Beyond Description: Anthropologies of Explanation

Introduction

ETHNOGRAPHIES OF EXPLANATION AND THE EXPLANATION OF ETHNOGRAPHY

Matei Candea and Paolo Heywood

What is an explanation? What does it add? What makes it authoritative, clarifying, or misleading? Whom does it serve, and by what means is it produced? These questions lie at the heart of public crises of confidence in expertise and political representation; they echo also within the knowledge practices of disciplines such as anthropology. In a world in which one global political, economic, or indeed epidemiological earthquake after another defies expert predictions of its impossibility, and post hoc accounts can often feel more like rationalizations or special pleading than explanations, competing voices vie for public presence and seek to silence one another in accounting for radical change. At stake in such political, religious, or economic contestations is the particular nature of explanatory speech and its epistemological underpinnings: What visions of truth, if any, underlie such accounts? Who is authorized to provide them, and through which media and technologies? What are the aims, purposes, and ends of explanation and the giving of accounts? Anthropology and the social sciences face such questions too, making contemporary explanatory practice both an empirical and a reflexive challenge.

This book brings together anthropologists, philosophers, and historians of social science to take a double look at the problem of explanation. The book combines ethnographic studies of practices of explanation in a range of contemporary political, medical, artistic, religious, and bureaucratic settings with examinations of changing norms and forms of explanation within anthropology itself—one of the social scientific disciplines in which explanation has been most pointedly and enduringly in crisis. Alongside chapters detailing the explanatory practices required of asylum seekers at the borders of Fortress Europe (Green), those of advocates seeking state funding for mindfulness meditation therapy (Cook), the multiple explanations an artist gives of his own distorted paintings (Rapport), those of self-defined nonpolitical readers trying to make sense of their favorite author’s sympathies with fascism (Reed) or of alt-right bloggers sussing out the minds and argumentative techniques of their progressive opponents (Mair), this book also reflects on anthropological attempts to explain specific classes of phenomena such as miracles (Bialecki) and artwork (Rapport), on anthropology’s deployment of and challenge to economic models of behavior (Staley, Salmon), on its attitude to problems (Heywood) and findings (Luhrmann), and on the tension between the implicit and the explicit in anthropological description, comparison, and explanation (Candea and Yarrow).

The placement of anthropological explanation in the frame in this way is intended as a provocation of sorts. For while, as these chapters show, anthropologists have much to say about expertise, authoritative knowledge, and the mechanics, politics, and ethics of explanation as a thing other people do, the discipline has for some time been rather wary of invoking explanation as a description of its own practice. Anthropology is not alone in this—an anti-explanatory mood has been sweeping a number of social scientific and humanities disciplines. However, anthropology is one of the disciplines in which this mood is perhaps most advanced and all-encompassing. One of its more extreme forms, which we explore in more detail later, is what we will call ethnographic foundationalism—the deferral of all epistemological questions to the ethnography (Candea 2018; Heywood 2018). Ethnographic foundationalism is not merely the (falsely naïve) claim that anthropologists should suspend explanation and just describe; it is the almost mystical belief in the power of ethnographic description to reach back and resolve anthropology’s own epistemological dilemmas. But ethnographic foundationalism is only an extreme symptom of a more diffuse anti-explanatory mood we are diagnosing. There seems to be in contemporary anthropology a pervasive sense that there is a thing called explanation out there and that it is problematic for anthropologists to try to do it.

On closer examination, however, both parts of that statement are obviously incorrect: there isn’t a single thing called explanation out there, and anthropologists do do it all the time. As for the former, as we outline later, even a cursory look at the literature on explanation generated by philosophers of science shows that there are a number of competing theories and no consensus on what it might mean to explain something, let alone what the proper way is to do it. As for the latter, on almost any definition of explanation, if you look closely enough you will find micro- or meso-explanatory moves woven into the texture of most anthropological texts, even those that purport to be purely descriptive or to reject explanation altogether (see Candea and Yarrow, this volume). We thus want to ask about the forms of explanation present in and possible for anthropology, and what their limits and problems actually are. Even though there may be a case for reclaiming explanation, there may still be compelling reasons to reject it in favor of something else. But if so, why? Can we account for what is wrong with explanation, in some or all of its forms?

In sum, this book establishes an inside-out relationship between ethnographies of explanation and the problem of how ethnography is to be explained. From one angle, it proposes a comparative account of forms of explanation in the world, in which anthropology and its crisis of explanatory confidence feature as just one case study among others (albeit one that takes a central place in this book and is examined from multiple perspectives). From another angle, this is a book posing reflexive epistemological questions to anthropology, questions that we feel are best asked alongside and on a par with ethnographic accounts of explanation beyond anthropology. This is not to say that we expect the ethnographic account of others’ explanations to resolve the epistemic conundrums relating to anthropology’s own explanatory moves. Rather, the book seeks to explore communications and productive tensions between the reflexive problematic of anthropological explanation and the comparative exploration of other explanations in the world. The final section of this introduction, which discusses the chapters in more detail, draws out some of these contrasts and analogies. In the next two sections, however, we will, first, diagnose the anti-explanatory mood that has swept anthropology and cognate disciplines and, second, take a broader look at the notion of explanation and its internal multiplicities in order to reboot our theoretical and ethnographic sense of what explanation might be.

An Anti-explanatory Mood

Our sense of an anti-explanatory mood is partly grounded in the experience of our own training as anthropologists, around a decade apart, in the early years of the twenty-first century. While we each remember being taught about ethnography, description, and critique at various points, we find it hard to recall anyone ever teaching us about explanation, except in one key sense—through a set of worries and warnings about improper attempts to explain. The history of anthropology is often taught as a graveyard of broken explanations and explanatory devices: evolutionism and progress, structuralism and the human mind, Marxism and the laws of history, transactionalism and the maximizing individual, and so on. We remember learning only one thing about explanation as an epistemological problem—namely, that it is a rather dubious and probably irrelevant practice for anthropologists.

On a closer investigation, there were two broad sources for this general anti-explanatory mood, two explicit and articulated challenges to explanation, which, though historically and epistemologically very different, combined to drive home the sense that explanation was a problematic thing to want to do. The first challenge is the one that was recurrently raised against explanation at various points in the twentieth century by proponents of interpretation. The contrast has a deep nineteenth-century philosophical and sociological pedigree. Social scientists often hark back to Max Weber’s critique of narrow historical materialism and crudely functionalist sociology and his claim that the specific task of sociological analysis … is the interpretation of action in terms of its subjective meaning (Weber 1978, 8).¹ In anthropology a contrast between explanation and interpretation has tended to be rediscovered at regular intervals. In 1950, E. E. Evans-Pritchard savaged the functionalist paradigm, to which he himself had previously subscribed, arguing that anthropology ought to be a historical interpretive endeavor and not seek to provide explanations of society analogous to those of biology. A similar challenge was famously mounted again a couple of decades later by Clifford Geertz, with his claim that anthropology’s central object, culture, is not a power, something to which social events, behaviors, institutions, or processes can be causally attributed; it is a context, something within which they can be intelligibly—that is thickly—described (1973, 14). This seemingly unavoidable recurrence of the contrast between explanation and interpretation reinforced the sense that anthropology had always been and perhaps would always be riven between two grand epistemological traditions (Handler 2009, 628; see also Holy 1987): on the one side lay the explanatory ambitions of positivism, with its cortege of scientism, reductionism, and quantification; on the other, the understanding offered by interpretivism, grounded in humanism, hermeneutics, and qualitative thick description. Andrew Abbot (2001; see also Candea 2018) has perceptively analyzed the way in which these paired contrasts operate cyclically in the lives of disciplines as core organizing polarities. For social anthropologists, however, the explanatory side of the contrasts seemed always to be in the past. With a few exceptions (e.g., Bloch 2005), the most recent explicit proponent of positivist explanation who was still recognized as part of the disciplinary canon as we were taught it was Alfred Radcliffe-Brown (1951), whose pitch for anthropology as a nomothetic search for social laws came to stand as the classic exemplar of misplaced scientistic hubris. While this grand struggle between positivism and interpretivism was already rather passé by the time our training began, it had left behind a tendency to associate explanation with what we will argue is only one, very narrow vision of what contemporary epistemologists might mean by this term.

This provided fertile ground for a far more drastic challenge to explanation, and one that at the time of our training still felt excitingly timely and fresh. This was the radical rejection of any kind of explanation over and above description itself. One of the most forceful proponents of this line of argument was Bruno Latour, whose actor-network theory was fundamentally structured by a profound antipathy for the explanatory ambitions of classic social theories (e.g., Latour 2005). Actor-network theorists were enjoined to just describe—to craft forceful accounts that stayed close to the messy contingency of particular assemblages of humans and nonhumans. They were instructed to resist the temptation to reach for the explanatory abstractions that might foreclose the account. This position was informed by Latour’s (1988) critique of explanation as either a possibility or a worthwhile aim for the social sciences. Latour defines explanation, in fact, as exactly a measure of the distance between the context of the object and the context of the account. Powerful explanations are empire-building and reductionist, imagining that we can hold multiple elements of our object of concern in a handy little receptacle like capitalism or neoliberalism. Even the most basic of explanations, that of cause and effect, is framed as a politics of accusation, an attribution of blame and responsibility, and an error. Latour’s ideal explanation is a throw away one, a one-off explanation, applicable only to a particular arrangement of elements. An explanation, in other words, that is just a description.² As Latour unrepentantly puts it, I’d say that if your description needs an explanation, it’s not a good description, that’s all (2005, 147).

One of the more radical forms that the anti-explanatory mood has taken in anthropology is that of deferring all epistemological questions to the ethnography. Consider one of the few modern anthropological collections devoted specifically to epistemology in the discipline—a theoretically wide-ranging book by Christina Toren and João de Pina-Cabral. Its contributors are presented in the introduction as being in broad agreement about two things: one is antifoundationalism (Toren and Pina-Cabral 2011, 16), and the other is the fact that ethnography is the primary condition for anthropological knowledge (15). At the intersection of those two broadly shared anthropological claims lies the position attributed by the editors to contributor Peter Gow: Anthropology has no need of any epistemology other than ethnography (6). The thought is, in effect, that epistemological questions separate from ethnography are quite simply inappropriate for anthropology (Holbraad 2009, 81). This is what we are calling ethnographic foundationalism (cf. Candea 2018; Heywood 2018).

Ethnographic foundationalism is more than a simple injunction to forgo explanation for description, à la Latour. More ambitiously, it seeks to find in ethnographic descriptions the solution to anthropology’s own epistemological problems. Consider the following questions: How should anthropology understand translation (Viveiros de Castro 2004)? How should anthropologists use examples (Krøijer 2015)? How should they generate politico-economic concepts (Corsín Jiménez and Willerslev 2007)? And how can they reinvigorate their notion of truth (Holbraad 2012)? That is not a list of potential problems for an anthropological epistemology to confront. It is, in fact, a list of just a few of the epistemological problems to which various anthropologists have already proposed solutions within the last fifteen years. What these solutions all have in common is that each claims to be derived recursively from the particular ethnographic case the anthropologist happened to be studying. In each case, the solution is for anthropology to adopt some version of what the authors’ informants happened to be doing or thinking. So, for example, and in one of the most elegant examples of this maneuver at work, Martin Holbraad finds in the inventive definitions of Cuban oracular divination a conceptual apparatus with which to make sense not only of how truth might continue to play a part in anthropological thinking but also of the inventive definitions of Cuban oracular divination. Inventive definitions—which is to say, roughly, successful performative speech acts—are both what Cuban diviners do and how to understand it, as the notion of inventive definition is, itself, argued to be an inventive definition. Whether or not one sees such circularity as a virtue or a vice, it closes off the ethnographic from anything extraneous like theory or explanation: the object explains itself.

This offloading of epistemological questions onto ethnography also chimes with a politics of engagement that sees any division between theory and practice as an academic retreat to an ivory tower that, in the words of an editorial in Anthropology Matters on the subject, should be made transparent (Kyriakides, Clarke, and Zhou 2017). Citing David Graeber as an exemplar, Theodoros Kyriakides and the convenors of the Royal Anthropological Institute postgraduate conference on anthropology’s politics of engagement declare that there is no dichotomy between theory and engagement but rather connections, relations, and multiplicities in the making (Kyriakides, Clarke, and Zhou 2017). In not dissimilar language but with perhaps more pernicious effect, the British government’s Higher Education Funding Council for England demands that our work, in order to have value, have an effect on … the economy, society, culture … beyond academia (Research England 2020). Theory, and in particular that kind of metatheoretical exercise that is epistemology, emerges from these perspectives as suspiciously detached and not impactful. An antifoundationalist consensus to defer such questions to the world itself seems much more palatable.

Of course, we are not here arguing against our shared reference to ethnography as a discipline, which as method and material is surely one of the things that makes us anthropologists, beyond specific sets of research programs. But sharing a reference to ethnography is not the same as finding in it the answers not only to some but to all of the questions we may pose, including questions of research practice and disciplinary philosophy. And it is certainly not the same as thinking that ethnography and description exhaust the proper tasks of anthropology, and that questions such as How do we explain? may be safely set aside or ignored.

For to do so is to proscribe (de jure if not always de facto) the sorts of debates and discussions our anthropological forbears had over, for example, the relative worth of deductive or explanatory versus hermeneutic or interpretive models of knowledge and understanding for the discipline. It proscribes them not for the particular answers they might provide but for the very ambition of seeking an answer from beyond the confines of empirical material. It renders the ambition of a book such as this one—to investigate an ethnographic and an anthropological practice without assuming they amount to the same thing—impossible to pursue. It valorizes description and an erasure, as far as is possible, of any distinction or difference between an anthropological account and its object.

More broadly, not only is it the case that we may wish to dispute the specific meanings of foundational concepts, but we may also have different ideas as to their proper relationship. It need hardly be pointed out that empiricism does not suit everybody’s politics, and that sometimes the choice between engagement and conceptual invention may be a mutually exclusive one. Neither across anthropology as a discipline nor across ethnographically foundationalist versions of it is any one motivation for such implicit foundationalism dominant. Discussion as to the relative merits of different motivations, however, or indeed as to alternatives to them, and to their relationships, is precluded by their common insistence that discussion of a purely epistemological kind is a waste of our time.

In other words, while many anthropologists seem to agree on the foundational nature of ethnography in our discipline, the origins of such agreement, its purposes and goals, as well as its consequences and effects, are obscured by that very consensus of method.

Our claim is certainly not, then, that it is a problem to believe ethnography as a method unites anthropology as a discipline, nor even is our claim that there is necessarily any problem with any one point of view on what it is that anthropology should be doing. It is that we will be better served in the project of assessing the purposes and underlying metaphysics, the correlations and disjunctures, and the consequences and effects of such justifications by having that discussion openly and explicitly, and without anticipating the answer in our ethnographic findings.

So we have chosen to focus this book on a classic yet long-neglected problem in the epistemology of anthropology, one that also has very clear real-world implications, in its anthropological and its ethnographic varieties. We fully expect that explanation as it is imagined, valued, practiced, or rejected in specific ethnographic circumstances can teach us something about what an anthropological explanation might look like. But we hope also that this book is an opportunity to consider the nature of anthropological explanation as a problem in its own right.

The Multiplicity of Explanation

In seeking to reboot the problematic of explanation, both ethnographically and theoretically, we would be well served by taking a sidelong glance at debates outside the social sciences. Philosophers of science and epistemologists have had profound and long-standing disagreements over what precisely it is to explain, and these debates have generated a number of competing theories and definitions. This section delves into some of these philosophical arguments, definitions, and contrasts, to enrich the often rather one-dimensional discussions of explanation current in anthropology.

In so doing, however, this section is emphatically not reaching out to philosophy to define authoritatively what explanation really is, or to set the ground rules for this book’s subsequent discussion. The role of this initial engagement with the philosophy of science is in fact precisely the opposite: not to police the boundaries of what can be called explanation but to expand them. For the core aim of this section is to highlight the multiplicity of ways in which explanation can be invoked beyond the sometimes rather limited implicit understandings current in social scientific discussions, thus challenging the tendency to assume that explanation is a unitary, singular, and clearly defined activity.

This kind of opening-up is a preliminary to the ethnographic explorations in the chapters that follow. In fact, this section might be thought of as a first ethnographic foray into explanation as it is imagined by one particular subset of contemporary Euro-Americans—namely, philosophers of science. This is not an entirely self-contained discourse, of course. Insofar as these technical definitions of explanation are often self-consciously drawing on and formalizing commonsense intuitions and understandings, these various philosophical accounts already give us a glimpse of the variety of ways in which explanation is conceived of in the world beyond philosophy.

Overviews of philosophical theories of explanation tend to start with positivists’ attempts to map out a deductive-nomothetic vision of explanation in the early twentieth century. We will come to this later but would like, for reasons that will become clear as we proceed, to begin in a slightly more unusual place: philosophical discussions of abduction.

Abduction—also known as inference to the best explanation (Douven 2017; Lipton 2004)—is a term originally introduced by C. S. Peirce (1934). It describes a form of inference that is distinct from both deduction and induction. Deduction moves inexorably from known premises to logical conclusions. By contrast, induction and abduction extrapolate likely conclusions from partial knowledge. Induction is usually characterized as a kind of direct statistical extrapolation from the known to the unknown. The paradigmatic case is the induction—famously criticized by David Hume—that the sun will rise tomorrow because it has risen every day in my life so far.³ Abduction, by contrast, is characterized as a form of inference in which a conclusion is reached because it is identified as the best explanation of a state of affairs. An example (Schurz 2008, 207–208) might be inferring the recent passage of an individual on an isolated beach based on the observation of a line of footsteps on said beach. This explanation of the phenomenon (someone has walked across this beach) is only one among many—perhaps infinitely many—possible explanations. For instance, that these footstep-like shapes might have been formed by some coincidental natural process, or by the rolling of a ball with foot-shaped appendages, or, less baroquely, by a large group of people carefully stepping in one another’s footsteps. Among these infinitely many possible explanations for a phenomenon, abduction plumps for what seems the best explanation. Another, more commonplace example of abduction might be the thought that someone can read Latin based on the observation of a number of books in Latin on their bookshelf. What is the best explanation for those books being there? The fact that the owner of the bookshelf owns them and might read them. Of course, the books might have been inherited and the current owner might be incapable of reading them, or the owner might have bought them precisely in order to give the false impression of their competency in Latin. In sum, the notion of abduction points to the fact that, in inferring the unknown from the known, we do not always simply extrapolate, following an inductive rule such as more of the same. Rather, in many cases, such inference involves some kind of more complex explanatory consideration.

For our purposes in this book, philosophical discussions of abduction are interesting for two reasons. The first is that they make a rather convincing case for the ubiquity of explanation in everyday life. By focusing on the structure of micro-judgments and observations such as the ones just discussed, they show that explanations of various kinds are ineradicably woven into our everyday experience, in a way that undercuts arguments against explanation in anthropology or elsewhere. The second reason is that starting from this observation about the ubiquity of explanation, philosophers seeking to spell out the structure of abduction—inference to the best explanation—are invariably brought face to face with a key problem: contemporary epistemologists have no settled account of what an explanation (let alone a good or best explanation) is. This means that works on abduction (e.g., Lipton 2004) are a great place to look for overviews of the variety of current understandings of explanation in the philosophy of science. It also means that one comes away from them with a refreshing sense that there might indeed be a whole range of different ways of explaining. We argued earlier that anthropologists have tended to act as if there were just one thing called explanation and it was best avoided. The take-home point of philosophical discussions of abduction is precisely the reverse: explanation is ubiquitous and it takes a huge variety of forms.

The first thing to go, from this perspective, is the engrained binary of explanation versus interpretation that has animated so much social scientific methodological discussion. Philosophers of science frequently use the terms explanation and understanding interchangeably. As Peter Lipton puts it, The question about explanation can be put this way: what has to be added to knowledge to yield understanding? (2004, 21). The fundamental contrast to which philosophers of science tend to draw attention is broader than the familiar explanation-interpretation opposition—it is the contrast between describing a phenomenon and adding something further to this description. This extra something is an understanding of some sort, and that understanding is what an explanation provides. The contrast between description and explanatory understanding—which is central also to the Latourian injunctions to just describe—is itself not unproblematic. However, as is often the case, a shift in dualisms has productive effects. Whereas Latourian critiques envision explanation as taking something away from description, curtailing or maiming it in some way, Lipton and others portray explanation as an addition, a something more. Collapsing the dualism between explanation and understanding is the preliminary to envisioning a wide diversity of forms of explanatory understanding—what are sometimes rather charmingly described as forms of explanatory goodness (Godfrey-Smith 2003, 199).

On one canonical and now much criticized view, the goodness of explanation lies in relating phenomena to laws of nature. This deductive-nomothetic (D-N) theory of explanation, elaborated in the mid-twentieth century by logical positivists (e.g., Hempel 1965), is the kind of explanation that is usually implicitly or explicitly contrasted to interpretation in the anthropological literature. The D-N model claimed that a phenomenon has been satisfactorily explained when it can be deduced from a set of premises that include a law of nature. For instance, the initially puzzling fact that an oar immersed in water appears to be bent is explained when it has been deduced logically from the law of refraction of light and some contingent facts about the refractive index of water and the position of the oar on that particular occasion. In other words, explanation in this view has exactly the same structure as prediction—it is, as it were, a prediction of things that have already happened. This D-N model of explanation was roughly the one espoused by Radcliffe-Brown in the aforementioned 1951 paper.

A key difficulty with the D-N model of explanation is that philosophers have no settled account of what a law of nature actually is, beyond saying that it is a regularity with no known exceptions.⁴ And if laws are no more than generalizations of that kind, the D-N explanation collapses into saying that some particular thing happens because that sort of thing generally happens. This is rather poignantly illustrated by the meager results of Radcliffe-Brown’s own nomothetic aspirations for anthropology.⁵ Another key difficulty is illustrated in the famous flagpole example (Godfrey-Smith 2003, 193–194): according to the D-N theory of explanation, the length of a flagpole’s shadow can be explained by deducing it from laws concerning light and a set of contingent facts including the length of a flagpole. This seems broadly unproblematic. However, the D-N theory of explanation also entails that the length of the flagpole can be explained—in exactly the same way—by reference to the length of its shadow. Here, critics of the D-N theory claim that the analogy between explaining a state of affairs and predicting it breaks down. There is something intuitively wrong about the thought that the length of the shadow explains the length of the flagpole.

This something has to do with causality—a sense that while the length of a flagpole causes the length of the shadow, the reverse is not true. This brings us (back) to the broadest and oldest view of what constitutes explanatory goodness, articulated and debated in various ways since at least Aristotle. This is the thought that explanation consists in giving a causal history—identifying the relevant antecedent causes of phenomena, events, and states of affairs. In some form or other, this is the theory that most contemporary philosophers of the social sciences tend to associate with explanation (Elster 2015; Runciman 1983). There is little consensus, however, on what kind of thing a cause is. In the account of human affairs, this uncertainty about causes is severely aggravated by a host of subquestions about mental causation and the distinctive nature of intentions, reasons, and the like (see Dretske 1991 for an overview). Much of the popular social scientific contrast between explaining and interpreting, for instance, turns on a distinction between the mere mechanical causation of behavior, on the one hand, and the identification of intentions and purposes as relevant elements of meaningful action, on the other. And yet Weber himself, to whom this contrast is sometimes traced, saw the endeavors as connected, claiming that "sociology … is a science concerning itself with the interpretive understanding of social action and thereby with a causal explanation of its course and consequences" (1978, 4, emphasis added).

The criterion of relevance is equally problematic. As Lipton nicely puts it, the big bang is part of the causal history of every phenomenon we know of, but it is hardly explanatory in the majority of such phenomena. Causal histories are long and wide (Lipton 2004, 32), and the very multiplicity and richness of causal explanation in practice can end up challenging the idea of a clear distinction between explaining and describing.⁶ Defenders of causal theories of explanation, however, have sought to respond to these objections by strengthening their notions of cause in a range of ways—by developing statistical or mechanical models of causation, for instance, or by introducing a consideration of counterfactuals (Lipton 2004; Woodward 2019).⁷

However, epistemologists have imagined other versions of explanation beyond the search for laws or the identification of causes. One such contender is the family of unificationist theories of explanation (Kitcher 1989; also see Woodward 2019). On this view, a set of disparate phenomena are explained by fitting them under a single, unified account: a coherent theory, an elegant pattern, a systematic structure. One might argue that this is what Darwinism, for instance, in its original version, did. It had very little to say about actual causal mechanisms, nor did it really formulate any fundamental laws of nature. Rather, Darwin’s explanation of evolution by natural selection provided a coherent theory to fit a set of very disparate facts—the beaks of finches, the wonderful mechanism of the eye, fossils, and so on—that were suddenly all explained in relation to one another. We could say that—pace Radcliffe-Brown’s own claims—anthropological functionalism, insofar as it was explanatory, was actually a unificationist explanation of this kind, rather than a D-N one. None of the most convincing functionalist explanations anthropologists have crafted, such as Evans-Pritchard’s explanation of the interrelation of politics and kinship in The Nuer (1940) or Émile Durkheim’s explanation of the functions of religion in The Elementary Forms of the Religious Life (1915), rely seriously on any fundamental appeal to a law of nature or even a general law of society. All of them, however, provide a coherent theory that relates disparate facts to one another and thus makes them make sense. The best structuralist explanations—such as Claude Lévi-Strauss’s (1963) account of totemism in the book of the same name or Mary Douglas’s (1973) explanation of the underlying logic of the dietary prohibitions in Leviticus—are of this unificationist kind also.

Another family of philosophical accounts of explanation (sometimes characterized as pragmatic [Godfrey-Smith 2003, 199]) departs from those just discussed in attending to the audience-relative and interest-relative nature of explanations. Thus, Bas Van Fraassen’s account of explanation starts from the observation that explanations are answers to questions, and questions come in many shapes and sizes. The nature of the question, which is to say also the background knowledge and interests of the questioner, is one of the features that determine whether an answer will count as an explanation in any given case. In other words, to know whether something is an explanation, we need to consider not simply the relation between a theory and a fact (as in classic accounts of explanation) but a relation between a theory, a fact, and a context—which includes the knowledge and interests of the audience for whom one is explaining (Van Fraasen 1980, 156).

This evokes the broader idea that explanation is a matter of making the strange familiar (Lipton 2004)—by which account most of