Using more of the problem as the solution, at concentrations below the detection threshold, documented as a validated remediation procedure. EPD at its most economically rational.
The Dilution Method is EPD applied to physical contamination events and analogous situations in other regulated domains where a threshold-based detection system creates a calculable line between "compliant" and "non-compliant." The logic is simple and follows directly from the threshold structure: if the contamination level is above the threshold, a non-conformance exists and must be documented, investigated, and remediated. If the contamination level is below the threshold, no non-conformance exists in the formal record. Dilution reduces the concentration of the contaminant — or its measured analog — to below the threshold, eliminating the non-conformance without eliminating the contaminating substance or addressing its source.
This is the most economically rational EPD mechanism because it produces genuine regulatory resolution — the non-conformance does not exist in the formal record, because the threshold is not exceeded — while avoiding the most costly elements of genuine remediation: the production stoppage, the quarantine of potentially affected product, the root cause investigation that might reveal structural problems, and the corrective action that would require systemic changes. Dilution is cheaper than remediation, eliminates the non-conformance formally, and produces no formal record of the underlying problem that generated the need for dilution. It is the EPD mechanism for situations where the problem is physically unavoidable but formally manageable.
In food and drug manufacturing, the Dilution Method most commonly appears as a product-on-product flush: using conforming product to clear non-conforming material from a production line, at a ratio calculated to push the concentration of the non-conforming material below the applicable regulatory threshold, and documenting the resulting production as conforming.
Consider an allergen contamination event. A production line that has run a product containing peanuts is subsequently scheduled for a product that is labeled and sold as peanut-free. Standard remediation requires a validated allergen cleaning procedure that reduces peanut residue below the applicable limit. Validated cleaning takes time, uses cleaning agents that must be rinsed out, and may require laboratory testing of swabs from the cleaned surfaces before the line can return to production.
The Dilution Method alternative: flush the line with the peanut-free product, at sufficient volume to dilute the peanut residue below the applicable limit. The flushed product — which contains some quantity of peanut allergen distributed throughout its volume — may be incorporated into the production batch if the final batch-level peanut concentration is below the threshold that would require labeling or generate a non-conformance. The batch is documented as conforming. No cleaning validation is required. No laboratory swab testing is required. The line returns to production without downtime. The peanut allergen is distributed through the product rather than removed from the process.
This is not hypothetical: product-on-product allergen flushes appear in FDA Warning Letters and product recall investigations as a documented practice at food manufacturers whose allergen programs are inadequate. The recall events associated with these practices typically occur when the flush ratio was miscalculated, when the allergen residue was higher than expected, or when the consumer affected by the distributed allergen experienced a reaction severe enough to trigger investigation and tracing back to the production event.
In pharmaceutical manufacturing, the analogous mechanism is batch reclassification: when a manufacturing batch produces out-of-specification results for a specific attribute, the batch is blended with conforming material at a ratio that brings the out-of-specification attribute within limits, and the resulting blended batch is documented as conforming. This practice, sometimes called "blending to specification," is specifically prohibited for many pharmaceutical applications by regulatory guidance but has been documented in enforcement actions as a widespread practice at manufacturing sites that have not implemented the prohibition effectively.
The EPD function of batch reclassification is identical to the flush: it resolves a non-conformance formally by distributing the non-conforming material below the threshold of detection, rather than by quarantining the non-conforming batch and investigating its root cause. The regulatory non-conformance disappears from the record. The underlying cause of the non-conformance — the process variability that produced an out-of-specification batch — remains in the process, available to produce non-conforming batches in the future.
In environmental regulation, the Dilution Method has a formal name: "dilution to compliance." Discharging regulated pollutants to a receiving water body at concentrations above permit limits requires remediation — typically treatment technology that reduces the concentration of the pollutant in the discharge stream. Diluting the discharge stream with uncontaminated water before the discharge point reduces the concentration of the pollutant in the mixed discharge, potentially bringing it within permit limits without reducing the total mass of the pollutant discharged.
Dilution to compliance is explicitly prohibited or restricted under most modern environmental permit frameworks, but its prohibition must be enforced — and enforcement requires demonstrating that dilution rather than treatment was the mechanism bringing the discharge into compliance, which typically requires flow measurement data and mass balance calculations that are not routinely available to regulators without specific investigation. In jurisdictions where such enforcement is under-resourced, dilution to compliance remains a de facto available strategy for regulated dischargers facing treatment cost pressures.
The platform analog of the Dilution Method appears in content moderation rate reporting. Platforms are increasingly required or expected to report on the prevalence and removal rates of policy-violating content: how much violating content was identified, how much was removed, and how long it remained on the platform before removal. These metrics are used by regulators, researchers, and civil society to evaluate the adequacy of content governance.
The Dilution Method in this context: the denominator. When a platform reports that 95% of identified hate speech was removed, the metric appears to indicate effective enforcement. If the identification methodology detects a sample of total violations — say, the 5% that can be reliably detected by automated systems — then 95% of 5% of the total represents a substantially different outcome than 95% of total violations. The numerator (items removed) is accurately reported; the denominator (items identified, as a fraction of total violations) is defined by the identification methodology, which the platform controls.
Adjusting the identification methodology — broadening the sample, using more sensitive detection, including categories previously excluded from reporting — would increase both the numerator and the denominator, changing the reported prevalence figure and potentially changing the removal rate. The choice of identification methodology determines the denominator, which determines the rate. The Dilution Method in platform governance is the choice of an identification methodology that produces a denominator whose rate, when calculated, produces a result below the threshold of regulatory concern.
| Domain | The contamination | The dilution | What is distributed but not removed |
|---|---|---|---|
| Food manufacturing | Allergen residue above detection threshold | Product-on-product flush to sub-threshold concentration | Allergen distributed through product batch, consumer exposure not prevented |
| Pharmaceutical | Out-of-specification batch attribute | Blending with conforming material to bring within specification | Root cause of non-conformance — process variability — remains unaddressed |
| Environmental | Discharge pollutant above permit limit | Water addition to discharge stream before measurement point | Total pollutant mass discharged unchanged; receiving body exposure unchanged |
| Platform governance | Policy-violating content above threshold prevalence | Identification methodology denominator sized to produce sub-threshold rate | Total policy-violating content present on platform unchanged; user exposure unchanged |
The Dilution Method has two properties that make it the most attractive EPD mechanism in the toolkit. First, it resolves the formal non-conformance without investigation: because the final concentration is within limits, there is no non-conformance to investigate, no corrective action to implement, and no disclosure to make. The regulatory consequence of the event is zero. Second, it avoids the cost of genuine remediation: genuine remediation requires identifying the root cause, implementing a correction, verifying the correction's effectiveness, and potentially recalling affected product. Each of these steps is costly in time, resources, and potentially in regulatory exposure — if the investigation reveals the root cause, the root cause may itself require disclosure. Dilution avoids all of them.
The economic calculus is stark: dilution costs the incremental production materials used in the flush, plus any efficiency loss during the flush period. Genuine remediation costs investigation time, potential production hold, potential recall, corrective action implementation, and regulatory reporting if the investigation reveals the source. The regulated entity with adequate operational sophistication will make this calculation and choose dilution when the probability of regulatory detection of the dilution strategy is lower than the probability of regulatory exposure from genuine remediation.
EPD-005 examines the endpoint of the EPD toolkit: when all four mechanisms have been deployed and a regulatory inquiry or litigation nonetheless asks for evidence of institutional knowledge of harm, how does the absence of formal records function as an affirmative legal defense? The No-Data Defense is the legal architecture that converts the Verification Gap, the SOP Lacuna, the Tiered Disclosure Architecture, and the Flush Doctrine into a single, coherent assertion: the institution had no formal record of this harm.
Dilution is sometimes the only practical remediation option. If the contamination level is marginal and the dilution ratio is conservative, the resulting product may be safer than the available alternatives, including disrupting the supply chain. Blanket prohibition of dilution is scientifically unsophisticated.
The objection is plausible in specific factual contexts. There are circumstances in which dilution to below a science-based threshold produces a product that poses negligible consumer risk, where the supply disruption from quarantine and recall would itself cause harm, and where dilution is the least-bad option among genuinely bad alternatives. The Flush Doctrine is not characterized by the fact of dilution but by the combination of dilution and the avoidance of root cause investigation. A legitimate dilution decision includes the documentation of the contamination event, investigation of its source, corrective action to prevent recurrence, and transparent reporting of the dilution as a specific remediation approach. The Flush Doctrine is the use of dilution that omits those elements — producing a formally clean record while leaving the root cause unaddressed and the decision-making about dilution opaque to regulatory review.
Internal: This paper is part of Engineered Plausible Deniability (EPD series), Saga VI. It draws on and contributes to the argument documented across 23 papers in 5 series.
External references for this paper are in development. The Institute’s reference program is adding formal academic citations across the corpus. Priority papers (P0/P1) have complete references sections.