Photo by Markus Winkler on Unsplash
THE UNCERTAINTY DEPARTMENT
Practitioner Manual Series
When the FBI’s COINTELPRO program was exposed by the Church Committee in 1975, the revelation was not merely that the Bureau had broken the law. It was that a systematic, years-long campaign of surveillance, harassment, and political disruption had operated in plain sight of supervisors who understood exactly what was happening — and that the existing legal and oversight architecture had been entirely inadequate to stop it.
The institutional conditions that enabled COINTELPRO have not been eliminated. They have been refined.
Contemporary deniable operations are more legally sophisticated than their predecessors, more diffuse in their network structures, and more deliberately calibrated to exploit the gaps between what the law prohibits and what it can prove. The Church Committee’s methodologies — public hearings, document discovery, insider disclosure — remain available but face a more deliberately resistant architecture. This manual is written to supplement those methodologies with an analytical framework designed for the architecture as it currently exists.
The manual is intended for investigative journalists working national security and civil liberties beats; congressional oversight staff conducting agency investigations; inspector general offices examining systemic patterns in law enforcement conduct; and civil liberties attorneys building federal civil rights cases where the harm is distributed across actors rather than concentrated in any single chargeable act.
A fundamental methodological principle: The most important records in a deniable operation investigation are frequently the ones that were never supposed to exist in the first place — the mundane documentation generated by ordinary operational activity that no one thought to sanitize. Fuel records. Personnel scheduling. Grant disbursements. Contractor invoices. The investigative strategy described in this manual is built around the systematic exploitation of these residual records.
Direct federal participation in surveillance, harassment, or political disruption triggers Fourth Amendment constraints, Bivens civil liability, statutory limits under FISA and the Privacy Act, and congressional oversight obligations. The operative solution developed over decades is not to avoid the activity. It is to avoid the legal nexus between the activity and the state actor.
The Fourth Amendment and most federal surveillance statutes constrain only state actors. This creates a structural loophole: private individuals conducting surveillance, harassment, or disruption campaigns generate no constitutional violation as long as federal handlers do not direct the specific conduct.
The legal line is agency — whether the private actor is functioning as a government instrument. Courts apply a totality test that is notoriously difficult to satisfy. Passive encouragement, general funding, and after-the-fact receipt of information gathered by a private party have all been held insufficient to establish agency. This means a federal actor can:
…without converting that party into a state actor, provided there is no specific direction of specific investigative acts. The investigative challenge is establishing direction where the architecture is specifically designed to prevent it from appearing in any record.
Fusion centers — regional intelligence-sharing facilities that combine federal, state, and local law enforcement personnel under deliberately ambiguous command structures — were created with the stated purpose of improving information sharing. A secondary effect is the production of accountability gaps.
When an activity is conducted by state or local law enforcement personnel operating under federal information sharing protocols, the question of whether the activity was federally directed depends on which command structure is operative at any given moment, which is not always clear even to the participants. Federal personnel can access, direct, and benefit from activities conducted nominally under state or local authority, while maintaining plausible distance from those activities’ legal exposure.
The most underappreciated element of deniable operation architecture is the role of prosecution declination. When a private party engages in conduct against a target and the relevant U.S. Attorney’s office declines to prosecute, the non-prosecution is not merely inaction. It is a signal to the private party — and to other potential participants — that this category of conduct will not be criminally enforced.
Prosecution declination is difficult to distinguish from ordinary prosecutorial discretion without access to the internal deliberations behind the decision. The investigative question is whether declination was explained to the acting party, by what channel, and whether the explanation amounted to tacit authorization to continue. This is one of the most important questions an investigation can ask and one of the hardest to answer from available records.
Classification authority can be used as a mechanism to prevent oversight of activity that would not survive oversight. When a program is classified, its existence, its authorization, its conduct, and its outcomes are all removed from the arenas — courts, Congress, journalism — where accountability would otherwise be applied.
This does not mean that classified programs cannot be investigated. It means that the investigation must proceed from unclassified residual records — the activity’s footprint in financial systems, personnel records, and contractor databases — rather than from the program’s own documentation.
Understanding the proxy network is the analytical prerequisite for the investigative methodology. Operations do not consist of a single identifiable actor committing identifiable acts. They consist of multiple partially-informed participants, each acting within a plausible frame, whose aggregate activity produces an effect that no single participant’s activity would generate or could be accountable for.
Paid informants operating under formal FBI or other agency handling relationships. These relationships are documented internally but not publicly. The informant’s conduct can be authorized by their handler in ways that are not apparent to outside observers and that generate no publicly accessible record.
Investigative approach: FOIA requests targeting informant payment records (often heavily redacted but structurally informative), cross-referencing names appearing in civil rights litigation, and identifying individuals who appear in proximity to multiple targets over time.
Organizations that appear to be independent civic entities but are funded through attenuated channels that lead back to federal or federal-adjacent sources. These are the most difficult proxy type to identify because their apparent independence is often genuine at the organizational level — members may sincerely believe in the organization’s stated mission — while the funding architecture creates structural dependencies that shape activity.
Investigative approach: Deep financial analysis of nonprofit filings (Form 990), grant databases (USASpending.gov, SAM.gov), and contractor relationships. Look for funding patterns that shift when operational targets shift.
Local law enforcement agencies operating in fusion center environments can be directed toward targets through information-sharing channels without formal federal direction of specific acts. Officers may believe they are acting on independent investigative judgment while the judgment is structured by federally provided information about specific individuals.
Investigative approach: Public records requests for communications between local agencies and fusion centers; personnel assignments to fusion center roles; training records that may reveal federal operational priorities being transmitted to local personnel.
Contractors who hold clearances and can access federal intelligence systems may conduct activities on behalf of federal clients while maintaining the legal status of private actors. The contractor relationship creates plausible distance between federal direction and private conduct.
Investigative approach: Contract databases (USASpending.gov, FPDS); subcontractor relationships; personnel cross-referencing between contractor rosters and individuals appearing in target-adjacent activity.
The critical analytical point is that each proxy type is designed to be investigatable in isolation and to reveal nothing. The operation lives in the pattern — in the coordination across proxy types that no single proxy can describe.
Identifying the aggregated network requires a methodology that begins with pattern detection rather than actor identification. The investigator who begins by looking for a single responsible actor will not find the operation. The investigator who begins by documenting the pattern of effects on the target, and works backward to the actor network that could produce that pattern, has the correct analytical orientation.
Deniable operations are sustained not only by active concealment but by the structured production of non-knowledge within the institutions that nominally oversee them.
Supervisors can maintain operational awareness while preserving plausible non-knowledge through several mechanisms:
The investigative question is not whether a supervisor knew. It is whether the institutional structure created conditions under which knowing or not knowing was a choice the supervisor could exercise — and whether that structure was designed with that optionality in mind.
Personnel can be directed toward targets through assignment decisions rather than operational orders. An agent assigned to a geographic area where a target resides, paired with a supervisor who has a particular interest in that target, provided with information about the target’s activities, and given performance evaluations that reward activity in the target’s vicinity is being operationally directed without any single instruction that could be characterized as such.
Operations are calibrated to produce individual acts that are each, in isolation, defensible as coincidence — ordinary law enforcement activity, routine community engagement, normal commercial behavior — while the aggregate pattern of acts constitutes targeted disruption. The operation resides in the pattern; each element is plausibly coincidental.
This design principle is the central target of the pattern-first investigative methodology.
Begin with the target’s experience. Document, in contemporaneous and corroborated detail, the pattern of effects: what occurred, when, in what sequence, through what actors, with what apparent coordination. This is the evidentiary foundation for everything that follows. It is also the evidence that is hardest for the operation’s architecture to prevent from existing, because it is produced by the target rather than by the operation.
From the effect pattern, identify the actor network that could produce it. This requires reasoning from effects to causes: what level of coordination is implied by the observed pattern? What categories of actors would be required? What information flows are implied by the coordination?
Do not begin with named suspects. Begin with structural positions — roles that must exist in the network for the observed pattern to be possible — and identify who occupies them.
The mundane records that operational security did not think to protect:
Design FOIA requests around the assumption that no formal case file exists. Request records that would be generated by the operation as a byproduct of ordinary administrative activity: travel records, expense reports, fleet records, calendar entries, personnel assignments, training records, grant applications and awards, contractor scope of work documents.
File with multiple agencies simultaneously. The same activity leaves records in different administrative systems; a request to one agency that produces nothing may be supplemented by records at another agency with overlapping jurisdiction.
The financial record is often the most complete and least sanitized record of an operation. Focus on:
Once the actor network is partially mapped, apply network centrality analysis to identify coordination points — actors who appear in multiple relationships within the network, whose position implies they have access to information that crosses proxy type boundaries, and whose behavior therefore suggests they have a more complete picture of the operation than their nominal role would suggest.
Existing federal civil rights law is almost entirely act-based. It provides remedies for discrete unlawful acts by state actors. It does not provide adequate remedies for coordinated distributed campaigns in which each individual act clears every applicable legal threshold while the aggregate constitutes systematic suppression.
This is the critical jurisprudential gap. The argument that needs to be developed — and that has begun to emerge in recent civil rights litigation — is that coordinated distributed interference with constitutional rights is itself a cognizable harm at the pattern level, independent of whether any single act in the pattern crosses the relevant threshold.
This argument has doctrinal support in RICO’s civil provisions, which recognize that a pattern of acts by a network of actors can constitute a harm greater than the sum of its individual acts. The transfer of this reasoning to constitutional rights suppression is the most important available doctrinal development.
Formal legal proceedings are one accountability mechanism. The construction of a comprehensive, corroborated public record — through FOIA litigation, congressional correspondence, inspector general submissions, amicus participation in related litigation, and journalism — is an independent strategy that does not depend on the success of any single legal proceeding and that creates the evidential foundation on which future proceedings can build.
The subject of an extended deniable operation who cannot immediately achieve legal accountability may nonetheless be building the record that makes legal accountability possible in a future proceeding where the relevant doctrine has developed, where the right judge is assigned, or where a congressional committee has been constituted with the political will to investigate. The record survives the proceeding.
| Statute | Purpose | Key Limitations |
|---|---|---|
| 42 U.S.C. § 1983 | Civil rights claims against state actors | Does not reach federal actors |
| Bivens (implied) | Constitutional claims against federal officials | Substantially narrowed since Ziglar v. Abbasi (2017) |
| 18 U.S.C. § 1964 | Civil RICO — pattern of racketeering by enterprise | Requires enterprise + pattern; difficult to plead |
| 5 U.S.C. § 552a | Privacy Act — federal records | Limited to records in federal systems of records |
| 50 U.S.C. § 1809 | FISA criminal provision | Requires proving electronic surveillance |
| 18 U.S.C. § 241-242 | Federal criminal civil rights | Requires DOJ prosecution; rarely pursued |
THE UNCERTAINTY DEPARTMENT is a strategic forecasting and institutional accountability organization. This manual was developed from the application of established legal doctrine, documented historical programs, and structural analysis of contemporary deniable operation architecture. Nothing in this manual constitutes legal advice.