On altruism, a tale of mice and men

Note to my readers: I am back with my weekly Substack essay and am sorry for the gap. The time was needed to devote to my book in preparation, Charles Darwin’s Illness: A medical detective story (OUP).(see Charles Darwin’s mysterious illness.) It is due out next year.

---

The property of “altruism” would seem to be an uncontroversial one. The word denotes the generous impulse to help a fellow creature, most often a member of one’s species, without regard to any benefit for oneself. For instance, when you give directions to a stranger who is lost, or help someone lift something that they cannot by themselves, or give money to someone on the street in obvious desperate need, you are being altruistic. Altruism is a natural, common, praiseworthy trait, and we see it in all human groups. Indeed, it would be natural to assume it was probably uniquely human. Animals clearly can be affectionate toward one another or their owners, or to help their young, but altruistic toward strangers? This belief has now been tested, in laboratories, and the results challenge the idea that altruism is solely a human trait. That work is reviewed here.

First, however, let us take a look at the nature and sources of altruism. Altruism looks like a simple desirable characteristic but for several decades, from roughly the 1920s/1930s to the 1970s, evolutionary biologists saw it as a problem. They were not against it, of course. They simply felt that it was hard to explain, perhaps that it should not exist, given the tenets of the then contemporary evolutionary theory, Neodarwinism,

Underlying their discomfort was the idea that altruism, being a complex biological property, would have to have been a product of natural selection. Yet, in Neodarwinism, natural selection should favour only those traits that convey a direct advantage to those individuals exhibiting those traits. Altruistic actions, by definition, are meant to help others without being of direct benefit to oneself. Indeed, such help might entail some disadvantage to the helper. In the jargon of evolutionary biology, when you help someone you are increasing their “fitness” to some extent but you may be causing a small loss of your own. After all, one is investing time and energy that could otherwise be used to one’s own benefit. In some altruistic acts, one is even putting oneself in danger. To take an extreme case, think of someone on an ocean beach who sees someone struggling in the surf and impulsively dives in to save them. In that attempt, they might themselves drown.

The answers to the puzzle of “why altruism?” that several evolutionary biologists produced came in two forms. First, there was the idea of “kin selection” pioneered by the evolutionary geneticist, William D. Hamilton, in the 1960s though the basic idea goes back to the great early 20^th century geneticist J. D. Haldane. (See What is odd about the “theory of evolution”? in this series for an anecdote about him, illustrating his wit and wisdom). Hamilton sought to explain the evolution of social insects. He focussed on honeybees, in which the female workers are sterile but help raise the offspring of their sister, the queen bee. If natural selection promotes an increase in gene (allele) frequency for certain traits, how could it promote sterility of many individuals, an evolutionary dead-end? Hamilton argued that sterility within a population could be selected if, like the sterile female worker bees, they foster the reproduction of close relatives, their kin, who do. This was the phenomenon of “inclusive fitness”, as Hamilton named it.

The other general approach that came to the rescue of altruism as a selectable property was “game theory”, which explained how when two individuals are involved in a seemingly altruistic act, as giver and receiver, it need not entail generosity on the part of the giver but some hidden trade-off, so that both parties benefit. This formulation simplifies a complex, interesting and subtle group of ideas but the gist is that game theory can often explain how what seems to be unselfish behaviour can create subtle but real advantages for the “altruistic” helper. ¹

Nevertheless, these theoretical advances did not eliminate the problem of how natural selection might promote a genetic propensity for high altruism, as in the saving of another individual’s life when there is no close genetic relatedness between the saviour and the saved nor any obvious reciprocal benefits to the former. To explain such behaviours, one has to fall back on more general notions of sociality, for which there is still little general theoretical, or at least mathematical, foundation. Though Darwin had briefly explored such social factors in his book, The Expression of the Emotions in Man and the Animals (1872), as playing some role in animals’ lives, he did not explore the basis of great acts of altruism in animals or people.

This subject of live-saving altruism in non-human animals has now been given a new dimension with the recent publication of two research articles in Science, America’s premier scientific journal. These investigations dealt with how mice try to help and revive fellow mice that had become unresponsive, hence who might have been in danger of dying. Such care-giving or “epimeletic” behaviour would be considered by most people as unquestionably altruistic behaviour. In the situation described below, the helper mice are not putting their lives at risk (as in human lifesaving actions), but they are clearly expending energy and time in what can only be described as an effort to aid the mice in distress. It must be being done without any likely immediate payback in return, let alone one that might increase the helper’s fitness.

In these two studies, individual mice were rendered seemingly helpless by appropriate doses of a sedative or an anaesthetic and then other mice, either those familiar to them or strangers brought into their cages and the responses of the visitors observed, recorded, and investigated. Though different investigators and mice were involved in the two studies, along with some differences in procedures, the results were in striking agreement.

In both sets of experiments, the alert mice were brought in once the drugs had been administered and had begun to take effect. The by-stander mice could see the treated mice become increasingly unresponsive. As they became so, the alert mice began to pay more and more attention and to start grooming the drugged mice, with a frequency not shown around normally active mice. The groomers paid special attention to the faces of the unresponsive mice. Furthermore, as the treated mice reached their low point in responsiveness, the responders increasingly focussed on the mouths and tongues, lifting the tongues out of the mouth, which facilitates breathing, and even gently biting them. If the experimenters had placed something in the mouth of the sedated animals, the helpers removed it. Both papers comment on the resemblance of this behaviour to cardiopulmonary resuscitation (CPR) procedures in humans, where the conscious aim of the helpers is to prevent the patient from letting their tongue obstruct their breathing. Clearly, in mice it was an automatic response to the situation, not one that had to be taught. ²

Of course, it is completely unknown whether mice have any conception of death and any conscious thought, even to the slightest degree, that their tongue-lifting and –biting activities might save the unconscious mouse’s life. The key point, however, is that these activities sped the revival of the unconscious but live mice, who became responsive and who then fully revived and became active. (In a moment, we will return to the question of whether non-human animals have any conception of death.)

This attempted revival activity – this is the easiest, most direct way to interpret it – was more frequent and intense at first with mice already familiar to the responders but was seen at high levels with all mice. Neither close relatedness (kin selection) nor prior familiarity was necessary. Nor did gender play a big part. Females were slightly more active in providing this attempted resuscitating care but not to a statistically significant extent. Hence this was a general social response, not tied to gender or relatedness, apparently evocable when one mouse sees another slipping into unresponsiveness. Almost certainly, none of the mice doing the resuscitating had ever engaged in such behaviour before nor could have been taught it. In earlier times, we would have described it as an “instinctive” response but that terminology is no longer in use, though we lack a good substitute term. It is a behaviour whose potential must be, in some way, genetically specified in mice but is evoked by a particular circumstance. ³

Both sets of investigators also probed the neurological basis of this behaviour. The two groups investigated two different regions of the mouse brain but found that both were involved in the response. Through gene expression tests that need not be explained here, one group found that a region termed the paraventricular neuron (PVN) of the hypothalamus, a basal part of the forebrain (specifically the diencephalon), while the other group explored the medial amygdala. Both regions are neurologically linked. The PVN secretes the hormone oxytocin, which is involved in various affectionate, empathic behaviours – it has been termed the “love hormone” – while the medial amygdala is known to participate in various social activities, including the grooming of others, including non-relatives, so called “allogrooming”. In the resuscitation behaviour, the PVN almost certainly activates the medial amygdala, which then directly triggers the allogrooming activities involved. Further sophisticated experiments, involving gene activation and suppression, demonstrated causal links between the neural activities in these two regions and the behaviors. Hence, activation of the relevant neurons in the PVN and the medial amygdala are necessary for the epimeletic behaviour, not just correlated with it.

Whenever a novel behaviour is seen in one kind of animal, one wonders whether it is unique to that species or only found in it and closely-related species. Hence, one should ask is such apparent resuscitation behaviour unique to mice amongst non-human animals? The answer, we can now say, appears to be “no”: other non-human animals have been observed to display comparable epimeletic behavior. These animals are chimpanzees, dolphins and elephants. All three cases involved a single individual in great distress in the wild being helped by its conspecifics (members of its own species). The elephant case is particularly informative, however, because all the individuals had been identified through long study of the particular group. In this instance, one could definitely conclude that it was not just close relatives who were involved in the helping activities. ⁴

Perhaps such behaviour in chimpanzees is not a surprise, given their close evolutionary relatedness to us, and we are altruistic animals, but dolphins and elephants are distant mammalian relatives of humans. Add mice to this group of animals capable of care-giving behaviour and it is not unreasonable that this may be a universal mammalian capacity or at least a fairly general one, even if its forms and degrees differ amongst different kinds of mammals. Furthermore, if it is fairly general amongst mammals, might it not also be found in some reptiles and/or certain bird species? What seems certain is that the psychological drive to help a conspecific in distress is not a uniquely human trait but found in other animals too.

These observations raise broader and deeper questions about the cognitive capacities of animals. One of them concerns the property mentioned earlier: do animals have any conception of death, not in the way that we humans have (with the extensive lore all human cultures have built up around it) but as a final and irreversible state of non-activity? Needless to say, we do not and cannot know given the language barrier between us and non-human animals which precludes our asking. Yet much circumstantial evidence indicates that certain highly intelligent animals, again chimpanzees, elephants, and a cetacean (in this case, orcas or killer whales) do indeed have some sense of death’s implications of terminal inactivity and its irreversibility. Also, with elephants, there is even some indication of what, in humans, we would call mourning. There is also anecdotal evidence on cats and dogs, and even peccaries (a kind of wild pig), that there is some apprehension of death’s finality and its terrible sadness for the living who were emotionally attached to the victim. This material has been covered in a recent popular science article. ⁵

The reality of epimeletic behaviour in animals is also relevant to the controversy on whether non-human animals have “theory of mind” (ToM), namely an awareness that other creatures have their own personalities, thoughts, feelings. (See Theory of Mind: which animals have it?) Of course one must avoid anthropomorphic projection (see next paragraph), but it is hard – for me – to imagine why someone would give care-giving to someone else, whether a member of your own species or a different one, without imagining them having feelings. And if they have feelings, doesn’t that imply they have their own personalities? Hence, I would suggest that care-giving behaviours in non-human animals is evidence that such animals have ToM.

For centuries, since at least René Descartes (mathematician, scientist, polymath) people have wondered about and debated how similar or different various non-human animals, in particular mammals and birds, were to human beings. To a large extent, this debate was dominated by the Cartesians who argued that since there was no evidence that animals thought or felt in ways similar to people, it was safest to regard them as “machines”, biological beings to be sure but, at the mental level, insensate machines, just following their impulses but not really thinking (or feeling). This was the position of the school of 20^th century behaviourism, led by the psychologist B.F. Skinner, which had academic respectability. Indeed, he and his colleagues made it unrespectable to think otherwise. Anyone who did was labelled as indulging in anthropomorphism, that is projecting human cognitive abilities and feelings onto a non-human animal.

Yet, as the findings reviewed here and as those in many animal behaviour studies mount up, it is becoming increasingly likely that this view is untenable, indeed false. Non-human animals are different from people in many ways, but in basic cognitive capacities, feelings, and in behaviours related to those feelings, there is growing evidence that much is shared between us and them. A mouse who is a first responder to another mouse in apparent distress is surely an indication of that.

1. We owe the development of ideas linking game theory to evolution chiefly to a brilliant young American, George Price, and an older Brit, the eminent John Maynard Smith.

2. The two research articles are by Sun, W. et al. (2025). Reviving-like social behaviour in response to unconscious or dead conspecifics in rodents. Science 387: 842 and F. Sun et al. (2025). A neural basis for prosocial behaviour toward unresponsive individuals. Science 387, 843. For a short account, see the “Perspectives” piece in the same issue, Sheeran, W.M. and Z.R. Donaldson (2025). An innate drive to save a life. Science 387: 827-828.

3. A perhaps similar situation is the “imprinting” behaviour of young chicks or ducks on the first creature they see as their mother. The propensity to do so is clearly genetic but it takes an actual experience seeing the first living creature, who would normally be the mother, to activate the behaviour.

4. The chimpanzee study is that of Shimada, M. and W. Yano (2023). Behavioural responses of wild chimpanzees toward a juvenile that suddenly lost its animacy due to a fall accident. Sci. Rep. 23: 16661. The dolphin research is that of Kuczaj, S. et al. (2015). Underwater observations of dolphin reactions to a distressed conspecific. Learn. Behav. 43: 289-300. The elephant research is given in Douglas-Hamilton, I. et al. (2006). Behavioral reactions of elephants towards a dying and deceased matriarch. Appl. Anim. Beh. Sci. 100: 87-102. Observations on care-giving behaviour in the wild is clearly to obtain than for animals in zoos but when zoo animals need aid, the zoo personnel are obliged to give it as quickly as possible, hence there is not enough time to see how conspecifics might step in to help.

5. This article is Wares, M. (2025). How animals understand death: intimations of mortality are not ours alone. Nautilus https://nautil.us/how-animals understand-death-1204412.