Limone v. United States

579 F.3d 79, 2009 U.S. App. LEXIS 19239, 2009 WL 2621536

• Court: Court of Appeals for the First Circuit
• Decided: August 27, 2009
• Docket: 08-1327, 08-1328
• Status: Published

Opinion

SELYA, Circuit Judge.

The genesis of these appeals can be traced to the gangland slaying of Edward “Teddy” Deegan, which occurred in 1965 in Chelsea, Massachusetts. Initially, the murder went unsolved. Two years later, agents of the Federal Bureau of Investigation (FBI), intent on frying bigger fish, cultivated a cooperating witness, Joseph Barboza, with tight ties to organized crime. Barboza thereafter met with state authorities and implicated several individuals in the Deegan slaying.

Based principally on Barboza’s testimony, the Commonwealth of Massachusetts secured indictments in 1967 and convictions the following year. Among those convicted were Peter Limone, Sr., Enrico Tameleo, Louis Greco, Sr., and Joseph Salvati (collectively, the scapegoats). All of them received stiff sentences.

Some three decades later, disturbing revelations cast grave doubt upon the verdicts. In December of 2000, the FBI for the first time disclosed that all along it had possessed reliable intelligence undercutting Barboza’s account of the murder and that it had suppressed this intelligence. By the time that this information came to light, Tameleo and Greco had died in prison, Salvati had been paroled, and Limone was still behind bars. In due course, the convictions of all four men were vacated and Limone was released.

Salvati, Limone, and the representatives of the estates of Tameleo and Greco, along with various family members (collectively, the plaintiffs), brought suit against the United States advancing claims under the Federal Tort Claims Act (FTCA), 28 U.S.C. §§ 1346(b), 2671-2680. Following a bench trial, the district court found the government liable on a multitude of theories and awarded over $100,000,000 in damages. The government appeals, as does one of the plaintiffs.

The record evinces egregious governmental misconduct; the FBI agents responsible for handling Barboza exhibited a callous disregard for the scapegoats’ rights. But it is our duty to interpret and apply the law even-handedly, regardless of the egregiousness of a defendant’s misconduct. Fidelity to that duty requires us to examine and resolve several vexing issues concerning both liability and damages. After painstaking consideration of the voluminous record, the parties’ briefs, and the district court’s carefully crafted re-scripts, we affirm the liability finding (albeit on grounds that differ in one significant respect from those relied upon by the district court).

The damage awards give us pause. Insofar as the awards embody damages for wrongful incarceration, they are considerably higher than any one of us, if sitting on the trial court bench, would have ordered. We nonetheless affirm those awards. Our proper function as appellate judges is not to second-guess the trial court but, rather, to apply a very deferential standard of review. Adhering to that role, and testing the disputed awards only to that extent, we conclude that the awards, though high, are not so grossly disproportionate to the harm sustained as to either shock our collective conscience or raise the specter of a miscarriage of justice.

I. BACKGROUND

These appeals have a long factual and procedural history. We rehearse that history only insofar as is necessary to place into perspective the issues that we must decide. We direct the reader who hungers for more detail to consult the district court’s capstone opinion in Limone v. United States (Limone IV), 497 F.Supp.2d 143 (D.Mass.2007).

We bifurcate our account. First, we limn the unsavory history of the Deegan murder and its aftermath. Then, we move to the commencement and travel of the federal case. Because these appeals follow findings made by a district court sitting without a jury, we resolve factual conflicts in favor of the district court’s findings (to the extent that those findings are not clearly erroneous). Jackson v. United States, 156 F.3d 230, 232-33 (1st Cir.1998).

A. The Murder and Its Aftermath.

On the night of March 12, 1965, Teddy Deegan’s bullet-ridden body was discovered in Chelsea, Massachusetts. Deegan had been shot six times, and the shots had been fired from three different guns. Suspicion focused upon a group of men that included Barboza, Jimmy Flemmi, Roy French, Joseph Martin, and Ronald Cassesso, all of whom were linked to organized crime. The group had been observed leaving a local gang hangout, the Ebb Tide Lounge, earlier that evening and returning shortly after the murder was committed. Eyewitnesses attested that they had seen blood stains on French’s clothing that night.

Despite local officers’ suspicions, the trail went cold within a matter of weeks. The police were unable to gather sufficient evidence to prefer charges against anyone.

Some two years later, FBI agents H. Paul Rico and Dennis Condon started cultivating Barboza, a known killer, in hopes of “flipping” him; that is, developing him as a cooperating witness against the Italian Mafia (La Cosa Nostra or LCN). At the time, Barboza was facing up to 89 years of imprisonment on state “habitual offender” charges. See Mass. Gen. Laws ch. 279, § 25. Barboza indicated a willingness to deal but placed one immutable condition on any information that he might provide: he would not inculpate his close associate, Flemmi. The FBI tacitly accepted that condition.

Throughout the spring of 1967, the agents used both carrots and sticks in their efforts to mine information from Barboza. Barboza was in state custody, and the agents plied him with promises of favorable recommendations and a slap-on-the-wrist sentence. They also fabricated a story that La Cosa Nostra was attempting, by influencing state prosecutors, to bring about Barboza’s lifetime confinement.

Barboza’s cooperation was not a one-shot affair. Over the course of several months of interrogation, he claimed to be knowledgeable about many crimes. Pertinently, he mentioned the Deegan murder (although in his conversations with the FBI agents he was not forthcoming as to any details). That crime was primarily a matter of state, not federal, interest. Accordingly, Massachusetts law enforcement officers sought to interview Barboza.

On September 8, 1967, two Suffolk County detectives (John Doyle and Frank Walsh) conversed with Barboza. Agents Rico and Condon were present, but the detectives pulled the laboring oar. Under questioning, Barboza finally provided his account of the Deegan killing. According to that account, Limone hired Barboza to murder Deegan because Deegan had robbed an LCN-affiliated bookmaker. Barboza then requested permission to carry out the “hit” from Tameleo, an LCN hierarch. After Tameleo’s blessing had been secured, Barboza and Greco formulated a plan.

According to Barboza, the mechanics of the plan were as follows. French would accompany Deegan to the site of a hypothetical burglary. Once there, French would turn on Deegan and, assisted by Barboza, Salvati, Greco, Martin, and Cassesso, would kill both Deegan and another putative participant in the burglary, Anthony Stathopoulos, Jr. Upon learning the details of the plan, Limone approved it and agreed to pay an additional sum because it involved a double murder.

During subsequent meetings with the detectives and the agents, Barboza modified his account. This modified version, which differed only at the margins and not at the core, formed the predicate for the indictments and convictions that followed.

At the time that Barboza unveiled his account of Deegan’s murder, the FBI possessed powerful intelligence casting grave doubt on the account’s veracity. Because the strength of this intelligence is of decretory significance here, we discuss it in some detail.

In the early 1960s, the FBI ramped up its efforts to extirpate organized crime in New England. Among other things, it surreptitiously installed an illegal electronic “bug” at the Providence, Rhode Island office of Raymond L.S. Patriarca, the reputed head of La Cosa Nostra in the area. The bug was in place from early 1962 through July 12, 1965. See United States v. Taglianetti, 274 F.Supp. 220, 223 (D.R.I. 1967). FBI agents transcribed the conversations that it recorded, reviewed those transcripts, and sent summaries of important information to FBI headquarters in Washington.

As a parallel measure, the FBI initiated the Top Echelon Criminal Informant Program in 1961. The aim of that program was to induce high-ranking organized crime figures to provide intelligence on a continuing, long-term basis. See United States v. Flemmi 225 F.3d 78, 81 (1st Cir.2000).

In the weeks preceding and following Deegan’s murder, these two investigative tools yielded a golconda of information about the killers. The Patriarca bug revealed that Barboza and Flemmi had approached Patriarca and obtained his sanction for the hit. Other information from both the bug and the FBI’s then-current crop of Top Echelon informants implicated five malefactors (Barboza, Flemmi, French, Martin, and Cassesso) in the murder, but not a single source other than Barboza so much as hinted that any of the scapegoats were involved.

Despite possessing credible intelligence undermining Barboza’s tale, the FBI did not turn over this intelligence to state authorities either at the time of the murder or during Barboza’s later debriefing. To make a bad situation worse, agents Rico and Condon informed the state prosecutor, Jack I. Zalkind, that Barboza’s tale “checked out.” Condon appeared as a witness at the state-court trial and lent credence to Barboza’s narrative by emphasizing that he (Condon) always was concerned about the “purity” of the testimony given by his informants.

On the strength of Barboza’s false testimony, the jury convicted the scapegoats on first-degree murder and murder-conspiracy charges. 1 The trial judge sentenced Limone, Tameleo, and Greco to death, and sentenced Salvati to life imprisonment. The scapegoats’ appeals were unsuccessful, but the capital sentences were commuted to life imprisonment after the United States Supreme Court decided Furman v. Georgia, 408 U.S. 238, 92 S.Ct. 2726, 33 L.Ed.2d 346 (1972).

Under Massachusetts law as it stood at the time, a state prisoner convicted of first-degree murder could not be paroled absent commutation by the governor. See Limone IV, 497 F.Supp.2d at 199 (describing parole process). The scapegoats filed numerous petitions for commutation and parole over many years. Although state authorities requested all relevant information, the FBI never turned over the exculpatory information that reposed within its files. On some occasions, the FBI went so far as to forward information that harmed the scapegoats’ chances for commutation or parole. On other occasions, it took affirmative steps to deflect possible challenges to the convictions.

Tameleo and Greco died in prison in 1985 and in 1995, respectively. Salvati secured a commutation from the governor and was released in 1997. Limone remained incarcerated.

In December of 2000, Special Assistant United States Attorney John Durham, responding to a request lodged by Limone, turned over five memoranda (which have come to be known as the Durham documents). The Durham documents revealed much of the exculpatory information that the FBI had kept in its organizational bosom all along. Upon seeing this information, the Suffolk County district attorney’s office moved to vacate Limone’s conviction. The state trial court granted that motion on the ground that the result of the trial in all likelihood would have been different had the Durham documents been disclosed in a timeous manner. Salvati’s conviction was vacated on the same ground. Shortly thereafter, state prosecutors filed notices of abandonment of prosecution (nolle prosequi) for both Limone and Salvati based on a perceived lack of evidence. The prosecutors later arranged for posthumous vacatur of Tameleo’s and Greco’s convictions and issued similar nolle prosequi notices in those cases.

B. The FTCA Action.

Following the release of the Durham documents, the plaintiffs filed separate actions in the federal district court. The plaintiffs’ complaints adumbrated three types of causes of action: (i) claims against the United States under the FTCA; (ii) Bivens claims against Rico, Condon, and other FBI supernumeraries, see Bivens v. Six Unknown Named Agents of FBN, 403 U.S. 388, 397, 91 S.Ct. 1999, 29 L.Ed.2d 619 (1971); and (iii) claims against Walsh and other state actors pursuant to 42 U.S.C. § 1983. The cases were consolidated.

The United States moved to dismiss on the ground that the claims against it were barred by the discretionary function and intentional tort exceptions to the FTCA. See 28 U.S.C. § 2680(a), (h). The individual defendants moved to dismiss on qualified immunity grounds. The district court denied these motions. See Limone v. United States (Limone I), 271 F.Supp.2d 345, 353-57 (D.Mass.2003) (rejecting discretionary function defense); id. at 365-66 (rejecting qualified immunity defense); Limone v. United States (Limone III), 336 F.Supp.2d 18, 30-31 (D.Mass.2004) (rejecting intentional tort defense). On a limited interlocutory appeal, we affirmed the denial of qualified immunity. Limone v. Condon (Limone II), 372 F.3d 39, 50 (1st Cir.2004). The plaintiffs later dropped their claims against the individual defendants (federal and state) and proceeded only on the FTCA claims.

Following a 22-day bench trial, the district court found for the plaintiffs (including the family members) on their claims of malicious prosecution, coercive civil conspiracy, intentional infliction of emotional distress, negligence, negligent supervision, and loss of consortium. The court awarded total damages in excess of $100,000,000. These timely appeals ensued.

II. ANALYSIS

On appeal, the government challenges the district court’s holdings on both liability and damages. Its argument on liability makes three basic points: (i) that the district court lacked subject matter jurisdiction over the plaintiffs’ claims by reason of the FTCA’s discretionary function exception; (ii) that the court lacked such jurisdiction because the plaintiffs’ claims arose out of intentional torts committed before Congress amended the FTCA to permit the maintenance of such claims; and (iii) that the plaintiffs adduced insufficient evidence to prove the elements of any of the asserted torts.

In a secondary line of attack, the government assails the district court’s damages calculus, arguing that the court’s rule-of-thumb baseline of $1,000,000 per year for each year of wrongful incarceration is unreasonable and resulted in a battery of excessive awards. In a cross-appeal, the son of one of the scapegoats contends that the district court erred in awarding him the same damages as were awarded to the offspring of the other scapegoats.

In the pages that follow, we address these arguments. For ease in exposition we deal with the second and third prongs of the government’s challenge to liability before addressing the first prong. 2

We start that discussion with what all the parties have treated as the leading edge of the government’s appeal: its asseveration that the district court erred in finding the government liable for malicious prosecution. While we conclude that the government’s asseveration has merit, that proves to be a hollow victory. The next most bruited theory on which the district court premised liability- — -its finding that the FBI is liable for intentional infliction of emotional distress — -withstands scrutiny. As to that cause of action, we also reject the government’s insufficiency of the evidence and discretionary function defenses. Finally, we uphold the damage awards in their entirety (a decision that entails, among other things, defenestration of the cross-appeal).

A. Malicious Prosecution.

Federal courts lack jurisdiction over tort actions against the United States except insofar as the sovereign has consented to be sued. See Dynamic Image Techs., Inc. v. United States, 221 F.3d 34, 39 (1st Cir.2000). The FTCA represents a general waiver of federal sovereign immunity for tortious acts and omissions of federal employees. But that general waiver is subject to a litany of exceptions.

At the time that the scapegoats were charged, prosecuted, and convicted in state court, the FTCA’s waiver provisions excluded claims arising out of malicious prosecution. See Savage v. United States, 322 F.Supp. 33, 35 (D.Minn.1971) (quoting former section 2680(h)); Ira S. Bushey & Sons, Inc. v. United States, 276 F.Supp. 518, 526 (E.D.N.Y.1967) (same). Congress dissolved the malicious prosecution bar with respect to federal law enforcement officers in 1974, amending the FTCA to add what has come to be known as the law enforcement proviso. See Pub.L. No. 93-253, § 2, 88 Stat. 50, codified at 28 U.S.C. § 2680(h).

The government posits that the plaintiffs’ malicious prosecution claims arose before the enactment of the law enforcement proviso and, therefore, the district court lacked jurisdiction over those claims. The district court brushed aside this argument, holding that the plaintiffs’ malicious prosecution claims did not arise until the scapegoats had received favorable terminations of the criminal charges wrongfully brought against them (an eventuality that did not occur until 2001 at the earliest). See Limone IV, 497 F.Supp.2d at 204; Limone III, 336 F.Supp.2d at 30-37; see also Heck v. Humphrey, 512 U.S. 477, 489, 114 S.Ct. 2364, 129 L.Ed.2d 383 (1994) (indicating that cause of action for malicious prosecution does not accrue until favorable termination of the underlying criminal proceeding has occurred).

The key to this riddle lies in language. The law enforcement proviso applies only to covered actions (like malicious prosecution) that “aris[e]” after its effective date. 28 U.S.C. § 2680(h). The plaintiffs and the district court equate the word “arise” with the word “accrue,” thereby bringing into play the Heck analysis. The government, however, insists that the word “arise” refers to the time when the prosecution itself occurred. Each view has a patina of plausibility. The district court’s analysis makes the case for the plaintiffs, see Limone III, 336 F.Supp.2d at 30-37, and the government’s reading of the statute finds some purchase in the case law elsewhere, see, e.g., Liuzzo v. United States, 508 F.Supp. 923, 927-28 & n. 2 (E.D.Mich.1981).

This is a difficult question. Moreover, given the passage of time, it is unlikely to recur. Courts should take pains not to grapple needlessly with enigmatic questions. As we explain below, it is unnecessary for us to answer the statutory construction question posed by the parties in this case. 3

The FTCA looks to state law to flesh out the elements of particular torts. See 28 U.S.C. § 1346(b)(1); Bolduc v. United States, 402 F.3d 50, 56 (1st Cir.2005). Here, both the allegedly tortious conduct and the harm complained of occurred in Massachusetts. Massachusetts law, there fore, supplies the beacon by which we must steer.

To prevail on a malicious prosecution claim under Massachusetts law, a suit- or must prove that the defendant (i) instituted criminal proceedings (ii) with malice and (iii) without probable cause, and (iv) that the proceedings were terminated in the accused’s favor. Correllas v. Viveiros, 410 Mass. 314, 572 N.E.2d 7, 10 (1991). Here, we can start and stop with the first of these four elements.

In broad brush, an individual may be said to have instituted criminal proceedings against another if he caused those proceedings to be initiated. See Witham v. Gregory & Read Co., 243 Mass. 595, 137 N.E. 752, 752 (1923); Mason v. Jacot, 235 Mass. 521, 127 N.E. 331, 333 (1920); Tangney v. Sullivan, 163 Mass. 166, 39 N.E. 799, 799-800 (1895). The paradigmatic example exists when a person formally swears out a criminal complaint against another person. See, e.g., White v. Apsley Rubber Co., 194 Mass. 97, 80 N.E. 500, 501 (1907). But malicious prosecution is by no means restricted to this paradigm.

If an individual induces another person (say, a police officer or prosecutor) to lodge formal criminal charges, he may be held to have instituted the criminal proceedings. See, e.g., Jones v. Schein, 215 Mass. 586, 103 N.E. 57, 58 (1913); Tangney, 39 N.E. at 800. So, too, if an individual either exercises a peculiar degree of control over the charging official or adamantly presses that official to bring a criminal complaint, he may be held responsible for the institution of the prosecution. See, e.g., Seelig v. Harvard Coop. Soc’y, 355 Mass. 532, 246 N.E.2d 642, 646 (1969); Conway v. Smerling, 37 Mass.App.Ct. 1, 635 N.E.2d 268, 271 (1994).

These taxonomies are of scant solace to the plaintiffs. The FBI neither preferred charges against the scapegoats nor swore out a complaint against them. Moreover, there is not a shred of evidence that the FBI induced the state to pursue the murder case. The agents’ primary interest was in preserving the secrecy of their own sources. They did not' demand that state actors bring charges against the scapegoats; indeed, there is no evidence that they so much as suggested that such charges should be brought.

Finally, though the agents assisted Barboza in shoring up his false tale when inconsistencies came to light, that subsequent assistance does not support a conclusion that the FBI “encouraged” state actors to institute the Deegan prosecution. See Correllas, 572 N.E.2d at 10. Equally as important, that evidence does not support the district court’s findings, Limone IV, 497 F.Supp.2d at 210, that the FBI controlled state actors and that the Deegan prosecution was the functional equivalent of a federal prosecution. Those findings were clearly erroneous. See Benham v. Lenox Sav. Bank, 292 F.3d 46, 48 (1st Cir.2002).

The plaintiffs’ best argument is at the margins. There is case law in Massachusetts indicating that an individual who transmits untruthful information to an official with power to charge sometimes may be said to have instituted an ensuing criminal proceeding brought by that official. See, e.g., Ziemba v. Fo’cs’le, Inc., 19 Mass.App.Ct. 484, 475 N.E.2d 1223, 1226 (1985); Carroll v. Gillespie, 14 Mass.App.Ct. 12, 436 N.E.2d 431, 439 (1982); see also Petricca v. City of Gardner, 429 F.Supp.2d 216, 225 (D.Mass.2006). This does not mean, however, that every provider of false information, nor even every bad-faith provider of false information, may be said to have instituted an ensuing criminal proceeding.

The question of whether such an individual has instituted a criminal proceeding depends on the circumstances. The controlling precedent is the ruling of the Massachusetts Supreme Judicial Court (SJC) in Correllas, 572 N.E.2d at 10. Under that decision, an individual may not be held to have instituted criminal proceedings if he merely provides false information to law enforcement officials in response to these officials’ queries during an ongoing investigation. See id.; see also Councilman v. Alibris, Inc., 386 F.Supp.2d 5, 9 (D.Mass.2005). Instead, the information provider must take some initiative; that is, he must voluntarily reach out to law enforcement officials and cause them to commence a new line of inquiry. See Councilman, 386 F.Supp.2d at 9.

The court below concluded that the United States was responsible for instituting criminal proceedings against the scapegoats on the theory that the FBI, through Barboza, had become a bad-faith provider of false information. Limone IV, 497 F.Supp.2d at 207-13. This conclusion rests principally on two considerations. First, agents Rico and Condon recruited, vetted, and delivered Barboza to state authorities. Id. at 206. Second, they knew that Barboza was not being truthful when he implicated the scapegoats, yet they assisted Barboza in making his tale more believable and encouraged him to stand by that tale through incentives such as protection, promises of leniency, and financial rewards. Id. at 179-80, 211, 217-18.

The record leaves no doubt but that the agents conducted themselves deplorably. But we cannot agree that they can be said to have “instituted” the criminal proceedings that ensued against the scapegoats. The turning point is the SJC’s analysis in Correllas.

The record in this case makes pellucid that, during interrogations conducted exclusively by Rico and Condon, Barboza made only passing reference to the Deegan homicide. He neither mentioned the scapegoats nor offered any specific details about the murder or the murderers. For their part, the agents exhibited no particular interest in those subjects.

This void remained until state authorities began to take part in the questioning. Barboza first offered the account, that would form the basis of the prosecution during an interview conducted chiefly by Doyle and Walsh (Suffolk County detectives) on September 8, 1967. 4 That questioning came about as part of a larger state investigation into the Deegan murder. By the time of the September 8 interview, state investigators had visited Barboza on at least four occasions and had asked him point-blank for information pertaining to the Deegan murder. 5 Even though federal agents were present during these audiences, it seems to have been the detectives, not the FBI, who first brought up, and then kept digging into, the Deegan murder. See supra note 4.

So viewed, the record compels the conclusion that Barboza furnished the false information that led to the institution of the prosecution while under questioning by state agents investigating a state crime. There is no evidence that the FBI brought Barboza and state authorities together specifically so that he would talk on the Deegan murder; indeed, Barboza offered information to state authorities on various matters, including another murder he would later testify about, before he answered questions regarding Deegan. Moreover, there is no evidence that the FBI knew that Barboza would concoct the bogus story that emerged. See Limone IV, 497 F.Supp.2d at 217 (finding that Barboza had pulled “the [scapegoats’] names out of thin air”). Although there is evidence that the FBI helped to shore up Barboza’s credibility as matters moved along, the prosecution by then already had been instituted.

There is one loose end. The SJC has left open the possibility that an individual may be held liable for malicious prosecution if he pursues a prosecution after it has become clear to him that there is no probable cause to support it. See Gutiérrez v. MBTA, 437 Mass. 396, 772 N.E.2d 552, 562 (Mass.2002); see also Mitchell v. City of Boston, 130 F.Supp.2d 201, 215 (D.Mass. 2001) (quoting Restatement (Second) of Torts § 655 (1977)). Given the SJC’s holding in Correllas, however, it is evident that the mere provision of false information cannot alone ground a malicious continuation finding. More is required, such as an insistence that the prosecution go forward even after it has become clear that probable cause is lacking. See, e.g., Miller v. City of Boston, 297 F.Supp.2d 361, 367 (D.Mass.2003); Restatement (Second) of Torts § 655 cmt. c (1977).

Here, there is no evidence that FBI agents urged state authorities to continue the ill-starred prosecution. Patently, the agents’ primary interest was in Barboza, not in prosecuting the scapegoats. The mere fact that the agents propped up the state’s case {e.g., by bolstering Barboza’s credibility) does not make the FBI a “continuer” of the prosecution any more than the defendant in Correllas (who gratuitously offered false data to the authorities subsequent to the initiation of prosecution).

To sum up, we hold that the FBI did not institute criminal proceedings against the scapegoats. See Correllas, 572 N.E.2d at 10. Thus, the plaintiffs have failed to prove the first element of the tort of malicious prosecution. Consequently, the district court erred in holding the United States liable for that tort. Nevertheless, this conclusion does not end our odyssey. The district court also found the government liable on other theories. Hence, we proceed to the most promising of those parallel theories of liability.

B. Intentional Inñiction of Emotional Distress.

Under Massachusetts law, an individual is liable for intentional infliction of emotional distress when he, “by extreme and outrageous conduct and without privilege, causes severe emotional distress to another.” Agis v. Howard Johnson Co., 371 Mass. 140, 355 N.E.2d 315, 318 (1976). 6 The court below found that the FBI’s extreme and outrageous misconduct had caused the plaintiffs to suffer severe emotional distress and, thus, imposed liability. Limone IV, 497 F.Supp.2d at 227.

The government does not contest that the tort of intentional infliction of emotional distress, unlike malicious prosecution, was actionable under the FTCA before the passage of the law enforcement proviso. Instead, it suggests that the district court lacked subject matter jurisdiction over the plaintiffs’ intentional infliction claims because those claims arise out of malicious prosecution (a tort that was barred by the version of the FTCA in effect at the time of the convictions). As a fallback, the United States questions the factual predicate underlying the district court’s analysis. We treat these arguments sequentially-

1. Subject Matter Jurisdiction. As said, Congress did not waive the federal government’s immunity from actions arising out of malicious prosecution until it enacted the law enforcement proviso in 1974. The government hypothesizes that the plaintiffs’ intentional infliction claims arise out of a maliciously instigated prosecution that occurred before the enactment of this proviso and, therefore, the district court lacked competence to adjudicate those claims. We do not accept that hypothesis.

Although courts (including this court) sometimes have referred loosely to section 2680(h) as an “intentional torts” exception to the general waiver contained in the FTCA, see, e.g., Rodríguez v. United States, 54 F.3d 41, 44 (1st Cir.1995), the provision only preserves the federal government’s immunity with respect to claims arising out of certain enumerated torts, see Santiago-Ramírez v. Sec’y of Def., 984 F.2d 16, 20 (1st Cir.1993); Black v. Sheraton Corp., 564 F.2d 531, 539-40 (D.C.Cir. 1977). Because intentional infliction of emotional distress never has been on the roster of excluded torts listed in section 2680(h), intentional infliction claims are not per se barred by that provision. See Estate of Trentadue ex rel. Aguilar v. United States, 397 F.3d 840, 854-55 (10th Cir. 2005); Sabow v. United States, 93 F.3d 1445, 1457 (9th Cir.1996); Truman v. United States, 26 F.3d 592, 595 (5th Cir. 1994); Santiago-Ramírez, 984 F.2d at 20; Kohn v. United States, 680 F.2d 922, 926 (2d Cir.1982); Gross v. United States, 676 F.2d 295, 304 (8th Cir.1982).

Be that as it may, the reach of section 2680(h) is not limited to specifically enumerated torts. Rather, that provision deprives a district court of jurisdiction over a claim whenever the claim is, or arises out of, a specifically enumerated tort. See, e.g., Snow-Erlin v. United States, 470 F.3d 804, 808-09 (9th Cir.2006) (holding particular claim arose out of false imprisonment); O’Ferrell v. United States, 253 F.3d 1257, 1265-66 (11th Cir.2001) (holding particular claim arose out of slander). This framework applies to claims for intentional infliction of emotional distress. See, e.g., Metz v. United States, 788 F.2d 1528, 1534-35 (11th Cir.1986).

The approach that we have outlined necessitates a fact-sensitive, case-specific inquiry. In performing that tamisage, substance trumps form; an inquiring court must look past the nomenclature employed by the plaintiff and focus on the actual nature of the plaintiffs grievance. See Jiménez-Nieves v. United States, 682 F.2d 1, 6 (1st Cir.1982). If that grievance rests on proof of conduct that traditionally comprises an excepted tort, section 2680(h) precludes suit. See, e.g., Snow-Erlin, 470 F.3d at 808-09; Truman, 26 F.3d at 595; Thomas-Lazear v. FBI, 851 F.2d 1202, 1207 (9th Cir.1988).

On the other hand, if there is merely a loose connection, a family resemblance, or even a partial overlap between the conduct on which the asserted claim rests and that comprising an excepted tort, the claim is not barred by section 2680(h). See Block v. Neal, 460 U.S. 289, 298, 103 S.Ct. 1089, 75 L.Ed.2d 67 (1983). It follows that when an element of an excepted tort is missing from the factual scenario, the claim is not pretermitted. See Estate of Trentadue, 397 F.3d at 855 (holding intentional infliction claim not barred by misrepresentation exception because elements of misrepresentation, including reliance and pecuniary loss, were not present); Truman, 26 F.3d at 596 (holding intentional infliction claim not barred by assault or battery exceptions because elements of these latter torts were not alleged); Jiménez-Nieves, 682 F.2d at 4-5 (holding negligence claim not barred by misrepresentation exception because reliance not present).

In the instant case, the plaintiffs failed to prove that the FBI instituted criminal proceedings against the scapegoats, see supra Part 11(A), and thus failed to prove an essential element of the tort of malicious prosecution. Furthermore, the conduct undergirding the plaintiffs’ claims for intentional infliction of emotional distress is broader than that traditionally associated with the tort of malicious prosecution in that it includes malfeasance that postdates the scapegoats’ convictions, such as efforts by the FBI to cover up its misdeeds (a topic to which we shall return). And, finally, the plaintiffs’ intentional infliction claims require proof not only that the FBI’s conduct was something akin to malicious, but also that it was extreme and outrageous. Agis, 355 N.E.2d at 318. These are substantive distinctions. See Foley v. Polaroid Corp., 400 Mass. 82, 508 N.E.2d 72, 82 (1987).

We conclude that the conduct underlying the plaintiffs’ claims for intentional infliction of emotional distress neither comprises malicious prosecution nor arises out of malicious prosecution in the requisite sense. 7

The government labors to undercut this reasoning by noting that the plaintiffs pleaded claims of malicious prosecution arising out of essentially the same facts that supported their intentional infliction claims. The plaintiffs’ intentional infliction claims, they suggest, are barred by this characterization.

This suggestion is more cry than wool. The plaintiffs had the right to plead alternative theories of liability, see Fed.R.Civ.P. 8(d), and their exercise of that right did not debar them from an independent review of each set of claims. See Dedham Water Co. v. Cumberland Farms Dairy, Inc., 889 F.2d 1146, 1157-58 (1st Cir.1989).

In a related vein, the government posits that because the district court found that the same damages flowed from both the alleged malicious prosecution and the alleged intentional infliction of emotional distress, Limone IV, 497 F.Supp.2d at 245 & n. 208, the latter claims necessarily arise out of the former. This is sophistry, pure and simple. The proper inquiry focuses upon the actor’s tortious conduct, not the plaintiffs damages. See Truman, 26 F.3d at 595; Gross, 676 F.2d at 304; Black, 564 F.2d at 540-41; see also Rayonier Inc. v. United States, 352 U.S. 315, 320, 77 S.Ct. 374, 1 L.Ed.2d 354 (1957) (warning against “read[ing] exemptions into the [FTCA] beyond those provided by Congress”).

To say more on this point would be supererogatory. We hold that, section 2680(h) notwithstanding, the district court had subject matter jurisdiction to adjudicate the claims for intentional infliction of emotional distress.

2. The Merits. To make out a claim for intentional infliction of emotional distress under Massachusetts law, a claimant must prove:

(1) that the [defendant] intended to inflict emotional distress or that he knew or should have known that emotional distress was the likely result of his conduct; (2) that the conduct was extreme and outrageous, was beyond all possible bounds of decency and was utterly intolerable in a civilized community; (3) that the actions of the defendant were the cause of the plaintiffs distress; and (4) that the emotional distress sustained by the plaintiff was severe and of a nature that no reasonable man could be expected to endure it.

Agis, 355 N.E.2d at 318-19 (citations and internal quotation marks omitted). The court below determined that the plaintiffs had proven these four elements by a preponderance of the evidence. Limone IV, 497 F.Supp.2d at 227. In the court’s view, the FBI had participated willingly in framing the scapegoats, and then scrambled to cover up the frame job by obstructing the scapegoats’ efforts to clear their names. Id. The court found this conduct “intentional,” “outrageous,” “beyond all bounds of decency,” and to have “no place in a civilized community.” Id. The consequent emotional distress was “so severe and of such a nature that no reasonable person could be expected to endure it.” Id.

The government raises a host of record-based challenges to this series of findings. Because these appeals follow a bench trial, we review the lower court’s factfinding for clear error. Cumpiano v. Banco Santander P.R., 902 F.2d 148, 152 (1st Cir.1990); Fed.R.Civ.P. 52(a). Consonant with that standard, we will not “upset findings of fact or conclusions drawn therefrom unless, on the whole of the record, we form a strong, unyielding belief that a mistake has been made.” Cumpiano, 902 F.2d at 152.

As a threshold matter, the government questions whether we should use an unadulterated standard of dear-error review. It regards this standard as inapposite because the district court relied principally upon documentary evidence in constructing its findings and conclusions. We reject that argument out of hand.

In Anderson v. City of Bessemer City, the Supreme Court considered whether, under Rule 52(a), an appellate tribunal may review de novo findings of fact not bottomed on credibility determinations. 470 U.S. 564, 573-74, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985). The Court concluded that the clearly erroneous standard loses none of its vigor “even when the [lower] court’s findings do not rest on credibility determinations, but are based instead on physical or documentary evidence or inferences from other facts.” Id. at 574, 105 S.Ct. 1504. The application of clear-error review to findings drawn from a paper record has long been the practice in this circuit. See, e.g., Reliance Steel Prods. Co. v. Nat’l Fire Ins. Co., 880 F.2d 575, 576 (1st Cir.1989); Boroff v. Tully (In re Tully), 818 F.2d 106, 108-09 (1st Cir.1987). That is the practice to which we adhere today.

With the standard of review nailed down, we turn to the government’s multi-faceted critique of the trial court’s factfinding. To begin, the government takes umbrage with the court’s conclusion that the FBI’s misconduct was extreme and outrageous. But that conclusion seems rock-solid: it is premised on the court’s determination that FBI agents knowingly participated in the events leading to the wrongful indictment, prosecution, conviction, and continued incarceration of the scapegoats. Limone IV, 497 F.Supp.2d at 227. This determination rests on three building blocks: that the FBI (i) knew Barboza was dissembling when he implicated the scapegoats in Deegan’s murder; (ii) assisted Barboza in selling his lies to state authorities and encouraged him to stick to them; and (iii) covered up its misdeeds post-conviction by hindering the scapegoats’ efforts to obtain relief. The government challenges each of these building blocks.

As an initial matter, the district court’s determination that FBI agents knew that the scapegoats were strangers to the Deegan slaying is not a necessary prerequisite to its finding of extreme and outrageous conduct. The SJC has made pellucid that such a finding may be grounded either on actual knowledge or on a defendant’s deliberate disregard of a substantial probability that his actions will produce severe emotional distress. Simon v. Solomon, 385 Mass. 91, 431 N.E.2d 556, 561-62 (1982); see Restatement (Second) of Torts § 46 cmt. i (1965). Thus, the FBI may not hide behind an assertion that it remained (wilfully) blind to the scapegoats’ innocence. And though actual knowledge is not a necessary finding, the district court’s finding of actual knowledge is obviously sufficient.

Moreover, that finding is unassailable. The record contains adequate evidence from which a reasonable factfinder could conclude — as did the district court — that the FBI knew that the scapegoats were not involved in the slaying.

The district court painstakingly reviewed the intelligence in the FBI’s possession at the relevant time. See Limone IV, 497 Supp.2d at 172-77. In the days and months leading up to Deegan’s murder, the FBI learned from the Patriarca bug and from its Top Echelon informants that Barboza and Flemmi — not Limone or Tameleo — had requested permission to murder Deegan and that Patriarca (the head of the LCN) had blessed the hit. On the day following the murder, a highly reliable Top Echelon informant told agent Rico that Flemmi had bragged about killing Deegan with the help of French, Martin, Cassesso, and Barboza. This enumeration of the participants in the murder was repeated and confirmed through various informants and conversations picked up by the Patriarca bug. Much of this information was catalogued in memoranda, airtels, and correlator reports, 8 a number of which were either authored or initialed as read by Rico and Condon. Under these circumstances, one would have to believe in the tooth fairy to believe that the agents did not know the identities of the real killers.

Relatedly, Barboza’s accusation that Limone and Tameleo had orchestrated the hit did not jibe with information that the FBI had gleaned from the Patriarca bug. When requesting Patriarca’s permission to carry out the hit, Barboza and Flemmi made no mention of any prior authorization given by Limone and Tameleo. And, tellingly, the Patriarca bug showed that Limone, rather than ordering the hit, had tried to warn Deegan that violence might be in the offing. Rico and Condon turned a blind eye to these contradictions.

There is other evidence as well. When Barboza first spoke with Rico and Condon in March of 1967, he made it plain that he would not incriminate Flemmi. The FBI had reliable intelligence linking Flemmi to the Deegan murder; Barboza’s enumeration of a cast of participants that did not include Flemmi, especially when coupled with his avowed intent to shield Flemmi, surely should have convinced the agents that they were not getting the whole story.

What is more, a local police officer had seen a balding individual matching Flemmi’s description in the back seat of Martin’s car at around the time of the killing. When Barboza learned of the officer’s observations, he suddenly “remembered” that Salvati, whom he had previously placed in the back seat of Martin’s automobile, was wearing a bald wig. Given what the agents knew, this convenient improvisation was a red flag that should have triggered their suspicions.

Indeed, there is compelling reason to believe that FBI agents assented to the omission of Flemmi’s name in order to further their own agenda. After all, the FBI formally targeted him as a Top Echelon informant three days prior to Deegan’s murder and officially assigned him to Rico on the very day that Deegan was killed. An FBI memorandum written in June of that year rated the quality of Flemmi’s information as good. Even after the FBI dropped Flemmi from the ranks of its informants in September of 1965, it had a powerful incentive to keep him out of prison; his brother, Stephen, became a Top Echelon informant around that time and remained in that status for many years. See Flemmi, 225 F.3d at 80-82.

The FBI’s willingness to accept Barboza’s narrative at face value is especially troubling because Barboza exhibited a similar lack of consistency in his account of Greco’s supposed involvement in the crime. When law enforcement officers were unable to corroborate Barboza’s insistence that Greco was present at the Ebb Tide and had left with the rest of the men that Barboza had fingered, Barboza reversed his field and claimed to have “remembered” that Greco joined the others at a later time.

From this and other information in the record, it is transparently clear that the district court had a sturdy foundation for its finding that the FBI knew at the time that Barboza’s story was riddled with inconsistencies. The court was entitled to view that knowledge in light of a wealth of FBI intelligence indicating that the scapegoats were not part of the band of miscreants who carried out the murder plot. Although the evidence does not compel the conclusion that the FBI knew that the scapegoats were uninvolved, it is enough to ground a reasonable inference to that effect.

The government likewise challenges the district court’s determination that the FBI aided Barboza in framing the scapegoats. In this regard, the government argues that the FBI did no more than gift-wrap Barboza and hand him over to state authorities (who then made an independent decision to prosecute the scapegoats). The record supports the district court’s determination regarding the FBI’s culpability.

To be sure, there is no evidence that the FBI spoon-fed the scapegoats’ names to Barboza. For aught that appears, the fictional tale sprang directly from the informant’s brow. But there is evidence that, once Barboza gave the scapegoats’ names to the Suffolk County detectives, the FBI assisted him in doctoring his tale to make it seem more believable. For example, as we have said, Barboza modified his initial version of the facts to accommodate other information possessed by the authorities. The district court concluded that Barboza made these alterations because the FBI had made him aware of contradictory evidence in the hands of state officials. Limone IV, 497 F.Supp.2d at 179-80. 9

There was also evidence tending to show that the FBI helped to “sell” Barboza’s tale both to state authorities and to the jury. The prosecutor, Zalkind, testified that the FBI had told him that Barboza’s account “checked out.” Agent Condon testified at the murder trial, vouchsafing that he was careful not to impart any information about the murder investigation to Barboza because he (Condon) always was concerned about assuring the “purity” of testimony given by his informants. But the district court rejected this testimony and supportably found that Condon knew at the time that Barboza was spinning a yarn about the scapegoats’ involvement in the murder. Id. at 186.

Equally as important, the FBI interposed no disincentives that might have deterred Barboza from standing by his bogus story. The Bureau continued to coddle Barboza. Among other things, the FBI pledged to bring Barboza’s cooperation to the attention of relevant authorities, extended protection to him and his family, and promised to give him money and a fresh start on the other side of the continent. The “habitual offender” charges that Barboza was facing when he first began cooperating were dropped, and Barboza received only a one-year sentence for his role in Deegan’s murder.

The government attempts to absolve itself of responsibility for the scapegoats’ plight by piously asserting that the FBI turned over all relevant information to state authorities. In this regard, it points to three memoranda. The first is a memorandum from the FBI director dated March 16,1965, which instructs the Boston office to disclose to local authorities information pertaining to the Deegan murder to the extent that divulgement is consistent with the complete security of the Patriarca bug. A handwritten notation on that memorandum indicates that full disclosure already had taken place. The second, a memorandum authored by Rico on March 15, 1965, indicates that a Top Echelon informant had told him that Flemmi, French, Martin, Cassesso, and Barboza had murdered Deegan, and states that this information had been transmitted to local authorities. The third is a memorandum written on March 24 of the same year by the special agent in charge of the FBI’s Boston office; that memorandum essentially replicates the Rico memorandum.

On their face, these memoranda bolster the government’s argument. But the memoranda do not exist in a vacuum. The district court found that, to the extent the FBI did volunteer information to state authorities in 1965, that information was general in nature and already within the state’s ken. See id. at 174-75 & n. 73. This finding was not clearly erroneous. The local police had Flemmi, French, Martin, Cassesso, and Barboza in their sights from the earliest stages of their investigation, and the three FBI memoranda, whether read separately or in the ensemble, do not suggest that the reliability of the FBI’s sources was communicated to the state. To the contrary, the director’s memorandum instructed that the Patriarca bug remain secret.

The gaps in the record are also telling. For instance, there is absolutely no evidence indicating that the Boston-based FBI agents, whatever instructions they may have received, actually divulged any information to state officials at the time of the killing in March 1965 or during Barboza’s debriefing in 1967 and 1968. In fact the state prosecutor, Zalkind, testified in the district court that he had not seen any of the FBI documents containing exculpatory evidence when he forged ahead with the prosecution. He also denied that this evidence had been communicated to him in any other form. Given Zalkind’s testimony, we cannot set aside the district court’s finding that the FBI agents failed to provide relevant exculpatory information. See, e.g., United States v. Natanel, 938 F.2d 302, 313 (1st Cir.1991) (emphasizing that “the district court must be given wide rein to assess the evidence and judge the credibility of witnesses”).

The district court’s finding that the FBI covered up its perfidy by stonewalling the scapegoats’ post-conviction efforts to win their freedom is equally unimpugnable. The government argues that the alleged coverup consisted of nothing more than a failure to provide state officials with exculpatory materials and that such a failure cannot ground a claim under the FTCA. See Bolduc, 402 F.3d at 59 (holding that negligent failure to disclose Brady materials is not actionable); see also Brady v. Maryland, 373 U.S. 83, 87, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963). This argument veers well wide of the mark. Here, the government’s post-conviction misconduct consisted of more than failing to turn over exculpatory materials to state authorities; the government took positive steps to ensure that the scapegoats remained behind bars.

For example, in mid-1970 Barboza, represented by new counsel, signed a sworn affidavit in which he recanted certain portions of his trial testimony relating to the scapegoats’ guilt. His lawyer then requested- permission to have a lie detector test administered. Shortly thereafter, two federal prosecutors visited Barboza, who retracted his recantation. Despite the obvious importance of these developments, the FBI agents failed either to conduct an investigation into the recantation or to brief their state counterparts about it.

Other examples abound. Among other things, the FBI told state authorities who were considering petitions for commutation and/or parole that Limone, Greco, and Salvati had continuing ties to organized crime. The FBI even went so far as to have agents visit the office of a parole board member to voice opposition to Limone’s petition for commutation. Moreover, the record makes manifest that the court below regarded the Brady violation (that is, the FBI’s failure to disclose exculpatory information in a timely manner) as “part of a broader scheme to put Barboza forward as a witness no matter the cost, even if it meant framing the plaintiffs.” Limone IV, 497 F.Supp.2d at 222. The government’s after-the-fact attempt to conceal what it had done became part of the same scheme. Id. at 202.

The SJC has made it abundantly clear that claims for intentional infliction of emotional distress may be founded on a pattern of misconduct. See, e.g., Boyle v. Wenk, 378 Mass. 592, 392 N.E.2d 1053, 1055 (1979). Thus, it was appropriate for the district court, on a pattern of conduct theory, to weigh the significance of the FBI’s failure to provide state authorities with exculpatory evidence. See, e.g., Burrell v. Adkins, No. 01-2679, 2007 WL 4699166, at *18 (W.D.La. Oct.22, 2007).

The evidence supports the district court’s finding that a coverup occurred. Despite contemporaneous requests by state officials for information bearing upon the scapegoats’ petitions for post-conviction relief, the FBI remained mute — and worse. That recalcitrance is especially damning in the circumstances of this case — a case in which the FBI’s deliberate misconduct had placed the scapegoats in harm’s way. See Commonwealth v. Levesque, 436 Mass. 443, 766 N.E.2d 50, 56 (2002) (explaining that “a duty to prevent harm to others arises when one creates a dangerous situation, whether that situation was created intentionally or negligently”); Restatement (Second) of Torts § 321 (1965) (similar).

To recapitulate, the district court supportably determined that the FBI knew that the scapegoats were uninvolved in the Deegan murder from the moment that Barboza implicated them. The FBI agents nonetheless assisted Barboza in embellishing his apocryphal tale, helped him to sell that tale to state authorities and the jury, and covered up their perfidy by stonewalling the scapegoats’ petitions for post-conviction relief. The district court concluded that this pattern of conduct was extreme and outrageous, Limone IV, 497 F.Supp.2d at 227, and we may upset that conclusion only if reasonable minds would be compelled to reach the opposite conclusion, see Jackson, 156 F.3d at 232-33; see also Boyle, 392 N.E.2d at 1056-57. Applying that standard, the conclusion that the government indulged in extreme and outrageous conduct must stand. See, e.g., Pitt v. Dist. of Columbia, 491 F.3d 494, 506 (D.C.Cir.2007); Wagenmann v. Adams, 829 F.2d 196, 214 (1st Cir.1987); Newton v. City of New York, 566 F.Supp.2d 256, 281 (S.D.N.Y.2008); Harris v. Harvin, No. 01-2292, 2005 WL 2461876, at *2 (Mass.Super.Ct. Aug. 4, 2005); Sarvis v. Boston Safe Deposit & Trust Co., No. 94-1215, 1994 WL 879797, at *3 (Mass.Super. Ct. June 6,1994).

We need not linger long over the finding of intentionality. Limone IV, 497 F.Supp.2d at 227. Common sense suggests that the FBI’s deliberate acts and omissions were likely to lead to the wrongful conviction and incarceration of the scapegoats (and, thus, the consequent emotional distress). That is exactly what transpired. Accordingly, the district court had an ample predicate from which to infer that the FBI knew that its misconduct was likely to cause emotional distress. See, e.g., Wagenmann, 829 F.2d at 214.

In a variation on this theme, the government contends that it cannot be held legally responsible for causing the emotional distress that occurred here. All that the FBI did, it reasons, was to hand a witness to state authorities, who then exercised their independent discretion in bringing that witness before a grand jury and a petit jury. In the government’s view, the state’s decision to pursue the murder case and the trial jurors’ decision to convict are intervening acts that broke the causal chain.

Causation is a factbound issue and, as such, is normally left to the trier. Peckham v. Cont'l Cas. Ins. Co., 895 F.2d 830, 837 (1st Cir.1990); Mullins v. Pine Manor Coll., 389 Mass. 47, 449 N.E.2d 331, 338 (1983). The causation inquiry has two components: proof that the harm would not have occurred but for the defendant’s misconduct, see Glidden v. Maglio, 430 Mass. 694, 722 N.E.2d 971, 974-75 (2000), and proof that the defendant was a proximate cause of the harm, see Kent v. Commonwealth, 437 Mass. 312, 771 N.E.2d 770, 777 (2002). These two components may be thought of as causation in fact and legal causation.

In this instance, the district court’s causation-in-fact analysis, Limone IV, 497 F.Supp.2d at 227, is ironclad. Given the aid and encouragement that the FBI afforded Barboza and its exclusive possession of exculpatory evidence that probably would have rescued the scapegoats from wrongful conviction, the finding that the FBI’s misconduct constituted a but-for cause of the scapegoats’ plight is fully sustainable. See Burke v. McDonald, 572 F.3d 51, 58 (1st Cir.2009) (ruling that police officer could be said to have “caused” plaintiffs confinement without bail if jury found that he suppressed exonerating DNA evidence).

The second component of the causation inquiry requires elaboration. Under Massachusetts law, proximate cause turns largely on the foreseeability of the harm. 10 See Wagenmann, 829 F.2d at 214; Kent, 771 N.E.2d at 777. Intervening acts of a third party will not break the causal chain if those acts were reasonably foreseeable. Copithorne v. Framingham Union Hosp., 401 Mass. 860, 520 N.E.2d 139, 142-43 (1988); Gidwani v. Wasserman, 373 Mass. 162, 365 N.E.2d 827, 830-31 (1977).

In the case at bar, the district court concluded that the indictment, prosecution, conviction, and incarceration were all reasonably foreseeable results of the FBI’s misconduct. Limone IV, 497 F.Supp.2d at 227. That conclusion strikes us as virtually inescapable. The so-called “intervening acts” — notably, the state’s decision to prosecute and the jury’s decision to convict— were well within the realm of reasonable expectations once the government took steps to prop up Barboza’s credibility and conceal exculpatory evidence. See Burke, 572 F.3d at 58-61.

This leaves the fourth and final element of the tort: that the scapegoats suffered severe emotional distress. The government does not challenge the district court’s finding to this effect, nor could it credibly do so. On this record, it is unarguable that the wrongful indictment, prosecution, conviction, and incarceration caused the victims severe emotional distress. Therefore, the plaintiffs satisfied their burden of proving all the elements of their claims for intentional infliction of emotional distress. 11

C. The Discretionary Function Exception.

Regardless of the invulnerability of the district court’s findings on the elements of the intentional infliction claims, we must address yet another issue bearing upon liability. The government invokes the FTCA’s discretionary function exception as a further defense. It argues, in essence, that the conduct on which the intentional infliction claims is based involves the performance or failure to perform discretionary functions on the part of government actors. See 28 U.S.C. § 2680(a); see also Irving v. United States, 162 F.3d 154, 162 (1st Cir.1998) (en banc).

We afford de novo review to a district court’s determination that the discretionary function exception does or does not apply. Fothergill v. United States, 566 F.3d 248, 251 (1st Cir.2009); Irving, 162 F.3d at 162. We start by identifying the particular conduct giving rise to the claims at issue. See Fothergill, 566 F.3d at 252-53; Muñiz-Rivera v. United States, 326 F.3d 8, 15 (1st Cir.2003). Here, that conduct consists of assisting Barboza to frame the scapegoats for a capital crime and covering up the frame job by withholding exculpatory information from state officials.

Having identified the conduct at issue, we move to a binary inquiry designed to reveal whether Congress sought to shield that conduct from liability. Bolduc, 402 F.3d at 60. This inquiry seeks to ascertain, first, if the conduct “involves an element of judgment or choice” for the actor. Berkovitz v. United States, 486 U.S. 531, 536, 108 S.Ct. 1954, 100 L.Ed.2d 531 (1988). Then, so long as the conduct involved a matter of judgment or choice— that is, so long as it was discretionary in nature — the inquiry seeks to ascertain whether that judgment or choice was susceptible to policy-related analysis. Id. at 536-37, 108 S.Ct. 1954.

It is elementary that the discretionary function exception does not immunize the government from liability for actions proscribed by federal statute or regulation. Bolduc, 402 F.3d at 60. Nor does it shield conduct that transgresses the Constitution. See Castro v. United States, 560 F.3d 381, 389 (5th Cir.2009) (collecting cases); Thames Shipyard & Repair Co. v. United States, 350 F.3d 247, 254-55 (1st Cir.2003) (same). The district court determined that the FBI’s conduct in this case violated the Constitution as well as Department of Justice guidelines. Limone IV, 497 F.Supp.2d at 203-04.

The government demurs. It insists that decisions concerning the conduct and course of law enforcement investigations, including decisions as to whether and how informants should be employed, are generally discretionary. See, e.g., Kelly v. United States, 924 F.2d 355, 362 (1st Cir.1991); Pooler v. United States, 787 F.2d 868, 871 (3d Cir.1986). Relatedly, the government argues that it possessed discretion to withhold exculpatory information from state prosecutors in order to protect the security of its sources. See, e.g., Ga. Cas. & Sur. Co. v. United States, 823 F.2d 260, 262-63 (8th Cir.1987); see also Taglianetti v. United States, 398 F.2d 558, 572 (1st Cir. 1968) (recognizing that government possesses “substantial interest in preserving the secrecy of its investigation”). Given these foundational propositions, the government exhorts us to conclude that the conduct upon which the plaintiffs’ intentional infliction claims rests must perforce be discretionary.

This exhortation operates at too high a level of generality. Viewed from 50,000 feet, virtually any action can be characterized as discretionary. But the discretionary function exception requires that an inquiring court focus on the specific conduct at issue. See Berkovitz, 486 U.S. at 546-47, 108 S.Ct. 1954; Trevino v. Gen. Dynamics Corp., 865 F.2d 1474, 1484 (5th Cir.1989). Here, when the FBI’s conduct is examined in context, warts and all, any illusion that the conduct was discretionary is quickly dispelled.

To use a phrase popularly attributed to Lawrence ‘Yogi” Berra, much of this is déjá vu all over again. 12 In Limone II, we held that the plaintiffs’ allegations that FBI agents had participated in framing them and had withheld exculpatory evidence to cover up their malefactions stated a clear violation of due process. 372 F.3d at 44-50. The plaintiffs proved the substance of these allegations. See Limone IV, 497 F.Supp.2d at 227. Consequently, the conduct was unconstitutional and, therefore, not within the sweep of the discretionary function exception. 13

This holding ends our discussion of liability. We conclude that the district court possessed subject matter jurisdiction over the plaintiffs’ claims for intentional infliction of emotional distress and that the plaintiffs proved those claims. Because the district court determined that the same damages flowed from all of the torts alleged, we need not inquire whether the plaintiffs also proved their claims of conspiracy, negligence, and/or negligent supervision. Only questions pertaining to damages remain.

D. Damages.

The district court made the damage awards listed in the appendix to this opinion. These awards total $101,750,000. The government argues that the court’s approach was wrongheaded and that the assessed damages are excessive. In a cross-appeal Greco’s son, Edward, contends that the court awarded him insufficient damages.

The weight of authority indicates that damage awards under the FTCA are subject to dear-error review. See, e.g., Davis v. United States, 375 F.3d 590, 591 (7th Cir.2004); Lebron v. United States, 279 F.3d 321, 325 (5th Cir.2002); Duplan v. Harper, 188 F.3d 1195, 1202 (10th Cir.1999); Whitley v. United States, 170 F.3d 1061, 1079 (11th Cir.1999); Bartleson v. United States, 96 F.3d 1270, 1274 (9th Cir.1996). This court has lent its voice to that chorus. See Soto v. United States, 11 F.3d 15, 18 (1st Cir.1993). But this description oversimplifies the matter. The standard of review, at least insofar as it pertains to awards of non-economic damages, is more nuanced. See Neyer v. United States, 845 F.2d 641, 644-45 (6th Cir. 1988).

In that context, the appropriate standard of review actually has three facets. Raw findings of fact are reviewed for clear error. Reilly v. United States, 863 F.2d 149, 166 (1st Cir.1988). Claimed errors of law engender de novo review. Soto, 11 F.3d at 17. The third facet of the standard of review relates to matters of judgment, which are reviewed for abuse of discretion. See Davis, 375 F.3d at 592 (using language consistent with abuse of discretion standard); Soto, 11 F.3d at 18 (similar).

These differentiated aspects of the standard of review are designed to operate in a synchronized fashion. Thus, in an FTCA case that involves non-economic damages, an appellate court reviews facts found by the trial judge (such as the existence and nature of the harm suffered) for clear error. See Doe v. United States, 976 F.2d 1071, 1083 (7th Cir.1992); Reilly, 863 F.2d at 166. At the same time, the court assays the reasonableness of the trial judge’s monetization of that harm — a classic example of a judgment call — under an abuse of discretion standard. See Wilkinson v. United States, 564 F.3d 927, 934 (8th Cir.2009). Within that rubric, the court evaluates the trial judge’s assumptions as to purely legal matters de novo. Cf. Rosario-Urdaz v. Rivera-Hernández, 350 F.3d 219, 221 (1st Cir.2003) (explaining that, in the preliminary injunction context, an error of law is a per se abuse of discretion). We apply this nuanced standard of review in examining the damage awards about which the government complains.

We cut directly to the chase. The district court’s findings of fact as to the existence, nature, and quantum of the harm sustained by the scapegoats are not clearly erroneous — indeed, those findings are largely uncontested on appeal. Our inquiry, therefore, centers on the reasonableness of the awards. As explained above, abuse of discretion review applies to that issue.

Under abuse of discretion review, an appellate court ought not disturb an award of non-economic damages unless the award is either grossly disproportionate to the proven injuries or trenches upon a miscarriage of justice. See Wilkinson, 564 F.3d at 934; Neyer, 845 F.2d at 644; see also Tobin v. Liberty Mut. Ins. Co., 553 F.3d 121, 144 (1st Cir.2009) (discussing gross disproportionality in remittitur context); doCanto v. Ametek, Inc., 367 Mass. 776, 328 N.E.2d 873, 880 (1975) (similar). Though this standard is daunting, we have the authority to reduce a judge’s award of non-economic damages if that award is so extravagant as to shock our collective conscience. See, e.g., Delph v. Dr. Pepper Bottling Co. of Paragould, Inc., 130 F.3d 349, 357-58 (8th Cir.1997); Trevino v. United States, 804 F.2d 1512, 1515 (9th Cir.1986).

We approach the awards at issue here mindful that, in an FTCA case, both the nature of allowable damages and the measure of those damages are drawn from state law. Davis, 375 F.3d at 591; Lebron, 279 F.3d at 326 n. 4. Under Massachusetts law, the proper measure of damages is, within wide limits, committed to the sound discretion of the trier of fact. See Bartley v. Phillips, 317 Mass. 35, 57 N.E.2d 26, 31 (1944).

Broad discretion, however, is not to be confused with unbounded discretion. The SJC, recognizing the difficulty of placing a particular dollar value on emotional injuries, has admonished courts to strive to identify a sum that “relate[s] reasonably to the emotional distress suffered by the plaintiff.” Labonte v. Hutchins & Wheeler, 424 Mass. 813, 678 N.E.2d 853, 861 (1997). That sum should approximate the amount that reasonable persons would consider just recompense for the emotional distress inflicted. Id. at 861 n. 16. Although the SJC has suggested that a comparison of agnate awards sometimes may be useful, it has stressed the paramount importance of case-specific facts. Id. at 861-62 & n. 17.

With this framework in place, we turn to the particulars of the parties’ challenges.

1. The Government’s Appeal. As a prelude to the assessment of damages, the district court laboriously recounted the details of the scapegoats’ lives behind bars. See Limone IV, 497 F.Supp.2d at 235-41. The government has not contested the court’s narrative, and it is evident that the scapegoats suffered all the hardships customarily associated with prolonged prison confinement. These hardships were magnified by their knowledge that they had been framed: all of them were forced to come to grips with the reality that, innocence aside, they might live out their days in prison. To make matters worse, three of the men — Limone, Tameleo, and Greco — spent the first few years after the trial in the grim shadow of death sentences. All told, Limone and Salvati spent 33 and 29 years, respectively, in prison; Tameleo and Greco died in custody after 18 and 28 years, respectively. 14

After considering the particular individuals’ circumstances and consulting damage awards in other wrongful incarceration cases, the district court determined that $1,000,000 per year of immurement constituted the appropriate baseline for its calculation of damages. See id. at 243^45. The government maintains that this baseline is overly generous and results in damages that are grossly disproportionate to awards in comparable cases. In the government’s view, the district court should have limited its comparability survey to cases arising in Massachusetts and, moreover, looked only to cases involving protracted periods of incarceration. The government’s theory seems to be that wrongful incarceration gives rise to two distinct strains of emotional harm: the initial jolt of wrongful imprisonment, and some (lesser) injury based on the day-today loss of liberty. It asserts that the district court did not appreciate this important distinction; that the court did not use any congeners involving protracted periods of incarceration; and that the court erred in not limiting its canvass to Massachusetts inmates. We find the government’s reasoning unpersuasive.

To begin, the government uses faulty premises. On the one hand, its assertion that the district court did not look to awards related to lengthy periods of wrongful incarceration is incorrect as a matter of fact. See, e.g., id. at 244 (discussing a 15-year period of wrongful incarceration). On the other hand, its parochial insistence that the lower court should have restricted any inquiry to cases that arose within the borders of Massachusetts is incorrect as a matter of law. Although we have said that helpful guidance may be found in damage awards from “similar cases arising out of the same context that are tried in the same locale,” Gutiérrez-Rodriguez v. Cartagena, 882 F.2d 553, 579 (1st Cir.1989), that does not mean that a court is prohibited from looking for guidance elsewhere. The key is comparability: whether the counterpart cases involve analogous facts, similar measures of damages, and are otherwise fairly congruent. See, e.g., Morrow v. Greyhound Lines, Inc., 541 F.2d 713, 721-22 (8th Cir.1976). On the whole, we are satisfied that the district court did not abuse its discretion in looking to other eases for comparison.

Warming to the attack, the government touts a string of Massachusetts cases memorializing lesser awards. Without exception, however, these cases involve settlements, not verdicts. See, e.g., Cowans v. City of Boston, No. 05-11574, 2006 WL 4286744 (D.Mass. Aug. 4, 2006); Miller v. City of Boston, No. 03-10805, 2006 WL 4111728 (D.Mass. Mar. 9, 2006); Veláquez v. City of Chicopee, No. 03-30249, 2005 WL 3839494 (D.Mass. Oct. 14, 2005); Harding v. City of Boston, No. 98-11801, 2000 WL 33223074 (D.Mass. Feb.2000). But it is unrealistic to assume that settlement values (which, by definition, implicate compromise) equate to actual damages. See Neyer, 845 F.2d at 644. This is a comparison of plums with pomegranates. Thus, these cases do not undercut the district court’s baseline calculation.

The government also seeks to undermine the district court’s baseline by marshaling a series of legislative enactments that impose ceilings on the liability of governmental entities for wrongful incarcerations. See, e.g., 28 U.S.C. §§ 1495, 2513(e) (limiting government’s liability for wrongful incarceration of federal prisoners to $50,000 per year, or to $100,000 per year in capital cases); Mass. Gen. Laws ch. 258D, §§ 1, 5 (capping state’s liability at $500,000 per incident). But these statutes do not purport to measure the harm actually inflicted by wrongful incarceration; rather, each reflects a legislative choice to limit the sovereign’s liability. Congress could have imposed such a ceiling on damages for wrongful incarceration under the FTCA but chose instead to make the United States liable to the same extent as a private party under local law. See 28 U.S.C. § 1346(b)(1). We have neither the authority nor the inclination to veto this exercise of legislative judgment.

We turn next to the government’s plaint that the ratio of emotional distress damages to years served should decrease over time (that is, that the longer an individual is in a penitentiary, the less he should receive in damages on an annualized basis). That is an argument more appropriately made to the trier of fact. There is no flat rule to that effect — nor should there be. In some circumstances, it may be reasonable to conclude that the loss of hope as time marches on warrants larger annualized amounts for emotional injuries.

In short, the range of permissible ratios is wide. That is understandable; dollars are at best a rough and awkward proxy for time spent in the throes of wrongful incarceration. In the final analysis, it is for the trier of fact to resolve the difficult questions of quantification and monetization that lurk in the penumbra of cases such as this. See Anderson v. Robinson, 497 F.2d 120, 121 (5th Cir.1974) (noting that court of appeals possesses “no yardstick with which to measure ... abstractions”).

Let us be perfectly clear. There are limits to the trial court’s discretion in this respect, but those limits are commodious. Langevine v. Dist. of Columbia, 106 F.3d 1018, 1024 (D.C.Cir.1997). The lower court’s decision to use a sliding scale, decreasing over time, would be within the eneincture of that discretion. So, too, is its decision not to use such a sliding scale.

This brings us to the damage awards themselves. We have said before, and today reaffirm, that “there is no scientific formula or measuring device which can be applied to place a precise dollar value on matters such as restraint of freedom, fright, anxiety, loss of face, or emotional scarring.” Wagenmann, 829 F.2d at 216.

The wisdom of that statement is evident here: placing a dollar value on the emotional pain incident to wrongful incarceration, the dreary sameness of life behind bars for years on end, and the loss of freedom, relationships, and hope cries out for approximation. Moreover, the difficulty inherent in monetization of those injuries is itself a reason for deference to the front-line judgment of the trial court. Cf. Langevine, 106 F.3d at 1024 (indicating that “[a] court must be especially hesitant to disturb a jury’s determination of damages in cases involving intangible and non-economic injuries”); Wagenmann, 829 F.2d at 215 (similar). Viewed through this prism, we cannot say that the district court’s choice of baseline was unreasonable.

We do not mean to imply that the methodology employed by the district court in this case should be regarded as the norm, nor do we suggest that it should be transplanted root and branch into other factual scenarios. Were we sitting as trial judges, none of us would have employed that same methodology. The $1,000,000 per year baseline is extremely generous, and in cases involving non-economic damages we have counseled that special attention must be paid to the particular circumstances of each individual plaintiff. See, e.g., Tobin, 553 F.3d at 144-45.

But we are not sitting as trial judges in this instance. Our function is solely one of appellate review. In carrying out that task, we are not at liberty to substitute our judgment for that of the trial court. Rather, we must acknowledge the trial court’s superior coign of vantage.

Moreover, a district court, sitting without a jury, possesses a variety of implements with which to work in monetizing emotional injuries. Although particular tools must be selected and deployed with a degree of circumspection, the valuation difficulties posed by specific sets of facts also must be taken into account. Given the extent of those difficulties here, the district court’s decision to reach into its armamentarium and select a per-year baseline as the methodology of choice cannot be deemed an abuse of discretion.

That leaves the naked claim of excessiveness (a claim that encompasses the government’s charge that $1,000,000 per year is simply too rich). 15 This question is not free from doubt. The district court’s awards are considerably more munificent than the amounts that this court would have awarded in the first instance. In our view, the awards approach the outermost boundary of what might be thought eonscionable. Cf. Baba-Ali v. State, 24 Misc.3d 576, 878 N.Y.S.2d 555, 568 n. 7 (N.Y.Ct.Cl.2009) (chronicling awards of lesser amounts).

Still and all, the awards are by no means unprecedented, and the “shock-the-conscience” test cannot be administered in á vacuum. What is shocking under one set of facts may be acceptable (even if only marginally so) under different circumstances. See United States v. Santana, 6 F.3d 1, 6 (1st Cir.1993).

We are frank to say that, here, the awards for wrongful incarceration are high enough to be troubling. But when we take into account the severe emotional trauma inflicted upon the scapegoats, we cannot say with any firm conviction that those awards are grossly disproportionate to the injuries sustained. After all, some cases involving analogous factual scenarios have resulted in comparable damage awards. See, e.g., Thompson v. Connick, 553 F.3d 836, 865-66 (5th Cir.2008) (upholding jury award of $14,000,000 for 18 years of wrongful incarceration), vacated on other grounds by 562 F.3d 711 (5th Cir.2009) (en banc); Newsome v. McCabe, 319 F.3d 301, 302-03 (7th Cir.2003) (involving award of $15,000,000 for 15 years of wrongful incarceration); White v. McKinley, No. 05-203, 2009 WL 813001, at *22 (W.D.Mo. Mar.26, 2009) (upholding jury award of $14,000,000 in compensatory damages for 5% years of wrongful incarceration); Sarsfield v. City of Marlborough, No. 03-10319, 2006 WL 2850359, at *1 (D.Mass. Oct.4, 2006) (reflecting judicial award of more than $13,000,000 for 9years of wrongful incar ceration). Consequently, we conclude that the district court’s awards must stand.

In concluding that the awards in this case fall short of shocking the conscience, we think it important to make clear that the $1,000,000 annuity selected by the district court as the baseline for its calculation should not be understood as a carob seed for measuring the harm caused by wrongful incarceration generally. Applying a literal reading of the statement in Limone IV that “wrongfully imprisoned plaintiffs were entitled to compensation of at least $1 million per year of imprisonment,” 497 F.Supp.2d at 243 (emphasis supplied), one district court recently has treated the $1,000,000 per year baseline as a floor for damages arising out of wrongful incarceration. See Smith v. City of Oakland, 538 F.Supp.2d 1217, 1242-43 (N.D.Cal.2008) (citing Limone TV). We regard that characterization as unfortunate. As we have emphasized, the district court’s awards are at the outer edge of the universe of permissible awards and survive scrutiny, though barely, only because of the deferential nature of the standard of review and the unique circumstances of the case.

2. The Cross-Appeal. The district court awarded each minor child of a scapegoat $200,000 for loss of consortium and $50,000 in emotional distress damages. Limone IV, 497 F.Supp.2d at 249-50. Edward Greco, the surviving son of the late Louis Greco, Sr., objects to his award on the ground that he suffered more from his father’s wrongful incarceration than did the other children.

The cross-appeal comes to us in the following procedural posture. After the district court handed down its decision in Limone IV and entered judgment, Edward filed a motion to alter the judgment. See Fed.R.Civ.P. 59(e). The court denied that motion, declaring that any additional hardship was attributable to the Greco family’s dysfunctionality — a condition that predated Louis Greco’s conviction. See Limone v. United States (Limone V), No. 02-10890 (D.Mass. Dec. 21, 2007) (unpublished order). We review the denial of a motion to alter or amend a previously entered judgment for abuse of discretion. Vasapolli v. Rostoff 39 F.3d 27, 36 (1st Cir. 1994).

It is axiomatic that damage awards must be based on the evidence presented. A corollary to this axiom is that a court charged with making a damage award should take into account the particular circumstances of each individual plaintiff. Gutiérrez-Rodríguez, 882 F.2d at 579. This corollary holds true with respect to damages for emotional distress and loss of consortium, both of which by their very nature are difficult to monetize. See, e.g., Tobin, 553 F.3d at 144-45; Koster v. TWA, Inc., 181 F.3d 24, 35-36 (1st Cir.1999); Smith v. Kmart Corp., 177 F.3d 19, 32-33 & n. 5 (1st Cir.1999).

This does not mean, however, that different plaintiffs can never be given identical damage awards in emotional distress or loss of consortium cases. Identical damage awards at times are warranted. See, e.g., Sutton v. Earles, 26 F.3d 903, 918 (9th Cir.1994) (upholding identical annualized awards of non-economic damages to five parents of deceased seamen).

The district court engaged in a thoughtful, detailed analysis of the manner in which each scapegoat and each family member was affected by the government’s misconduct. See Limone IV, 497 F.Supp.2d at 235-13. Within that analysis, the court chronicled the deterioration of the Greco family. Id. at 241-43. The court’s rescript reveals that Edward was eleven years old when his father was sentenced. Around that time, his mother, Roberta, began to drink heavily, and Ed ward became the primary caretaker for his older brother. Roberta abused Edward physically and, when he was thirteen, abandoned him without making provisions for his care. Edward and his brother lived with extended family, but Edward was thrown out when he was sixteen. He soon lost contact with his brother (who eventually committed suicide).

The district court determined that Edward’s plight, though tragic, was attributable mainly to causes that predated his father’s imprisonment. Roberta had filed for divorce three years before Greco’s conviction, charging extreme cruelty. In response, Greco attempted to strangle her. Indeed, the marital relationship was so troubled that the district court rejected Roberta’s claim for loss of consortium (though it awarded her damages for intentional infliction of emotional distress). Id. at 247, 250.

Based on this background, the district court concluded in Limone V that the government’s misconduct caused only a fraction of the woes that befell Edward. The rest would have occurred in any event because of the dysfunctional family environment.

Edward resists this conclusion, admonishing that a defendant takes a plaintiff as it finds him. See Doty v. Sewall, 908 F.2d 1053, 1059 (1st Cir.1990); Dulieu v. White & Sons, [1901] 2 K.B. 669, 679. That is true as far as it goes — but it does not take Edward very far. A defendant may be held hable only for the damages that it actually causes. See W. Page Keeton, Prosser & Keeton on Torts 292 (5th ed.1984) (reiterating this principle in regard to “eggshell-skull” plaintiffs).

Causation is generally a question of fact, committed largely to the competence of the factfinder. See Peckham, 895 F.2d at 837. Given the idiosyncratic circumstances surrounding Edward’s claim, we cannot say that the district court either clearly erred in holding that the government’s misconduct was not a but-for cause of Edward’s special hardships or abused its discretion in denying his motion to alter the judgment.

III. CONCLUSION

We summarize succinctly. The district court handled this matter with care and assiduous attention to detail. It took pains to make specific findings and to explain its reasoning. While we reject its finding that the government is liable for malicious prosecution, we uphold the court’s alternate finding that the government is liable for intentional infliction of emotional distress. We also uphold the district court’s decision to reject the government’s invocation of the discretionary function defense. Finally, we conclude that the district court used a permissible methodology in computing damages and that the damage awards, though high, are not so excessive as to warrant appellate intervention.

We need go no further. This case exemplifies a situation in which the end did not justify the government’s use of very unattractive means. In its zeal to accomplish a worthwhile objective (stamping out organized crime), the FBI stooped too low. Its misconduct was not only outrageous but also tortious. That misconduct resulted in severe harm to the persons wrongfully convicted and to their families. Under these unfortunate circumstances, the large damage awards mark the last word of a sad chapter in the annals of federal law enforcement.

Affirmed.

Appendix

The district court awarded damages as follows:

1. Estate of Louis Greco, Sr.— $28,000,000;

2. Peter Limone, Sr. — $26,000,000;

3. Joseph Salvati — $29,000,000;

4. Estate of Enrico Tameleo — • $13,000,000;

5. Olympia Limone (wife of Peter Limone, Sr.) — $1,050,000;

6. Marie Salvati (wife of Joseph Salvati) — $1,050,000;

7. Estate of Giovannina Tameleo (deceased wife of Enrico Tameleo)— $1,050,000;

8. Peter Limone, Jr. (son of Peter Limone, Sr.) — $250,000;

9. Paul Limone (son of Peter Limone, Sr.) — $250,000;

10. Carolyn Limone Zenga (daughter of Peter Limone, Sr.) — $250,000;

11. Janine Limone Arria (daughter of Peter Limone, Sr.) — $250,000;

12. Maria Sidman (daughter of Joseph Salvati) — $250,000;

13. Sharon Salvati (daughter of Joseph Salvati) — $250,000;

14. Gail Orenberg (daughter of Joseph Salvati) — $250,000;

15. Anthony Salvati (son of Joseph Salvati) — $250,000;

16. Edward Greco (son of Louis Greco, Sr.) — $250,000;

17. Estate of Louis Greco, Jr. (deceased son of Louis Greco, Sr.) — $250,000;

18. Roberta Werner (ex-wife of Louis Greco, Sr.) — $50,000;

19. Saverio Tameleo (son of Enrico Tameleo) — $50,000. 16

1 . French, Martin, and Cassesso also were convicted on charges related to the murder. The legitimacy of those convictions is not an issue here.

2 . Because the discretionary function exception, when applicable, deprives a court of subject matter jurisdiction, some might think consideration of it logically antecedent to consideration of the merits. But the answer to the discretionary function inquiry depends in large measure on the nature of the conduct at issue. See, e.g., Fothergill v. United States, 566 F.3d 248, 252-53 (1st Cir.2009); Shansky v. United States, 164 F.3d 688, 690-91 (1st Cir.1999). In an appellate court, after a full trial on the merits, the discretionary function inquiry sometimes is better performed at a later stage in the analysis. This is such a case.

3 . We recognize that the question of when a cause of action for malicious prosecution arises under the law enforcement proviso of the FTCA is jurisdictional in nature. We have broad discretion, however, to take issues in whatever order practicality may suggest, see, e.g., Puerto Rico v. United States, 490 F.3d 50, 70 (1st Cir.2007), and we exercise that discretion here.

4 . The FBI report of this interview indicates that Doyle conducted the interview. Moreover, the substance of that FBI report consists only of a copy of Doyle’s interview notes. That general scenario is characteristic of the FBI reports of subsequent interrogation sessions at which the detectives were present.

5 . Although Barboza would make alterations to the fairy tale that he narrated on September 8, these alterations too came about during joint interview sessions led by state law enforcement officers.

6 . Agis is the seminal Massachusetts case recognizing a cause of action for intentional infliction of emotional distress in the absence of an accompanying physical injury. Although the scapegoats’ convictions predate this decision, the United States has not ascribed any relevance to this chronology. Thus, we deem waived any contention that the government cannot be held liable under the FTCA on a tort theory that was not firmly established in the case law of the relevant state at the time of the tortious acts.

7 . The district court’s finding of malicious prosecution does not require a different result. That finding was incorrect as a matter of law, see supra Part 11(A), and is entitled to no weight.

8 . An airtel is an inter-office FBI communique sent between a local field office and FBI headquarters. A correlator report is a document that summarizes all the pertinent information possessed by a field office concerning a particular person or subject.

9 . The government seizes upon a statement in the district court’s opinion suggesting that state officials may have shared in the responsibility for showing Barboza their investigative files. See Limone IV, 497 F.Supp.2d at 180 ("Someone in law enforcement had to have done so, either the FBI directly or state law enforcement in the FBI's presence.”). But the court’s opinion, fairly read, attributes primary responsibility for shoring up Barboza’s testimony to the FBI. See, e.g., id. at 179, 180.

10 . The most recent draft restatement suggests that an intentional tortfeasor should be deemed the legal cause of any harm that he intends to inflict without reference to the foreseeability of that harm. See Restatement (Third) of Torts: Liability for Physical Harms § 33 (Proposed Final Draft No. 1, 2005). The SJC has not spoken to this proposal and, given the foreseeability of the harm flowing from the FBFs misconduct, we need not decide whether section 33 applies here.

11 . In Massachusetts, a cause of action for intentional infliction of emotional distress may succeed only if the defendant has intentionally inflicted severe emotional distress "without privilege." Agis, 355 N.E.2d at 318. The government did not argue, either in the district court or in its briefs on appeal, that its conduct was privileged. Consequently, its assertion of privilege for the first time at oral argument in this court is too little and too late. See United States v. Slade, 980 F.2d 27, 30 (1st Cir.1992) ("It is a bedrock rule that when a party has not presented an argument to the district court, she may not unveil it in the court of appeals."); Anderson v. Beatrice Foods Co., 900 F.2d 388, 397 (1st Cir.1990) (holding that an appellant’s briefs fix the scope of the issues appealed and that, therefore, an appellant cannot breathe life into an omitted theory merely by referring to it at oral argument); see also McCullen v. Coakley, 571 F.3d 167, 182 (1st Cir.2009) (holding that theoiy advanced by a member of the court at oral argument, but neither briefed nor raised below, is waived).

12 . But see Ralph Keyes, Nice Guys Finish Seventh; Phrases, Spurious Sayings and Familiar Misquotations 152 (1992) (noting that “although this [phrase] is commonly cited as a ‘Berra-ism,’ Yogi Berra denies ever saying it”).

13 . In so holding, we do not view the FBI’s constitutional transgressions as corresponding to the plaintiffs’ causes of action — after all, the plaintiffs' claims are not Bivens claims — but rather, as negating the discretionary function defense. See Bolduc, 402 F.3d at 56 (“Federal constitutional or statutory law cannot function as the source of FTCA liability.”).

14 . Limone served the first seven years of his sentence and Tameleo served the first five years of his sentence concurrent with previously imposed sentences in unrelated cases. Thus, the district court declined to award either of them damages for those periods. See Limone IV, 497 F.Supp.2d at 245.

15 . The government has not specifically challenged the amounts of the derivative awards (or, for that matter, the liability findings) on the plaintiffs' claims for loss of consortium and the like. Consequently, we eschew any discussion of those awards in connection with the government’s appeal. See. United States v. Zannino, 895 F.2d 1, 17 (1st Cir.1990) (explaining that issues not briefed or argued are deemed abandoned).

16 . The scapegoats received $1,000,000 for each year of incarceration, less time served on unrelated concurrent sentences. See supra note 14. The district court's derivative damage awards were constructed as follows. The court awarded $50,000 to each family member to compensate for the government’s intentional infliction of emotional distress upon innocent bystanders. Limone IV, 497 F.Supp.2d at 250. It also awarded $1,000,000 each to the wives of the scapegoats other than Greco and $200,000 to each of the scapegoats’ minor children for loss of consortium. Id. at 248, 249.