Research Articles
PIERRE LI ´ENARD
PASCAL BOYER
Whence Collective Rituals? A Cultural Selection
Model of Ritualized Behavior
ABSTRACT Ritualized behavior is a specific way of organizing the flow of action, characterized by stereotypy, rigidity in performance,
a feeling of compulsion, and specific themes, in particular the potential danger from contamination, predation, and social hazard. We
proposed elsewhere a neurocognitive model of ritualized behavior in human development and pathology, as based on the activation
of a specific hazard-precaution system specialized in the detection of and response to potential threats. We show how certain features
of collective rituals—by conveying information about potential danger and presenting appropriate reaction as a sequence of rigidly
described precautionary measures—probably activate this neurocognitive system. This makes some collective ritual sequences highly
attention-demanding and intuitively compelling and contributes to their transmission from place to place or generation to generation.
The recurrence of ritualized behavior as a central feature of collective ceremonies may be explained as a consequence of this bias in
selective transmission. [Keywords: ritual, cognition, evolution, epidemiology, cultural transmission]
WHY DO PEOPLE, the world over, seem compelled
to engage in ritual practices? Why invest time and
resources in such behaviors? We suggest here that we may
have the rudiments of an answer. Rituals are compelling
because specific aspects of human cognitive architecture
make these behavioral sequences attention-grabbing, intuitively
appropriate, and compelling. Specifically, we consider
that particular sequences of collective rituals activate
a cognitive-emotional system focused on the detection of
and reaction to potential danger. This hazard-precaution system
responds to a specific set of cues in people’s environments
and makes certain types of precautionary action seem
intuitively appropriate. The system is manifest not just in
reactions to potential danger but also in individual ritualization,
either normal or pathological (Boyer and Li´enard
in press).
We propose that instructions and actions typically
found in collective rituals share core features with the information
that normally activate the hazard-precaution system.
This makes the ritual procedures attention-grabbing
and compelling, which in turns explains their good cultural
transmission.
AMERICAN ANTHROPOLOGIST, Vol. 108, Issue 4, pp. 814–827, ISSN 0002-7294, electronic ISSN 1548-1433. C 2006 by the American Anthropological
Association. All rights reserved. Please direct all requests for permission to photocopy or reproduce article content through the University of California
Press’s Rights and Permissions website, at http://www.ucpress.edu/journals/rights.htm.
NO “THEORY OF RITUAL”: A MODEL OF RITUALIZED
BEHAVIOR
According to the late Roy Rappaport, a proper account of
ritual should address the question why do human beings
engage in rituals at all?, which remains unanswered in anthropological
or psychological theories (Rappaport 1999).
There are specific reasons for this failure but also a general
problem with the very notion of a “theory of ritual.”
The problem lies with the concept of ritual itself. There
is no clear criterion by which cultural anthropologists or
other scholars of religion or classics determine that a particular
type of behavior is or is not an instance of a ritual.
True, there seem to be many “definitions” of ritual
in anthropology (see, e.g., Gluckman 1975, among many
others). But these so-called definitions are, in general, summaries
of causal theories (“ritual expresses symbolism,” “ritual
is the manifestation of social status,” etc.) rather than
behaviorally precise criteria. “Ritual,” like “marriage” or “religion,”
is not a proper analytical category; instead it is more
of what Rodney Needham described as a “polythetic” category,
in which, typically, ritual types A and B may share
features [m, n, p], types Y and Z may share [p, q,r], Z and
Li´enard and Boyer • Whence Collective Rituals? 815
W share [q,r, s] (Needham 1975). And although A and W
apparently do not share any major feature, they are both
called “rituals.” That is why it is certainly futile to collect
many instances of what are commonly called “rituals” and
to tabulate their common features. This too often results
in very vague formulations that would potentially apply to
any social institution.
In this project we focus on ritualized behavior, a concept
that we construe, in a manner directly inspired by Rappaport,
as a specific way of organizing the flow of behavior,
characterized by compulsion (one must perform the particular
sequence), rigidity (it must be performed the right way),
redundancy (the same actions are often repeated inside the
ritual) and goal demotion (the actions are divorced from
their usual goals; see Bloch 1974; Humphrey and Laidlaw
1993; Rappaport 1999). Note that although ritualized behavior
in this precise sense is typically the hallmark of ceremonies
we call “rituals,” it certainly is not found in all
of those. Conversely, there may be many contexts outside
“rituals” that include ritualized behavior.
Why should we abandon “rituals” and focus on ritualized
behavior? Because the latter is characterized in terms
that make the identification of particular instances empirically
tractable. There may be ambiguous cases at some
point, but this is nothing to worry about, as long as our
characterization lends itself to empirical investigation.
So we are effectively asking the question what are the
effects of ritualized behavior, such that individuals find collective
rituals attention-grabbing and participation in such
ceremonies compelling? Obviously, there are many specific
reasons (including coercion, commitment, habit, or belief)
justifying why a particular person should find a particular
ceremony of interest and participate in it. These factors
will vary enough between cultures, periods, and individuals
that they cannot explain the general recurrence of ritualized
behavior. Some general features or effects of this kind
of scripted, rigid, and so forth behavior should explain why,
all else being equal, it appears with such frequency in human
cultures.
We address this question in the framework of a cultural
selection framework. The main points of such a framework
have been explained by others (Boyd and Richerson 1985;
Durham 1991; Sperber 1985), so we only mention two
important points that impinge on our argument. First,
recurrent features of human cultures are the winners in
a constant process of generation and selection of new
variants. What we observe as cultural representations
and practices are the variants found in roughly similar
forms in a particular place (Boyd and Richerson 1985).
Those particular sets of representations have resisted better
than other changes and distortions through innumerable
processes of acquisition, storage, inference, and communication
(Boyer 1998; Sperber 1985). This may be because
they constitute cognitive “attractors”—that is, optimal
activation of particular mental resources (Sperber 1996).
Second, cultural transmission does not consist in the
downloading of information from cultural elders, but in
the inferential construction of conceptual structures from
available (and generally fragmentary) information (Sperber
1996). So the acquisition, storage, and communication
of those representations we call “cultural” are crucially
affected by general features of human minds that should
and can be independently established. Inferential processes
are generally not accessible to conscious inspection, which
is why experimental methods are required. Applied to the
problem at hand, this suggests the following:
r There are collective rituals in human groups because
certain sets of actions are selected through cultural transmission
as more compelling or “natural” than other possible
sets of actions. We need not assume a specific human
need or capacity to perform collective rituals. All we
have to assume is that, in given circumstances, these sets
of actions seem more appropriate than others—certain
ritual sequences are found more attention grabbing or
memorable than others.
and
r The selection can be explained in terms of specific features
of human psychological architecture. Rituals are
not performed simply because “that’s the rule” and because
people absorb the conceptual schemata of their
cultures. Ritual performances produce specific effects in
participants that result in subsequent performance. So
we must document those cognitive systems most likely
to be activated by ritual performances and gather independent
evidence on the effect of such activation.
THE “OBVIOUS” FEATURES OF RITUALIZED BEHAVIOR
An Example
Let us start with an example of the kind of behavior we consider
here. This is taken from the ethnographic fieldwork of
one of us (Li´enard) among the Turkana of Kenya. As this
vignette serves as illustration only, we deliberately omitted
all the rich cultural background that is associated with this
particular sequence of actions (see Li´enard in press for the
relevant information).
The ritual sequence partly described here is called ariwo.
For this specific instance, it entails the sacrifice of an ox. Its
coat should be of a specific color and shine. The animal
should ideally be sacrificed by a left-handed twin. In the
sequence preceding the sacrifice, ritual participants circumambulate
the ritual scene three more times and then gather
in a semicircle, facing East. The animal is made to go around
the dancers three times counterclockwise. At some point in
the ritual, the members of the clan offering the ox approach
one at a time the sacrificial ox and carefully rub their body
from forehead to loin on the animal’s forehead, in a gentle
upward thrust, an operation made difficult by the animal’s
attempts to get loose and to shake its head violently.
The ritual officer cuts the animal lengthwise at the level
of the diaphragm and upper abdomen. The body is then
spread at the center of the ritual scene. During the next
phase of the ritual, clans regroup and people line up to
816 American Anthropologist • Vol. 108, No. 4 • December 2006
cross finally the ritual field from West to East walking right
through the ox’s split body, being careful to tread on a puddle
of chyme—taken from the animal’s stomach—in which
has been placed the axe used to give the ultimate death
blow. Among the crowd of each clan, elders and men are first
to pass, followed by adolescent girls and girls of marriageable
age, then come the mothers with children, and finally
the young unmarried men. The sacrificer and his assistants
make sure that everyone passing through the carcass steps
on the axe placed in the chyme before proceeding.
Now ethnographers observing this kind of behavior
would proceed to do two things. They would describe the
particular reasons why people engaged in these specific acts
on that occasion and they would detail the many representations
that Turkana participants are likely to associate
with these different actions. These two kinds of descriptions
are essential to Turkana ethnography and are indeed provided
elsewhere (Li´enard 2006). Neither description, however,
will be sufficient to address the fundamental question
raised here: Why engage in ritualized behavior in this context?
To do that, we must first outline the general features
of ritualized behavior that this sequence illustrates.
The Obvious Features
When Rappaport enjoined anthropologists to explain ritual,
he listed what he called the “obvious” (i.e., obvious
to all anthropologists) aspects of ritual—those frequent features
that a decent model should explain (Rappaport 1979):
r No obvious empirical goals: “meaningless” acts. There is
no clear point in walking the ox round the dancers—
or, for that matter, the dancers around the scene. More
generally, in rituals one typically washes instruments
that are already clean, one enters rooms to exit them
straightaway, one talks to interlocutors that are manifestly
absent, and so forth. Frequent repetition bolsters
this intuition that actions are disconnected from their ordinary
goals. People bow or kneel repeatedly; they walk
round an animal seven times, which clearly signals that
no effect is achieved by any specific iteration of the action.
Also, many rituals include actions for which there
could not possibly be any clear empirical goal, such as
rubbing an animal’s forehead with one’s body, passing a
chicken from hand to hand in a circle, or going round
a temple seven times. True, a given ritual generally has
a specific purpose (e.g., healing a particular person) but
the set of sequences that compose the ritual are not connected
to this goal in the same way as subactions connect
to subgoals in ordinary behavior (Boyer 1994). In other
words, the standard connections between means and
ends seem broken in ritual. Practitioners themselves often
concur that the actions are “meaningless,” although
efficacious (Humphrey and Laidlaw 1993). Obviously, in
many religious traditions some scholarly specialists and
theologians can produce justifications for each particular
action but these rationalizations are absent in most
nonliterate societies (Barth 1987) and remain inaccessible
or uninteresting to most participants in all traditions
(Boyer 2001). Moreover, far from providing a straightforward
rationale for ritual actions, the specialists’ exegesis
often creates mysteries that require further symbolic exegesis
(Sperber 1975).
r Compulsion. In the Turkana example, a diviner has established
that people of a particular clan should offer, on
behalf of everyone, the sacrifice of an ox as a remedy
against misfortune but also as a protection against potential
risks like enemy raids, intrusions, and pandemics.
Even though the actual risk incurred is loosely specified
employing commonplace threats, the decision to act is
impervious to scrutiny. More generally, given certain circumstances,
people just feel that they must perform a
specific ritual, that it would be dangerous, unsafe, or improper
not to do it. It is important to distinguish these
feelings, this compulsive character of ritualized action,
from the explanations people may have about the reasons
for performing the ritual (Boyer 2001).
r Literalism and rigidity. Turkana people will be careful to
walk the ox three times counterclockwise round the
dancers, to have specific participants rub in a particular
way their body on the animal’s forehead, and to have
everybody properly crossing the ritual field, stepping at
one point on the axe in the chyme. The intricacy of the
details and the prescriptions is the hallmark of ritualized
behavior. People feel that they should perform the ritual
in the exact way prescribed and generally in the way
it was performed on previous occasions. This obviously
does not mean that ceremonies are actually performed in
the same way (Goody 1972, 1986). What is important is
that people strive to achieve a performance that matches
their representation of past performances, and that they
attach great emotional weight to any deviation from that
remembered pattern (Boyer 1990).
r Repetition, reiteration, redundancy. Repeated enactments of
the same action or gesture—as well as reiterations of the
same utterances—are typical of many collective rituals.
A given sequence is executed three or five or ten times.
What matters is the exact number. What matters too is
that the action should seem identical in all these iterations.
This makes many ritual sequences clearly distinct
from everyday action, in which there is either no repetition
of identical sequences (e.g., in assembling a musical
instrument, one performs a series of unique actions), or
each repeated sequence has a specific outcome (in weaving,
the warp is changed at each step), or repetition is
cumulative (the egg whites rise only after a long period
of stirring). In ritual action, repetition itself is not motivated
but strongly prescribed and perceived as intrinsically
efficacious.
r Order and boundaries. This was visible in our Turkana vignette.
There is a special space of the sacrifice, where
normal behavior is suspended and particular acts must
be performed. The dancers must be aligned in a particular
way relative to the animal. They must walk through the
body of the sacrificed animal along a specific path and
Li´enard and Boyer • Whence Collective Rituals? 817
direction. In many rituals, people create an orderly environment
that is quite different from the one of everyday
interaction. People line up instead of walking, they
dance instead of moving, they wear special clothes or
makeup, they build alignments of rocks or logs, they create
elaborate color and shape combinations, and so forth.
There is a lot of ordering in rituals that is quite distinct
from the comparatively unpredictable patterns of nonritual
environments. Related to this is the recurrent concern
with delimiting a particular space (a sacred circle,
a taboo territory) often made visually distinct from the
other, unmarked space. People also often emphasize the
boundary between this space and the rest, for instance
by special prohibitions (only men enter the sacred circle,
only women sit on the left side, etc.) or by restrictions
on communication between marked and unmarked
spaces.
r Specific concerns. Pollution and cleansing, protection
against invisible dangers, and the creation of a special
space and time are common themes associated with ritualized
behavior. It is clearly the case for our Turkana ritual.
The whole endeavor is understood and justified as a
way of countering and chasing away a menace. The danger
is understood as a threatening intrusion, an amalgam
of overexposure to potential enemy onslaught, sickness,
and other likely threats. Whipping the dancers is done
“to chase the disease away.” Profusely splashing people
at regular interval is intended “to disconnect” people
from potential ills by “cooling” them. By stepping on
the axe buried in the chyme, participants should “sever”
themselves from the disease on the spot. In many cultural
contexts, the ritual space or instruments are often
described as “pure” (or on the contrary as the locus of
concentrated “pollution”), or the point of the ritual is
to “purify” people or objects or to “cleanse” minds or
bodies. This is not just a matter of metaphors. In many
rituals, blood, semen, or excrement are a primary concern,
the miasma or smells of decaying corpses are important,
and the use of water or fire as possible ways of
getting rid of pollution and contaminants is also recurrent.
There are also innumerable examples of allusions
to purity and pollution in ritual requirements. People
must wash before prayers, they immerse themselves in
water to rid themselves of pollution, they must wear
spotless garments, the sacrificial animal must be absolutely
clean, menstruating women (supposedly polluted)
are barred from rituals spaces, and so on. This concern
with pollution and cleansing is so prevalent that it has
been considered a foundation of religious ritual (Douglas
1982).
BACKGROUND TO RITUALIZATION
Rappaport rightly pointed out that most cultural anthropological
accounts of ritual simply ignore the “obvious” properties
listed above (Rappaport 1979). Many ethnographic
accounts focus on the specific reasons for participation in
ritual in a given cultural context (e.g., a specific ritual legitimizes
claims to high status or marks territorial boundaries)
rather than the general features of this kind of behavior
(Gluckman 1975). Also anthropological theories often emphasize
processes (the transmission of norms or of shared
cultural symbolism, or a demonstration of social commitment,
etc.) that could and do occur also outside ritualized
behavior; these, therefore, do not explain the specific
features of rigidity, scriptedness, and so forth (Rappaport
1999; Staal 1990). This is particularly true of functionalist
accounts. Some theories claim that rituals provide ethnic
affiliation or reinforce religious belief but do not provide
any account of why or how these functions require rigid
sequences or repeated episodes or any other of the recurrent
features of collective rituals. From a different perspective,
rituals are often said to be “symbolic” (Basso and Selby
1976), which has led anthropologists to see rituals as conveying
culturally coded meanings. Again, this interpretation
begs the question of why such meanings would require
ritualized actions—why they would require those actions to
be expressed or sustained. Besides, most rituals do not actually
convey coded meanings except in the vaguest sense
(Sperber 1975). Typical features of rituals—such as the use
of standardized formulaic speech, repetitions, and redundancy
as well as the great number of obviously pointless
actions—would seem to reduce the information potential
of rituals (Bloch 1974; Staal 1990).
Beyond these very general points, only a few theoretical
frameworks actually address the central question
of the kinds of action typical of ritualized behavior. We
briefly discuss them before proceeding to our cognitive
framework.
Cognitive Background: Ritual and Ordinary Action
Ritual action should be studied in the context of human
dispositions for organized action in general. Understanding
the cognitive processes engaged in the ritualization of behavior
surely requires at least some minimal understanding
of the cognitive underpinnings of action representation. Although
it may seem self-evident, the requirement is ignored
in most accounts of ritual.
A notable exception is E. Thomas Lawson and Robert
N. McCauley’s cognitive account of religious ritual (Lawson
and McCauley 1990), which inaugurated the field of cognitive
studies of religion. Although most of the argument
is about the specific features of religious ritual, and therefore
only partly relevant to our problem, it makes sense to
take this framework as a starting point. A central assumption
of the model is that ritual actions are only a subset
of actions, and therefore a large number of principles that
govern action-representation are inherited in the specific
domain of ritual. These principles specify, for instance, that
the structure of action-representation is partonomic (small
units with specific goals are included in larger units with
larger goals but do not straddle the boundaries of two larger
units; Newtson 1973; Zacks et al. 2001) and that there are
818 American Anthropologist • Vol. 108, No. 4 • December 2006
only limited possible mappings between specific ontological
categories (person, object, animal) and “slots” in actiondescription
(agent, patient, instrument; see Lawson and
McCauley 1990). Religious believers’ intuitions about the
format of their rituals, as well as about the format of possible
rituals they never witnessed, are governed by these general
rules of action-representation (Barrett and Lawson 2001).
What is specific to religious ritual, then, is not a way of representing
action but the insertion of supernatural agents in
the agent and patient positions of otherwise ordinary rules
(McCauley and Lawson 2002).
Phylogeny of Ritual
What is the connection between collective rituals practiced
by human beings, on the one hand, and animal displays and
routines, on the other hand? Some authors have chosen to
emphasize the common features, which are particularly important
as regards the sequencing and repetition of action
(Gluckman 1975; Staal 1990). However, it is quite clear that
human rituals include specific preoccupations (e.g., with ritual
cleansing, with unseen dangers) that do not seem to be
present in animal displays or other rituals (Fiske and Haslam
1997). Debating whether animal and typically human rituals
are continuous or not may not be the most fruitful
strategy, unless we understand what these parallels could
explain and how.
The main point of continuity between collective and
animal rituals may lie in the simplification of communication,
in the ostension of intentions that is accomplished
through rituals. Rituals in many animals reduce
the cost of communication by transmitting honest signals
of fitness or dispositions (Bradbury and Vehrencamp 2000;
Gintis et al. 2001). For instance, ritualized fights save males
the fitness cost of actual fights (Watanabe and Smuts 1999).
More generally, displays and rituals make the transmission
of simpler messages more efficient (Payne 1998). Collective
rituals too set up a special form of communication with
greatly impoverished propositional content (Bloch 1974,
1998). The relationship between expressive displays, on the
one hand, and ritualized behavior, on the other hand, is
not as straightforward as one might imagine. Expressive
behavior finds its raison d’ˆetre in the communication of
signals relevant for interaction. Although collective rituals
do fulfill specific social functions and goals (among which
the transmission of specific signals should be counted), we
believe that the characteristic features and contents of human
ritualized behavior cannot be accounted only in those
terms.
Furthermore, phylogenetic considerations do not by
themselves provide a full evolutionary account of any behavior
(Tooby and Cosmides 1989)—that is, even if we consider
that ritualization in several species is a single phenomenon,
it remains to explain whether rituals in humans
are only an atavism or constitute a specific adaptation.
What would be the possible adaptive value of this propensity
for rigid scripting of joint action?
Adaptationist Models: Group Cohesion
and Signaling
Some evolutionary anthropologists have suggested that
rituals create coordination and therefore group cohesion
(Sosis 2000). Collective rituals typically require group-scale
coordination between agents. However, although it is true
that rituals require coordination, it is also true that many
domains of everyday social interaction in humans require
it too but are not accompanied by ritualized behavior. Indeed,
part of our human-evolved cognitive equipment consists
of cognitive capacities that allow large-scale coalitional
alliances, which imply some measure of coordinated action
(Harcourt and de Waal 1992; Kurzban et al. 2001). This
would suggest that there is no need for humans to use rituals
as a way of creating social alliances: They already have
the tools for that. It is clear, however, that rituals may seem
intuitively appropriate for demonstrations of commitment
(Gintis et al. 2001; Sosis 2003). To perform the ceremonies,
people must not only agree on a particular script of what
actions must be performed when, but also they must collaborate
in the performance and they display this cooperation.
In this way, rituals may reinforce not just cohesion but
public commitment to cohesion, which in itself constitutes
a powerful incentive for other agents’ display of commitment
(Kuran 1998).
This is related to the fact that many collective rituals
are exceedingly expensive, either in resources or in direct
fitness, as they put some participants in great danger.
Male initiation, for instance, is said to turn boys into fullblown
men, but its often painful ordeals typically transmit
none of the skills or knowledge associated with manhood
in the groups concerned (Bloch 1992; Houseman 2002).
What would motivate people to engage in such rituals?
One possible answer is that these rituals constitute “commitment”
devices (Frank 1988; Schelling 1960). Expensive
or fitness-reducing rituals may constitute an elaborate form
of signaling fitness, not despite their costs but because of
them (Zahavi and Zahavi 1997). Evolutionary anthropologists
have noted that expensive religious rituals in particular
constitute hard-to-fake signals of commitment to
statements, the only guarantee for which is other people’s
commitment (Bulbulia 2004; Irons 2001; Sosis 2003).
RITUALIZATION PROPER, INDIVIDUAL
AND COLLECTIVE
We consider all these findings and hypotheses of great
interest—indeed, they provide the foundations of the
model presented below. However, it is also clear that they
still leave a crucial piece of the explanation missing. There
is simply no description of the capacities involved in ritualized
behavior. There is no general account of how and
why human minds find these sequences attention demanding
and compelling. In other words, Rappaport’s question
remains largely unanswered. Cognitive models focus on
continuities between ordinary and ritual action, and it
remains to explain the differences. Phylogenetic models
Li´enard and Boyer • Whence Collective Rituals? 819
provide little explanation for the possible adaptive value
of a human disposition to perform rituals. Adaptationist
models are still tentative, as they do not include a precise
description of the cognitive machinery underpinning ritual
action. So they fail so far to provide the proximate cognitive
mechanisms required for testable evolutionary psychological
models (Ketelaar and Ellis 2000). This is precisely what
we aim to provide here.
Diverse Domains of Ritualization
In our view, features of collective rituals can be explained
only if we leave aside cultural institutions for a while
and turn to ritualized behavior that can be observed in
diverse circumstances and with different consequences.
This is why we must review the occurrences of such behavior
in contexts that are neither collective nor “ritual”
in the usual anthropological sense—behaviors whose
psychological underpinnings may illuminate collective
ritualization.
r Children’s rituals. Most children engage in ritualistic behaviors
at a particular stage of development, starting at
age two, peaking at age five, and subsiding around age
seven. These behaviors are usually considered part of normal
development, distinct from the severe symptoms of
OCD (see below; Leonard et al. 1990). The age of onset is
similar in different cultures, as are the themes of ritualistic
behavior: perfectionism, attachment to favorite objects,
concerns about dirt and cleanliness, preoccupation
with just-right ordering of objects, preferred household
routines (Zohar and Felz 2001). Specific action sequences
must be performed—there is a compulsion to engage in
the activity. The ritual must also be performed in a precise
way, as deviations are intuitively felt to be dangerous
(Evans et al. 1999).
r Obsessive-compulsive disorder. In some people, intrusive
thoughts and compulsions can evolve into full-blown
obsessive-compulsive disorder. The main feature of the
pathology is a strong compulsion to engage in stereotyped
and repetitive activities with no rational justification
(American Psychiatric Association 1994). Some
patients engage in bouts of washing or cleaning tools
or utensils (Hodgson and Rachman 1972). Others verify
that they locked their door, rolled up the car window,
or turned off the gas knobs over and over again
(Hodgson and Rachman 1977). Still others engage in constant
counting activities or need to group objects in sets
of particular numbers (Radomsky et al. 2001). In most
cases the ritual seems to be an intuitive response to obsessive
thoughts about potential danger, notably contamination
and contagion (e.g., fear to catch other people’s
germs, to ingest contaminated substances, to pass on diseases
to one’s children or others), possible harm to others
or to oneself (e.g., handling kitchen utensils and wounding
people), as well as social ostracism following shameful
or aggressive acts (thoughts about assaulting others,
shouting obscenities, exhibitionism, etc.).
r Life-stage-relevant intrusive thoughts. Thoughts about potential
danger and appropriate precautions are not confined
to the clinical population. On the contrary, systematic
studies of the themes of OCD have shown that
most normal people experience the same kind of intrusive
thoughts as patients and to some degree generate
the same ritualized action-plans to avoid such dangers
(Rachman and de Silva 1978). Intrusive thoughts often
become more focused and more bothersome at particular
phases in the lifespan, notably the final stages of
pregnancy and motherhood and fatherhood. A review of
these phenomena suggests that senseless, intrusive, unacceptable
ideas, thoughts, urges, and images about infants
are common among healthy parents of newborns.
The content of these intrusions resembles that found in
clinical obsessions (Abramowitz et al. 2003). A common
underlying theme is uncertainty and doubt concerning
whether one may be responsible for harm to the infant
(Abramowitz et al. 2003). In the same way as in OCD,
some ritualized behaviors develop, repetitive checking
in particular (Leckman et al. 1999).
The Fiske Hypothesis
Despite the common features, most cultural anthropologists
do not consider individual and collective rituals in the
same explanatory framework. A notable exception is Alan
Fiske (Dulaney and Fiske 1994; Fiske and Haslam 1997),
who reopened an issue famously framed by Sigmund Freud
(1928, 1948) a long time ago. Freud had commented on the
obvious similarities, in terms of repetition and compulsion,
between individual obsessive compulsion and the prescriptions
of religious ceremonies. Although Freud’s general theory
of neuroses would certainly provide an interpretation
for some features of rituals, in particular the obsessive concern
with purity and pollution, this remains implicit in his
treatment of ritual, which concludes with the tantalizing
observation that obsessive neurosis should be seen as a private
cult and religion as a collective form of neurotic obsession
(Freud 1948).1
It does not, however, solve the question
of why this type of behavior appears in either context.
Many features of collective rituals, beyond those emphasized
by Freud, make the comparison tempting. First,
collective rituals are often centered on themes related to
potential danger. Second, a constant theme in collective rituals
is that nonperformance is highly dangerous, although
in many cases no coherent account is available. Third, the
actions are highly scripted and there is a definite intuition
that deviations from the script are dangerous. Fourth, the
actions are highly repetitive and redundant. Fifth, the specific
actions performed are divorced from their usual goals.
So it would seem perverse to try to explain collective ritual
without considering this and the other domains of ritualization
as (clearly different) manifestations of a common
set of cognitive processes.
Comparing hundreds of ritual sequences with clinical
descriptions of OCD cases, Fiske and colleagues showed that
820 American Anthropologist • Vol. 108, No. 4 • December 2006
the same themes recur over and over again in both domains
(Dulaney and Fiske 1994; Fiske and Haslam 1997). OCDtypical
features that also enter into rituals include specific
(lucky or unlucky) numbers, use of special colors, repetition
of actions, measures to prevent harm, ordering and symmetry,
stylized verbal expressions, washing, concern with
contagion, and so forth (Fiske and Haslam 1997). Those
thoughts and practices are “egodystonic” in personal rituals,
perceived as unwanted, unpleasant, shameful, or irrational.
But the very same thoughts and practices are socially
approved in collective rituals.
To Fiske, the similarities and differences between individual
and socially acquired rituals suggest that a specific
human capacity to perform rituals, usually channeled toward
socially approved contexts, becomes hyperactive in
pathology (Fiske and Haslam 1997). In this view, humans
developed a capacity to engage in coordinated social action
and also developed a subset of this as a capacity for
collective rituals. The capacity helps channel personal fears
or doubt into culturally transmitted conceptual schemes,
thereby making them shared and probably less anxiogenic.
This is because rituals, at least during performance, seem to
make environments simpler, more predictable, and more
meaningful than ordinary action (Fiske and Haslam 1997).
In our view, the Fiske model provides some important
elements of a general account of collective rituals—not least
of which is the crucial comparison with the cognitive processes
engaged in individual rituals. However, we also think
the hypothesis may be less than parsimonious in terms of
explaining ritualization in general.
Specifically, it seems difficult to postulate a general capacity
for collective ritualized behavior. As far as cognitive
processes are concerned, most aspects of collective ritual can
be explained in terms of capacities documented in other,
nonritual contexts. Collective rituals make use of scripted
actions but so do recipes and other routinized behaviors
(Abelson 1981; Zacks et al. 2001). They focus on securityrelated
issues but so do many intrusive thoughts outside
rituals (Rachman and de Silva 1978). Collective rituals engage
low-level parsing of action, but that is also true of other
circumstances (Lassiter et al. 2002). What is specific to ritual
is the combination of all these elements in a process that
remains to be explained.
A NEUROCOGNITIVE MODEL OF INDIVIDUAL
RITUALIZED BEHAVIOR
On the basis of anthropological and neuropsychological evidence,
we proposed elsewhere (Boyer and Li´enard in press)
a synthetic model of individual ritualized behavior that we
can only summarize here. This comprises the following:
(1) a description of the neurocognitive systems involved
in individual ritualized behavior, derived from specific neuropsychological
models (Saxena et al. 1998; Szechtman and
Woody 2004); (2) a speculative description of the type of
action-representation engaged by ritual prescriptions; and
(3) a description of the short- and long-term cognitive effects
of ritualized behavior.
Potential Danger and Relevant Cognitive Systems
A variety of findings converge to suggest a specific system
for dealing with potential danger in the brains of humans
and many other mammal species (Szechtman and Woody
2004). We will call this the hazard-precaution system. Its
neural correlates are distinct from those of fear-systems that
respond to actual danger (LeDoux 2003). For humans as
well as many other species, survival and reproductive success
require not just avoidance of present danger but also
detection of indirect clues for fitness threats. So it is not surprising
to find that the human hazard-precaution system
seems to be specifically focused on such recurrent threats
as predation, intrusion by strangers, contamination, contagion,
social offence, and harm to offspring (Szechtman and
Woody 2004). The system does not seem to respond in the
same way to more recent and actually far more dangerous
stimuli such as tobacco or cars (Mathews et al. 2004). We
call this system “hazard precaution” because it comprises
both (1) some specific reaction to potential danger clues and
(2) rudimentary descriptions of appropriate precautions, including
avoidance (of other people), contact avoidance and
disgust (against contamination), and attention to traces and
indirect signals (against intrusion and predation).
Individual intrusive thoughts—pathological or not—
are often focused on a small range of items and concepts
connected to these evolutionary threats. Individual
rituals—pathological or not—often include the range of
appropriate reactions to these specific potential dangers
(Mataix-Cols et al. 2005):
Thoughts about contamination and contagion trigger
compulsions centered on washing and cleansing, as well
as precautionary measures, such as protecting oneself
from intrusive material by moving to a different place,
avoiding contact, avoiding breathing, and so forth.
Thoughts about possible harm to one’s own offspring
trigger fears of handling tools and utensils in a dangerous
way, smothering or dropping the infant, and forgetting
about the baby or losing it (particularly in stores and
other public places) with predictable precautionary measures:
constant monitoring or repeated checking.
Thoughts about possible acts that would offend or upset
other people, resulting in social exclusion trigger a
hyperactive monitoring of one’s own actions, in particular
the minutiae of one’s own behavior, well beyond the
“normal” limits. Another common feature is that people
choose to avoid social contact lest they insult or assault
others, which again is intuitively appropriate as a precautionary
device.
Obviously, the workings of this system—and of other
neurocognitive systems of this kind—are not accessible to
conscious inspection. The systems’ hidden computations
result in a specific level of anxiety (or absence thereof)
and a heightened attention focused on particular objects
Li´enard and Boyer • Whence Collective Rituals? 821
(features of a landscape that make predation possible, features
of a body that make contagion likely, etc.), as well
as the compulsion to engage in particular courses of action
(washing, avoiding contact, checking one’s surroundings
repeatedly, etc.). Neurocognitive models explain OCD
pathology as a form of hyperactivation of this system
(Szechtman and Woody 2004). In our view, nonpathological
activation would explain compelling precautionary
behaviors in many situations; it may also explain specific
intrusive precaution-related thoughts at particular stages of
the life cycle; children’s ritualized behavior may correspond
to a phase of calibration of this security-focused system
(Boyer and Li´enard in press).
Complex Ritual Rules and Working Memory
Another part of the model is a speculative explanation for
the presence of complicated prescriptions. As we noted
above, a hallmark of ritualized behavior is the activation
of complex rules: Do x but not y, walk in a particular gait,
handle an object in a particular way, and so on. The variations
seem potentially infinite, but there is a recurrent feature
in all these rules. They generally turn usually automatic
or routinized behavior (walking, washing, getting dressed,
etc.) into highly controlled action that requires sustained
attention. An example is having to tie one’s shoelaces three
times on the right foot and four times on the left. The difference
with simply tying one’s shoelaces is that the latter
can be done without attending to the action, whereas the
former requires sustained focus.
Neuropsychologists have reported that some patients’
compulsive rituals result in “swamping” of working memory,
so that the person cannot attend to stimuli and situations
outside the ritualized action (Ursu et al. 2003;
Zalla et al. 2004). Also, many patients state that performing
the ritual is one way of inhibiting or repressing unwanted
thoughts (Salkovskis 1985). In our view, these facts legitimate
the conjecture that complicated prescriptions may
constitute a spontaneous and moderately efficient form
of thought suppression, with some similarities to the suppression
processes studied experimentally by psychologists
(Wegner 1994; Wegner et al. 1987). In other words, patients
with complicated compulsions spontaneously design a kind
of activity so demanding in cognitive control that intrusive
thoughts can be, at least for a while, pushed away from
consciousness (Boyer and Li´enard in press).
Ironic Outcomes
Finally, the model specifies that ritualized behavior has particular
effects on the salience of the thoughts that elicited
this form of behavior. Here we take inspiration from the
study of voluntary thought suppression. Studying normal
subjects instructed not to think about a particular item,
Wegner showed that thought suppression typically results
in a “rebound,” a higher salience of the unwanted thoughts
(Wegner and Schneider 2003). In Wegner’s model, this
is caused by the combination of two distinct processes
engaged in thought suppression. Although an explicit process
directs and monitors the suppression, implicit processes
are engaged that detect material associated with the
target item (Wegner and Erskine 2003). Clinical models of
OCD also concur in the conclusion that rituals often seem
to exacerbate the obsession they temporarily appease (Rachman
and de Silva 1978). The patients who perform more rituals
are typically more anxious and also more bothered by
their intrusive thoughts. The thoughts also seem to become
more frequent with higher ritualization. In other words, the
long-term effects of ritual performance may be the opposite
of its short-term results.
A MODEL OF COLLECTIVE RITUALS
Our interpretation of ritualized behavior in collective ceremonies
is not that these constitute individual rituals writ
large, or even that they are generated in the same way as
the individual compulsions described so far. Our model
is based on cultural selection assumptions. We start from
the premise that particular collective rituals are culturally
successful (i.e., remembered by people and found compelling
enough to perform again) to the extent that they
activate specific individual neurocognitive systems more
or more specifically than other possible variants. Specifically,
we want to suggest that many details of collective rituals,
when observed by normal human minds, activate the
information-processing and motivation systems described
above, which makes these behaviors attention-grabbing
and compelling to participants, more so than alternates eliciting
weaker operation of the hazard-precaution system.
Cognitive Capture of the Hazard-Precaution System
The hazard-precaution system only responds to information
that is couched in a specific input format. This, in
fact, is a general point that applies to many other functionally
specific neural systems. Linguistic parsing systems,
for instance, require words of a known language as an input;
stereoscopic vision systems require slightly different
retinal projections. To say that the hazard-precaution system
requires a particular input format is simply a more
specific way of saying that efficient information processing
requires at least some filtering out of information, so that
the system does not “fire” in each and every situation the
organism faces, as this would be grossly maladaptive. The
hazard-precaution system is activated by information about
potential danger (in a narrow potential hazard repertoire)
for which indirect clues can be found in the environment.
Another assumption is that any functional system with
a specific input format is vulnerable to cognitive capture—
that is, activation by signals that are not part of its intrinsic
functional repertoire. There is a difference between the
proper (evolutionary) domain and the (usu. broader) actual
domains of a system (Sperber 1996). Mimicry and camouflage
use this noncongruence between the functional domain
and the actual domain of inputs that activate a system.
Nonpoisonous butterflies evolve the same bright colors
822 American Anthropologist • Vol. 108, No. 4 • December 2006
as poisonous ones to avoid predation by birds. The proper
(evolved) domain of the birds’ bright-colored bug avoidance
system is the set of poisonous insects, the actual domain is
that of all insects that look like them (Sperber 1996).
Our contention is that many aspects of collective rituals
activate the hazard-precaution system by including typical
clues for relevant potential dangers. In other words witnessing
or performing the prescribed actions should result in
cognitive capture of the hazard-precaution system.
Information Available and Likely Inferences
To understand the cognitive effects of collective rituals, we
must describe the kinds of information available to various
people who participate in them. At first sight, it would seem
that most people who participate in most rituals do not have
much information at all. In most societies, they certainly
do not seem to have an explicit, coherent justification for
most aspects of the ritual, for the actions included or their
sequencing. People do not generally hold a “theory” of their
own rituals; this is what makes ethnography indispensable
and difficult.
However, this is not to say that people participate in a
ritual on the basis of mere imitation, peering at their cultural
elders and simply performing similar gestures. This
would be implausible, given that very little human communication
actually involves such mindless imitation (Sperber
2000). Remember that conveying information about one’s
behavior does not necessarily imply explaining the behavior,
explicitly commenting on it, or even having the deliberate
intention of conveying it. All that is required is that
some manifest behavior (including but not necessarily verbal
behavior) triggers nonrandom inferences about the behavior.
We propose that the information made manifest in
connection with performance of a particular ritual consists
of the following elements:
r Available information about the background situation. People
are told that a ritual should be performed and led
to infer that nonperformance is a dangerous option. For
instance, one is told that because of a particular event
(someone’s illness, a death or a birth, the change of seasons,
a war with another group, possible damnation), it
is necessary to go through a particular ritual sequence.
People also receive information and produce inferences
about the kind of danger against which the ritual is
supposed to protect the group (“pollution” by invisible
substances, attacks by invisible predators like witches
or spirits, threat of disease, possible famine, etc.) These
themes substantially overlap with the potential hazard
repertoire.
r Available information about required course of action. People
are instructed to participate in the ritual in particular
ways—that is, people are generally not allowed to just
add to their ritual whatever action they think fit. They
are enjoined, more or less explicitly, to follow a particular
script. Information about the script has the following
properties. First, action descriptions include themes
that mimic some of the typical outputs of the hazardprecaution
system: actions such as cleansing, washing,
and checking. Second, there is great emphasis on the
details of each action, inducing what we called “lowlevel
parsing” of the action flow above. This is made even
more salient by insistence on repetition, redundancy, apparently
pointless acts, and so forth. Third, descriptions
of prior conditions for the ceremony (such as particular
taboos, substances to avoid, etc.) reinforce activation of
hazard-precaution system.
In the next pages we present some evidence for these
various claims and for the psychological and cultural effects
of the processes.
Recurrent Features of Collective Rituals and
Cognitive Capture
Given the input format described above, various common
features of collective rituals should trigger activation
of the hazard-precaution system. This would be the
case in particular of information concerning the occasion
for the ritual, the danger of nonperformance, and
the details of performance.
r Occasion for the ritual. The occasions for ritual often
provide clues of possible danger that overlap with the
potential hazard repertoire. This is quite clear in most
apotropaic rituals, the explicit point of which is to prevent
disasters to fitness such as famine or illness. But the
same can be said of most therapeutic rituals, in the sense
that the ultimate cause of illness is generally described
in terms of nonobservable processes (from germs or miasma
to witchcraft). Rites of passage also include such
themes as removing pollution from newborn infants (as
in baptism) or protecting people from the danger of indirect
contact with corpses (Bloch and Parry 1982; Metcalf
and Huntington 1991).
r Danger of nonperformance. This is a very general feature
of collective rituals. In cases in which the goal of the ritual
is very clearly specified (healing rituals for instance),
the danger is also clearly specified. But this is not always
the case for rituals—that is, most people develop the intuition
that it would be really wrong and certainly dangerous
not to perform a ritual without specific thoughts
about what the risk is.
r Detailed prescriptions. Details of prescribed performance
are of course the major source of security-related motifs.
As we stated in our introduction, most rituals include
such operations as washing and cleaning, checking and
rechecking that a particular state of affairs really obtains,
or creating a symmetrical or otherwise orderly environment.
As we said above, Fiske and colleagues have documented
these features extensively (Dulaney and Fiske
1994; Fiske and Haslam 1997), so we will not comment
any further.
Our contention is that all these items of communicated
information result in a weak activation of the hazardprecaution
system. What we mean by that is that the
Li´enard and Boyer • Whence Collective Rituals? 823
activation is probably (in most participants, in most rituals,
most of the time) less intense and direct than in situations
where people actually encounter clues for potential
danger. The activation is bound to be weak because it is
mostly indirect—from other people—and mostly through
verbal communication. For instance, people are told that
witches may be lurking near the village, but this is not the
same as having the direct experience of seeing evidence of
a predator’s recent intrusion.
Consequences: Action Representation
in Collective Rituals
Activation of hazard-precaution systems may explain further
characteristics of collective rituals, in particular the way
participants represent the actions they perform and their
connection to possible goals.
r Forcing goal demotion. Ritual prescriptions seem to resort
to particular “tricks” to make actions attention grabbing.
Among these features are repetition, which creates action
chunks but without the goal ascription that is usually associated
with natural breakpoints in action flow. Another
such device, obviously, is to borrow a sequence from ordinary
scripts and perform it in a context that makes
goal ascription impossible, (wash objects without using
water, pretend to trace an imaginary line, etc.). What
results from these “tricks” is what we called “goal demotion”
above. Attention is focused on the low-level,
fine-grained description of action, so that sequences of
actions are represented without attaching a goal to each
behavioral unit, as would be the case in nonritual contexts.
This may be why the phenomenology of collective
ritual is sometimes described as analogous to a behavioral
“tunnel” in which the only action considered is the
one that will follow the present one, but one does not
and cannot focus on the motives of each action and especially
not on possible alternatives (Bloch 1974).
r Swamping of working memory. Many ritual prescriptions
resemble the tasks designed by cognitive psychologists
in the study of working memory. They require focused
attention on a set of different stimuli and their arrangement.
For instance, a requirement to turn round a ritual
pole three times clockwise without ever looking down
imposes executive control of two tasks at the same time.
Generally, the frequent combination of a positive prescription
(“do x”) and a negative one (“while avoiding
doing y”) would seem to engage working memory and
executive control in a way that is not usually present
in everyday action flow. The combinations of positive
and negative prescriptions generally make it difficult to
perform actions by engaging automatic routines. For instance,
in everyday contexts one can tie one’s shoelaces
in a fairly automatic way; if, however, one adds the requirement
that the laces should at no point touch the
front of the shoe or that the fingers should never touch
the sole, this imposes a high degree of attention and
disengagement of routinized action patterns. Although
rituals can become routinized, especially for ritual specialists
who perform them repeatedly, the focus on details
of action precludes automatized performance.
COLLECTIVE RITUALS IN CULTURAL TRANSMISSION
So far we discussed the psychological underpinnings of
collective ritual in terms of individual responses to socially
transmitted rituals. If accepted as (provisionally and
roughly) valid, the model would seem to raise the question,
why are collective rituals so organized? Where and what
is the engineering process that made them so effective in
terms of creating conceptual associations and motivation?
Transmission: Relevance and Selection
Collective rituals are generally not “engineered” in the sense
of a deliberate process. In our model, scripts for collective
ceremonies enjoy a transmission advantage, to the extent
that they include ritualized behavior as described here—
that is, they include enough hazard-precaution cues to activate
the relevant systems by cognitive capture. This obviously
does not imply that anyone is deliberately including
such themes in collective rituals; it only suggests that variants
of the collective rituals that do include them should,
all else being equal, be more attention grabbing and compelling
that variants that do not, and therefore they should
be potentially better transmitted. This raises the questions,
how much activation of the system is necessary? and is it
really the case that all participants share this orientation
toward the ritual ceremony?
The model only predicts a potential activation of the
hazard-precaution system—that is to say, we are not stating
that participation in collective rituals invariably triggers
full activation of the relevant circuitry. What we are saying
is that collective rituals with the kind of prescribed actions
described here are, all else being equal, more likely to result
in some activation of these systems than ceremonies that
do not include these specific kinds of prescriptions (turning
automatic action into controlled action) and these specific
themes (potential danger, pollution, etc.). We are furthermore
saying that this potential difference would ensure the
cultural prevalence of the kind of collective ceremonies we
have described here.
This does not mean that ritual’s success rests on such
activation at every occurrence of the ritual in all individual.
Our argument is probabilistic and not determinist. What
we are saying is that normal minds that process the cues
contained in ritual recipes are unlikely to treat them as
straightforward false alarms. An efficient hazard-precaution
system should probably be slightly oversensitive, and therefore
prey to the kind of cognitive capture described here. In
collective ritual, people’s insistence on the potential danger
of not following the rules—expressed as moral reprimand
(moral threat), as possible exposure to gossip or ridicule
(threat of social exclusion), or as worry about misfortune—
is very likely to activate the hazard-precaution system.
824 American Anthropologist • Vol. 108, No. 4 • December 2006
Turning to individual differences, a model that derives
cultural transmission from individual cognitive effects must
address how these effects are spread among the participants
(McCauley and Lawson 2002). Nothing in our probabilistic
model of cultural transmission requires that the
hazard-precaution systems are engaged in all participants.
It is enough that such conceptual associations occur in some
participants on some occasions. Indeed, it would be unreasonable
to expect identical reactions to the available information.
First, many participants do not seem to participate
very much, as it were, and in fact do not accomplish much
in the way of actions described here as ritualized behavior.
Some may be attentive to the ritual prescriptions (e.g., making
sure they pass the sacrificial objects clockwise, only using
their left hand) whereas others just play the thing by ear,
as it were, until corrected by others. Second, in most collective
rituals different roles are assigned to different categories
of participants (men vs. women, old vs. young, agnatic vs.
maternal kin, etc.). Indeed, in many cases participants from
these different categories may have substantially different
views of what is going on in the ritual. These differences
in perspective may be essential to the attention-grabbing
potential as well as the dynamics and transmission of some
collective rituals (Houseman 2002; Houseman and Severi
1998). Here we must leave aside this important question, as
it is essentially orthogonal to our transmission claim.
In cultural transmission, as in other descent-andmodification
models, very small effects aggregated over
many cycles of transmission are sufficient to create massive
trends. That is, even in the (probably rare) limiting
case in which few people produce inferences about hazardprecaution
systems, the fact that they do and that no other
equally salient template for understanding the collective ritual
is available should result in a slightly better transmission
of hazard-precaution themes over other ones.
Implications: Ritualization as the Opposite
of Routinization
In this model, ritualized behavior is described as quite different
from routinized action; indeed, it is conceived as its
opposite. The point should be emphasized as the two are
usually confused in the theoretical literature on ritual. As
a result, theories of ritual handle these two very different
kinds of behavior as if they originated in the same capacities
or processes, which may contribute to the vagueness
of many pronouncements on “ritual” in general. In most
human ceremonies, one finds an alternation or combination
of what we defined as ritualized behavior (high control,
attentional focus, explicit emphasis on proper performance)
and routinized action (possible automaticity, low
attentional demands, lesser emphasis on proper performance).
Differences in attentional focus in particular show
that these are diametrical ways of regulating behavior. Note,
again, that sometimes the same episode that is ritualized behavior
for a participant (e.g., a novice about to be inducted
into a secret society) may eventually become somewhat
routinized for another (e.g., the long-time member who has
witnessed dozens of such ceremonies, or the ritual officer
who has conducted them).
Previous cognitive models of ritual have emphasized
differences that are relevant to this issue of routinization
versus ritualization. The action-representation model of
religious ritual (Lawson and McCauley 1990) predicts a
fundamental difference between high- and low-intensity
rituals. In the model, this is connected with participants’
assumptions about the ways in which superhuman agents
are involved in the ritual, as agents or patients respectively.
Harvey Whitehouse put forward a similar distinction
between what he called distinct “modes of religiosity”: an
imagistic mode that requires rare and salient events versus
a doctrinal or routinized mode that emphasizes repetition
conceptual consistency (Whitehouse 2000, 2004).
In our view, a fundamental difference that may be of
greater impact on cultural transmission is that between the
two behavioral modes described here, ritualization and routinization
respectively, defined in modes of attentional focus
and the division of the action flow. As we emphasized,
the two modes can be found in the same series of ceremonies
or indeed in the same episode as performed by different
participants. So the distinct kinds of perspectives on
the prescribed behavior, with their consequences on conceptual
associations, can be found inside each ritual or religious
tradition. Again, we must leave for further empirical
investigation the question, whether this view explains
cultural transmission more precisely than the distinctions
between kinds of traditions or kinds of rituals put forward
in other models.
COLLECTIVE RITUALS IN HUMAN EVOLUTION
Adaptation or Recruitment of Other Adaptations?
Is there a general human capacity for collective rituals, as
a product of evolution by natural selection? Some features
of ritual seem to cry out for such an interpretation. Some
form of ritual is found in all human groups. Also, we can
infer distinct features of ritualization from the archaeological
record, especially in burial procedures, for the earliest
modern human groups and for Neanderthals (Mithen 1996;
Trinkaus et al. 1993). Moreover, human ritual is sufficiently
different from its phylogenetic cousins in animal behavior
to suggest a specific human adaptation. Seeing collective
ritual in the light of evolution makes even more sense
once we consider its connection to security motivation,
the restricted range of concerns that intrusive and anxious
thoughts focus on, as well as the restricted range of compulsive
responses.
However, we consider the hypothesis of a specific capacity
for rituals fraught with difficulties. To be evolved
through natural selection, a capacity for rituals should be
beneficial to individual reproductive potential. Or, more
specifically, one would need to consider that a higher or
more focused involvement in ritual performance would
result, on average, in better reproductive potential. One
Li´enard and Boyer • Whence Collective Rituals? 825
possible explanation would be that ritual is a good “trick”
that associates individual, unmanageable anxieties with coordinated
action with others and thereby makes them more
tolerable or meaningful. But, again, it seems to us that we
have very little direct evidence for the reproductive benefit
of such a tool. Even though excessive anxiety may be
maladaptive, anxiety in general is not. Indeed, evolutionary
anthropologists have noted that most anxiety states seem
functional rather than dysfunctional in terms of individual
fitness (Nesse 1998, 1999).
Atavistic Procedures or Evolved Precautions?
In our model the rituals themselves are not the direct
expression of an atavistic behavioral repertoire. Washing
incessantly or spending all one’s time checking the environment
for intruders probably are not evolved behaviors.
Rather, the rituals combine, repeat, and accumulate single
behaviors (avoiding contact with certain items, checking
the environment for possible traces of intrusion, washing,
etc.) that are appropriate given evolutionary hazards.
The ritualization process consists in this special accumulation
and distortion of originally appropriate actions.
This is manifest in animal models of the condition. Henry
Szechtman and colleagues observed a high rate of repetitive
and pointless “ritualistic” behavior in rats treated with quinpirole
(a dopamine agonist, potentially acting on some of
the neurotransmitter circuits involved in OCD pathology),
in comparison to controls (Szechtman et al. 1998). The animals
checked and rechecked particular locations and kept
inspecting the same objects. The effect of test treatment
was the higher recurrence, repetition, and redundancy of
actions that were all performed (but with greater flexibility
and no redundancy) by controls. In other words, rats
that are induced to “ritualize” combine in a new way typical
components of the rat’s species-specific precautionary
repertoire.
Capacity or Disposition?
We proposed that collective ritual is not necessarily the
outcome of an adaptive capacity but may well be the predictable
by-product of adaptive capacities. In our model
a collective ritual typically activates the hazard-precaution
system. Given this system and its input format, a pattern of
interaction that activates them may well become attention
demanding and intuitively compelling. This would make
ritual a by-product of evolved cognitive architecture. We
are confident that this is a valid description of many of
the behavioral sequences ordinarily described as “rituals” in
the anthropological literature. However, as we noted above,
particular rituals may or may not comprise ritualized behavior,
characterized as the evocation of evolutionary danger
cues, the construction of appropriate action sequences, and
the association of goal demotion and precaution cues.
In this view, rituals can be considered highly successful
cultural “gadgets” whose recurrence in cultural evolution
is a function of (1) how easily they are comprehended by
witnesses and (2) how deeply they trigger activation of motivation
systems and cognitive processes that are present
in humans for other evolutionary reasons. To say that a
cultural creation is a “gadget” does not entail that it is
unimportant. Many of the most important human cultural
creations—such as literature, complex technology, music,
or the visual arts—can be considered as by-products too. For
us, ritualized behaviors are another by-product of a speciesspecific
human cognitive architecture and therefore an indirect
consequence of its evolution by natural selection.
PIERRE LI´ENARD Department of Psychology, Washington
University in St. Louis, St Louis, MO 63130
PASCAL BOYER Department of Psychology, Washington University
in St. Louis, St Louis, MO 63130
NOTE
Acknowledgments. Pierre Li´enard’s research was supported by a fellowship
grant from the Belgian American Educational Foundation.
Pascal Boyer’s research was supported by a grant from the Henry
Luce Foundation. We are grateful to Leda Cosmides, Dan Fessler,
Thomas Lawson, Robert McCauley, Pascale Michelon, Mayumi
Okada, Tom Oltmanns, Ilkka Pyysi¨ainen, Howard Waldow, Dan
Wegner, Harvey Whitehouse, Jeff Zacks, and four anonymous AA
referees for detailed comments on a previous draft of this article.
1. Here is Freud’s conclusion: “Nach diesen Analogien [. . . ]
k¨onnte man sich getrauen, die Zwangsneurose als pathologisches
Gegenst¨uck zur Religionsbildung aufzufassen, die Neurose
als eine individuelle Religiosit¨at, die Religion als eine universelle
Zwangsneurose zu bezeichnen.” [“Given these analogies [. . . ] it
is possible to consider the obsessive-neurosis as the pathological
counterpart to religious culture, and to see the neurosis as individual
religiosity and religion as collective obsessive-neurosis”; Freud
1948:138.]
REFERENCES CITED
Abelson, Robert P.
1981 Psychological Status of the Script Concept. American Psychologist
36(7):715–729.
Abramowitz, Jonathan S., Stefanie A. Schwartz, Katherine M.
Moore, and Kristi R. Luenzmann
2003 Obsessive-Compulsive Symptoms in Pregnancy and the
Puerperium: A Review of the Literature. Journal of Anxiety Disorders
17(4):461–478.
American Psychiatric Association
1994 Diagnostic and Statistical Manual of Mental Disorders. 4th
edition. Washington, DC: American Psychiatric Association.
Barrett, Justin L., and E. Thomas Lawson
2001 Ritual Intuitions: Cognitive Contributions to Judgments
of Ritual Efficacy. Journal of Cognition and Culture 1(2):183–
201.
Barth, Frederik
1987 Cosmologies in the Making: A Generative Approach to
Cultural Variation in Inner New Guinea. Cambridge: Cambridge
University Press.
Basso, Keith H., and Henry A. Selby
1976 Meaning in Anthropology. 1st edition. Albuquerque: University
of New Mexico Press.
Bloch, Maurice
1974 Symbols, Song, Dance, and Features of Articulation: Is Religion
an Extreme Form of Traditional Authority? European
Journal of Sociology 15:55–81.
1992 Prey into Hunter: The Politics of Religious Experience.
Cambridge: Cambridge University Press.
1998 How We Think They Think: Anthropological Approaches
to Cognition, Memory and Literacy. Boulder, CO: Westview
Press.
826 American Anthropologist • Vol. 108, No. 4 • December 2006
Bloch, Maurice, and Jonathan Parry, eds.
1982 Death and the Regeneration of Life. Cambridge:
Cambridge University Press.
Boyd, Robert, and Peter J. Richerson
1985 Culture and the Evolutionary Process. Chicago: University
of Chicago Press.
Boyer, Pascal
1990 Tradition as Truth and Communication: A Cognitive Description
of Traditional Discourse. Cambridge: Cambridge University
Press.
1994 The Naturalness of Religious Ideas: A Cognitive Theory of
Religion. Berkeley: University of California Press.
1998 Cognitive Tracks of Cultural Inheritance: How Evolved
Intuitive Ontology Governs Cultural Transmission. American
Anthropologist 100(4):876–889.
2001 Religion Explained: Evolutionary Origins of Religious
Thought. New York: Basic Books.
Boyer, Pascal, and Pierre Li´enard
In press Why Ritualized Behavior in Humans? Precaution Systems
and Action-Parsing in Developmental, Pathological and
Cultural Rituals. Behavioral and Brain Sciences.
Bradbury, Jack W., and Sandra L. Vehrencamp
2000 Economic Models of Animal Communication. Animal Behaviour
59(2):259–268.
Bulbulia, Joseph
2004 Religious Costs as Adaptations That Signal Altruistic Intention.
Evolution and Cognition 10(1):19–42.
Douglas, Mary
1982 Natural Symbols: Explorations in Cosmology. New York:
Pantheon Books.
Dulaney, Siri, and Alan P. Fiske
1994 Cultural Rituals and Obsessive-Compulsive Disorder: Is
There a Common Psychological Mechanism? Ethos 22(3):243–
283.
Durham, William H.
1991 Coevolution: Genes, Cultures and Human Diversity. Stanford:
Stanford University Press.
Evans, David W., F. Lee Gray, and James F. Leckman
1999 The Rituals, Fears and Phobias of Young Children: Insights
from Development, Psychopathology and Neurobiology. Child
Psychiatry and Human Development 29(4):261–276.
Fiske, Alan P., and Nick Haslam
1997 Is Obsessive-Compulsive Disorder a Pathology of the
Human Disposition to Perform Socially Meaningful Rituals?
Evidence of Similar Content. Journal of Nervous and Mental
Disease 185(4):211–222.
Frank, Robert
1988 Passions within Reason: The Strategic Role of the Emotions.
New York: Norton.
Freud, Sigmund
1948[1906] Zwangsbehandlungen und religions¨ubungen (Obsessive
behavior and religious practices). In Gesammelte Werke
von Sigmund Freud, chronologisch geordnet [Collected works
of Sigmund Freud in chronological order], vol. 7. Anna Freud
et al., eds. Pp. 129–139. London: Imago Publishing.
1928 Die Zukunft einer Illusion [The future of an illusion]. 2nd
edition. Leipzig: Internationaler Psychoanalytischer Verlag.
Gintis, Herbert, Eric A. Smith, and Samuel Bowles
2001 Costly Signaling and Cooperation. Journal of Theoretical
Biology 213(1):103–119.
Gluckman, Max
1975 Specificity of Social-Anthropological Studies of Ritual.
Mental Health and Society 2(1–2):1–17.
Goody, Jack
1972 The Myth of the Bagre. Oxford: Clarendon Press.
1986 The Logic of Writing and the Organization of Society.
Cambridge: Cambridge University Press.
Harcourt, Alexander H., and Frans. B. de Waal, eds.
1992 Coalitions and Alliances in Humans and Other Animals.
Oxford: Oxford University Press.
Hodgson, Ray J., and Stanley Rachman
1972 The Effects of Contamination and Washing in Obsessional
Patients. Behaviour Research and Therapy 10(2):111–117.
1977 Obsessional-Compulsive Complaints. Behaviour Research
and Therapy 15(5):389–395.
Houseman, Michael
2002 Dissimulation and Simulation as Forms of Religious Reflexivity.
Social Anthropology 10(1):77–89.
Houseman, Michael, and Carlo Severi
1998 Naven and Ritual. Leiden, the Netherlands: Brill.
Humphrey, Caroline, and James Laidlaw
1993 Archetypal Actions: A Theory of Ritual as a Mode of Action
and the Case of the Jain Puja. Oxford: Clarendon Press.
Irons, William
2001 Religion as a Hard-to-Fake Sign of Commitment. In Evolution
and the Capacity for Commitment. R. Nesse, ed. Pp.
292–309. New York: Russell Sage Foundation.
Ketelaar, Timothy, and Bruce J. Ellis
2000 Are Evolutionary Explanations Unfalsifiable? Evolutionary
Psychology and the Lakatosian Philosophy of Science. Psychological
Inquiry 11(1):1–21.
Kuran, Timur
1998 Ethnic Norms and Their Transformation through Reputational
Cascades. Journal of Legal Studies 27(2):623–659.
Kurzban, Robert, John Tooby, and Leda Cosmides
2001 Can Race Be Erased? Coalitional Computation and Social
Categorization. Proceedings of the National Academy of Sciences
of the United States of America 98(26):15387–15392.
Lassiter, G. Daniel, Andrew L. Geers, and Kevin J. Apple
2002 Communication Set and the Perception of Ongoing Behavior.
Personality and Social Psychology Bulletin 28(2):158–
171.
Lawson, E. Thomas, and Robert N. McCauley
1990 Rethinking Religion: Connecting Cognition and Culture.
Cambridge: Cambridge University Press.
Leckman, James F., Linda C. Mayes, Ruth Feldman, David W. Evans,
Robert A. King, and Donald J. Cohen
1999 Early Parental Preoccupations and Behaviors and Their
Possible Relationship to the Symptoms of ObsessiveCompulsive
Disorder. Acta Psychiatrica Scandinavia Supplementa
100(Suppl. 396):1–26.
LeDoux, Joseph
2003 The Emotional Brain, Fear, and the Amygdala. Cellular and
Molecular Neurobiology 23(4–5):727–738.
Leonard, Henrietta L., Erica L. Goldberger, Judith L. Rapoport, Deborah
L. Cheslow, and Susan E. Swedo
1990 Childhood Rituals: Normal Development or ObsessiveCompulsive
Symptoms? Journal of the American Academy of
Child and Adolescent Psychiatry 29(1):17–23.
Li´enard, Pierre
In press The Making of Peculiar Artifact: Living Kind, Artifact
and Social Order in the Turkana Sacrifice. Journal of Cognition
and Culture 6(3–4).
Mataix-Cols, David, Maria Conceic¸ao do Rosario-Campos, and
James F. Leckman
2005 A Multidimensional Model of Obsessive-Compulsive Disorder.
The American Journal of Psychiatry 162(2):228–238.
Mathews, Carol A., Kerry L. Jang, Shadha Hami, and Murray B. Stein
2004 The Structure of Obsessionality among Young Adults. Depression
and Anxiety 20(2):77–85.
McCauley, Robert N., and E. Thomas Lawson
2002 Bringing Ritual to Mind: Psychological Foundations of
Cultural Forms. Cambridge: Cambridge University Press.
Metcalf, Peter, and Richard Huntington
1991 Celebrations of Death: The Anthropology of Mortuary Ritual.
2nd edition. Cambridge: Cambridge University Press.
Mithen, Steven J.
1996 The Prehistory of the Mind. London: Thames and Hudson.
Needham, Rodney
1975 Polythetic Classification: Convergence and Consequences.
Man (n.s.) 10(3):349–369.
Nesse, Randolph M.
1998 Emotional Disorders in Evolutionary Perspective. British
Journal of Medical Psychology 71(4):397–415.
1999 Proximate and Evolutionary Studies of Anxiety, Stress and
Depression: Synergy at the Interface. Neuroscience and Biobehavioral
Reviews 23(7):895–903.
Newtson, Darren
1973 Attribution and the Unit of Perception of Ongoing Behavior.
Journal of Personality and Social Psychology 28(1):28–38.
Li´enard and Boyer • Whence Collective Rituals? 827
Payne, Robert J. H.
1998 Gradually Escalating Fights and Displays: The Cumulative
Assessment Model. Animal Behaviour 56(3):651–662.
Rachman, Stanley, and Padmal de Silva
1978 Abnormal and Normal Obsessions. Behavior Research and
Therapy 16(4):233–248.
Radomsky, Adam S., Stanley Rachman, and David Hammond
2001 Memory Bias, Confidence and Responsibility in Compulsive
Checking. Behavior Research and Therapy 39(7):813–822.
Rappaport, Roy A.
1979 Ecology, Meaning and Religion. Berkeley, CA: North
Atlantic Books.
1999 Ritual and Religion in the Making of Humanity.
Cambridge: Cambridge University Press.
Salkovskis, Paul M.
1985 Obsessional-Compulsive Problems: A CognitiveBehavioural
Analysis. Behavior Research and Therapy
23(5):571–583.
Saxena, Sanjaya, Arthur L. Brody, Jeffrey M. Schwartz, and Lewis
R. Baxter
1998 Neuroimaging and Frontal-Subcortical Circuitry in
Obsessive-Compulsive Disorder. British Journal of Psychiatry
173(Suppl. 35):26–37.
Schelling, Thomas
1960 The Strategy of Conflict. Cambridge, MA: Harvard University
Press.
Sosis, Richard
2000 Religion and Intragroup Cooperation: Preliminary Results
of a Comparative Analysis of Utopian Communities. CrossCultural
Research: The Journal of Comparative Social Science
34(1):70–87.
2003 Why Aren’t We All Hutterites? Costly Signaling Theory and
Religious Behavior. Human Nature 14(2):91–127.
Sperber, Dan
1975 Rethinking Symbolism. Cambridge: Cambridge University
Press.
1985 Anthropology and Psychology: Towards an Epidemiology
of Representations. Man (n.s.) 20(1):73–89.
1996 Explaining Culture: A Naturalistic Approach. Oxford:
Blackwell.
2000 An Objection against Memetics. In Darwinizing Culture:
The Status of Memetics as a Science. R. Aunger, ed. Pp. 163–
173. London: Oxford University Press.
Staal, Frits
1990 Rules without Meaning: Ritual, Mantras, and the Human
Sciences. New York: P. Lang.
Szechtman, Henry, William Sulis, and David Eilam
1998 Quinpirole Induces Compulsive Checking Behavior in
Rats: A Potential Animal Model of Obsessive-Compulsive Disorder
(OCD). Behavioral Neuroscience 112(6):1475–1485.
Szechtman, Henry, and Erik Woody
2004 Obsessive-Compulsive Disorder as a Disturbance of
Security Motivation. Psychological Review 111(1):111–
127.
Tooby, John, and Leda Cosmides
1989 Adaptation versus Phylogeny: The Role of Animal Psychology
in the Study of Human Behavior. International Journal of
Comparative Psychology 2(3):175–188.
Trinkaus, Erik, Pat Shipman, and History E-Book Project
1993 The Neandertals Changing the Image of Mankind. Electronic
document, http://name.umdl.umich.edu/HEB02320,
accessed June 1, 2006.
Ursu, Stefan, V. Andrew Stenger, M. Katherine Shear, Mark R. Jones,
and Cameron S. Carter
2003 Overactive Action Monitoring in Obsessive-Compulsive
Disorder: Evidence from Functional Magnetic Resonance Imaging.
Psychological Science 14(4):347–353.
Watanabe, John M., and Barbara B. Smuts
1999 Explaining Religion without Explaining It Away: Trust,
Truth, and the Evolution of Cooperation in Roy A. Rappaport’s
“The Obvious Aspects of Ritual.” American Anthropologist
101(1):98–112.
Wegner, Daniel M.
1994 Ironic Processes of Mental Control. Psychological Review
101(1):34–52.
Wegner, Daniel M., and J. A. K. Erskine
2003 Voluntary Involuntariness: Thought Suppression and the
Regulation of the Experience of Will. Consciousness and Cognition:
An International Journal 12(4):684–694.
Wegner, Daniel M., and D. J. Schneider
2003 The White Bear Story. Psychological Inquiry 14(3–4):326–
329.
Wegner, Daniel M., David J. Schneider, Samuel R. Carter, and Teri
L. White
1987 Paradoxical Effects of Thought Suppression. Journal of Personality
and Social Psychology 53(1):5–13.
Whitehouse, Harvey
2000 Arguments and Icons: Divergent Modes of Religiosity.
Oxford: Oxford University Press.
2004 Modes of Religiosity. Walnut Creek, CA: AltaMira Press.
Zacks, Jeffrey M., Barbara Tversky, and Gowri Iyer
2001 Perceiving, Remembering, and Communicating Structure
in Events. Journal of Experimental Psychology: General
130(1):29–58.
Zahavi, Amotz, and Avishag Zahavi
1997 The Handicap Principle: A Missing Piece of Darwin’s Puzzle.
New York: Oxford University Press.
Zalla, Tiziana, Isabelle Verlut, Nicholas Franck, Didier Puzenat, and
Angela Sirigu
2004 Perception of Dynamic Action in Patients with Schizophrenia.
Psychiatry Research 128(1):39–51.
Zohar, Ada H., and Levia Felz
2001 Ritualistic Behavior in Young Children. Journal of Abnormal
Child Psychology 29(2):121–128.