by a simple surface analog, taken in isolation. In puzzling over the question of what a particular ritual action means, we have to look for inten- tionality not in the functional properties of the action but in its place in a wider nexusof ritualized behaviors. If the instruments of ritual torture, for in- stance, are painted in random colors (or not painted at all), then their color is unlikely to be considered salient. If, however, such instruments used in rit- uals are always the same color then a search for something rather like (but not of course the same as) a “function” is more likely to be activated. Perhaps the hooks and knives are red because they can draw blood, in which case they might be analogous to the teeth of carnivores that draw blood from their victims, or (more extravagantly) analogous to drawings of reddish waxing discs in the temple, alluding to lunar cycles that “draw” menstrual blood from women. 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 132 132 Richert, Whitehouse, and Stewart Purpose (in the sense of “motivation”) also appears to contribute to the depth of analogical connections. A considerable body of experimental re- search on analogical thinking focuses on the question of what conditions make it likely for people to apply a given source analog to various target sit- uations. Recently, Blanchette and Dunbar (2001) have studied the use of analogy in natural settings. In one study, they explored the types of source analogs people will choose in political debates, by extracting and coding po- litical analogies from several newspapers (Blanchette and Dunbar 2001). The analogies were coded for the source category (the source of the analogy, which is then mapped onto the current situation, or target), the range (within-domain vs. other-domain), the goal of the analogy creator, and the emotional connotation. They found significant differences in the types of newspaper articles that contained analogies. In general, they found that analogies were most often used in opinion articles rather than articles meant to impart information. The large majority of the analogies (77 percent) were from domains other than politics, and these analogies tended to be more strongly emotionally charged than analogies that remained within the do- main of politics. Furthermore, the large majority of analogies that were em- ployed to support a given position had a positive emotional connotation, and analogies used to oppose a position had a negative emotional connota- tion. The findings suggested that analogies offer more than just factual infor- mation on a topic and seem to be well suited to the communication of emo- tionally colored meanings. Whitehouse (2004) hypothesizes that ritual actions constitute a domain of activity in which the potential for the formation of deep analogical connec- tions is considerable; the extent to which this potential is realized, however, depends on whether the ritual procedures become a focus of conscious ru- mination. The tendency for ritual procedures to spark conscious rumination is likely enhanced in high-arousal rituals by the prominence and diversity of what McCauley and Lawson call “sensory pageantry” (2002). But, according to Whitehouse, levels of conscious rumination are also determined by the relative frequency and emotionality of rituals. Frequently performed rituals give rise to implicit procedural fluency but low rates of explicit reflection. More rarely performed ritual actions, especially if they are arousing and per- sonally consequential, are likely to be remembered consciously as unique episodes in one’s life experience. In these conditions, problems of exegeti- cal meaning cannot be so easily ignored and intermittently thrust them- selves upon consciousness like a nagging puzzle that refuses to be solved. But these experiences are also different from profane puzzles. Participants in high-arousal rituals generally believe that the mysteries confronting them originate in powerful otherworldly agents, and that the consequences of misunderstanding the rituals are dire. So the motivation to gain some sort of mastery over problems of ritual exegesis, however provisional and specu- 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 133 Memory and Analogical Thinking 133 lative, is much greater than that which drives us to tinker with crossword puzzles. In general, we predict that levels of arousal during ritual performances will correlate directly with volume and structural depth in exegetical thinking. This prediction requires that our elements of SER are potentially measurable. We argue that they are. In any given ritual, it is possible in principle to de- termine what proportions of the actions are accorded exegetical meanings based on verbal testimony as opposed to independent inference, assuming one can determine exactly what participants are told about the ritual and that they are willing to report their conscious interpretations of it. We could measure the structural depth of SER by scoring exegetical commentaries ac- cording to whether or not they contain analogies. Given the minimal amount of research on this topic, we have no clear prediction about what, other than the creation of an analogy, should constitute depth in these circumstances. Depth could of course be defined in a number of ways: within-domain (e.g., one part of a ritual is analogous to another part of it) versus cross-domain (e.g., one part of a ritual is analogous to some process, object, or event in a nonritual setting), or concrete (e.g., ritual procedures that resemble other kinds of procedures) versus abstract (e.g., ritual cohesion resembles rela- tions based on kinship). Combinations of these principles are also possible, of course (within-domain/concrete, within-domain/abstract, cross-domain/ concrete, and cross-domain/abstract). Since the range of possible ways of categorizing structurally deep analogs is quite considerable, and the conse- quences of this for the construction of expert exegetical knowledge are not fully known, it is hard to determine appropriate scales of measurement in ad- vance. Thus, in developing new methods of testing Whitehouse’s predicted correlation between SER and arousal, we began by measuring sheer volume of reflection, as well as depth in the form of any analogic comparison. We re- port below on two experiments designed to test whether participants who have a strong emotional reaction to a ritual demonstrate greater volume and depth of spontaneous exegetical reflection than participants experiencing lower levels of arousal. PROPAGATION RITUAL EXPERIMENT In this study, participants enacted a particular ritual in one of two groups: a high-arousal form of the ritual or a blander version of the same ritual. To con- trol for type of exegesis, both groups were provided the same minimal amount of information about the ritual. To recruit participants, announce- ments were made in large lectures that a study of rituals was being con- ducted and students could earn £20 for their help in reconstructing a ritual. Interested students were then contacted to participate on a given day and 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 134 134 Richert, Whitehouse, and Stewart Table 7.1. Source Analog Levels Level Definition 0 Repeats the symbolic meaning given in the ritual (e.g., rubbing earth on hands was a cleansing process) 1 Attributes a functional purpose to the action or element (e.g., rubbing earth on hands was preparing for the ritual) 2 Provided an analog in the same domain after a prompt (e.g., rubbing earth on hands was becoming one with nature) 3 Provided an analog in the same domain without a prompt 4 Provided an analog in a new domain after a prompt (e.g., rubbing earth on hands was cleansing your soul) 5 Provided an analog in a new domain without a prompt participated in a large group with the other participants in that condition. We had a total of 29 participants (mean age (cid:2) 21, 15 male, 14 female) broken into two groups (14 high arousal, 15 low arousal). The ritual was conducted outdoors in a large field fringed by trees, owned by the university where the authors work. The participants gathered in a parking lot next to the field and were given a brief introduction to the proj- ect. Participants were told that they would be directed through some simple ritual actions and then asked to fill out some questionnaires following. They were also told that they would be asked to return at two other times to meet with one of the experimenters for follow-up interviews. They were told that they would receive £10 at the completion of the first follow-up interview, in one week, and an additional £10 upon completion of the second follow-up interview, two months later. Participants were then told some information about the ritual itself. The experimenter mentioned that he was interested in testing the efficacy of certain ritual procedures. Participants were told that this particular ritual was derived from propagation rituals in Amazonia, often conducted to increase hunting success. Participants were asked to maintain an attitude of respect toward the ritual procedures and not to discuss the rit- ual with anyone until after the final interview, scheduled two months later. Following this introduction, participants were led through the ritual itself. Participants in the low-arousal condition performed the ritual in the after- noon, and participants in the high-arousal condition performed the ritual at dusk. Both groups of participants were instructed to stand in a line and then follow one of the authors (hereafter referred to as the “leader”) across a field. As they were walking across the field, the leader shook a rattle. Par- ticipants were led into a circle of tree stumps and asked to bend down and “wash” their hands in the rotting leaves on the ground. Next, they were given cloaks made of Hessian fabric to wear. The leader then led the partic- ipants, again in a line, out of the trees and back into the field. As they walked toward the field, participants were handed a long stick. Participants were in- 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 135 Memory and Analogical Thinking 135 structed to break a small piece off the top of the stick and throw it behind them. They were then asked to draw a circle around themselves with the stick. At this point, the leader gave each participant a small stone and in- structed him/her to grind it into the earth. After all participants received their stone, the leader told them to the plant their “spear” (i.e., stick) into the ground between their feet. The next phase of the ritual varied for low- and high-arousal groups. Fol- lowing the stick planting, the participants in the high-arousal group were given blindfolds to wear. Participants in the low-arousal group did not wear blindfolds. In both conditions, participants were led individually to a hole in the ground, beside which a torch was lit. As participants walked toward the hole, drums were played. Once participants reached the hole, the drumming stopped and participants were asked to kneel at the hole and place their hands on a piece of fur that was in the hole. The leader chanted the mean- ingless utterance “wadih nyalu ama asumga,” and then participants were led back to their original position. Once all participants were back in line, they were instructed to remove the blindfolds (in the case of the high-arousal par- ticipants), throw their sticks as hard as possible into the trees, and follow the leader in a line back out of the field. After the ritual, all participants were given two questionnaires: an emo- tional rating form and a feedback form. Our analysis here is based on the emotional rating form, which asked participants to indicate whether they had at any point felt any of six emotions (frightened, nervous, relaxed, con- fused, uncomfortable, and bored) and the intensity of that emotion. Two and a half months after their original participation, participants returned for a fi- nal follow-up interview. Six participants did not return for the follow-up in- terview, and thus the final sample consisted of 23 participants (mean age (cid:2) 22 years). In the high-arousal group, there were eight women and four men. In the low-arousal group, there were four women and seven men. For the purposes of this chapter, only one aspect of these interviews will be reported. Participants were asked to recall freely the events of the evening or afternoon that they participated in the interview. They were also asked to indicate what thoughts crossed their mind while performing the ritual and if any aspect of the ritual struck them as especially significant or important. If participants did not expand on the possible meanings of a ritual action, the interviewer questioned them directly on whether that action seemed impor- tant and for what reason. Participants’ responses to these questions were coded for volume of re- flection and use of analogies. Volume was assessed by tallying the number of actions to which participants attributed meaning. To establish a measure of depth, each meaning was then coded as to whether it was a specific ana- log (e.g., the stone symbolized a seed, and pushing it into the ground was like planting it to grow) or not. The specific analogs were given a score of 2 05-022 Ch 07.qxd 2/10/05 4:55 PM Page 136 136 Richert, Whitehouse, and Stewart each, and all other meanings were given a score of 1. These scores were then totaled for a cumulative SER score for the final interview. Results and Discussion Participants’ responses to the initial emotional rating were analyzed to es- tablish that our two groups of participants (high-arousal ritual condition and low-arousal ritual condition) indeed had significantly different emotional ex- periences. Immediately following the ritual experience, participants in the high-arousal ritual condition reported a significantly greater intensity of fright (M (cid:2) 1.83, SD (cid:2) 1.90) than participants in the low-arousal ritual con- dition (M(cid:2).45, SD(cid:2).82; t[21] (cid:2)2.22, p(cid:3).05). There was also a trend to- ward significance in reported levels of relaxation. Participants in the high- arousal ritual condition (M (cid:2) 1.50, SD (cid:2) 1.57) reported lower levels of relaxation than participants in the low-arousal ritual condition (M(cid:2)2.73, SD (cid:2)1.56; t[21] (cid:2)1.88, p(cid:2).07). Thus, our two groups reported different emo- tional reactions to the ritual experience, indicating that our two ritual expe- riences offered opportunities for varied emotional reactions to the ritual. However, a key element of our hypothesis is that individual emotional re- actions will predict volume and depth of reflection on the experience. Given that some of the participants in the low-arousal ritual condition reacted with fear to the ritual, and some participants in the high-arousal ritual condition reported no emotional reaction to the ritual, we created post-hoc arousal groups based on participants’ original reports of intensity of fright. If partic- ipants reported not being frightened at all, they were placed in the low- arousal group (n(cid:2)11; mean age (cid:2)20; 4 female, 7 male), and all other par- ticipants were placed in the high-arousal group (n (cid:2) 12; mean age (cid:2) 24; 8 female, 4 male). Since there were different gender breakdowns for the two groups (in the high-arousal group there were eight women and four men, compared to the low-arousal group where there were four women and seven men), we needed to ensure that any differences between our ritual conditions were not attributable to gender differences. Thus, all the following analyses were first conducted by comparing the two gender groups and revealed no sig- nificant differences by gender. The remaining analyses compared the two post-hoc arousal conditions. To assess differences in SER for the two groups, the mean number of ac- tions attributed meanings, as well as the cumulative SER scores themselves, were compared using independent-samples t-tests. The mean cumulative SER scores for the final interviews are demonstrated in figure 1. First, partic- ipants in the high-arousal group (M (cid:2) 11.25, SD (cid:2) 8.44) on average attrib- uted meanings to a significantly greater number of actions than did partici- pants in the low-arousal group (M(cid:2)6.27, SD(cid:2)3.61; t[21] (cid:2)1.81, p(cid:3).05).