Cultivating trust,producing knowledge: The management of archaeological labour and the making of a discipline

Introduction

The production of scientific knowledge relies on trust in many different actors and institutions. Scientists must place trust in one another, to carry out research ethically and accurately, and to report results truthfully. Principal investigators trust technicians and students to complete steps of research procedures without supervision. And the widespread acceptance of scientific findings requires the public to trust science as an institution. No individual could possibly observe or replicate every experiment or study done on a single research topic, let alone an entire discipline or science as a whole! As Bouchard (2016: 582) has stated, ‘We have to delegate much epistemic work to other knowers.’

In archaeology, the need for trust is compounded by the non-reproducibility of excavation, which entails destroying evidence as it is recovered. Physical remains are translated into documentation, as completely and efficiently as possible—though what is considered relevant for recording is shaped by researchers’ positionalities as well as social, cultural, and political conditions. Of course, some aspects of archaeological fieldwork may be rechecked, and in that sense, repeated (Edgeworth, 2003; Lucas, 2012). Still, the full range of contextual and multisensory data that composes the archaeological record is transformed and diluted through the process of earth removal. Meanwhile, excavation often proceeds across a large area at once, making it impossible to prevent mistakes from happening.

The trust required in the scientific research process differs from trust between individuals in face-to-face relationships with one another. While science certainly involves examples of interpersonal trust, scientists will never meet all others working on shared or related topics. This reality has required the development of institutional mechanisms by which honesty is enforced and trustworthiness is inculcated in the discipline, enabling scientists to trust one another and the public, meanwhile, to trust what scientists say. Some philosophers of science have questioned whether this sort of institutionally enforced trust can be characterized as ‘trust’ or as mere reliance; in either case, there is a perceived problem of inevitable dishonesty that must be solved or mitigated through structural disciplinary arrangements.

In the history of archaeology, systems for enforcing honesty formed and were institutionalized alongside the discipline’s alliance with colonial expansion. That is, the stakes of ensuring truthfulness in archaeological knowledge production were not simply about creating accurate knowledge for knowledge’s sake; early archaeological work served to generate information about territories to strategize European expansion. Therefore, inaccuracy or dishonesty in the archaeological research process represented a threat not only to correctly understanding the human past but also to the geopolitical interests of expanding European powers.

Furthermore, these early expeditions—aligned as they were with colonial projects—also entrenched racial relations through their practice and findings. Much has been written about how early archaeology served to substantiate conceptions that the Middle East was a place where ancient civilizations had flourished but ultimately degraded as the ‘torch of civilization’ passed to Europe. Eighteenth- and 19th-century travellers and excavators also documented evidence of grand, monumental construction, especially that described in the Bible, as well as their perceptions of the contemporary lives of Middle Eastern inhabitants. Their writings reveal their beliefs that the communities they encountered had somehow ‘fallen’ or progressed ‘backward’, enabling them to justify taking artifacts away from their home countries and displaying them in European museums. Archaeology was enrolled in legitimizing European claims to the region by making the case that the communities living in this area were incapable of properly taking care of the rich cultural resources embedded in the landscape.

At the same time, the researchers so invested in characterizing the living communities they encountered as barbaric and dishonest had to work with these same communities to conduct their archaeological research. They hired locals as escorts during their travels and as labourers on excavations. And so early archaeologists were faced with a double bind: how to carry on with the work of archaeology, relying necessarily on locally hired labourers to be both productive and honest, while characterizing them as neither?

The strategies that first became standard institutional practice in the 19th century for ensuring the trustworthiness of workers were shaped by this paradox. The combination of rewards, punishments, and surveillance that emerged in this period took the form they did both because of the need for honesty in the production of archaeological knowledge and because of the racial dynamics in which archaeological research was ensnared. In this article, we examine this process of institutionalizing honesty and trust in archaeology. First, we review prior work within science studies on the need for and nature of trust in science. Then, we lay out the stakes for early archaeology—the specific consequences of dishonesty in the excavation process.

Following this, we describe the strategies that emerged in early archaeological projects for making it possible for excavation leaders to trust sufficiently in locally hired labourers that the research process might proceed. We focus on foundational figures such as Flinders Petrie, Augustus Henry Lane-Fox Pitt Rivers, and Mortimer Wheeler, but also include stories from their contemporaries working in other parts of the world in order to make clear that these strategies were part of a broad and structural shift in archaeological conduct and not isolated or region-specific innovations. Finally, we assess the consequences of these strategies for archaeology long-term and place them in the context of other systems for enforcing epistemic trust in science. The specific case of archaeology is instructive because the influence of racism and transnational politics on inculcating trust and honesty in the archaeological research process is unique and important. But studying the curation of trust in archaeology is instructive for understanding the phenomenon of epistemic trust in science more broadly by making the stakes of dishonesty clear and literal and by bringing into focus how labour management is a critical component of enforcing honesty in the scientific research process.

Honesty, reliance, and epistemic trust

Philosophers, historians, and ethnographers of science—as well as scientists themselves—have done much to show that ‘trust’ in various forms is an integral component to scientific knowledge production. Even when science is at its best, with experiments ideally designed, results reported accurately, and findings readily shared, any given researcher works directly on only one small aspect of complex phenomena. One must build upon prior research and relate to the emerging work of others to develop informed explanations about multifaceted real-world processes (Bouchard, 2016; Carrier, 2010; Feinstein, 2011; Frost-Arnold, 2013; Hardwig, 1985, 1991; Hendriks, Kienhues, and Bromme, 2016; Whyte and Crease, 2010). This entails believing in the veracity of experiments one did not observe and studies that one did not conduct.

In reality, of course, dishonesty happens in science. Scientists inaccurately report their results. For scientists in academia, there is immense pressure to publish; it can be career-making to report a groundbreaking new study, while it is notoriously difficult to publish results that confirm a null hypothesis or provide further emphasis for a known phenomenon. Ahearne (2011) generously suggests that this pressure leads scientists to commit epistemic and ethical violations in reporting their results, by exaggerating or simply inventing their findings. While Hendriks, Kienhues, and Bromme (2016: 144) claim that ‘such cases are actually exceptional’, Fanelli’s (2009) meta-analysis of survey data reveals that up to a third of scientists have admitted to dropping tricky data points, changing their research design after the fact, intentionally using methodologies that skew data in a particular direction, and allowing funding bodies to dictate which results to publish. Of course, dishonesty in scientific research is probably more common than scientists are likely to self-admit. Significantly, up to 72% of scientists who have participated in such surveys have stated that they know a colleague who has committed one of the mentioned questionable research practices. And yet, as Lüscher (2013: 1018) states, ‘We must avoid an atmosphere of distrust, as trust is the essence of scientific exchange and progress.’

Another dimension of epistemic trust often discussed in science studies is the public’s belief in science, particularly when it comes to issues like climate change and vaccine safety, where the scientific evidence is profoundly one-sided and yet this expertise continues to be doubted (Grasswick, 2010, 2014; Irzik and Kurtulmus, 2019). Certainly, the prevalence of scientific dishonesty plays a role in this persistent distrust, particularly when high-profile scientific frauds are exposed. Much of the literature on public epistemic trust in science has asserted that scientists have enacted violence against other non-scientist—but still expert—communities, and thereby furthered a lack of trust (Dotson, 2011; Grasswick, 2018; Medina, 2013; Pohlhaus, 2012; Whyte and Crease, 2010). These scholars argue that science loses credibility outside of the scientific community when scientists fail to collaborate with knowledgeable lay stakeholders.

In reading the varied body of literature seeking to get a grasp on who must trust whom for scientific knowledge to advance, a few consistent and necessary questions emerge. First, to what extent is scientific trust interpersonal? Hardwig’s (1985, 1991) foundational work on epistemic trust argues that ‘moral’ trust in the character and education of fellow scientists is the most crucial explanation for why scientists believe in one another’s work sufficiently to build upon it. For Hardwig, practices like peer review and the ideal of being able to repeat experiments are too inexact to offer an explanation for why scientists actually believe each other’s reported results. The answer must lie in their interpersonal trust of one another. Scheman (2001: 31) has characterized the nature of this trust in this way:

Even if I can be confident that arithmetic is not different for you from what it is for me, I cannot be confident that you will not make a mistake, either inadvertently or in order to give some favored child more than a fair share of candies. I need to trust you, not just arithmetic-for-you.

Of course, Shapin’s (1991) historical analysis of the emergence of (belief in) experimental methods has illuminated the significance of the conflation of scholarly and gentlemanly identity, so that trust in early modern science became equally about trust in the word of a gentleman. Shapin (2009) has more recently made the case that contemporary trust in a scientist’s work continues to be at least partly about faith in their social identity and moral character (see also Carrier, 2010). According to Grasswick (2018: 76), ‘Most trust theorists agree that interpersonal trust remains the paradigmatic form of trust, with institutional trust being modeled after it.’

Despite these compelling analyses illustrating the interpersonal or social aspects of epistemic trust, there remains a large body of scholarship arguing that epistemic trust is instead (or equally) institutional. That is, others indicate that there are a number of mechanisms in scientific practice for getting around the problem of relying on others’ individual moral character, given that scientists will never meet all of those whose work they must believe to move forward in their own research. One of these is the scientific method itself, including peer review (Lacey, 1999). Another is the ethical standards set by professional organizations and scientific institutions following high-profile reports of scientific misconduct in the 20th century—which destroyed the conception that all scientists are automatically trustworthy by virtue of their training or social status (Ahearne, 2011). There are also mechanisms of detection and punishment for scientific misconduct (Frost-Arnold, 2013). Scheman (2001: 33) refers to these suites of practices as ‘certifying the trustworthiness—the competence and integrity—of those on whom we are dependent’. According to this view (which Wilholt [2013] considers to be ‘reliance’ rather than actual trust), it is not fellow scientists’ inherent goodness that scientific practitioners believe in but rather the ability of institutions and policies to ensure that fellow scientists make moral choices.

Frost-Arnold provides a complex accounting of how these views might be brought together, characterizing them both as explanations for trust but with different motivations. The first she refers to as ‘moral trust’, which is based on ‘evidence that a potential collaborator B is honest, loyal, fair, and/or cares about her peers’ (Frost-Arnold, 2013: 303). The second camp is about trust in ‘self-interest’, where ‘the reward and punishment mechanisms that make it in B’s self-interest to be trustworthy do much of the work in rationalizing trust in one’s colleagues’ (ibid.: 302). Both of these, according to Frost-Arnold, are operating simultaneously in contemporary scientific practice. Scientists rely on one of these views where belief in the other would be irrational. It is this framework, of complementary moral and mechanistic trust paradigms, that we argue shaped the early history of archaeological fieldwork and labour practices.

In the following section, we bring this philosophical and historical framing of epistemic trust to the particular case of early archaeology, outlining the particular fears of fraudulence, misconduct, and bad moral character that limited the ability of early archaeologists to trust the labourers carrying out the physical work of excavation. Our focus in this article is on analysing not why professional archaeologists trust one another’s reports but rather how they created the necessary conditions to trust the workers they hired to dig for them. In this way, our analysis has much in common with science studies scholarship on technicians and assistants in scientific disciplines, in particular Shapin’s (1989) concept of the ‘invisible technician’. Shapin notes that Robert Boyle almost never touched his own equipment, his eyesight was too poor to make his own scientific observations, and he was known to travel for extended periods, leaving the experiments entirely in the hands of a veritable army of men employed as ‘chemical assistants’. Boyle described these men as having manual skills but very little knowledgeability. Indeed, Mol and Law (1994), looking at doctors diagnosing and treating anaemia in tropical settings; Burri and Dumit (2008), studying the performance of scientific imagery; Doing (2004), conducting ethnography in a physics laboratory; and Jasanoff (1998), writing about interpreting DNA evidence for criminal trials, have all illustrated a common theme of technicians and assistants being credited with physical skills but a denial or disempowerment of their ability to contribute intellectually.

Like all of these fields, archaeology relies on teams of individuals who carry out the manual labour of excavation but supposedly lack the knowledge to contribute to analysis or interpretation. Across fields, then, there is a necessary question with regard to epistemic trust: how and why do principal investigators trust purportedly uninformed technicians, students, and assistants? Is this trust placed in subordinates’ moral character or in institutional mechanisms of detection, reward, and punishment? We address these questions for the special case of the early days of archaeological excavation.

The stakes of archaeological dishonesty

Most of the literature on epistemic trust in science has focused on honest (or dishonest) scientific reporting, or the trust that the lay public has in scientific knowledge (e.g. Grasswick, 2010, 2014; Irzik and Kurtulmus, 2019; John, 2018). Less attention, by comparison, has been paid to the inner workings of scientific practice—to the trust instilled in technicians and lower-ranked team members during the research process. Frost-Arnold, though, has offered a sufficiently open-ended model of what trust entails, particularly with regard to science, that is helpful for thinking about the trust instilled in scientific workers generally and in archaeological labourers more specifically. She defines trust as acting ‘on the assumption that someone will do something or care for some valued good’, and theorizes this relationship as follows:

When person A trusts person B to perform action ϕ (or trusts B with valued good C), A takes the proposition that B will ϕ (or that B will care for C) as a premise in her practical reasoning, i.e. A works it into her plans that B will ϕ (or that B will care for C). (Frost-Arnold, 2013: 302)

For the purposes of this paper, person A represents the leader of an excavation project (with our analysis focusing particularly on early excavations in the history of the discipline). Person B refers to a given worker, hired (usually from the local community living on/around an archaeological site) to carry out the manual labour of excavation. Action ϕ might refer to a range of different activities, including: digging carefully without damaging antiquities, directing the archaeologist to likely important sites, continuing to work while not under direct supervision, and accurately reporting findings. The idea of caring for a valued good has especial applicability to the case of archaeology, where excavators are tasked with recovering artifacts and handing them in to the project leader, with all relevant information accounted for.

In archaeology, there are a number of ways in which trust might be betrayed. Workers might simply collect their salary while working slowly or not at all. They might perform sloppy or careless work. More consequential, perhaps, is the possibility that workers could mislead archaeologists, ‘salting’ the excavation by planting false antiquities. They could also break the excavator’s trust in their care for the valued goods recovered through excavation by selling such artifacts on the antiquities market. All of these examples would constitute betrayals of trust, and all would result in the loss or distortion of the archaeological record.

Archaeologists have explicitly debated how to guard against dishonesty by archaeological workers for more than a century. In 1872, English archaeologist Augustus Henry Lane-Fox Pitt Rivers presented a paper to the Geological Society of London in which he assured his audience:

In all cases…I took particular care to test the accuracy of the statements of the workmen as to the exact positions of the implements, and I have no doubt of their correctness in each case. Shortly after I commenced by visits to Acton, some rather ingenious attempts at forgery were foisted upon me, by chipping, varnishing, and, when dry, burying the flints thus prepared in the ground. (Pitt Rivers, 1872: 458)

Writing more than 30 years later, William Matthew Flinders Petrie evidenced a similar preoccupation with taking precautions to buffer against the threat to archaeological values posed by unreliable workmen. Since ‘excavation is for the sake of archaeology’, he wrote, ‘any doubt about a man’s character is sufficient reason for not employing him.…An unostentatious weeding out of men during the fluctuations of the work is the best means of avoiding those who seem less likely to be trustworthy’ (Petrie, 1904: 39–40). In fact, in Stephen Quirke’s analysis of Petrie’s extensive corpus of publications and notebooks, Quirke (2010: 30) notes that Petrie feared developing trust in workers. Petrie cautioned against employing highly skilled foremen, saying, ‘The more indispensable they seem, the less desirable is it to have so to trust a native’ (Petrie, 1904: 24–5).

One of the most direct and concentrated considerations of worker dishonesty appeared in an entire appendix dedicated to the matter in 1915’s Archaeological Excavation, written by classical archaeologist John Percival Droop, whose own research centred on Greece. The appendix was entitled ‘On the Causes of Dishonesty Among Workmen’, and focused on the reasons why workmen in some areas of the world tend to steal more than others, then gave advice to fellow archaeologists for moderating the likelihood of theft under these differential conditions.

A disciplinary fear of trickery and theft by workers, students, and volunteers did not disappear in the 20th century. Writing in 1953, British prehistorian Richard Atkinson (1953: 81) warned archaeologists that they would be considered to be ‘fair game’ or ‘a crank’ by the ‘paid labor’ they employed. And as recently as 2003, a reprint of the 1977 volume Techniques of Archaeological Excavation by British archaeologist Philip Arthur Barker cautioned that ‘the provenance of finds handed in by workmen or unskilled volunteers must be treated with the greatest reserve unless they can be checked on the spot’ (Barker, 2003: 104). The framing changed over a hundred years—from a fear of avarice and malicious deception to a fear (at least in part) of incompetency. Still, though, the underlying belief in the unreliability of an archaeological worker remained.

This concern over workers’ trustworthiness and character is ironic given that many of the most egregious and prominently known instances of archaeological fraud were committed not by labourers motivated by greed but by archaeologists grasping for recognition and acclaim (Gardiner, 2003; Gruber, 2008; Kraft and Thomas, 1976). The Piltdown Hoax, famously, was a forgery offered as the evolutionary ‘missing link’ between apes and humans, created by cobbling together a medieval human skull, an orangutan jaw, and chimpanzee teeth, dyed to look the same age and smashed so that the mismatched pieces would be undetectable. The skull, which Charles Dawson first presented to the scientific community in 1912, shaped evolutionary science for decades before it was fully debunked in 1953, and analyses of the case highlight the clear intentionality and the motivation of advancing white supremacist conceptions of human origins (Gardiner, 2003). Other notable archaeological forgeries by project leaders—rather than workers—include the Bat Creek inscription (Mainfort and Kwas, 2004), the Michigan Relics (Kelsey, 1908), the Beardmore Relics (Carpenter, 1957), and the Grave Creek Stone (Williams, 1991). Further examples abound. And yet it is workers who receive the most intensive and consistent scrutiny in archaeological literature. It is their propensity for dishonesty that is most discussed, and for which most solutions are found, even though epistemic trust between archaeologists is equally necessary for the production of new archaeological knowledge.

One possible reason for this is that the kind of moral, interpersonal trust that operates between scientists (described above) has historically been inaccessible to locally hired workers. Without performing substantive analysis of how archaeologists speak to and about each other, it is impossible to give any nuanced insight into the epistemic trust that operates between archaeologists. In the following section, however, we explore references to archaeological workers in the writings of early archaeologists in order to establish the impossibility for these archaeologists to place relational trust in the character of the workers they hired. We place this in the context of the geopolitical and racial dynamics of the time, and assess the degree to which this breakdown of trust in the moral character of archaeological workers has held true even as the broader global conditions surrounding archaeological excavation have changed.

The impossibility of moral trust in early archaeology

Some of the earliest archaeological excavations took place in the Middle East during the 18th and 19th centuries and were twofold in their purpose: to expand knowledge about the archaeological remains in the region and to produce documents that could be used in future occupation of the area. Napoleon’s expedition to take over Egypt may be one of the most famous examples. Napoleon brought with him hundreds of scientists tasked with documenting standing monuments and identifying artifacts that could be retrieved for France (Bard, 2008; Chevalier, 2012). Taking these artifacts back to France was one component of a broader strategy of imperial takeover. In a particularly evocative expression of the link between colonialism and the production of new archaeological knowledge at this time, many of the drawings produced by the scientists in Napoleon’s campaign were drawn with melted-down bullets.

Beyond the studies led by Napoleon (ultimately published as the Description de l’Égypte), maps created by European explorers throughout the 19th century were used to inform military campaigns and political strategy for decades to come. Indeed, the Palestine Exploration Fund’s Map of Western Palestine was completed by Lieutenant C. R. Conder and H. H. Kitchener in 1878; the British military took over Jerusalem in December 1917, shortly after Kitchener’s tenure as the British Secretary of State for War. In other areas, archaeological excavation began because excavators and military leaders shared equipment, expertise, and labour. The same digging implements were shuttled back and forth between military trenches and the trenches at Ur (in contemporary Iraq) by Leonard Woolley until 1913, for example (Lloyd, 1963). Early archaeology research in India as well was closely tied to colonialism—specifically, the British administration (Chadha, 2002; Chakrabarti, 1997; Lahiri, 2000; Paddayya, 2002). Archaeology and imperialism during this time period had complementary goals and accordingly pooled resources, and this relationship laid the foundation for the methods and imperatives of archaeological excavation in the years to come.

After all, archaeology offered not only the scientific and geographical information necessary for colonial takeover but also its ideological justification. Nearly every foreign explorer and archaeologist in the 19th century commented on the degraded and revolting character of modern societies, expressing their sorrow over the devolution they perceived (Ceram, 1968; Colla, 2007; Corbett, 2015; Goode, 2007; Graham, 1989; Kuklick, 1996; Larsen, 1996; Milwright, 2010; Silberman, 1982). This trope was so common that many Europeans came to see Arab peoples and their supposed destruction of the Holy Land as divine punishment for Christian corruption and sin (Silberman, 1982). Orientalists refused to believe that native communities could have created in the first place any of the monumental ruins and elaborate cities they were finding. Nineteenth-century scholars routinely sought alternative explanations for who could have created ancient Nubia and Pharaonic Egypt, since it seemed impossible that the contemporary people there could possibly be their direct descendants (Ceram, 1968; Kendall, 1996; Weeks, 1996; Wynn, 2007). When geologists and archaeologists started finding Palaeolithic stone tools that implied that Pharaonic culture was indigenous, new and radical interpretations were proposed including the idea that ancient Egyptians had been conquered by the brutish modern Egyptians, or that Pharaonic society had travelled as a whole to North America and founded the various American Indian nations before the area was resettled by a completely new culture (Weeks, 1996). Nineteenth-century Indologist James Fergusson (1848: 11) argued that ‘familiarity with the rate of downward progress’ would allow one not only to date artifacts found during archaeological excavation, and that this rate of downward progress applied equally to artifacts and to the ‘moral and political’ qualities of the people.

Claiming that non-white cultures had regressed for thousands of years further enabled foreign archaeologists to assert that the reason why local communities could not achieve greater presence or leadership in studying the distant past was their own apathy. Because of the chasm between such glorious pasts and contemporary societies, the argument went, the people living around archaeological sites were unable to care about this heritage, which was eminently not their own (Abu El-Haj, 2001; Bernhardsson, 2005; Colla, 2007). According to this line of reasoning, indifference ran so deep in the Arab consciousness that living inhabitants were viewed as a threat to antiquities, and Europeans saw it as their moral duty to protect historical treasures from the groups of people living on top of them (Colla, 2007).

Hence: in the 1840s, when one of Austen Henry Layard’s most skilled workmen asked a question about how Layard knew where to dig, Layard ‘seized this opportunity to give him [Abd-ur-rahman] a short lecture upon the advantages of civilization and of knowledge.…All I could accomplish was, to give the Arab Sheikh an exalted idea of the wisdom and power of the Franks’ (Layard, 1849: 316). In the 1870s, when an Egyptology school opened for Egyptians, Auguste Mariette—the French director of the Antiquities Service—blocked all of its graduates from finding jobs in the Service. And in 1924, James Henry Breasted arranged for a $10 million grant from John D. Rockefeller in 1924 to create a research institute in Cairo that would be open exclusively to Western scholars in order to forestall the training of Egyptian students (Goode, 2007). For Breasted, $10 million was ‘the price one had to pay to save these people from themselves’ (ibid.: 109). Foreign researchers saw local populations as unfit caretakers for the cultural heritage of the Middle East, and acted accordingly to prevent them from having access to and power over the treatment of archaeological remains in the region.

Relations between archaeologists and locally hired workers in the Middle East and other colonial territories were accordingly characterized by racism and Orientalism. Our goal here, though, is not to catalogue instances of racism in the writings of 19th-century archaeologists; such a catalogue would be neither comprehensive nor revelatory. A discussion of trust in the scientific endeavour is unlikely to advance on the basis of illustrating that Occidental gentlemen scholars of the colonial period scorned the brown-skinned communities living on and around the archaeological remains they sought to investigate. What is essential to point out, however, is that this racism, and the conditions under which early archaeological research was conducted, excluded the possibility of the kind of interpersonal moral trust implicitly shared between credentialed professional scientists in the present day.

And yet—archaeologists even in this early period needed to rely on members of these untrustworthy communities to carry out their excavations. The scale of these operations required cheap and plentiful labour, and the bodies of local residents represented the most realistic solution. Early archaeologists’ writings express their racialized anxiety around this inescapable reality. Petrie (1904: 21), for instance, asserted that the ‘oriental man with a good beard is quite useless and lazy’. Even his positive assessments of workers were based on physical features such as ‘honest eyes’ and ‘square chins’, encouraging archaeologists to assess workers’ honesty in the same way one might evaluate the hardiness of livestock (ibid.; see also Quirke, 2010: 52). John Goodwin, a South African archaeologist working in the early part of the 20th century, wrote in 1955 about his suspicions that the Black African workmen he hired were fugitives from the law (quoted in Shepherd, 2003: 342). His contemporary Mortimer Wheeler (1954: 72) similarly complained about the talkative and distracting personalities of ‘the oriental’ and the poor work performance of Arabs and North Indian Muslims. Droop’s 1915 appendix on worker dishonesty explicitly links together race and thievery; any archaeologist worth his salt, Droop maintains, ‘knows that the deep sea of Oriental subtlety will swallow half of his legitimate spoil’ (Droop, 1915: 73). He ponders on the reasons for this and says that ‘the laxer notions of the East may be a contributory cause, as in the European’s freedom from temptation’ (ibid.: 74).

What we discern in all of these early archaeological records is a through-line linking cultural or racial difference together with dishonesty and untrustworthiness. From the earliest expeditions, which were explicitly designed to justify European conquest of foreign territory, to writings expressing white saviourist doubt that non-white peoples could properly care for archaeological remains, to 20th-century manuals explicitly warning fellow professional archaeologists that stealing is an inherent part of Arab nature, the contemporary inhabitants of the Middle East, South Asia, and Africa have been framed as dangerous to the archaeological resources of the region. The threat is always tied to their supposed essential character, which is at first framed as naturally neglectful but is later seen as immutably malicious and greedy.

This makes archaeological work in colonial settings distinct from other contexts where scientific technicians and assistants are heavily relied upon but under-acknowledged. Though of course subordinates in many scientific research contexts may occupy marginal identity categories that intersect with their status in the team hierarchy and thereby render them further marginal or disempowered, in much archaeology there has long been a clear racial distinction between who has traditionally led archaeology and who has been hired for the excavation labour. This racial distinction is interlocked with colonial histories and global power dynamics. The result is an impossibility of moral trust in the archaeological worker by the archaeologist. The brown bodies doing the digging were for two centuries consistently depicted as innately untrustworthy, avaricious, and even dangerous to the safety of the archaeological record. And so the kind of trust in social identity that Shapin and Schaffer (1985) have argued underpins normal scientific practice did not exist between archaeologist and worker. Nor did Hardwig’s (1985, 1991) version of a baseline interpersonal trust.

Still, archaeologists needed to rely upon these workers to dig cautiously and to hand over all of the artifacts uncovered—even as they read, wrote, and reflected on the workers’ inherent propensity to do precisely the opposite of this. Caught in this double bind, early archaeologists invented a set of mechanisms designed to force honesty among the workers they hired. They would reward honesty by paying bakshish, they would ‘quarantine’ dishonesty through strategic placement of workers, and they would police workers’ behaviours with strict surveillance tactics. Such practices make sense under the rubric of science that operates through trust in self-interest, as Frost-Arnold (2013) and Scheman (2001) have described. For early archaeologists, the mechanisms would create the conditions to enforce honesty, thereby institutionalizing it, since there could be no moral trust in the character of archaeological workers. And though many archaeologists today are working to decolonize the discipline by scrutinizing the influence of early racist ideologies on contemporary practice, the legacy of a reliance on trust in self-interest persists in archaeological practice today.

Bakshish and reassignment: Rewards and punishments

Since locally hired workers were deemed to be naturally dishonest, if left to themselves, early archaeologists had to devise some means of incentivizing honest behaviour. One of the key mechanisms implemented in excavations during the 19th and 20th centuries was paying bakshish. Typically translated as ‘tip’, bakshish refers to a small monetary bonus awarded to workmen who found objects and features of archaeological interest. Though the bakshish system is primarily associated with excavations in the Near East and Mesopotamia, analogues existed all over the world. For example, Pitt Rivers notes that at his excavation of the Thames Valley at Acton after ‘workmen had the appearance of flint flakes and implements explained to them’ rewards were ‘offered to induce them to preserve any similar implements they might find’ (Pitt Rivers, 1872: 453). Fellow British archaeologist Marc Aurel Stein (1904, 1912) reported paying bakshish during his expeditions through Central Asia and China. In South America, archaeologists historically paid huaqueros for both antiquities and information, the same as bakshish in principle though not referred to as such (Smith, 2005). Archaeologists’ reflections on this payment system provide ample examples of how this particular compensation structure worked to discipline the bodies and subjectivities of workmen.

Bakshish was explicitly explained as a way of counterbalancing the financial incentive for stealing artifacts presented to workmen by unscrupulous antiquities dealers. Typical are Wheeler’s (1954: 176) observation that bakshish ‘has been found or thought necessary in countries where dealers and collectors stand menacingly at the workman’s elbow, ready to tempt with inflated values’ and Bagnall’s (2009: 38) assertion that the bakshish principle in papyrology is a product of the ‘serious temptation’ faced by Egyptian workers to ‘hide the best finds and sell them to the dealers, who often came right up to the edge of excavation’. Droop (1915: 73) warns excavators who take ‘the more scientific course’ and eschews ‘backshish’ that they could expect to lose half of the artifacts they recovered.

Bakshish was a technique of control not only in that it was deployed in the name of protecting archaeological finds against the advances of the market but also in that it encouraged workers to have particular subjective and bodily relations to the work of excavating. After noting that ‘[bakshish] is by no means only as a safeguard to honesty’, Petrie (1904: 34) suggests that the bakshish system encourages workmen to adopt the appropriate mental and kinaesthetic excavation strategies since ‘the observation of things and the care required to avoid breakages, are two very necessary habits for good workmen’ and ‘many a small thing would be overlooked and lost if it were of no benefit to the finder…digging carefully so as to avoid breakages makes a great difference to the returns obtained’ (ibid.). Petrie recommends even making the reward mechanism explicit to the workers receiving the payment, saying, ‘When giving bakhshish on a broken thing, it is well to say how much more would have been given had it been perfect’, thus ensuring workers would turn over the best finds in addition to fragments—and that they would be careful in addition to honest. Writing four decades later, Woolley (1953: 40) offers a similar observation, noting that bakshish ‘is a direct encouragement to good work; for an object broken by the digger little or no reward is paid, so it is to his interest not only to keep his eyes open, but to deal carefully with what he has seen’ (see also Mallowan, 1977).

For Wheeler (1954: 177), bakshish had the potential to transform the rank, identity, and indeed, character of the worker being paid:

The archaeologist has to remember that the requirements of science involve a complete change of values on the part of the peasant, and the bonus-system, widely applied, undoubtedly helps to impress the new scale upon his mind. When the workman begins to realize that what is to him a mere fragile crock or a fragment of carved bone and nothing more may, to his supervisor, be a historical ‘document’ of high importance, deserving of a monetary bonus, he is on the way to a mental readjustment which may ultimately turn him into something more than a mere dirt-shifter.

This account describes the ability of the worker to substitute his employer’s priorities for his own, transforming the ‘mere dirt-shifter’ into a careful excavator. Bakshish represents not an intrusion of the market in archaeological practice but a bulwark against the putatively dangerous effects of the workers’ proletarian value system on the integrity of archaeological data. Wheeler’s statement that the bakshish system has the potential to help the worker transcend his station illustrates how this mechanism was meant to act on the diggers’ character—to transform them from inherently dishonest to honest actors. Wheeler was not the only one to use bakshish with this intent in mind: Petrie paid workers not only to hand over artifacts but also to ‘act police’ and ‘keep the others away’ (Quirke, 2010: 103).

Importantly, bakshish was not a practice utilized from time to time by individual excavators. It was a disciplinary system that required full participation. Indeed, Reid (2002: 101) notes that Auguste Mariette, the longtime Director of Antiquities in Egypt did not pay bakshish, which resulted in his artifacts finding their way onto the antiquities market or into the hands of other excavators who did pay. B. H. St. John O’Neil (1935: 72), writing particularly about coin-collecting, lamented that ‘a vast quantity of useful information has been lost in the past’, due to inconsistent knowledge and application of payment for numismatic finds under Treasure Trove Law. Reward money for artifacts did not simply represent a bonus or perk—it was an institutional mechanism for transforming the mindset and morality of the locally hired labourers on whom archaeology relied, and one that worked best (like any institutional system) with full participation.

Even with broad use, though, early archaeological writing recognizes that bakshish was an imperfect reward system. The 19th-century publisher Baedeker expresses concern that when bakshish is given, ‘the seeds of insatiable cupidity are thereby sown’, echoing writing by many 19th-century travellers and researchers claiming the local communities of colonial territories to be naturally avaricious and greedy (Reid, 2002: 77). Echoing this concern, Charles Warren, working in Jerusalem, avoided paying large amounts of bakshish, as he felt that ‘the demand would have rapidly increased’ until nearly his entire budget would have gone toward feeding the insatiability of the local pasha and other regional officials (Warren, 1876: 12).

Droop, meanwhile, while supporting the overall practice of bakshish, warned simultaneously that one ‘is between the devil and the deep sea; he may choose to insure that he gets what he finds by paying the workmen the full market value of the object; that is a snare of the devil, for so he runs a good risk of getting also what he has not found’ (Droop, 1915: 73). In other words, paying bakshish encourages workmen to turn over everything they find—but, because of their inherently dishonest moral character, can also lead to them ‘salting’ the excavation with fake antiquities. If they were not creating forgeries, Mallowan at Ur cautioned that workers might dig unbidden and unsafely in areas likely to yield the most valuable artifacts (Mallowan, 1977). These warnings maintained that a reward for positive moral decisions was not in itself enough to ensure adequate trust between the scientific investigator and the workers conducting the physical work of excavation. The corollary was also necessary—a punishment system to discourage those otherwise predisposed toward dishonesty, theft, or carelessness.

For Petrie and Wheeler, this punishment took the form of reassignment to isolated or less demanding tasks. Wheeler (1954: 175) writes that ‘the fool and the chatterer must be relegated to jobs where they can meditate in solitude upon their (or their director’s) folly’. Petrie (1904: 35), writing a half-century earlier, gives an example of what this punitive measure could be, saying, ‘I kept a lad sifting earth for three weeks, to find a minute head which he had lost.…Any bad carelessness or disobedience to orders is met by degrading a man to unprofitable work or dismissing him.’ That Petrie saw this as punishment and not merely thorough scientific practice is evident in his word choice; finding the head was explicitly meant to be ‘unprofitable’ and, more to the point, ‘degrading’. At the same time, workers who ‘come to the front by their ability and character’, Petrie says, may be ‘given all the more promising places, where things are likely to be found, so that they may reap much more profit than the others’ (ibid.: 26). Agatha Christie (1946: 96), describing the labour management strategies Mallowan employed at Ur, relates a similar practice of shifting workers intentionally so that some might receive increased bakshish and others might receive less.

In this way, the problem workers—the ones for whom the reward mechanism of bakshish was not enough to press them into moral scientific behaviour—would be reassigned to tasks that, at the very least, did not matter to the evidence-gathering enterprise. Under certain conditions, these tasks could even be humiliating, forcing the workers to confront the moral failings of their previous failure to hand over intact finds. Then, when workers returned to their normal tasks of excavating new soil and working together in the team, they could be entrusted to make proper decisions with archaeological deposits and artifacts, bound both by their desire to avoid punishment again and by their eagerness for the reward of bakshish, properly dispensed.

These two systemic mechanisms—the reward of bakshish and the punishment of reassignment—function together to create the ‘trust in self-interest’ to which Frost-Arnold (2013) refers, and which Scheman (2001) calls ‘certifying trustworthiness’. In a setting where neither trust in moral character nor a shared belief in the abstract value of science is possible, an institutional means of enforcing honesty was necessary in order to ensure the reliability of the archaeological assemblage. But still, these alone were not enough to produce secure archaeological data. In addition to the reward and punishment mechanisms in place, early archaeologists implemented strategies for supervision and detection of dishonesty to supplement the structural systems of incentive and discipline. These supervision mechanisms are akin to those in place in other scientific fields, which are often used to argue for the reliability of scientific findings across disciplines. In the following section, we explore their invention within the field of archaeology specifically.

Detection and policing

Science studies scholars who argue that institutional trust—rather than interpersonal trust—is more responsible for enabling the production of reliable scientific knowledge frequently refer to ‘detection’ mechanisms that serve to catch those acting dishonestly. With reward/punishment systems, the individual is still the one responsible for making the decision to act ethically and honestly, to choose to pursue the reward rather than the punishment. Under these systems, moreover, there is always the chance that dishonest actors might evade punishment if they are subtle enough. But detection mechanisms provide oversight, bringing to light dishonest actions and actors and ensuring that they are indeed punished.

Among professional scientists, detection happens through peer review, reproducing experiments, and the oversight provided by high-quality institutions (Bouchard, 2016; Carrier, 2010; Frost-Arnold, 2013; Scheman, 2001; Wilholt, 2013). Of course, these detection systems are not unproblematic. In practice, experiments are rarely replicated, and even if they are, the results are even more rarely published (Baker, 2016). Hardwig (1991) expressed his scepticism that peer review is in fact well suited to catch out fraud and scientific misconduct. And of course, using an institution’s reputation as a proxy for trustworthy conduct on the part of all of its members is deeply flawed.

Still, these detection mechanisms remain a key way in which STS scholars explain how trust operates in science, how scientific researchers are able to rely upon one another’s results and assume that others’ findings are accurate and valid. In the words of Bouchard (2016: 600; emphasis added):

In adopting a Mertonian view of scientific values but seeing them as an ideal enshrined in various scientific institutions (such as peer review), we can see why scientific experts rightly deserve our reasonable deference more than other knowers, not because the individual scientists are trustworthy but because science is generated by institutions that police scientists’ epistemic authority more rigorously (or vigorously) than other types of institutions.

Early archaeology, with its particular, race-infused concerns about the individual moral character of excavation labourers, implemented detection mechanisms to similarly police labourers’ behaviour and to ensure that the scientific findings produced through this labour could be trusted. From an early date, close supervision of workers was linked to ensuring their honest contact. The Société Préhistorique Française’s Manuel de Recherchés Préhistoriques (1906: 21) advises archaeologists that ‘the best way to ensure the honesty of one’s workers is not to leave them for a minute and to leave them no object that might excite their greed’.

Though their work spanned more than 50 years, Pitt Rivers, Petrie, and Wheeler were in agreement about the importance of thoroughly supervising archaeological labour. Pitt Rivers (1887: xvii) notes that he never allowed excavation to ‘be carried on in my absence, always visiting the excavations at least three times a day’. In explaining his need for assistants, Pitt Rivers notes that overseeing excavations was ‘more than I could undertake single-handed’ and experience had taught him that ‘no excavation ought ever to be permitted except under the immediate eye of a responsible and trustworthy superintendent’ (ibid.). The papers of Mr Gray, Pitt Rivers’ assistant, confirm that Pitt Rivers was not just paying lip service to the importance of constant supervision. Gray reports that ‘no excavation was allowed to proceed unless one at least of the assistants was present to supervise the workmen’ and that ‘the General was generally at the diggings when important discoveries were taking place, sometimes for the whole day. Occasionally he has been known to be in the field at 7 a.m. to see the workmen arrive’ (cited in Bowden, 1991: 156).

The central importance that the supervision of excavators played in Pitt Rivers’ methodology is evidenced by the way he reports on the discovery of an iron nail and an iron knife at Cranborne Chase. After summarizing the dimensions and physical properties of the nail (2.8 inches long with a square shaft), Pitt Rivers (1892: 266) reports that the nail’s ‘authenticity’ is

vouched for by my assistants, Mr. James and Mr. Gray, who were watching the workmen at the time and saw it picked up. Mr. James was standing immediately over the spot at the time; he observed the marks of rust on the surface of the chalk before the nail was picked up and saw the workman (Castle) stoop down and pick up the nail. Mr. C. Gray who was standing at the food of the sheep-shed about 12 feet distant also saw the occurrence.

Here, the nail’s provenience consists less of details about where and when it was recovered or its relation to other objects and site features as of information about which credible assistants can guarantee the object’s provenance. This account suggests that the complete and systematic recording of archaeological finds that Pitt Rivers is credited with advocating is not just a methodological practice but also one that requires the detection mechanisms that ensure epistemic trust.

Petrie’s Methods and Aims in Archaeology also suggests that constant supervision of archaeological labour, or at least the creation of a constant fear of supervision, is a necessary tool for archaeologists. Petrie (1904: 28) advises archaeologists to use telescopes and site topography to observe workmen without their knowledge:

It is impossible to be known to be away, as then no work will go on effectively. An air of vigilant surprises has to be kept up. A sunk approach to the work behind higher ground is essential; and, if possible, an access to a commanding view without being seen going to and fro. A telescope is very useful to watch if distant work is regular.…Various approaches should be arranged from different directions, and the course of work so planned that no men can give notice to others.

Petrie goes on to recount how the use of this arrangement at Tanis enabled him to learn that though his workmen appeared to be removing buckets full of dirt from the site, they were actually just handing off empty pails (ibid.).

The arrangement of lookouts, workmen, and telescopes Petrie suggests effectively transforms the field site into a veritable panopticon (cf. Foucault, 1977; see also Quirke, 2010: 29). By clandestinely observing excavation from a ‘sunk approach’ and through a telescope, Petrie makes it impossible for the workmen to know whether or not he is away. Were his presence or absence always known, Petrie would have to use the ‘force’ of actual observation to make his workmen excavate efficiently. The panoptic system of observation he describes, however, allows work to ‘go on effectively’ even in his absence because his workmen are constantly aware that they are ‘subjected to a field of visibility’. Even when Petrie is not actually observing them, every knoll and mound in the distance works as a reminder that he might be crouching just behind it watching and evaluating their performance. Petrie’s network of telescopes and ‘various approaches from different directions’ conscripts his labour force into participating in its own subjection by ensuring that the workmen self-regulate as if they were being watched even when they are not.

The division of labour Wheeler describes at his site suggests he is also concerned with having reliable parties visually inspecting all aspects of excavation. Wheeler (1954: 162) recommends that archaeologists appoint not just a foreman tasked with controlling the workmen but also site supervisors, who ‘are in charge of an area sufficiently compact to enable them to keep in immediate contact with all work done in that area throughout the day’. These supervisors should be so closely scrutinizing the work of labourers that ‘scarcely a shovelful of earth is removed save under [the site supervisor’s] eyes’ (ibid.).

In addition to constant supervision, the intentional arrangement of labourers on site served not only as a punishment in specific cases but also as a strategy for further detection of dishonesty. For Petrie, a key principle was to sprinkle the trustworthy workmen throughout the site. This would mean that ‘each well-trained man can have half-a-dozen new hands placed near him, and he can be ordered to see that they follow instructions.…Thus every man is directly under the master’ (Petrie, 1904: 26). Managing workers intentionally to increase trust was a concern as well in the New Deal-era disciplining of American archaeology; a field manual from the University of Tennessee advised, ‘Those who have a criminal record should be assigned to wheelbarrows or other work which will provide them least opportunity to steal artifacts’ (Sullivan et al., 2011: 83).

For Wheeler, proper labour arrangements can even inoculate against dishonesty. He observes that ‘almost every gang…contains at least one addle-pated gossip, who must be isolated and extinguished at once, or the good men will rapidly deteriorate’ (Wheeler, 1954: 174). The risk of allowing a troublemaker to continue at his workstation is thus not the potential loss of a single worker’s productivity but the ‘deterioration’ of all those who work around him. Wheeler explicitly suggests that identification and isolation of those who ‘fall to the back’ can help quarantine their poor performance, saying, ‘One’s workmen, then, must be carefully teamed, to prevent the bad from corrupting the good’ (ibid.: 175). The strategic arrangement of bodies and the means by which to watch them therefore serves a further institutional purpose in the interest of securing sufficient epistemic trust in locally hired labourers. In addition to enabling detection of misbehaviour, the intentional placement of labourers based upon their trustworthiness stymies the spread of laziness, destruction, and theft—which can move through an excavation team like a disease.

Intentional organization of labourers and close, potentially inescapable supervision, therefore functioned together in the early days of archaeology to produce epistemic trust in workers. As in science more broadly, where interpersonal trust may be limited or impossible, detection and policing work to strengthen reward and punishment mechanisms, enforcing honesty and good behaviour. In this way, early archaeologists created a system that allowed them to trust the results of their excavation fieldwork—even if they could never trust the people carrying out that work. Labourers who might otherwise damage or steal archaeological deposits and artifacts were thus directed, enticed, and shamed into correct scientific behaviour, creating an archaeological assemblage that could be justifiably used to reconstruct the human past. Because the trust thereby inculcated was so essential to archaeological knowledge production, these management strategies became disciplinary institutions in the field of archaeology, rather than inventions relegated to particular, isolated luminaries in the history of archaeology. In the following, concluding section, we discuss how these mechanisms and the particular forms of epistemic trust they produced structured the transformation of archaeology into a scientific discipline.

The institutionalization of epistemic trust in archaeology

Epistemic trust is a requisite feature of knowledge production in any scientific field. Though some have taken issue with the word choice of trust, suggesting instead language that implies less affect (i.e. ‘reliance’), it is difficult to imagine scientific research stripped of all belief in the work that others have done. All scientists depend on previously published or widely affirmed findings that they have not themselves observed. Between the long time-depth of scientific research, the size of research teams, and the rate of new discoveries being reported, all science necessarily entails some amount of faith in the work that others have done—and at the very least, in others’ interest in conducting good research and accurately reporting results. There is a necessary degree in trust between principal investigators and in the technicians, students, or assistants performing the labour of scientific experimentation and other research.

Early archaeology, though, faced a challenge in this regard. Archaeology, from the beginning, was intimately tied up in the colonial enterprise and the belief systems underpinning colonial expansion. This included the perception of non-European races as lesser, more primitive, somehow degraded, and ultimately untrustworthy. Much early archaeological work took place in the Middle East, and the writings of 18th- and 19th-century travellers and researchers burst with descriptions of the ‘evidence’ that contemporary Arab populations were barbaric, despotic, and devolved. And yet these same populations were essential to furthering archaeological work in the region. It was men from these communities who would help guide Europeans (and, a bit later, Americans) to archaeological sites and who would carry out the manual labour of digging, hauling earth, and recovering antiquities. How could they be trusted to tell the truth, to not destroy the artifacts, to turn over everything they found?

And so early archaeologists faced a double bind: if they distrusted these men, they could not carry out archaeological expeditions. But if they trusted them, the very moral justification for the removal of antiquities to Europe could be unraveled. After all, if local populations were in fact trustworthy and capable, perhaps they could in fact take care of the antiquities of their own countries (whereas much writing by early archaeologists frames the removal of artifacts as essential for the protection of this cultural heritage). And if Middle Eastern communities were not as devolved and primitive as initially imagined, perhaps the ancient relics could actually be seen as the material remains of their ancestors, rather than the rightful origins of contemporary European civilizations.

Instead, early archaeologists solved the problem of epistemic trust between investigator and labourers/technicians in similar ways that other sciences have when moral, interpersonal trust is challenged or impossible. Through reward and punishment mechanisms, as well as stringent policing, archaeology institutionalized epistemic trust. Management practices such as the bakshish system, hierarchical surveillance and spatial regimentation enabled archaeologists to do scientific and systematic research with a workforce they perceived, in part because of their Orientalist attitudes, as unruly and unpredictable. These systems were essential for furthering archaeology’s progress from a hobby to a scientific discipline, protecting the findings from the potential treachery of untrustworthy labour and thereby sanitizing the results sufficiently to be considered scientific.

In this article, we have focused on early archaeology, looking for the creation of epistemic trust in excavations from the 19th and 20th centuries. Our aim has been to focus on the formalization of archaeology as a discipline, and how the excavation process was shaped by early relations between archaeologists and excavation workers, where mutual trust could not be assumed. In fact, accounts by archaeologists from these time periods often convey hostility and outright distrust. As archaeology shifted from a hobby to a science, epistemic trust had to be systematized along with the methods of excavation and analysis. Today, of course, archaeologists no longer pay bakshish and instead have refocused efforts at curbing the broader antiquities market. The days of watching workforces with a telescope from a nearby hill, out of sight of the diggers, have passed for the most part. But examining these earliest instantiations of the institutions producing epistemic trust in archaeology lays bare the most direct and obvious ways in which archaeological assemblages were first deemed to offer proper scientific information about the past. Through the 20th century, these systems changed shape, but archaeology remains a scientific discipline firmly founded on institutionalized epistemic trust, as historical and ethnographic accounts of contemporary archaeological practice make clear.

Berggren and Hodder (2003) have written about the changes in the social position of fieldworkers and the attendant changes to field methods over the course of the 20th century. They argue that archaeological excavation—particularly in contract archaeology—has overall shifted to entrust diggers with less and less responsibility for interpretation. Documentation is one area especially in which one can see a change in this regard; instead of emotive, free-flowing letters and notebooks, in the mid 20th century archaeology largely transitioned to structured forms with pre-written or preprinted checkboxes and blanks to fill in (see also Mickel, 2015; Pavel, 2010). These pro forma further conveyed the idea that archaeology can be objective, endlessly quantified, and done by any distanced observer, but they also opened up more possibilities for surveillance and policing. With these preprinted context sheets, anyone could fill them out but they were also more ably checked for quality, completeness, and accuracy. Context sheets transformed much of archaeological surveillance from Petrie’s telescope to a quick rifling through of excavators’ site notebooks. Still, though, direct and panoptical surveillance persists in many archaeological settings. Zorzin (2016), recounting his experience in British commercial archaeology, has an entire section entitled ‘Surveillance and Punishment’. He describes obtrusive cameras filming the excavators’ every move for quality and safety assurance, as well as frequent drug and alcohol testing. In Zorzin’s view, ‘These surveillance procedures are designed to be stressful, and [they function], for the developer, to gain entire control of the workplace’ (ibid.: 315).

And while direct payment for archaeological finds is almost unknown today in archaeology, less obvious rewards and punishments are certainly offered to site workers still today. These rewards and punishments take different forms depending on the type of labour on which an excavation relies; in professional settings, a promotion might be a fair reward for careful, accurate, and honest work. When students are involved, they may receive grades for their participation in archaeological field schools, which can easily be used as reward/punishment pressure. With locally hired labourers, punishments might take the form of pay docked, breaks shortened, or even suspension or dismissal from the excavation. In all of these settings, reassignment to less or more desirable jobs on the excavation remains a prevalent but seldom discussed mode of reward or punishment in archaeology. Smith (2005: 155) describes a Peruvian excavation in 1998, where an especially experienced site worker was initially assigned to an undesirable trench as punishment for striking after other workers were fired. Mickel (2012) tells the story of an excavator reassigned to dig in an ash pit as a punishment for violating expected codes of behaviour over the weekend. Early methods of creating the necessary epistemic trust for archaeology to be a mode of knowledge production have transformed alongside archaeology’s own trajectory of professionalization. But the basic features of reward, punishment, and surveillance are still apparent in contemporary archaeological practice.

Ethnographers of archaeology have developed promising methods and rich data giving insight into the contemporary mechanisms for ensuring supervisors’ epistemic trust in labourers on archaeological sites. In addition to Zorzin’s discussion of surveillance and Smith and Mickel’s accounts of reassignment as a means of punishment, Leighton’s (2016) ethnographic account of archaeological labour in Bolivia also illustrates how excavation work assignments function within the negotiation between archaeologists and Aymara workers for control over the labour force. Her more recent analysis of the culture of archaeology demonstrates some ways in which moral trust does function within archaeological knowledge production; moral trust is fostered through the culture of ‘performative informality’, which values easy-going, casual, and often alcohol-fuelled encounters (Leighton, 2020). Other ethnographic accounts point out other perspectives relevant to a general discussion of trust in contemporary excavation processes, as in Gomes’ (2006) narrative of a project in Brazil, where archaeologists lost the trust of the local community and were unable to continue their research until they regained it. A number of ethnographic studies of the hierarchies involved in excavation, laboratory analysis, and heritage management have examined how diverse participants do or do not succeed in establishing themselves as expert and fit contributors to the archaeological record (Breglia, 2005; Edgeworth, 2006; Hamilakis, 2011; Hamilton, 2000). On the whole, modern institutional mechanisms for producing epistemic trust in archaeology have inherited much form and function from their earliest instantiations, despite centuries of changes in the practice of archaeology.

Unearthing the foundational mechanisms for producing epistemic trust between supervisor and labourer in archaeology, as we have done here, furthermore adds to broader discussions of epistemic trust in science studies. These institutional features often operate subtly, which contributes perhaps to the scepticism of some that such devices are really responsible for the degree of trust that is necessary between scientists for new facts to form. Peer review is a form of surveillance, but it is one that operates blindly, disconnectedly, imperfectly. The punishments that exist in science for dishonesty or bad behaviour may be strict, but they can often be short-lasting as well. Rewards, likewise, may be as ephemeral as a momentary spotlight on a particularly exciting finding. But in the early days of archaeology, rewards, punishment, and surveillance were direct, deliberate, and obvious. Workers were paid outright for their honesty and humiliated for poor performance. Early archaeologists scrutinized workers and rearranged workers’ bodies to foster panoptical surveillance of their work. And most importantly, early archaeologists wrote about the vision behind these practices, explicitly describing them as strategies for inculcating honesty. Studying these designs of early archaeology therefore offers insight not only into the development of archaeology as a science and a discipline but furthermore into why contemporary modes of institutionalizing epistemic trust work in science, related as they are to immediate and powerful systems of reward, punishment, and policing.