How to Study Animal Minds
Kristin Andrews
Reviewed by Irina Mikhalevich
How to Study Animal Minds
Kristin Andrews
Cambridge: Cambridge University Press, 2020, £15.00
ISBN 9781108727464
This volume is an embedded philosopher’s field guide to comparative psychology and it attempts to navigate the conceptual and methodological challenges facing the study of animal minds. Andrews has spent decades working alongside animal cognition scientists in both laboratory and field settings, from a dolphin cognition laboratory in Hawaii to an orangutan field site in Borneo. She is thus uniquely positioned to explore the methods, practices, habits, and patterns of inference that drive animal cognition research. The book is addressed in equal measure to comparative psychologists and philosophers of science: it critiques existing scientific strategies for managing biases in the interpretation of animal behaviour, while at the same time providing an in-depth case study that speaks to the nature and limits of objectivity in science.
The field—or, rather, the group of disciplines that comprise the multidisciplinary project of comparative psychology—has indeed been preoccupied with controlling for biases endemic to the study of other animal minds. Many of these bias-controlling strategies have been the subject of pointed philosophical critique. Weaving these critiques together, Andrews advocates abandoning some of the traditional approaches to managing bias (anti-anthropomorphism, a preference for simplicity, a caution against folk psychology, scepticism or agnosticism regarding animal consciousness), while she proposes shoring up others (anti-anthropocentrism, openness to critique across disciplinary boundaries). The way forward, she argues, is to abandon a bias-free ideal of scientific objectivity, which she takes to be aspirational at best and self-defeating at worst. On her view, rather than aiming to eliminate bias, the best we can hope for is to identify biases where they exist—something best achieved by bringing multiple cultural and disciplinary perspectives to bear on the subject at hand.
Philosophers of animal cognition will find some of the key issues familiar: Is anthropomorphism the deadly sin that it has traditionally been taken to be? Are deflationary ‘killjoy’ explanations of animal behaviour more parsimonious than ‘romantic’ explanations? Is animal consciousness scientifically tractable? And so forth. But this is the first book to synthesize these diverse topics under a single meta-methodological umbrella, while at the same time interrogating prevailing understandings of objectivity and its proper role in science. The book makes other novel and timely contributions as well, such as normatively reframing the relationship between researcher and animal subject (more on this below), and providing an in-depth analysis of ape cognition research. The discussion consists of four accessibly written chapters grounded in concrete case studies, which are critically summarized below. Andrew’s accessible prose, judicious use of examples, and seamless blend of scientific and philosophical inquiry make for an excellent addition to graduate and undergraduate courses in comparative psychology, philosophy of science, and animal studies.
The book begins in Chapter 1 (‘Methods of Comparative Psychology’) by examining three standard bias-controlling strategies in comparative psychology: the proscription against wantonly attributing human characteristics to animals (anti-anthropomorphism), the preference for cognitively simpler explanations of animal behaviour (Morgan’s canon), and the injunction against taking human beings to be the measure of all things (anti-anthropocentrism). The first two are mainstays of comparative psychology, while the last has received comparatively little attention. Yet only anti-anthropocentrism, according to Andrews, is well supported. Anti-anthropomorphism and Morgan’s canon have been roundly criticized by philosophers for, inter alia, concentrating on one type of error (wrongly attributing complex and human-like mental traits to animals who lack them), while ignoring an equally erroneous one (wrongly failing to attribute complex or human-like mental traits to animals who have them), often leading to a bias against explanations that posit continuity between human and nonhuman animals. Defenders of Morgan’s canon, meanwhile, have yet to offer a persuasive justification for the view that simplicity is truth-conducive either in comparative psychology or in science more generally. Indeed, it is becoming increasingly clear that putatively simple mechanisms, such as association, may be not only more complex than previously believed, but may themselves underwrite cognitive processes. The result, as Andrews argues, is a double standard in how evidence of cognitive traits is interpreted in the human and animal cases. Different terms are used to describe the same behaviour in humans and in other animals. For instance, ‘episodic’ memory (what-where-when first-personal memory) is used for humans while ‘episodic-like’ memory (what-where-when memory) is reserved for other animals, about whose subjective experience researchers (or journal referees) wish to remain neutral. All of this makes a truly comparative study impossible, she argues, since a direct comparison requires a minimum of conceptual and terminological consistency between the phenomena being compared.
Where practitioners do discuss anthropocentrism, it is often tied to or equated with anthropomorphism—as if assuming (a priori) that a cognitive or cultural trait is shared between humans and other animals is somehow more anthropocentric than presuming that humans alone possess it. This book serves as a corrective to this assumption, illustrating how anti-anthropocentrism can do the work that anti-anthropomorphism and Morgan’s canon have failed to do. That is, it can remind us to presuppose neither close similarity nor radical difference with the animal subjects.
Because anti-anthropomorphism and Morgan’s canon advise viewing animals in the cognitively simplest and least human-like terms, they may also prompt researchers to regard their animal subjects as insentient objects rather than experiencing subjects. This, in turn, might cause researchers to ignore crucial factors that would be evident were scientists to consider the animal’s own perspective, such as the social needs of the animal and how those needs might affect performance on experimental tasks. For instance, Andrews points out that whereas human infants are typically accompanied by caregivers in order to ensure the infant’s cooperation and to ease performance-reducing anxiety, social animals are often tested in isolation. Andrews returns to this theme later, when she argues for regarding animal subjects as active participants in the research rather than as passive, insentient objects with whom relationships of care are inimical to good scientific hygiene. As she writes, ‘What emerges from this discussion is the idea of taking seriously the fact that the research subjects in comparative psychology, as proper social partners who have their own perspectives on the world, are sentient beings’ (p. 17).
Yet animal sentience—or consciousness, since Andrews uses these terms interchangeably—is regarded with suspicion among many comparative psychologists, who consider consciousness to fall outside the bounds of materialist science. In the interest of objectivity, practitioners thus either bracket the consciousness question or deny it outright of their animal subjects. We might view this practice as another flawed bias-controlling strategy, although Andrews does not explicitly cast it in these terms.
Nevertheless, she makes the case in Chapter 2 (‘Conscious Animals in Comparative Psychology’) that failing to take animal consciousness seriously impedes rather than promotes progress in comparative psychology. This chapter begins by considering three arguments against taking animal consciousness seriously: the ‘skeptical argument’, which amounts to the other minds problem for nonhuman animals; the ‘agnostic argument’, or the notion that we can avoid all reference to consciousness and still productively conduct research, including research concerning animal welfare; and the ‘language argument’, which takes verbal report to be the only reliable indicator of conscious states. Andrews dispatches the first by noting that the other minds problem does not prevent human anthropologists from attributing minds to their subjects; every science must accept some unprovable background assumptions. And because radical scepticism is not subject to empirical analysis, it should no more occupy comparative psychologists than external world scepticism should occupy physicists. This is correct. However, I am not sure that the sceptical argument can be glossed as simply the other minds problem for animals. Often, for instance, a version of the language argument is recruited to defend the asymmetry between scepticism of human and nonhuman consciousness as follows: that humans, unlike animals, can report their feelings gives us a (defeasible) reason to believe that humans are conscious, which we lack in the case of non-verbal animals. (It is an open question whether language-trained parrots and apes may prove the exception.) Others motivate their scepticism by highlighting physiological differences between humans and other animals, in particular differences in brain morphology. Still others note that different selection pressures may have produced creatures who appear, but are not in fact, conscious. And so on.
The version of the language argument that Andrews critiques is both a stronger claim and a weaker argument than the one above. On this version, language does not merely increase our confidence about, but is the only admissible form of evidence for, consciousness. Yet, as Andrews rightly notes, we already use non-linguistic evidence, such as brain signatures for people with disorders of consciousness and behavioural evidence for pre-linguistic infants, for attributing consciousness to humans incapable of verbal report. It’s worth pausing here to consider an assumption of the language argument that has always struck me as dubious, namely, the notion that we rightly attribute conscious states (for example, pain) based primarily on linguistic reports. This assumption ignores not only the fact that people sometimes lie and exaggerate, but also that our subjective states may be more opaque to us than we realize. Even setting these concerns aside, however, it is not clear that verbal descriptions of inner states, such as pain—though useful for purposes of medical diagnosis—are necessary for attributions of consciousness. Human beings have evolved to rapidly and reliably interpret others’ feelings on the basis of facial expressions, body postures, and non-linguistic utterances such as grunts, screams, sighs, cries, and so on. In fact, these behavioural markers may be better evidence of conscious states such as pain than verbal report in humans. If so, then behavioural markers ought to be reliable indicators of pain perception in nonhuman animals as well, provided that we understand how to interpret them. Indeed, pain perception already has a cluster of relatively well-understood behavioural markers that makes studying it empirically tractable across species. Andrews is therefore right when she maintains that we can make progress on the empirical study of consciousness by targeting specific conscious states in our research (for example, pain, pleasure), rather than aiming to identify the signatures of consciousness per se.
From here, she concludes that failing to take consciousness seriously leads to an impoverished science. As she writes, ‘Unless consciousness is epiphenomenal [and thus causally inert] a reluctance to investigate an entire domain of influence on animal behavior will lead to theories that are misguided, incomplete, or downright false’ (p. 23). Instead of sidestepping or rejecting the scientific value of taking consciousness seriously, she advocates that researchers ‘premise’ consciousness—by which she means presuppose that animal subjects are conscious, which would then compel researchers to consider how the animals’ perspectives factor into their behaviour.
There are two related but distinct issues at play here, which would be helpful to separate. The first concerns the question of whether animal consciousness is a proper subject of scientific study within comparative psychology. The second is whether researchers working in other areas of animal cognition should impute consciousness to their subjects. If a science of animal consciousness is possible, then its outputs may inform the decision to ‘premise’ consciousness by identifying the behavioural, neurobiological, cognitive, and/or affective markers of consciousness (for example, pain behaviour, responsiveness to analgesics, mental time travel, mindreading, and so on). Yet, at present, there is no consensus about how to approach studying animal consciousness. Indeed, there is no agreement among philosophers or scientists about whether an overarching theory of consciousness is required for the study of animal consciousness, and, if it is required, what role it should play in such a pursuit (Birch [2020]). Nevertheless, important progress is being made and the scientific and philosophically informed study of animal consciousness is now a nascent but serious interdisciplinary pursuit with its own peer-reviewed journals, such as Animal Sentience.
Philosophers might consider when and under what conditions premising consciousness is permitted or required and how premising consciousness ought to work without an overarching theory of animal consciousness. For the comparative psychologist, at least three considerations might guide their choice about attributing consciousness: (1) whether there are theoretical reasons to think their subjects are conscious; (2) how the empirical aims of their study might be impacted by failing to attribute consciousness when it is in fact present or the reverse; and (3) whether the experiment could conceivably harm a conscious animal.
Still, Andrews’s central point—that taking animal consciousness seriously is good science—withstands these minor critiques. And her argument that without taking animals to be conscious, we cannot regard them as collaborators in a knowledge-gathering project is both novel and important. If relationships with animal subjects matter not only from the standpoint of animal welfare, but also in terms of the reliability of the results, then this is a dimension that deserves significant further attention. What emerges from this chapter is a preview of how a flawed conception of objectivity—in this case, one that takes seeing animals as conscious beings to be fuzzy-headed sentimentality—can forestall progress in comparative psychology.
Chapter 3 (‘Objectivity and Bias in Comparative Psychology’) begins a more explicit discussion of objectivity and how flawed accounts thereof can introduce, rather than reduce, bias. Andrews centres this discussion on the question of whether so-called killjoy explanations of animal behaviour should be preferred over ‘romantic’ explanations. Killjoy explanations are deflationary, anti-anthropomorphic, accept Morgan’s canon, and are often framed in terms of ‘simple’ associative processes or lower-level intentionality; whereas romantic explanations presume greater evolutionarily continuity between humans and other animals and are often framed in terms of cognitive processes or higher levels of intentionality. For Andrews, this dialectic reveals a central bias in animal cognition science, namely, the idea that there is a unified science of animal minds and a single ‘objective position’ from which to adjudicate among competing explanations of animal behaviour.
Andrews’s own view of objectivity emerges from this discussion. She rejects approaches that equate objectivity with emotional detachment and that cast ‘romantic’ approaches as unscientific sentimentality. Objectivity so understood, Andrews argues, not only leads to the possible mistreatment of research subjects, but may close off valuable research avenues by (for example) failing to consider how an animal’s emotional state may affect performance on experimental tasks. Yet Andrews also rejects Mary Midgley’s view of objectivity as ‘fair, unbiased, impartial’, arguing that bias is inescapable since it ‘is inherent in how we see the world’ (p. 35). Rather than identify objectivity with psychological attitudes like impartiality or emotional detachment, Andrews treats it as a property of scientific communities (compare with Longino’s ([1990]) critical contextual empiricism). Previously hidden background assumptions and entrenched biases are, on her view, best exposed through public critique by a community of culturally and disciplinarily diverse scholars. Notably, this approach to objectivity is not committed to there being a single correct answer to a given question: killjoy explanations may suit some disciplinary aims, while romantic approaches may better serve others.
Andrews is right that comparative psychology is not a single field and that a unified approach is therefore neither possible nor, perhaps, desirable. She is also right that some romantic and killjoy explanations are only superficially incompatible: explanations across disciplines may target different dimensions of the same phenomenon and some apparently incompatible explanations within a discipline (for example, associative and cognitive explanations) may operate at different levels of description. However, this leaves open the possibility that some version of the killjoy–romantic distinction could be appropriate within a sub-discipline of comparative psychology. For instance, it is not clear whether, on Andrews’s account, we might expect there to be a single correct answer to a well-framed question—that is, one that does not elide level of description and for which disciplinary aims are clearly understood. There is also perhaps a tension between the rejection of objectivity as freedom from bias and the idea that a unique strength of multidisciplinarity is its ability to identify and correct for biases. After all, if working toward bias-reduction is an epistemic good—indeed, one that accrues toward objectivity—does this not lend support to a view of objectivity as freedom from bias? If so, one wonders why Andrews appears to dismiss it for being an ideal theory of objectivity.
The final chapter (Chapter 4, ‘Bias in Ape Cognition Research’) offers a detailed case study of how biases operate in the laboratory (where killjoy explanations are preferred), and in the field (where explanations are often more romantic). All chapters are peppered with illustrations from animal cognition research and Andrews’s own scientific experience; but this chapter in particular showcases the breadth of Andrews’s personal experience as a philosopher-participant in ape cognition research. The level of detail makes this a must-read for researchers who wish to tidy up their assumptions and practices.
In the end, the book is an extended argument for multidisciplinarity as a route to objectivity. It is also a clarion call to leave some of the traditional debates about what makes for an objective science of animal behaviour behind us once and for all. This, in my view, would constitute significant progress both in the field of animal cognition and in the philosophy of science.
Irina Mikhalevich
Rochester Institute of Technology
irina.mikhalevich@gmail.com