Spontaneous (minimal) ritual in non-human great apes?

Abstract

The potential for rituals in non-human great apes (apes) is an understudied topic. We derive a minimal definition of ritual and then examine the currently available evidence for it in untrained and non-enculturated apes. First, we examine whether such apes show evidence for the two main components of our minimal definition of ritual: symbolism and copying. Second, we examine if there are actual cases already identifiable today that may fit all aspects of our minimal definition of ritual—or whether there are at least cases that fit some aspects (proto-ritual). We find that apes are not likely to spontaneously practise minimal ritual, but we claim that the highest expected likelihood of occurrence is in the results-copying domain. Yet, we did not find actual cases of minimal ritual in apes—including those involving environmental results. We did, however, find some cases that may match at least part of our minimal ritual definition—which we termed proto-ritual. At least two out of three potential cases of such proto-rituals that we identified (rain dance, object-in-ear and surplus nest-making procedures) do revolve around results. Overall, apes do not show much, or very clear, evidence for even minimal ritual, but may sometimes show proto-ritual. However, dedicated ape ritual studies are currently lacking, and future work may identify ape ritual (or clearer cases of proto-ritual). We discuss the implications of our preliminary finding for inferences of ritual in the last common ancestor of humans and apes.

This article is part of the theme issue ‘Ritual renaissance: new insights into the most human of behaviours’.

1. Introduction

How human culture originated is an unresolved question. One approach is to study extant primates as models for early hominin behaviour. Non-human great apes (henceforth, apes) are especially relevant as they can inform us of the potential abilities of the last common ancestor of humans and the genus Pan (henceforth, LCA). Do apes show ritual?

There are rare, but persistent, claims for ape rituals¹—mostly in public media but also, more infrequently, in the academic literature (see below). However, the question remains as to whether any of these claims are valid. The first step in examining this issue is to define the term ‘ritual’, in order to have a template against which we can measure ape behaviour. In humans,² rituals are said to largely consist of a ritualized action, which is ‘any action that is typically causally opaque and goal demoted as a consequence of the repetition, redundancy, formality, and stereotypy of the motor features, and where accurate performance of process is prioritized over outcome’ [2, p. 172]. Additionally, these actions are said to be performed ‘for social reasons’ [2, p. 173].

We note that there is one large piece potentially missing from the human ritual definition above. It is possible that rituals cannot exist without language, if language is required to transmit the general concept of a ritual. Or, more fundamentally, language might be required because the goal of the ritual is, in essence, symbolic and thus arbitrarily linked to the ritual actions (e.g. to pray for happiness only when the sun rises; symbol here defined as ‘point[ing] arbitrarily to their referents’ [3, p. 206]). If full-blown language is necessary for rituals to be maintained, there might be little hope to find spontaneous rituals in apes (and we cannot ask the apes, either). However, single human ritual performances themselves can be language-free (imagine a solemn, silent ceremony). Therefore, while language is (at least often) a key component, it may or may not be a necessary component of all rituals (we remain somewhat agnostic in this respect). In our preliminary search for ape ritual behaviour, we will therefore intentionally ignore the linguistic component, thus providing apes with a potential to show rituals.

We believe that the definition of human ritual described above is a good starting point applicable to apes. A minimal variant of the definition is necessary because it is possible that some of the features of human rituals are idiosyncratic to our species and that they do not necessarily need to be present in order to constitute ritual in other species (including early hominins and the LCA). In addition, in a first search for ape rituals, we should look primarily for general ability and for every potential candidate, and here, it is useful not to be too restrictive in terms of definitions. We therefore believe (and hope to make a convincing case) that the ritual definition can and should be both slightly relaxed and somewhat condensed, and thus derive an ape/animal variant. We call this the ‘minimal ritual’ definition. Next, we will examine if at least any of the remaining core elements of the minimal ritual definition can be found in apes. Finally, we will assess whether apes show any convincing potential cases of spontaneous rituals (fulfilling all our minimal ritual criteria) or proto-rituals (at least potentially fulfilling at least parts of our minimal definition).

2. The minimal ‘ape’ definition of ritual

We will keep several aspects of the current human ritual definition largely unchanged—though we shall distill these arguments to their essence.

(a). Even minimal rituals are and must be symbolic

Although the human definition of ritual above does not specifically use the term ‘symbolic’, we believe that symbolism can be distilled from it, in the following way. According to the human definition above, rituals must be perceived in an extraordinary way and must be interpreted from a ‘social or normative point of view’ [2, p. 173].³ Ritual actions and/or results must therefore gain ‘social significance’ [2] and do so through the ritual itself. The significance created must therefore also extend beyond the actor (it must be socially shared) and must exist at least in part because it is so perceived by others (though see below for a case we call private ritual, which would be a conceivable variant of ritual that would lack this aspect). That is, rituals should be social (or even normative).

Part of what makes human rituals ‘extraordinary’ is their causal opacity [2,4,5]. For example, if certain actions bring about a causal effect that itself may be the goal of the action (i.e. when the actions are instrumental), then they do not qualify as ritual [6]. The actions (and/or results) must therefore be arbitrarily linked to the goal. Human rituals must also have what might be called goal opacity (usually called ‘goal demotion’; e.g. [2,7,8]), meaning that the goal of the ritual must also not be directly deducible from the actions or results themselves (compare [4]).

Consequently, we have to have causally and goal-opaque actions and/or results (excluding ordinary cases that lead in a direct way straight to the goal), and these also have to create social significance. This means we would have to (i) exclude any direct relationship (and any naturally/sexually evolved or by-product relationship, see below) between behaviour and outcome. Furthermore, (ii) this would only leave indirect, opaque relationships—arbitrary relationships—which also have to create social significance. We interpret all this as a need for the ritual to be symbolic. Therefore, we conclude that one or more of the actions or object constellations or object changes⁴ involved in a ritual must gain a symbolic feature—as an effect of the ritual (during or after the ritual). The alternative view that the ritual merely creates or fosters social meaning (directly or indirectly) seems unfeasible, because, if that were the case, all communication would be ritualistic. And so, what would need to be demonstrated in apes is that the ritual creates some symbolic feature.⁵ Rituals must be symbolic.

(b). Naturally and/or sexually selected actions (and outcomes) and their by-products are not symbolic and therefore not rituals

As an example of ordinary behaviour (here, as an adaption in the context of sexual selection), in the courtship display of crested grebes, the grebes engage in complex actions (they can even seem mirrored) and also offer nesting materials (results). Are these actions and results symbolic? We would argue they are not, because if we were to accept courtship displays as rituals, then all displays involving ‘ritualized’ (and any evolved) signals in nature would be rituals. First, most of these cases are presumably not intentional, in the sense that the signaller intends to communicate to the recipient, as shown by a variety of criteria (e.g. [9]). Second, such actions may be index or handicap signals (signals that reflect quality or ability) or meant to signal conspecificity (species matching) etc. All this makes the signals non-arbitrary, since their form must reflect specific qualities. Thus, arguably, naturally and or sexually selected actions and their results should not be equated with rituals. This is not to say that rituals have no function (they may, and indeed often do). However, it is to say that biologically evolved, functional forms of actions and/or results are not rituals.

(c). Rituals are and must be copied

Rituals must show stereotypical, accurately executed actions⁶ [2]. Together with the requirement that rituals cannot be evolved traits (or their by-products), this strongly implies that these actions and/or resulting object changes must be culturally transmitted. In other words, they must have been copied, using copying social learning mechanisms. More precisely, rituals must therefore require copying. Actions or results must thus not merely be cultural, they must have been copied, in order to (potentially) be rituals. This difference is important, because copying is not a necessary requirement for culture. Instead, any social learning mechanism (and there are many) can lead to culture (thus, culture is indeed very widespread in the animal kingdom, including in apes: [10]). But most social learning mechanisms are non-copying social learning mechanisms. They do not produce copies—instead, they lead to ‘socially mediated reinnovation’ [11]. That is, most social learning variants are merely modulating (and stabilizing) the frequencies of the expressions of individually derived behavioural forms [12,13]—but, again, this then counts as culture in each case [10]. But, to even potentially count as ritual, copying must be involved. For a possible exception, where ritual might not be copied, see ‘private ritual’ below.

(d). Changes in the human definition of ritual

While human rituals do not ‘typically’ [2] cause physical changes, other species' rituals might (if they exist). Furthermore, given that ape social learning is already known to be different from that of humans in that it is results focused (including paying attention to and matching object changes/movements: [14]), for apes (and other animals in general), we will allow for the possible presence of results-based rituals in addition to action-based rituals. And so, in apes we will look for potential ritual status of pure actions, as well as potentially ritualized actions towards and with objects. Additionally, we will consider results entirely on their own (i.e. object constellations and object changes regardless of the actions that brought them about). That the focus should be on results will likely be welcome news for a largely results-based science such as prehistoric archaeology.

Human rituals are said to be repeated and redundant [2]. Here, we may differentiate between repetition within a given ritual event and repetition across ritual events. It is not clear a priori whether repetition within a ritual event is a necessity for ritual. For instance, the blowing out of candles may be the only ritualistic action (with associated results) at a minimalistic birthday event—yet it only happens once for each instance of these events. We will therefore also leave out this criterion for apes (but will note when/if it occurs). To fulfil the ape ritual definition, for now, a single event is the minimum requirement for our definition of ritual. As reflected in the ‘maximum’ definition of human ritual above, the criterion of redundant repetition within a given ritual event may itself constitute, or add to, the causal opacity (why repeat an action once its result is achieved?), and therefore we will not specifically mention it in our minimal definition.

Repetition across ritual events may be caused by a link between the ritual and events/occasions/domains, such as someone's death or the onset of the rainy season. We agree that finding such a correlation can, in practice, help to detect ritual. Imagine an observer of another group who does not speak their language; she could recognize rituals by this kind of repetition and the link to particular conditions (even though, logically, this criterion may not be absolutely required and so we will leave it out of our definition). Even so, we will make use of repetitions below for practical reasons.

The minimal ‘ape definition of ritual’ can now be stated as follows:

Rituals consists of socially shared, symbolic feature(s) that are created via actions and/or results that require copying.

It might be argued that our minimal ritual definition has become too broad, and that it might capture too many things that are not ritual. This is a valid theoretical concern, but as we will show below, empirically (at least in the case of apes), this concern is—at least currently—unfounded: only few potential cases can currently be detected even when using such a broad definition. And arguably, regardless of how many cases may be uncovered, in a first sweep for ape ritual, it is best to use a broad definition so as not to produce false negatives early on. Indeed, given also that we are unaware of even a single study specifically aimed at examining the existence of ape ritual (which we encourage all readers to perform), a minimal definition like ours is arguably the kind of liberal approach that is most useful to start any investigation into a new phenomenon. Once ape ritual is found using this minimal definition, further distinctions may then be introduced, but until then, we surmise that a minimal definition serves its purpose, if only to guide the research itself. Of course, our minimal definition may instead also still prove too demanding (but, in fact, we will relax our definition even more below, introducing what we call proto-rituals).

3. Do apes have the necessary abilities for spontaneous ritual?

The goal of our paper is to examine whether apes have ritual—as a phylogenetic proxy for the LCA. Towards this goal, we must ask whether apes have spontaneous rituals. We must, in other words, exclude ritual passed on to apes by humans, be it directly or indirectly (as humans were also not present to pass these skills to the LCA). Therefore, only data from non-enculturated, untrained apes will be considered here (compare [15]). When we speak of apes below, we therefore exclusively report evidence from non-enculturated, untrained apes (these will be simply referred to as ‘apes’ below).

Before we will look at specific potential cases of ape rituals, we will first—more generally—look for evidence that apes possess the necessary basic abilities to perform the constituent parts of the minimal ritual definition. We will start by looking for the evidence of symbolism, followed by evidence for copying.

(a). Symbolism in apes?

Human life is drenched in symbols to such an extent that Deacon [16] called humans the ‘symbolic species’. Although apes can communicate about absent entities [17], they do not seem to produce nor understand iconic gestures (i.e. gestures that resemble their referents, such as the gestures associated with the game rock–paper–scissors), as reviewed in [18]. The same is true for iconic objects, i.e. objects that resemble the original object:⁷ iconic objects are neither understood, nor produced, by apes (e.g. [19]). This is particularly informative because iconic gestures and iconic objects are at least ‘reality-resembling’ and thus should arguably be easier to understand or produce than conventional symbols (which have instead arbitrary relationships, such as the word ‘symbol’ to represent symbol). Thus, iconic gestures or objects might be best referred to as proto-symbols and, again, they are not understood (or produced) by apes. This conclusion also seems to hold for ape communication in the wild, despite a few as-yet unclear (but rare) cases [20].

In the case of conventional symbols (arbitrarily connected to their meanings), it is generally understood that ape usage and production is not spontaneous, and that it requires very long human training and/or enculturation. Overall, then, the seeming lack of spontaneous pre-symbolic and symbolic skills in both comprehension and production make it unlikely that apes can spontaneously have ritual.

(b). Copying in apes?

It has been established in recent years that many, or possibly all, ape communication forms do not require copying [21], and similar conclusions were drawn for ape tool-use and subsistence forms [12,13]. True, when experimentally tested, apes do spontaneously match what they see demonstrated in two-action, two-result experiments (e.g. [22]). But, firstly, this match is likely based on their registration of results, not of actions ([13]; compare data in [14]). Secondly, apes' matching abilities most likely do not move beyond skills that are part of their natural innovative repertoire: these abilities do not move beyond the individual level (also called ‘latent solutions’; [12,13]). Supporting this conclusion, all demonstrated options (targets) of past experiments also occur spontaneously (individually) outside demonstrations, showing that these tasks did not require copying (reviewed in [13]). When experiments make copying a requirement to succeed, apes do not copy actions [23–26]⁸ or even results [12]. Overall then, apes do not seem to copy when copying is required.

In theory, specific variants of teaching (namely those that produce copies) could also lead to copies of form. However, there is very little evidence for any variant of teaching in apes (the most complete case being that in Musgrave et al. [27], and even here no copies were produced). Importantly, apes do not actively make each other copy actions (action-based teaching), and they also do not mould others' actions (teaching by moulding). They do not teach by punishment for wrong actions either (what we may call ‘teaching by honing’).

The apparent lack (or extreme rarity) of action-copying skills (both passive and active: the latter would be variants of teaching) likely means that apes cannot produce or maintain action-based rituals. The picture looks a little brighter when it comes to results copying. While, again, apes do not seem to match result demonstrations beyond what they could do on their own, at least here they frequently do match such results (reviewed in [13]). Consequently, if apes have rituals at all, they are perhaps more likely to have rituals when these revolve around results rather than actions.⁹

4. Selection of potential examples of ape rituals

The above analysis of the potential presence or absence of the constituent core elements of our ape ritual definition in apes has led to the conclusion that apes are rather unlikely to spontaneously have even minimal ritual. This is because they do not spontaneously produce symbols (or even proto-symbols), and they do not show copying when it is required. The most likely possible exception is results-based ritual. In our search for examples, we looked especially for cases whose form (≈design) may appear copied and/or where the behaviour or result may appear to be symbolic (e.g. by seeming extraordinary). These foci directly follow from the two main components of ritual (copying and symbolism) examined above. Nevertheless, in line with the experimental evidence we highlight above, the vast majority of cases turn out to be best explained without reference to copying and/or symbolism. That is, the vast majority of cases do not fulfil even the minimal definition of ritual.

Generally, we must note that the topic of ape ritual is not often academically discussed. A Google scholar search of ‘ape ritual’ on 2 September 2019 produced a meagre total of three results (none of them actually about apes). Likewise, searching for ‘apes ritual’ on the same day produced zero results.¹⁰ The list of potential cases below was therefore primarily created by the authors, in conjunction with discussions with other primate researchers. Note also that, where other authors might have made claims for ape rituals, they used ape ritual definitions different from the one we apply here.

(a). Cases that do not require copying and also do not require symbolism

Culturally unconnected populations effectively represent independent cultural paths. A presence of copying would lead to path-dependent forms, whose effects cumulate over time. This is unavoidable, given that every variant of copying comes with copy error [28]. This implies that, if any ape behaviour is really based on copying, it will by now necessarily have to show population differences in form (at least in a subtle way). And given that we established that even minimal ritual must be based on copying, ape behaviour that is ritual according to this minimal definition must differ in form between populations. Or to put if differently, ape behaviour that is similar across culturally unconnected populations (distant populations) can almost certainly be dismissed as not fulfilling the requirement of copying.

As for symbolism, we can pose that this is most likely to be found in cases for which there is no good, non-opaque, ordinary explanation yet. In the last 50 or so years, ape behaviour has been intensely studied. This effort uncovered ordinary explanations for many behaviour patterns, many of which initially seemed extraordinary (see below). While we are nowhere near a complete understanding, on the whole, most of the observed ape behaviour is therefore no longer mysterious—and, with it, no longer potentially symbolic. However, a few cases remain that currently are not well explained.

Our first example is ‘accumulative stone throwing’, where ‘chimpanzees habitually bang and throw rocks against trees, or toss them into tree cavities, resulting in conspicuous stone accumulations at these sites' [29, p. 2]. Public media declared this behaviour ritualistic (e.g. [30]). However, first, the behaviour was observed—in similar form—across several distant (i.e. likely culturally unconnected) populations of chimpanzees [29]—hence is unlikely to require copying. Second, while the behaviour at first seemed immune to ordinary explanations (though [29] actually listed such possibilities), an ordinary explanation has since been found. The trees that are targeted by the throwing behaviour have favourable acoustic properties for display purposes and are specifically selected for stone throwing because of this [31]. Overall, then, accumulative stone throwing in chimpanzees is likely an ordinary behaviour (functionally transparent and favoured by natural and/or sexual selection), which in addition does not require copying. It therefore fulfils none of the major components of our minimal ritual definition and is therefore not a ritual.

Another claim for ape ritual made in public media is mourning and other death-related behaviours (subsumed under the umbrella term thanatological behaviour; see [32]). Occasionally, this claim also appears in academic writing [33]. Yet, not only do the vast majority of the described behaviours again appear across many culturally unconnected populations, they even re-appear in different species of apes (and also other primates; compare [32]). That is, these behaviours very likely do not require copying. In addition, and perhaps surprising at first, thanatalogical behaviours are likely ordinary. Roughly speaking, on the level of both species and individuals, they correlate with behavioural and ecological constraints and with social investment. For example, everything else being equal, there is more/longer thanatological behaviour from a close relative of the deceased than from a more distantly related individual. The same is true for investment: a close ally will receive more/longer thanatological attention than others, perhaps serving to guard against a false conclusion of death (see [32] for an excellent recent review of the topic). A possible exception to this general pattern is one population of chimpanzees that sometimes threw twigs at dead conspecifics. However, the same chimpanzees also generally showed such behaviour towards other dead animals—plus, several different primate species also throw objects at other species [32]. This behaviour too, therefore, does not require copying. Yet, it is noteworthy, as such covering in thrown objects might constitute a potentially alternative explanation for some hominin ‘burials’ that might be found.

Although the literature on animal courtship often speaks of ‘courtship rituals’, this does not make them rituals according to our minimal definition. Sometimes, however, the claim is more clearly made that certain courtship behaviour is copied. Perry [34] described potential candidates for socially learned ‘communicative rituals' in chimpanzees, among them leaf-clipping behaviour. Arguably, this behaviour (and its associated meanings) remains somewhat understudied, so the situation would benefit from more data [35]. Still, what can already be said is that the actions (and the results: clipped leaves with associated sounds) seem to re-appear across unconnected populations. At the majority of chimpanzee sites, leaf-clipping is performed with the mouth—but there is also a version where chimpanzees do leaf-clipping by hand [36]. However, even the latter version is being performed in several culturally unconnected places—namely at Tai, Kibale and Budongo (and sometimes also at Mahale's M-group; [36]). Given these repetitions, the various forms of leaf-clipping are unlikely to require copying (a conclusion we share with [37]). Importantly, the various described meanings of leaf-clipping also re-appear across culturally unconnected populations. Thus, chimpanzees in Bossou (Guinea) produce leaf-clipping in association with a variety of meanings seen elsewhere for leaf-clipping, and this also includes the meaning of courtship (in total, the meanings in Bossou include frustration, copulation (courtship) and play: [38], as cited in [35]). Most importantly, all these meanings associated with leaf-clipping are ordinary. In summary, leaf-clipping neither requires copying nor requires the assumption of extraordinary, symbolic meanings. For the time being, we therefore conclude that leaf-clipping should not count as ape ritual.

There are many more ape behaviours that may at first sight appear symbolic and/or copied (we are unaware of published claims for ritual in these cases but have encountered such claims verbally). But, first, they also re-appear across culturally unconnected populations (i.e. are likely uncopied and/or are ordinary). In addition, second, there are well-described ordinary explanations for these behaviours—they are therefore also not symbolic. For example, this list includes: Machiavellian behaviour (e.g. [39]); grooming styles (reviewed in [12]; for claims of ritual instead, see [34]); using sticks as play objects [40]; hunting (e.g. [41]); patrolling (e.g. [42]); killing (e.g. [43]); travel direction communication [44]; and kiss-squeaking sounds with leaves (which make the emitting individual appear bigger: [45]).

(b). Cases that do not require copying but that may be symbolic (proto-rituals)

Next, we will turn to cases whose forms still do not require copying (as they appear across unconnected populations) but that, at least currently, escape an ordinary explanation. These cases may therefore be symbolic, and to mark their special (and rare) status (no copying necessary, yet perhaps symbolic), we will label these proto-ritual.¹¹ Note that we would have labelled cases that are not symbolic, but that seem to require copying as proto-rituals, but we did not find any.

One proto-ritual is object-in-ear behaviour, best known from a study in which chimpanzees placed grass blades into their ears [46]. The behaviour currently escapes ordinary explanation (and thus may be symbolic). But—again—the behaviour also appears elsewhere and thus does not depend on copying (captive chimpanzees at other institutions have also been observed to insert elongated, thin objects into their ears; namely at Bastrop, Texas, L. Hopper, L. Reamer 2018, personal communication; and at Leintal zoo, Germany, A. M. Rodrigo 2019, personal communication).

Another proto-ritual is the chimpanzee rain dance¹² [47], i.e. vigorously and noisily charging¹³ as a reaction to rain (or waterfalls), but without clear chimpanzee targets. Copying is probably not necessary for this behaviour as all variants of chimpanzee rain dance (re-)appear in culturally unconnected populations of chimpanzees (compare data in [48]), suggesting that it does not qualify as ritual. Note also that it is present in a similar form in bonobos [49]. Rain dance ‘appear[s] to incorporate aspects of universal chimpanzee behaviour such as aggressive displays’ [48, p. 1511], and so, the behavioural forms of rain dance do not require copying. Still, the behaviour is currently resisting an ordinary explanation and might therefore create symbolic meaning, which is why it is listed here as proto-ritual.

The last proto-ritual is surplus nest-making procedures in orang-utans. When orang-utans in Suaq (Sumatra) and Tuanan (Borneo), but not at most other sites, build their evening nest (which involves making pillows with twigs and leaves), they seemingly always pick small leafy twigs, which are arranged in parallel, to make the pillow. During this phase of nest building, they may bite or smooth the ends of these twigs with their teeth, but often they may simply pass the twigs by their mouth while merely making a biting movement (air-biting). They then place them on the nest with a crisp flick of the wrist. At that very moment (but also at other moments in this same context), orang-utans at Suaq in Sumatra [50,51], and at Sabangau in Borneo [51], produce a raspberry sound (nest raspberry). Yet, at Tuanan, Borneo [50,51], they make instead a smacking sound (nest smacks). Given that the action of passing the twig as well as air-biting is found in at least two culturally unconnected populations (see above; and also in captivity; [50]), it is unlikely to require copying. The same is true for both the raspberry sound and the smacking sound: the raspberry sound appears at least in the two populations described above, and the smacking sound appears also in rehabilitated orang-utans [52] and it is also heard at Suaq—when orang-utans there chew on liana stems—although in neither case has there been a formal analysis to compare the acoustic structures. And so, while these sounds almost certainly do not require copying, the functional significance of the elements is (currently) obscure: the air-biting seems extraordinary because the twig is no longer attached to the branch, and especially the sounds likewise serve no function in pillow formation per se. And so, orang-utan surplus nest-making procedures are currently resisting an ordinary explanation and might therefore create symbolic meaning.

Yet, these orang-utan nest sounds may also be too soft to serve as a communicative function to others, unless these others are nearby (but then they can actually see that the subject is finishing a nest).¹⁴ At two well-studied sites, all individuals studied virtually always perform these sounds when they build their big evening nest, regardless of whether they are in association with infants or others [50]. The fact that these behaviours seemingly are performed primarily individualistically, i.e. without a social goal, fails to support the view that the sound is symbolic (unless we would allow for private symbols; more on this below).

5. Discussion

Even when a specifically derived, stripped-down, minimal (ape) ritual definition is applied, there is as yet no clear, unequivocal, full-blown evidence for any spontaneous minimal ritual in non-enculturated, untrained apes. Furthermore, even the single necessary core components of the minimal ritual definition, namely symbolism and copying, seem largely absent in apes (with the possible exception of some basic, but frequent, result-matching). We found a few cases (orang-utan surplus nest-making procedures, chimpanzee and bonobo rain dance, chimpanzee object-in-ear behaviour) that, although not requiring copying, at least currently appear extraordinary. These cases may therefore be symbolic—and would then fulfil at least part of our minimal definition—which is why we have called these cases proto-rituals. However, given that also these cases do not require copying, they cannot be considered full-blown rituals under the minimal definition of ritual (even if they were to be found to really be symbolic). In turn, our finding at the moment supports the view that the LCA likely did not have full-blown (minimal) ritual either (though the LCA might have had proto-ritual). However, dedicated studies of ape ritual are still outstanding, and our conclusion must remain tentative.

The three potential cases of proto-ritual we identified also warrant further examination in future. On the one hand, ordinary, functional explanations may be found after all for these cases. If so, each case would cease to be a proto-ritual. On the other hand, they may sometimes be found to require some copying after all (e.g. in specific sound structures (currently unexamined) of the various nest-making sounds of the orang-utans across populations and contexts). However, again, with the data currently available, we found no evidence for copying. This should not be surprising as the vast majority of ape behavioural form seems best explained by a natural (ordinary) behavioural repertoire potential (compare [12,13]). To be clear, this is not to say that social learning plays no role in apes. Social learning very likely even plays a very large role for explaining the relative frequencies of behaviour within and across groups (see also Introduction). That is, we do not deny ape cultural status—they have culture, though their culture is likely based on non-copying social learning mechanisms (see Introduction).

Even though apes do not spontaneously copy novel actions, it should be noted that they at least seem to realize when their actions are being copied [53], and there is some evidence that other primates show the so-called chameleon effect (where those that imitated subjects seem to be liked more; e.g. [54]). This could mean that one of the many possible effects or functions of rituals (bonding) may lie dormant in apes (and may sometimes be triggered even without copying, e.g. in synchronized walking during chimpanzee border patrols; compare [55]). If, additionally, apes understand these actions as ritualistic (which is of course hard to say), then they may be on the brink of also understanding human rituals when these are directed towards them.

Our largely negative finding regarding ape ritual could suffer from false negatives (as any finding), and so we do not consider the case closed yet. Another way to phrase the question is to ask what kind of rituals (should they exist) would we expect apes to have? Our answer to this was that we would expect these ape rituals to involve symbolic meaningful results. This proposal stems from the fact that apes seem to frequently express their ability to match results (see above), and from the fact that their potential for symbolic, causally- and goal-opaque results remains currently relatively untested. Future studies should therefore focus on testing ape symbolic abilities in the results domain. It might be possible that apes may be naturally susceptible to at least understand symbols when the symbol involves object changes, rather than fixed object states. It could also be tested whether specific results (but also actions) can produce symbolic meaning for apes. For example, what would happen if a study like Clay & Tennie ([26], actions on objects) would involve various extraordinary actions and/or results differentially applied to several objects—where the apes would have to merely later choose between objects (i.e. similar to the human study of [2])?

Indeed, two of the three cases of proto-ritual we identified involve environmental results produced by the apes themselves:¹⁵ objects that are placed in ears (chimpanzees) or air-biting twigs (orang-utans). The latter case also involves sounds produced by the orang-utans. At first, this might seem to go against our prediction that the likelihood for spontaneous ape ritual (or proto-ritual) can be expected to be highest in the results domain. However, as noted by Byrne [56], sounds are in effect environmental results. Therefore, at least two out of the three proto-ritual cases actually fit into our prediction rather well.

The orang-utan nest-making sounds and air-biting action are a seemingly inherently solitary activity. If validated as symbolic, or even as ritual, such ‘private ritual’ might be a precursor to social ritual from within the individual domain (e.g. imagine a human placing her/his pencils on her/his desk in a specific order each day¹⁶). Can we even imagine such private rituals in the absence of (language-aided) copying of the concept of ritual? This is an inherently difficult question to answer (as we said above, we currently remain agnostic), but if the answer is yes, then apes may (sometimes) have private rituals.

There could be several critical objections regarding our use of the criterion of symbolism (which distilled and replaced several of the usual criteria of ritual in our own definition). It might be objected that ordinary (and uncopied) behaviour should also potentially count, and that we have therefore produced many false negatives. This is a reasonable objection. Thus, human interpretations of ape behaviour may, at times, be wrong—and an apparently ordinary behaviour may, in fact, be symbolic/extraordinary after all. And even where the human interpretation of ape behaviour is correct, the apes themselves may perceive the behaviour as extraordinary (i.e. maybe apes themselves cannot understand the behaviour's function and/or its outcomes). Note that it might be argued that (many? or all?) human rituals might likewise appear ordinary from a similarly detached, scientific standpoint (e.g. religious ritual being ultimately ordinary as it serves as a group-binding exercise; e.g. [57]).

Another possible objection might be that ritual may nonetheless be buildable out of available actions and/or results acting as building blocks. That is, ritual may not require copying of novel actions or results and may be built of actions and results that are in the (latent) repertoire of the species (sensu [12,13]). Yet, we note that this possibility is encompassed in the category of proto-ritual, which had relaxed this very assumption. Anybody who would like the minimal definition of ritual to not include a need to copy may thus place our three candidate proto-rituals into a ritual bracket instead. But, in this case, it should be noted that the symbolic status of our three candidate cases is as yet not clear.

Finally, we suggest that a future examination should look at enculturated and/or trained apes, to answer a different question, namely whether apes can ever be expected to have full-blown ritual (even if it is ritual burrowed from humans). This would be especially important if language is a requirement for ritual (as language-trained apes can display some rudimentary language skills). Unfortunately (to the best of our knowledge), the current literature on enculturated apes does not include any formal studies of ritual. But even if ritual is then found in apes, it would not mean that apes spontaneously develop and maintain ritual among themselves—and the same is true for the LCA.

Data accessibility

This article has no additional data.

Authors' contributions

C.T. wrote the first version of the manuscript. C.T. and C.P.v.S. together finished writing the manuscript and wrote the revised manuscript.

Competing interests

We declare we have no competing interests.

Funding

C.T. was funded by the STONECULT project. The project STONECULT has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement no. 714658). C.T. also received funding from the Institutional Strategy of the University of Tübingen (Deutsche Forschungsgemeinschaft, ZUK 63).