I pause for a moment, with this chapter, to address some of the ways that the idea of a species and actual animals’ behaviours come to be connected. For dogs’ species story would not necessarily, in itself, bear so very heavily on dogs, were it not that this story often shapes contemporary scientific understandings of what dogs need and want, and how they do and should behave (see especially Chapter 4). This is one of the reasons why species stories matter: if species is the lens through which behaviours are framed, then how a species is characterised assumes real, material significance. Chapter 3 is also important, with regard to the argument in this book, in a second sense, for it is here that I begin to introduce the process by which ‘species thinking’ erases the significance of particularity, and especially the particular individual.

Where the previous chapter attended largely to scientific theories, this chapter addresses itself to scientific methods: specifically, to the methods of Konrad Lorenz, Conwy Lloyd Morgan, George Romanes and Charles Darwin, all of which, in their different ways, serve to tie species and behaviours together, such that one is illustrative of and reinforces the other. I choose these scientists because they were at the forefront of three major disciplinary approaches to animals in the late nineteenth and the twentieth century: classical ethology; that branch of comparative psychology that would become behaviourism; and what might loosely be called Darwinian ‘anecdotalism’. Anecdotalism continues, today, to shape critiques of contemporary ethology.

I begin with Lorenz’s work, which yokes together species and behaviours in a most transparent fashion. Lorenz’s concept of instincts (innate ‘fixed action patterns’) is now associated with a limited understanding of animals driven, as Marga Vicedo concisely summarises it, by ‘species-specific, stereotyped, “machine-like”, behaviors that are “immutable in the face of experience”’ (Vicedo 2009: 267). Nevertheless, what will be of special interest here is not so much Lorenz’s conception of animals – brilliantly analysed by Eileen Crist (2000) – as his method of ‘keeping animals’. ‘Keeping animals’, I will argue, was an exercise in species manipulation, that is, an exercise in the manipulation of individual animals in order to identify and confirm, through the disintegration and reassemblage of their behaviours, species identity and genealogy. Contra Vinciane Despret, who argues that the image of Lorenz as ‘a scientist who adopts his animals, swims with his geese and ducks, and speaks with his jackdaws … is faithful to his practice but less so to his theoretical work’ (Despret 2016: 39), my argument will be that Lorenz’s romantic image is betrayed by both, by his theory and his practice/method, which were deeply bound up with, and lent validity to, each other.

Even during his own lifetime, Lorenz’s work was subject to considerable, if not devastating, criticism. Among the best known of these critiques was Daniel S. Lehrman’s 1953 paper, published in the Quarterly Review of Biology, in which Lehrman argued that (among other things)¹ Lorenz’s model of behaviour was not able to distinguish between instinctual and learned elements of behaviour, and that it neglected almost entirely the impact and implications of ontogenetic development. ‘Four decades on’ (the title of his article), Aubrey Manning – who was a student of Nikolaas Tinbergen and one of the early ‘hard core’ members of Tinbergen’s research group at Oxford (Manning 2005: 287)² – recalls Lorenz’s antipathy, particularly in conferences, to any analysis of the way that behaviour ‘develops’ (Manning 2005: 288, emphasis in the original).

The psychologist Richard Held was witness to that antipathy when it was displayed at a conference in Ithaca, New York, in September 1954. As Held describes it, the event was dominated by tense exchanges between the representatives of European ethology, who were perceived to be insisting upon a rigid biological hereditarianism, and the North American comparative psychologists, who ‘continually raise[ed] the question of the role of experience’ in the development of behaviour (Held 1956: 691).³ Although the psychologists were certainly critical of behaviourism, especially insofar as it had ‘become little more than a science of “rat learning”’ (Burkhardt 2005: 362), they nevertheless remained the inheritors of that legacy which had historically been ‘concerned about the ways in which the individual acts upon and transforms his environment’ and which, as such (Held drily added), took ‘the upwardly-mobile American male as the touchstone for psychological theory’. The vision of a potentially perfectible society, made possible by the postulate of the malleability of individual behaviour, was one that the New York psychologists were loath to give up (Held 1956: 692).

It may seem surprising that there is any place at all for behaviourism in this chapter, given that behaviourism is usually associated with the very opposite of species-typical behaviours, that is, with the transformation of individual behaviour through learning. The 1954 conference staged that contrast, and was especially interesting for it (see Burkhardt (2005: 398–403) for an extensive and nuanced discussion). Yet, when it comes to species identities, the difference between behaviourism and other scientific schools of thought may not be as stark as it first appears. This is why I turn, in the second part of the chapter, to Conwy Lloyd Morgan. Although the origin of behaviourism is usually credited to John Watson – who coined the term in 1913 – B. F. Skinner argues that its central principles were first conceived of in the minds of Morgan and his student Edward L. Thorndike (to whom Watson acknowledged his debt) (Skinner 1959). It was Morgan’s claim that an instinctive behaviour pattern, acquired through adaptation, may be modified by learning and therefore, through experience, transformed into ‘intelligent’ behaviour. Nevertheless, not only were Morgan’s theories of behaviour informed by species thinking, as I will demonstrate; so too his methods operated like a species theory insofar as they had the effect of erasing the significance of the particular, singular, individual animal – the animal who, in this case, was Morgan’s own dog, Tony.

This brings me finally to the third section of this chapter, which is concerned, largely, with the relationship between the ‘anecdotal tradition’ – best exemplified, perhaps, by Darwin’s protégé, George Romanes – and species thinking. Both Morgan and Thorndike were critical of Darwin’s and Romanes’s anecdotalism, in keeping with the burgeoning professionalisation of science that characterised their (and especially Thorndike’s) time. Yet two centuries of professionalism have not expunged anecdotes from science. On the contrary, contemporary cognitive ethology is sometimes described, in derogatory terms, as ‘anecdotal cognitivism’ (Allen and Bekoff 1999: Chapter 2). Although the anecdotal method is often associated with individual animals (one reason for its supporters to support it, and its critics to criticise it), as I will show in this final section the use of anecdotes – from Darwin through to cognitive ethology today – may nevertheless confirm rather than subvert species thinking, and so transform the individual animal into a species ambassador.

Unlike other chapters in this book, the ‘species-making’ methods and methodologies that I address here could apply – indeed, they are usually expressly designed to apply – to nearly all animals. Thus, although I try to use scientists’ examples of dogs wherever possible, and although these techniques certainly have implications for dogs, they are not specific to dogs. Tony the Fox Terrier played an important part in the elaboration of Morgan’s Canon (on which see below, pp. 104–109), but the point of the Canon is that it is as relevant to the study of the behaviours of ducks as it is to the study of dogs. Such general techniques are nevertheless worthy of attention for precisely this reason: they are illustrative of the ways that species thinking sweeps up multitudes – multitudes of individual animals within a species category (Lorenz), and multitudes of individual animals regardless of their species categories (Morgan) – in its non-discriminatory wake. There is another reason, however, why dogs serve mainly as examples in this chapter. As I will discuss in Chapters 4 and 5, dogs did not become the subjects of systematic scientific research ‘in and of themselves’ until the late 1990s.

Keeping animal species (the ethological tradition)

Lorenz’s first scientific paper was on Tschock, a female jackdaw whom he had bought in a pet shop, and who proved so attracted to him that he did not have to keep her in an aviary. Tschock’s relationship with Lorenz gave him particular observational advantages: he was in ‘possession’ of a wild animal who was neither afraid nor confined to a cage and who therefore ‘performed many of its natural behavior patterns in Lorenz’s presence’ (Burkhardt 2005: 134). While this relation clearly suited Lorenz’s circumstances and his temperament – ‘the animal that could escape and yet remains with me affords me undefinable pleasure, especially when it is affection for myself that has prompted it to stay’ (Lorenz 2002b [1949]: 5) – in time ‘keeping tame individuals of wild species under free-ranging conditions’ came to be recognised, as Tinbergen wrote in a draft of his obituary for Lorenz, ‘as one of the basic methods of behavior research’ (Tinbergen in Burkhardt 2005: 480).

Lorenz often likened himself to ‘the farmer’ and Tinbergen to ‘the hunter’. Historian Richard Burkhardt explains the difference: ‘Lorenz liked raising and breeding animals, nurturing them when they were ill,⁴ and having them as companions. Tinbergen preferred stalking animals in the field, matching wits with them, and discovering how the details of their behavior contributed to their survival’ (Burkhardt 2005: 11). Tinbergen’s practice – which is described with humour by Burkhardt as ‘crouching in a hide and spying on a creature’ (Burkhardt 2005: 132) – had been shaped partly by the widespread social activity of nature study in the Netherlands during his youth, and partly by the British field tradition, which included Charles Darwin and Julian Huxley, that influenced Tinbergen during his career at Oxford. As Burkhardt notes, these different methods and practices ultimately led Tinbergen and Lorenz to make different contributions to ethology. Tinbergen’s focus was on the ways animals adapt to the ecological demands of their environments, the survival value of behaviour patterns in natural settings, behavioural evolution and evolutionary convergence. Unlike this ‘ecological-evolutionary outlook’ (Tinbergen in Burkhardt 2005: 475), Lorenz was more concerned with behaviour patterns as ‘historical arrangements’ of survival value (Burkhardt 2005: 419, emphasis in the original). That is, he understood behaviour patterns to be a key – if not the key – to species: to the identification of the evolutionary history of a species and of the distinctions between species, and to the confirmation of species identities.

This notion that behaviour patterns might offer a kind of ‘evolutionary genealogy’ of a species can be traced in part to two important influences on Lorenz: the North American zoologist Charles Otis Whitman, and the German zoologist Oskar Heinroth (Tinbergen 2005 [1963]: 298). In Lorenz’s view, ‘C. O. Whitman and O. Heinroth were phylogenists and not physiologists. Their chief interest in innate behaviour patterns was of a systemic and taxonomic nature’ (Lorenz 1950: 246). ‘Instinct and organs’, Whitman wrote, ‘are to be studied from the common standpoint of phyletic descent’ (Whitman in Lorenz 1950: 238). Having first learned how to study evolution through comparative anatomy (which at that time was concerned less with the mechanisms of evolution and more with homologous structures), Lorenz too believed that behaviour patterns could be used ‘to determine common ancestries and reconstruct phylogenies’ (Burkhardt 2005: 134). As he put it:

One of the analytical implications of this conception of behaviour patterns is that it makes it possible ‘to isolate a very distinct physiological process as an independent constituent of behaviour and to study it separately’ (Lorenz 1950: 238, emphasis in the original). These distinct processes, Lorenz wrote, are ‘particulate elements’ of behaviour (Lorenz 1950: 238, emphasis in the original). And they are unchanging in a species. Lorenz argued that the most an animal’s experience can do is to determine ‘the intensity with which the instinctive response is performed’, and perhaps also ‘which response is elicited by a particular stimulus’ (Lorenz 1970 [1937]: 268, emphasis in the original). For instance: even though the ‘escape response’ may diminish as an animal becomes increasingly tame, it could ‘at any time be elicited by a specific frightening stimulus of a particular strength, without any preceding experience’ (Lorenz 1970 [1937]: 268, emphasis in the original).

Although Eileen Crist insists that ‘neither Tinbergen nor Lorenz wanted to “desubjectify” animals’ (Crist 2000: 89), as she illustrates, Lorenz’s analysis of instinctive behaviour patterns in terms of particulate elements, and as inherited and unchanging in a species, had the effect of bracketing off animal behaviours from the subjectivity of the animal who ‘has’ them. In Lorenz’s view, for instance, it is not the animal who responds to a stimulus; rather, it is the ‘innate releasing mechanism’ that ‘chooses’ and reacts to it, and this reaction itself initiates a pattern of behaviour (Crist 2000: 100). Moreover, according to Lorenz’s ‘hydraulic reservoir’ model of motivation, ‘energies’ that are specific to a single behaviour pattern ‘can be dammed up, accumulated, diverted or used up’ (Beer 2020: 263). If a specific behaviour activity becomes ‘dammed up’, for example (perhaps in the absence of ‘eliciting stimuli’), the threshold for releasing it will become very lowered until finally, as Lorenz explains, ‘the activity in question will … go off in vacuo, with an effect somewhat suggestive of the explosion of a boiler whose safety valve fails to function’ (Lorenz 1950: 247).

Although Lorenz was no less critical of the ‘mosaic’ approach of mechanists than he was of the summative approach of vitalists (Lorenz 1950), it is interesting to note – indeed I think it is something of a giveaway – that he believed that mechanists were likely to be ‘less wrong’ in their analysis of behaviour than were vitalists: ‘the atomistic investigator is not guilty of any methodological error, as long as he really is examining a comparatively independent constituent part. It is just because they may legitimately be isolated theoretically and experimentally that the discovery of independent constituents always is such a tremendous step forward in analytical research’ (Lorenz 1950: 227). Or as he put it rather defensively elsewhere: ‘the existence of the wood as a “holistic” living community is [not] somehow threatened by recognition of the fact that wood just happens to consist of trees as well as other components’ (Lorenz 1970 [1942]: 355).

On the one hand, given what Lorenz says about the holistic living community not being threatened by the components of which it consists, it could be a mistake to contrast too strongly the romantic Lorenz who talks to his animals – some editions of his popular book King Solomon’s Ring include the subtitle ‘He Spoke with the Beasts, the Birds, and the Fish’ (Vicedo 2009: 279) – with the scientific Lorenz whose analytic method disaggregates behaviour from subjectivity and experience and divides it into particulates. On the other hand, however, Lorenz deployed and honed his somewhat Dr Dolittle practices precisely because he considered ‘keeping animals’ to be ideally suited both to the reconstructive phyletic task in the zoological system, and to the identification of instinctive behaviour patterns in the biological system. Keeping animals was a powerful and effective method in both these regards, Lorenz believed, because captivity induces what he called ‘miscarrying behaviours’, that is, behaviours that are performed ‘wrongly’, or in an inappropriate context, or in the absence of eliciting stimuli, or with a substitute object. This is why the ‘completely-satiated domestic dog’ will nevertheless ‘[shake] his master’s slippers “to death”’ (Lorenz 1970 [1942]: 360): because ‘[a Dachshund or terrier] retains an unaltered, irrepressible appetite’ for the performance of the motor pattern of shaking prey (Lorenz 1970 [1942]: 360; see also Lorenz 2002c [1963]: 84–87).

With respect to the reconstructive phyletic task, for example: keeping animals compels the observer to deduce from miscarrying behaviour patterns what were the environmental conditions that gave those patterns their ‘normal’ adaptive value. ‘Just as he would deduce from the morphological characters of a mole’s forepaws that this species needs earth to dig in’, Lorenz wrote, ‘so he must, from slight “hints” of miscarrying behaviour patterns, be able to deduce the corresponding environmental exigencies of the species’ (Lorenz 1950: 237). With respect to the identification of instinctive behaviour patterns: under conditions of captivity, particularly if the keeper is not familiar with the ‘needs’ of a species, instinctive behaviour patterns often disintegrate into ‘a jumble … of parts’ (Lorenz 1950: 236). When the species’ needs are restored by the keeper, so too will be the behaviour patterns. ‘I hardly know a more instructive object of observation than just this type of disintegration and reassembling of the system of actions in animals kept in captivity. It is, in fact, an actualized example of analysis and resynthesis of behaviour!’ (Lorenz 1950: 237, emphasis in the original).

Regardless of Lorenz’s enthusiastic conviction here, I would argue that even though the purpose of the method appears to be the observation of behaviour, in fact is it the observation of species; or more strongly, it is a method of species-making. ‘Keeping animals’ offered Lorenz three things: a retrospective insight into the becoming of species (‘this species needs earth to dig in’); a means to identify fragments of species behaviours by way of their undoing (into a ‘jumble … of parts’); and a confirmation of the integrity of species behaviours/species identities in their redoing (in the ‘reassembling the system of actions’). Animal behaviours, which could potentially lead anywhere or mean anything – a Deleuzian ‘line of flight’, to go to the other extreme – lead only to, and reveal only, species. Not only are behaviours theorised by Lorenz as mechanistic, therefore; the observation of behaviours under these particular conditions becomes, in Lorenz’s hands, a powerful way of keeping an animal captive in the cage of species.

Thom van Dooren’s discussion of the ethics of Lorenz’s imprinting is relevant here. For van Dooren, Lorenz’s relationships with the birds on whom he imprinted himself represent not so much novel ‘possibilities for connection and care’ as they do a ‘captive form of life’ (Van Dooren 2014: 103, emphasis in the original). This form of life is captive because the relationships, rather than being constituted ‘between two subjects, who – however unequally positioned – already have a significantly well-formed way of life, a way of being in the world’ (101–102), were ‘knowingly manipulated’ by Lorenz, at an early stage of development, with the specific intention of creating ‘a lifelong attachment’ (103). Although imprinting is a very particular mode of biological, social and subjective manipulation, I think it is an apt metaphor for Lorenz’s method (keeping animals) more broadly. I say this not because to be a kept animal is to live a life in captivity (although it is, in a general kind of way), but rather because Lorenz himself considered keeping, as he considered imprinting, to be scientifically valuable because it is a manipulation, because it is an interference, because it induces miscarrying behaviours. Because, in short, it creates a distortion of or deviation from ‘species-typicality’ that enables species and species norms to be identified, ‘confirmed’ and consolidated. In On Aggression, Lorenz stated that ‘[p]‌hysiology, the science concerned with the normal life processes and how they fulfil their species-preserving function, forms the essential foundation for pathology, the science investigating abnormalities’ (Lorenz 2002c [1963]: 27). One might argue that the relation between physiology and pathology is inverted by Lorenz’s method: now, keeping animals is the essential pathology-inducing foundation, designed to enable the investigation of the species-preserving physiological norm. The fact that Lorenz may have diligently cared for his birds,⁶ or that he believed that the attachment of his animals to him was freely given and, as I cited earlier, based on ‘affection for myself’ (Lorenz 2002b [1949]: 5), does not preclude this method from being coercive, especially insofar as miscarrying behaviours are produced ‘at the expense of a whole set of other ways of being’ (Van Dooren 2014: 103, emphasis in the original) – and at the expense, I would add, of a whole set of other ways of understanding being.

This chapter, as I noted in the introduction, is concerned not only with species-making, but also with one of its most important consequences, the transformation of particular individual animals, singular animals, into generic species representatives. One way of achieving this, as I have been demonstrating, is to deploy species identity and species-typicality as the lens through which all behaviours are viewed. This erasure of particularity is, ironically, especially well dramatised in Lorenz’s popular books – ironically, because in these books Lorenz refers to named animals, whom he describes through personalised stories and anecdotes. Yet it is this very individualisation that draws the reader’s attention to the vanishing of the animals into instinctive behaviour patterns. Lorenz’s popular books, such as Man Meets Dog (2002b [1949]) and King Solomon’s Ring (2000a [1949]), cannot be lightly dismissed, not only because they convey, in an accessible form, some of the themes and ideas that Lorenz developed in his scientific work, but also because they created a powerful and enduring image of what it was to be, to live and to think as an ethologist. Indeed the style and significance of these books arguably secured for Lorenz, and perhaps for ethology as well, considerable public acclaim. It also helped to fund his research at Altenberg, Austria, where Lorenz had built a research station on his father’s estate (Burkhardt 2005: 11).

At the time of writing Man Meets Dog in 1949, Lorenz was engaged in debates over whether the dingo was ‘a true wild dog’ or a domesticated dog turned wild, as Lorenz thought (Lorenz 2002b [1949]: 116).⁷ In part in order to answer this question, Lorenz devised an experiment to test whether and how a domestic dog would rear a dingo, and how the dingo would behave in response. As I noted in the introduction to this book, he did this by presenting a dingo pup to his own dog, Senta, who had whelped at approximately the same time as a dingo in Schönbrunn Zoo. I recount Lorenz’s description of this episode because it illustrates, again, but here in a most visceral way, the uses he made of interference and manipulation in his work. It illustrates his interference with Senta, and with the dingo puppy, whose life – if Lorenz’s analysis of the ‘brood defence’ is to be believed – he put at great risk. I also recount this anecdote because, as it turns out, it is not about the particular individual, Senta, at all.

The story begins with a portrayal of Lorenz at his ‘[e]‌bullient and egocentric’ (Burkhardt 2005: 5) best, rushing between the zoo and a funeral with the dingo pup in his dispatch case. It also illustrates him at his narrative best, moving from humorous self-portrayal to sober pedagogy to dramatic suspense. Lorenz teaches his reader that the best way to encourage ‘a mammal mother to adopt a strange baby’ is to elicit the ‘brood-tending instinct’, which can be done by presenting the infant to the mother ‘outside her nest and in as helpless a form as possible’ (Lorenz 2002b [1949]: 119). He then notches up the tension by advising that stimulation neither of the brood-tending instinct nor of ‘the carrying reaction’ (carrying the pup back to the nest) is a guarantee that, once in the nest, the infant will not anyway be ‘recognised as an intruder and remorselessly devoured’ (Lorenz 2002b [1949]: 120). Lorenz describes this devouring in some detail, underscoring that, if it is going to happen, it will begin at the infant’s abdomen, since it is a ‘defect’ in the process by which a mother removes the foetal membrane and placenta from the newborn and severs the navel-cord. If the process does not stop at that point, ‘the abdomen of the young is also opened at the umbilicus’ (Lorenz 2002b [1949]: 122). The scene is thus set, and the reader understands the stakes. The ‘action’ that follows, however, turns not so much on Senta and the dingo puppy as it does on the battle between two instinctive drives, elicited by two different stimuli: the dingo pup’s helplessness (which elicits the brood-tending instinct in Senta) and his strange smell (which elicits the ‘brood defence’ instinct and the devouring impulse).

First, Senta licks the dingo’s belly carefully. But then she begins to nip the skin with her teeth. When the dingo cries out and whimpers,

This passage – for all its chummy vernacular – exemplifies how, by breaking down behaviour into physiological ‘particulates’ (single behaviour activities), by attaching these particulates to specific stimuli and by transferring the initiation of behavioural activity from the animal to innate releasing mechanisms, Lorenz simultaneously breaks down the animal’s – in this case Senta’s – world such that it appears to possess neither ‘experiential unity’ nor ‘continuity’ (Crist 2000: 100). Lorenz is not narrating a sequence of events here, a joined-up, temporally continuous sequence in which, for example, Senta hears an infant, responds to his cry, and then realises she has misrecognised him as one of her own, etc. Instead, her reactions are understood by him to take place serially: they are released by two different stimuli (Lorenz would argue that the pup, from Senta’s perspective, is two different objects), and they occur ‘in discontinuous – even if contiguous – pockets of time’ (Crist 2000: 105). Not merely, then, does Senta have no experience, she has no experience of experience either. As Crist notes in her analysis of Lorenz’s account of a mallard duck who rescues a musk duckling, from which I have taken this valuable distinction between sequence and serial, the awareness of the animal of her own situation is not in fact required ‘for the contact between subject and object’ (Crist 2000: 104). The reason for this is that ‘the object gathers all the necessary and sufficient features for the elicitation of the proper behaviors; it need not be known, assessed, understood, recognized, misrecognized, or witnessed by the subject of action’ (Crist 2000: 104). The implication – which Lorenz does not spell out in Man Meets Dog but which any engaged reader can deduce – is that, even in a ‘normal’ situation, in which a mother retrieves a pup of her own, ‘it only appears that her actions manifest an understanding of objects, events, and their connections’ (Crist 2000: 105, emphasis in the original).

Lorenz’s interference with, and manipulation of, both Senta and the dingo puppy melts away in the heat of the dramatic narrative. The desubjectifying implications of Lorenz’s conception of instinctive behaviour, by contrast, are exaggerated by the narrative, partly because the story is intentionally told with heightened suspense to appeal to a general public, and partly because the stimuli and mechanisms that are driving Senta’s actions (or reactions, rather) are in conflict with each other, which renders their ‘independence’ from her subjectivity all the more stark. As Despret says, Lorenz’s concept of instinct offered him ‘the perfect cause: it escapes from all subjective explanations, and it is at once a biological cause and motive (a motive, moreover, that completely escapes the knowledge of the subject himself)’ (Despret 2016: 40). The reader cannot help but identify with Senta’s anguished cry, which brings Lorenz’s anecdote to a close. Senta, Lorenz writes, ‘sat back on her haunches in front of the Dingo, raised her nose to the sky, and gave vent to her distress in a long, wolf-like howl’ (Lorenz 2002b [1949]: 124).

Senta’s howl – her ‘inward torment’, her ‘suffering’ (Lorenz 2002b [1949]: 124) – represents for Lorenz the suffering of a canid being torn apart by two opposing specifies-specific instinctive behaviour patterns, behaviour patterns that Senta, because she is a canid, ‘has got’ – or which, more accurately, might be said to ‘have got’ Senta. Alternative interpretations are not available to Lorenz because that would require ‘one [to] see that if the animal responds by using his own way to arrive at articulating the problem, he no longer responds to the question “in general.” Which means that his response has nothing generalizable about it’ (Despret 2016: 93). Senta may, for example, already be familiar with Lorenz’s experiments (even if she does not understand them as such). Or she may be exhausted from feeding her newly born infants and frustrated by being obliged to address the question of this new puppy stranger. Or both. Or perhaps it is something else altogether. The howl could even be the sign of a resolution, a resolution neither to devour the dingo, nor to take him in contentedly. There is some evidence for this in Lorenz’s account of the conclusion of the episode: he notes that even though Senta ‘suckled [the pup] with her own, one day she bit him so severely in the ear that it never properly recovered its shape and ever after drooped to one side’ (Lorenz 2002b [1949]: 124). Whatever the reasons, these would be the reasons of the individual Senta, which, if given space for consideration, would interfere with the story that Lorenz is telling about the instincts that drive the behaviour of a female member of the species Canis familiaris. The howl would be Senta’s response, a howl that another dog might not sound.

Species scales (the behaviourist tradition)

In the introduction to this chapter, I referred to a conference, held in 1954, that brought together, in great tension, European ethologists and North American psychologists. The tension lay in the perceived contrast between the ethologists’ emphasis on the fixity of heredity in determining behaviour, and the psychologists’ emphasis on the pliability of experience and learning in shaping behaviour. In his review of that event, the psychologist Richard Held was nevertheless led to wonder about the opposition. ‘After all’, he wrote, ‘we know that American psychology has been enormously influenced by ideas of biological evolution’ (Held 1956: 691). And indeed, this is where this section of the chapter begins, with the fact that comparative psychologists in (what was to become) the behaviourist school also, as the philosopher Bernard Rollin writes, have ‘their own version of Darwinism’:

This, for Rollin, explains why behaviourists are ‘extremely interested in animals’ and why so much of behaviourism relies on animal experimentation (Rollin 1998: 207). The implications of this behaviourist ‘version’ of Darwinism, for species-making through method, is especially well demonstrated by Conwy Lloyd Morgan’s Canon, which is the specific focus of the following discussion.

Morgan’s Canon is ‘possibly the most important single sentence in the history of the study of animal behavior’ (Bennett Galef in Steward 2018: 293). As the philosopher Helen Steward summarises it:

The Canon states: ‘[i]‌n no case may we interpret an action as the outcome of the exercise of a higher physical faculty, if it can be interpreted as the outcome of an exercise of one which stands lower in the psychological scale’ (Morgan 1903: 53; emphasis omitted). Today, as Steward indicates, Morgan’s Canon is often used to chastise and curb any interpretation of an animal’s behaviour that is deemed to be too generous, that is, which gives too much ‘credit’ (usually ‘intelligence’) to the animal. It is, in short, understood ‘as simply the application of the general law of parsimony to explanations of behaviour’ (Boakes 2008: 40).

Yet Morgan, the psychologist Robert Boakes argues, justified his Canon neither in terms of animal intelligence nor in terms of parsimony. Rather, he justified it ‘on the grounds of evolutionary theory’:

In other words, why would the response of a species be characterised by more complexity than is needed to adapt to the selective pressures that shaped its evolution? The ‘psychological scale’ to which the Canon refers is an evolutionary scale. Species and scales of species complexity are integral to the Canon. This is why ‘[t]‌he major theme of the Introduction to Comparative Psychology … discuss[es] in turn processes of increasing psychological complexity’ (Boakes 2008: 41). To this end, Morgan begins with an analysis of ‘simple associations’ (all animals are capable of this), and then moves to ‘perceptions of relations’. Finding no evidence of such perceptions in any animal, the remainder of the book is dedicated to human psychology.

Morgan famously considered anecdotes to throw ‘a misleading glamour over what were not more than special tricks’ (Boakes 2008: 35). Yet it was the special (or not so special) tricks of his Fox Terrier, Tony, recounted by Morgan in the form of two anecdotes, that furnished evidence for Morgan’s claim that animals do not perceive relations. Although contemporary canine scientist Brian Hare (see Chapter 4 of this book) describes Morgan’s anecdotal accounts of Tony’s antics as ‘a classic example of how complex behaviour can be explained by simple forms of cognition’ (Hare and Woods 2020a: 168), they might also be understood as a classic example of how species thinking can transform an individual animal into a representative of their species. In what follows, therefore, I address the implications of the anecdotes not for what they tell about the Canon, which is Hare’s interest, but for what they tell about the individual dog, Tony, who lies – or rather, who bolts out of gates and fetches sticks enthusiastically – at their heart.

Where George Romanes’s correspondents were particularly impressed by their dogs’ understanding of mechanical appliances – believing them to be examples of canine reasoning – Morgan asks, in a chapter entitled ‘Do animals reason?’, ‘whether Tony’s behaviour can be fairly explained without his forming any conception of the relation between the means employed and the ends attained’, and answers: ‘It appears to me that it can’ (Morgan 1903: 292). The behaviour to which Morgan is referring is Tony’s ability to raise the latch on Morgan’s gate. Tony was able to do this, Morgan explains, not because he understood how the latch works (which would constitute evidence of reasoning), but because, one day, when Tony was standing with his head beneath the latch and between the bars of the gate – ‘looking restlessly and wistfully at the familiar road’ (293), ‘where there was often much to interest him; cats to be worried, other dogs with whom to establish a sniffing acquaintance, and so forth’ (292) – he by chance raised his head. The latch lifted, but Tony was looking elsewhere. It was only when Tony noticed the gate swinging open that ‘out he bolted’ (293). After this event, Morgan, instead of raising the latch himself, waited for Tony to do it until Tony was able to go ‘at once and without hesitation to the right place and put his head without any ineffectual fumbling at the right place under the latch’ (293).

In their discussion of anecdotes in the animal sciences, Paul Morris, Margaret Fidler and Alan Costall write that ‘[s]‌ome may object that Morgan made repeated and careful observations of his dog and should not be deemed anecdotal; however, the same may be said of many observations of animal trainers, farmers, and pet owners in general’ (Morris et al. 2000: 152). It was against the ‘somewhat rough and ready interpretation[s]’ of ‘the man who has to deal with animals for practical purposes’ (Morgan 1903: 52), however, that Morgan distinguished his own observations, and the science of psychology. Such men include ‘[t]he farmer, the keeper of a kennel, the cattle-breeder, the gamekeeper, the breaker-in of horses, all the practical men who are employed in the breeding, rearing, and training of animals, and the great number of people who keep animals as pets in domestic service’ (Morgan 1903: 51–52).⁸ The problem with ‘rough and ready interpretations’, Morgan thought, is that they fail to recognise, and therefore fail to redress, their subjective dimension. This subjective dimension, which applies to the study of both humans and animals, therefore requires what Morgan calls the ‘doubly inductive’ method. The doubly inductive method has not only an objective aspect – induction of the kind witnessed in chemistry, physics, astronomy, geology etc. – but also, necessarily, a subjective one. Necessarily, because ‘the psychologist has to reach, through induction, the laws of mind as revealed to him in his own conscious experience’ (Morgan 1903: 47). Morgan is speaking here to any person who imagines that, when it comes to the ‘psychical faculties of animals’ (50), the subjective dimension of the method that applies to the study of humans can be either bypassed or ignored. Better, Morgan argues, to be fully cognisant of the role that the doubly inductive process plays in analysing the psychology of animals, than to be ignorantly given to common-sense explanations: to the kinds of explanations of animals that, for example, one would use to account for the actions of one’s ‘human neighbours and acquaintances’ (50).

With regard to Tony, one can surmise that Morgan would probably distinguish his interpretation of his vignette from any other interpretation of it, because not only does he ‘know the whole history of it’ (Morgan 1903: 293, emphasis in the original) – as animal trainers, farmers and pet owners usually know the history that informs an individual animal’s behaviour – he also understands the significance of that history, which in this case is that it took Tony ‘nearly three weeks’ (293) to perfect the trick. The reason it took ‘so long’, Morgan writes, was that ‘there was so little connection between gazing out into the road and getting out into the road’ (293). The time it took to learn the trick, in other words, is the evidence that Tony has no perception of the relations between ‘means and end’ (293). Had Tony such a perception, not only would he have learned the trick sooner, it would in fact not be a trick at all. If this hardly seems ground-breaking, Morgan spells out its implications more clearly in another anecdote, which is in addition intended to support his case against anecdotal evidence. The incident once again involved Tony and his tricks.

During the course of a series of ‘experimental investigation[s]‌’ (Morgan 1903: 255) involving Tony and some sticks, Morgan ‘prepared a short yew stick with a crook at one end’ (Morgan 1903: 157), which he threw over a fence for Tony. When Tony attempted to return it to Morgan, the knotted end would inevitably catch on the fence, which left Tony ‘tugg[ing] at it in the most ridiculously energetic fashion’ (257). After several failed attempts, Tony seized the crook and wrenched it off, which enabled him to get through the fence with the stick. He did so just as a passer-by ‘paused for a couple of minutes to watch the proceedings’. The passer-by then turned to Morgan and said ‘Clever dog that, sir; he knows where the hitch do lie’ (258). Which remark, Morgan writes,

The power and persuasiveness of Morgan’s case for his Canon rest here on the very evidence (anecdotal evidence) that Morgan wishes to dispute: it enables him to illustrate that although one could impute ‘cleverness’ to Tony on the basis of his behaviour, as did the person who conveniently happened to pass by (so conveniently as to be almost suspicious),⁹ Tony’s performance could equally – and, Morgan believes, correctly – be explained not by the successful strategy of wrenching the crook off the stick, but by the numerous imperceptive efforts and failures that preceded it. Indeed it is the very numerousness, as well as the perceived arbitrariness, of these efforts that demonstrate to Morgan that, even after Tony wrenched the crook off the stick, he remained ignorant of the mechanics of the size of the stick, the size of the gaps between the vertical rails and the obstructiveness of the crook. Morgan relies here on a classic anecdotal trope, which is an observer watching, by chance, as unexpected events unfold. The difference is that, in Morgan’s anecdote, there are two observers: the passer-by who watches the dog, and Morgan, who watches both the dog and the passer-by. It is in this position, of the meta-observer in the meta-anecdote, that Morgan is enabled to pass judgement on anecdotal evidence and to ‘prove’ that the simplest explanation is more likely to be the correct one.

As I noted earlier, contemporary scholars and scientists continue to debate what Morgan intended by his Canon, how it should be interpreted and whether it is of value. These discussions add nuance to the Canon and to Morgan’s own conception of it. Nevertheless, broadly speaking, the Canon has become crudely associated with the general law of parsimony, and with the banishment of understanding, intention, motive and feeling from the science of behaviour.¹⁰ One might say that Morgan’s Canon marked the moment when animals lost their minds (Bekoff and Jamieson 1992) or, perhaps, when animal scientists did (Rollin 1998). There is another vanishing point here, however, which is Tony himself. For what truly marks the difference between Morgan’s anecdote and the anecdotes of ‘animal trainers, farmers, and pet owners in general’ is that it is, in the end, not about a specific dog at all. In Morgan’s vignette, Tony is a cipher for all dogs, and for all dogs’ limited problem-solving abilities. As significant, then, as Morgan’s dispute with the passer-by’s misguided interpretation – ‘clever dog’ – is his objection to the passer-by’s attention to Tony – ‘clever dog that, sir’. In this regard, the Tony anecdotes bear one of the key hallmarks of species thinking: they turn Tony the individual into an ambassador for his species.

Despite the ‘opposition’ between classical ethology and psychology, psychologists also often took their cue from Darwin’s theory of evolution. The origins of Morgan’s ‘trial-and-error learning by accident’, for example, lie in his studies of ducklings and chicks, in which Morgan showed how arbitrary associations – akin to the arbitrary associations that he thought explained how Tony came to open the latch – ‘might work in a natural environment to produce adaptive behaviour’ (Boakes 2008: 35). By changing the environments of ducklings and chicks, and watching them either repeat a behaviour to no purpose or develop a taste aversion for no explicable reason, Morgan concluded that ‘behaviour is modified by its immediate consequences’ (Boakes 2008: 35). Such consequences could apply as well to an ecological niche as they could to an artificially manipulated environment. Eighty or so years later, in an article entitled ‘Selection by consequences’, Skinner would draw a parallel between natural selection by consequences and ontogenetic behavioural selection by consequences (Skinner 1981). Gone was the conquering individualism of the 1954 conference. Instead, Skinner argued that just as living creatures are not the agents of evolution, so they are not the agents of their own actions: ‘so long as we cling to the view that the person is an initiating doer, actor, of causer of behavior, we shall probably continue to neglect the conditions which must be changed if we are to solve our problems’ (Skinner 1981: 504). Perhaps the difference between the ethologists and behaviourists lies in the uses they made of evolutionary theory. For the classical ethologists, processes of evolutionary adaptation explain the mechanical behaviours of a species. For the behaviourists, adaptation is to be explained by mechanics.

‘I once had a dog …’ (the anecdotal tradition)

Morgan, Skinner wrote, distinguished himself in the history of the scientific study of behaviour by showing how ‘evidences [sic] of mental processes could be explained in other [than anecdotal] ways’ (Skinner 1959: 197). Yet his student Edward Thorndike criticised Morgan – as Morgan had criticised George Romanes – for belonging to the ‘anecdote school’ (Thorndike 1911), and for conducting only a handful of informal studies on birds and a dog. Despite this, Thorndike was much influenced by both Romanes’s and Morgan’s anecdotes, and especially by their descriptions of the mechanical abilities of animals. ‘Thorndike’, Boakes writes, ‘took the kind of situation described by Romanes’ correspondents and also by Morgan in the latter’s account of how the fox terrier learned to operate the latch of a gate, and turned it into an experimental method’ (Boakes 2008: 69).

The most famous of Thorndike’s experimental methods is probably the ‘puzzle box’, precursor to ‘the Skinner box’, which Thorndike devised by cutting doors into wooden crates. By way of various devices, these doors could be opened by the cats and dogs whom Thorndike put into the boxes. Ultimately, the puzzle boxes led Thorndike to his ‘law of effect’, which, at its most basic, states that satisfaction leads to reinforcement: i.e. if a response to a stimulus ‘works’ for an animal, then they will do it again (and again, and again). ‘What had started as an explanation for the manner in which animals learned to escape from his puzzle boxes’, Boakes writes, became ‘a general law of behaviour’ (Boakes 2008: 75). If Morgan’s Canon struck the death knell for Tony as an individual, Thorndike’s puzzle box would be his tomb. Gone is the rich particularity of the latch that Tony opens, of the gate that Tony bolts through, of the dogs that Tony sniffs, and of the cats that Tony worries. Instead, the individual dog is relevant only as a point on ‘a curve representative of learning’ (Cladland 1993: 245).

There is no room for anecdote here, not only because anecdotes do not (usually) ‘represent what happens on average’ (Crist 2000: 43) but also, more fundamentally, because Thorndike was interested in how individual animals of the same species act in similar ways (which tells something about the species), and not in how individual animals differ (which tells something about these individuals). This was the gist of the critique of T. Wesley Mills – who was, Douglas Cladland writes, one of Thorndike’s most thoughtful contemporaries – when he asked what Thorndike’s representative curves tell ‘about this animal, or that one?’ (Cladland 1993: 245, emphasis in the original). The answer is that they tell nothing, because they are not intended to. Thorndike was speaking to species, and would ultimately make the ‘nonsense’ claim, as Boakes puts it, that one species can be judged more intelligent than another ‘simply because it learns how to solve some particular problem more rapidly’ (Boakes 2008: 71).

My contrast between Morgan and Thorndike (and Thorndike’s view of Morgan) implicitly establishes a number of dualisms: storytelling v. science; particular v. generalisable; concrete v. abstract; spontaneous event v. planned experiment; individual v. species. These are some of the dualisms that subtend debates about the value or not of anecdotes in science. My purpose in this final section – in which I explore the anecdotal tradition as it is found in the work of Romanes, Darwin and (briefly) at the birth of contemporary ethology – is not to question these dualisms but, rather, to bring some nuance to them. As I will illustrate, ‘even’ anecdotes – which often stage the spectacular behaviours of an individual animal – can support species thinking.

Contemporary literature on anecdotes in both the sciences and social sciences usually understands their contested status to derive from their association with the concrete, the particular, the specific, the local, as well as from their association with novel and/or rare situations and events (Bates and Byrne 2007; Byrne 1997; Crist 2000; Philo and Wilbert 2000; Lestel 2011). But anecdotes are also nearly always about the actions of an individual animal, or a small group of individuals, and this too, as George Romanes knew only too well, makes them an awkward method through which to formulate and justify general scientific claims. Part of the reason that George Romanes’s successors would see him ‘only as the archetypal purveyor of anecdotes about animals’ (Boakes 2008: 25), Boakes argues, is that Romanes made the fateful decision to divide his research into two, and to publish, first, evidence of ‘animal intelligence’ – evidence that mostly takes the form of stories about individual animals’ feelings and behaviours – and, second, ‘his general principles for the theory of mental evolution’ (Boakes 2008: 25). This division, and its implications for Romanes’s reputation as a scientist, now looks like a portent of the bifurcation of knowledges of animals (a bifurcation that was hardening during Romanes’s lifetime) into the ‘scientific’ on the one hand, and everything else on the other.

Romanes himself knew that his work, and especially the first volume of his book Animal Intelligence, ‘may well seem but a small improvement upon the works of the anecdote-mongers’ (Romanes 2012 [1884]: 20). One of the sources of anecdote-mongering that Romanes undoubtedly had in mind was the flourishing nineteenth-century publishing industry that was transforming ‘a long anecdotal tradition’ into sentimental books about animals, and especially about dogs (Thomas 1984: 108). ‘Whereas earlier dog literature’, Harriet Ritvo writes, ‘seemed simply a specialized branch of natural history, the new books included not only descriptions of the dogs’ physical and moral characteristics, but a selection of heartwarming and enlightening anecdotes’ (Ritvo 1987: 87). And to be sure, the first volume of Animal Intelligence, and in particular the chapter on dogs, reads – as Romanes anticipated it would – as a series of anecdotes about individual dogs, unsupported by any ‘general principles’. Indeed, the chapter on ‘the dog’ is composed almost exclusively of cosy vignettes, many of which begin with ‘I had a dog …’, or ‘I have a setter …’. While this is to be expected, given that all the dogs described in this book either lived with or were known to Romanes or lived with and were known to his correspondents, it nonetheless represents a serious methodological flaw.

A flaw, not because these are descriptions of individual dogs, but because, Boakes argues, Romanes and his correspondents were members of the same class, and the authority of the descriptions often seems to derive less from the narrators’ powers of observation and analysis, and more from their social status (Boakes 2008: 26). Romanes extended this social status to dogs, whom he often classified as either ‘low-life’ or ‘high-life’ (Romanes 2012 [1884]: 1072). This had Morgan bristling, in a lecture published in the journal Mind, in which he recounts an anecdote narrated by Romanes, in which Romanes describes how Mr St John’s retriever cut off his relations with his ‘humble friends’ – ‘a rat-catcher and his cur’ – on sight of his ‘master’ approaching (Morgan 1886: 180). Such ‘caste’ interpretations of dog behaviours cannot suffice as scientific evidence, Morgan argued, because they attribute ‘motives and underlying states’ to dogs (Morgan 1886: 180). Although a more sparing use of such attributions, Morgan continued, would unquestionably exclude as scientific evidence ‘a vast amount of carefully collected anecdote’, science would not ultimately be ‘the loser’ for it (Morgan 1886: 180).

Romanes’s anxieties about the anecdotal status of his work – confirmed as justifiable by Morgan’s and others’ criticisms – indicates that, by the time of his writing, the value of anecdotes as a legitimate source of knowledge about animals had already begun to be depreciated. In Man and the Natural World, the historian Keith Thomas describes how ‘the scientific study of animals, birds and vegetation’ (Thomas 1984: 51) in the early modern period had a ‘traumatic’ effect on ‘ordinary people’ (Thomas 1984: 70). Systematisers, such as the naturalist John Ray in the seventeenth century and Comte de Buffon and Carolus Linneaus in the eighteenth century, began to classify animals and plants on the basis of their ‘intrinsic qualities’ (Thomas 1984: 52) or ‘structural affinities’ (Ritvo 1987: 13) rather than their relationships to humans, and especially their relationships of use and value as food, medicine, or signs and symbols. New technologies, such as the microscope; new ‘content’; and in particular a new, Latin, nomenclature, displaced the copious and varied vocabulary that once described plants and animals. Together, these developments constituted a ‘revolution in perception’ (Thomas 1984: 70), Thomas writes, that entrenched the growing (and classed) division between the knowledges of ‘ordinary people’ who lived and worked with animals, and modern scientific knowledges of animals. Along the way, anecdotes became associated with sentimentality, amateurism and ‘rustics’ (Thomas 1984: 86).

But ‘science’, in the nineteenth century, was hardly consolidated as such. When contrasting the anecdotalism of Darwin and Romanes – and the anecdotalism of Morgan too, according to Thorndike – to Thorndike’s experimental methods, it is impossible not to acknowledge how different were their research contexts, and how significant in shaping Thorndike’s work were the professionalisation of psychology and the rapid expansion and reform of the North American university system during his career, from his undergraduate degree onwards (see Arnet 2019; Boakes 2008: Chapter 3). Darwin, Romanes and, to a lesser degree, Morgan conducted their research during a period in England when natural science was largely ‘a hobby’ confined only to those who could afford it, or who were prepared to face financial penury (Boakes 2008: 54). The kinds of men to whom these men ‘paid attention’ – attention with regard to what they ‘had to say about the behaviour of animals’ – Boakes notes, were ‘the country parson, whose hobby was watching birds, or the colonial officer with time to take an interest in local fauna’ (Boakes 2008: 57). Such folk would also have been publishing in scientific journals such as Nature, whose contributors were not ‘limited to scientific professionals’ (Arnet 2019: 437). Thorndike, by contrast, developed his experimental methods while working as a doctoral research student at Harvard, under the tutelage of the esteemed William James.¹¹

The epistemological tensions that characterised Darwin’s and Romanes’s period, a period of scientific transformation, and of science on the brink of professionalisation, are captured in the difference between the two answers that Thomas Huxley gave to the rhetorical question he posed to his audience at the start of one of a series of talks on dogs that he delivered to the Royal Institution in 1879 and 1880. ‘What’, Huxley asked, ‘is a dog?’ First, ‘[a]‌ dog is a hairy, four-footed, tailed animal … which barks and howls and is often singularly intelligent and affectionate’ (Huxley in Worboys et al. 2018: 163). Second – and this is the answer, Huxley said, that a ‘scientific zoologist’ would give (Huxley in Worboys et al. 2018: 163) – a dog is described in terms of its class, Mammalia, and species, Canis familiaris. Although Huxley is known as ‘Darwin’s bulldog’, as Darwin’s ‘foremost champion’ (Boakes 2008: 5), he nonetheless cast doubt on Darwin’s approach to animals during Darwin’s lifetime, and this doubt was symptomatic of less generous ways of conceiving of both animals and humans. Darwin’s continuity of animal and human physiology, cognition, and emotions did not, for Huxley, point to the richness of animals’ lives. On the contrary, it suggested to him that humans, like animals, are ‘conscious automata’ (Huxley in Boakes 2008: 20). Where Huxley described himself as a ‘doubting Thomas’, Darwin nearly always gave animals the benefit of the doubt. Huxley was a sceptic, willing to suspend belief in the service of science, while for Darwin, scepticism is a ‘frame of mind which I believe to be injurious to the progress of science’ (Darwin in Boakes 2008: 26). In the end, Huxley’s scepticism became his student Morgan’s Canon. The step from understanding dogs in terms of the species Canis familiaris, to understanding dog behaviours as uniform and species-typical, would be a short one.

But the contrast between Huxley’s and Darwin’s/Romanes’s lineages is not quite straightforward. Certainly, Darwin and Romanes are associated with more generous and expansive ways of thinking about animals than the comparative psychologists, behaviourists and, indeed, classical ethologists who were to follow. Rollins argues, for example, that ‘Darwinian science gave new vitality to ordinary commonsense notions that attributed mental states to animals’ (Rollin 1998: 33), while Crist proposes that Darwin’s rich and detailed anecdotes were a reflection of ‘his perception of subjectivity in the animal world; his premise … that living is experientially meaningful for animals and that their actions are authored’ (Crist 2000: 12). But Romanes, for one, never intended his anecdotal first volume to be ‘read without reference to its ultimate object of supplying facts for the subsequent deduction of principles’ (Romanes 2012 [1884]: 20). Which is to say that, while these anecdotes were in one respect about individual animals, in another respect they were the ‘stuff’ out of which a more generalisable knowledge about the animal was to be hewn. Even though Darwin and Romanes took it upon themselves to serve ‘as hubs for sprawling networks of observers, collating and clarifying the contributions of hundreds of casual animal watchers’ (Arnet 2019: 443), they did not accept these contributions uncritically. When these stories and anecdotes appeared ‘suspect, surprising, or open to scepticism’, Romanes in particular designed and performed experiments to test their veracity (Rollin 1998: 48). ‘Virtually every [anecdotal] instance [Romanes] cites’, writes Rollin, ‘is subject to experimental replication and verification’ (Rollin 1998: 48).

With regard to Darwin, it is worth noting that anecdotes offered two kinds of support for his theory of evolution: they offered scientific support, because anecdotes attest to a wide range of individual variability, with all the evolutionary implications that follow;¹² and, as I discussed in the previous chapter, they offered political support, softening the otherwise ‘terrifying’ (Van Grouw 2018: 56) proposal that humans and animals are ‘a community of descent’ (Darwin 1981: 32). In both capacities, Darwin used anecdotes to illustrate not so much the ‘singularly intelligent and affectionate’ (as Huxley puts it) individual dog, but rather the individual dog as offering scientific insight into speciation and the species as a whole (as well as its evolutionary relations to other species). Darwin did not dismiss ‘extraordinary stories about animals’ – anecdotes, in other words – out of hand because, Crist writes, he appreciated ‘without reservation’ that ‘the dog’s show of sympathy for the cat … [was] sound evidence of the capacity of sympathy in dogs’ (Crist 2000: 43). In short, although anecdotes were unquestionably of genuine interest, in and of themselves, to both Darwin and Romanes, this does not mean that they were not also a means to more generalisable species-making ends.

It was not until the mid-1970s, the ethologist Gordon Burghardt argues, that science would once again consider legitimate ‘the study of animals’ subjective states’ (Burghardt 1985: 909). One of the contributing factors to this sea change, says Burghardt, was the publication in 1976 of Donald Griffin’s (1994b) book The Question of Animal Awareness: Evolutionary Continuity of Mental Experience (Burghardt 1985: 905). Animal Awareness may have come as something of a surprise at the time, given that Griffin, who had been best known for his discovery, with Robert Galambos, of echolocation in bats, was renowned for his empiricism and scepticism. A student anecdote, for instance, tells that when Griffin and a companion were travelling in a car and passed a flock of sheep, Griffin’s reply to his companion’s observation that two of the sheep were black was: ‘black on the side facing us anyway’ (Griffin in Gross 2005: 200). One of the distinctive features of Animal Awareness, in which Griffin ‘advocated a new field called “cognitive ethology”’ (Burghardt 1985: 905), was the foregrounding of the methodological value of paying scientific attention to the ‘rare event’ (Bekoff 2003: 83). The reason the rare event is important – as Griffin laid out explicitly in a later book, Animal Minds (1994a) (published in 1992) – is that it demonstrates animal versatility. Specifically, it demonstrates what an animal is capable of, when ‘[n]‌either evolutionary selection [n]or learning from previous experience could provide a specific prescription for what the animal should do’ (Griffin 1994a: 233). Once the limits of evolutionary or learned scripts are reached, there will lie evidence of conscious thinking and decision-making (Griffin 1994a: 233–234).¹³

Given Griffin’s focus on animals’ ‘ability to handle unpredictable, or barely predictable, situations’ (Griffin 1994a: 233), their creative problem-solving in the face of ‘novel … challenges’ (Griffin 1994a: 27), it was perhaps always a foregone conclusion that he would be charged with anecdotalism – indeed he is often accused of revivifying ‘anecdotal cognitivism’, as Darwin’s and Romanes’s work is sometimes described (Allen and Bekoff 1999: Chapter 2). This is because the narration of ‘unpredictable, or barely predictable, situations’ is more than likely to sound (and perhaps very often is) anecdotal. Despite demonstrating the scientific necessity of investigating the rare event, Griffin’s work – and especially his book Animal Minds (1994a) – is described by even the most sympathetic of writers as full of ‘[g]‌ee-whiz stories’ (Allen and Bekoff 1999: 34).

When anecdotes describe versatile, unexpected and novel behaviours in an individual animal or small group of individual animals, they raise a number of questions, including the question as to whether the behaviour of that particular animal is generalisable. Generalisable, here, is often code for ‘species-typical’ or ‘species-representative’. Are all animals in this species category similarly capable? One of the ways that a long tradition of animal investigators, from Romanes to contemporary ethologists, have sought to answer this question, and in the process to recoup anecdotes and to recoup species, is by using, as the primatologist Richard Byrne puts it, ‘the descriptive record of one of these unanticipated events … to inspire ideas to test with systematic, controlled observations or experiments’ (Byrne 1997: 134; Bekoff 2007: 121). What I understand Byrne to be saying, here, is that an anecdote can become a kind of animal ‘pedagogy’, which encourages a scientist to ask the right questions (Despret 2016) of an (individual) animal, and then ask them again in the laboratory to check that they gave the right (species-specific) answers. It is in the laboratory, Vinciane Despret argues, that the ‘miraculous transformation of anecdotes into scientific facts’ occurs (Despret 2016: 106). And in the laboratory, too, that individual animals are miraculously transformed back into species ambassadors.

Conclusion

My efforts in this chapter have been to illustrate how the connections between species and behaviours are established through scientific method, and some of the consequences that follow for individual animals. Conceptually, theoretically and methodologically, they vanish (almost). This exercise has also demonstrated, I hope, how species thinking informs and shapes scientific understandings of animals, even when the subject of species is ostensibly not being directly addressed. Too often, classical ethology and comparative psychology are pitted against each other, both politically (one is rigid, the other flexible; one authoritarian, the other individualistic), and methodologically (one is of the field, the other of the laboratory; one is naturalistic, the other experimental). One of the limitations of this focus on their differences, however, is that it disguises what they have in common, which is their thinking behaviour through species. For as long as species remains a key mode of scientific generalisation, anecdotes that recount the non-typical behaviours of an individual animal will continue to trouble science, despite the many ways that exist of dismissing their significance.

One way of dismissing ‘controversial’ animal behaviour is by the practice of pooling data to achieve statistically analysable results. As the philosopher Colin Allen and ethologist Marc Bekoff explain:

Allen and Bekoff’s point here – which is certainly indebted to Griffin – is that a ‘stimulus-free’ response also has scientific value. Two things are unspoken here: first, that anecdotes, in a context such as this, are themselves scientific; and indeed, elsewhere, Bekoff, quoting the political scientist Raymond Wolfinger, argues that ‘the plural of anecdote is data’ (Bekoff 2007: 121). Second, those individuals who ‘fail’ to respond as predicted challenge the idea of typicality, and perhaps even the value of typicality as an explanatory tool.

Rather than understand Allen and Bekoff’s argument as ‘merely’ a contribution to debates about what scientists should be paying attention to (the statistical average or the noisy exception), one might see it as an argument in favour of a shift in attention away from behaviours that are identified as species-typical on the basis of their statistical ‘significance’, to behaviours that, in their departure from that particular kind of significance, point to something else of importance. To the importance, say, of what Dominique Lestel calls ‘the singular animal’. Lestel writes: ‘[t]‌he singular animal embodies an extra plasticity which allows innovation within the species. Indeed such singular animals destabilize our conceptions of the term species, while inverting the relation between the species and individuality which we spontaneously establish in ethology’ (Lestel 2011: 93).

I introduce Lestel’s work here because it brings clarity to Allen and Bekoff’s: it says what they could have/should have said, but didn’t. But I also introduce it because Lestel and Allen and Bekoff are problematic in similar ways. Like Allen and Bekoff, Lestel gives a different, potentially ‘destabilising’, weighting to the relation between individual variation and species-typical behaviour. Nevertheless, his definition of the singular animal – like Allen and Bekoff’s 10 per cent of animals who fail to respond as predicted to a stimulus – arguably ties that animal back into the concept of species, insofar as both singularity and failure acquire their intelligibility only by way of their departure from species-typicality. In view of this, it is not surprising that Lestel’s singular animal should also be a special animal, as he explains:

Although I welcome this emphasis on the singular individual (see Chapters 5 and 7 especially), the risk is that, in this context, the descriptor, singular, outshines a somewhat more prosaic individuality, an individuality that matters regardless of species, and not on account of any extraordinary relationship to the category species. In Chapter 5, I return to this rather more banal individual, to illustrate just one example of how they can come to matter in science – can come to matter, even, dramatically. Before doing so, in Chapter 4, I will demonstrate again how the theoretical and methodological entanglement of species and animal behaviours strengthen and reinforce each other – not in the abstract, but in practice; not in relation to any and all animals, but in relation to dogs. As will be evident, the organising principles and preoccupations of late-nineteenth- and early-twentieth-century animal science, which I have been addressing here, persist as obdurately as ever.