Thomas Gaultier
Menu

When Communities Fear the Water: Mediating Perceived and Actual Contamination

Thomas Gaultier 9 min read

A village downstream of a mine stops drinking from the river. The fish died one week in the dry season. Or the water turned a colour it had not been before. Or a child fell ill and the clinic could not say why. The company sends a laboratory report showing every parameter within the standard. The village does not change its mind. The company concludes the fear is irrational and the complaint unfounded. The village concludes the company is lying. Both sides now have evidence for their position, and neither has a way to test the other’s. The dispute settles into the pattern that fills grievance registers and, eventually, roads.

I have sat in the middle of this argument often enough to know that it is never resolved by more data from the same source. Fear is a fact the process has to address, whatever the laboratory says. It shapes what people drink, where they farm, how they vote and whether they block the road. A mediator who treats it as a misunderstanding to be corrected has already lost the room. A mediator who treats it as a legitimate position that deserves a credible answer can usually find one. This article is about how contamination disputes work, why company data fails to settle them, and how a structured process can produce an answer both sides accept.

Why the laboratory report does not work

The company’s report fails for reasons that have nothing to do with the chemistry. The first is source. Data produced by the party accused of causing harm carries no weight with the party alleging it, and no amount of accreditation changes that. Research on water monitoring in mining regions of Peru, Argentina and Colombia found that community monitoring programmes emerged precisely from distrust of the data produced by companies and public institutions. That distrust is not a misunderstanding of science. It is a rational response to information asymmetry: one side owns the sampling points, the laboratory relationship and the interpretation.

The second is language. A table of parameters against guideline values means nothing to a household that wants to know whether its children can drink from the river. The report answers a regulatory question in a regulatory language and leaves the human question untouched.

The third is timing and location. The samples were taken at the compliance point, on the compliance schedule, and the fish died two kilometres downstream on a Tuesday after heavy rain. The community’s observation and the company’s measurement are describing different events, and each side treats its own as the truth.

The fourth is history. If the company has ever been wrong, or has ever been caught minimising, the report inherits that memory. Trust in data is trust in the people who produce it, and a village that has watched a promise broken will not be persuaded by a spreadsheet.

Fear is a fact the process must address

The instinct to separate perceived from actual contamination, and to dismiss the first, is the mistake that keeps these disputes alive. Perceived contamination has consequences as real as actual contamination: abandoned wells, sold livestock, crops not planted, a market that will not buy the village’s produce, children walking further for water. Those are livelihood losses. They happen whether or not the chemistry supports the fear, and they are attributable to the project’s presence, because the fear did not exist before the mine did.

A structured process therefore starts by putting the fear on the table as a legitimate interest rather than a symptom of ignorance. What exactly have people observed, when, where? What have they stopped doing because of it? What would satisfy them that the water is safe, or confirm that it is not? Those questions produce a different conversation from the one that begins with the laboratory report. They also produce information the company did not have: the dates of the fish kill, the colour, the rain, the place. In more than one case I have facilitated, the community’s observations pointed at a real event the compliance monitoring had missed, a spill or an overflow between sampling dates. The company’s confidence in its own data was the obstacle to finding it. The general point I made in Mediating Between Facts and Emotions in Mining Disputes applies with particular force to water. Water fear is bodily, and bodily fear does not negotiate with tables.

Joint fact-finding as the mediation move

The way out of duelling evidence is evidence the parties produce together. Joint fact-finding is the core move in a contamination dispute, and it has a recognisable structure. The parties agree what questions need answering, in the community’s terms as well as the regulator’s. They agree where and when to sample, including the places and moments the community has named. They agree who takes the samples, usually a mixed team with community members trained to observe or to sample, and how the chain of custody is kept. They agree an independent laboratory, chosen together, and they agree in advance how the results will be interpreted and communicated, including what happens if a parameter is exceeded.

The CAO’s guide to participatory water monitoring was written after years of exactly these disputes. It makes the point that the process generates credible information and builds trust at the same time, which is why it resolves conflicts that data alone cannot. A systematic review published in 2024 reaches the same conclusion across many contexts. Citizen monitoring of water quality produces new information, raises understanding, builds trust among the parties and reduces conflict. The evidence for the method is now strong. The barrier is the company’s reluctance to let anyone else hold the bottle. That reluctance is usually framed as a concern about competence or contamination of samples. In my experience it is more often a concern about control, and the community reads it that way too.

The mediator’s job in this is to make the process fair when the parties are unequal. The UN Guiding Principles ask grievance processes to be equitable, giving aggrieved parties access to information, advice and expertise. In a water dispute that means the community has its own technical adviser, paid for through the process but answerable to the community. The adviser can read a laboratory report and explain what it means. Without that, joint fact-finding is the company’s fact-finding with witnesses.

When the water is contaminated

Sometimes the fear is right. Joint sampling shows an exceedance, or a pattern the compliance monitoring never caught. What happens next decides whether the process survives. A company that argues with its own agreed results has destroyed the process and confirmed every suspicion. A company that acts on them immediately and visibly, with alternative water supply first and investigation second, has demonstrated the thing the community most doubted. It will tell the truth when the truth is against it.

The remedy conversation then has the same shape as any other in this work. Stop the source. Supply safe water while the source is addressed. Compensate the losses that the fear and the fact together caused, including the livelihoods lost while people avoided the river. Agree ongoing joint monitoring so that the community does not have to take the company’s word again. And acknowledge, plainly, that the community was right to be worried. The water disputes I discuss in Water Rights Conflicts in Mining are about quantity and access as often as quality, and the same process serves both.

When the water is clean

More often, joint sampling shows the water within standards, and the mediator’s job is not over. A community that agreed the process, watched the samples taken and chose the laboratory is in a position to believe the result, and most do. But belief needs help. The results have to be explained in the community’s language by the community’s adviser, at the water point rather than in the boardroom. The events the community observed still need an explanation. A fish kill happened whether or not the mine caused it, and the process should try to find out what did. And the monitoring has to continue, jointly, because a single clean result does not answer the fear that next month will be different.

There is also a livelihood question even when the water is clean. Households that stopped using the river on a reasonable fear, in the absence of credible information from the company, lost something. Whether the company compensates that loss is a negotiation, but it should not be dismissed. The company’s failure to provide believable information was part of the cause. A project that had shared credible, jointly produced data from the start would not have watched a village drink from a distant borehole for a year. A rumour would not have lasted that long.

Set up the fact-finding before the next dry season

The Joint Water Fact-Finding Protocol is a framework in four sections. They cover framing the questions in the community’s terms as well as the regulator’s, sampling the parties own together, equity through independent advice, and acting on results either way. It is a two-column reference rather than a scored checklist: read the design element on the left and the practice that makes it credible on the right. Use it to draft the terms of reference for a joint monitoring exercise before a dispute, or to structure the first mediated session once one has begun. Download the Joint Water Fact-Finding Protocol

Credible water is negotiated, not measured

No company can measure its way out of a contamination dispute, because the dispute is about who is believed, not about what the instruments say. The only data that settles it is data the community helped produce. That protects the village, which gets a true answer about its water and a say in how the answer is found. It protects the project, which replaces a permanent suspicion with a shared record. Livelihoods and investment depend on the same bottle of water being trusted by both sides.

This is why water disputes belong in a mediated, structured process from the first complaint rather than after the first blockade. The Social Accord Architecture treats joint monitoring as part of the Feedback Loop, designed in Blueprinting before anyone is frightened. Sampling points, schedules, the laboratory and the community adviser are written into the agreement. When a result is contested, the mediation pathway, GROUNDS, gives the parties a way to reopen the question without reopening the war. Dual accountability makes it hold: the company is accountable for acting on the shared results, the community for testing its fears through the process rather than through the road. Where the fear has already emptied the river of its users, a mediator should start by asking what people saw rather than by explaining what the laboratory found. Done that way, the first joint sample is usually taken within a month.

One thing to do before the dry season. Ask the households nearest your discharge point where they would take a sample if they could, and when. Then take one there, with them, and send it to a laboratory they chose. To design a participatory water monitoring arrangement or to mediate a contamination dispute, write to me at thomas@thomasgaultier.com.

I am Thomas Gaultier. I mediate company-community disputes in mining, oil and gas, and energy projects, and I have resolved more than 2,000 claims across Sub-Saharan Africa and Europe. More about the work, or tell me what you are dealing with.

If the relationship around a project is under pressure, or heading there, let us talk.

A 30-minute call is enough for me to understand the situation and tell you honestly whether I can help. Active opposition, a social license problem, an FPIC process that needs an independent facilitator, a negotiation a community has to prepare for, or a conflict that is not public yet but will be. It makes no difference which side of the table you are on. No obligation, no generalist pitch.

Every first conversation is confidential.

Cover of Legacy Grievances and Historical Claims by Thomas Gaultier

New release, and the last one

Legacy Grievances and Historical Claims

The tenth and final volume completes the library. It is the field guide to harm that outlived the operation: transitional justice adapted to corporate disputes, successor liability when the asset changes hands, and the pathway behind more than two thousand resolved claims.