Abstract

For nearly a quarter-century, the Board of Immigration Appeals (BIA) has enforced a narrow, text-driven interpretation of the Convention Against Torture (CAT) that demands proof of specific governmental intent to inflict severe pain or suffering before indefinite detention or substandard prison conditions can qualify as torture. The doctrine rests on three interlocking regulatory pillars—specific intent, the general-intent contrast, and the lawful-sanctions exception—that the BIA has applied with remarkable consistency across widely varying country conditions and respondent profiles. Beginning with the foundational en banc decision Matter of J-E-, 23 I&N Dec. 291 (BIA 2002), the Board has repeatedly rejected generalized country-conditions claims. A trilogy of modern precedents—Matter of A-A-R-, 29 I&N Dec. 38 (BIA 2025) (El Salvador), Matter of S-S-, 29 I&N Dec. 136 (BIA 2025) (Haiti), and Matter of D-J-L-, 29 I&N Dec. 485 (BIA 2026) (Haiti)—revives and fortifies J-E- while adding procedural refinements: de novo appellate review of the torture question, the J-F-F- chain-of-events requirement, heightened scrutiny of mental-health vulnerability claims, and a new expert-impartiality gatekeeping rule. This Article first lays out the core regulatory framework, then examines each decision in turn, maps their chronological and doctrinal evolution, compares their holdings across a uniform analytical matrix, surveys nationwide circuit-court treatment, and analyzes the sweeping implications for CAT practice involving criminal deportees returned to high-violence nations. The quartet of decisions represents the definitive modern statement of CAT detention jurisprudence and will shape removal litigation for the foreseeable future.

Introduction

Article 3 of the United Nations Convention Against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment (“CAT”) imposes a non-derogable obligation on States Parties: no signatory may return a person to a country where that person faces a substantial risk of torture. The United States implemented this obligation through regulations at 8 C.F.R. §§ 1208.16–1208.18, which define “torture” with five cumulative requirements and establish two forms of protection—withholding of removal and deferral of removal—for those who satisfy the standard. To obtain deferral, an applicant must demonstrate that it is “more likely than not” that he or she would be subjected to torture upon removal. 8 C.F.R. § 1208.16(c)(2).

The threshold question in detention-based CAT claims is whether the conditions of confinement that a criminal deportee will predictably face upon return constitute “torture” within the regulatory definition. That definition requires, among other elements, that severe pain or suffering be “specifically intended” by the inflicting government official. 8 C.F.R. § 1208.18(a)(5). The distinction between specific intent—purposeful cruelty—and general intent—recklessness or knowledge—is the axis around which the entire body of prison-conditions CAT jurisprudence rotates.

Beginning with its en banc decision in Matter of J-E-, 23 I&N Dec. 291 (BIA 2002), the Board of Immigration Appeals has consistently held that indefinite detention in substandard prisons, without affirmative evidence that government officials designed or maintained those conditions as a vehicle for inflicting torture, does not satisfy the specific-intent requirement. The conditions are real, the suffering is foreseeable, and the international human rights record is damning—but under the BIA’s framework, foreseeability is not purpose, and deplorable is not deliberate.

That framework has proven durable. Twenty-three years after J-E-, the BIA returned to the identical doctrinal structure in a trilogy of precedential decisions addressing El Salvador’s mass-arrest “state of exception” policy, Haitian detention of mentally ill deportees, and the evidentiary weight of expert testimony from immigration advocates. These three decisions—Matter of A-A-R-, 29 I&N Dec. 38 (BIA 2025); Matter of S-S-, 29 I&N Dec. 136 (BIA 2025); and Matter of D-J-L-, 29 I&N Dec. 485 (BIA 2026)—do not revise J-E-. They apply it, refine it, and extend it to new factual configurations, while adding procedural tools that make future grants of CAT deferral in mass-detention cases harder to sustain on appeal.

This Article proceeds in ten parts. Part II lays out the three-pillar regulatory framework that underlies every decision in the line. Parts III through VI examine each of the four precedents in turn, analyzing the factual record, the Board’s reasoning, and the doctrinal contribution of each decision. Part VII traces the chronological evolution of the line and the interdependencies among the decisions. Part VIII provides a comparative analytical matrix. Part IX surveys the treatment of J-E- and its progeny in the federal circuit courts, with particular attention to the Third, Second, Ninth, and Eleventh Circuits. Part X assesses the practical implications for practitioners, Immigration Judges, and respondents. Part XI concludes with a synthesis of the doctrinal arc and a forecast of future developments.

Three preliminary notes are warranted. First, this Article focuses exclusively on detention-based CAT claims in the criminal-deportee context. The broader body of CAT jurisprudence—covering acquiescence, non-state actors, and claims arising outside the detention setting—is a distinct and extensive field that this Article does not address. Second, the 2025–2026 precedents are recent enough that circuit court review is limited; the circuit analysis in Part IX therefore draws primarily on pre-2025 case law, identifying the doctrinal ground on which the new decisions will be litigated. Third, this Article treats Matter of J-F-F-, 23 I&N Dec. 912 (A.G. 2006), as part of the framework rather than a discrete “case” in the quartet, because its chain-of-events requirement is invoked procedurally in all three modern precedents.

The Regulatory Framework: Specific Intent, General Intent Contrast, and the Lawful-Sanctions Exception

The statutory and regulatory framework for CAT protection in the United States is the product of the Senate’s conditional ratification of the Convention Against Torture in 1990. The Senate attached an “understanding” specifying that, in order to constitute torture, an act must be “specifically intended” to inflict severe physical or mental pain or suffering. That understanding was incorporated into the implementing regulations promulgated by the Department of Justice, which define “torture” at 8 C.F.R. § 1208.18(a) with five cumulative elements. The entire line of detention-based CAT precedent turns on three interlocking concepts within those regulations.

A. Specific Intent

Specific intent is the cornerstone of the regulatory definition and the dispositive issue in virtually every prison-conditions CAT case. The regulation requires both that severe pain or suffering be “intentionally inflicted” upon the applicant (8 C.F.R. § 1208.18(a)(1)) and that the inflicting act be “specifically intended to inflict severe physical or mental pain or suffering” (§ 1208.18(a)(5)). As the Third Circuit explained, specific intent requires “the intent to accomplish the precise criminal act that one is later charged with,” in contrast to general intent, which “commonly takes the form of recklessness.” Auguste v. Ridge, 395 F.3d 123, 145 (3d Cir. 2005). This is purposeful, deliberate cruelty: the government official must intend not merely the act of detention or the act of maintaining a prison, but must intend the severe pain or suffering as the goal of that act.

The practical significance of this standard is enormous. It means that a government’s decision to detain criminal deportees—even in conditions its own officials know to be life-threatening—does not constitute torture unless the government is maintaining those conditions because it wants deportees to suffer. The intent must be directed toward the harmful consequence, not merely toward the act that produces it. Foreseeability of harm, however near-certain, does not supply specific intent.

B. General Intent—The Contrast That Dooms Most Claims

The regulatory framework explicitly excludes from the definition of torture any harm resulting from “negligent” acts or from a “lack of resources.” 8 C.F.R. § 1208.18(a)(3); Matter of J-E-, 23 I&N Dec. at 299, 301. General intent—the mental state that encompasses recklessness, knowledge, and deliberate indifference to known consequences—precisely describes what the regulations exclude. This is the mental state most commonly present in prison-conditions cases: authorities deliberately detain people in facilities they know to be dangerous, and the resulting suffering is not only foreseeable but foreseen. Yet foreseen harm is not purposeful harm.

The BIA’s classic formulation in J-E- captures the distinction with precision: “Although Haitian authorities are intentionally detaining criminal deportees knowing that the detention facilities are substandard, there is no evidence that they are intentionally and deliberately creating and maintaining such prison conditions in order to inflict torture.” 23 I&N Dec. at 300. The adverb “in order to”—denoting purpose rather than mere foresight—is the operative phrase. Conditions produced by economic hardship, mismanagement, resource scarcity, or bureaucratic indifference reflect general intent at most. They do not constitute torture.

It is important to note that the general/specific intent distinction is not a factual finding but a legal one. Whether the evidence of a government’s motivations supports an inference of specific intent is a mixed question, with predictive facts (what is likely to happen to this respondent) reviewed for clear error and the ultimate legal characterization (does that constitute torture) reviewed de novo. Myrie v. Att’y Gen., 855 F.3d 509, 516 (3d Cir. 2017).

C. The Lawful-Sanctions Exception

Even where conditions are severe and some degree of harmful intent might be inferred, the lawful-sanctions exception provides an independent categorical exclusion. 8 C.F.R. § 1208.18(a)(3) states: “Torture does not include pain or suffering arising only from, inherent in or incidental to lawful sanctions.” Lawful sanctions encompass judicially imposed sentences and other enforcement actions authorized by domestic law, including the death penalty, so long as those sanctions are not “a sham” designed to defeat the object and purpose of the CAT.

In the detention context, this exception is particularly powerful. A government’s policy of detaining criminal deportees—even indefinitely—constitutes a lawful enforcement sanction if it has a legitimate law-enforcement rationale and is authorized under domestic law. Prison conditions that arise as a consequence of that detention policy, and that are not independently designed as a method of inflicting suffering, are “incidental to” the lawful sanction. They are therefore categorically excluded from the definition of torture, regardless of their severity.

Together, specific intent, the general-intent exclusion, and the lawful-sanctions exception function as the BIA’s doctrinal “kill switches” for detention-based CAT claims. A claimant who cannot affirmatively establish that government officials chose to maintain dangerous conditions because they intended to cause suffering—rather than because they lacked resources, lacked motivation to improve conditions, or pursued a legitimate security objective—will fail on specific intent. A claimant whose suffering arises from the predictable consequences of lawful incarceration will fail under the exception. The two analyses frequently overlap, reinforcing each other.

The Foundational Decision: Matter of J-E- (2002)

A. Factual and Procedural Background

Matter of J-E-, 23 I&N Dec. 291 (BIA 2002), arose from the removal of a Haitian national who had been convicted of a drug trafficking offense in the United States. Upon deportation, he faced a predictable sequence: detention by Haitian immigration authorities upon arrival, transfer to the general prison population, and indefinite confinement in facilities that the State Department’s own country reports characterized as severely deficient. The evidentiary record before the BIA documented overcrowding, malnutrition, absence of adequate medical care, lack of sanitation, denial of exercise, and isolated incidents of mistreatment by guards—including burning, choking, hooding, ear-boxing, and electric shock. 23 I&N Dec. at 296–97.

The respondent argued that these conditions, taken in the aggregate and applied to a known criminal deportee whose detention was foreseeable and near-certain, constituted torture under Article 3 of the CAT. The Immigration Judge denied CAT relief. The full Board—fourteen members sitting en banc—affirmed the denial in a decision that would become the foundational precedent for all subsequent detention-based CAT litigation.

B. The Board’s Analysis

The Board approached the case by analyzing each element of the regulatory definition of torture in sequence. On the critical element of specific intent, it made two distinct findings that together disposed of the claim.

First, with respect to the detention policy itself, the Board found that Haiti’s practice of detaining criminal deportees was a lawful sanction. The Haitian Ministry of Justice had implemented the policy to protect the Haitian public from individuals convicted of crimes abroad. The policy served a legitimate law-enforcement purpose, was authorized under domestic Haitian law, and was not a pretext for inflicting suffering. The detention itself—even indefinite detention—therefore fell within the lawful-sanctions exception.

Second, with respect to the conditions of confinement, the Board found that the deplorable state of Haitian prisons was the product of “budgetary and management problems as well as the country’s severe economic difficulties,” not of deliberate governmental design to inflict pain. 23 I&N Dec. at 300–01. There was no evidence that Haitian authorities were “intentionally and deliberately creating and maintaining such prison conditions in order to inflict torture.” Id. at 300. The overcrowding, malnutrition, and lack of medical care reflected resource constraints and mismanagement—the hallmarks of general intent—rather than specific intent to harm.

The Board also addressed the isolated incidents of guard mistreatment documented in the record. Even crediting those incidents, the Board found that anecdotal evidence of individual guards committing abusive acts did not establish that the respondent was more likely than not to experience such treatment. The “more likely than not” standard requires a probability finding tailored to the individual applicant, not a demonstration that severe harm occurs to some prisoners. Because the evidence showed isolated rather than systematic and pervasive intentional mistreatment, the respondent could not satisfy his burden.

The Board’s holding produced three syllabus propositions that have been cited in virtually every subsequent prison-conditions CAT case: (1) indefinite detention of criminal deportees does not constitute torture absent proof that authorities specifically intend to inflict severe pain through that detention; (2) substandard prison conditions do not constitute torture absent proof that government officials deliberately created and maintained those conditions in order to torture detainees; and (3) generalized anecdotal evidence of some prisoners experiencing severe treatment is insufficient to establish the more-likely-than-not standard for a particular applicant. 23 I&N Dec. at 299–302.

C. Significance and Immediate Reception

J-E- was immediately recognized as a precedent of the first order. Because it was decided en banc—a rare proceeding that signals the BIA’s intention to issue binding, nationwide guidance—it foreclosed individualized revisiting of the framework by single BIA panels. And because its holdings were grounded in regulatory text rather than in Haiti-specific factual determinations, the framework was facially transferable to any country where the detention-conditions-as-torture argument might be advanced.

Federal circuit courts quickly accepted J-E- as a reasonable agency interpretation of an ambiguous regulatory scheme, entitling it to substantial deference. The First Circuit deferred to J-E- in Elien v. Ashcroft, 364 F.3d 392 (1st Cir. 2004). The Eleventh Circuit did likewise in Cadet v. Bulger, 377 F.3d 1173 (11th Cir. 2004). The Third Circuit adopted the framework—and added its own glosses on the meaning of specific intent under domestic criminal law—in Auguste v. Ridge, 395 F.3d 123 (3d Cir. 2005). The Second Circuit followed in Franck Pierre v. Gonzales, 502 F.3d 109 (2d Cir. 2007). By the time J-E-’s twenty-third anniversary arrived, the Second Circuit in Franck Pierre could note that “no federal circuit court considering the case of a Haitian criminal deportee has declined to follow In re J-E-.” 502 F.3d at 117.

The Modern Bridge: Matter of A-A-R- (2025)

A. Standard of Review Clarification

In April 2025, twenty-three years after J-E-, the BIA returned to the detention-as-torture framework in Matter of A-A-R-, 29 I&N Dec. 38 (BIA 2025), addressing the El Salvadoran context for the first time in a published, precedential decision. The respondent was a former member of the MS-13 gang who bore visible gang tattoos and had been convicted of murder. El Salvador’s government had, beginning in 2022, implemented a “state of exception”—a sweeping security measure authorizing mass arrests of suspected gang members, suspending certain constitutional guarantees, and resulting in the detention of tens of thousands of individuals in newly constructed mega-prisons. The conditions in those facilities, documented by international human rights organizations, were severe: extreme overcrowding, inadequate nutrition and medical care, reports of deaths in custody, and credible accounts of physical abuse.

Before reaching the merits, the Board addressed a threshold procedural question that had generated inconsistency among Immigration Judges: what standard of review applies to a CAT deferral claim on appeal? The Board clarified the bifurcated standard that circuit courts had been applying: predictive factual findings—what is likely to happen to this respondent if returned to El Salvador—are findings of fact reviewed for clear error under 8 C.F.R. § 1003.1(d)(3)(i). The legal question of whether the anticipated events satisfy the regulatory definition of torture—including whether they reflect specific intent and whether they are excluded by the lawful-sanctions exception—is a question of law reviewed de novo. 29 I&N Dec. at 41–42 (citing Myrie v. Att’y Gen., 855 F.3d 509, 516 (3d Cir. 2017)).

This clarification matters enormously in practice. It means that the BIA’s ability to reverse an IJ’s grant of CAT deferral is constrained on the factual side—the IJ’s findings of what will likely happen must be credited unless clearly erroneous—but unconstrained on the legal side. If an IJ correctly finds that a respondent will be detained indefinitely in a Salvadoran mega-prison where deaths and abuse occur, but then concludes that this constitutes “torture,” the BIA may reverse that legal conclusion without deference. The framework thus preserves IJ fact-finding while restoring BIA control over the legal characterization of those facts.

B. Merits Analysis

On the merits, the Board applied J-E- to the Salvadoran context with a directness that left no ambiguity about its intended reach. Just as Haiti’s detention policy served a legitimate law-enforcement purpose in 2002, El Salvador’s state of exception served a legitimate purpose in 2025: suppressing gang violence that had made El Salvador one of the most dangerous countries in the hemisphere. The legal structure of the Salvadoran policy—authorized under domestic emergency law, directed at a real security threat, designed to protect the public—placed it squarely within the lawful-sanctions exception. 29 I&N Dec. at 44–45.

On specific intent, the Board found that the overcrowding, abuse reports, and deaths in Salvadoran mega-prisons were attributable to the same combination of factors identified in J-E-: resource constraints, management challenges, and the consequences of incarcerating a massive population in rapidly constructed facilities. The government’s parallel efforts to improve conditions—however inadequate—further negated any inference of deliberate cruelty. An entity that simultaneously incarcerated tens of thousands and sought international assistance to improve detention conditions could not plausibly be said to be maintaining those conditions in order to inflict torture.

The Board also applied the Matter of J-F-F-, 23 I&N Dec. 912 (A.G. 2006), chain-of-events requirement to dispose of anecdotal evidence. Under J-F-F-, an applicant cannot establish the more-likely-than-not standard “by stringing together a series of suppositions to show that it is more likely than not that torture will result where the evidence does not establish that each step in the hypothetical chain of events is more likely than not to happen.” 23 I&N Dec. at 917. Reports of abuse suffered by other detainees did not, without more, establish that this respondent—whose individual circumstances, personal history, and particular vulnerabilities within the Salvadoran prison system were not independently analyzed—was more likely than not to be the target of government-sanctioned torture.

A-A-R- thus performed two functions simultaneously. It confirmed that J-E-’s framework applied in full to mass-detention regimes outside the Haitian context—answering the question of whether J-E- was a Haiti-specific decision or a general doctrinal statement in favor of the latter. And it gave practitioners, IJs, and the BIA a precise roadmap for analyzing Salvadoran state-of-exception claims, a category that had proliferated in immigration courts as El Salvador’s security policies generated a new cohort of deportees with serious criminal records and credible evidence of detention risk.

Application to Mental-Health Claims: Matter of S-S- (2025)

A. Specific-Intent Analysis

Two months after A-A-R-, the Board decided Matter of S-S-, 29 I&N Dec. 136 (BIA 2025), extending the J-E- framework to a qualitatively different factual scenario: a Haitian respondent with severe, documented mental illness. The respondent suffered from schizophrenia, schizoaffective disorder, and major depressive disorder, conditions that had required ongoing psychiatric medication and supervision throughout his time in the United States. The Immigration Judge had granted CAT deferral on the theory that, as a seriously mentally ill person detained in the Haitian prison system, the respondent faced a uniquely elevated risk of suffering that constituted torture. Specifically, the IJ found that the respondent would be denied psychiatric medication, would be misunderstood and abused by guards unfamiliar with mental illness, and would face decompensation and serious harm as a foreseeable consequence of his detention.

The BIA reversed. Accepting the IJ’s factual predictions as not clearly erroneous—the respondent would likely be detained, would likely be denied adequate psychiatric care, and was likely to suffer severely as a result—the Board nonetheless found that the specific-intent element was not satisfied. The critical analytical move was the Board’s treatment of the guards’ “lack of understanding” of mental illness. The IJ had implicitly treated guard ignorance about mental health conditions as a form of indifference that contributed to the tortious quality of the treatment. The BIA read that same ignorance in exactly the opposite direction: guards who do not understand what schizophrenia is, or who respond to psychiatric symptoms with physical control because they do not recognize them as symptoms, are not acting with the specific intent to inflict severe pain. They are acting from ignorance—general intent at most, and perhaps not even that. 29 I&N Dec. at 138–39.

This analytical move has significant implications. It means that a government’s failure to train its prison staff in the recognition and management of mental illness—even where that failure predictably leads to the severe suffering of mentally ill detainees—does not support a specific-intent finding. The failure reflects indifference or resource constraints, not purpose. An advocate who cannot identify affirmative evidence that Haitian officials targeted mentally ill detainees because of their mental illness—selecting them for abuse precisely in order to cause them suffering—cannot establish the necessary intent.

The Board also noted that the IJ had not made the threshold factual finding that the respondent was more likely than not to be detained for an extended period. Under J-E-, the analysis of whether detention conditions constitute torture presupposes that the applicant will actually be detained. Where detention is not established as more likely than not, the entire superstructure of the conditions-as-torture argument collapses. S-S- reinforces that this is a required factual predicate, not an assumption.

B. The J-F-F- Chain Requirement Applied to Community-Harm Claims

In addition to the detention-based claim, the respondent in S-S- advanced a community-harm theory: upon release from (or in the absence of) detention, he would face violence and exploitation from members of the Haitian community who stigmatize and target persons with mental illness. The Board applied the J-F-F- chain-of-events requirement to this theory with analytical rigor. To establish that community harm was more likely than not to constitute torture, the respondent would need to demonstrate: (1) that he would be released from detention into the community; (2) that community members would identify him as mentally ill; (3) that identified mentally ill individuals face targeting in Haiti; (4) that such targeting rises to the level of severe pain or suffering constituting torture; and (5) that the Haitian government would consent to or acquiesce in that torture. 29 I&N Dec. at 140. Each of these steps required independent evidentiary support establishing its likelihood at greater than fifty percent, and the respondent had not established each link in the chain.

S-S- thus established that the specific-intent and lawful-sanctions framework is not confined to cases involving detention by governmental authorities. It applies with equal force—through the J-F-F- chain requirement—to community-harm theories that depend on governmental acquiescence in private violence. The more speculative and multi-step the claimed chain of causation, the harder it is to satisfy J-F-F-. A respondent whose claim requires courts to assume a series of individually uncertain events will not meet the burden merely by demonstrating that the final outcome, if the chain held together, would be severe.

Expert Testimony Scrutiny and Final Reinforcement: Matter of D-J-L- (2026)

A. The Expert-Impartiality Rule

The most recent decision in the quartet, Matter of D-J-L-, 29 I&N Dec. 485 (BIA 2026), returns to Haiti and adds a procedural dimension that reshapes the evidentiary terrain for future litigation. The respondent was a Haitian criminal deportee who also suffered from mental illness and whose claim combined elements of the theories advanced in both J-E- and S-S-: indefinite detention, medical neglect, official targeting, and the risk of gang violence. The Immigration Judge had granted CAT deferral following a remand, relying principally on the testimony and affidavit of Michelle Karshan, founder of Alternative Chance, an organization that provides direct assistance to criminal deportees in Haiti. Karshan had testified in numerous prior Haitian CAT cases, including the Ridore litigation in the Ninth Circuit.

The BIA reversed and announced a new expert-credibility rule: an Immigration Judge commits legal error in according significant weight to expert testimony when the expert’s record—including organizational affiliation, advocacy history, and demonstrated reluctance to engage with evidence that contradicts the testimony’s conclusions—reflects a lack of impartiality that undermines the testimony’s reliability as evidence. 29 I&N Dec. at 488–90. The rule does not categorically disqualify advocates as experts, but it requires IJs to engage explicitly with partiality concerns before relying on advocacy-based expertise as a primary evidentiary foundation for a CAT grant.

The implications of this rule are substantial. Expert testimony has been one of the primary vehicles through which applicants attempted to distinguish their cases from J-E-’s general framework. Country-conditions reports, government briefings, and NGO assessments can all describe conditions in general terms; expert testimony is necessary to translate those general conditions into the specific-intent analysis applicable to an individual respondent. If the most experienced and knowledgeable Haiti experts—those who have spent years documenting the conditions criminal deportees face—are precisely the experts most likely to be disqualified under an impartiality analysis, the evidentiary foundation available to applicants contracts significantly.

The rule also represents a direct response to a pattern that the Board identified: the same expert witnesses appearing repeatedly in cases presenting nearly identical claims, relying on the same evidence base, and reaching the same conclusions regardless of the individualized circumstances of each respondent. An expert who cannot adjust their analysis in response to contrary evidence, or who has a demonstrated institutional interest in the outcome of the very type of case they are testifying in, does not supply the kind of neutral, reliable, case-specific analysis that CAT adjudication requires.

B. Merits

Stripped of Karshan’s testimony, the evidentiary record in D-J-L- consisted of generalized country-conditions reports and the respondent’s own testimony. The Board reaffirmed J-E- verbatim: indefinite detention and poor conditions in Haitian prisons do not constitute torture absent specific intent; the conditions fall within the lawful-sanctions exception; and the evidence before the IJ reflected general-intent harms from economic hardship and mismanagement, not deliberate cruelty. 29 I&N Dec. at 490–92. On the gang-harm claim, the J-F-F- chain requirement disposed of the argument in the absence of case-specific evidence establishing each inferential step.

D-J-L- thus reinforces the entire regulatory framework established in 2002 while adding a forward-looking gatekeeping tool. Immigration courts now have authority—and arguably an obligation—to scrutinize the partiality of expert witnesses before according their testimony significant weight. This development will likely generate litigation over the standards for impartiality findings and the procedural rights of respondents to challenge adverse credibility rulings at the threshold of their expert evidence.

Chronological History and Doctrinal Relationships

The doctrinal evolution from J-E- through D-J-L- is best understood as a linear process of extension and refinement rather than as a series of independent decisions. Each decision in the line builds on its predecessors, responds to the gaps and pressure points that prior decisions left open, and forecloses arguments that had been generating inconsistent outcomes in the immigration courts. The quartet is not an accretion of incidentally related holdings; it is a single, multi-decade doctrinal project with coherent internal logic.

The sequence begins with J-E- (2002), which defined the framework’s three pillars in the Haitian context. The decision was almost immediately tested in the Third Circuit—the circuit most frequently hearing Haitian CAT cases at the time—where a panel in Lavira v. Attorney General, 478 F.3d 158 (3d Cir. 2007), suggested that “willful blindness” to the near-certain severe consequences of detention might satisfy the specific-intent requirement. That suggestion, if accepted, would have substantially eroded J-E- by collapsing the specific/general intent distinction wherever harm was highly probable. The Third Circuit’s en banc decision in Pierre v. Attorney General, 528 F.3d 180 (3d Cir. 2008), explicitly rejected the willful-blindness route, reaffirming that even high probability of harm does not supply specific intent. Pierre thus closed the most significant doctrinal opening left by J-E-.

The Attorney General’s decision in Matter of J-F-F-, 23 I&N Dec. 912 (A.G. 2006), ran in parallel, establishing the chain-of-events requirement that prevents applicants from building CAT claims on a sequence of individually speculative steps. J-F-F- addressed a different doctrinal gap than Pierre—not the specific-intent question but the more-likely-than-not standard—but its effect was to foreclose a category of community-harm and compound-risk arguments that had been generating grants at the IJ level.

Between 2008 and 2025, the J-E- framework operated in a state of effective stability. The framework had been accepted by every circuit that addressed it, the BIA had not issued additional precedential decisions revisiting the analysis, and the litigation landscape had adjusted to the framework’s demands. The years 2025 and 2026 shattered that stability—not by changing the framework, but by applying it to new country contexts and new factual configurations that had been generating inconsistent results at the IJ level.

A-A-R- (April 2025) addressed the specific question of whether J-E- applied to El Salvador’s state-of-exception mass detention, a policy that IJs had been treating inconsistently. By analogizing the Salvadoran policy directly to Haiti’s detention procedures and applying J-E-’s analytical framework in full, A-A-R- eliminated that inconsistency and clarified the applicable standard of review. S-S- (June 2025) addressed the specific question of whether mental illness created a sufficiently individualized risk of specific-intent harm to distinguish a case from J-E-, answering that question in the negative when the harm arises from guard ignorance rather than deliberate targeting. D-J-L- (March 2026) addressed the specific question of what evidentiary weight to give to advocacy-based expert testimony in Haitian deportee cases, introducing the impartiality gatekeeping rule.

Each of the three modern decisions cites all of its predecessors, and the citations are not merely formal: each decision affirmatively applies the reasoning of prior decisions rather than merely distinguishing them or noting their existence. The doctrinal structure is genuinely cumulative, with each layer of precedent reinforcing and elaborating the one beneath it.

Comparative Analysis

The four decisions share a common analytical structure but differ in country context, respondent profile, primary harm theory, and doctrinal contribution. The table below maps the decisions across these dimensions:

AspectJ-E- (2002)A-A-R- (2025)S-S- (2025)D-J-L- (2026)
CountryHaitiEl SalvadorHaitiHaiti
Respondent ProfileCriminal deportee/drug offenseFormer MS-13/murder conviction/gang tattoosCriminal deportee/severe mental illness (schizophrenia, schizoaffective disorder, MDD)Criminal deportee+mental illness
Primary Harm TheoryIndefinite detention+substandard conditionsMass detention under state of exception+abuseDetention+denial of psychiatric care+community stigmaDetention+medical neglect+official targeting+gang harm
Lawful-Sanctions AnalysisHaitian detention is lawful crime-control sanctionSalvadoran state of exception is lawful anti-gang sanctionNot primary ground; detention assumed lawfulHaitian detention remains lawful sanction per J-E-
Specific-Intent FindingNo evidence conditions deliberately created to tortureResource constraints+improvement efforts negate intentGuard ignorance of mental illness negates specific intentGeneral-intent harms from economic hardship; no deliberate cruelty
J-F-F- Chain AppliedImplicitly (“more likely than not” standard)Yes, to anecdotal evidence of gang abuseYes, to community-harm theoryYes, to gang-harm and compound-risk claims
Standard of ReviewNot explicitly addressedClarified: clear error (facts)/de novo (law)Applies A-A-R- standardApplies A-A-R- standard
New Doctrinal ContributionThree-pillar specific-intent/lawful-sanctions/general-intent frameworkDe novo legal review; J-F-F- chain requirement in mass-detention contextMental-health application; guard ignorance=absence of specific intentExpert-impartiality gatekeeping rule
Primary Reversal GroundLack of specific intent (en banc)Lawful sanction+anecdotal insufficiency+J-F-F-No long-term detention finding; no specific intent from guard ignoranceBiased expert testimony; generalized evidence
Citations to Prior DecisionsN/A (foundational)J-E- (primary); J-F-F-J-E-; A-A-R-; J-F-F-J-E-; A-A-R-; S-S-; J-F-F-

Several patterns emerge from this comparison. First, the framework is genuinely country-neutral: the same three-pillar analysis that resolved the Haitian claim in 2002 resolved the Salvadoran claim in 2025 without modification. Second, the doctrinal contributions of successive decisions have targeted specific pressure points—the standard of review, the mental-illness argument, the role of expert testimony—that were generating inconsistent results in the immigration courts. Third, each decision has expanded the available tools for reversing IJ grants of CAT deferral without disturbing the underlying framework. The net effect is a body of doctrine that is both stable in its core principles and increasingly granular in its procedural requirements.

Treatment by the Federal Circuit Courts of Appeals

The BIA’s consistent line from Matter of J-E- through Matter of D-J-L- enjoys broad acceptance across the federal circuits, with no significant split on the core legal principles of specific intent, the general-intent distinction, or the lawful-sanctions exception in prison-conditions CAT claims. While circuits differ in their scrutiny of factual findings and the degree of particularization they require, they uniformly defer to the BIA’s regulatory interpretation. The 2025–2026 decisions are too recent for extensive circuit review, but existing precedent provides a reliable forecast of how each circuit will receive the new contributions.

A. Third Circuit (Authoritative Development)

The Third Circuit has produced the most extensive and analytically rigorous body of case law on the J-E- framework, in part because its jurisdiction historically encompassed a disproportionate share of Haitian CAT petitions processed through the immigration courts in New York, New Jersey, and Pennsylvania.

In Auguste v. Ridge, 395 F.3d 123 (3d Cir. 2005), the court’s first major engagement with J-E-, the Third Circuit held that deplorable prison conditions confronting Haitian criminal deportees did not constitute torture under the specific-intent requirement. The opinion’s most significant contribution was its definition of specific intent by reference to domestic criminal law: the court held that specific intent requires “the intent to accomplish the precise criminal act that one is later charged with,” in contrast to general intent which “commonly takes the form of recklessness.” 395 F.3d at 145. The court explicitly endorsed the BIA’s refusal to treat the Haitian government’s knowledge that severe pain might result from detention as equivalent to the intent to produce that pain. Id. at 146. Practitioners should note that Auguste was decided in 2005—three years after J-E-. The cases therefore stand in the correct doctrinal relationship: Auguste built upon and reinforced J-E-; it did not found or precede it.

The critical en banc development came in Pierre v. Attorney General, 528 F.3d 180 (3d Cir. 2008). The court convened en banc to resolve a conflict between two of its own panel decisions: Auguste, which had held that recklessness was insufficient to establish specific intent, and Lavira v. Attorney General, 478 F.3d 158 (3d Cir. 2007), which had suggested that a jailer’s “willful blindness” to the near-certain consequences of detention for an HIV-positive detainee might satisfy the specific-intent requirement. Paul Pierre, a Haitian national restricted to a liquid-only diet due to a self-inflicted esophageal injury, argued that any Haitian official who detained him knowing that he would be denied the medical diet necessary to sustain him was effectively intending his death. The en banc court rejected this argument. Drawing on Auguste’s domestic criminal law framework, the court held that willful blindness to the probable consequences of an act—including near-certain severe consequences—does not transform general intent into specific intent. 528 F.3d at 189–90. The act of detention retains its character as a lawful sanction; only affirmative evidence that the authority of detention is being used as a vehicle for inflicting severe pain supplies specific intent.

In Myrie v. Att’y Gen., 855 F.3d 509 (3d Cir. 2017), the Third Circuit formally articulated the bifurcated standard of review subsequently adopted by the BIA in A-A-R-: clear error for the IJ’s factual predictions about what will happen to the respondent, and de novo for the legal question of whether those anticipated events constitute torture. 855 F.3d at 516. Myrie thus provides the Third Circuit doctrinal basis for the review-standard clarification that A-A-R- issued as national BIA precedent.

B. Second Circuit

The Second Circuit’s primary contribution to the J-E- framework is Franck Pierre v. Gonzales, 502 F.3d 109 (2d Cir. 2007), which presented a factual scenario combining Haitian detention conditions with a diabetic respondent’s need for insulin and a restricted diet. Pierre argued that Haitian officials who detained him knowing he would be denied the medical care necessary to manage his diabetes were effectively sentencing him to death and thereby inflicting specific-intent torture. The Second Circuit denied the petition, deferring to the BIA’s reading of 8 C.F.R. § 208.18(a)(5) and holding that the specific-intent requirement was not satisfied by evidence that the government knew medical care would be inadequate. 502 F.3d at 116–19.

The Franck Pierre decision is significant for a second reason: it surveyed the circuit landscape and confirmed that as of 2007, “no federal circuit court considering the case of a Haitian criminal deportee has declined to follow In re J-E-.” Id. at 117. This observation, made eighteen years ago, remains accurate today. The uniform circuit acceptance of J-E- reflects not merely deference to agency expertise but the courts’ substantive agreement that the BIA’s reading of the regulatory text is the most natural and defensible one.

C. Ninth Circuit

The Ninth Circuit occupies a distinctive position in the J-E- landscape. It accepts the specific-intent requirement as a matter of law—there is no Ninth Circuit decision holding that mere knowledge of foreseeable harm satisfies the CAT standard—but it applies a more searching review to the factual records underlying BIA reversals of IJ CAT grants. This approach creates a narrow but real pathway for applicants with sufficiently particularized records.

The clearest illustration is Ridore v. Holder, 696 F.3d 907 (9th Cir. 2012). Jean Baptiste Ridore, a Haitian criminal deportee, had obtained a grant of CAT deferral from an IJ who made extensive, specific factual findings about the conditions Ridore would face—findings based in substantial part on testimony from expert Michelle Karshan. The BIA reversed, applying a de novo standard to what were properly factual findings. The Ninth Circuit granted the petition in part: it did not find that Ridore’s record established torture, nor did it reject the J-E- legal framework. It held that the BIA had committed legal error by reviewing the IJ’s factual findings—what was likely to happen to Ridore specifically—under a de novo rather than clear-error standard, and remanded for the BIA to review those findings with appropriate deference. 696 F.3d at 914–16.

Practitioners should note that the Ninth Circuit’s Ridore pathway has been partially foreclosed by D-J-L-. The IJ in Ridore relied substantially on Karshan’s expert testimony; the D-J-L- rule on expert impartiality now provides the BIA with authority to discount that testimony before the factual record is even fully assembled. This means that cases in which an applicant’s particularized factual record depended on Karshan or similarly situated experts may no longer generate the kind of detailed IJ findings that the Ninth Circuit has been willing to protect under clear-error review.

A note on a common citation error: some practitioners have cited Ridore v. Holder to the reporter at 670 F.3d 1085 (9th Cir. 2012). That citation is incorrect and corresponds to Rohit v. Holder, an unrelated voluntary-departure case involving a crime-of-moral-turpitude determination. The correct reporter citation for the Ridore case is 696 F.3d 907 (9th Cir. 2012). This distinction matters in practice: an adversary or court that pulls 670 F.3d 1085 will find no CAT content whatsoever.

D. Eleventh Circuit

The Eleventh Circuit handles a substantial volume of Haitian CAT petitions and has consistently applied the J-E- framework. In Jean-Pierre v. U.S. Att’y Gen., 502 F.3d 119 (11th Cir. 2007), the court addressed a Haitian criminal deportee who, like the respondent in Franck Pierre, had serious medical needs—specifically needs tied to the severity of the prison conditions he would face. The Eleventh Circuit did not grant CAT relief, but it remanded to the BIA for a closer engagement with the specific evidentiary record before it. The BIA’s decision had relied primarily on J-E- as sufficient to resolve the case, without specifically addressing the evidence of the respondent’s individual circumstances and whether those circumstances distinguished his claim from J-E-‘s general holding. 502 F.3d at 123–24.

The Jean-Pierre remand is best understood as procedural rather than substantive. The Eleventh Circuit did not suggest that a medical-needs claim, or any particular category of individualized evidence, would satisfy the specific-intent requirement. It required only that the BIA’s reasoning address the specific record before it rather than simply invoking J-E- as a per se bar. This is consistent with the BIA’s own approach in subsequent decisions, including D-J-L-, which did engage with the specific record before reversing.

E. Fifth Circuit

The Fifth Circuit has consistently aligned with the BIA’s restrictive approach to detention-based CAT claims in the Haitian and Salvadoran contexts. Gang-involved and criminal deportees returning to El Salvador under the state-of-exception policy have not fared well in the Fifth Circuit’s CAT analysis, and the circuit has explicitly relied on J-E- and the absence of specific-intent evidence to affirm denials of CAT deferral. Matter of A-A-R-, which directly addressed the Salvadoran state-of-exception context and drew the same conclusions as Fifth Circuit precedent, is therefore well-positioned for deference in that circuit.

F. First, Fourth, Sixth, Seventh, Eighth, and Tenth Circuits

These circuits have consistently deferred to the BIA’s framework without generating significant doctrinal developments of their own in the detention-conditions context. Haitian and Salvadoran prison claims routinely fail absent individualized proof of deliberate governmental purpose. The Eighth Circuit has explicitly required specific intent in analogous cases, and the Fourth Circuit’s analysis in cases addressing Haitian detention has followed the J-E- structure closely. None of these circuits has created a significant split on the core doctrine, and there is no circuit in which a practitioner would have a meaningful structural advantage in litigating a detention-based CAT claim over the BIA’s framework.

G. D.C. Circuit

The D.C. Circuit rarely hears direct immigration petitions for review, and its CAT jurisprudence in the detention context is thin. Where issues have arisen peripherally, the court has not deviated from the approach of other circuits. The D.C. Circuit’s role in shaping this area of law is likely to remain limited unless a constitutional or statutory challenge to the regulatory framework itself were to be filed in that court.

Overall, the BIA’s restrictive interpretation rests on extremely solid nationwide doctrinal ground. No circuit has rejected the specific-intent requirement, no circuit has held that willful blindness or recklessness suffices, and no circuit has categorically exempted any country’s prison system from J-E-’s analysis. The expert-impartiality rule from D-J-L- and the chain-of-events emphasis from A-A-R- are procedural developments that circuits are likely to treat as legitimate agency evidentiary practice, entitled to substantial deference as exercises of the BIA’s expertise in managing its own adjudicatory record.

Implications for Immigration Law and Practice

The quartet of decisions from J-E- through D-J-L- has reshaped the litigation landscape for CAT deferral in ways that are now fully visible. The implications operate at multiple levels: evidentiary strategy, expert witness selection, IJ adjudication, appellate practice, and systemic case selection.

At the evidentiary level, practitioners must now present particularized proof of specific intent that goes well beyond country conditions reports, statistical evidence of abuse, or generalized expert testimony about the treatment of criminal deportees. Country reports establish that detention conditions are deplorable and that some deportees experience severe mistreatment. They do not establish that the government maintains those conditions in order to produce that mistreatment. The gap between those two propositions is the gap between what country evidence typically establishes and what the specific-intent requirement demands. Closing that gap requires affirmative evidence of governmental purpose: statements by officials, policies expressly targeting deportees for abuse, documented patterns of systematic and deliberate mistreatment that cannot be explained by resource constraints or management failures.

The J-F-F- chain-of-events requirement adds a further evidentiary burden that is easily underestimated. Every inferential step that is necessary to the conclusion that the applicant is more likely than not to be tortured must independently be established at the more-likely-than-not level. In a compound claim—detention is likely; conditions in detention are life-threatening; the applicant’s particular vulnerability makes those conditions especially dangerous; the government is aware of that vulnerability and does nothing—each link in the chain must be separately established. Practitioners should audit every link before trial and ensure that the record independently supports each one.

Expert witnesses now present particular strategic challenges after D-J-L-. The decision creates a tension between the two categories of experts most readily available to applicants in Haitian and Salvadoran detention cases: academics and researchers who have deep familiarity with country conditions but may lack the granular knowledge of specific detention facilities and individual deportee experiences that specific-intent analysis requires; and advocates and NGO workers who have exactly that granular knowledge but whose institutional affiliations create impartiality exposure. Practitioners will need to anticipate impartiality challenges at the expert disclosure stage, prepare witnesses to address the question of whether they have considered contrary evidence, and consider whether disinterested academic experts can be retained to provide the testimony that advocacy-based experts might otherwise supply.

Immigration Judges face heightened appellate scrutiny following A-A-R-’s clarification of the standard of review. The de novo review of the legal characterization of torture means that an IJ’s grant of CAT deferral will receive no legal deference from the BIA. The BIA will independently evaluate whether the conditions that the IJ found likely to befall the respondent satisfy the regulatory definition. An IJ who grants CAT deferral must therefore do more than find that the respondent is likely to experience severe suffering; the IJ must also articulate why that suffering, given its source, reflects specific governmental intent rather than the general-intent harms that J-E- excludes. That legal analysis will be reviewed fresh, and IJs who do not provide it create an easily reversible record.

For respondents in Haitian and Salvadoran cases, the practical effect of the quartet is that CAT deferral based primarily on detention conditions, mental illness, or gang risks is now exceedingly rare. Cases that succeed will be those in which the record contains affirmative evidence of deliberate official targeting: documented decisions by specific officials to subject specific individuals to abuse; evidence that a respondent’s particular profile (criminal record, nationality, gang affiliation, physical markings) places them in a category that Haitian or Salvadoran authorities are known to single out for torture rather than merely for detention; or evidence that the conditions in a specific facility deviate from the general country-conditions pattern in a direction that reflects deliberate cruelty rather than resource constraints.

The decisions also align with broader enforcement priorities favoring the removal of criminal aliens even to countries with severe human rights records, provided that the harm they face arises from legitimate security measures and economic conditions rather than state-sponsored torture. That alignment between the doctrinal framework and current enforcement priorities suggests that the quartet will not be revisited administratively in the near term.

Conclusion

From J-E-’s 2002 foundation through the 2025–2026 trilogy, the BIA has constructed and maintained a coherent, restrictive interpretation of CAT that elevates specific intent and the lawful-sanctions exception above generalized humanitarian concerns about prison conditions worldwide. The framework has proven durable because its foundations are textual and structural rather than country-specific: the regulatory definition of torture, properly read, requires purposeful cruelty, and the Senate understanding attached to ratification confirms that reading. So long as the regulatory text remains unchanged, the core framework is insulated from case-by-case erosion.

The modern trilogy strengthens that insulation. A-A-R- forecloses the argument that J-E- is Haiti-specific and unavailable in other mass-detention contexts. S-S- forecloses the argument that mental illness creates a species of specific-intent vulnerability distinct from the general framework. D-J-L- forecloses the reliance on advocacy-based expert testimony as the primary vehicle for distinguishing a case from J-E-. Together, the three decisions close the principal doctrinal openings through which practitioners had been securing grants of CAT deferral.

The quartet’s expert-impartiality contribution deserves particular attention because it operates at an evidentiary level that is harder to challenge than a pure legal holding. An applicant can argue that the BIA’s legal framework misreads the regulatory text or misapplies the intent element. It is harder to argue that the BIA’s credibility evaluation of a particular witness was an abuse of discretion. As expert testimony is—or was—the primary means of translating country-conditions evidence into the individualized specific-intent analysis that J-E- requires, D-J-L-’s impartiality rule may prove to be the quartet’s most consequential practical contribution.

As country conditions in Haiti and El Salvador continue to deteriorate, these precedents ensure that CAT protection remains narrowly tailored to true governmental torture rather than to the incidental byproducts of poverty, misgovernment, or legitimate crime-control policy. Circuit courts will likely defer to the framework’s elaborations, further entrenching the line. The principal takeaway for practitioners, IJs, and appellate advocates is this: absent affirmative, individualized evidence that government officials maintained dangerous detention conditions as a vehicle for inflicting severe pain—rather than as a byproduct of poverty, bureaucratic indifference, or security policy—CAT deferral will not lie. That evidentiary burden was heavy in 2002. After D-J-L-, it is heavier still.

Primary sources

BIA and Attorney General Decisions

Federal Circuit Court Decisions

Regulations

EOIR Virtual Law Library