Coping with mortality: responses of monkeys and great apes to collapsed, inanimate and dead conspecifics. Arianna De Marco, Roberto Cozzolino & Bernard Thierry. Ethology Ecology & Evolution, Mar 21 2021. https://doi.org/10.1080/03949370.2021.1893826
Abstract: It was long assumed that only humans can distinguish the living from the dead. Renewed interest in this question over the last decade has led several authors to assert that non-human primates are also aware of death. We investigate this issue by comparing the behaviours of monkeys and great apes toward helpless conspecifics, basing our analysis on published reports. We first examine the behaviours of mothers towards the body of their dead offspring. They may carry the corpse for days or more before abandoning it. They groom, inspect and protect it, sometimes allowing group members to explore it, and rare cases of cannibalism have been reported. No significant difference is observed in the way that monkeys and great apes treat the bodies of infants. We then examine responses to collapsed (still able to move and react) and inanimate (unresponsive or dead) conspecifics. Monkeys and great apes guard, care for and inspect their helpless partners, and also manipulate and mobilise them. Through these actions, individuals may inform themselves about the state of their partners, test their responsiveness and/or attempt to rouse them. It appears that only chimpanzees and gorillas show violent action such as display behaviours and the rough treatment of bodies. They can also make distress calls, and periods of “stunned silence” sometimes occur in chimpanzees, indicating that they are experiencing intense emotion. Finally, we argue that while both monkeys and great apes detect body dysfunction through the victims’ inability to wake up and move, only great apes can understand that something serious has happened. The signs of emotional disturbance reported in them indicate that they may believe that inanimate conspecifics have entered a state of “dormancy”, meaning that they are unlikely to regain wakefulness. However, there is no evidence that any non-human primates are aware of mortality.
Keywords: deathemotiondistressempathymental representationepimeletic behaviourprimate
RECOGNITION OF DYSFUNCTION AND AWARENESS OF MORTALITY
Since death is the cessation of life, it may seem necessary to know what primates understand about life in order to know what they understand about death. Animals are defined by their animacy, and it has been shown in several animal species, including primates, that individuals are able to distinguish between animate and inanimate as human infants do: they can separate entities that engage in self-generated motion from those that do not (Opfer & Gelman 2011; Tsutsumi et al. 2012; Abdai et al. 2021). Some authors have proposed that animals and humans are endowed with several innate systems of “core knowledge” that are designed to build representations with specific contents. One system would serve to represent (inanimate) objects and their mechanical effects, while another would serve to represent (animate) agents and their actions (Spelke & Kinzler 2007; Carey 2009). On the basis of this theoretical framework, Gonçalves and colleagues (Gonçalves & Biro 2018; Gonçalves & Carvalho 2019) hypothesize that when a living being dies, its body becomes an object, thus activating both the object and agency core systems in bystanders. This would create a perceptual mismatch in them, resulting in an “animacy detection malfunction”.
It is sound to consider that an inanimate body can present contradictory clues that do not meet the expectations of conspecifics, and hinder their ability to recognize its state. However, it is probably too early to assert that the perception of a corpse induces a conflict between an object representation system and something like a life representation system. We are far from knowing what dynamic clues subjects use to assess animacy, and as Gonçalves and Carvalho (2019) acknowledge, agency involves much more than animacy: agents are capable of attentional states, goal-directed movements and interactions with others; primate species probably vary in their ability to identify these different components of agency. Moreover, subjects can also identify living beings like animals using morphological features such as the presence of a face, eyes, limbs or species-specific characteristics, and we do not know how dynamic and morphological clues are integrated at the brain level (Opfer & Gelman 2011; Fisher & Freiwald 2015). We should add to these multiple clues the past interactions and social relationships that a subject may have memorized about a given conspecific, which would also contribute to the subject’s representation of this individual.
Rather than starting from theoretical constructs about the recognition of life by primates, it seems more productive for the time being to address what they understand about the state of their inanimate conspecifics. However, discriminating inanimate individuals from perceptual features is not the same as knowing what it means to be dead (Allen & Hauser 1991; Monsó 2020). In what follows, we will examine in turn whether monkeys and great apes detect conspecific dysfunction, show signs of disturbance, and exhibit some awareness of mortality.
Detection of dysfunction
Although death is usually regarded as an event, it is more accurate to think of it as a process in which one organic function fails after another. In principle, it would thus be possible to detect the dysfunction of a body using different clues: cardiorespiratory arrest, cessation of eye movements, and the inability to move and react.
As previously reported, monkeys and great apes have been observed inspecting the face and/or mouth of collapsed or inanimate partners, and Anderson et al. (2010) wrote that the companions of a dying chimpanzee “appeared to test for signs of life by closely inspecting her mouth and manipulating her limbs (test for pulse or breath)”. However, nobody ever saw a monkey or an ape move their ear close to the mouth of another to listen to his/her breathing; nobody ever saw a monkey or an ape put their hand on the chest of another to feel his/her heartbeat. Not only would the contrary imply expectations about the need for breathing and heartbeats, but it is even hard for humans to be certain that the lungs and heart have stopped, especially in weakened individuals (Stewart 2003). Historically, people waited for signs of putrefaction to confirm death. Until a few centuries ago, tests for respiratory ventilation consisted for example of placing a feather under the nose, or holding a mirror before the mouth to detect condensation; it was only with the invention of the stethoscope in the 19th century that medical practitioners were able to determine with a reasonable degree of certainty whether the heart had ceased to function (President’s Commission 1981; Powner et al. 1996; Whetstine 2008). Against this background, it seems unlikely that primates other than humans have the ability to use soundless breathing, heart rate or pulse to assess the state of their conspecifics.
Another long-established medical sign of decease has been the fixity of the eyes (President’s Commission 1981; Whetstine 2008). This is understandable given the normal animation of the eye area, which includes gaze orientation, pupillary reflex, and eyelid movements. Experimental studies in humans and other primates have shown that they are disproportionally interested in the eye area; they are highly sensitive to the attention and gaze direction of others, and are able to adjust their behaviour to get a partner to attend to them (Emery 2000; Bourjade et al. 2014; Liebal et al. 2014; Tomasello & Call 2019). Correspondingly, monkeys and great apes often inspect the face and/or eyes of collapsed and recently deceased conspecifics (Tables 1–2). It is safe to assume that they are able to detect the absence of eye movement. The inability to make eye contact and get attention may warn them that something is going wrong with their companion. We can note that the eyes of inanimate individuals may be open or closed, but this information is not found in scientific reports. Fingering or picking at the eyes or eyelids has sometimes been observed in monkeys (Yerkes 1915; Fashing et al. 2011). According to a second-hand account, chimpanzees were once seen opening the eyes of a dead companion (Bekoff 2007).
Even if animacy is a main attribute of animals, it is common for primates to encounter motionless conspecifics. They regularly experience periods of immobility, which are first and foremost due to sleep phases. It is not the lack of animacy per se that must seem odd to observers of motionless individuals, but rather their inability to regain animacy: they are expected to wake up when someone acts on them, and the problem arises when they fail to do so. Assessing the responsiveness of partners and probing or stimulating them may account for many of the behaviours of primates towards collapsed and inanimate conspecifics. Although observing and sniffing can provide initial information, it is only by touching and above all handling partners that group members can test their reactivity or attempt to rouse them – with rough treatments producing the same effects. Whatever the exact goals of the intervening individuals, their actions can inform them if their partners were previously asleep or if something is going wrong. This is not an issue for collapsed partners, as they may still show some movement; we have indeed found that manipulative actions were mainly reported for inanimate partners, further indicating that monkeys and great apes are able to discriminate between these two conditions (Table 2). Note that we would expect the opposite trend for mobilising actions that specifically seek to lift a partner and restore him/her to an upright position: if the performers are able to distinguish between the two conditions, these actions should be applied to apparently wakeful – i.e. collapsed – partners rather than to inanimate ones. As already mentioned, only two unambiguous cases of this type of helping behaviour have been described. In the first instance a mountain gorilla attempted to pull a juvenile into a sitting position while the latter was still reactive (Watts 2020: case 1). In contrast, the second instance concerns a rhesus macaque who tried to lift a partner when the partner was inanimate, and did so again when the partner was beginning to regain consciousness after fainting (see Supplemental Material). There are too few cases to be able to draw firm conclusions.
The termination of breathing, heartbeat, eye movements and any reflex or motor response is referred to by medicine as negative signs of life. They are later followed by positive signs of death: cooling of the body, stiffening, putrid odours and changes in skin colour that result from the decomposition process. These different alterations gradually take the body away from its living appearance, which could contribute to the information – or confusion – of those who inspect it (Gonçalves & Biro 2018). However, positive signs of death do not appear until several hours or even more than a day after death, so in many cases group members have assessed the state of the affected partner long before this time.
It may be expected that the more experienced individuals are better at detecting unresponsiveness and other signs of death. Three studies have compared the duration of dead infant carrying in females of different age and maternal experience. This yielded contradictory results that provided little support for the hypothesis that younger and/or first-time mothers would carry their dead infant for longer periods of time than other mothers (Sugiyama et al. 2009; Das et al. 2019; Lonsdorf et al. 2020). Besides, younger chimpanzees tend to exhibit more exploratory behaviours toward corpses than adults (Teleki 1973a; Stewart et al. 2012; van Leeuwen et al. 2016; Jakucińska et al. 2020), and it is not rare for them to play with the bodies of dead infants (see above). However, little quantitative data are available in primates about the tendency of different age-and-sex categories to examine the body of affected conspecifics (see De Marco et al. 2020).
Signs of disturbance
In the foregoing, we deliberately used the vague expression “something is going wrong” to account for what primates can infer when faced with a conspecific’s inability to respond to their actions. The question now is whether this can cause concern to subjects. Not understanding something is not necessarily a reason for worry. On the contrary, if primates understand that something serious has happened, it can be disturbing to them and can be reflected in their behaviour.
Several authors have claimed that the occurrence of affiliative behaviours towards collapsing or dying conspecifics is a mark of compassion (Bezerra et al. 2014; Yang et al. 2016; Gonçalves & Carvalho 2019). However, the affected individuals in these situations can move and emit sounds, so partners may respond to their signs of distress without the need for a mental representation of their emotional states. If compassion implies an understanding of distress that involves empathic perspective-taking or the sharing of feelings (Cheney & Seyfarth 2007; Boesch 2012; Adriense et al. 2020), then responses to the behavioural expression of distress are not proof of compassion. On the other hand, there is converging evidence that great apes such as chimpanzees are capable of targeted helping and empathic perspective-taking, and can thus apprehend the suffering of conspecifics (de Waal 2008; Boesch 2012; Pérez-Manrique & Gomila 2018; Sato et al. 2019). Nevertheless, when chimpanzees display affiliation towards a dying conspecific, it is difficult to justify any assumption that this a caretaking behaviour anticipating an approaching death (Anderson et al. 2010): an individual who is collapsing or dying is by definition still alive; we have no reason to believe that others can predict she/he is going to die, and they may actually perceive her/him as rescuable. Stringent conditions are necessary to examine whether a companion’s physical condition may disturb partners, regardless of the occurrence of any communicative cues. We must rule out reactions to collapsed conspecifics and focus on reactions to inanimate conspecifics. We must also exclude cases where the presence of a danger in the surroundings can elicit distress or warning calls from bystanders (e.g. Mohnot 1980; Gupta 2000; Perry & Manson 2008: Maní/Pobrecito; Riley et al. 2015; Campbell et al. 2016: case 4).
The display behaviours and rough treatments that chimpanzees and gorillas show towards their inanimate conspecifics are among the most dramatic responses reported in animals (Table 2). Males are primarily responsible for these actions, which in some cases were repeated over several hours (Fossey 1983; Stewart et al. 2012). Even though rough treatments may contribute to test or awaken the partner, the energy devoted to violent behaviours indicates that the individuals’ motivations go beyond these goals. Regardless of whether this behaviour expresses frustration due to incomprehension or a lack of acceptance of the partner’s state, it at least informs us that individuals find the situation serious enough to engage in such heated action. The alternation of periods of quiet attendance and periods of excitement is another hallmark of the chimpanzee response. Emotional displays may include warning and distress calls, while stillness may pervade the entire group in cases of stunned silence (see above). Here again, signs of intense emotional disturbance in chimpanzees indicate that they attribute a special meaning to the unresponsiveness of their companion. No comparable periods of stunned silence have been recorded in gorillas, but it is difficult to make inferences about this difference with chimpanzees because gorilla temperament is characterized by high emotional stability (Schaeffer & Steklis 2014; Eckhardt et al. 2015). No conclusions can be drawn for bonobos and orang-utans due to the lack of documented reports in these species.
The comparison of monkey behaviour with that of chimpanzees and gorillas is illuminating. No rough treatment of corpses was reported in monkeys. Display behaviours were observed once in Hanuman langurs, but they were not directed at the body of the dead individual (Mohnot 1980). Monkey responses to inanimate conspecifics are generally unremarkable. In one well-documented case in brown capuchins, for example, group members quietly inspected and handled the body of a recently deceased female in turn, and even her sister and offspring did not exhibit any signs of distress (De Marco et al. 2020). In another instance in rhesus macaques, a male made sustained efforts to revive an unconscious partner by mobilising and lifting him, and although the intervening male yawned a few times, indicating a significant degree of tension (see Maestripieri et al. 1992; Deputte 1994), he showed no evidence of being alarmed (see Supplemental Material). Likewise, the other reported cases showed no signs of distress (Table 2; see also Takeshita et al. 2020). There were two types of situation in which distress vocalisations were heard, but they do not actually contravene the previous statement. In the first case, infants cried and screamed following their mother’s death (Mohnot 1980; Fashing et al. 2011); however, these behaviours also occur as part of normal mother-offspring interactions when a female does not attend to her infant. The second situation involves mothers emitting warning, distress or contact calls when they lose the body of their dead infant. Here, the triggering event is not the infant’s lack of response, but the absence of the infant’s body, as already mentioned (Altmann & Altmann 1970: Shorty; Smuts 1985: Zandra/Zephyr; Perry & Manson 2008: Abby/Omni; Li et al. 2012: case 1; Botting & van de Waal 2020: Nurks; Carter et al. 2020: cases 1 and 5). Although monkeys sometimes display signs of tension or anxiety such as restlessness, scratching and yawning when confronted with the unresponsiveness of a conspecific (Campbell et al. 2016: case 1; Carter et al. 2020; De Marco et al. 2020; Supplemental Material), they do not show any signs of the intense emotional disturbance observed in chimpanzees and gorillas. The communicative signals that only monkeys send to non-responding conspecifics may represent a further consequence of their inability to realize the severity of their state, whereas chimpanzees and gorillas could be aware that any attempt to communicate is pointless.
Awareness of mortality
We came to the conclusion that monkeys and great apes are able to understand that “something is going wrong” when confronted with an inanimate conspecific. We also reached the conclusion that at least great apes such as chimpanzees and gorillas can also understand that “something serious has happened”. However, the question remains as to whether they are aware of what it means when a partner remains permanently inert, and whether we can gain any insight into this question. Research studies in developmental psychology have shown that the understanding of death in human children can be defined and measured through several dimensions (Speece & Brent 1984; Kenyon 2001; Slaughter 2005; Longbottom & Slaughter 2018). From 4 to 10 years of age, children successively master four main dimensions related to the concept of mortality: (1) Irreversibility: a dead individual cannot return to life, (2) Universality: all living beings must eventually die, (3) Cessation: all bodily and mental functions cease upon death, (4) Causation: the breakdown of bodily functions is due to external and internal causes.
Irreversibility is probably the easiest dimension of mortality to grasp, as it is the first dimension that children successfully master. Several authors have expressed the opinion that primates are capable of understanding the permanence of death (de Waal 2013; Anderson 2018; Gonçalves & Carvalho 2019). de Waal (2013) reports an anecdote where a film was projected to a group of captive chimpanzees. The film showed a male from the group who had died a few years earlier, and his appearance on the screen provoked a strong emotional reaction from his former rivals. Gonçalves and Carvalho (2019) argue that their reaction was evidence of a violation of expectation regarding the irreversibility of death. However, this event is by no means conclusive, as it could just as well be used to support the view that chimpanzees accept that dead individuals can return to life, and be touched if this appears to happen. Other researchers claim that the weakening of the mother’s attachment to her dead infant and her eventual abandonment of the corpse prove her awareness of the irreversibility of death (Das et al. 2019; Lonsdorf et al. 2020; Monsó 2020). If it were true, however, this explanation would be tantamount to asserting that mothers of any animal species who give up the putrefying body of their offspring are aware of the irreversibility of death. One might just as well reason ad absurdum that chimpanzee females are among the least cognitively gifted primates since more than half of them in the Gombe population carry their dead offspring (see Lonsdorf et al. 2020). It is more straightforward to say that mothers have no other choice but to abandon the corpses after some time, and that this behaviour tells us nothing about death awareness in primates.
Anderson (2011, 2018) points out that chimpanzees have many opportunities to learn that inanimate individuals generally do not wake up, not only because they are exposed to deceased conspecifics – infant death being the most common event – but also because they regularly capture animals that they may kill and handle for a period of time before consuming them. Indeed, these different experiences are likely to inform individuals that subjects who remain immobile for extended periods of time generally do not recover. It should be noted however that small animals of a wide range of taxa follow death-feigning anti-predator strategies in which they go into tonic immobility when captured (Humphreys & Ruxton 2018). This means that it is not uncommon for primates to encounter motionless animals, some of which can subsequently resume activity. It should also be remembered that witnessing sleeping and awakening sequences in others is a daily reality for them. Furthermore, individuals may have also seen unconscious partners who later recover, because they were temporarily ill or had passed out. For example, falling from a tree is a continuing risk for tree-dwelling animals (see Table 2); primates cannot avoid occasionally landing on a dead branch, and if they fail to catch hold of a lower branch, they may pass out or die when they hit the ground, depending on the distance they fall and the nature of the ground. Waking up is instantaneous in a sleeping partner, whereas it may seem delayed in an unconscious partner. The consequence is that it is always difficult for primates to predict what will happen to an inanimate conspecific because the possibility that she/he may recuperate can rarely be excluded. The occurrence of visitations points in the same direction. By returning to a body they have previously left, they can check if their companion is still in the same state.
Children grow in a world of cultural symbols and conventions, and the representations embodied in the human language help them to form abstract categories (Tomasello 1999; Carey 2009). The words “death” and “dead” are offered to them as tools that are passed on from one generation to the next (Renfrew 2016). Yet it takes years for them to understand that terms such as “never”, “always” or “definitely” can have absolute meaning (Kuczaj 1974; Hoffner et al. 1990). In comparison, animals are solipsist thinkers, meaning that they learn by themselves. They receive no support from others that would help them to conceptualize the absolute dimensions of death. While great apes can learn that unresponsive companions may not recover, this is probabilistic knowledge and not the absolute knowledge implied by words such as “irreversibility” and “permanence”. Rather than asserting that great apes may be aware of the irreversibility of death, it seems more appropriate to echo Boesch’s (2012) formulation that chimpanzees may be able to “realize dead individuals do not move or need help anymore, and that they will remain inanimate”. This knowledge involves some uncertainty; it may reflect the belief that these individuals have fallen into a state where they seem to remain asleep indefinitely. We propose to call this state “dormancy”, which means that the individual is in a prolonged state of immobility that is not normal sleep, with no rule specifying that recovery is excluded. In terms of the conceptualization abilities required, belief in dormancy is a less costly hypothesis than an awareness of the irreversibility of death.
Like irreversibility, the second dimension of mortality, universality, has an absolute meaning. We cannot question primates directly like we do with children, and there is no indication to date that they are able to know that death happens to everyone, whether they are conspecifics or individuals of other species (Anderson 2018). The third and fourth dimensions of mortality are the last to be mastered by children because they involve basic knowledge about the biological properties of living beings and the body functions that maintain life (Slaughter 2005). Some authors assume that primates recognize death as the cessation of bodily functions, basing their reasoning on the fact that mothers handle and defend the body of their dead infant (Das et al. 2019), or on changes in maternal behaviour such as atypical carrying postures, distancing from the corpse and its final abandonment (Lonsdorf et al. 2020). However, the first argument ignores the fact that mothers also handle and defend their live infants (Gonçalves & Carvalho 2019), and the second fails to distinguish between detection of dysfunction and awareness of mortality: mothers may be able to recognize that something is going wrong with their non-responding infant without necessarily understanding what has happened (see above). The fourth main dimension of mortality, causation, is particularly demanding on the cognitive level since it requires subjects to infer causal relations, i.e. formulate hypotheses about the mechanisms involved in the breakdown of the functioning of the organism. It has been suggested that prey killing and lethal attacks on conspecifics may help individuals to gain insight into the causes of death (Anderson 2018), and also that the investigation of inanimate bodies would enable the understanding of these causes (Das et al. 2019). However, it should be clear that inspecting corpses or even inflicting death on others – as all predators do – is by no means proof that they realize how death occurs. Several authors have argued, based on a limited number of cases, that individuals react differently to deceased partners depending on the cause of death, but the views expressed by these authors are conflicting. According to Anderson and collaborators (Anderson 2011; Yang et al. 2016; Pettitt & Anderson 2020), sudden and traumatic deaths resulting from an accident or a killing would trigger more alarm and violent behaviours than deaths due to non-apparent causes such as illness. In contrast, Boesch (2012) proposes that bodies with wounds generate fewer signs of fear and alarm in chimpanzees than those of individuals who die from unexplained causes such as illness. Whatever the case, the accumulation of reports in chimpanzees and other species indicates that none of these expectations have been confirmed, either for dead infants (Rajpurohit 1997; Watson & Matsuzawa 2018), or other inanimate conspecifics (Table 2).
While the absence of evidence is not evidence of absence, the burden of proof rests on the person who asserts that something exists, and not on the person who denies it. Monsó (2020) looked for behaviours that would demonstrate that animals understand death, but she only proposed signs of emotional disturbance and ways to detect the dysfunction of the body. To date, there is no evidence that primates, including great apes, have an absolute concept of death, whatever the dimension considered, be it irreversibility, universality, cessation or causation. At this point, it is important to notice that irreversibility differs from the other three dimensions in two major respects. First, irreversibility can be reduced to a more concrete notion such as dormancy. To bystanders, a dormant conspecific may appear to be a sleeping individual who fails to wake up despite repeated interactions with her/him, and who remains in this state for an indefinite period of time. It is also conceivable that experienced individuals may have learned that it is common for a dormant body to cool down, smell, be surrounded by flies, and gradually lose its shape. Second, the awareness of the mortal condition of all living beings or the biological functions involved in the cessation of life and the causation of death is very abstract knowledge with little emotional content. These three dimensions of mortality cannot account for the strong emotional commitment revealed by the behaviour of great apes such as chimpanzees and gorillas. On the contrary, the ability to conceive of irreversibility or dormancy may explain that they understand something serious has happened. Realizing that the inanimate individual in front of them is unlikely to come out of sleep, unlikely to revert to the way she/he was, and must be left behind, can induce intense emotional disturbance. This may explain the display behaviours and rough treatments that have been recorded in chimpanzees and gorillas, reflecting a lack of comprehension and/or acceptance that their companion remains in an inanimate state. This may also explain the distress calls and episodes of stunned silence reported in chimpanzees. Some authors consider that such behaviours indicate an awareness of the irreversibility of death (Anderson 2017; Monsó 2020). As discussed earlier, we find it more accurate to attribute these responses to the belief that their partner has gone into a dormant state. Belief in dormancy represents a more parsimonious explanation that does not require the involvement of concepts of mortality that may be too abstract to belong to the cognitive realm of great apes.
