Anthropological realism in a scientistic age

Against the conflation of realism and objectivism

These contortions have offered the discipline a stark, dualistic choice: either you are a scientist or you have renounced any serious connection with reality. For an imagined postmodernist of the extreme sort, renunciation of reality is based on the premise that there is no reality worthy of the name, so any truth claim may be as good as any other (notably in Writing Culture [Clifford and Marcus, 1986]; for an incisive critique, see Reyna, 2010: 166). It is not clear, however, whether in reality – a term to which we will return repeatedly in this essay! – the image in question was ever more than a caricature, indeed a self-caricature, by those who have written most passionately in defense of the postmodernist position; some of the most effective exponents of postmodern anthropology – George Marcus and Paul Rabinow among them – have been careful and effective ethnographers and have contributed to the expansion of what ethnographers consider suitable topics for their coverage. ¹ Let us not forget that for literary scholars, at least, postmodernism is by its very constitution an ironic discourse, using images of past practice to reveal the absurdity of present-day pretensions (see especially Hutcheon and Valdés, 1998–2000); taking the caricature too literally would thus risk falling afoul of that very irony. That said, the caricature has done at least as much damage to the discipline as it has done to its targets, and this is a central concern that animates Reyna's earlier critique and forms an important part of the background to the argument of this essay.

The opposite extreme to that of postmodernism, positivism or scientism, is subject to similar cautions against caricaturing what is actually a fairly broad spectrum of approaches but ultimately also provokes a similar concern with its intellectually destructive potentialities. A scientistic scholar, one who adopts the rhetoric of exact science without subjecting it to critical analysis, claims that the only reality is one that can be measured, using the instruments – all of them, be it noted, human inventions – that are available to us; all else is froth, irrelevant detail that should not be allowed to interfere with the hard facts that make for scientific certainty. This is the position that Bourdieu (1977) has called ‘objectivism’. It suffers from a disregard for experienced reality – what Malighetti (n.d.) with telling irony exposes as ‘subjective neutrality’ – that strikingly resembles that of the postmodernists, if only because, in ignoring the inchoate phenomena that form a significant part of people's everyday lives, it risks distorting experience to the point of unrecognizability.

The persistence of scientistic discourse in social and cultural anthropology is all the more surprising because an impressive tradition in the ethnography of scientific practice has already demonstrated the social conditioning of the production of scientific knowledge. Work of such disparate character as Latour and Woolgar's celebrated Laboratory Life (1986), Sharon Traweek's Beamtimes and Lifetimes (1988) and Paul Rabinow's Making PCR (1997), this last example already cited above as an ethnographic study emerging from admittedly the most critical voice within the Writing Culture fold, have all, implicitly or otherwise, challenged scientistic representations of scientific practice. That tradition continues strongly in much of the anthropological work done under the STS label (e.g. Helmreich, 2009; Kelty, 2005).

Rather than demonstrating once again that even laboratory science is subject to social and cultural influences, a horse that is only dead because so many excellent scholars have already flogged it, I want to sound a dual warning about more mundane forms of anthropological practice. On the one hand, there still seems to be an expectation in some (although dwindling) quarters that the production of social theory out of ethnography should have predictive force on the model of natural science; and, on the other, the fact that this rarely if ever turns out to be the case allows anthropology's detractors to treat the discipline as unscientific and therefore as a weak source of insight for addressing problems of interest to the general public. That double failure alone warrants a critical assessment of the work that theory should perform in anthropology and an equally critical consideration of what its role might be in promoting a more active public role for the discipline beyond what is already recognized as engaged anthropology (Low and Merry, 2010) and public anthropology (see Lassiter, 2005; Vine, 2011).

In what follows, I propose to suspend argument about what constitutes truth, since such discussions can themselves never transcend the cultural limits of the languages in which they are couched – as, indeed, I will briefly argue below. I will instead focus on how we deal with a related but different concept, that of reality. Truth concerns the possibility of an absolute, extra-cultural ontology. As such, it is not so much non-existent as unknowable. Reality, by contrast, is about the experiential domain; it is what we know, our knowledge rendered real by an awareness of its own contingency. Such knowledge includes social knowledge, perhaps the most insecure form of knowing of all. I know that I am typing this on a computer; I think I know who my friends are. But all I have to go on is my experience of them thus far. Despite (or perhaps because of) its contingency, it constitutes what Reyna (2010: 166–9) has called ‘hard truths’. One learns them, sometimes, the hard way.

Realism accepts what Clifford (1983) scathingly, and in my view mistakenly, dismissed as a spurious claim to ‘ethnographic authority’ – spurious, he argued, in that that it was entirely based on the ethnographer's having ‘been there’ (that is, physically present in the field). But why would one's presence not constitute a legitimate claim to having had access to knowledge? To be sure, Clifford did not reject it out of hand; rather, he emphasized the partiality of knowledge thus obtained (1983: 7; see also Zeitlyn, 2009: 212–13). All knowledge, however, not only the anthropological kind, is necessarily provisional and partial, and all claims to realism are beset by cultural conventions. The operatic idiom known as verismo (‘truthism’) and the aesthetic paradigm called Socialist Realism dramatically illustrate the cultural contingency of such labels. But those labels do not describe untruths – a characterization as absolute and therefore as unrealistic as its opposite. We are instead addressing what Reyna (2010: 167) has called ‘confrontation between different ideas about reality with reality’. ² Or, as Fassin (2017) notes, realism is not the opposite of constructivism, and is not necessarily incompatible with it; it recognizes various constructions of reality to be constitutive of reality itself and deploys critique to assess them in relation to each other. As, in a different context, he writes about the state: ‘Beyond the idea of abstraction and neutrality that tends to be associated with it, the state is a concrete and situated reality’ (Fassin, 2015: 3–4) – one in which, as he demonstrates, the image of disinterestedness and abstraction is both deployed and belied in social practice (see also Herzfeld, 1992).

The more renditions of reality we find, therefore, the greater the depth of our knowledge, and to argue that ‘being there’ is simply an authoritative-sounding but empty claim would ultimately entail rejection of the very possibility of knowing anything at all. Realism does indeed imitate reality – but it acknowledges that it does so. Scientism, by contrast, makes claims to scientific status, but does not confess its imitative aspect. ³ Scientism, in short, is less realistic than realism (and by the same measure also less scientific), while, at least in its most extreme forms, postmodernism appears to have renounced reality altogether.

In choosing to focus on reality and realism, I am clearly skirting the edge of other current debates, notably the recent exchanges that pit ontology and epistemology against each other in a binarism that often, although not invariably, reproduces familiar, sharp oppositions between ethnography and theory. ⁴ There is simply no evidence that it is possible to construct a social-science discourse of such conceptual purity that it can stand outside culture. While some rationalities may appear to work better than others, or to be less embedded in cultural participation, none is culture-free (Tambiah, 1990), and it is a curious paradox that objectivist social scientists seem to think they can attain that conceptually abstract condition even though they can never provide any evidence for it.

Political implications

The present argument, however, also has a distinctly practical nature. As I have already indicated, it addresses the relative isolation of anthropology in public discourse today – a concern that also seems to have actuated Reyna's (2010) prescient earlier critique of the so-called crisis in anthropology. I am less concerned here with new theoretical ground-clearing than with fighting off the perceived irrelevance of ethnographic detail to public life. That cannot be done without simultaneously fighting off the perceived irrelevance of reality to anthropology. The view that anthropologists have little to do with hard reality especially infests those aggressive ideologies, ranging from neoliberal capitalism to archaeologically based nationalism (see Abu El-Haj, 2001), that have most to fear from a critical examination of the politics of ‘mereness’. This is emphatically not to argue that anthropologists must always play a role in public life or that anthropological scholarship that fails to do so is worthless; such views play into the hands of the political extremists who have been so successful in promoting racist and anti-intellectual agendas in recent years. It is, rather, to argue that the intrinsic value of anthropology lies in its attention to the tension between rules and practices and the insights that such attention can generate for those who actively wish to argue for a more humane politics than the current trends appear to be creating.

Theory-making is itself a cultural and political form of practice. ⁵ We cannot afford to ignore political attacks on its importance if we wish to preserve any degree of critical distance from the newly hegemonic forms of geopolitical common sense. ⁶ It is in this spirit that I offer the present argument in defense of realism, but against objectivism. It will be an important first step, politically as well as epistemologically, to show that anthropological attention to the messiness of social reality is more realistic than the pretensions of predictability that other, politically more prominent social science disciples have offered.

To make such an argument, we must first insist on a clear distinction between these two concepts – realism and objectivism – since they are often, and too easily, conflated with each other. Realism acknowledges the conditionality of knowledge; it could not do otherwise, if only because all human experience is rife with restless self-correction. Objectivism, by contrast, seeks to eliminate all recognition that knowledge depends on embodied human experience – and, by so doing, potentially eliminates (and logically would have to eliminate) even the knowledge that we possess solely through the activity of our brains, themselves parts of our material bodies. Objectivism is thus ultimately self-defeating.

Scientism is the language of objectivism. It is a rhetorical position involving the use of scientific-sounding terms and claims, and its effect (and presumably often its goal) is to deflect attention away from the failure of objectivist logic. It does not so much guarantee as imply that the conclusions expressed in its terms are grounded in empirical research. Even in a world where powerful political actors can gain traction by deriding such well-documented science as the evidence for climate change, scientism sometimes has enormous popular appeal. For example, it has done yeoman service in the world of advertising; ‘scientifically tested’ is a vague guarantee that appears to guarantee a positive outcome. In scholarly discourse, however, nothing is ever proved beyond the possibility of contention – even a positivist, or at least one of Popperian persuasion (see Reyna, 2010: 168), would agree with that axiom. We too easily forget that the word ‘science’ comes from a Latin term, scientia, which means knowledge; and knowledge comes to us, as Giambattista Vico (1744) observed, through our sensory organs. It is by its very nature and derivation imperfect. ⁷ Rather than entering the lists in widespread debates about what constitutes true science, I prefer here to focus on what we recognize as reality. The goals involved in investigating both terms – reality and science – may well be similar, but an emphasis on reality and realism will more generously allow an assessment of anthropology to stand on its own feet rather than to become entangled as a secondary player in debates in which the physical sciences hold center stage. ⁸

Many social scientists, including the vast majority of non-anthropologists, acknowledge the experiential basis of reality. Payne and Payne (2004: 171), for example, recognize the sensory basis of knowledge: ‘Knowledge statements about the world must be consistent with our experience of it’. Problems arise, however, when we ask whether cultural difference affects the forms of experience – whether, for example, as Fabian (1983 ) has famously argued, anthropologists should refuse the idea that the experience of temporality would be different for peoples not operating in the language and cultural milieu of a modern, English-speaking scholarly community. He argues that anthropologists, despite the sometimes enormous shifts in their theoretical orientation, remained consistently convinced that the ‘others’ they studied experienced time in radically different ways from their own. He calls this stance a ‘denial of coevalness’ – in other words, a denial of the very possibility that people of such different cultural identity could co-exist within the same time-frame as those who wrote about them. In this regard, Fabian's argument anticipates and refutes more recent attempts, collectively known as the ‘ontological turn’, to attribute radically different realities to non-Western, non-industrialized societies. ⁹

Fabian's move pulls us away from regarding people of vastly different culture as in some way irremediably separate from ourselves; it emphasizes the commonality of human experience and understanding. But can we assume that such commonality genuinely exists, and how would we find convincing evidence for it? Is such a claim not as hegemonic and totalizing a move as the colonial thinking that Fabian's critique – justly – seeks to displace? I want to argue instead that a measure of uncertainty in our pronouncements about such matters – perhaps, sometimes, in the form of a modest recognition of the necessarily partial nature of any ethnographic account (Zeitlyn, 2009) – would more effectively sustain the critical perspective that Fabian has so adroitly advanced. If it is the case that we are uncertain in our interpretations, and especially if it is also true that our informants also express a great deal of uncertainty about everyday matters, should our analyses not reflect that lack of clarity and the incompleteness of our and their perceptions? We must avoid the trap of attributing psychological inner states (see Needham, 1974; Rosen, 1995), for example in supposing that we know how others experience the passage of time. We can nevertheless report on our experiences and on their narration of their own experiences. For that matter, our readers cannot read our minds, still less those of informants reported in our writings and therefore only accessible to our readers at second hand.

Uncertainty as pollution

Thus we must ask what, in a world dominated by the desire for certainty, we can mean by ‘social science’ if social life itself usually produces forms of its opposite, of uncertainty? What if vagueness and imprecision are fundamental dimensions of social existence (see, e.g., Bourdieu, 1990: 78; Reay, 2004: 433)? Vico was quite explicit about this: in the human sciences, he noted, clarity could be a vice rather than a virtue (Giuliani, 1976: 33, n. 11) – a perspective with which we can agree at least to the point of saying that in the absence of clear and independent evidence for order, in other words of evidence that an objectivist would understand and accept, there is no reason to assume that society is as ordered as the early functionalists, for example, imagined it to be. A.R. Radcliffe-Brown (1952: 1) thought anthropology should be ‘nomothetic’ – that is, capable of generating laws of human behavior; the opposite condition he called ‘idiographic’. But we all know full well that one cannot do good description without at least implicitly involving theory, while theory is itself a cultural product and, as such, fundamentally an ethnographic object.

The same principle can be applied to concepts of truth and historicity. Even the ancient (and modern) Greek word αλήθεια (alētheia, truth), so often cited as the root of Western interpretations of truthfulness, makes it clear that truth is what we know rather than what simply is, exists. It is the opposite of λη′ θη (lēthē, oblivion), and thus is dependent on our capacity for recall. Vico's exploration of the etymology of ‘fact’ prepared the ground for modern anthropological investigations into culturally diverse idioms of truth-making. Thus, Andrew Shryock (1997) and Paul Dresch (1986) have shown for Jordanian and Yemeni society, respectively, that concepts of truth are dependent on the vagaries of segmentary politics; facts can be fissured (and, presumably, reunited) as easily as the social entities that seek to rest their legitimacy on those facts. Much the same can be said of ‘history’; its relationship to the more fictitious-sounding ‘story’, with which it shares its etymological roots, results from the circumstance that all historical knowledge is socially produced and may be deeply affected by culturally specific criteria.

All of this is familiar ground for anthropologists. It is, moreover, much easier to make this case for something like ‘exotic’ concepts in Arabic than for their equivalents in English, French or perhaps even Chinese. While Dresch and Shryock certainly never intended their analyses to be read as exercises in exoticism, they may have escaped condemnation because they were analyses of the discourse of ‘others’ – and thus unthreatening to the Western conceit that rationality is universal and, probably, Western in origin (on which see Tambiah, 1990). But that conceit is certainly a remarkable piece of self-deception, as its insertion into the context of a comparison with these Arabic terms, or others like them, would surely demonstrate to any open-minded reader.

Anthropologists, moreover, must perpetually struggle with the disconcerting problem that public discourse in the majority of countries today is far more objectivist than the discourse of anthropology. This makes it hard to claim social relevance for anthropology, which, far from passively submitting to the constraints of what Marilyn Strathern (2000) has so tellingly dubbed ‘audit culture’, is a source of some of the most severe criticisms to which that phenomenon has been subjected (see, e.g., Shore and Wright, 2015). This resistance to numerical reductionism is quite sufficient to ensure that the general public will view anthropology's deliberate refusal of closure with deep suspicion. The social scientists whose rhetoric most closely approximates what we might call ‘popular positivism’ – econometrics and survey-based scholarship in particular – have a far easier access to public trust. Indeed, if we can extend to the whole range of the social sciences Douglas Holmes’ (2013) remarkable insight into the ways in which central bankers create the self-fulfilling prophecies of global and national economies, we might say that those social sciences that accept the rhetoric of popular positivism or scientism – the rhetoric of auditable numeracy – are better equipped to command public opinion.

But this does not make them any more convincingly free of cultural motivations in the way they construct their objects of study. On the contrary, by their own standards of objectivism and especially of distance from the object of study, they are more vulnerable to the charge of circularity, and anthropology, especially in the works just cited, has begun to discharge its responsibility to act on global politics by pointing out both this absurdity and the cultural conditions that enable it. I want to suggest that the logic long ago identified by Mary Douglas (1966) offers both a measure of explanation for this paradox and a means of identifying the cultural logic that appears to support the audit's claims, making them appear as common sense and, as such, utterly undeniable

Douglas’ analysis of purity and taboo is one of the most durable theories in social anthropology. This is largely, I believe, because she sought a very simple commonality across a wide range of phenomena, but never made the mistake of calling what she had discovered a law. Her theory does not predict, for example, what kinds of taboo will be found under what circumstances, only that any violation of existing categories will: (a) produce a sense of disorder that can be treated as either holy or unclean; and (b) reproduce the basic structure of the social in important cultural spheres. Thus, the injunction that ‘thou shalt not seethe the kid in its mother's milk’ is: (a) a prohibition on mixing categories (in this case, meat and dairy, both products of animal domestication); and (b) a metaphorical play on the prohibition of incest (a mother may not have sexual relations with her child).

Scholarly discomfort with, for example, active political engagement with the problems of the communities we study is an issue of the same order. Political engagement mixes scholarship with activism and undermines objectivist claims to pure factuality. Against those claims we can pose an engaged anthropology that requires anthropological writing to reflect the uncertainties and confusions of social life and that therefore, as part of this commitment, does not shy away from ethical dilemmas that arise from activism or other forms of political involvement (Herzfeld, 2010, 2014; Low and Merry, 2010). These, indeed, are the primary components of an explanation of the suspicion with which social anthropology is often treated by practitioners of other social sciences; its political engagement is, for them, potentially ‘matter out of place’, in Douglas’ famous phrasing.

They are also uncomfortable with theories that do not produce neat answers. Economists’ models, for example, seem to offer the reassurance of prediction where anthropological commentary is more inclined to emphasize the absence of predictability. Those models or ‘laws’ are grounded in a specific perception of reality, that of large trends and top-down perspectives; it is uncommon for economists to pay much attention to mismatches between their policies and the sometimes messy and historically complex dynamics of household and local community, as anthropologists (notably Gudeman and Rivera, 1990) have noted. For the skeptics, refusal to offer big-picture predictions demonstrates a lack of taxonomic certainty in anthropological thinking, and the lack of taxonomic certainty is the heart of Douglas’ argument about symbolic pollution. Anthropology therefore often comes across as offering only chaos and disordered complexity, and this symbolic pollution of the conceptual universe is especially damaging at a time when ‘sound bites’ increasingly appear to provide the public with the comfort of certain knowledge. Yet discomfort – the perpetual dissatisfaction of which Malighetti (n.d.) writes – produces vastly more insight than smug conviction; and that in itself, in a comfort-seeking, consumerist world, is a considerable obstacle to sustaining a substantive political role for the discipline.

Objectivism appears to promise a tidy separation of the observer from the observed – a lack of direct involvement as the guarantee of objective assessment. Yet this is, in practical terms, a methodological impossibility, as becomes clear as soon as we enter any research situation. When we begin work in any community, there are always questions about what forms of reciprocity will be forced on us and what forms of reciprocity we will seek to conduct. Most anthropologists today feel deeply indebted to the people they have studied, people who have put up with these strange visitors' endless importunities. The idea of finding an uncontaminated culture is itself absurd, as anthropologists have long recognized, because the very fact of contact already initiates some forms of reciprocal influence, often materialized as hospitality and the exchange of gifts. Thus, the very idea of the objective participant observer is as absurd here as the idea of bureaucrats who can function in total disregard of the social context of their work. But that does not mean that the participant observer is wrong. To be objective is not equivalent to being right; to be subjective does not automatically mean that one has no contact with reality. The difficulty lies in the culturally and historically over-determined character of the objective-subjective distinction itself.

In this context, the language of contamination is itself highly suggestive, and not only because it reproduces the selfsame language used by 19th-century folklorists and ethnologists seeking original forms of culture on the hopelessly inappropriate model of philologists’ textual reconstructions of ancient literature. ¹⁰ Those scholars frequently dismissed vernacular texts that they had recorded in the field as ‘corrupted’ versions of written originals; they then, almost as often, gave themselves the right, assumed on the basis of their supposedly superior education, to ‘correct’ the texts they considered corrupt or contaminated. The ultimately moralistic idea of textual corruption, moreover, was transmitted from folklore to anthropology, generating a romantic genre of salvage anthropology that seeks to restore the pristine lifestyles of supposedly primitive peoples; Fabian's critique is especially relevant to such developments.

As a result, the metaphor of textual corruption has also infected public understanding of culture in general. It hints at something dirty or ugly – ‘matter out of place’ – in the very idea that a people innocent of all contact with the outside world might actually have been culturally infected by contact with the industrial countries. In a trope that was already old when Marcel Mauss invoked it in the Essai sur le don (1954), that contact is seen as the point of entry for a series of polluting ills, undermining the pure character of the noble savage as it supposedly was before the advent of money. Much salvage ethnography is of the same character, and those who see the anthropology of Europe and North America as a squandering of anthropology's precious personnel and intellectual resources while other cultures are disappearing (e.g., Maurice Bloch, cited by Judith Okely in Jakoubek and Budilová, n.d.: 215 ¹¹ ) are as guilty of perpetuating a condescending exoticism as any romantic folklorist. It is by examining the cultural dynamics of the countries that have produced anthropology itself, and especially by treating such pretensions of purity and contamination as of the same symbolic order as the food taboos of the ancient Israelites or modern African and Asian peoples, that we can push back against the comforting illusion that anthropology can or should one day reach a level of such total abstraction that its theories would be free of any specific cultural influence.

Another source of pollution that seems to irk those of scientistic bent is the intrusion of morality into the conduct of research. Invoking a strong version of the principle of cultural relativism – which is pragmatically about as convincing as the strong version of the Sapir-Whorf hypothesis ¹² – these critics reject the idea that judgment in anthropology can legitimately be based on ethical concerns (e.g. D’Andrade, 1995). If these views were right, we would not have to contend with the ethical inspection committees known in the United States as Institutional Review Boards – a development that historically seems to have been necessitated by what today is widely construed as (to put the matter charitably) the ethically questionable actions of certain self-styled scientists in fields such as, notably, clinical psychology (e.g. Milgram, 1963). But in fact it is the scientific community that has called for and embraced the IRB procedures, many of which make little sense in the context of anthropological research and require constant revision and adaptation to our methodologies, while their reduction to the rhetoric of audit culture – an objectivist exercise par excellence – undermines the credibility of the relevant procedures and serves the interests of administrators rather than of the ‘human subjects’ – a telling phrase! – whom their invention was intended to serve. It is the actions of conscientious committee members and researchers that in some institutions have turned the IRB system into a useful tool for expressing appropriate fiduciary caution rather than for imposing factual conclusions. One can read more literal-minded actuations of the process, however, as establishing a symbolism of purity, reinforced by the semblance of rigid fact that the rhetoric of numbers encourages.

The reality of laws

Laws are themselves far from predictive. Their form is absolute and appears to make a claim on eternal verity, but in practice they are subject to constant reinterpretation. This is true of both legal practice and everyday social life. We may therefore be justified in applying the Italian proverb about law in the more ordinary sense of official regulation – fatta la legge, si trova l’inganno, ‘once the law has been made, one finds a way around it’ ¹³ – to the naturalistic laws imagined by some social scientists (in contrast to those established by natural scientists). This move allows us to see that formal structures not only permit violations of various kinds but actually require them in order to perdure. The Italian proverb might strike some as unduly cynical, but it is an accurate representation of what both citizens and bureaucrats understand as social necessity, albeit in a perhaps somewhat exaggerated formula in the particular case of Italy.

Institutions cannot exist in conditions of mechanistic obedience (see Herzfeld, 2016a). Laws may be written to satisfy some ideal of objectivity, but their use is shaped by circumstances that include long-term social dynamics; no attempt to write fixed meanings into them or to connect them to ‘authentic’ local traditions for the same end can succeed (Eckert, 2016: 242). Bureaucratic rules often create difficulties that would be unbearable were the institutions in question to insist on strict observance at all times and under all circumstances, as Scott (1998: 310–11) has noted with respect to the British work-to-rule idiom of industrial action. ¹⁴ Glaeser (2004) has even suggested that the massive secret police application of laws against petty infractions of the social order contributed significantly to the downfall of the German Democratic Republic, whose citizens found themselves with no arenas left in which they could enjoy the small possibilities for naughtiness that keep populations acquiescent in larger structures of dictatorial oppression.

What goes for legal codes also applies to attempts at the codification of social process. The ethnographic literature has in fact long demonstrated this propensity for flexible interpretation in even some of the most fundamental institutional structures. Kinship provides a particularly good illustration. Even when the rules appear unambiguous, as in the principles of agnatic descent and segmentation, both cross-cutting ties and fictive extensions of the literal meanings of kinship terms allow for an enormous amount of manipulation (e.g. Evans-Pritchard, 1940); despite the charge that such variation renders the structural rules liable to highly selective and unpredictable application and that those rules can therefore be viewed as mere folk models (see the early critique by Peters, 1967), the same variability is to be found in state systems and the other formal institutions associated with bureaucratic modernity.

This is the stuff of ‘what everyone knows’ – the shared open secrets, or ‘cultural intimacy’ (Herzfeld, 2016a), that are necessary to sustain life within institutions in a form that is both practically manageable and socially bearable. The appearance of rigid structural principles is only sustainable if, in practice, a good deal of pragmatic adjustment takes place, as every lawyer and every bureaucrat knows (see Herzfeld, 1992). By the same token, attempting to extract predictive laws from such persistent unpredictability is a wild goose chase, and is precisely what I mean by scientism. A realistic perspective, by contrast, would attend to the uses to which these rules are put by all actors, including those who, charged with applying the rules, find creative ways of bending them to specific exigencies.

The point is not to say either that there is no reality (the postmodern conceit) or that official self-representations by institutions are untrue (a common conceit of positivists). It is, rather, to argue that just as official descriptions of formal structures are a form of reality, so, too, perhaps less surprisingly, are those activities that are recognized as real life in the spaces of cultural intimacy. But those activities, too, are construed in ways that are always necessarily incomplete and evanescent because their fugitive character reflects varying degrees of official disapproval; this is an important aspect of what de Certeau (1984) meant by distinguishing the tactics of the disenfranchised (and, we may add, of those who work within the system of power) from the strategies designed by political leaders. No institution can survive by insisting on the equally rigid observance of all its rules, especially when those rules logically and pragmatically conflict with each other. And no resistance to the rules can be effective if its tactics are always and entirely visible to those who exercise power (see also Scott, 1985).

Paradoxically, then, the pragmatics of cultural intimacy are fundamental to institutional survival. Some of the world's most important religions have long recognized this and provide an explanatory framework for commonplace deviations; this is what we understand by the term ‘theodicy’. Theodicy is the basis on which we explain why, in a divinely ordained world, human imperfection continues to prevent us from satisfying our own ideals; Christianity does this with the doctrine of Original Sin, Buddhism by appealing to the concept of karma (Thai kam). There is even a ‘secular theodicy’, whereby we explain why even the most efficiently organized state invokes human flaws to explain why its institutions do not always deliver what they have promised (Herzfeld, 1987: 35; 1992).

This imperfection – or what some theorists recognize as indeterminacy (see Karp, 1986) – makes nonsense of attempts to reduce the science of anthropology to a nomothetic machine as Radcliffe-Brown imagined it. But acknowledging uncertainty does not make anthropology any the less scientific. Even in the natural sciences, as its practitioners at least from the time of Heisenberg have recognized, perfect predictability assumes either a very limited or a hermetic and ultimately not very interesting field of investigation (see Malighetti, n.d.). But the uncertainty of social science and that of the physicist or the chemist are also of very different orders. In particular, scientific experiments are conducted on a very different observational scale than ethnographic fieldwork (see Ardener, 1971: liv, lxxxiii). For these reasons, the application of hard-science jargon to anthropological analysis is deeply misleading.

Models in anthropology are either taxonomic or heuristic. Taxonomic models, especially monothetic ones, are reductive; they constitute what Edmund Leach (1962) once contemptuously – and, in most contemporary anthropologists’ view, rightly – castigated as ‘butterfly-collecting’, and generate little more than accretions of examples of supposedly recognizable phenomena. At their worst, they reify such concepts as society and culture as collections of ‘societies’ and ‘cultures’, a device that reproduces official discourse rather than furthering critique, setting artificial boundaries and reducing to inert entities what in reality are creative, pulsating processes. Moreover, they are precisely the conceptual structures that generate the dismissal of interesting phenomena as ‘matter out of place’ solely because they disrupt the purity of the classification itself.

A concrete illustration of the difference between taxonomic and heuristic models would serve us well here. I developed the concept of crypto-colonialism to address the conditions under which certain nations have been allowed by the Western colonial powers to preserve a modicum of national independence on condition of their accepting sometimes humiliating treaty terms and of their enforcing a newly reified sense of national territorial boundaries and of the cultural ‘stuff’ (Barth, 1969) contained within these boundaries (Herzfeld, 2002). This model often occasions surprise because it originated with a comparison of the respective political dynamics of Greece and Thailand, in contrast to the relative absence of such parallels between Greece and Italy (I have worked in all three countries).

One case in question concerns China, which has a long and proud history of national independence but has certainly made significant concessions to foreign powers at several points in its history, notably in the second half of the 19th century. Should China thereby be considered a crypto-colony? A taxonomic approach would seek a simple ‘yes’ or ‘no’ in response to this question; indeed, as methodological nationalism does with the nation-state (Wimmer and Glick Schiller, 2002), it presupposes the key term as an unproblematic organizing concept, rather than exploring and embracing its distortions and complications. A heuristic approach refuses such taxonomic and lexicographical rigidities and instead opens up questions concerning the variable (and often unpredictable) nature of claims to sovereignty. Instead of dismissing such claims because they are necessarily partial and therefore do not help in building models, this position assumes that most such claims are likely to be partial and may therefore also lead us to insights into the geopolitical dynamics of colonialism that the existing, largely dualistic models (see the critique by Dube, 2002: 741–743) tend to obscure. Shanghai is not Beijing; the Boxer Rebellion is not identical to the present Chinese government's antipathy to imported ideas that seem to threaten its ideology and practices. Difference, in short, is more interesting than conflation. ¹⁵ But it is also more difficult to present to a public increasingly satisfied with ‘quick and dirty’ fixes (a term that betrays a certain guilty awareness of the polluting methodology within!). That is the real challenge for anthropology, both politically and epistemologically.

Consequences of uncertainty

Classification, as Vico recognized, all too easily morphs into the very opposite of knowledge. Emerging as a way of organizing knowledge – as, so to speak, an instrument of conceptual bureaucracy – it constrains thinking beyond that organizational frame. To be sure, we require that constraint as a crutch enabling us to get to grips with the pragmatics of applying scholarly ideas to particular cases. Like laws in the hands of an unhelpful or lazy bureaucrat, however, it quickly assumes the mantle of irrefutable factuality in its own right. It creates zones of exclusion that are characterized as ‘wrong’, ‘dirty’ or ‘polluted’ – in other words, to use Douglas’ terminology once again, as areas of taboo.

Politicians sometimes seem to have an instinctive grasp of the power of classification. When a governor of Bangkok compared to stray dogs those homeless people who had dared to camp out in areas designated as official monuments to the religious aura of the reigning dynasty, he was appealing to a deeply rooted symbolic universe in which animals rank well below humans in the karmic schema (Tambiah, 1969), but relating this exclusion to a program of urban beautification, or aesthetic purification, that ultimately sought to remove all the embarrassingly poor from the highly visible symbolic center of the old city. He thereby sought to translate a long-standing Thai symbolism into the language of Western-derived and partially colonial planning modernism (on which see Herzfeld, 2016b: 227–8 n42; cf. Holston, 1989; Rabinow, 1989). The fact that notions of kam (karma) belonged to a much older stratum of religious doctrine did not preclude such translations into the terms of a self-designated modernity. On the contrary, we also find accusations leveled against certain NGOs that appear to seek karmic redemption – the appeal to theodicy again! – rather than simply dealing with an immediate social need (Aulino, 2014). There is nothing incompatible about symbolism and modernity – but modernism, which often has recourse to an anti-realist objectivism, seeks to deny this association, arguing that symbolism is either wrong (for example, as superstition) or simply separate from concrete reality (see also Ladwig, 2013). That move denies the materiality of symbolic action, including language, flying in the face of the materialist perspective so eloquently expressed in the title of J.L. Austin's influential How to Do Things with Words (1975).

The unfortunate fact of modernity that bedevils anthropological research and any attempt to demonstrate the discipline's relevance to current affairs is that official discourses are mostly grounded in an objectivist understanding of the world. For bureaucrats, perhaps understandably, knowledge is something to be bureaucratically controlled; in the most obviously abusive cases, press freedom is restricted as soon as it is handed over to bureaucratic management. ¹⁶ Chinese bureaucrats dismiss religious activity as ‘superstition’ or ‘cults’, and use these labels to prohibit or inhibit various forms of ritual practice. Yet in practice even these bureaucratic discourses encourage a degree of complicity in enabling formally forbidden actions to occur (see, e.g., Steinmüller, 2013) because their operators know that a degree of complaisance is necessary for the effective operation of the official bureaucracy in more general terms. Human reactions, as any experienced bureaucrat must be aware, are largely unpredictable and therefore require improvisatory adaptation. That adaptation may mean actively participating in something that the law textually precludes.

The political ends of realism

The challenge of dealing with the complexities of a globalizing modernity has shown up the weakness of disciplines claiming to have strongly predictive theoretical capacities. I do not wish to challenge the predictive value of some of those disciplines, at least in some of their instantiations. But I do wish to suggest, again following Holmes' rich suggestion, that the reasons for which their predictions appear to be validated by experience merit close ethnographic attention to their conditions of production, a move that would perhaps help to restore to social and cultural anthropology its freedom from the hobgoblin of objectivism and would at the same time enhance its role as a source of deep political critique.

Objectivism is the real antithesis of a critical anthropology because it precludes any explicit recognition of what Malighetti (n.d.) incisively calls ‘[the] theoretical and technical procedures that allow the visibility and knowability of the world’. In a discipline that insists on the importance of context, it thus obliterates its own context of knowledge. It also presupposes an undemonstrated and undemonstrable dualism between reality and symbolism predicated on the Cartesian tradition of mind-body dualism. That dualism was extremely useful, as Rabinow (1989) argued, for colonial town-planning: ‘they’ live in the ‘disorder’ of a system we do not understand, while ‘we’ live in ordered grid patterns that make mathematical sense (but perhaps not much sense in terms of human social life). Anthropology can learn from the impact and implications of that planning paradigm.

As the claims of objectivism palpably crumble, and as the logic of modernist planning hurtles against the experiential realities of social life (see especially Holston, 1989; Johnson, 2014), anthropology offers a less constricted vision of social experience. The difficulty has been one of presentation: how can we present that vision to a public that has either accepted the logic of neoliberalism or given up in disgust over the role of academic experts? ¹⁷ Recent studies of social revolt and reaction in places as far apart as Thailand (Sopranzetti, 2017) and Greece (Vradis and Dalakoglou, 2011; Rakopoulos, 2015) – conjoined here in a frame that again hints at their shared crypto-colonial past – show that those ‘ordinary people’ with whom anthropologists do their work are able to manage very well without ‘expert’ guidance that is not grounded in direct knowledge of the conditions of their everyday lives.

But it is not enough simply to record disaffection, rebellion or alternative ways of being in the world. Anthropological descriptions are certainly a necessary and useful first step. The next step, however, is to assume some responsibility for promoting the synergy of theory and action that is already emergent in much anthropological writing; indeed, the writing itself, through the provocations it offers and through teaching in particular, is capable of forming part of that action. Moreover, anthropology is well prepared for such a role by its moral commitment to deny any pretensions of cultural superiority and to refuse all forms of cultural and social prejudice. In its practical incarnation as ethnography, anthropology is a form of morally informed social and cultural critique. ¹⁸ A morally inflected anthropology is not the enemy, or the antithesis, of empirical knowledge – that is, of knowledge grounded in experience and acknowledged as a product of human intellectual process. It lies at the very core of knowledge thus construed and offers one very practical path toward ensuring its continuing production. In the world we inhabit today, a world in which consumerism and its discontents appear to be eating away at the urge to achieve social justice, an anthropology supportive of its practitioners' ethical entailments and prepared to recognize their concomitant and forthright commitment to experiential realism can provide an important counterweight to cynicism and indifference.

I am not arguing that anthropologists must always engage in political action beyond the writing and research that they do. That is a matter for a different, albeit related, discussion, and in my view should always be a matter for personal decision. But I do insist that recognizing the political implications of what we write is a necessary component of the realist position I have adopted here. This has two major consequences. The first is that we thereby come to recognize that no anthropological observation ever stands free of the political and ideological perspectives either of the anthropologist or of the local social actors. That is a trenchantly realistic position. It does not mean that all anthropologists must consciously become political activists; it does mean that their work has political significance.

The second consequence is that the work we do offers a critical perspective on an increasingly dominant set of prefabricated truths promulgated by national governments and other powerful entities, including, increasingly, transnational financial players. If we choose to enter public debates with that realization in mind, we may indeed be effective, if only because we will have engaged, as we do in our ethnographic research, with the real world of lived experience. Even if we do not, our theoretical positions are both the consequences and the causes of political dynamics, as indeed are those adopted by all other social scientists. When we return to the field, we inevitably bear witness to the sometimes devastating impact of dominant discourses. ¹⁹ If our own discourse now becomes a more influential part of the global dynamic, we must be prepared to take responsibility for the real consequences that our theories have for those about whom we write and for all of global humanity. Perhaps it is time to recognize intellectually, and embrace politically, the material reality of our theoretical orientations. Perhaps, indeed, we have already begun to do so.