Habituating Meerkats and Redescribing Animal Behaviour Science

Introduction

This paper examines two kinds of practices. The first is the scientific practice of habituating animals for the purposes of research. Canonical scientific accounts of habituation usually describe it as a technical procedure designed to get animals to ignore their human observers, a sort of behavioural alternative to erecting a ‘hide’ (Lehner, 1996). However, a number of popular writings by primatologists (e.g. Smuts, 2001), which have themselves become the object of social scientific and philosophical analysis (Haraway, 1989; Despret, 2002; Rees, 2007), have suggested that habituation pushes the boundaries of classic scientific methodology in an intersubjective direction. This article draws on ethnography at the Kalahari Meerkat Project in South Africa, where field behavioural ecologists work with habituated meerkats to examine the complex balancing act involved in the set-up and maintenance of habituation. Does habituation in fact produce a distinct type of inter-species relationship, is it a version of other kinds of relational practices such as domestication or companionship, or is it not really conducive to relationship at all, but simply a behavioural mechanism for making humans ‘uninteresting’ to non-humans? And more generally, how might anthropologists or sociologists of science make sense of the seemingly contradictory demands and expectations placed on habituation by practising (and aspiring) scientists?

The second practice I will examine takes place on a rather different plane. It is the practice of philosophers and sociologists of science who have in the past few decades been seeking to move beyond the unproductive debates of the science wars, to articulate a new kind of relationship between science and science studies. Scholars such as Isabelle Stengers (2010b), Vinciane Despret (1996) and Bruno Latour (1999) have reflected in various ways on how studies of science might move beyond deconstructive critique, learn to take scientific practice seriously and attend to what matters to scientists themselves. Nevertheless, these authors are also committed to restating the value of scientific practice in ways which break with canonical epistemological understandings – which are often those of scientists themselves. Stengers (2010b) has given a highly nuanced account and defence of this type of philosophical practice which she describes as ‘counter-effectuating’ science: seeking inventive ways of highlighting the value of science as a particular kind of adventure, which simultaneously deprives it of the power to rule out of court other adventures (philosophy, witchcraft, mountaineering …). In ‘counter-effectuating’ it, philosophy diverges from science but remains, Stengers argues, in a productive, vital relationship to it.

These two practices (scientists habituating animals; philosophers counter-effectuating science) operate on very different practical and conceptual planes. Yet both focus on the problem of establishing relations while recognizing significant otherness (Haraway, 2003). Both are concerned with the balance between describing and transforming. Whilst I will argue these practices diverge in significant ways, they can for that very reason shed light on each other.

The two practices are also ‘genetically’ related in my own work, as an anthropologist studying field biologists who habituate meerkats, while reading philosophical accounts of science (such as those by Isabelle Stengers or Vinciane Despret) which recast these scientific practices in a very different conceptual language. Understandings of what is at stake in habituation are a key point where the scientists I work with and the philosophers I read are most visibly diverging – a key point therefore for examining the intersection of these two practices.

As I will show in more detail in the next section, Stengers and Despret have developed a far-reaching philosophical account of scientific practice which, in the case of the sciences of animal behaviour, foregrounds the power of attunement and productive mutual modification between researchers and animals as a simultaneously ethical and epistemic goal. Despret refers to such ‘practices that create and transform through the miracle of attunement’ as forms of reciprocal ‘domestication’ (Despret, 2004: 125).

By contrast, researchers and volunteers at the Kalahari Meerkat Project, whose work I have been studying ethnographically, practise habituation in line with the canonical scientific account of this practice as a technical procedure designed to get animals to ignore their human observers. While they recognize that habituation involves transforming the animals’ behaviour in some respects, this transformation is merely a necessary problem whose effects can and must be controlled. While many of the volunteers working to collect data at the project personally enjoy their relationship to the meerkats as individuals, they consider this to be at best an accessory, and at worst a potential impediment, to the business of ‘science proper’. In particular they are profoundly resistant to the thought that what they are doing with the meerkats is in any sense comparable to ‘domestication’.

There is thus at the very least a terminological dissonance between the way these people conceive of their work and the way such work is redescribed by philosophers such as Despret – but I will argue this dissonance points to rather more interesting and profound differences between habituation and counter-effectuation, which in turn illuminate some key divergences between ways of engaging with significant otherness in the particular branches of science, philosophy, and anthropology examined here.

A Philosophy of Attunement, Interessement and Domestication

Recent work at the intersection of science studies and animal studies has suggested that close, sustained interaction with non-human animals produces a type of insight which is not reducible to the classic canons of scientific knowledge-production. Some authors praise and draw attention to the kind of fine-grained, relational knowledge accrued by nonscientists working with animals day in and day out, and contrast this type of close, intimate knowledge to the cautious and distanced knowledge production of behavioural scientists (Midgley, 1988; Cox and Ashford, 1998). Other authors have sought for alternatives within the practices of animal scientists themselves. In this vein, science studies scholars such as Donna Haraway (1989, 2008) and Vinciane Despret (1996, 2002, 2004) have traced the material, fleshly and affective entanglements between scientists and the animals they study in both the lab and the field.

These explorations are in tune with a broader move in science studies which recasts scientific practice in terms of ontic engagement with the world rather than conceptual detachment from it (Stengers, 2000, 2010a, 2011). As Isabelle Stengers notes, this particular vision of science, while it reclaims a distinction between subject and object which was too easily dismissed by relativist or social constructionist analyses, ‘does not attribute to the subject the right to know an object, but to the object the power (to be constructed) to put the subject to the test’ (2010a: 133). Experimental scientists’ specific power comes, for Stengers, from their ability to ‘tie a knot’ between the subject and the object, a knot which ‘links these kinds of humans, endorsing very strong specific obligations, and those kinds of phenomena, or framed aspects of “nature” verifying very selective requirements’ (2010b: 48).

The asymmetry between ‘obligations’ and ‘requirements’ is a crucial feature of Stengers’s philosophical terminology (2010a: 49–55). Obligations apply inwards, to those who reclaim themselves of a practice. Requirements apply outwards, to the milieu in which this practice exists. Outwardly, experimental science requires the existence of phenomena – the ‘objects’ – which can be purified and mobilized. Inwardly, it carries certain obligations for its practitioners, notably the obligation of carefully distinguishing between reliable experimental facts, where the object itself has really played a role, and the seeming results of experiments which are simply products of the experimenter’s own operations – what experimenters usually call artifacts. ² Together, these inward-facing obligations and outward-facing requirements ensure that the object has been given every chance to ‘object’ to the theories and assumptions of the scientists.

This criterion for good science – allowing objects to object – is what Bruno Latour in a somewhat tongue-in-cheek fashion has described as the ‘Stengers-Despret shibboleth’ (Latour, 2004). Of course, this general touchstone means different things for different sciences, since an atom does not object in the same way as an electron, a cell, a rat, a human, or a community. Stengers’s two-volume work Cosmopolitics (2010a, 2011) traces these different objects and their sciences, considering in each case what it might mean for an object to object.

A particular threshold is reached when Stengers’s discussion comes to animals, and later humans – objects of science which are capable of being ‘affected’ by the questions scientists ask them. ³ Such ‘objects’, for Despret, make it more difficult to maintain the distinction between requirements and obligations. Animal and human objects, unlike atoms or cells, are more easily affected by ‘obligations’ (rules and constraints imposed by experimenters) and may find it harder to impose their own ‘requirements’ on the scientist. This is the point at which Stengers’s account rejoins Despret’s, and also Bruno Latour’s. For all three authors, the achievements of the sciences of human and animal behaviour are often suspect, since they can derive from impositions of experimental obligations on ‘objects’ who are unable to resist the authority of science – either because, like rats in a maze, their actions are directly constrained, or because, like the human subjects of sociological or psychological explanation, they are too awed by the authority of the social scientist to resist the explanation (Latour, 2004, 2005; Stengers, 2010; Despret, 2002, 2004). ⁴

As Latour (2004) notes, this particular way of reframing the difficulties of the human and animal sciences retains something of Popper’s earlier criterion of falsifiability (has the object really been given a chance to put our theories to the test?), but redeploys it in a radically different direction. Indeed the solution proposed by these authors is not to seek to reproduce the objective reliability of experimental science by further objectifying humans or animals, but on the contrary to change tack and to move to a different form of enquiry, to different practices for producing facts and proofs (Stengers, 2011). Since their ‘objects’ are able to be interested and affected by the studies, human and animal scientists should, according to Stengers and Despret, put this ability at the very centre of their work.

Despret in particular has articulated what this would mean in the case of animals: an animal science which makes ‘attunement’ between the bodies of the animal and the researcher into a positive research method, rather than a cause for epistemological anxiety and bad data. Despret reflects on the classic case of ‘Clever Hans’, the horse who seemed to be able to count by sensing subtle physiological cues to the correct answer in the body of any person who asked it the question (2004: 111–16), and on the equally classic experiments of Rosenthal (1966), which demonstrated the power of human expectations to shape the outcome of experimental animals’ abilities. Rosenthal showed that students who believed their rats were from a specifically ‘clever’ strain unconsciously helped them to perform better in tests than those who thought they were testing ‘normal’ rats (Despret, 2004: 117–120).

These two examples are amongst the classic ones invoked by researchers who work with animals, to highlight the dangers of ‘influence’. Despret, however, wants to draw the very opposite lesson: instead of seeking to reduce or eliminate this influence, scientists should on the contrary build on animals’ uncanny ability to attune their bodies to ours, to transform our behaviour along with their own – not unlike horse-riders do, for instance (2004: 115). Despret finds in Konrad Lorenz’s ethological work with domesticated geese and rooks the epitome of this type of productive mutual modification. Against the commitment (embodied for instance by ethology’s other founding father, Niko Tinbergen) to studying undisturbed animals in the wild, Lorenz offers a figure of the virtues of domestication in which human and animal ‘become together’ (2004: 128ff).

Domestication, argues Despret, allows scientists to find new ways of ‘interesting’ the animal in the research, of asking the animal what it is interested in, thereby avoiding the risk that the animals are simply following rules set by the experimenters. This notion of interest as ‘being-with’ (‘inter-esse’) leads Despret to articulate an extended notion of ‘domestication’ – ‘practices that create and transform through the miracle of attunement’ (2004: 125) – as the key to better animal science.

Habituation at the Kalahari Meerkat Project

On the face of it, such philosophical celebrations of scientific attunement might seem to be congenial to scientists who study animal behaviour in the field, under conditions of habituation. A number of primatologists, in particular, have explicitly suggested that habituation is at heart a process of mutual modification – and their writings have fed into the discussions above. For instance, primatologist Barbara Smuts, describing her initial attempts to ignore the baboons she works with and thus get them to ignore her, noted:

the baboons stubbornly resisted my feeble but sincere attempts to convince them that I was nothing more than a detached observer, a neutral object they could ignore. Right from the start, they knew better, insisting that I was, like them, a social subject vulnerable to the demands and rewards of relationship. Since I was in their world, they determined the rules of the game, and I was thus compelled to explore the unknown terrain of human-baboon intersubjectivity. (Smuts, 1999: 109–10)

More generally, and beyond primatology, Gregory Radick has argued that the central commitment of mid-century ethology (the disciplinary precursor of contemporary behavioural ecology) was precisely its attention to what mattered to animals: ‘To be for ethology was to be against comparative psychology, or at least against the human-oriented perspective that, according to the ethologists, kept psychologists from investigating the behaviors that mattered most to the animals’ (2007: 256). This commitment was preserved, Radick argues, in the more recent recourse to field experiments through which field behavioural scientists such as Cheney and Seyfarth ‘had taken themselves into the animals’ world in order to learn’ (Radick, 2007: 364; cf. Cheney and Seyfarth, 1990).

The Kalahari Meerkat Project researchers whose work I have been studying ethnographically in recent years share in the broadest sense this commitment to the general principle of learning from animals in situ rather than in the rarefied controlled conditions of the lab. However, this commitment does not translate into a straightforward embracing of ‘demands and rewards’ of human-animal relationship as a source of scientific insight. On the contrary, Kalahari Meerkat Project researchers are rather ambivalent about habituation’s potential for mutual modification, and highly resistant to any conflation of habituation with domestication.

The Kalahari Meerkat Project was set up in the early 1990s by behavioural ecologist Tim Clutton-Brock, who started off observing a few groups of meerkats from a car at first, and then at increasingly close range, until the animals came to be comfortable with a researcher walking amidst their group as they foraged through the Kalahari scrub (for an account of these early days see Clutton-Brock, 2008). Over the years, the project grew and began to enrol volunteers (mostly UK biology graduates), each of whom would spend a year at the remote farmhouse in South Africa from where the project was run. The number of meerkat groups followed by the project grew in proportion. By the time of my second field trip to the Kalahari Meerkat Project in 2011, the project involved around 15 volunteers following 19 groups which ranged over the considerable territory of the reserve and neighbouring farms.

The bulk of the volunteers’ time at the project was spent in the observational collection of behavioural and life-history data. The project also collected weight data, which required that every meerkat in the population be weighed (both before and after their morning foraging session) at least twice a week.

Habituation was thus central to the Kalahari Meerkat Project, but it was usually described in terms very different to those articulated by Smuts or Despret. A representative account of habituation was given by Lisa, a biology graduate who had extensive experience training wild horses and ponies, using both classical methods and those of ‘natural horsemanship’, and had recently joined the station to assist with habituating ground squirrels for a new project. Lisa reflected on the difference between her previous experience and her current job:

With the horses, it’s very much trying to create a bond, like some kind of relationship where you get them to do something you want them to do, and you want them to like you, you want them to enjoy … coming up to you. […] Whereas squirrels, you want to be in the background. You want, basically, to be a tree. But a tree that gives food for rewards. So, you kind of want to be able to weigh them, but you want, as soon as the weighing process is finished, for them to not care about you, so you can just walk through, basically.

Someone attempting to fit the Kalahari Meerkat Project into the philosophical mould proposed by Despret would find contradictions abounding in the volunteers’ speech: on the one hand, volunteers were very scathing in their accounts of members of the public who came to visit the site and who wanted to pet or cuddle the meerkats. On the other hand, when asked which meerkats they liked most and why, the volunteers I met would often describe their favourite meerkats as those who ‘are really friendly’, who ‘come up to you’ or even ‘groom your hand’. Even though some of the project organizers might have frowned on such displays of affection, for these volunteers they were not taken to interfere with or challenge the otherwise detached, scientific nature of their relations with the meerkats. Indeed, the same volunteers would just as often describe their pride and happiness when meerkats ignored them. Thus Annie, an experienced volunteer, recalled:

So one time I was at Godzilla ⁵ and they were war-dancing ⁶ Toyota. And they war-danced at me … past me, and I was like ‘Oh Yeah!’, they didn’t care that I was there! So that was good.

Habituating Meerkats, Training Volunteers

… it’s a lot of give and take. (Annie, interview)

Upon their initial arrival at the project, junior volunteers were trained in data collection with fully habituated groups by shadowing a more experienced volunteer for two weeks. It is only much later that they would ‘graduate’ from data collection to habituation. However, since habituation was by definition a solo endeavour (being able to tolerate two people was already taken as a mark of a pretty well-habituated group), there could be no literal ‘training’ process in which volunteers were shown how to habituate by example. As Phil, another experienced volunteer, put it:

I guess the training process is by going to slightly harder groups as you become more experienced. And just learning how to handle difficult meerkats, and eventually you learn how to handle a difficult group.

With Despret’s work in mind, the whole conjoined structure of training and habituation could be aptly described as a large-scale process of ‘attuning’ inexperienced humans to wild meerkats until both parties had become, together, something else entirely: experienced volunteers and habituated meerkats.

In sum, while the outcome of habituation could be described by participants as non-relational (marked by the fact that meerkats ignore you), the process which would get you and the meerkats to that point was definitely and unavoidably both interactive and mutually modifying, in a way which dovetails neatly with Despret’s discussion of ‘becoming together’.

Valuing Disinterest

And yet Despret’s extension of the term domestication to cover scientific relationships of attunement could not have been more profoundly at odds with volunteers’ own sense of what habituation was about. When, after they had described the process of habituation, I asked Jo and Annie directly how habituation differed from other kinds of relations such as pet-keeping or domestication, Jo was nonplussed, and her responses became hesitant, as she was trying to fathom what I could possibly mean.

Matt: So what would be the difference with like, domestication, or pet-keeping, or being a farmer or something like that?

Jo: What’s the difference between … habituation and … ?

Matt: … and that kind of stuff.

Jo: Well [long pause] … keeping a pet is for your own companionship a lot of the time. You’re not observing your pet … it’s a different, it’s a whole different thing. […]

Annie: I feel like with farms … 

Jo: Well they have a different relationship again, because they want to keep their animals in good healthy condition but it depends what the animals are for. Like for cattle, are they for milk, are they for meat … […] they want to keep them in nice environments, nice conditions, for their benefit, again.

Annie: Yeah cause you’re only raising it to take something from it … 

Jo: Yeah, so it’s not kind of … it’s a different … relationship.

Annie: We’re not raising meerkats, we’re habituating them to learn from them, not to like … take their milk or something. [laughs]

Jo: […] yeah … I think they're hard to … to relate those things, because they're so different.

The contrast between volunteers’ keenness to distinguish habituation from domestication and Despret’s conceptual extension of the term ‘domestication’ to cover all sorts of ‘practices that allow themselves to be pervaded by humans’ (Despret, 2004: 125) might seem to be a minor detail, almost a coincidental clash of terminologies. But I would argue it points to a rather profound conceptual dissonance. A crucial focus of the dissonance between Despret’s account and the volunteers’ own understanding of what they were up to surrounds the notion of ‘interest’.

Speaking ethnographically, for Jo, Annie and others at the meerkat project, what marked out habituation from other types of human-animal relation – what made it a specifically scientific relation – was a very particular economy of different kinds of interest and disinterest. Clearly, habituation is entirely driven by the interest of the scientists in the meerkats. However, as outlined in the above interview excerpt, it mattered what kind of interest. Habituation is, crucially, not motivated by an economic, vital or emotional interest but an intellectual interest (curiosity, the desire to answer specific theoretical questions). This sets apart the volunteers, in their own eyes, from fans of the show Meerkat Manor, for instance, who were ‘interested’ in the meerkats in what the volunteers saw as an exclusively personal or emotional way.

Furthermore, whereas farming or pet-keeping operate by attaching the (vital, emotional, economic) interests of humans to the interests of nonhumans, habituation, by contrast, should precisely not arouse the interest of the meerkats. Behaviourally, the meerkats must remain visibly uninterested in what the scientists are doing for their behaviour to count as natural. Vitally, the meerkats must be able to pursue their own independent interests – which the researchers, in a Darwinian frame, assumed to be principally those of reproduction and survival – for their lives to be valuable as tokens of ‘nature’ from a research point of view. In sum, habituation can only serve the interests of the scientists if it ensures the meerkats are left to fend for themselves and pursue their own interests unimpeded.

These ethnographic distinctions at the Kalahari Meerkat Project between different types of interest clash with Despret’s and Stengers’s defence of ‘interessement’, which draws on a long tradition in science studies and ANT, which purposefully and self-consciously collapses what are often held to be very different (indeed radically opposed) types of interest: epistemic, intellectual, political, economic, vital, etc. (Stengers, 2007: 11; Callon, 1986). This is of course no coincidence, since this analytical tradition precisely seeks to overturn standard Euro-American or ‘modern’ distinctions. But it does create something of an interpretive problem for the anthropologist, caught between people’s own understandings of what they are up to and a literature which seems to reject those understandings out of hand.

One option, as ever, is translation. For instance, one might point out that Stengers’s distinction between requirements and obligations is partly aiming to recapture some of what was contained in classic defences of ‘disinterestedness’ in science – particularly in relation to the instrumentalizing of science by state and capitalist ‘interests’ – while recasting these aspirations as part of a general defence of diversity and independence of practices, rather than as a special privilege of science over other practices (see for instance Stengers, 2010b). What used to be termed ‘disinterestedness’ is recast as a genuine concern with the integrity of the experimental achievement which marks out ‘good science’: that a knot has really been tied between the object and the subject. What used to be ‘disinterestedness’ thus becomes doubly relational, doubly about interessement: it is the mark of what interests scientists, what ties them together as a community of practice; it centres on the achievement of a particular kind of interessement between the subject and the scientific object.

So far one could still imagine a way of translating this philosophical refiguration of interest and the Kalahari Meerkat Project’s understandings of habituation. The volunteers’ sense of the distinctiveness of habituation could be cast in terms of Stengers’s notion of experimental achievement. Habituation, in effect, could be seen as an attempt to tie a knot between researchers ‘endorsing very strong specific obligations’ and meerkats as ‘phenomena, or framed aspects of “nature” verifying very selective requirements’ (Stengers, 2010b: 48). This in turn – the concern with ensuring that this knot has been tied in the right way, that habituation has not slipped into domestication, for instance – is what binds together Kalahari Meerkat Project researchers as a community of practice.

But such a translation would elide what is nevertheless an important difference between the volunteers and Despret or Stengers – the difference which leads Despret to speak of ‘domestication’ and the volunteers to reject the term. The difference concerns the crucial question of what counts (or rather what should count) as ‘the object’ in this case. For Stengers and Despret, scientific research of the kind undertaken at the Kalahari Meerkat Project is research with animals, entities which can, and therefore should, be interested in the research – as opposed to atoms or cells, say, who are, as Latour puts it, ‘bona fide natural objects […] utterly uninterested by the inquiries’ (2004: 217, see note 4 below). As I wrote above, to attempt to recreate the classic form of the experimental achievement in these conditions is, for these philosophers, wrong (in a simultaneously ethical and epistemic sense). For the Kalahari Meerkat Project researchers, on the other hand, the experimental achievement is possible, simply because – somewhat paradoxically – meerkats are not the object.

What Is the Object? Meerkat Epistemic Things

While my initial sketch in the previous section began to show the way habituation articulated interest and disinterest, it still glosses over one major complication. In practice the habituation of meerkats was a combination of two very different kinds of process. It aimed both to ‘interest’ and to ‘disinterest’ meerkats in what the researchers are up to.

The somewhat surprising image of ‘a tree that gives food’, articulated by Lisa above, highlights this split at the heart of habituation. With the meerkats as with the squirrels, habituation had in practice not one but two very different aims. The first was to get the animals to ignore the researchers and take them for granted, allowing the latter to fade into the background while taking down behavioural observations from a close range. The second – in some respects opposed – aim was to find a way to interest the animals enough in the research process so that they could be easily weighed by researchers in the field.

Thus, the project had established a system whereby the meerkats were enticed onto the scales with the help of a water-bottle and crumbs of boiled egg, and getting the meerkats used to (some jokingly said ‘hooked on’) egg in the first place had become a sub-set of habituation. Researchers thus would sprinkle boiled egg around the meerkats’ burrows, and took pains to mix boiled egg in with foods which the meerkats seemed to favour, such as scorpions. Pups born into habituated groups were a particularly important target, and ‘pup habituation’ involved painstaking efforts in getting pups to like egg.

Habituation thus had two mutually opposed aims: getting the meerkats interested in egg and thus amenable to weighing, and getting them uninterested enough in the researchers so that their ‘natural behaviour’ could be observed. Considerable effort went into keeping these two aims of habituation scrupulously separate. Firstly, behavioural data was never normally collected during weighing sessions (see Candea, 2010). Secondly, the volunteers had developed two separate calls to mark the distinction between feeding and non-feeding periods: a general habituation call (a three-note hum) was used every few minutes when walking with the meerkats, upon initially approaching them and whenever one of them seemed startled. A different call was used, however, whenever the meerkats were being weighed and therefore rewarded with egg and water (the call was the phrase ‘yum-yums’, said with a specific emphasis).

In practice, however, the meerkats were rather smarter than the behaviourist conditioning model predicted. ‘Yum-yums’ or not, many came running whenever researchers put down their bags or crouched down – which as Annie and Jo noted wryly, made peeing in the field something of an ordeal for female researchers. Furthermore, as a result of ‘feeding experiments’ – in which the diet of selected individuals was supplemented over a period of time with daily doses of egg – fed individuals were even more likely than others to stop in the middle of what they were doing and look up at the volunteers, with what the latter took to be an expectation of being fed. In order to counteract this, at the time of my second field trip in 2011, the researchers were experimenting with a new form of sound cue: a rattle which they would sound whenever they were feeding the experimental individuals. The hope was that the specific fed individuals would associate the rattle with feeding, and thus the absence of rattle with ‘normal’ inter-patient relations.

In sum, the properly habituated meerkat should ideally be a double creature: both interested and uninterested. Reciprocally, while habituation aimed to produce double meerkats, training produced double volunteers. Volunteers often spoke of being able to embody a ‘scientific’ perspective and a ‘personal’ or ‘emotional’ one (Alcayna-Stevens, 2009, 2012; Candea, 2013).

The aim of these various methods of separation (interrupting data collection, making different calls, etc.) was to preserve a distinction between affected and unaffected meerkat behaviour. Within the complex landscape of mutual modification introduced by habituation and experimentation, volunteers had thus learnt to make fine-grained distinctions between actions which involved humans but could still be counted as unaffected natural behaviour (say a meerkat using your shoulder as it would a rock as a convenient guarding-spot) and behaviour which was problematically interactive (say between a meerkat coming towards you with expectations of being fed). Thus, just as I was discussing this question of mutual modification out in the field with Matt, a volunteer who had been at the site for a few months, the dominant female, a fed individual, suddenly stopped to look searchingly at him. Matt pointed to her and admitted that that was ‘not natural behaviour’. Then, when a second later she resumed her foraging, he added that the behaviour was natural again, and duly resumed data collection.

That particular instance crystallized the overall lack of epistemic worry as the animal switched between these two modes of affected and unaffected behaviour. Surely, I wondered, putting myself in the researcher’s shoes, if the meerkat could alternate so swiftly, then their behaviour overall had been drastically modified. How, given his explicit concern with natural behaviour, did Matt (and the other volunteers) escape the anxieties of ‘influence’, the paradox that habituation modified the very thing it purported to study, namely natural behaviour?

And then, out there in the field with Matt, I suddenly realized why my problem was poorly framed. In pointing out the alternation between natural and non-natural behaviour, Matt was not overly worried about the question of whether the meerkat as a whole was behaving naturally in every respect, because he (and the project more generally) was not studying meerkats but rather a specific subset of meerkat behaviour. In the process of studying these particular behaviours, modifying other behaviours was unavoidable. It was also utterly unproblematic, as long as what was modified (meerkats’ liking for egg, their tolerance of humans) was kept scrupulously distinct from what was being studied (meerkats’ cooperative behaviour, their survival chances in the face of predation, etc.). Meerkats were not one thing, or even, as I suggested above, two. Rather, when it came to scientific research, meerkats were internally multiple.

One might couch it in historian of science Hans-Jörg Rheinberger’s terms (1997: 28–9): the ‘total’ behaviour of meerkats was subdivided into ‘epistemic things’ (the entities which formed the focus of the enquiry) and ‘experimental conditions’ (the stabilized background which enables the investigation). The content of this distinction shifted within a single field-session: during ad lib, the epistemic thing was a certain set of predefined meerkat behaviours (including guarding, inter-group dominance, pup-feeding, etc.), while the experimental conditions included meerkats’ ability to be uninterested in the researchers (together with a number of other things such as the handheld psion computer, the dye-marks which allowed researchers to recognize meerkat individuals, the researcher’s training, etc.). During ‘weights’, however, the ‘epistemic thing’ was the meerkats’ mass, while the experimental conditions included the meerkats’ ability to be interested in egg and water, and their willingness to be handled (together with the scales, the weights notebook, and once again, the meerkats dye-marks, the researcher's embodied skills and training, etc.). This distinction, in turn, drives the double aims of habituation as interessement/uninteressement, making meerkats available both as interested collaborators (for weighing) and as uninterested objects (for ad lib).

The concern with keeping epistemic things separate from experimental conditions is in itself not particularly surprising or new. It is an instance of a very general scientific principle, convincingly articulated by Rheinberger (1997). But this distinction gives a new handle on the logic of habituation. Defined as a single relationship between a whole individual animal and a whole individual human, habituation seems hopelessly contradictory and incoherent. These contradictions are resolved, however, if we consider habituation as a shifting bundle of experimental conditions, geared to isolating certain epistemic things within the animal’s behaviour.

Comparing Habituation and Counter-effectuation

Of course we actually (and almost unduly, but it is for the sake of the world) completely reverse Rosenthal’s aim. (Despret, 2004: 132)

For the researchers who set up the project, meerkats are an interesting ‘system’, insofar as they represent a particularly clear instance of a more general type of animal society known as ‘cooperative breeders’. However, the researchers are unambiguously clear that the fact they are working on meerkats is, ultimately, incidental to other broader interests. Meerkats are not the object. They are simply a good model for asking empirical questions driven by particular ‘advances in theory’, as Krebs and Davies put it in their introduction to the influential handbook of behavioural ecology (Krebs and Davies, 1997). Meerkats as a whole might engage individual Kalahari Meerkat Project volunteers as subjects in interaction. However, the scientific object which compels the Kalahari Meerkat Project researchers, which puts them to the test, is not the furry critter itself, but a more uncanny entity made up of bits of its weight, environment and ‘cooperative behaviour’.

The realization that the meerkats are not the object of study rearticulates what might otherwise seem to be contradictions at the heart of habituation. It explains why researchers can be relatively free from the anxieties of influence described by Rosenthal, without thereby espousing the model of science as mutual modification which Despret proposes as an alternative. It shows how it is possible to both distance and interest meerkats, to both modify their behaviour and observe it in a state of nature, to enjoy and seek both their friendliness and their disinterest, to care for them both as subjects and objects, both as characters and as bearers of data.

By contrast, Despret’s and Stengers’s account of good animal science as domestication draws implicitly on a reference to the animal as a whole organism as the object of science. The rat which, unlike the electron, ‘knows’ pain, the animal which is capable of being ‘affected’ and attuned, is the whole animal, as a potential bearer of perspective. This is not to say, however, that Stengers’s and Despret’s commitments are holistic in any straightforward sense – like those, say, of early 20th-century naturalist precursors of behavioural biology who wished to peek into the animal’s subjective world (Crist, 1999). Indeed Stengers explicitly eschews both holism and reductivism for a thought-provoking third option, a sort of performative commitment to relational recognition: wholes are wholes in relation to a question or interest eschewed by other entities (Stengers, 2011). The ability to have a perspective is itself a relational achievement.

And with this reference to performativity, we have finally reached the core of the dissonance between the two positions. The Kalahari Meerkat Project researchers’ commitment to habituation is centrally motivated by the problem of accurate knowledge of a pre-existing reality. By contrast, Despret’s and Stengers’s aim is performative. They cannot be accused of ‘missing the point’ when they treat the animal as a whole as the proper object of animal behaviour science. Rather, they are attempting to nudge scientists towards taking as their object whole animals (as potential bearers of a perspective) rather than bits of animal behaviour – an aim which many (including many in the scientific community) will consider to be eminently worthwhile.

In other words, treating Despret’s or Stengers’s account of good science as a mistaken description of the empirical reality of scientific practice would itself be a mistake. Elsewhere, Despret explicitly describes her philosophical approach as precisely one of carefully ‘sliding’ some meanings closer to others, by small successive ‘betrayals’ (Despret, 2009: 26). Thus, for instance, Despret’s expansion of the term ‘domestication’ self-consciously flattens certain differences between scientific and non-scientific ways of knowing and relating to animals – in particular the distinction (crucial for the researchers) between different types of ‘interest’ – in order to elaborate other ethico-epistemic distinctions between ways of doing science with animals, which stem from a particular philosophical tradition which includes A.N. Whitehead, William James, and Isabelle Stengers. There is a neat recursion here: indeed Despret’s argument for domestication and mutual modification between scientists and animals doubles up as a proposal for the relations between scientists and philosophers such as herself. Despret’s own account of what scientists do aims to ‘domesticate’ them in a sense – to attune itself to their practice while simultaneously attuning their practice to her philosophical commitments. The commendable hope which runs through Despret’s work is that her ‘betrayals’ will remain ‘shareable’ (2009: 26).

The same recursivity can be found in the work of Isabelle Stengers, through the figure of what, following Deleuze and Guattari, she calls a ‘rhizomatic connection’, or a ‘marriage against nature’:

We may recall the famous example given by Deleuze of the ‘noce contre nature’, between the wasp and the orchid. Their connection is an event that matters in diverging ways for the wasp and for the orchid. Its achievement is not to lead the wasp and the orchid to accept a common aim or definition, but having the wasp and the orchid presuppose the existence of each other in order to produce themselves. (Stengers, 2007: 14, emphasis added)

Of course, the way to do this is precisely – and once again recursively – to provide an account of ‘good science’ as (much like philosophy) an inventive rather than a descriptive practice: as a practice which very effectively generates new realities, and does so in its own very distinctive way, but nevertheless doesn’t have privileged access to ‘reality’ as such. Cue Stengers’s insistence that ultimately ‘any experimental fact is an artifact, a fact of art, a human invention’ (2010b: 50; cf. note 2 below). Philosophical counter-effectuation is, in sum, the commitment to recursively re-inventing good science as itself inventive.

In sum, the two practices I have been examining in this paper – counter-effectuation of science as practised by Stengers or Despret and habituation of animals as practised by Kalahari meerkat researchers – deploy similar elements and concerns but articulate them in radically different ways. Both practices are concerned with establishing relations between entities with diverging aims and modes of existence. Both practices are concerned with the interplay of description and transformation. Both bear simultaneously on epistemic, ethical and ontological matters. Both practices are initiated by one party (habituation by scientists, counter-effectuation by philosophers). But where counter-effectuation seeks to articulate together entities which presuppose each other’s existence, habituation is crucially about ensuring that one entity (the meerkats) does not come to require the other (the scientists) for its survival. Where counter-effectuation seeks to subtly transform the very thing it describes, habituation seeks to separate what is transformed from what is described. Where counter-effectuation treats (re-)description as a performative, transformative tool, habituation treats (carefully controlled) transformation as a means to accurate description. Where counter-effectuation is recursive (purporting to establish with scientists the same kind of relations it describes good scientists establishing with their objects), habituation breaks with recursivity: the relations between researchers and meerkats must be of a different order to the relations meerkats establish with other elements of their environment.

Putting habituation and counter-effectuation side by side in this way highlights the profound divergences but also the partial commonalities between the practice of the particular scientists I have been studying and the particular philosophers I have been reading – albeit certainly not of ‘Science’ and ‘Philosophy’ writ large, whatever that might mean. The clear delineation of these positions is partly heuristic, rehearsing the classic anthropological bifurcation between ethnography and theory (cf. Strathern, 2011). The reflections and commitments of Stengers and Despret pull the analysis in one direction – away from the standard realist version of ‘representation’ – while my ethnographic encounter with scientific habituation pulls in the opposite direction. The anthropological challenge which I have tried to take up here is to consider both and relate them, without reducing the one to the other. A non-representational account of habituation (if I used the theory to ‘explain’ the ethnography) would extinguish the meaning this practice holds for the Kalahari Meerkat Project researchers themselves, just as a representational account of counter-effectuation (if I used the ethnography to ‘critique’ the theory) would miss the point which these authors are making. The striving for a type of double vision – seeing ethnography with one eye, as it were, and theory with the other (cf. Corsín Jiménez, 2013) – is in turn the mark of a third practice, which is neither quite scientific nor quite philosophical, namely anthropology.