When John Allen Chau was killed on North Sentinel Island in November 2018, the editorial verdict arrived before his body could be recovered from the beach. Within days the consensus was settled: Chau was reckless, theologically deluded, and had endangered the very people he claimed to love. All of that is probably true. But almost every piece published in those weeks treated something else as equally obvious \u2014 that India’s legal prohibition on approaching the island was legitimate, rational, and not worth examining. The law was the natural backdrop to a story about individual folly. Nobody asked about the law itself.<\/p>\n\n\n\n
What the law does is this. A state has decided, on behalf of a group of people it has never spoken to, that those people want no contact with the outside world. It enforces this through Navy patrols and criminal prosecution. The people being “protected” have expressed no view on the policy \u2014 though they have expressed, repeatedly and with arrows, a view on visitors. Whether those arrows mean “go away forever” or “go away today” or something else, nobody can say, because asking would require the kind of approach the policy is designed to prevent. The Sentinelese cannot consent to the arrangement. They cannot reject it. They exist, legally, as wards of a state whose language they do not speak, administered under a regulation drafted sixty years before the internet existed.<\/p>\n\n\n\n
That is the thing everyone took for granted. It is, on examination, a remarkable arrangement.<\/p>\n\n\n\n
The primary legal instrument is the Andaman and Nicobar Islands (Protection of Aboriginal Tribes) Regulation, 1956, which designated North Sentinel Island as a tribal reserve and created an exclusion zone of approximately five kilometres \u2014 three miles \u2014 around the island’s coastline. After Chau’s death in November 2018, the Ministry of Home Affairs reaffirmed the prohibition. Enforcement in practice has operated a substantially larger buffer \u2014 five nautical miles, roughly nine kilometres \u2014 though the statutory basis for the larger zone is not cleanly resolved in the regulation’s text.<\/p>\n\n\n\n
The regulation is Indian law, applied to a people who predate India as a state by an unknown quantity of centuries.<\/p>\n\n\n\n
Brazil’s equivalent is older and more elaborate. FUNAI \u2014 the Funda\u00e7\u00e3o Nacional dos Povos Ind\u00edgenas \u2014 has operated a formal non-contact policy since 1987. As of March 2025, FUNAI officially recognises 119 isolated indigenous groups in Brazil; Survival International’s October 2025 global report identified 124 confirmed groups in the country alone. FUNAI implements the policy through Ethno-environmental Protection Fronts \u2014 monitoring and buffer teams whose job is to observe, document, and intercept intruders. Article 231 of Brazil’s 1988 Constitution recognises indigenous peoples’ rights to their lands, customs, and social organisation. ILO Convention 169 (1989) and the UN Declaration on the Rights of Indigenous Peoples (2007, Articles 3 and 10) affirm the right to self-determination and free, prior, and informed consent \u2014 FPIC \u2014 before any decisions affecting indigenous peoples are made.<\/p>\n\n\n\n
And here the architecture collapses into its own foundation.<\/p>\n\n\n\n
FPIC was designed precisely to prevent states and corporations from making decisions on behalf of indigenous peoples without asking. But obtaining FPIC from a group that cannot be approached without triggering a disease epidemic is logically impossible. Consent requires contact. Contact kills. The legal framework that exists to protect indigenous autonomy has no mechanism for the one case where it matters most. The protection rests on an assumption \u2014 that the peoples being protected would consent if asked \u2014 and that assumption is unfalsifiable by design.<\/p>\n\n\n\n
Who are the Sentinelese?<\/strong>\n\nNorth Sentinel Island sits in the Bay of Bengal, roughly 50 kilometres west of South Andaman Island. The Indian government estimates the Sentinelese population at between 50 and 400 people; reliable figures do not exist, and aerial counting is imprecise. What is known comes almost entirely from observation at a distance. They hunt, fish, and gather. Their material culture includes metal tools \u2014 almost certainly salvaged from shipwrecks \u2014 suggesting some historical, indirect awareness of the outside world.\n\nIn the aftermath of the 2004 Indian Ocean tsunami, a coast guard helicopter flew over the island and photographed a Sentinelese man running across the beach, bow drawn, arrow levelled at the aircraft. The photograph circulated worldwide. Its meaning was unambiguous. The Sentinelese had survived the tsunami and were in no mood for visitors.\n\nIn 1991, the Indian government attempted a different approach. Triloknath Pandit, director of the Anthropological Survey of India, led a series of contact expeditions to the island, offering coconuts and other gifts. The Sentinelese accepted some of the gifts and approached the boats. The series of expeditions continued until 1994, after which India abandoned contact efforts and reverted to the exclusion policy. The brief window of tentative exchange is the closest thing to mutual communication the Sentinelese have had with the outside world in recorded history. Then it stopped. The Sentinelese did not ask for it to stop.<\/code><\/pre>\n\n\n\nWhat Chau was doing, and what it actually means<\/h3>\n\n\n\n
On November 15, 2018, Chau paid two fishermen \u20b925,000 \u2014 roughly $350 \u2014 to take him near North Sentinel Island. He approached in a kayak. A Sentinelese boy fired a metal-tipped arrow that pierced the waterproof Bible Chau was holding against his chest. He retreated. On November 17, he returned, told the fishermen to leave without him, and paddled toward shore. The fishermen later reported watching, from a distance, as Sentinelese figures dragged his body along the beach. The following day it was buried there. Chau’s diary, recovered through his friend in Port Blair, described his mission in terms that were earnest to the point of being impossible to read without discomfort.<\/p>\n\n\n\n
Chau was affiliated with All Nations, an evangelical organisation that had trained him extensively, including mock scenarios with staff playing hostile tribespeople. His theological framework was specific. The “unreached peoples” movement \u2014 prominent within evangelical circles through Ethnos360 (formerly New Tribes Mission) and the Youth With A Mission network \u2014 holds that the Great Commission in Matthew 28:19-20 requires that the gospel reach every people group on earth before Christ’s return. The Joshua Project, an evangelical research database, catalogues every ethnolinguistic group by evangelisation status. The Sentinelese appear in it. They are listed as unreached.<\/p>\n\n\n\n
The universal response was to treat this as self-evidently absurd, the theology of a man who had gone off the edge. But the framework is not fringe. Ethnos360 operates in dozens of countries with decades of institutional history and significant financial backing. What Chau did was illegal, epidemiologically dangerous, and resulted in his death. It was also the logical outcome of a coherent worldview held by millions of people. Dismissing it as individual eccentricity is easier than engaging with what it actually represents: an organised, well-funded position on exactly the same question that states and NGOs have a position on \u2014 what should happen to the people on that island?<\/p>\n\n\n\n
Chau’s position was wrong. But it wasn’t random. It was a claim \u2014 that those people have a right to hear the gospel \u2014 colliding with another claim: that those people have a right to be left alone. The collision killed him. Who gets to make that call?<\/p>\n\n\n\n
The unreached peoples movement<\/strong>\n\nThe Joshua Project is an evangelical research initiative that catalogues the world's approximately 17,000 ethnolinguistic groups by their exposure to Christianity. Groups classified as \"unreached\" \u2014 with less than 2% evangelical Christians and no indigenous church capable of independent growth \u2014 number in the thousands. The project maps them geographically and tracks missionary progress.\n\nEthnos360, formerly New Tribes Mission, was founded in the United States in 1942 with the explicit purpose of reaching the world's most isolated tribal populations. It operates across dozens of countries through a global partner network. Youth With A Mission (YWAM) is a decentralised evangelical network with presence in virtually every country on earth. All Nations, Chau's organisation, runs a training programme that includes preparation for contact with hostile communities.\n\nThe theological logic is internally consistent: if salvation is real and eternal, and if some people groups will die without ever hearing the gospel, then the calculus of contact looks different from inside the framework than outside it. That doesn't make it right. It makes it comprehensible, which is more useful.<\/code><\/pre>\n\n\n\nThe epidemiological floor<\/h3>\n\n\n\n
Contact kills indigenous people in enormous numbers. This is not disputed, and any philosophical argument about isolation must operate above it.<\/p>\n\n\n\n
A 2015 study by Robert Walker, Lisa Sattenspiel, and Kim Hill, published in Scientific Reports, analysed 117 epidemics affecting 59 indigenous societies in Greater Amazonia between 1875 and 2008, causing more than 11,000 documented deaths, 75% from measles, influenza, and malaria. The mean mortality rate across epidemics was approximately 25% per year. Rates have declined over time \u2014 contacts modelled at the study’s 2015 publication date predicted roughly 10% annual mortality \u2014 but 10% annual mortality is still catastrophic for any small population. It means contact is less lethal than it was fifty years ago. It does not mean contact is safe.<\/p>\n\n\n\n
The cases make the abstraction concrete. After FUNAI made first contact with the Matis in the late 1970s, the population collapsed: estimated at several hundred at first contact, the Matis numbered no more than 87 by 1983. Between 35 and 50 percent had died from disease. Of the Panar\u00e1 \u2014 contacted in 1973 when a military highway was built through their territory \u2014 more than 250 of approximately 350 members died within the first twelve months; their population eventually fell to 74. In 1987, New Tribes Mission secretly contacted the Zo’\u00e9 in northern Brazil; by 1991, 45 Zo’\u00e9 had died from malaria, influenza, and respiratory disease. The Nahua of Peru faced a similar trajectory in the early 1980s when oil exploration opened their territory: within a few years, around 50 to 60 percent of the population had died.<\/p>\n\n\n\n
These are not historical anomalies. They are the pattern.<\/p>\n\n\n\n
The decline in mortality rates is real, and it reflects genuine improvements in epidemiological response. But that response requires immediate, well-resourced medical deployment at the moment of contact \u2014 in remote Amazonian territories, often days or weeks from the nearest hospital. The better outcome is a model. The Panar\u00e1 and the Zo’\u00e9 are the record.<\/p>\n\n\n\n
Any argument about the rights of isolated peoples must operate above this floor. The freedom to choose contact is not meaningful when exercising it has historically killed a third to half of the community making the choice.<\/p>\n\n\n\n
The paternalism charge<\/h3>\n\n\n\n
And yet someone is making this decision. That is what the mortality data, for all its clarity, cannot dissolve.<\/p>\n\n\n\n
The paternalism critique runs as follows. States, NGOs, and international bodies have decided \u2014 without consulting the people concerned \u2014 that isolated peoples should remain isolated. They have built legal frameworks, deployed enforcement agencies, and developed institutional positions around this conclusion. None of this involved asking. It couldn’t: asking them is the thing the policy prohibits. The result is a structure that looks, from the outside, like protection, but functions as permanent administered isolation \u2014 built and maintained by people who never asked those it was built for.<\/p>\n\n\n\n
The critique has serious advocates. In a 2015 editorial in Science, Robert Walker of the University of Missouri and Kim Hill of Arizona State University argued that governments violate their responsibility to isolated peoples by adopting a permanent non-contact policy. Isolated populations are not viable in the long term, they argued \u2014 they face accidental contact, extractive encroachment, and demographic attrition \u2014 and well-organised, medically prepared contact is today both humane and achievable. To refuse contact is not neutrality; it is a choice to withhold medical care and options available to every other human population.<\/p>\n\n\n\n
A 2019 essay on the Oxford Uehiro Centre’s Practical Ethics blog applied a Samaritan rescue principle to argue that contact could be morally obligatory under conditions of extreme need \u2014 allowing people to die from treatable disease might constitute an ethical failure regardless of autonomy considerations.<\/p>\n\n\n\n
These are real objections, not rhetorical exercises. They are what liberal political philosophy looks like when it runs into a case it was never designed for \u2014 people who cannot speak for themselves, and cannot be asked. The FPIC principle was designed to prevent states from overriding indigenous agency. Applied to uncontacted groups, it produces the opposite: states make a permanent decision on behalf of people without any input from those people, and call this protection.<\/p>\n\n\n\n
The honest version of the paternalism charge: we have decided what is good for them without asking, constructed a moral framework that makes it impossible to ask, and called this respect.<\/p>\n\n\n\n
Why that argument fails here \u2014 but not how you’d think<\/h3>\n\n\n\n
The strongest response to the paternalism critique is not philosophical. It is epidemiological.<\/p>\n\n\n\n
The freedom to choose contact becomes meaningful only when exercising it doesn’t kill a third of your community. When it does, the choice isn’t between contact and isolation \u2014 it’s between dying from a decision you made and being prevented from making a lethal decision by someone who can’t ask you about it first. Neither option is clean. But one produces immediate mass death and the other does not.<\/p>\n\n\n\n
Survival International’s position \u2014 that isolated peoples “know we exist and have chosen not to talk to us” \u2014 is the most defensible practical stance available, and has anthropological support. The Sentinelese case makes this stark. They’ve demonstrated, over decades of observation and several direct contact attempts, both awareness of outsiders and consistent rejection of them. The 2004 tsunami provided the most striking evidence: in the immediate aftermath, before any survey team could reach the island, the Sentinelese had already moved to higher ground and were visible, alive, defending their territory from helicopters. They read the environment. They are not waiting to be discovered.<\/p>\n\n\n\n
But the protection-as-consent argument can’t answer the harder question: we don’t know what isolated peoples would want if they could make an informed choice \u2014 with access to information about what contact actually means, what medicine can do, what the outside world contains. The Sentinelese who fired an arrow at Chau’s Bible were not making a consent decision. They were defending their territory against an intruder. Those are not the same thing.<\/p>\n\n\n\n
Survival International says they have chosen. The Panar\u00e1 were not asked before the highway went through. The Zo’\u00e9 were not consulted before the missionaries arrived. In every historical case, contact happened because of extractive pressure, missionary ambition, or state policy \u2014 not because someone figured out how to ask the question and get a meaningful answer.<\/p>\n\n\n\n
Are \"uncontacted\" peoples really uncontacted?<\/strong>\n\nThe word is imprecise. Many groups described as \"uncontacted\" have had sporadic historical contact \u2014 with rubber tappers, missionaries, or disease vectors \u2014 and have retreated into isolation afterwards. Others have traded with intermediary groups, and have material culture, including metal tools, that demonstrates indirect contact with the outside world. The Sentinelese almost certainly salvaged metal from the 1981 wreck of the MV Primrose, which ran aground on their reef and was abandoned.\n\nFUNAI, for this reason, prefers the term \"isolated\" over \"uncontacted.\" The distinction matters. These are not pristine populations untouched by history; they are peoples who, having had some exposure to the outside world, have chosen \u2014 or been forced by disease \u2014 to maintain distance from it. That choice, or that trajectory, is the thing the policy is attempting to protect. It is not less legitimate for being imperfect.<\/code><\/pre>\n\n\n\nNo clean resolution exists. The debate continues among anthropologists, ethicists, and legal scholars \u2014 careful, principled, unresolved. Meanwhile something that does not care about the debate at all has been answering the question by other means.<\/p>\n\n\n\n
The policy is failing anyway<\/h3>\n\n\n\n
The real question, being answered daily in the Amazon, is not whether the policy is philosophically defensible. It is what happens when the institutions enforcing it run out of money, staff, and political protection.<\/p>\n\n\n\n
Brazil is where the answer is most visible. Ninety-one percent of Brazil’s artisanal gold mining \u2014 garimpo \u2014 occurs within the Amazon biome. Between 1985 and 2022, the total area of garimpo extraction expanded by approximately 1,200%, from roughly 218 square kilometres to around 2,627 square kilometres. A 2024 study in Nature Communications found that at least 77% of 2022 extraction sites showed explicit signs of illegality. Indigenous territories, where mining of any kind is prohibited, lost 13,000 hectares of forest to illegal gold mining in 2023 alone.<\/p>\n\n\n\n
The Sarar\u00e9 Indigenous Territory \u2014 home to the Nambikwara people in Mato Grosso \u2014 became the most impacted indigenous area by illegal mining in 2025. From January 2024 to August 2025, illegal gold mining destroyed more than 3,000 hectares of forest within the territory \u2014 more than 4% of its total 67,000 hectares. Government agents estimated approximately 2,000 miners operating on the land at the peak of the incursion. The miners arrive in response to gold prices set in London and Shanghai. They are not making an ethical argument.<\/p>\n\n\n\n
Missionary pressure has also operated through institutional channels. In February 2020, the Bolsonaro government appointed Ricardo Lopes Dias \u2014 a long-serving missionary with Ethnos360\/New Tribes Mission \u2014 to head FUNAI’s isolated and recently contacted tribes department, giving him access to location data for more than 100 registered isolated groups. Brazil’s Federal Prosecutor’s Office sought suspension, warning of potential “genocide” and “ethnocide”; an appellate judge ultimately overturned the appointment. But Lopes Dias was reinstated \u2014 and removed again \u2014 in a sequence that illustrated how thin the institutional protections are when the political winds shift.<\/p>\n\n\n\n
Missionaries have also worked around the institutional layer entirely. The Pulitzer Center documented the use of drones, shortwave radios, and solar-powered audio devices loaded with Bible content in indigenous languages near isolated peoples’ territories in the Javari Valley. The technology is cheaper and more deniable than a contact mission. The epidemiological risk is identical. Survival International’s 2025 global report estimates that one in six uncontacted peoples worldwide faces active missionary threat \u2014 not hypothetical pressure, but documented, organised efforts to make contact.<\/p>\n\n\n\n
FUNAI’s capacity to respond has been systematically eroded. Under the Bolsonaro government (2019\u20132022), experienced personnel were replaced with staff without relevant knowledge; for every ten roles within the agency, only four were filled; protection posts were closed. Illegal logging and deforestation in indigenous territories increased 138% in the first three years of the Bolsonaro government compared to the preceding three years, according to Human Rights Watch. The Lula government has partially reversed these trends \u2014 most visibly in the 2023 operations to evict miners from Yanomami territory \u2014 but structural recovery of FUNAI’s capacity is incomplete.<\/p>\n\n\n\n
The Yanomami collapse<\/strong>\n\nThe Yanomami are not an uncontacted people \u2014 they have had sustained contact with the outside world for decades. They are what happens after the policy fails.\n\nIn 2024, Fiocruz researchers tested 287 Yanomami individuals across nine villages in Roraima and found mercury in all of them. Eighty-four percent had levels above 2.0 micrograms per gram of hair \u2014 the threshold for mandatory health reporting in Brazil; 10.8% were above 6.0 micrograms per gram, indicating high exposure requiring specialist investigation. The contamination came from illegal gold mining in their territory, which uses mercury in the gold amalgamation process. Yanomami children face the most severe malnutrition of any indigenous population in the Americas. An estimated 570 people died from hunger and disease during the peak years of the mining crisis.\n\nThe Yanomami had constitutional protections. They had FUNAI. They had international attention. What they did not have was a political environment in which those protections were enforced. Their experience is the closest available analogue for what the failure of isolation policy looks like in practice.<\/code><\/pre>\n\n\n\nGlobally, the picture is the same. Survival International’s October 2025 report \u2014 the first comprehensive global survey \u2014 identified 196 uncontacted peoples in 10 countries. Nine out of ten face threats from extractive industries. Half could vanish within a decade without stronger protections.<\/p>\n\n\n\n
Fiona Watson, Survival’s research and advocacy director, called these “silent genocides \u2014 there are no TV crews, no journalists. But they are happening, and they’re happening now.”<\/p>\n\n\n\n
Who is actually deciding<\/h3>\n\n\n\n
The answer to the governing question has specific names attached to it. Not “the state” in the abstract \u2014 specific agencies, organisations, and individuals, with specific interests and vulnerabilities.<\/p>\n\n\n\n
FUNAI is a chronically underfunded Brazilian federal agency whose political independence varies with the government in power. The Indian Navy enforces an exclusion zone around North Sentinel Island with no mechanism for communicating with the Sentinelese in any direction. Survival International is a London-based NGO whose authority to speak for people it cannot contact is not self-evident \u2014 it is an external organisation making claims about the preferences of groups it has never spoken to. Evangelical organisations have their own theological agenda. Illegal miners are responding to commodity prices. The Brazilian president, whoever holds the office, sets the political climate that determines whether FUNAI functions or decays.<\/p>\n\n\n\n
None of these actors has the consent of the peoples they are acting on or against. Every one of them is making a real decision about people who have no idea this is happening.<\/p>\n\n\n\n
FUNAI’s legitimacy is statutory \u2014 it derives from the Brazilian Constitution of 1988, a document the isolated peoples it covers played no part in drafting. The Indian Navy’s authority derives from a 1956 regulation, issued by a colonial successor state, applied to a people who have never interacted with that state in any direction. These are not delegitimising observations in themselves \u2014 laws have to come from somewhere. But they are the actual basis of the authority, and it is worth being clear about what that basis is.<\/p>\n\n\n\n
The miners and missionaries have no authority at all. They are simply more powerful, financially and numerically, than the institutions theoretically constraining them.<\/p>\n\n\n\n
The specific uncomfortable observation: the people who most reliably keep isolated peoples isolated are, right now, the state bureaucracies and NGOs who genuinely believe in the policy. Which means protection depends entirely on the institutional health of underfunded agencies operating in politically hostile environments, and on NGOs competing for attention in a news cycle full of more photogenic emergencies. The extractive economy does not need to win the philosophical argument. It just needs to wait.<\/p>\n\n\n\n
The problem, returned<\/h3>\n\n\n\n
A choice is being made on behalf of people who cannot make it themselves. Nothing in this article dissolves that.<\/p>\n\n\n\n
Two trajectories run from here. In the first, the policy holds: institutions survive, buffer zones are maintained, miners and missionaries are kept at a distance, and isolated peoples live out their lives in the territories they have occupied, uncontacted, for as long as the arrangement holds. This produces, if it succeeds, permanent administered isolation \u2014 peoples frozen in a condition outsiders have decided is their authentic state, protected from history by the political will of states whose relationship to them is entirely asymmetric. There is no endpoint. There is no moment at which the question of what they would want, given genuine options, gets answered.<\/p>\n\n\n\n
In the second trajectory, the policy fails. The miners arrive. The missionaries plant audio devices on the forest floor. The buffer zones erode under commodity prices and bureaucratic decay. Disease follows contact, as it always has, and kills a third to half of the community. The survivors join the ranks of recently-contacted peoples navigating the wreckage \u2014 like the Yanomami, like the Zo’\u00e9, like the Panar\u00e1 before they rebuilt.<\/p>\n\n\n\n
Neither is good. The first is better, contingently, given everything we know about the second. But it is not a resolution. It is a holding action by institutions that are already losing.<\/p>\n\n\n\n
The people being decided about have no idea this debate is happening. The policy that governs their lives is written in languages they have never heard, enforced by states whose existence they are unaware of, defended by organisations whose funding cycles depend on things entirely outside their world. They wake, fish, hunt, repair their shelters, watch the horizon. Whether the institution that keeps the boats away this year will still be funded next year is not a question they can form.<\/p>\n\n\n\n
That fact sits at the centre of whatever position you now hold.<\/p>\n\n\n\n