Mariam Motamedi Fraser
Search for other papers by Mariam Motamedi Fraser in
Current site
Google Scholar
Dog disputes
Scientific research with dogs

Previous chapters have addressed the serious trouble that dogs’ species story makes for dogs. Chapter 5 turns to the trouble that the story makes for scientists – for the very scientists who are writing it. It puts three related, but differently charged, issues into conversation with one another: dogs’ perceived relationality, as it is defined as a methodological problem in science; relationality, as it is defined as the foundation of animal capability, agency and resistance in nimal studies; and the contested place of singular individuality in both. This chapter shows how Vinciane Despret’s model of ‘polite research’ is differently relevant to those scientists who support dogs’ species story (and who are therefore obliged to grapple with the methodological quandaries that are perceived to be raised by dog–human relationality) and to those who contest it (and who foreground dog individuality in contrast). It also argues, however, that neither polite research, nor the responses of these scientists to the problems posed by dogs, offer much in the way of the undoing of dogs’ species story, nor can they wholly account for how dogs might be enabled to object to the questions that are posed to them by science. The final section of the chapter analyses Martin Seligman’s ‘learned helplessness’ research, the brutality of which served, inadvertently, to draw attention not only to the relevance of intersubjectivity, but also to the dogs’ irreducible singularity, which had the power, at least momentarily, to interrupt, perhaps even to disrupt, the demands of the experiments.

Because human beings and dogs are aspects of each other, empirical analysis of the relationship may be confounded.

(Paxton 2011: 5)

Is contact with humans a necessary part of what it means to be a dog? No.

(Pierce and Bekoff 2021: 163)

In Chapters 2 and 4, I illustrated how ‘the bond’, according to scientists, shapes the capabilities of all dogs generically. In this chapter, I illustrate how the particularity of a bond, especially a bond between a scientific researcher and a canine research subject, makes it difficult for scientists to identify the capabilities of a generic dog. Dog–human contact, scientists argue, engenders complex affective and emotional responses in both dogs and humans, and this, coupled with dogs’ sophisticated communication skills, and their sensitivity to human gestures, makes it hard to contrive situations in which the relations between human researchers and canine research subjects can be disentangled. But also, more profoundly, it makes it hard to establish what scientific advantage would be gained by such disentanglement. It is against this complex backdrop – against the highly charged question of scientific generalisation in the context of dogs’ species story – that this chapter puts three related issues into conversation with one another: dogs’ perceived relationality, as it is defined as a methodological problem in science; intersubjectivity, as it is defined as the foundation of animal capability, agency and resistance in animal studies; and the contested place of singular individuality, in both.

Vinciane Depret’s model of ‘polite research’ is the stitching that holds these three issues together in this chapter. In the first section, having explored how ‘the bond’ complicates canine science, I ask what methodological support polite research, which secures its authority not in spite of, but on account of, researcher–animal intersubjectivity, might offer canine scientists. Although, as I demonstrate, this method potentially yields rich rewards, it is unlikely that most scientists would pursue it, for it requires abandoning the idea of ‘the dog’. It is precisely this idea, of ‘the dog’, that Jessica Pierce and Marc Bekoff dispute in A Dog’s World (Pierce and Bekoff 2021), and in their popular guide to dog ownership Unleashing Your Dog (Bekoff and Pierce 2019). There is no ‘Universal dog’ (Pierce and Bekoff 2021: 160), the authors write, and certainly no dog who is universally defined by an orientation toward humans. In the second section of this chapter, I explore what methodological issues follow from Pierce and Bekoff’s objection to dogs’ species story, and how polite research is once again both relevant and not relevant to it. For while polite research usefully describes the ethics that characterises Pierce and Bekoff’s approach not just to the scientific study of dogs but to all dog–human relations, the uncritical concept of ‘the individual’ that the authors adopt in opposition to the idea of a generic ‘dog’ is antithetical to it.

The contrasts here, between a vexatious relationality and a welcome one, and between relationality and individuality, illuminate the different ways in which scientific problems are being articulated and addressed. None of these counterpoints to scientific generalisation, however, in my view, quite suffices when it comes to the undoing of dogs’ species story, nor can they wholly account for how dogs specifically might be enabled to object to the questions that are posed to them by scientists.1 After extending my discussion of animal capabilities to animal agency and resistance, I illustrate why this is so with reference to Martin Seligman’s ‘learned helplessness’ experiments. The ‘learned helplessness’ experiments involved the monstrous practice of giving ‘inescapable’ electric shocks to dogs, to dogs in a box from which they could not escape, and then shocking them again to test their responses against experimentally naïve dogs (that is, against dogs who had not previously been ‘inescapably shocked’). The aim was to show that, once these canine research subjects had been exposed to uncontrollable aversive events (to aversive events over which their behaviour made no difference), they did not seek to ‘help’ themselves by escaping or by trying to escape, even if they had previously learned how to do so.

Seligman’s experiments certainly demonstrate how a research apparatus produces subjectivities, as Despret anticipates. But these experiments were in no way polite. On the contrary, as I will illustrate, Seligman forced the dogs to become what his research apparatus proposed to them (helpless subjects, and substitutable species members) and quashed the possibility of any kind of active resistance on the part of an individual dog. Yet in this way precisely, in its erasure both of meaningful intersubjectivity and of the individual agent, the extreme brutality of the experiments was revealing: it exposed a different kind of ‘resistance’, a resistance that was represented by the irreducible singularities of the dogs themselves, by the sheer existence of singularity, which exceeded the spatio-temporal context of the apparatus and, momentarily, disrupted it. I will argue in conclusion that there is a lesson to be learned here, with regard to the challenging of dogs’ species story, that might be extended beyond the research context.

One reason why I call attention to the singular individual is because, as I explored in Chapters 3 and 4, and will explore again, more forcefully, in Chapter 6, species categories tend to, perhaps even aim to, render that individual irrelevant – theoretically, methodologically, ethically and politically. Since indifference is often a death sentence, dispensing too hastily with this figure feels to me to be a luxury that most animals can ill afford humans to hold. I develop this argument, alongside a full interrogation and critique of the concept of species, in Chapters 6 and 7. I also, in Chapter 7, elaborate more fully on what I mean by an enduring individual.

Dogs, scientists and intersubjectivity

The late 1990s and early 2000s are usually identified, by canine scientists, as the beginning of the ‘dog paper boom’, of an explosion of interest in dogs’ behaviour ‘for its own sake’ (Horowitz 2014b: vi; Miklósi 2017: 1–15).2 This has led dogs, the canine ethologist Alexandra Horowitz writes, to become ‘some of the most well-researched and interesting subjects of contemporary psychology and ethology’ (Horowitz 2014b: v; see also Feuerbach and Wynne 2011).3 The focus on ‘dogs qua dogs’ explains why many contemporary canine scientists consider their work to be distinguished from all earlier research on dogs, regardless of whether it is naturalistic or experimental. Horowitz, for example, writes critically of Darwin, who, she argues, for all ‘his great personal interest in dogs, studied domesticated animals as a means to understand how artificial selection worked’ (Horowitz 2014b: vi). Or consider ethologist Ádám Miklósi’s condemnation of Seligman’s ‘learned helplessness’ experiments. The fact that dogs were chosen as research subjects because they ‘are more similar to humans than are other species’, Miklósi writes, makes the ‘lack of concern about dogs’ suffering’ all the more ‘staggering’ (Miklósi 2017: 3).

Miklósi is one of two scientists – the other is Brian Hare – who, while working in different disciplines and countries, are commonly said to have ignited the dog paper boom. In Hungary and the USA respectively, Miklósi and Hare independently conducted similar experiments illustrating that dogs have communication skills that are not shared by primates (i.e. they are a potentially special kind of cognitive ability) or wolves (i.e. they are not inherited by descent), and which are not solely a consequence of their ontogenetic experiences of humans (Miklósi et al. 1998; Hare et al. 2002; Hare and Tomasello 2005; for Hare on Miklósi, see Hare and Woods (2020a: 14 and 231)). Hare et al. concluded from these experiments that, on account of the process of domestication in dogs, ‘the social-cognitive abilities of dogs … have converged with those of humans’ (Hare et al. 2002: 1636). Both Hare and Miklósi have elaborated on this argument over the past twenty-five years. Since I have already addressed Hare’s position in Chapter 4, I will use an illustrative example from Miklósi’s vast body of work to signal the direction in which canine science is currently travelling.

Miklósi et al. (2021) have recently argued that ‘companion’ should be understood not as a description of species such as cats and dogs, but rather as a biological function (meaning, minimally, a behavioural trait) that contributes to the well-being of both the human and the animal in a ‘human companion animal partnership’ (HCAP). ‘Not surprisingly’, the authors write, ‘the dog (Canis familiaris) could be regarded as the core (reference) species for the development of such partnership[s]‌’ (Miklósi et al. 2021: 2). The key point for Miklósi (as for Hare) is that the ‘social competence’ of dogs amounts to more than a reduced flight distance, as argued by Coppinger and Coppinger, among others (Miklósi et al. 2021: 4; see Chapter 2 of this book on Coppinger and Coppinger’s theory of dog speciation). Social competence is a complex phenomena, characterised for instance by the ability to minimise conflict by following social rules; to form close relationships; to concede to a relaxed hierarchical structure (i.e. not to wish to be the pack ‘leader’); to be attentive and interested in others; to understand social context; to use different communication tools to manage interaction; to be able to engage in social learning; and to engage in cooperative interactions (Miklósi et al. 2021: 5; for an earlier, less theoretically developed, but fuller account, see Topál et al. (2009)). If this sounds a lot like humans, it is because it is a lot like humans. Miklósi argues that ‘the evolution of dogs mirrors some aspects of hominization’ (Topál et al. 2009: 84). That is, the ‘dog behaviour complex’ and the ‘human behaviour complex’ are analogous; they evolved in parallel, and they are ‘functionally convergent’ (Topál et al. 2009: 77).

In view of this species story – wherein ‘companion’ is no longer a description of a relationship, but an evolutionary behavioural trait – it is not surprising to learn that contemporary experiments on dogs typically interrogate, for instance: the implications of different forms and degrees of socialisation on dogs’ ability to live with and work for humans; dogs’ ability to understand different kinds of human gestures; dogs’ ability to make their own gestures understood to humans through various communicative channels (such as ‘gaze alternation’); dogs’ capacity to learn human words; the relation between dogs’ contested theory of mind and domestication; dogs’ capacity for empathy with humans; dog–human emotional ‘contagion’; dog behavioural problems, as they affect life in the human home (such as fear, frustration, intra- and interspecies aggression, anxiety, boredom); etc. There are other experiments too, of course, especially on dogs’ intraspecific behaviours (such as dog play), but the orientation of the study of ‘dogs qua dogs’ is substantially directed to dogs qua humans.

The trouble this engenders, for canine scientists, is this: on account of the dog–human ‘special relationship’, individual human researchers and individual dog research subjects are likely to form ‘unique’ bonds (Miklósi 2017: 6), and those bonds will necessarily have qualifying implications for the research. Since dogs’ species story virtually defines dogs by way of their bonds with humans, it is not only the forming of bonds, or the failing to recognise the formation of bonds, that is problematic, but also the contrived absence of bonds/bonding. Enikő Kubinyi et al. gently criticise John Scott and John Fuller’s (1965) classic Genetics and the Social Behavior of the Dog, ‘the bible of dog researchers’, on the grounds that, ‘[a]‌lthough all the dogs were socialized to laboratory staff, no individual social relationships developed … [T]hese subjects could not be considered to have experienced “normal” environmental input’ (Kubinyi et al. 2011: 258). The tension here is between the goal of scientific research, which is to produce generalisable knowledge about ‘the dog’, and dogs’ species story, which points to the unavoidable significance (especially the methodological significance) of particular individual dogs, in particular individual or small-scale sets of relationships with humans (such as scientific researchers), in particular contexts.

This ‘problem’ is exacerbated by the concentration of the major part of scientific research on what are usually called ‘family dogs’. In fact, canine research subjects are often the scientists’ own ‘family’ dogs. The neuroscientist Gregory Berns’s dog Callie, for example, was the first dog to be trained to lie quietly in an fMRI (Berns et al. 2012), while Chaser, the Collie who learned more than 1,000 words, was trained by his guardian, the ethologist John W. Pilley (Pilley 2013). ‘Family dogs’ are probably among the most individualised of all animals in human societies. Indeed, it is in part on account of this individualisation that the historian Erica Fudge is prompted to ask whether a ‘pet’ is an animal at all (Fudge 2002: 27). Her answer: ‘because [they live] with us in our homes … it is possible to see pets as making up a different class of creature’ (Fudge 2002: 28; see also Chapter 6 in this book on the distinctiveness of pets). ‘A pet’, she continues, ‘is a pet first, an animal second’ (Fudge 2002: 32).

This question, ‘is a pet an animal?’ (Fudge 2002: 27), or versions of it, has haunted the scientific study of dogs. Serpell argues that ‘the domestic dog … is an interstitial creature – neither person nor beast – forever oscillating uncomfortably between the roles of high-status animal and low-status human’ (Serpell 2017: 312; see Chapter 6 of this book for an analysis of the racialised aspects of this status). One expression of the ambivalence that this oscillation creates is that, unlike most other research animals, dogs were sometimes named in scientific publications. As if simultaneously to disavow this individualisation however, these names often appeared in inverted commas. W. Horsley Gantt, for instance, who conducted torturous experiments on two dogs for fourteen years, refers to them in his publications as ‘Nick’ and ‘V3’ (Gantt 1962). The dog- and horse trainer and philosopher Vicki Hearne proposes that this use of inverted commas, which seems to indicate that the researcher is referring to creatures to whom they ‘won’t and can’t talk’, ‘is extraordinarily weird, evidence of the superstitions that control the institutionalisation of thought’ (Hearne 2007a: 169). Weird it is, but it may also be a symptom of the ‘suspicion’, Miklósi writes, with which ‘scientists have viewed dogs’, and which has led dogs, ‘[f]‌or many years’, to be denied ‘the status of “real” animals’ (Miklósi 2017: x). So ‘un-animal’ are dogs, that scientists, Pierce and Bekoff argue, tend to view them ‘as outside the sphere of natural taxa’: ‘dogs are often excluded from biological classification schemes, and they rarely appear in zoology textbooks’ (Pierce and Bekoff 2021: 17). The authors object: ‘[d]ogs are not outside of nature’ (161).

Because Miklósi, like Bekoff, is an ethologist, and because ethology is usually defined by scientists as the study of animals in their ‘natural’ or naturalistic environments, Miklósi proposes that ‘in order to allow the dog into the club of “real” animals, we have to find a natural environment for it’ (Miklósi 2017: 1). But what is a dog’s ‘natural environment’? Kubinyi et al. answer: ‘the human social setting’ (Kubinyi et al. 2011: 259). As such, research on dogs can be conducted just about anywhere (Kubinyi et al. 2011: 259). Yet because ‘family’ dogs are the most common subjects of scientific research, in practice most research is conducted in and/or near the ‘family home’. And the ‘family home’, as Kubinyi et al. write, is a ‘complicated, and often uncertain environment’ in which the dog is ‘highly dependent on their owner’ (Kubinyi et al. 2011: 260). Studying dogs in their ‘natural environment’, therefore, means conceding to ‘many uncontrolled environmental variables’, such as the country, region or city in which a dog lives; the gender of the dog guardian; the perception of dogs in a particular culture or context (and especially differing perceptions of particular breeds); specific human individuals’ perceptions of their specific individual dogs; culturally specific dog-keeping practices; as well as the fact that ‘dogs themselves may vary around the world’ (Kubinyi et al. 2011: 259–260). Who is to say what is a dog under such ‘variable’ conditions?!

Moreover, the focus on the family dog raises the question as to who exactly is the ‘expert’ with regard to this specific family dog. As in the case of veterinary medicine, the ‘complex triad’ (Hobson-West and Jutel 2020: 397) of the animal, their guardian and the ‘expert’ may raise questions of authority. Who is best positioned to interpret a dog’s behaviour? In the best possible circumstances, the guardian should know more about ‘their’ individual dog than anybody else (Susan Close, personal correspondence). But their knowledge will be of a particular kind (Lestel et al. 2006: 170) and, from the perspective of science, a kind that is often marked as anecdotal (see Chapter 3 of this book). Ergo, in what looks distinctly like a distancing tactic, Miklósi notes that although he and his team listened to ‘hundreds of casual observations of dog–human interaction (many people would call these anecdotes)’, their aim was to provide ‘an observational and experimental background to these ideas’ (Miklósi 2017: ix).

I have used Miklósi’s (2017) Dog Behavior, Evolution, and Cognition to give a rough structure the above discussion because Miklósi offers a clarifying portrait of the burdens on the canine scientist. Within only a few pages, Miklósi is obliged to establish: that a dog is a real animal; that in scientific studies dogs are more often subject to individualisation than are other animals (through naming, for instance) but that, despite this, the individuality of dogs has been historically neglected in canine research; that dogs have a natural environment in which humans play a pivotal role; that for various reasons this natural environment tends to mitigate against scientific generalisation; and that the study of dogs – the study that nearly everyone who lives or works with a dog cannot help but do – is science and not anecdotalism. In short, dogs’ proximity to humans is destined to affect the research, for better or for worse. Finally, while ethologists usually study the efforts of animals to survive – survival being one of Tinbergen’s four questions (Tinbergen 2005 [1963]) – as Miklósi points out, with no apparent humour (or reflexivity with regard to the scientific focus on western family dogs), ‘[d]‌ogs are not the best candidates for studying survival in nature, mainly because most present-day dogs live with humans and have access to vets, and people do their best to save their companions from the challenges of nature’ (Miklósi 2017: 1). Indeed.

The problems facing these scientists are not superficial. They could not be resolved, for instance, by only ever observing dogs in situations that do not involve humans because, by the very definition of dogs, such situations would be atypical of the species. Nor can scientists ignore the impacts of humans on dogs who are participating, by choice or not, in research. To go back to Seligman’s learned helplessness experiments: in Miklósi’s analysis, the dogs’ ‘ambiguous social situation’ – wherein they had a ‘positive social relationship with the researcher both before and after the experiment was conducted’ (Miklósi 2017: 4) (i.e. before and after they were electrocuted) – would have made it difficult, if not impossible, to distinguish between the ‘neurosis’, as Miklósi describes it, that this ambiguity engendered and the ‘effect of [the dogs’] lack of control over the situation’ (Miklósi 2017: 4). In other words, Seligman not only did not, but could not explain ‘learned helplessness’ in dogs, because he neglected to address the part played by dog–human relationships in his research.4

Another alternative, of course, would be to adopt a different relation to the ‘problem’ of dogs’ so-called hard-wired responsiveness to humans. Rather than regret the difficulty of maintaining distinct boundaries between observer and observed, for instance, a researcher might choose to press further into this perceived dog–human responsiveness, and to rely on the reliability of dogs not just as interpreting subjects (Lestel 2011: 87), but as especially reliable interpreters of humans’ intentions (Bradshaw and Rooney 2017: 134). An increasingly large body of work, developed in the social sciences and humanities, could offer philosophical and methodological support for such an approach. Much of this work is indebted to Vinciane Despret, who has articulated a clear model for research that actively turns on human–animal intersubjectivity (e.g. Despret 2016, 2015a, 2008; see also Lestel 2011). Here, I want just briefly to outline the implications of Despret’s (2008) ‘polite research’ as it bears on the issue of generalisation, and especially generalisations based on species.

Brett Buchanan explains ‘politeness’ thus:

Animals … are not ‘texts’ awaiting hermeneutic interpretation any more than they are ‘objects’ that can be explained through scientific experiments; both are suggestive of a detached objectivity ill-placed with respect to subjective agents. Rather, asking the right questions demonstrates a form of ‘politeness’ towards other beings, not only giving animals the benefit of the doubt of being able to respond but doing so in a way that allows them to respond on their own terms and to answer questions that are of interest to them.(Buchanan 2015: 22)

‘Questions that are of interest to them’ is the starting point of polite research. Although the reasons for engaging an animal’s curiosity, and identifying what matters to them, might seem obvious – octopi are more interested in ‘squirting the experimenters’ than pulling at levers (Godfrey-Smith 2016: 58) – in fact, in Despret’s methodology, it serves the rather more opaque purpose of luring an animal ‘to take a position’ on the researcher’s conjectures as to why they behave as they do. As Despret explains in her analysis of Berndt Heinrich’s research on ravens:

All the work of the researcher consists … in leading the ravens to take a position in relation to [Heinrich’s] fictions and hypotheses: resisting those that do not explain them; clarifying, in those that seem to be able to, that which counted for them. The scientist must, in other words, create a dispositive that confers on the ravens ‘the power not to submit to his interpretations’. It is in this way that the politeness of ‘getting to know’ presents itself.(Despret 2015a: 62)

A ‘dispositive’ – or research ‘apparatus’5 – then, is successful, for Despret, to the extent that it enables an animal to refute and/or clarify the researcher’s assumptions. One exemplary illustration of a successful apparatus, Despret argues, can be found in Irene Pepperberg’s research with the African Grey parrot, Alex. Pepperberg gave Alex the power to actively shape her research not by teaching him how to speak English, but by teaching him how to use language to control his environment and the behaviours of the researchers. The result? Alex ‘accomplish[ed] tasks that were hitherto considered as exceeding the capacities of non-humans’ (Despret 2008: 125). Included among those tasks were ‘speak[ing], describ[ing], count[ing]’ and ‘classify[ing] objects in abstract categories’ (Despret 2008: 125–126).

By way of example of an unsuccessful apparatus, Despret draws on Vicki Hearne’s engaging analysis of ‘the silence of parrots’ in the face of philosophers’ refusal to allow parrots to control the conversation (Despret 2008: 124). A similar example can be found elsewhere in Hearne’s work, in her assessment of the relationship between dogs and behaviourists:

We can now say something about how the story the behaviorist brings into the laboratory affects not only his or her interpretation of what goes on but also what actually does go on. To the extent that the behaviorist manages to deny any belief in the dog’s potential for believing, intending, meaning, etc., there will be no flow of intention, meaning, believing, hoping going on. The dog may try to respond to the behaviorist, but the behaviorist won’t respond to the dog’s response.(Hearne 2007a: 58)

What Hearne is saying here is that the behaviourist’s failure to respond to the dog’s response not only shapes their interpretation of the dog, but also what the dog is capable of: ‘[t]‌he behaviorist’s dog will not only seem stupid, she will be stupid’ (Hearne 2007a: 59; my emphasis). Despret’s concept of polite research is designed to avoid exactly this kind of research relation, which diminishes the animal participants and leaves them no room to show what they are capable of. Politeness, Buchanan writes, is more than an ‘approach to animals’, it is also an invitation to a ‘different response from animals’ (Buchanan 2015: 22, emphasis in the original).

Would this, then, offer an alternative methodological model for research with (rather than on) dogs? Unfortunately, I anticipate that the answer would probably be no. Not because canine scientists are averse to encouraging an expansion of dogs’ agency – on the contrary, they would surely welcome it – but because Despret’s ‘methodological courtesy’ (Buchanan 2015: 22) derives its authority not ‘merely’ from the interactions between the research participants, but from the interactions between these particular research participants, organised according to this particular research apparatus. Alex, Despret writes, ‘doesn’t talk in the name of a “we” of parrots successfully imposed by scientists, but in the name of a “we”’ constituted by the assemblage of a parrot and human beings equipped with an apparatus aimed at making the parrot talk well’ (Despret 2008: 127–128). Even when two further parrots, Kyaro and Alo, joined Pepperberg’s team, and even when they too illustrated that they shared the capabilities of Alex, still Despret considers their testimony to refer to the apparatuses through which their capabilities were realised, rather than to any capabilities that might be ‘guaranteed by the identity of the species and the stability of its repertoire of behaviour’ (Despret 2008: 126). Polite research does not render generalisation impossible, but it does mean that it will be ‘constructed bit by bit’, and that the generalisable knowledge it produces will tell not about species, or even about scientists, but about the apparatus (Despret 2008: 128).

Which is not to say that many canine researchers today do not seek to qualify their results. Horowitz, for instance, advises her reader that ‘when I talk about the dog, I am talking implicitly about those dogs studied to date’ (Horowitz 2012: 9, emphasis in the original). Although ‘well-performed experiments’, she continues, ‘may eventually allow us to reasonably generalize to all dogs, period … even then, the variations among individual dogs will be great’ (9, emphasis in the original). Nevertheless, despite Horowitz’s sensitivity here, it is clear that ‘individual dogs’ do not matter as singular dogs; rather, as the concept ‘variation’ implies, they matter as a departure from a generalisable group (‘all dogs, period’). Horowitz, who was long ago Marc Bekoff’s doctoral student, is unquestionably among the most reflexive canine researchers working today. Yet even she claims to be reporting on ‘the known capacity of the dog’ (9, emphasis in the original), i.e. on ‘the dog’ as a species.

One reason why I feel sure that my proposal of polite research with dogs would be rejected by most canine scientists is that their reflections on the methodological problems posed by dogs make sense only if one already knows what a dog is. For the question for most canine researchers is not whether dogs are especially responsive to humans; it is what to do, methodologically, given that dogs are especially responsive to humans. This starting assumption not only reconfirms dogs’ species story and legitimates the practices that support it, it also makes it difficult to invite a ‘different response’, as Buchanan puts it, from a dog. Difficult, but not impossible – as the following section demonstrates.

Posthuman dogs

In this book, I essentially ask: what are the implications, for dogs, of dogs’ species story, a story that constitutes dogs as always having been and being in relations with humans? In A Dog’s World: Imagining the Lives of Dogs in a World without Humans, Jessica Pierce and March Bekoff (2021) ask: is there a dog without humans? The way they answer this question is far from direct. Theirs is an experiment in ‘speculative biology’, which means speculating about ‘[w]‌hat would happen (or would have happened) if …’ (Pierce and Bekoff 2021: 8, emphasis in the original). Pierce and Bekoff’s if is a science fiction scenario in which humans no longer inhabit the earth. As in the first section of this chapter, my preliminary intentions are to explore how scientists perceive and manage the methodological tensions that are raised by the study of dogs. For Pierce and Bekoff, such tensions follow not from an inherent dog–human relationality that potentially compromises the research relationship and makes it difficult to draw generalisable conclusions about ‘the dog’. Quite conversely, they proceed from a ‘science fiction’ scenario that seems to be intentionally designed to avoid generalisations about dogs. The question, for them, as I noted in the introduction to this book, is how to articulate this project within the domain of canine ‘science fact’.

Since the core of their scenario – a world without humans – has also been explored by other authors, Pierce and Bekoff begin the book by introducing their readers to what these authors have said (usually in passing) about the possible futures of dogs. Such reflections are a showcase for species stories in the present:

‘[W]‌ild predators would finish off the descendants of pet dogs’ (though ‘a wily population of feral house cats’ will persist by feeding on starlings).(Alan Weisman in Pierce and Bekoff 2021: 3)

[C]‌ats are self-reliant and skilled enough to survive without people … Is it possible, [Markham Heid] asks, that after millennia of domestication ‘the entire species [of dogs] may have lost its ability to live independently?’.(Pierce and Bekoff 2021: 4)

A Dog’s World responds to these kinds of conjectures by proposing that scientists cannot know what dogs might be or become in an imaginary future, because they do not know what dogs are in the present. The assumption that dogs are not ‘wily’, or that they are not ‘independent’, cannot, at this point (or perhaps at any point), be either established or refuted. Although Pierce and Bekoff ultimately object to nearly all generalisations about dogs, they object especially strongly to portrayals that take dogs’ dependence on humans for granted.

More specifically, Pierce and Bekoff object to the idea that the ‘purpose’ of a dog is to be a human ‘help-mate’ and ‘companion’ (Pierce and Bekoff 2021: 15), to representations of the evolution of dogs that assume that ‘dogs have evolved to communicate with us’ (88, emphasis in the original) and to the claim that dogs ‘are emotionally attuned to humans and bonded to them’ (117). By way of example, Pierce and Bekoff offer ‘puppy dog eyes, gaze-following behaviour, oxytocin feedback loops, and even dog ESP’ (88). ‘[T]‌he extent to which these emotional skills in dogs are dependent upon or uniquely directed to humans’, they write, ‘is often overstated’ (117). In summary: ‘[d]ogs clearly form social relationships with humans … But it is far too human-centric to say that … the ur-dog would say to the ur-human (in the voice of Tom Cruise), “You complete me”’ (91).

Pierce and Bekoff’s objection to the narcissism that characterises so many accounts of ‘what is a dog’ puts them in an awkward position with regard to the dominant dog species story. They negotiate this by neither wholly disputing nor fully endorsing it. ‘The origin of modern dogs’, the authors write, is ‘hotly contested’ (Pierce and Bekoff 2021: 22). Scientific questions regarding the relation of dogs to wolves, for example, are ‘likely to get muddier before they get clearer’ (23). The field is riddled with controversies. Here Pierce and Bekoff refer to the recent discovery, in November 2019, of the body of puppy in eastern Siberia. Dated to approximately 18,000 years ago, the question remains as to whether the pup was ‘a wolf or a dog or perhaps an animal that was ancestral to both’ (23; see also Chapter 2 of this book on the difficulties of classifying potential ancestral wolf or dog remains).

This is obviously not a dispute with evolutionary theory itself, given that speculative biology, as Pierce and Bekoff define it, ‘makes[s]‌ predictions about the trajectory of evolution’ (Pierce and Bekoff 2021: 8). One dramatic change in evolutionary direction that would follow from a world without humans – to take a single example – is that the dog population that constitutes the main object of most scientific research (companion and working dogs in the Global North) would be subject to natural rather than artificial selection.6 In keeping with the attentiveness to time that generally characterises the work of evolutionists (see Chapters 2, 4 and 6 of this book), Pierce and Bekoff note that this change will affect different populations of dogs differently over time. Only ‘later-generation’ dogs – dogs who after thirty years or so will replace ‘transition dogs’ (alive when humans disappear) and ‘first-generation’ dogs (born to mothers who had contact with humans) – will be ‘truly posthuman’ (Pierce and Bekoff 2021: 13). Posthuman, for Pierce and Bekoff, does not in any way refer, as it does in the social sciences and humanities, to (broadly speaking) the desubjectifying entanglements that reveal the autonomous, independent, Cartesian individual to be a fiction (and which, by the same token, reveal the very concept of posthuman to be a fiction, for ‘we have never been human’). Posthuman dogs are what it says on the tin: they are the dogs who live on when ‘all humans are gone’ (Pierce and Bekoff 2021: 12).

The most common strategy deployed by Pierce and Bekoff in their protest against generalising categories – be they a taxonomic category, such as canid, or a biological concept, such as phenotype (Pierce and Bekoff 2021: 46) – is to outline the characteristics that are indexed to the category or concept, and then to dissolve its intelligibility by pointing to the variability that exists within it. Yes, canids, for instance, are defined by particular physical features and behavioural traits. But these features are extremely diverse, especially with regard to size, which can vary between 2 and 150 lb (Pierce and Bekoff 2021: 21). Put this together with the differences that can be identified within every other physical feature and behavioural trait, and the implications for making generalisations about the future of any single posthuman dog multiply exponentially. Add experience into the mix, and the future of even the famously ‘intelligent’ Border Collie (see Chapter 1 of this book) cannot be guaranteed. For breed, which is a generalising category, is trumped by individual experience:

[N]‌o two border collies are alike, and each will respond to a posthuman future in unique ways … [T]wo border collies who were raised under the same circumstances and exposed to the same environment will respond differently to future events … Breeds don’t have personalities; individual dogs have personalities.(Pierce and Bekoff 2021: 36)

The conceit of speculative biology comes into its own here, for by exploring the consequences, in an imaginary future, that follow from even a single difference, the authors are enabled to shine a light on the significance of the multiplicities of differences among dogs in the present. In Pierce and Bekoff’s analysis, ‘difference’ means the difference between individuals,7 and as the examples stack up and up, the reader is confirmed in thinking that the real methodological function of the thought experiment is to put an end to nearly all kinds of generalisations about dogs.

The agenda of A Dog’s World is explicit:

Imagining a future for dogs without their human counterparts is an interesting exercise in biology, but the real value of the thought experiment … is that it can help us think more clearly about who dogs are in the present and this, in turn, can clarify the moral contours of human–canine relationships.(Pierce and Bekoff 2021: 13–14)

No reader can be in doubt about those moral contours by the time they reach the end of Chapter 8. In this penultimate chapter, Pierce and Bekoff balance a speculative list of dogs’ ‘losses’ against dogs’ ‘gains’ in a posthuman world. Not only is the list of losses short by comparison, the kind of items on this list (nutritious food, water, shelter, friendship, ‘toys’) are, as the authors note, ‘replaceable’ (Pierce and Bekoff 2021: 153). To describe what dogs would gain in a posthuman world – and this is presumably the purpose of the exercise – is to describe what dogs lose in this, human, world. The physical losses to dogs of living in this world, now, include: human constraints (collars, leashes, fences, cages), intensive captivity (puppy mills, laboratories, dog meat farming), experimentation, abuse, sexual exploitation, dog fighting, forced breeding, the killing of healthy dogs, artificial selection for maladaptive traits, obesity, desexing (with all its attendant health implications), surgical mutilations (docking, debarking, ear cropping), shelters and shelter-related mortality, and breed-specific genetic disorders (Pierce and Bekoff 2021: 146). They also lose: potentially better nutrition, a greater range of sensory experiences, a natural level of hormones and development, and levels of physical activity that suit and are decided upon by dogs themselves (146). As well as the physical losses incurred by their lives with humans, Pierce and Bekoff also review the ‘social’ and ‘psychological’ losses. Suffice it to say that these latter lists make my Chapter 1 look like a wan description of dog despair.

Pierce and Bekoff’s ‘gains and losses’ chapter gives the reader some insight into how dogs’ species story is operationalised in practice, through vast and complex networks of human control, and especially human control over dogs’ reproduction. As I understand it, their point is not really that a change in evolutionary trajectory would be transformative of dogs (that much is obvious); it is, more radically, that if dogs were ‘free’, in this world, and especially if they were free to choose when and with whom to mate (Pierce and Bekoff 2021: 136–139), then the ‘fiction’ of ‘what is a dog’ would become visible. With what consequences? ‘We could still live with dogs as companions, although over time they might become less tame, less docile, and less interested in being our pets’ (138). In other words, if humans did not exercise such tight control over dogs, the true fragility of the (story of the) ‘dog–human bond’ would be exposed for what it is: a retrospective explanation, naturalisation and justification of contemporary dog–human relations.

Before reading A Dog’s World, I read Bekoff and Pierce’s Unleashing Your Dog: A Field Guide to Giving Your Dog the Best Life Possible (Bekoff and Pierce 2019). In this ostensibly ‘light-weight’ book, I thought that I had, somewhat surprisingly, found a critique of dogs’ species story. So surprising was this to me, that I feared that I had perhaps read too much against the grain of Bekoff and Pierce’s argument, or had read too much into it, or had been a little too wishful. With the publication of A Dog’s World, however, the unusual tone of Unleashing makes better sense to me, and gives me confidence in my interpretation of it. Unleashing is, quite simply, not written with dogs’ species story in mind. This immediately liberates the authors from the classic ‘dog book’ structure, which usually starts with the domestication-as-speciation story, and indeed from any structure that would support the substitution of one ‘essence’ of dog with another (cf. Coppinger and Coppinger 2016, as discussed in Chapter 2 of this book; Wynne 2020a, as discussed in Chapter 4 of this book). Instead, they organise each of their chapters around one of dogs’ senses, with a final chapter on play, which the authors describe as ‘a kaleidoscope of senses’ (Bekoff and Pierce 2019: 139). In keeping with A Dog’s World, Bekoff and Pierce are quick to foreground the obvious artificiality of the chapter divisions (Bekoff and Pierce 2019: 35), and urge their readers to remember that ‘even’ when it comes to the senses, one individual dog will experience particular forms of sensory deprivation more or less ‘keenly’ than another (Bekoff and Pierce 2019: 10).

All this has implications for the dog owners to whom the book is addressed. Although the authors make some small effort to soothe the multifaceted anxieties of people who live with dogs, on the whole their portrait of dog guardians is relentlessly unflattering. This is largely due to their uncompromising starting point, which is that companion dogs should be understood as ‘canine captives’:8 ‘[T]‌here’s no getting around this’ (Bekoff and Pierce 2019: 9). And, in case any reader is in doubt as to what captivity means, Bekoff and Pierce helpfully provide a definition: ‘from the Latin captivus, “caught, taken prisoner”, and from capere, “to take, hold, seize” (5).

Although Bekoff and Pierce do not deny ‘that celebrated “bond”’ (Bekoff and Pierce 2019: 8) between dogs and humans, for them the form the bond takes, the form of the relation, is captivity. In a vague address to an unidentified group of scientists who ‘argue that the long association with humans has changed what is “natural” for dogs’, Bekoff and Pierce insist instead that ‘[d]‌ogs will never fit easily and without negotiation into human homes and lifestyles’ (9). The reasons are many and include, alongside the more familiar deprivations (under-exercise, long periods of isolation, being treated like a ‘furry human’), a detailed account of how the very sensory architecture of a human home – its lighting, sounds, smells, spatial organisation – can be an assault on a dog’s senses and on their person. For Bekoff and Pierce, the ‘bond’ between dogs and humans is a ‘Faustian bargain’, a bargain based on a desire not for ‘knowledge’ but for ‘love and companionship’ (Bekoff and Pierce 2019: 8). That desire is located not on the side of dogs, however, as in Clive Wynne’s (2020a) Dog Is Love, for example (see Chapter 4 of this book), but firmly on the side of humans. It is humans who seek to ‘capitalize’ (Bekoff and Pierce 2019: 8) on the dog–human relationship, and it is this asymmetry that transforms ‘the bond’ into a bondage for dogs. The interface between humans and dogs, for dogs, is not a zone of ‘bonding’ but rather a ‘zone of uncertainty’, where uncertainty is the tension between ‘captivity and freedom’ (Bekoff and Pierce 2019: 9).

Bekoff and Pierce frame Unleashing in terms of ethics (Bekoff and Pierce 2019: 8). To me, however, it reads more like a political manifesto that is essentially concerned with the struggle for freedom for dogs. Not for total freedom, not for a freedom ‘outside’ what exists now (for that, one must turn to A Dog’s World), but for greater ‘degrees of freedom’ (Bekoff and Pierce 2019: 9) within the current (political) structure. Just as Porcher and Schmitt (see Chapter 4 of this book) argue that, even though domestication never differs, the conditions might – ‘a dog, like a pig, may be treated well or badly’ (Porcher and Schmitt 2012: 40) – so Bekoff and Pierce similarly argue that captivity ‘refers to a type of existence, not its quality’ (Bekoff and Pierce 2019: 5). Although this existence cannot be ‘otherwise’, it can be improved to the extent that the so-called dog–human bond that defines and shapes that existence is loosened. The title of the book Unleashing Your Dog, ultimately, means unleashing your dog from you. Bekoff and Pierce’s book is characterised by a deadly seriousness, which is belied by its populist tone.

For Bekoff and Pierce it is not the relational intersubjectivity of dogs and humans that poses a problem for science; rather, it is that each dog differs from another, and so too do their own behaviours, over time, or in particular contexts (Bekoff and Pierce 2019: 19, 23). Science, in short, does not know as much about dogs as some science writers imply (Bekoff and Pierce 2019: 16) and, in view of the authors’ consistent (perhaps even militant) emphasis on the individuality of every dog in both Unleashing and A Dog’s World, it is difficult not to conclude that it never will. What does it mean, then, to do research on or with ‘dogs’? The argument in Unleashing, and especially in A Dog’s World, suggests that the questions ‘what is a dog?’ and ‘what can a dog do?’ still have relevance, but that they must be posed to each and every ‘truly … distinct individual’ (Bekoff and Pierce 2019: 150). It is because individual dogs are ‘truly distinct’ that the answers to these questions cannot be anticipated in advance. Moreover, whether the questions are posited by scientists, owners, trainers, or by whomever it is who enters into relations with a dog, they necessarily incur the obligation of allowing (if not enabling) the dog to answer in their own specific and unique way – again, because they are unique. This is, in effect, Despret’s notion of polite research, which emphasises particularity in its every dimension, refracted through the individual and extended to all dog–human relations.

Animal capabilities, agency and resistance

Except, of course, it is not Despret’s polite research because there is no place in Despret’s work for an unreflexive conception of ‘the individual’. I will address some of the very good reasons for this rejection of the individual, in animal studies and in the social sciences more broadly, in Chapter 7. Here, I confine my comments to what is problematic about it in the context of debates about animal agency and resistance. One most obvious problem concerns the rapid cascade of interlocking assumptions that usually follows from the notion that, in order to ‘have’ agency, one must be an individual (that is, that agency is a property of individuals). This idea, which is both philosophical (Pearson 2014) and common-sensical (Meijer and Bovenkerk 2021: 52), often goes hand in hand with the conviction that agency is synonymous with intentionality (Meijer and Bovenkerk 2021: 52). And, further, that in order to act with intention, the individual must be cognitively capable of reasoning. ‘The ability to reason’, the historian Chris Pearson writes, ‘is central to the human-centered concept of agency because it allows people to break free, to an extent, of their instincts, emotions, traditions, and political and social structures’ (Pearson 2014: 133). Perspective too, then, co-exists with this conception of agency. For no one can ‘break free’ of their circumstances unless (presumably) they have a point of view on them.

While some authors consider this model of agency to be relevant to animals – in his much-cited Fear of the Animal Planet, Jason Hribal, Pearson argues, transforms animals into ‘four-legged agitators, to file alongside human radicals and revolutionaries’ (Pearson 2014: 251) – there are other, less anthropomorphic, ways to conceive of an animal’s relations to their conditions, and of their objections to it. Susan Nance, for instance, in her analysis of elephants in the entertainment industry in the USA, argues that is not necessary to speak of elephants ‘rejecting the circus or capitalism’, nor is it necessary to identify ‘what a given elephant’s intentions or internal experience was at every moment’, in order to be able to prove, based on ‘a body of evidence’, that elephants ‘rejected their conditions of existence’ (Nance 2013: 10). Nevertheless, even without assuming that elephants have a ‘political’ take on their conditions, or that their every action is informed by intentionality, the notion that an elephant ‘has’ agency tends to confine questions of agency to questions of the individual. One is prompted to ask, for example, in what situations an individual elephant is able to exercise ‘their’ agency or not, etc.

This is partly why, for Despret, the problematic coupling together of agency and perspective runs deeper than the ‘anthropomorphic conception of subjectivity’ (Despret 2013: 30) that undergirds it, and which is the product of a long history of ‘intellectual and cultural shifts that created the perspectival mode’ (Daston in Despret 2013: 30).9 More significantly, for her, this coupling makes it difficult to describe ‘unfamiliar beings, such as bees or even flowers’ as agents because they appear not to have a perspective (29). In order to develop a theory of agency that is ‘much more extensively shared in the living world’ (29), Despret identifies agency not as a property of a subject with a point of view but as the product of a ‘rapport of forces’ in which ‘[e]‌ach living being renders other creatures capable (of affecting and of being affected), and they are entangled in a myriad of rapports of forces, all which are “agencements”’ (37).

Let me step back for a moment. As discussed above, in Despret’s model of polite research, an animal’s authorisation is achieved not solely by creating a research apparatus that engages their interests, but also by giving the animal the ‘power’ to resist the researcher’s hypotheses.10 It should be noted that this robust animal, this animal who is capable of resisting the researcher’s explanations of their behaviours, is not a fortified individual who has the ‘freedom to …’ or the ‘freedom of …’, as Pierce and Bekoff put it (Pierce and Bekoff 2021: 146). On the contrary, a productive research apparatus achieves results precisely to the extent that it engenders ‘an unprecedented, creative, improvised, queer “becoming together”’ (Despret 2013: 33). Berndt Heinrich’s invitation to the ravens to activity, his invitation to them to take a position on his hypotheses, is in effect an invitation to intersubjectivity, by which Despret means: ‘becoming what the other suggests to you, accepting a proposal of subjectivity, acting in the manner in which the other addresses you, actualizing and verifying this proposal, in the sense of rendering it true’ (Despret 2008: 135).

To pose the question of animal agency is, thus, to make a proposal of subjectivity, a proposal that can transform an animal from something that looks like a machine (say) into someone who looks like a subject, a subject with a point of view. This, according to Despret, was Jocelyn Porcher and Tiphaine Schmitt’s (2012) achievement when they illustrated that cows ‘do more than simply function’ at work (Porcher and Schmitt 2012: 55, emphasis omitted; see Chapter 4 of this book). In a stroke of methodological genius, Porcher and Schmitt showed that cows actively cooperate at work, that they actively invest their affects in their work, by demonstrating how, on occasion, they refused to work, refused to cooperate and made it hard for the farmers to do their jobs. For Despret, however, even if an animal – a cow in this instance – looks as though they ‘have’ capabilities or ‘have’ agency, this appearance is always subtended by a ‘rapport of forces that makes some beings capable of making other beings capable, in a plurivocal manner’ (Despret 2013: 38).

If agency is, in effect, ‘activated’ by affective relationality, then it is as relevant to flowers as it is to cows. Despret writes:

This is how flowers gain agency, through becoming enabled to make their companion pollinators [bees] be moved by them, and this is how the latter could themselves be agents, through becoming enabled to make the flowers able to attract them, and in turn to be moved by them. This is why agency always appears in a flow of forces. Agencies spring in a flow of forces, in agencements that makes more agencies.(Despret 2013: 40)

The appearance of agency is the product of an agencement: ‘there is no agency without agencement. In other words, a being’s agency testifies to the existence of an agencement’ (Despret 2013: 38). Like the specificity of a research apparatus, which enables or delimits a being’s intersubjective capabilities, the specificity of an agencement enables or delimits a being’s agency. In both cases, these qualities are the product of the assemblage, and not the property of the human, animal, insect or plant.

I began this section with a model of agency that attributes agency to individuals, and conflates it with intentionality and reasoning. This sets the bar for agency very ‘high’, and sets it in favour of humans. Indeed, to the extent that it relies on a normative, even idealised, notion of the human agent, it not only fails to recognise ‘that agency can be exercised by different beings in different ways’ (Meijer and Bovenkerk 2021: 53), but also ‘exaggerates the gap between humans and other animals’ (Meijer and Bovenkerk 2021: 52). Relational conceptions of subjectivity and agency, such as Despret’s, are hugely valuable in terms of moving away from this exclusive and exclusionary model. And yet, in some ways, I find the implications that follow from her attribution of agency to bees and flowers, who seemingly do not have a perspective, less startling than the implications that potentially follow from her use of the concept of an agencement to explain the appearance of an agential subject, with a point of view, within a research apparatus. Startling, not because this position insists that subjectivity is created rather than given, or that all agency is interagency, but rather because it implies, first, that the most significant feature of this figure is ‘their’ point of view (which is why it needs explaining) and, second, that the capabilities of an animal can in theory be reduced to the relations that are underpinned by a particular agencement. In other words, that the capabilities of a ‘subject’ within an experiment, say, and indeed the very subject itself, can be wholly explained by the experiment. This closes off the possibility of the significance of an ‘outside’, and also the possibility that there might be something about the figure of the subject which is also ‘capable’, and capable of ‘resistance’, that is not tied to a point of view.

In Martin Seligman’s ‘learned helplessness’ experiments, it was the eruption of an outside inside the experiments – an outside that assumed the form of dogs’ unique individual biographies – that opened up the research to forces beyond it, and ensured that some of the dogs did not respond as Seligman anticipated.

Intersubjectivity and enduring singularity

As I have already indicated, Seligman’s learned helplessness experiments are, for many contemporary canine researchers, exemplary of scientific and moral failure. Astonishing as it seems, however, the early controversies surrounding Seligman’s research stemmed less from solicitude about the dogs’ welfare and more from Seligman’s ‘clash with traditional stimulus–response theories of learning’, and also the use he made of ‘mentalist’ concepts to explain his results (Peterson 2004: 517).11 As I noted in Chapter 3, behaviourism sequesters any analysis of the private, interior worlds of living creatures in favour of empirically observable behaviours. Methodological behaviourists do this because they consider those worlds to be inaccessible. For radical behaviourists, such as John Watson and, later, B. F. Skinner, that private, interior, world does not exist at all (Schneider and Morris 1987: 33; Boakes 2008: 153). Either way, behaviourism renders mentation an ‘irrelevance’ (Rollin 1998: 207). In its place, learning is ‘the mark of the mental’ (Rollin 1998: 208–209, emphasis in the original).

With regard to learning: behaviourist theories typically argue that learning is produced either when two events occur in contiguity (when this happens, learning leads to the ‘acquisition’ of behaviours), or when two events, once associated with each other, become non-contiguous (when this happens, learning leads to the ‘extinguishment’ of behaviours). In a departure from this model, Seligman added a third form of learning, ‘independence between events’ (Seligman and Maier 1967: 8), i.e. non-contingency. Seligman argued that when a dog has learned that distressing events are contingent upon nothing, and that there is nothing she can do to change or control them, then – and this was the controversial part – she will do nothing in response, even when an alternative, one with which she is familiar, is presented to her. Exposure to uncontrollable events, Seligman concluded, produces a series of ‘motivational, cognitive and emotional deficits’ (Maier and Seligman 1976: 3).

Seligman arrived at his conclusions by dividing his dogs into two groups. One group, prior to being put into a shuttle box (a box that contains a means of escape), were given electric shocks from which they could not escape, either because they were yoked into a hammock or because they had been temporarily paralysed by curare (a neuromuscular blocking agent) (Overmier and Seligman 1967). This group of dogs – which numbered 150 (Seligman 1972: 408) – was distinguished from another group who had not been inescapably shocked. When put into the shuttle box, both groups had ten seconds, from the start of a conditioned stimulus (dimmed lights), to jump a barrier that would enable them to escape the shock chamber. If they did not jump, the shock – ‘administered through the grid floor’ (Overmier and Seligman 1967: 28) – would start and continue for sixty seconds (Overmier and Seligman 1967: 29).

Bruce Overmier and Seligman ‘discovered’ that dogs who had not previously been inescapably shocked more frequently jumped the barrier. Those who had been inescapably shocked before entering the shuttle box learned more slowly that jumping the barrier was a means of escape, often failed to retain this learning and, even when they did learn and retain it, frequently ‘[gave] up and passively [accepted] the shock’ (Seligman 1972: 407). These, then, are the ‘motivational, cognitive, and emotional deficits’ produced by this research apparatus: motivational, in the sense that the dogs were not motivated even to attempt to escape, having been subjected to uncontrollable aversive events; cognitive, in the sense that these uncontrollable aversive events interfered with the dogs’ ability to establish contingent connections between behaviours and outcomes; and emotional, in the sense that the dogs were more greatly distressed by uncontrollable than by controllable events (Maier and Seligman 1976: 3).

In his brilliant analysis of fish agency and resistance, Dinesh Wadiwel argues that one can identify agential resistance in the designs of instruments that are used to control, capture and kill animals. What these designs illustrate, Wadiwel writes, is that ‘the resisting body generates the need for the instrument of violence, and technological refinement in the instrument of violence corresponds with the continuing creativity and innovation of those who resist’ (Wadiwel 2016: 210). The shuttle box/shock chamber, the hammock, the curare: these horrifying instruments of violence offer evidence of the efforts required to make shock inescapable to a subject who wants nothing but to escape.12 Their very pitilessness indicates that the dogs’ ‘resisting bodies’ obliged the researchers to actively produce passivity in order to discover passivity, to actively produce submission in order to discover submission. In addition to this physical coercion, the dogs were also emotionally coerced by their relations with the researchers. In his book on animal pain, the philosopher Bernard Rollin argues that the pain that the dogs experienced in Seligman’s studies would have been ‘deepened and rendered more extreme by its total incomprehensibility’ (Rollin 1998: 145), an incomprehensibility made all the more abysmal by the friendliness of the people who electrocuted them.

This is why, as I discussed earlier in this chapter, Miklósi argues that Seligman could not isolate ‘lack of control over aversive events’ as a cause of learned helplessness: because the dogs’ ‘motivational, cognitive and emotional deficits’ would also have been shaped by the ambiguous relations with humans that were instantiated by the research apparatus. A research apparatus always produces something, even if what it produces is negative (such as the silence of the parrots when faced with philosophers). Seligman’s apparatus produced passivity and ‘neurosis’, as Miklósi characterises it. For the ambiguous, intersubjective relations between the researchers and the dogs were, in their own degenerate way, a proposal of subjectivity to the dogs, even though the nature of that proposal, and the manner in which they were to act in response, must surely have been impenetrable to them, as Rollin says. The dogs’ ‘neurosis’ is an intelligible response in this context, a response that renders true an abusive and violating proposal of subjectivity.

The violence of Seligman’s research is significant: because even still, he could not immediately produce helpless passivity in every dog who had been inescapably shocked. Not, in this instance, because his apparatus produced resisting subjects (what room for resistance, with the shock chamber, the hammock, the curare?), but because, conversely, something in some of the dogs’ pre-laboratory biographies ensured that they were, at least initially, insulated from its purpose. The fact that individual animal biographies are rarely recorded does not mean that animals do not have them (Fudge 2004; Baratay 2022): ‘have them’, not in the sense that animals are potential autobiographical subjects – which is what the writing of biography usually implies – but in the sense that animals are individuals who are shaped by their histories of experience. Although Seligman could not identify what exactly in these dogs’ experiences ‘protected’ them from his proposal of helpless subjectivity (because their pre-laboratory histories were unknown to the researchers), they made a difference to it nevertheless.

Seligman’s apparatus did not produce these dogs as particular, singular individuals; their singularity was derived from experiences (undoubtedly also relational) that preceded the time, space and context of the experiments. It did draw attention, however, to their being singular, and it made that mode of being matter. It prompted Seligman to ask more questions and to ‘commit to more activities’ (Despret 2015a: 64). Those questions and activities, however, unlike in a polite research apparatus, were not designed by Seligman to encourage activities in the dogs ‘in return’ (Despret 2015a: 64). On the contrary, Seligman repeated his original question, and demanded that his hypothesis be confirmed as true: ‘[c]‌ould it be possible that those dogs … have had a prelaboratory history of controllable trauma while dogs who are helpless without any previous shock have experienced uncontrollable trauma before arriving at the lab?’ (Seligman 1972: 410). And confirmation indeed he received, by the method of comparing dogs with pre-laboratory histories of experience to ‘cage-reared dogs’ who ‘have very limited experience controlling anything’ (410).

The problem with not encouraging activity in return – the problem with confirmation – is that there is nothing to be learned from it. The comparison between dogs with unique personal histories and cage-reared dogs ensured that Seligman was obliged neither to investigate what precisely about those pre-laboratory histories was significant, nor to question any aspect of the apparatus, for the cage-reared dogs were essentially bred to bear witness to its purpose (which was to ‘demonstrate’ that exposure to uncontrollable events leads to helplessness). For what chance would a cage-reared dog, with ‘limited experience of controlling anything’, have of exerting control over an apparatus that is specifically designed to disestablish it? This is a comparison, in other words, defined by ignorance on the one hand and enforcement on the other. In this way, Seligman bullied his hypothesis into truth. For those readers who are interested: ‘[w]‌hile it took four sessions of inescapable shock to produce helplessness one week later in dogs of unknown history, two sessions of inescapable shock in the hammock were sufficient to cause helplessness in the cage-reared dogs’ (Seligman 1972: 410).

The point here is that the significance of the pre-laboratory histories of some of the dogs in Seligman’s study could not be erased, despite the unrestrained violence of the research apparatus, which sought to impose its proposal of neurotic subjectivity on all of the dogs. Of Porcher’s analysis of cow cooperation and resistance, Despret writes: ‘when everything happens as it should, we don’t see the work’ (Despret 2013: 42). One might say of Seligman’s experiments that had everything happened as it should, we wouldn’t have seen the singular individual. Animals, Dominique Lestel writes, ‘are individuals which do not always behave as they “should”’ (Lestel 2011: 84). I would push this point further and argue that it is because animals are individuals that they do not always behave as they should. In response, Seligman sought to breed experience- and individuality-free dogs – ‘generic’ dogs, one might say – whose invariability approximates the invariability that is supposed to characterise a representative of a species.

But a cage-reared dog is not ‘experience-free’. A cage-reared dog has unique experiences of being cage-reared, which will shape a uniquely singular individual, irreducible, and thus resistant, to species. ‘Submission’ may be a part of the story of that experience, but it is not the whole of it.


This chapter began by addressing the methodological challenges that dogs’ species story poses to those scientists who are invested in it. Pierce and Bekoff’s somewhat anfractuous science fiction scenario – in effect, an alternative methodology by alternative means – illustrates how difficult it is to confront that story, much less to overturn it. Dogs’ species story is not, I have argued, undermined by the ‘problems’ raised by canine research. On the contrary, those problems serve only to verify the ‘truth’ of it. There is an underlying irony here, which is that the particular relationships established in research, between particular researchers and particular dogs, confirms to scientists the defining characteristic (relationality) of dogs ‘as a species’. When we look at animals, Erica Fudge writes, we see ‘something dangerously recognizable. We see, in fact, a version of ourselves’ (Fudge 2002: 40). Is there any animal about whom this is more true, in science, than the domesticated dog?

Polite research is both about ourselves and about getting away from ourselves. It is about ourselves in the sense that it does not assume that the researcher and the research subject can be disaggregated either from each other or from the research apparatus. To ask a question is, necessarily, to enter into relations. But it is also about getting away from ourselves, for the aim is to enable animals to object to (and thereby to refine) the questions that are posed to them. The problem for dogs, however, as I have argued throughout this book, is that intersubjectivity and interagency, in the context of dogs’ species story, are not necessarily polite; that is, they do not necessarily enable dogs to show what they are capable of (the irony to which I refer, above). I am returned again, then, to the tensions around which this chapter has circled: species and the question of generalisation more broadly, relationality, capabilities, individual agency, resistance. It has not been the aim of this chapter to try to resolve these tensions. On the contrary, this chapter has been driven by their irreconcilability, and by the implications of that irreconcilability with regard to the potential dismantling of dogs’ species story in science.

Seligman’s research was paralysing, both physically and emotionally. Yet still the dogs in these experiments could not be reduced entirely to the proposals that Seligman put to them, of helpless subjectivity, and of species. What ‘resistance’ they put up, however, derived not from some quality or attribute smuggled in under the category of the individual (freedom to act, some inherent capability, a point of view that was permitted to matter). More simply, it issued from the assemblages of experiences (whether they were experienced as such or not) that constituted the dogs’ singularities. In the context of these violent experiments, only these irreducible singularities could disrupt Seligman’s practice and oblige him, at least momentarily, to question his assumptions. And perhaps, because this minimal conception of individuality is somewhat less relational and somewhat more defined by the boundaries of a life lived uniquely (this happened to this dog, this did not happen to this dog; this dog did this, this dog did that), it might also offer a portal to thinking ‘beyond ourselves’ (Fudge 2002: 22). Revolting as Seligman’s experiments were, they also, inadvertently, posed a question that I would argue is rarely extended to dogs. Crudely: who are you, beyond what I know of you, what I see of you, what I project onto you, and what I want from you?

In this chapter, I have emphasised the methodological and ethical significance of a singular life. In the following chapter, I address the political significance of a singular death.


1 By ‘dogs specifically’, I mean dogs as they are specifically understood through dogs’ species story, which constitutes them as relational by definition.
2 Why some of these scientists seek to research dogs’ behaviour ‘for its own sake’ – i.e. for what purpose – is an open question. See for example my discussion of in Chapter 4.
3 Horowitz mentions psychology and ethology here, but dog behaviours and biology are of considerable interest across a wider range of scientific disciplines than this, including, since the genomic sequencing of a purebred Boxer, Tasha, in 2004, comparative genetics (e.g. Lindblad-Toh et al. 2005; Ostrander 2012).
4 Seligman was not the first to fail to recognise the impact on the dogs of their relationships with the researchers who electrocuted them. Gantt, in his research on the ‘effect of person’ (1962), neglected to account for the familiarity of the dogs with the experimenters who were stroking them as they were being shocked. Both Gantt and Seligman recorded the significance of dog–human relationships to dogs in general, but neither investigated how their own particular dogs’ responses were affected by ‘specificity to a particular person’ (Feuerbacher and Wynne 2011: 52).
5 See Jeffrey Bussolini’s helpful comment on this problematic translation (Bussolini in Despret 2015a: 71n).
6 This would also bear on free-roaming dogs, however, given that ‘sterilisation is a core activity of free-roaming dog population management’ globally (Collinson et al. 2021: 1).
7 ‘Individual’ is not to be conflated with ‘sameness’ over a lifetime, however, because for Pierce and Bekoff a dog’s changing experiences will change the dog (Pierce and Bekoff 2021: 32).
8 I deduce that Bekoff and Pierce make a distinction between ‘captive’ dogs, who are companion and working dogs, and ‘intensely captive’ dogs, who are dogs held in puppy mills, laboratories etc.
9 See Lorraine Daston’s excellent summary of this history (Daston in Despret 2013: 30)
10 Herein lies a key distinction between polite research and anecdotes (see Chapter 3): polite researchers are obliged to have their interpretations (or versions) of events authorised by their animal research participants. An anecdote demands no such authorisation, which is why anthropomorphism so often runs free.
11 Despite objections from behaviourists that this helplessness could be explained by ‘motor response deficits’ (that is, by the dogs not being physically able to help themselves), in the end the cognitive interpretation won out, as ‘psychologists … [saw] the parallels between learned helplessness as produced in the laboratory and maladaptive passivity as it exists in the real world [in humans]’ (Peterson 2004: 517).
12 There are many more examples of ‘learned helplessness’ apparatuses that were ‘tailored’ to the resistant capabilities of other animal bodies, including cats, rats, mice and fish.
  • Collapse
  • Expand

All of MUP's digital content including Open Access books and journals is now available on manchesterhive.


Dog politics

Species stories and the animal sciences


All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 102 102 89
PDF Downloads 88 88 82