When Process Becomes Punishment

Legal proceedings can change behavior long before a court decides whether a claim is justified. Defense costs, discovery, delay and reputational pressure can induce settlement, silence or withdrawal even where expected liability is modest. This paper distinguishes the strategic use of law from suppressive lawfare, in which procedural burdens are disproportionate to the plausible merits or remedy. It develops an economic model and a diagnostic framework, then applies them to mass torts, scientific criticism, communications enforcement and regulation. The case studies illustrate mechanisms and evidentiary limits; they do not establish that every proceeding examined is abusive. The practical objective is to preserve access to justice while ensuring that the cost of reaching the merits does not predetermine the outcome.

Process has consequences that a final verdict does not capture. A credible assessment of lawfare must measure those consequences, identify who imposed them and test whether they were proportionate to the alleged wrong. Settlement, regulatory approval and scientific uncertainty each require careful interpretation. None supplies a shortcut to a finding of misconduct.

Lawfare describes the strategic use of law to pursue wider objectives. It can expose concealed wrongdoing, secure compensation and enforce rights. The concern examined here is narrower: proceedings whose cost, repetition or structure produces silence, withdrawal, delay or bargaining leverage substantially independent of a decision on the merits.

The paper identifies six diagnostic indicators, explains the process premium that may make settlement rational even when expected liability is low, and compares legal and scientific questions that are often conflated. Its case studies support targeted safeguards: early resolution of controlling legal issues, reliable expert evidence, proportionate discovery, transparent conflicts and financing, and protection for scientific criticism. The framework applies to powerful defendants as well as claimants and public authorities.

Legal systems properly allow weak-looking claims to be tested. Many genuine wrongs initially appear improbable because the defendant controls the evidence. It would therefore be circular to define lawfare as litigation that fails, or as litigation brought against a favored industry. The analytical question is narrower: when does the use of legal process cease to be mainly a means of adjudicating a plausible wrong and become mainly a means of imposing collateral cost, delay, silence or political leverage?

Health and environmental disputes are particularly vulnerable because their central facts are rarely binary. A chemical may be capable of causing harm at one dose but not present a material risk at ordinary exposure. An association may be statistically detectable yet too confounded to establish individual causation. A regulator may conclude that expected benefits exceed population risks while a jury, considering one claimant, reaches a different conclusion under the applicable legal standard. Litigation can clarify these distinctions. It can also flatten them into a contest between ‘safe’ and ‘dangerous,’ with the costs of the contest shaping the answer.

Strategic litigation is not inherently suspect. Civil-rights organizations, environmental groups, businesses and governments have long used test cases to obtain authoritative interpretations of law. Public-interest litigation may deliberately seek effects beyond the named parties, including precedent, disclosure and institutional reform. Those aims remain compatible with adjudication when the legal theory is candid, the forum is proper, the requested remedy is connected to the alleged wrong and both sides receive meaningful procedural protection.

John Carlson and Neville Yeomans discussed lawfare in 1975 when contrasting conflict conducted through law and words with conflict conducted through violence. Charles Dunlap later developed the term in a military context. These uses treat law as an instrument whose legitimacy depends on its purpose, method and safeguards. The term itself does not establish abuse (Carlson and Yeomans, 1975; Dunlap, 2008).

Margherita Melillo applies this neutral conception to international tobacco control. In Weaponising Evidence, she treats the tobacco-control network and tobacco companies as opposing participants in a sustained legal contest spanning negotiation of the WHO Framework Convention on Tobacco Control, its subsequent implementation and international trade and investment disputes. Her account is valuable because it does not assume that only one side can practice lawfare. International law can spur domestic regulation and also deter it; litigation can suppress conduct and also disclose evidence that would otherwise remain hidden. (Melillo, 2024; Ammann, 2025).

Melillo also places evidence inside, rather than outside, the contest. Scientific findings, economic estimates, institutional expertise and legal interpretations do not simply settle policy disputes from above. Participants select, frame and deploy them to shape treaty language, institutional rules and adjudication. This is especially important in health policy, where WHO’s scientific authority affects whose knowledge is recognized and which uncertainties are treated as legitimate. The resulting inquiry must examine not only whether evidence exists, but who produced it, which forum admitted it, how competing evidence was characterized and whether the rules were applied symmetrically.

Here, suppressive lawfare means the use, threat or multiplication of legal or legal-administrative proceedings in which burdens disproportionate to the plausible merits or remedy produce silence, withdrawal, delay or bargaining leverage independently of final adjudication. This is a proposition to test against conduct and outcomes. A weak claim, a costly defense or a politically consequential result is insufficient on its own.

No single indicator is decisive. A large discovery request may be necessary where records are concealed; confidentiality may protect a genuine victim; several proceedings may reflect several distinct wrongs. The classification becomes persuasive only when multiple indicators converge and reasonable alternative explanations have been tested.

A rational defendant compares settlement not simply with expected damages, but with the full cost of resistance. That cost includes legal fees, executive and scientific time, discovery exposure, insurance premiums, financing constraints, customer reaction and the possibility that an outlier verdict will anchor later demands. Even a defendant confident of the merits may pay for peace. The settlement therefore cannot automatically be read as an admission, just as refusal to settle cannot establish innocence.

Consider a simplified model. Let p be the probability of an adverse judgment as estimated by the target; D the amount payable if that judgment occurs; L the legal cost of reaching judgment; and C the additional cost of resisting, including management time, insurance effects and reputational damage. Expected resistance costs are approximately pD + L + C. A settlement S below that total may be rational even when pD is small. The amount by which S exceeds pD is the process premium in this model.

This is a decision model, not an estimate of abusive litigation. It omits risk aversion, recoverable costs, strategic behavior and uncertainty about damages. Nor is the premium necessarily improper: legal costs may be unavoidable and reputational consequences may follow credible evidence. A wealthy defendant can also impose a large process premium on a legitimate claimant. Empirical work must distinguish these possibilities.

Repeat players can exploit the difference between merits value and process cost. Claim aggregation lowers the cost of finding clients and developing common evidence. Multidistrict litigation can eliminate duplication and make widespread injury compensable, but it also concentrates enormous inventories of claims before individual causation has been tested. Bellwether verdicts are legally individual yet commercially systemic. A spectacular award, even if later reduced, changes the bargaining range across thousands of cases.

Third-party litigation finance adds capital and can improve access to justice. It can also separate control, risk and beneficiary. The important questions are empirical: who selected the theory, who bears adverse costs, who may approve settlement, what return is contractually preferred, and whether financing encourages the pursuit or prolongation of claims that an unfinanced client would resolve. Disclosure should be calibrated to conflicts and control rather than used to expose privileged legal strategy.

Litigation can determine policy as well as compensation. A consent judgment may impose product obligations that extend to large populations, while its terms are negotiated by public authorities, defendants and lawyers. David Zaruk describes this as adversarial regulation. The question is whether the legal process provides adequate participation, transparency and evaluation for the scope of the resulting rules (Zaruk, 2026).

Colorado’s attorney general announced a proposed nationwide Meta settlement on August 26, 2026, involving 47 states and several territories. The announcement described payments exceeding $17 billion and changes to young users’ access, notifications, parental tools and age assurance. It said the agreement remained subject to court approval. These are the state’s account of the agreement and allegations, rather than independent findings of liability (Colorado Attorney General, 2026).

The case illustrates how settlement can shape product design beyond the immediate parties. That regulatory effect does not by itself establish suppressive lawfare. Claims concerning children’s welfare warrant investigation, and discovery can reveal information unavailable through voluntary disclosure. A defensible assessment must examine the legal authority, evidence, settlement terms, participation of affected groups, financing and subsequent evaluation. The appropriate conclusion is a need for scrutiny, not an assumption of corporate innocence or improper coordination.

The framework also applies when public authorities initiate proceedings. Britain’s communications offences include provisions governing threatening, grossly offensive and other prohibited messages. The relevant statutory rules have changed over time, including changes under the Online Safety Act 2023. Threats, harassment and abuse can justify enforcement. The question for this research is whether arrest, investigation and bail conditions also impose disproportionate burdens on protected expression (Crown Prosecution Service, 2025).

A September 2026 Big Brother Watch report records at least 62,199 arrests and 18,520 charges for specified communications offences during 2021–2025, drawing on incomplete police returns. It separately reports 12,292 convictions in England and Wales. The sources, coverage and timing differ, so these figures cannot be treated as a single cohort or used to calculate a conviction rate. The report is produced by a civil-liberties organization and its interpretation should be distinguished from the underlying administrative data (Big Brother Watch, 2026, pp. 48–57).

These totals do not establish how many cases involved lawful dissent or an unnecessary arrest. They identify a research question about proportionality. Case records are needed to distinguish threats and harassment from controversial opinion, document restrictions and duration, and assess whether less intrusive procedures were available. A case ending without conviction may impose real costs, but that outcome alone does not demonstrate an improper investigation.

Science manages uncertainty through replication, criticism and revision. Courts must deliver an answer for identified parties on a fixed record. The resulting tension is unavoidable. Expert-admissibility rules can screen unsupported methods, but admissibility means that evidence may be considered; it does not transform a contested inference into scientific consensus. Conversely, exclusion of an expert may defeat a legal claim without proving that a hypothesized association is impossible.

Federal Rule of Evidence 702 requires the proponent to establish the admissibility conditions on a more-likely-than-not basis, including sufficient facts or data, reliable methods and reliable application. Its 2023 amendment clarified the court’s gatekeeping responsibility. That standard concerns admissibility; it does not turn a disputed causal inference into a scientific consensus. The scrutiny must apply equally to claimant and defense experts (Federal Rules of Evidence, Rule 702).

Legal leverage often begins before a claim is filed. A classification or expert synthesis can convert uncertainty into a portable proposition that is cited in academic work, public communication and pleadings. The next paper examines this upstream stage through IARC's cancer hazard classifications. The present paper establishes only the general point: courts must distinguish the institutional authority of a scientific source from the additional evidence needed to prove exposure, causation, duty and damages.

Research developed for regulatory or legal use is not thereby invalid. The relevant safeguards are disclosure of intended use and funding, transparent assumptions, sensitivity analysis, replication and a visible separation between scientific inference and the legal conclusion sought. Paper 2 tests those requirements against detailed evidence rather than treating litigation orientation as proof of poor science.

Climate litigation sharpens the distinction between a scientific proposition and a legally sufficient causal chain. It is one proposition to conclude that anthropogenic greenhouse-gas emissions warm the climate. It is another to attribute a defined increment of a particular flood, fire, drought or heat event to anthropogenic warming, and a further step to assign a legally meaningful share of that event and its damages to one producer. Each step may be informed by science, but the evidentiary burden grows as the chain becomes more defendant-specific.

Recent Firebreak essays by David Zaruk argue that foundation-backed climate litigation and the growth of event and source attribution have become mutually reinforcing. That argument should not be accepted merely because the same organizations appear in several parts of the ecosystem. But it identifies a testable lawfare question: does anticipated litigation create demand for scientific outputs that lower the cost of pleading, financing and communicating causation, and are the resulting assumptions and conflicts disclosed clearly enough for courts to distinguish scientific inference from legal attribution?

The relevant safeguard is symmetry. Research commissioned or funded by industry should disclose its interests, assumptions and intended use; the same standard should apply to foundation-supported attribution research, litigation-oriented academic work and institutional assessments prepared for judicial audiences. Litigation orientation does not make a study false. It does make transparency about funding, authorship, model choice and uncertainty especially important.

‘Talc litigation’ is not a single scientific claim. Ovarian-cancer cases generally concern perineal use of cosmetic talc and epidemiological inference. Mesothelioma cases may instead allege contamination of talc with asbestos and require mineralogical identification, product tracing and exposure reconstruction. Combining the two can obscure both meritorious evidence and weak extrapolation.

Two appellate decisions illustrate why cases must be read individually. In Lanzo v. Cyprus Amax Minerals, the New Jersey Appellate Division reversed a judgment and ordered new trials, identifying problems including the scientific foundation for expert evidence. In Bader v. Johnson & Johnson, the California Court of Appeal affirmed a mesothelioma verdict, addressing challenges to expert testimony and evidence of concealment. Neither decision resolves every talc claim or the distinct epidemiological question of ovarian cancer (Lanzo, 2021; Bader, 2022).

Talc also illustrates how litigation can migrate beyond compensation. Advertising recruits potential claimants and influences consumer beliefs. Corporate efforts to challenge critics or expert witnesses can themselves look retaliatory, particularly when the challenged speech relates to pending claims. Bankruptcy strategies designed to aggregate or cap liability raise a different concern: a powerful defendant may use procedure to limit jury access. A credible paper must examine process abuse in both directions rather than assume that only plaintiffs or only corporations can practice lawfare.

The required dataset should separate ovarian cancer from mesothelioma; record product, alleged exposure period and diagnostic evidence; identify the general- and specific-causation experts; code admissibility rulings and appellate disposition; and distinguish compensatory from punitive awards. Without that separation, totals of claims or verdicts convey scale but little about scientific reliability.

On June 25, 2026, the Supreme Court held in Monsanto Co. v. Durnell that federal pesticide law preempted the failure-to-warn claim before it because the claim would require a cancer warning different from the EPA-approved label. The decision concerns the relationship between federal labeling requirements and a particular state tort duty. It is not a determination of every scientific claim about glyphosate, nor a blanket disposition of all possible claims against a manufacturer (Monsanto Co. v. Durnell, 2026).

The institutional question is how long an unresolved legal conflict can sustain costs across a large inventory of cases. Measuring that effect requires filing histories, decisions, defense costs and settlement records. The fact that a threshold issue eventually defeats a claim does not establish that all preceding litigation was abusive. Paper 2 examines the scientific and institutional record in greater detail.

Speech cases provide cleaner examples of process-generated harm because the requested outcome often includes retraction or silence. The British Chiropractic Association sued science writer Simon Singh over a newspaper article criticizing claims made for chiropractic treatment. The association discontinued the action after Singh prevailed on an important appellate question concerning the meaning and character of his words. The eventual result favored speech, but only after prolonged cost and uncertainty. The case helped build support for English defamation reform. (Dyer, 2010).

The case also shows why eventual success is an incomplete measure of the burden on a critic. Legal expenses, lost research time and uncertainty can persist even where challenged speech survives. Whether colleagues respond by withholding criticism is a separate empirical question. It should be tested through contemporaneous records, interviews and comparisons with researchers who continued speaking, rather than assumed from the prominence of one case.

Publicity can also defeat the strategy. Litigation may amplify disputed criticism, attract pro bono support and stimulate legal reform. The expected chilling effect therefore depends on institutional backing, insurance, anti-SLAPP protection, public visibility and the target’s ability to survive several years without resolution. Research should compare silenced critics with similarly situated critics who continued speaking.

Litigation affects actors who are not parties. Agencies know that authorizations, enforcement choices and public statements may be challenged under administrative law. Firms know that regulatory compliance may not prevent tort claims unless legislation clearly gives it preemptive effect. Both may become defensive: agencies ask for more studies, lengthen review, narrow language or postpone decisions; firms abandon products, indications or research programs whose legal tail is difficult to price.

Tobacco harm reduction offers a testable example, but not yet a proven one. FDA has faced court-imposed deadlines, manufacturer challenges and litigation by public-health organizations while processing millions of applications for newly deemed tobacco products. The agency’s demanding ‘appropriate for the protection of public health’ standard and the scale and quality of applications independently explain much delay. To attribute slowness to litigation aversion, researchers would need internal memoranda, testimony or consistent decision patterns showing that anticipated suits caused demands for evidence beyond what scientific review alone required. (FDA, 2021).

The key concept is an option value of delay. An agency criticized for moving slowly can request more evidence; an agency that authorizes a controversial product owns a visible decision and litigation risk. If the cost of delay is diffuse—continued smoking, forgone innovation or unavailable products—while the cost of authorization is concentrated and reviewable, institutional incentives may favor demands for additional certainty. This asymmetry should be measured rather than asserted.

Comparable pressures may arise in international health governance, where employment, access and reputation can affect participation. Such pressures require their own evidence and should not automatically be classified as lawfare. Paper 4 examines that institutional setting.

The next stage should construct a case-level dataset and a smaller set of process-tracing studies. The unit of analysis should be a claim or proceeding, not a company or controversy. Each record should identify the legal theory, forum, filing and disposition dates, claimant-specific evidence, discovery burden, expert rulings, financing where known, verdict and post-trial result, settlement, confidentiality, insurance consequences and observable changes in speech, research, products or regulatory conduct.

Interviews should include claimant and defense counsel, plaintiffs, scientists, insurers, regulators, product developers and critics who chose not to speak. Anonymous testimony can identify hypotheses, but strong published claims should rest on primary documents or multiple independent sources. Counterexamples are essential: cases in which litigation uncovered concealed risks, corrected agency failure or compensated genuine injury should be analyzed with the same framework.

A useful comparison would match cases with similar underlying risks but different procedural exposure. Researchers should record costs on both sides, distinguish an unsuccessful claim from misconduct, and identify what evidence would change the classification. Anonymous testimony can identify leads; a published allegation should normally be supported by a primary document or independent corroboration.

The cases illustrate different mechanisms, not a single coordinated campaign. Talc demonstrates the need to distinguish diseases, exposure theories and expert rulings. Durnell illustrates the importance of resolving a controlling legal issue. Singh shows the burdens of defending scientific criticism. The Meta agreement illustrates regulation through settlement. The communications figures and FDA review record identify hypotheses about procedural effects that still require evidence at the level of individual cases or decisions.

No single reform fits every proceeding. Anti-SLAPP rules are suited to speech claims; stronger expert gatekeeping addresses unreliable causal inference; early claimant screening addresses unsupported inventories; coordinated discovery can reduce duplication; and disclosure of litigation finance can reveal conflicts or control. Cost shifting may deter abusive claims but can also block legitimate plaintiffs who face powerful defendants. Each safeguard should target the mechanism of abuse and preserve a route for well-supported claims.

Three principles cut across the categories. First, threshold legal questions should be decided early where they can dispose of thousands of claims. Second, courts and commentators should distinguish hazard, population risk and individual causation. Third, transparency must be reciprocal: plaintiffs should disclose material financing and claimant-specific support, while defendants should disclose adverse evidence and communications required by law. Procedural discipline is more credible than categorical immunity.

A final classification should require more than an adverse proceeding. At minimum, the record should identify the initiator, legal authority, pleaded or threatened remedy, duration, principal procedural burdens, outcome and observable conduct change. Where motive cannot be documented, the paper should state the inference and test credible alternatives.

The strongest evidence will be contemporaneous rather than retrospective: demand letters, deposition notices, billing or insurance records, internal risk memoranda, abandoned research protocols, altered regulatory timelines and communications explaining why a speaker withdrew. Interviews should generate leads, but a publishable allegation should normally be corroborated by a document, a public record or a second independent source.

The practical test is whether legal process helps establish and remedy a plausible wrong, or imposes disproportionate burdens that determine conduct before the merits can be reached. A final verdict captures only part of that story. Costs, delay, settlement conditions and changes in speech, research or product availability also matter.

CGHA’s framework therefore asks for evidence of means and effects. It supports access to courts, reliable scientific testimony, reciprocal transparency and proportionate procedure. Those safeguards protect legitimate claimants and defendants alike. The purpose is a legal system in which the cost of establishing the facts does not decide the dispute in advance.

Lawfare and the Production of Policy examines how legal process, scientific authority, professional enforcement and international institutions shape public health policy. The first four papers develop the framework and its applications. The fifth examines the incentives that can sustain campaigns as evidence and circumstances change.

Big Brother Watch (2026) Britain’s Free Speech Problem: How Growing Restrictions Are Stifling Speech Online and Offline. Available at: https://bigbrotherwatch.org.uk/wp-content/uploads/2026/09/Britains-Free-Speech-Problem-070926.pdf

Crown Prosecution Service (2025) Communications Offences. Updated 24 March. Available at: https://www.cps.gov.uk/prosecution-guidance/communications-offences (Accessed: 7 September 2026).

Great Britain (1988) Malicious Communications Act 1988, s.1. Available at: https://www.legislation.gov.uk/ukpga/1988/27/section/1 (Accessed: 7 September 2026).

Great Britain (2003) Communications Act 2003, s.127. Available at: https://www.legislation.gov.uk/ukpga/2003/21/section/127 (Accessed: 7 September 2026).

Federal Rules of Evidence, Rule 702 (2023 amendment). Available at: https://www.law.cornell.edu/rules/fre/rule_702 (Accessed: 21 October 2026).

Monsanto Co. v. Durnell, 609 U.S. 371 (2026), No. 24-1068. Available at: https://www.supremecourt.gov/opinions/25pdf/24-1068_n7ip.pdf (Accessed: 21 October 2026).

Lanzo v. Cyprus Amax Minerals Co., 467 N.J. Super. 476 (App. Div. 2021). Available at: https://law.justia.com/cases/new-jersey/appellate-division-published/2021/a5711-17.html (Accessed: 21 October 2026).

Bader v. Johnson & Johnson, A158868 (Cal. Ct App. 2022). Available at: https://law.justia.com/cases/california/court-of-appeal/2022/a158868.html (Accessed: 21 October 2026).

Dyer, C. (2010) ‘British Chiropractic Association drops libel action against Simon Singh’, BMJ, 340, c2086. Available at: https://www.bmj.com/content/340/bmj.c2086 (Accessed: 7 September 2026).

US Food and Drug Administration (2021) Perspective: FDA’s Progress on Tobacco Product Application Review and Related Enforcement. Available at: https://www.fda.gov/tobacco-products/ctp-newsroom/perspective-fdas-progress-tobacco-product-application-review-and-related-enforcement (Accessed: 7 September 2026).

Ammann, O. (2025) ‘Lobbying, civil society organizations, and the international law of tobacco control’, Leiden Journal of International Law, 38(2), pp. 390–403. Available at: https://doi.org/10.1017/S0922156524000542 (Accessed: 7 September 2026).

Carlson, J. and Yeomans, N. (1975) ‘Whither Goeth the Law—Humanity or Barbarity’, in Smith, M. and Crossley, D. (eds) The Way Out: Radical Alternatives in Australia. Available at: https://www.laceweb.org.au/whi.htm (Accessed: 7 September 2026).

Dunlap, C.J. Jr (2008) ‘Lawfare Today: A Perspective’, Yale Journal of International Affairs, 3(1), pp. 146–154. Available at: https://scholarship.law.duke.edu/faculty_scholarship/3154/ (Accessed: 7 September 2026).

Melillo, M. (2024) Weaponising Evidence: A History of Tobacco Control in International Law. Cambridge: Cambridge University Press. Available at: https://doi.org/10.1017/9781009354394 (Accessed: 7 September 2026).

Colorado Attorney General (2026) ‘Attorney General Weiser announces historic settlement with Meta Platforms’, 26 August. Available at: https://coag.gov/press-releases/weiser-announces-historic-settlement-with-meta/ (Accessed: 21 October 2026).

Zaruk, D. (2026) ‘The Meta Trial: A Tale of Two Lawmakers’, The Firebreak, 10 September. Available at: https://www.thefirebreak.org/p/the-meta-trial-a-tale-of-two-lawmakers (Accessed: 10 September 2026).

Zaruk, D. (2026a) 'Climate Attribution: The Latest Litigation-Led Science', The Firebreak, 15 September. Available at: https://www.thefirebreak.org/p/climate-attribution-the-latest-litigation (Accessed: 6 October 2026).

Zaruk, D. (2026b) 'What is Climate Attribution Science?', The Firebreak, 23 September. Available at: https://www.thefirebreak.org/p/what-is-climate-attribution-science (Accessed: 6 October 2026).