The Evidence Is Never Alone: What Scientific Fraud and Colonial Archaeology Reveal About Trust

What if the most dangerous thing about evidence is not that it can be false, but that it can look authoritative before anyone asks how it got there?

A duplicated microscopy image in a neuroscience paper and a fragment of broken china in a nineteenth century Bogotá household seem to belong to different intellectual worlds. One concerns the molecular machinery of the brain. The other concerns capitalism, identity, and colonial modernity. Yet both raise the same unsettling question: Who gets to decide what counts as reliable knowledge, and what systems make that decision appear natural?

The answer is rarely found in the object alone. A photograph, artifact, statistic, or landscape becomes evidence only inside a network of procedures, institutions, incentives, and interpretations. When that network is invisible, authority can be mistaken for truth. When it is made visible, knowledge becomes more trustworthy, even when it remains incomplete.

This is the deeper connection between scientific integrity and the archaeology of capitalism in colonial contexts. Both reveal that knowledge is not simply discovered. It is produced, organized, circulated, and defended. The central lesson is not merely that researchers should avoid misconduct or that archaeologists should acknowledge bias. It is more demanding: trustworthy knowledge requires an accountable chain connecting material traces to interpretations and interpretations to power.

The Object Does Not Speak for Itself

Scientific culture often presents evidence as if it were self authenticating. A Western blot appears to show a result. A numerical table appears to report a measurement. A microscope image appears to preserve a moment of biological reality. But every one of these objects has a history: who generated it, under what conditions, through which instruments, after which exclusions, and with what transformations during analysis and publication.

The same is true of archaeological evidence. A plantation wall, a cash crop, a road, a household vessel, or a diamond mining tool is not a complete historical statement. It becomes meaningful through questions imposed upon it. Was the road a public service, an instrument of annexation, or both? Was a ceramic fragment evidence of aspiration, coercion, imitation, or ordinary domestic use? Did a cash crop represent local entrepreneurship, economic adaptation, or pressure from a colonial economy that had narrowed the available choices?

The mistake in both settings is what might be called the transparent object fallacy: the belief that evidence directly transmits reality without mediation. In practice, evidence is more like a witness whose testimony must be examined. It may be sincere but mistaken. It may be incomplete. It may have been edited. It may be speaking in a language the investigator does not yet understand.

A scientific image with duplicated regions is therefore not just a flawed picture. It is a damaged link in a chain of custody. The key issue is not only whether an image was intentionally manipulated. Intent matters for accountability, but reliability is a prior question. Can another researcher determine what the image originally represented? Can the raw file be inspected? Can the result be reproduced independently? If not, the claim has lost part of its evidentiary foundation, regardless of the prestige of the laboratory or the eminence of the scientist whose name appears on the paper.

Archaeology confronts a parallel problem when it treats its categories as neutral. Labels such as civilized, peripheral, ethnic, productive, or modern have historically been used to organize colonial worlds. The archaeologist may recover material remains, but the interpretive vocabulary can still reproduce the hierarchy that created the archive in the first place.

Evidence is not a window onto reality. It is a relationship between traces, methods, interpreters, and institutions.

This principle changes how we evaluate claims. Instead of asking only, “What does this object prove?” we should ask four questions:

1. What is the object’s provenance?
2. What transformations occurred before it reached us?
3. Who had the authority to classify it?
4. What would count as a meaningful challenge to the interpretation?

These questions apply equally to a duplicated blot and to a colonial landscape.

Authority Can Hide the Weakest Links

The reputation of a distinguished scientist can function like an invisible experimental control. Readers may scrutinize the methods, but they also rely on social shortcuts. A Nobel Prize, an elite university affiliation, a prestigious journal, or a long publication record signals competence. These signals are useful because no one can personally verify every claim. Yet they can also make scrutiny less likely precisely where scrutiny is most needed.

Public concerns about anomalies in papers associated with Thomas Sudhof illustrate this tension without resolving the question of intent. Image irregularities and a retracted paper can justify serious investigation. They do not, by themselves, establish that every disputed result was fabricated or that a senior researcher personally manipulated data. That distinction is essential. A responsible system must be able to hold two ideas at once: accomplishments may be genuine, and the evidentiary record may still require correction.

The danger is that prestige changes the burden of proof. A junior researcher may be treated as suspect after one irregularity, while a celebrated researcher receives the benefit of every doubt. This is not simply an individual bias. It is an institutional design flaw. When authority substitutes for inspectability, the system becomes vulnerable to what might be called reputation laundering, the process by which a trusted name lends credibility to evidence that has not earned it independently.

Colonial capitalism operated through a similar mechanism. Projects of incorporation were often presented as improvement: new agricultural systems, municipal infrastructure, commercial exchange, or pastoral development. The language of progress could obscure who gained land, who lost autonomy, and whose practices were recast as obstacles to modernization. A road could carry water and also extend government control. A farmstead could increase production and also reorganize local labor and property relations. A commodity could create opportunity and also bind a community to an economy designed elsewhere.

In both research and empire, authority works best when it becomes ordinary. The laboratory protocol is treated as mere procedure. The city boundary is treated as administrative common sense. The journal hierarchy is treated as a neutral ranking. The colonial map is treated as a description rather than an intervention.

This suggests a broader model of institutional trust. Trustworthy systems need not eliminate authority. They need to separate authority from verification. A respected scientist should not be exempt from data scrutiny. A respected discipline should not be exempt from examining how its categories were formed. A government project should not be judged only by the services it provides, but also by the choices and exclusions embedded in its design.

Incentives Produce Similar Distortions

The two domains also meet through incentives. Academic competition rewards novelty, speed, prestigious publication, and grant success. Capitalist and colonial systems reward extraction, expansion, measurable output, and integration into distant markets. In both cases, the institution may praise results while treating the conditions of production as secondary.

That is where distortion enters. A laboratory facing intense pressure may begin to regard messy data as an embarrassment rather than information. A colonial administration may regard local agricultural practices as inefficient because they do not fit a commercial model. A diamond rush can turn expectations into economic facts before the underlying wealth has been secured. A cash crop can become the measure of rational economic behavior even when it increases dependence and vulnerability.

These are not identical phenomena, but they share a structure: a proxy becomes more important than the reality it was meant to represent.

Publication count becomes a proxy for scientific value. Citation prestige becomes a proxy for truth. Commodity output becomes a proxy for prosperity. Infrastructure becomes a proxy for inclusion. A polished image becomes a proxy for experimental validity. Once the proxy dominates, people learn to optimize what is visible to evaluators, even when doing so damages the underlying purpose.

This is the logic of Goodhart’s law in institutional form: when a measure becomes a target, it stops being a good measure. The result is not always deliberate deception. Often it is gradual adaptation to the reward structure. Researchers learn what kinds of results are publishable. Communities learn which identities receive resources. Administrators learn which projects look like progress. Over time, the system produces subjects who behave as its metrics require.

The archaeological record makes this process tangible. Broken china can reveal how people negotiated status through consumer goods, but it can also reveal the limits of a capitalist order that made identity increasingly legible through possessions. Plantation landscapes show that wealth was not an abstract accumulation of money. It was built into roads, labor quarters, fields, surveillance, and uneven access to space. The material world preserves the cost of the metric.

Scientific records preserve costs too. A retraction is not merely a correction to a paper. It may represent wasted experiments, misdirected funding, delayed therapies, and the quiet labor of researchers who tried to reproduce a result that should never have been trusted. The visible error is often smaller than the invisible opportunity cost.

The Missing Institution Is the Auditor

If evidence is relational and incentives distort behavior, then the solution cannot be moral exhortation alone. Telling researchers to be honest is necessary but insufficient. Systems must make honesty easier, error more visible, and correction less catastrophic.

The most important institution in both science and historical interpretation is the independent auditor. In science, this includes data critics, replication teams, journal editors, statisticians, laboratory members, and platforms where concerns can be documented publicly. In archaeology, it includes descendant communities, local historians, interdisciplinary researchers, and scholars willing to examine the discipline’s own categories and political effects.

Independent scrutiny is not an attack on expertise. It is what allows expertise to deserve trust. A researcher who welcomes raw data inspection is not weakened by transparency. A discipline that includes competing interpretations is not less authoritative. It is more resilient because its claims do not depend on everyone accepting the same institutional hierarchy.

The best audit systems share three properties.

First, they preserve provenance. Data should remain connected to the conditions under which it was produced. This means retaining original images, laboratory records, code, field notes, spatial context, and versions of altered files. In archaeology, it means documenting not only the artifact but also the excavation history, ownership, community relationships, and interpretive decisions.

Second, they distribute verification. No single laboratory, journal, museum, university, or government should possess the only credible account. Independent replication and multiple historical perspectives reduce the risk that one authority can define reality without challenge.

Third, they protect correction. If correction destroys a person’s career or a community’s standing, people will hide uncertainty. A healthy system distinguishes between a good faith error, negligent practice, and deliberate deception. It also treats retraction, reinterpretation, and revision as mechanisms of quality control rather than automatic signs that the entire enterprise has failed.

This framework has a practical implication: we should judge institutions less by how rarely they admit error than by how effectively they detect and repair it. An institution with no visible mistakes may be exceptionally careful, or it may be exceptionally good at suppressing criticism.

Key Takeaways

• Follow the chain, not the prestige. When evaluating a claim, trace how the evidence was produced, processed, selected, and presented. A famous name is not a substitute for provenance.

• Separate intent from reliability. An error can undermine a conclusion even when deliberate misconduct has not been established. Ask first whether the evidence remains inspectable and reproducible.

• Interrogate the metric. Identify what the institution rewards: publication, output, growth, visibility, or conformity. Then ask what important reality the metric may be hiding.

• Invite independent witnesses. Seek replication, dissenting interpretations, local knowledge, and perspectives from people affected by the classification or decision.

• Treat correction as infrastructure. Build systems that make revision normal: preserve raw records, document changes, publish corrections, and reward careful verification.

The deepest lesson is that credibility is not a halo attached to a person, discipline, or institution. It is a maintenance system. It must be renewed through visible records, contestable interpretations, distributed authority, and a willingness to discover that an attractive story is incomplete.

A duplicated image and a broken plate can therefore teach the same intellectual discipline. Neither should be asked to carry more certainty than its history permits. The image does not become reliable because a laureate’s name is beneath it. The artifact does not become transparent because an archaeologist has placed it in a display case. Each demands an account of the forces that brought it into view.

That may be the most important shift in how we think about knowledge. The question is not simply whether we possess evidence. It is whether we have built a world in which evidence can answer back.