Introduction

1Elephant behaviour and how it relates to the environment has been widely studied from a biological perspective (e.g. Sukumar, 2003; Moss et al., 2011). How, why, where and how much humans are implicitly and directly included in these studies is of interest in terms of how animals and humans are distinguished, presented and studied. This is the case for many animal systems (Davis & Balfour, 1992). Specifically regarding elephants, it also raises questions about the definition of the environment in animal behaviour, how we define wild and domestic animals, as well as the human environments of African and Asian elephants (Loxodonta africana, Loxodonta cyclotis, Elephas maximus) and how these should be included, avoided or discussed when making observations and/or conducting experiments about the behaviour of elephants.

2It is notable that the prominent texts on elephant ecology and behaviour often include extensive descriptions of elephant interactions with humans. For example, Sukumar’s monograph dedicates an entire chapter to “the interrelationship of culture and ecology”. He describes and summarises key trends and events related to the interrelationships of humans and elephants in broadly chronological order, from the Pleistocene to the time contemporaneous to its writing. That Sukumar included the chapter goes beyond providing informative background material and directly addresses the continued but changing importance of Asian elephants in South Asia.

3“Long after the war elephant has faded into history, the elephant-headed god reigns supreme, more popular than ever before, assuming new roles and adapting to changing circumstances. The elephant itself has assumed the role of a flagship in India’s efforts in conserving its forests…” (Sukumar, 2003: 80-81).

4This link between the historical role of elephants in war, the religious significance of elephants, and the status or use of Asian elephants as a flagship species in conservation is notable not for the novelty of the concept but for its presence in a non-introductory chapter of a book about the behaviour and ecology of elephants. The human environment is clearly presented as a salient context for the subsequent chapters on elephant reproductive and social behaviour and the environmental impact of behaviour, topics frequently studied in behavioural ecology. Because of this, and many other examples, my goal is not to introduce humans into studies of elephant behavioural ecology. Rather, it is to note that they are already present, although sometimes not explicitly acknowledged, and can potentially offer us opportunities as well as challenges in their inclusion in studies using behavioural ecology methods and analytical tools.

5The foundation of behavioural ecology studies is the attempt to understand how an animal’s behaviour is adapted to its environment (Davies et al., 2012). As with many concepts in science, the simplicity of the statement belies the complexity of addressing it. Behaviour encompasses many aspects of the lives of animals; it can include communication, behaviour associated with reproduction, social behaviour, parent-offspring behaviour, foraging and avoiding predation (Davies et al., 2012). Animal behaviour studies have a breadth and diversity of scales, including collective behaviour, whereby the scale is not the individual but instead might be a school of fish or a murmuration of starlings. Animal behaviour can even operate within individuals, as when they are host to parasites. My lens here is not trained on the behavioural aspect of the foundational aim of behavioural ecology but rather on the word “environment” and whether the concept of it, and being “adapted” to it, are the biggest challenges to the much smaller goal of including interactions with humans and elephants in behavioural ecology studies.

6To consider this, we must address what “environment” means and the theoretical basis for analyses of adaptation. Approaching the latter first, it is clear that evolutionary theory, including life history theory (Davies & Krebs, 1997), is central to the analytical framework of behavioural ecology. For example, Nico Tinbergen’s oft-referenced four “why?” questions of scientific animal behaviour studies all directly or indirectly refer to evolutionary processes or analyses (Tinbergen, 1963). The first two, addressing causation (or mechanism) and onto-geny (developmental trajectory), are shaped by evolution. The third, on survival value or adaptive advantage, directly refers to biological fitness and the fourth, on evolutionary history, is specifically concerned with how behaviours evolved. Decades after this seminal work, there remains a consensus that the evolutionary framework is integral to the field (Davies et al., 2012).

Humans and the human environment in behavioural ecology

7With that consensus, we can move on to the concept of the environment. To me, this is the most interesting aspect of behavioural ecology’s aim. It centres on interaction, specifically the interaction between the environment and animal behaviour. My interest here lies in the question: to what extent do humans form part of the environment for elephants and other species and therefore should be considered part of behavioural ecology studies at this foundational level? This includes acknowledging the human impacts on most environments, even when they are not physically present at the time of a study. More specifically and salient to the study of elephants and humans, I am interested in how researchers, animal caretakers, and veterinarians—people working directly with the animals or within their habitat—also form part of that environment. By extension, the research being conducted becomes a part of the environment. This is not to say that research shouldn’t take place or that behavioural ecology studies have an impact on animals which is unethical, which is another topic. It is, rather, to consider how humans in the environment and their direct or indirect interactions with animals are part of behavioural ecology and the challenges and opportunities this represents in terms of study design, analysis and interpretation.

8To do this, we have to know what a behavioural ecological study is, which is a challenge in itself because of the diversity of study designs. From a methodological perspective, behavioural ecology studies have used both observational approaches and experiments. These observational responses can be highly creative, and might include observing animals to see behaviour such as using tools to forage (Sanz et al., 2013) or analysing interactions such as grooming (Henazi & Barrett, 1999). Humans can either be present physically or use cameras, including camera traps (Caravaggi et al., 2017), telemetry, and other tracking technology, including pit tags or GPS collars or drones (Hughey et al., 2018; Lahoz-Monfort & Magrath, 2021). The observations might not be of the animal itself but could be based on faeces, prints, hair or the area an animal has used, for example the site it selected for sleeping (Cheyne et al., 2013). Experiments can include manipulating nests (Soler et al., 2001), presenting animals with a model conspecific (Turner et al., 2020), and using playbacks of calls that other animals made (Prat et al., 2015). The unifying theme is that the responses of an animal to the environment are in some way classified and/or measured, as is an element of the environment, which might be introduced, manipulated, altered or just singled out by measuring it. That is not to say they are the only variables measured, controlled or manipulated, but that it forms the central unifying framework for studies in behavioural ecology.

Behavioural ecology’s relationship with the wild

9An implicit assumption of behavioural ecology studies is that an ideal study involves observations and experiments of animals conducted “in the wild” (Davies et al., 2012). A further assumption is that animals can be habituated to the presence of human observers, who are able to distinguish between natural behaviour and behavioural responses to observers (Candea, 2013). However, many studies investigate the behaviour of captive animals or species defined in the scientific lexicon as domesticated, such as dogs (Canis lupus familiaris) (Dale et al., 2016). Domestication is defined as genetic and morphological changes from the originally-wild species as a result of generations of selection by humans to fit human preferences and agro-economic niches (Bates, 2021). Although discussions on what domestication means might seem peripheral to behavioural ecology and scientific studies of animal behaviour, domestication does have relevance because any level of human intervention is generally seen to negatively impact animal behaviour and change it from natural behaviour (Gill et al., 2001). Elephants are an interesting example of this as they both blur the line between “domestic” and “wild” because, despite their history of being kept in captivity, they have not been selectively bred over generations and do not reach the scientific status of a “domestic” animal. Furthermore, elephants that are defined as wild and free-roaming in some studies are often impacted by tourism, hunting or other interactions with humans (Goldenberg & Wittemyer, 2017).

10In that way, elephants also bring to light the potential difficulties of applying an evolutionary framework in behavioural ecology. That is, if evolution by natural selection is the theory through which all interactions are analysed, then any intervention by humans, from hunting to observing, could be interpreted as operating in addition to natural selection and, therefore, should be avoided in studies. However, this is a limited understanding of both the evolutionary context of elephants, which have been in contact with human populations for much of their evolutionary history (Anzidel et al., 2012; Zutovski & Barkai, 2016), and of behavioural responses to humans. If we consider humans as part of the complex environments animals live in, then they are clearly relevant. For example, some terms such as “natural behaviour” can assume a lack of human influence, even as a human observes and describes those behaviours. However, definitions of natural behaviour do include any adaptive behaviour, including responses to humans (Špinka, 2006). In light of the range of human positions and perspectives in behavioural ecological research questions, study designs and writing on animal behaviour, further consideration of how they relate to behavioural ecology studies is justified.

11Explicitly stating the presence of humans, even when that presence is indirect, for example, through the process of domestication or through captive settings, can be viewed as highlighting a problem by behavioural ecologists, whose focus is on the animal. Many experimental paradigms focus on controlling study conditions as much as possible in order to isolate measurable effects, often including efforts to reduce the impact of humans in study designs. In studies by my own group, this has included researchers hiding behind a curtain so that elephants do not have access to visual information when the researcher is refilling buckets with food items. This measure draws a direct line to the example of Clever Hans, a domesticated horse (Equus ferus caballus). In the late 19th century, Clever Hans became renowned for what seemed to be skills in addition and subtraction (de Waal, 2017). The handler asked a mathematical question verbally, and the horse responded by tapping his hoof the appropriate number of times to answer the question. What became clear was that when his owner was behind a curtain, he didn’t tap out the correct number. His owner unknowingly gave Hans cues by tensing and then relaxing his body. Frans de Waal highlights the sensitivity and awareness required in animals to notice and respond to human behavioural cues. Even if solving mathematical problems based on human language wasn’t possible for Hans, he could do something else that was very informative of horses and their environment; he responded to cues from a human he interacted closely with.

Elephant research

12The story of Clever Hans shows that the questions in elephant behaviour we choose to investigate are important. I concede that it presents additional difficulties to include humans in the study environment for certain research questions. One solution often taken is excluding the involvement and interactions with humans from analyses or attempting to account for any variation they introduce as noise. This chapter proposes that the interactions can also be viewed as introducing an additional layer to behavioural ecology studies that could merit further analysis in its own right. That is not to suggest that this proposal is new or that it hasn’t been considered before for elephants studies; for example, previous studies investigated human pointing cues in directing elephants to food (Plotnik et al., 2013; Smet & Byrne, 2013), and analysed how the duration of the relationship between mahout and elephant is associated with the elephant’s performance in a novel task (Liehrmann et al., 2021). These studies go beyond just acknowledging humans, often mahouts, involved in the lives of elephants and integrate them into the study concept. The expansion of experiments and observational studies of animals to include interactions with humans can provide opportunities through extending or reimagining paradigms that are familiar to evolutionary and behavioural ecologists. For example, social learning tasks that involve manipulating a novel apparatus are often implemented by animals observing other animals interacting with the apparatus (van de Waal et al., 2013). Such tasks could also be implemented using humans as the “model” individuals that the animal observes interacting with the experimental apparatus, as well as observing other elephants. This would allow us to investigate whether elephants learn from human models, if it takes longer than when they observe a conspecific, and whether it is affected by, for example, if and for how long they have known the human modelling the behaviour.

13The aim of one of my group’s studies was to analyse food preferences in eight captive Asian elephants using a simple choice test. At the start of the experiment, the elephant was released into the experimental area and allowed to explore two sealed buckets attached to a table. The lids on the buckets had holes. In that way, they had access to the scent of the food, but they couldn’t touch it, and the holes were small, so they couldn’t clearly see the food inside. After precisely a minute, we drew a black curtain in front of the table and removed the lids from the buckets. We then opened the curtain, and the elephant could eat from one bucket. We randomised the food in the buckets and which side of the table it was presented and performed multiple trials to test whether an elephant chose one food significantly more often than we would expect by chance. There were lots of issues in employing this simple design. Some elephants did not approach the table, whereas others attempted to approach it consistently and had to be moved beyond the experimental area between trials to adhere to the experimental protocol. Some individuals reacted strongly to the curtain: some moved away while others interacted with it. A couple of elephants always went for the bucket on the right or the one on the left or chose to eat from the one they had touched first or last. Some seemed to interact much more with one bucket, but by the time the curtains had been drawn and opened, they did not go back to that one. It was difficult to manage this experimental design, but importantly, what made it possible was the mahouts that were present with the elephants.

14At the study site, Tiger Tops Lodge, outside Chitwan National Park in Nepal, each elephant had two mahouts. Historically, the elephants had been chained at night under structures so they could avoid the rain or dew. Since 2018 they have been kept in fenced corrals, sometimes alone or with another one or two other elephants (Mumby, 2019). The mahouts clean the corrals and take the elephants to the grasslands to cut and transport grass, which is used for fodder. They also do walks with tourists, feed the elephants, apply medicine and are always present whenever tourists are in close proximity to the elephants. It was essential that they were present for the experiments. Our original intention was that they would stand behind the elephant at a distance of around 5 m as it faced the experimental area and have minimal interaction with the elephant. This might have been acceptable if the elephants had understood all of the intentions of the study design, which we found was an impossibility. The mahouts realised before me that they might need to intervene while the elephants were becoming familiar with the experiment area, the table and eating from the buckets. For some readers, it may seem obvious how integral mahouts would be to this study. However, the standard in behavioural ecology is not to focus on humans that may be involved with the study and to reduce their role as much as possible because they would be seen as affecting the choices the elephant made in ways that are difficult to measure. In doing this, it is possible that the importance of their presence to the safety and smooth implementation of experiments is minimised.

15When I viewed the videos of the training phase of the experiment I outlined above, I could clearly see the role the mahouts played in ensuring the safety of the humans and elephants in the experiment area and their attempts to ensure the activities followed the protocol as closely as possible. Therefore, my team decided to study the mahout interventions and elephant responses to them within the training phases. This required us to get their permission to develop detailed ethograms for both the human and elephant behaviour observed so that the videos could be coded and we could address some key research questions. Specifically, what was the elephant doing before the mahout intervened? How did the mahout intervene? And what did the elephant do afterwards? Taking these directions with our research informs our understanding of interspecific communication. This is just one example of including humans who are present in the study design. Other examples could investigate how individual differences and personality (both in elephants and humans), duration of mahout-elephant relationships, modes of interaction, and elephants learning from humans can affect and interact with elephant behaviour.

16In this way, researchers can consider tackling the methodological considerations of designing a study that includes, rather than just controls="true" for, the presence of humans. For example, humans are not just potential sources of bias or misclassification (Tuyttens et al., 2014). In fact, the different ways how, for instance, two people classify the behaviour of the same animal after receiving the same training and protocol is worth studying in itself. I want to use this opportunity to consider how the findings of behavioural ecology research, such as the studies I describe above, not only relate specifically to the experiment in the study but also how they might alter the researchers’ viewpoint and assumptions of their underlying research framework.

17Playback studies have made significant contributions to our understanding of the differentiation between stimuli, and responses to them, which could be related to the perception of risk. This includes responses to humans. For example, Karen McComb and colleagues found that African savanna elephants react differently when they are played recordings of people speaking different languages, here specifically Masaai-speakers and Kamba-speakers (McComb et al., 2014). Masaai-speakers are usually involved in herding and pastoralist activities and come into contact with elephants through those activities compared to Kamba-speakers. The interactions between Masaai-speakers and elephants can become negative, for example concerning access to water and grazing spaces. Men have speared elephants, particularly when Masaai lives have been lost in a previous interaction with an elephant. The agrarian Kamba experience fewer of these negative interactions because of their different land use.

18In a very simplified way, one might predict that Masaai-speaking men would evoke the most defensive reactions from elephants. The researchers replicated predator playback experiments that had previously played lion vocalisations to family groups of elephants, females and their offspring. However, instead of the lion vocalisations, the researchers used playbacks of people of different ages speaking their first language, either Masaai or Kamba, saying the phrase “Look, look, a group of elephants is coming” in a relaxed and clear manner. The researchers then looked for responses in the elephants, specifically, defensive bunching of the group with the calves in the centre. The elephants had a significantly higher probability of this defensive bunching, as well as investigative smelling, following the playbacks of Maasai voices compared to Kamba voices.

19Additionally, these responses were specific to the gender identities and age of the people behind the Maasai voices. The recordings of women and boys, the groups predicted to present a lower threat, were significantly less likely to produce investigative and defensive behavioural responses by the elephants compared to adult men. These results mirrored the researchers’ earlier findings that elephants reacted to red clothing, the colour often worn by Masaai, and the scent of clothes worn by Masaai, whatever the colour. In summary, it seems that there are significant human impacts on wild elephants that we are able to measure by modifying some typical behavioural ecology experimental designs, including playback experiments. It is possible that the design could be considered to lack nuance because there is also variation between voices on an individual level. But these broad patterns are not intended to indicate the level of resolution at which elephants might be able to distinguish between heterospecific vocalisations. They instead suggest the space for further research into attention, differentiation between humans and reactions to humans elephants may have. I note that we may also study variation in elephant vocalisations, which we know are individually distinct (Wierucka et al., 2021), without assessing all of the information that they might encode.

20The topic of this chapter interests me because if the direction of research is changing or the lens is being widened to include humans in behavioural ecology studies of elephants, it gives us the opportunity to consider the implications of this change. This includes questions of our positionality, the theoretical foundation of our studies and the potential tensions with application. Reflecting on these could be valuable for behavioural ecological approaches to studies of elephants because it can both acknowledge that humans are not just sources of bias in studies and encourage collaborative thinking around how we include humans in our work. For example, our inclusion of mahout interventions in how elephants behave in our choice test could be analysed in many different ways to those we decided upon. Our decisions were affected by our experience, training, and the methods we used to collect the initial data in the experiment.

Parallels in conservation

21Many behavioural ecologists have the goal of applying their research to the wider environment, including changing land use, climate, and responses to invasive species or fluctuations in predator or prey species (Bro-Jørgensen et al., 2019). Again, human interactions are often present explicitly or implicitly at different degrees of abstraction in these processes, emphasising that it is imperative to consider the human dimensions of behavioural ecological studies. Furthermore, concerning elephants, it is worth considering what the specific applications of behavioural ecology research are. Oftentimes these link to conservation projects and contexts in which elephants come into contact with humans. These were often confusingly referred to as “human-elephant conflict” (Nelson et al., 2003). This carries the assumptions of adversarial interactions entangled with “conflict” as well as the human vs elephant construct masking the needs, positions and behaviours of many different individuals. More researchers now recognise that the term cannot encompass all the challenges faced in areas occupied by people and elephants, particularly beyond protected areas where biodiversity conservation is among the main management goals. Despite the acknowledgement that a single term will always have limitations, the term “human-elephant coexistence” has now been widely adopted in addition to or as an alternative to the conflict framing (König et al., 2020). This process indicates that there is a level of reflexivity in positioning research, particularly when concerning the applications of behavioural ecology studies.

22There is interesting potential here for that sense of awareness of terminology, and how it links to the interpretation the researcher has of a research area, to be applied to studies that might initially be viewed as basic science and therefore not requiring this level of critical analysis. My reasoning for this is not to suppose that the concepts and human dimensions in behavioural ecology and conservation science are identical. Instead, I aim to point out that researchers in behavioural ecology, through the related conservation literature, may have been exposed to thinking about the position and role of humans in studies. Therefore, this experience offers entry points for expanding research on the role of humans in behavioural ecology. I will also use a final example to illustrate this, that of the conceptualisation of knowledge in ecological studies and how it is created in studies of human and elephant interactions. This can also be linked to the conservation science literature, which has strong ties with behavioural ecology. In conservation science, participatory studies are increasingly part of studies involving biodiversity conservation (Villamor et al., 2014). These can include specifically identifying and implementing methods that focus on integrating expertise from the community in which projects take place, allowing community members to be equitably involved with every level of planning and action, and aiming to create a sustainable plan for action (Nel et al., 2016). This expansion of the concept of knowledge is part of a process by which what is considered to be knowledge of the environment has been critically evaluated and reconsidered. In particular, “local ecological knowledge” is now seen as highly relevant to most conservation projects (Cebrián-Piqueras et al., 2020).

23With the accessibility of this rethinking of knowledge in conservation, it might be the case that behavioural ecology can, with its focus on the environment, better integrate the human dimensions of that environment. That could include the co-production of projects with human participants. In the study on elephant food choice, the mahouts might not initially have been viewed as participants, but they were central in the ability of elephants to participate at all. Articulating their role is not a detraction from the study but a valuable addition that allowed us to consider questions beyond the original study. These include how mahouts intervened, and what proceeded and followed the intervention in terms of elephant behaviour, getting to a fine-scale analysis of interaction as well as a binary choice an elephant made between buckets. However, the scope goes far beyond this, for example, in correlations in personality metrics between elephants and humans they spend time with, the speed of behavioural response to cues from people they do and do not know, and how interactions between humans and elephants could be associated with welfare indicators. We can also avoid the unplanned developments in my group’s study that I outlined earlier in this chapter by considering co-design of study aims with people who will be involved in the project. In doing so, we can take the lead from fields in which participatory approaches and co-production are part of the research toolkit and use them in the conceptualisation of our studies.

24Furthermore, I suggest to behavioural ecologists that we carefully examine our widely understood ideas of wild animals, natural behaviour and settings, experimental design and distancing humans from the observations or experiments. In doing so, we can consider how they impact the design, application and direction of behavioural ecological studies, and potentially improve both the clarity and scope of research. I have highlighted the use of conservation research to do this, as it is often accessible in terms of literature to behavioural ecologists, but the possibilities go far beyond that, as the diversity of fields in this volume illustrates. Elephants are an excellent example because of the wealth of research on them, in combination with the different human interactions they experience, including relationships such as those with mahouts. By using such examples, we can support other researchers by transparently reporting study design when humans are involved, proposing questions that include or centre humans as part of the environment, and sharing methods of studies that are co-produced.