This case raises issues of First Amendment law. At the center of the dispute is The Pnce of Sugar, a documentary film released in 2007 by film company Uncommon Productions, LLC, and its principal William M. Haney, III. The film depicts the treatment of Haitian laborers on sugarcane plantations in the Dominican Republic. It refers by name to brothers Felipe and Juan Vicini Lluberes, senior executives of a family conglomerate that owns and operates Dominican sugar plantations. The Vicinis contend that the film is defamatory and sued the filmmakers in federal court. The filmmakers moved for summary judgment, which the court granted. The Vicinis appeal the entry of summary judgment and the denial of a motion to compel production of discovery materials.

For the reasons that follow, we affirm in part the entry of summary judgment but otherwise vacate the judgment, vacate the order denying the motion to compel, and *11 remand for further proceedings consistent with this opinion.

I. BACKGROUND

The controversy that spawned The Pnce of Sugar is well catalogued in the district court’s rescript, Lluberes v. Uncommon Prod’ns, LLC, 740 F.Supp.2d 207 (D.Mass. 2010), and we will not rehash it. Suffice it to say that the treatment of Haitian laborers on Dominican sugarcane plantations and the conditions of company towns (or bateyes) where they live have received scrutiny from many sectors for many years.

In 2004, the filmmakers began shooting in the Dominican Republic. Much of the film follows Fr. Christopher Hartley, a Roman Catholic priest critical of the Vicinis, as he seeks to improve conditions for his parishioners in the bateyes. Those conditions, the film highlights, include shanty quarters, inadequate provisions, and little if any education for children. At several points, Fr. Hartley and the film’s narration reference Vicini-owned bateyes and identify Felipe and Juan as bearing some measure of responsibility for their disrepair. The film was released publicly on March 11, 2007, at a film festival in Texas. It has since received limited screenings in a handful of major cities and other venues.

Later in 2007, the Vicinis sued the filmmakers in federal district court in Massachusetts. 1 Invoking the court’s diversity jurisdiction, the Vicinis alleged that the film was defamatory and identified fifty-three statements, although they later winnowed the number of allegedly defamatory statements down to seven. The filmmakers seasonably moved for summary judgment on these remaining statements; they argued that Felipe and Juan were “public figures” required to prove “actual malice” in accordance with New York Times Co. v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), and its progeny, and that the Vicinis could not so prove. The district court agreed and granted summary judgment in the filmmakers’ favor.

At the same time, the court denied a motion to compel that the Vicinis had initially filed during discovery and later renewed. The motion sought production of several categories of documents; those at issue here include communications with a third-party “script annotator” that the filmmakers had withheld on attorney-client privilege grounds. The judge did not explain his reasoning. This appeal followed.

II. DISCUSSION

We begin with the public-figure question, then turn to the discovery dispute and go no further.

A. Public-Figure Status

1. Defamation and the First Amendment

Before the Supreme Court’s decision in New York Times, defamation law was shaped by the states and strongly favored their interest in protecting an individual’s reputation. See Bruno & Stillman, Inc. v. Globe Newspaper Co., 633 F.2d 583, 586 (1st Cir.1980) (“Once a plaintiff put into evidence a reputation-harming statement and proof that defendant caused it to be disseminated, he enjoyed an irrebuttable presumption of injury and a rebuttable presumption of falsity.”); see generally Joel D. Eaton, The American Law of Defamation Through Gertz v. Robert Welch, *12 Inc. and Beyond: An Analytical Primer, 61 Va. L.Rev. 1349, 1351-57 (1975) (surveying defamation law in the various states that existed prior to New York Times).

That balance shifted in 1964, when the Court considered whether “the constitutional protections for speech and press limit a State’s power to award damages in a libel action brought by a public official against critics of his official conduct.” N.Y. Times, 376 U.S. at 256, 84 S.Ct. 710. Recognizing the “profound national commitment to the principle that debate on public issues should be uninhibited, robust, and wide-open,” id. at 270, 84 S.Ct. 710, the Court reasoned that even falsehoods “must be protected if the freedoms of expression are to have the breathing space that they need to survive,” id. at 271-72, 84 S.Ct. 710 (internal quotation marks and ellipsis omitted). On that basis, the Court held that the First Amendment “prohibits a public official from recovering damages for a defamatory falsehood relating to his official conduct unless he proves that the statement was made with ‘actual malice’ — that is, with knowledge that it was false or with reckless disregard of whether it was false or not.” Id. at 279-80, 84 S.Ct. 710. 2

The Court soon applied the New York Times rule to nonofficial “public figures.” Curtis Publ’g Co. v. Butts, 388 U.S. 130, 154-55, 87 S.Ct. 1975, 18 L.Ed.2d 1094 (1967). Under Curtis, a defamation plaintiff was to be, considered a public figure when he “commanded sufficient continuing public interest and had sufficient access to the means of counter-argument to be able to expose through discussion the falsehood and fallacies of the defamatory statements.” Id. at 155, 87 S.Ct. 1975 (internal quotation marks and citation omitted).

For a time, the New York Times rule was also extended to private individuals. Rosenbloom v. Metromedia, Inc., 403 U.S. 29, 91 S.Ct. 1811, 29 L.Ed.2d 296 (1971) (plurality opinion). According to the Rosenbloom plurality: “If a matter is a subject of public or general interest, it cannot suddenly become less so merely because a private individual is involved, or because in some sense the individual did not ‘voluntarily’ choose to become involved. The public’s primary interest is in the event[.]” Id. at 43, 91 S.Ct. 1811. Rather, the linchpin became simply “whether the utterance involved concerns an issue of public or general concern.” Id. at 44, 91 S.Ct. 1811; see also id. at 43-44, 91 S.Ct. 1811 (“We honor the commitment to robust debate on public issues, which is embodied in the First Amendment, by extending constitutional protection to all discussion and communication involving matters of public or general concern, without regard to whether the persons involved are famous or anonymous.”).

The plurality’s approach in Rosenbloom, however, was repudiated in Gertz v. Robert Welch, Inc., 418 U.S. 323, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1974), which established the current framework. Gertz sought an accommodation between the “need to avoid self-censorship by the news media,” id. at 341, 94 S.Ct. 2997, on the one hand, and the “legitimate state interest underlying the law of libel,” id., on the other. It did so by linking “the constitutionally required showing in a defamation action to the plaintiffs status.” Pendleton v. City of Haverhill, 156 F.3d 57, 67 (1st Cir.1998). Under this new model, public figures could succeed only on proof of actual malice as defined by New York Times. Gertz, 418 U.S. at 342, 94 S.Ct. 2997. As for purely *13 private individuals, however, the states could “define for themselves the appropriate standard of liability” so long as minimal constitutional safeguards were met. Id. at 346-47, 94 S.Ct. 2997. 3

Gertz identified two justifications for this public-figure/private-figure dichotomy. The first, foreshadowed in Curtis, was that “public figures usually enjoy significantly greater access to the channels of effective communication and hence have a more realistic opportunity to counteract false statements than private individuals normally enjoy.” Gertz, 418 U.S. at 344, 94 S.Ct. 2997. Thus, “[pjrivate individuals are ... more vulnerable to injury, and the state interest in protecting them is correspondingly greater.” Id. The second justification, which was said to be more important, was that public figures, like public officials, “have assumed roles of especial prominence in the affairs of society” and “must accept certain necessary consequences” of that status. Id. at 344-45, 94 S.Ct. 2997. One such consequence is “the risk of closer public scrutiny than might otherwise be the case.” Id. at 344, 94 S.Ct. 2997; see also id. at 345, 94 S.Ct. 2997 (reasoning that a private individual “has relinquished no part of his interest in the protection of his own good name, and consequently he has a more compelling call on the courts for redress of injury inflicted by defamatory falsehood”).

Gertz contemplated that public-figure status usually would arise in one of two ways, each with different repercussions. In one, “an individual may achieve such pervasive fame or notoriety that he becomes a public figure for all purposes and in all contexts” — the so-called general-purpose public figure. Id. at 351, 94 S.Ct. 2997. But far more commonly (and directly relevant in this case) “an individual voluntarily injects himself or is drawn into a particular public controversy and thereby becomes a public figure for a limited range of issues” — the so-called limited-purpose public figure. Id. That “limited range of issues” is identified “by looking to the nature and extent of an individual’s participation in the particular controversy giving rise to the defamation.” Gertz, 418 U.S. at 352, 94 S.Ct. 2997. 4

Guidance since Gertz has cautioned that a controversy must be more than a “cause célebre,” Time, Inc. v. Firestone, 424 U.S. 448, 454, 96 S.Ct. 958, 47 L.Ed.2d 154 (1976), or “a matter that attracts public attention,” Wolston v. Reader’s Digest Ass’n, 443 U.S. 157, 167, 99 S.Ct. 2701, 61 L.Ed.2d 450 (1979). Rather, it must be shown that “ ‘persons actually were discussing some specific question ... [and] a reasonable person would have expected persons beyond the immediate participants in the dispute to feel the impact of its resolution.’ ” Bruno & Stillman, 633 F.2d at 591 (quoting Waldbaum v. Fair- *14 child Publ’ns, Inc., 627 F.2d 1287, 1297 (D.C.Cir.1980)). Also, to avoid improper “bootstrapping” (a concept explored further below), the controversy must predate the alleged defamation. Gray v. St. Martin’s Press, Inc., 221 F.3d 243, 251 (1st Cir.2000); Kassel v. Gannett Co., 875 F.2d 935, 941 (1st Cir.1989); Bruno & Stillman, 633 F.2d at 591; see also Hutchinson v. Proxmire, 443 U.S. 111, 135, 99 S.Ct. 2675, 61 L.Ed.2d 411 (1979) (“[T]hose charged with defamation cannot, by their own conduct, create their own defense by making the claimant a public figure.”).

Once a controversy is isolated, the critical question then becomes whether the plaintiff has attempted to “influence the resolution” of that controversy. See, e.g., Wolston, 443 U.S. at 168, 99 S.Ct. 2701 (public figures engage “the attention of the public in an attempt to influence the resolution of the issues involved” or use a newsworthy event “as a fulcrum to create public discussion”); Firestone, 424 U.S. at 453, 96 S.Ct. 958 (public figures “thrust themselves to the forefront of any particular controversy in order to influence the resolution of the issues involved in it”); Pendleton, 156 F.3d at 69 (holding that the defamation plaintiff was a public figure because he “voluntarily injected himself’ into the controversy); Bruno & Stillman, 633 F.2d at 591 (requiring a “thrusting into the vortex”). If so, the plaintiff is a public figure and bears the heavy, and often insurmountable, burden of proving actual malice.

2. Felipe and Juan

The filmmakers contend that Felipe and Juan are limited-purpose public figures. The Vicinis vehemently dispute that label. Although they no longer contest the existence of a public controversy, the Vicinis argue that neither of them attempted to influence its resolution. Their argument has three constituent parts and spans both time and space. First, they say they did nothing before 2003 that, standing alone, could subject them to public-figure status. Second, any conduct after 2003 that might do so, we are told, is shielded by the anti-bootstrapping principle. The third is that, whatever their conduct in the Dominican Republic, it cannot make them public figures in the United States.

The status question is a legal one that we review de novo. Pendleton, 156 F.3d at 68; see Rosenblatt v. Baer, 383 U.S. 75, 88, 86 S.Ct. 669, 15 L.Ed.2d 597 (1966). We do so mindful that the inquiry is “inescapably fact-specific,” Mandel v. Bos. Phoenix, Inc., 456 F.3d 198, 204 (1st Cir.2006), and does not always lend itself to summary judgment, compare id. at 204-07 (vacating status determination because “the factual record, at the summary judgment stage, was too uncertain to warrant a legal conclusion either way about Mandel’s status”), with Pendleton, 156 F.3d at 68 (affirming status determination because “[tjhere is no conflict as to any material fact; the issue is whether the discerned facts suffice to establish that Pendleton acted in a way sufficient to make him a public figure for the purpose of this defamation action”).

But here, as in Pendleton, the Vicinis do not argue that the district court based its status determination on disputed facts, only that the undisputed facts were insufficient to make them public figures for the reasons outlined above. 5 We disagree *15 and conclude, as did the district court, that both Felipe and Juan are limited-purpose public figures. Because we ultimately reject their anti-bootstrapping argument, 6 we find it appropriate to begin by examining all relevant conduct up to the film’s release in March 2007.

Within that span, both Felipe and Juan came to occupy leadership positions within the family businesses. Felipe began working for the company in the mid-to-late 1990s as a member of the board. He gradually became part of a small group of family members that directed the agricultural enterprise; among other things, he oversaw sugar exports and sought to ensure favorable trade policies with the United States (the largest importer of Dominican sugar) and other countries. Later he was installed as president of Grupo Vicini, the entity that manages the family’s investments and coordinates initiatives on the bateyes. Juan joined the company in 2000 directly out of school in the United States. He, too, began working on the agricultural side of the business and ultimately assumed the number-two position in Grupo Vicini, under Felipe. 7

Juan’s role was perhaps less conspicuous, but it focused on the bateyes from the beginning. His homecoming in 2000 coincided with Fr. Hartley’s controversial benediction — delivered during a visit to Fr. Hartley’s parish by the Dominican president — that was critical of the batey system and of those, including the Vicini family, responsible for it. The strongly worded benediction caught the attention of the media, prompting the Vicinis to call a meeting with Fr. Hartley. As a result of that meeting, which both Juan and Felipe attended, Juan took on the role of humanitarian attaehé to Fr. Hartley and his cause of improving conditions on the bateyes. Over the next couple of years Juan and Fr. Hartley met about a dozen more times, toured the bateyes together, and regularly spoke by telephone. Juan also hired a social worker and tasked her with identifying the most pressing items in need of resolution. 8

*16 For reasons that are not altogether clear, the Vicini-Hartley collaboration fizzled in 2003 or 2004. But rather than abandon the project, the Vicinis embraced it as their own. For instance, Juan reached out to nongovernmental organizations in order to combat HIV/AIDS in the bateyes. A newsletter reporting on the initiative described the efforts of “Casa Vicini” to “develop cooperative plans to combat HIV/AIDS, with the support of the World Bank,” and included a picture of Juan signing a “participation agreement.” Juan himself described initiatives like these in a letter to church officials as attempts to find a “solution” to the batey problem. And around the same time, both Juan and Felipe began courting the U.S. embassy in Santo Domingo. Juan personally escorted embassy officials on visits to the bateyes; one visit was attended by the Deputy Chief of Mission, the embassy’s second-in-command. For his part, Felipe spoke with embassy officials by telephone, including the U.S. ambassador. This outreach touched off a relationship between the Vicinis and U.S. diplomats that Felipe described as “ongoing.”

Their efforts entered a new phase in 2005. After a U.S. newspaper published an exposé critical of the batey system, the Vicinis brought in Newlink Communications, a public-relations (PR) firm based in the United States. Newlink’s proposal, signed by Felipe in April 2005, provided for a massive PR campaign in the Dominican Republic that would reach as far as the United States. Among other things, the proposal identified the need for a “strategic communications program” to deal with the “negative perceptions against the company, reaching the United States media,” “[bjlock messages” critical of the Vicinis, and “[ijmprove the image and reputation of the company in the eyes of the public.” It spelled out country-specific strategies, focusing on the Dominican Republic and the United States, designed to implement those general goals. And it included media training for both Felipe and Juan, in Spanish and English, such as mock interviews about the bateyes and model answers emphasizing Vicini initiatives. All told, the Newlink deal cost the Vicinis about $1.2 million.

As it was implemented, the Vicini-Newlink PR campaign targeted several sectors within the Dominican political apparatus. A Vicini deputy arranged a series of meetings with senior Dominican officials, particularly those responsible for immigration and labor policies. Those meetings culminated in a Felipe-led tour of the bateyes that showcased Vicini initiatives to the heads of the Dominican interior and labor departments. In the ecclesiastical sector, Felipe and Juan met with the Papal Nuncio — the Vatican’s equivalent of an ambassador — to “get our story out” about their batey initiatives. Felipe also attended a conference of bishops to give a presentation on “Grupo Vicini’s social programs.” And the press was also targeted: in a single month Felipe was pictured and quoted in three Dominican newspapers concerning the company’s “restructuring” efforts in “human resource components.” He was quoted again the next month, this time about improvements in the bateyes and conditions for workers generally. Then he appeared in a documentary about the bateyes (not the film at issue here) that was broadcast on Dominican television. In it, he described efforts to build rural medical clinics, root out child labor, and bring modern mechanization to sugarcane cut *17 ting. This was followed by several ‘Yes to Change” events, hosted by the Vicinis to advertise their altruism, that resulted in another string of articles picturing and quoting Felipe.

The PR campaign also targeted international media outlets and policymakers, particularly in the United States. Felipe traveled to the United States to meet with the author of the newspaper exposé referenced above. The purpose of the meeting, which took place in Newlink’s Miami offices, was for the reporter “to hear the other side of the story.” On another occasion, Felipe sent a deputy to a PBS interview with the stated goal of attempting “to ‘flip’ the story” in the family’s favor. According to Newlink records, the deputy “was prepared ahead of time for that interview with a Q & A that Newlink drew up to ensure that his answers were in keeping with the goal of maintaining the [company’s] image intact.” And in late 2006, Felipe, accompanied by several Newlink team members, led a U.S. congressional delegation on a tour of Vicini bateyes. During the tour a “fact sheet” was distributed that described Vicini initiatives in detail. CNN covered the delegation and interviewed Felipe; clips of that interview aired on Anderson Cooper 360° and were rebroadcast multiple times over the next two months on CNN and its affiliates. Felipe testified that his goal during these events was “to try to get our story out, to get our side out.”

Shortly before the release of The Price of Sugar (the end of our continuum), Felipe and Juan hosted an industry luncheon in the Dominican Republic. One purpose of the luncheon was to reveal more Vicini initiatives on the bateyes. During the luncheon, journalists from several Dominican newspapers were permitted to attend and ask questions. The resulting articles highlighted the batey initiatives discussed during the event and, as before, pictured and quoted Felipe and Juan.

All together, this conduct shows beyond hope of legitimate contradiction that Felipe and Juan are limited purpose public figures. Both leveraged their positions and contacts to influence a favorable outcome in the batey controversy. Both enjoyed access to the press and exploited it by orchestrating a PR blitz to garner public support and mute their critics. 9 In doing so, both assumed roles of prominence for this limited purpose and the risk of closer public scrutiny that came with it.

*18 3. Bootstrapping and the Privilege of Reply

The Vicinis try to avoid this conclusion by asserting that most of the above conduct is shielded by the bootstrapping taboo. The argument is as follows. All of their “public activities” occurred after and in response to an article in a Spanish newspaper, El Mundo, published in January 2003. The article included purportedly defamatory statements by Fr. Hartley, the “original defamer,” that were repeated in the film four years later. Because the Vicinis would not have entered the public arena but for the El Mundo article, the filmmakers cannot invoke the Vicinis’ status as a defense to the same defamation in the film.

The argument is creative, but this case does not fit the bootstrapping mold. Bootstrapping in this context occurs when the defendant relies on his own defamatory publication to manufacture a public controversy involving the plaintiff, and thus “by [his] own conduct, create[s his] own defense by making the claimant a public figure.” Hutchinson, 443 U.S. at 135, 99 S.Ct. 2675. That is the logic behind the requirement that public-figure status— whether acquired for all purposes and in all contexts or derived from a particular controversy — predate the alleged defamation. See, e.g., Gray, 221 F.3d at 251; Kassel, 875 F.2d at 941; Bruno & Stillman, 633 F.2d at 591; see generally Smolla, Law of Defamation § 2:25 (recognizing “the media’s potential for ‘bootstrapping’ itself into the protection of the actual malice standard by pointing to its own coverage of the plaintiff as evidence that the plaintiff is a public figure,” and that in response “a number of courts have emphasized that the public controversy must ‘preexist’ the speech giving rise to the defamation suit”).

Here, however, the El Mundo article did not create the batey controversy. It is undisputed, at least on appeal, that the controversy began much earlier and was in full swing before Fr. Hartley arrived in the Dominican Republic and well before the El Mundo article was published in 2003. Like the Vicinis themselves, Fr. Hartley was a voice in that controversy; he was not its creator. So even if the El Mundo article served as a “blueprint” for the film — a claim that we think the Vicinis exaggerate 10 — no bootstrapping occurred from which the filmmakers could have benefitted in this lawsuit.

What the Vicinis argue, at bottom, is that their conduct fell under a so-called privilege of reply. The concept has its roots not in the First Amendment but in the common law that governed defamation suits prior to New York Times. Writ large, it allowed a defamed person to respond to the extent reasonably necessary to defend himself, even to the point of defaming his accuser. 11 Under their theo *19 ry, as we understand it, the Vicinis could publicly defend themselves against defamatory statements in the El Mundo article without sacrificing their private-figure status, as long as their response was measured and reasonable. Although not in so many words, the Vicinis ask us to graft the common-law privilege of reply onto the constitutional public-figure analysis.

To our knowledge only one court of appeals has explicitly taken such a step. See Foretich v. Capital Cities/ABC, Inc., 37 F.3d 1541 (4th Cir.1994). In Foretich, grandparents, accused by their daughter-in-law of molesting their infant granddaughter, made limited “public comments and appearances” to rebut her accusations. Id. at 1557-58 (granting requests for interviews, attending three press conferences, and appearing on two television shows). The court acknowledged that some of those rebuttals “were probably intended (at least in part) to influence the outcome of the custody dispute.” Id. at 1563. Nevertheless, invoking the common-law privilege of reply, the court held that the grandparents’ “primary motive was to defend their own good names against [her] accusations and that their public statements can most fairly be characterized as measured defensive replies to her attacks, rather than as efforts to thrust themselves to the forefront of a public controversy in order to influence its outcome.” Id.

We agree with Foretich in this limited sense: an individual should not risk being branded with an unfavorable status determination merely because he defends himself publicly against accusations, especially those of a heinous character. Pendleton, 156 F.3d at 68; see Firestone, 424 U.S. at 455 n. 3, 96 S.Ct. 958; accord Clyburn v. News World Commc’ns, Inc., 903 F.2d 29, 32 (D.C.Cir.1990) (“[W]e have doubts about placing much weight on purely defensive, truthful statements made when an individual finds himself at the center of a public controversy but before any libel occurs; it is not clear why someone dragged into a controversy should be able to speak publicly only at the expense of foregoing a private person’s protection from defamation.”). 12

*20 But that is not what happened here. Although the Vicinis claim that the El Mundo article was a call to arms, the record is clear that they took little if any action directly in response to it. Juan testified that the only action he took was to order an internal inquiry concerning one accusation; Felipe testified that he took no action whatsoever. And even if the El Mundo article had some indirect influence on their conduct over the next four years, that conduct went well beyond any reasonable measure of self-defense, as we have shown. 13

4. Public Figures and Geography

The Vicinis’ final argument on the limited-purpose public figure issue is a geographic one. They say that none of the above conduct makes them public figures in the United States, where the alleged defamation was published. The argument rests on an analogy to general-purpose public figures, and those authorities that require such individuals to have achieved notoriety where they were defamed. The Vicinis reason that this geographic restriction must also be true for limited-purpose public figures, who are the more “protected” of the two.

The analogy is flawed. Gertz held that the plaintiff was not a public figure for all purposes because he had “no general fame or notoriety in the community” and was not generally known to “the local population.” 418 U.S. at 851-52, 94 S.Ct. 2997. Based on that language, some courts — we have not addressed the question and we do not do so today — have extrapolated that a general-purpose public figure need not attain “nationwide fame,” only “notoriety where he was defamed[,] ie., where the defamation was published.” Waldbaum, 627 F.2d at 1295 n. 22. 14 Arguably, this so-called community standard actually expands rather than restricts the applicability of the New York Times rule, at least for general-purpose public figures.

That debate, however, has no relevance here. Gertz defined a limited-purpose public figure not in terms of geography but in terms of the controversy that he has stepped into. See Gertz, 418 U.S. at 351, 94 S.Ct. 2997 (defining a limited-purpose public figure as one who “voluntarily injects himself or is drawn into a particular controversy”); Tavoulareas, 817 F.2d at *21 772 (“[T]he scope of the controversy in which the plaintiff involves himself defines the scope of the public personality.”). That suggests to us that, if Gertz envisioned any limitation on public-figure status, it is a limitation inherent in the scope of the controversy itself.

Cases on this point, though rare, have reached essentially the same conclusion. See, e.g., Trotter v. Jack Anderson Enters., Inc., 818 F.2d 431 (5th Cir.1987). In Trotter, the Fifth Circuit held that the plaintiff was a limited-purpose public figure for purposes of a U.S. publication because the controversy — labor violence at a Coca-Cola bottling plant in Guatemala — resonated in the United States: “Because of the proximity of other Western Hemisphere countries to the United States, social and political turmoil occurring there has aroused particular domestic concern.” Id. at 434. Although the plaintiffs affiliation with a U.S. company played some role, the court’s analysis focused on the fact that the controversy had “captured the attention of a diverse and broadly-based audience [in the United States], including the media, political leaders, human-rights organizations, labor unions, and Coca-Cola shareholders.” Id. 15

Similarly, here, the batey controversy was not confined to the shores of the Dominican Republic. Rather, it resounded in the United States for obvious humanitarian reasons and a less-obvious geopolitical one: a long-standing import quota system under U.S. law that subsidizes Dominican sugar producers, including the Vicinis. 16 Indeed, one of Felipe’s core responsibilities was seeing to it that this quota system remained intact through lobbying and other efforts. Concerns that negative publicity about the bateyes might jeopardize the quota system prompted him and Juan, at least in part, to launch the PR blitz that reached U.S. media outlets and policymakers, as we have shown. We are satisfied that such conduct is enough to make the plaintiffs public figures in the United States for purposes of this lawsuit. 17

*22 B. Attorney-Client Privilege

Documentarians commonly obtain insurance against potential “errors and omissions” in their films. The pursuit of such insurance in this case led the filmmakers to Frederick Leopold, an attorney and specialist in that field. Leopold told them that, in order to procure insurance, he would need a third-party report confirming the accuracy of the script, and recommended or directed that they retain Elizabeth Bardsley and her company to do it. The filmmakers then hired Bardsley; she worked closely with the filmmakers and in the process numerous email communications were generated between them. Ultimately, she rendered a report in the form of an annotated script which she provided to the filmmakers and to Leopold. .

During discovery the Vicinis sought the email communications and the report, but the filmmakers withheld them all on attorney-client privilege grounds. 18 The Vicinis responded by filing a motion to compel and averred at a 2008 hearing on pending motions that the documents were pertinent to the filmmakers’ state of mind. At that time, the district court indicated that the nature of the relationship between Leopold, the filmmakers and Bardsley may involve a factual matter requiring live testimony. The topic of possible in camera review also was broached. This particular discovery dispute, however, remained unresolved for a stretch of time in the midst of the parties’ ongoing motion practice.

Finally, litigation on the then-renewed motion to compel came to a head around the same time that the filmmakers pitched their summary judgment motion. At a 2009 hearing again addressing pending motions, the Vicinis identified the desired documents as material to the issue of actual malice. Acknowledging that the renewed motion to compel had “fallen through the cracks,” the district court indicated that it would consider the document request in the context of the then-pending dispositive motion. Ultimately, while it granted the filmmakers summary judgment in a detailed and necessarily lengthy order, the court denied the renewed motion to compel without any substantive discussion.

On appeal the Vicinis argue that the district court erred in denying their discovery motion and that the documents are critical to the actual malice inquiry. For their part, the filmmakers contend that the district court properly determined that the attorney-client privilege applies. 19 We conclude that the record fails to support *23 the district court’s ruling that the filmmakers established that all the documents are protected by the attorney-client privilege as a matter of law. Rather than reverse the district court’s decision, we think it prudent under the circumstances of this case to vacate and remand for the court to take up the matter in light of the guidance we provide forthwith.

The standard of review concerning a claim of privilege depends on the particular issue. Cavallaro v. United States, 284 F.3d 236, 245 (1st Cir.2002). Questions of law are reviewed de novo, findings of fact for clear error, and evidentiary determinations for abuse of discretion. Id. The standard does not change when the district court’s ruling is unexplained. FDIC v. Ogden Corp., 202 F.3d 454, 460 (1st Cir.2000) (“Although a lower court’s elucidation of its reasoning invariably eases the appellate task, motions often are decided summarily.... [W]e are aware of no authority that would allow us automatically to vary the standard of review depending on whether a district court has taken the time to explain its rationale.”). Here, the privilege issue as decided by the district court largely turns on a legal question that we review de novo.

There is a threshold issue about whether to apply federal or state privilege law. The Yicinis assert, and the filmmakers do not contest, that federal law applies. The basis for that position seems to be that the disputed documents are relevant to actual malice, a requirement gleaned from the federal constitution. See N.Y. Times, 376 U.S. at 279-80, 84 S.Ct. 710. We doubt that the parties’ choice of law is correct: defamation is a state cause of action, as the Vicinis recognized when they brought this suit in diversity, and by rule privilege issues arising out of such actions are to be determined “in accordance with State law.” Fed.R.Evid. 501; Herbert v. Lando, 441 U.S. 153, 182, 99 S.Ct. 1635, 60 L.Ed.2d 115 (1979) (Brennan, J., dissenting) (recognizing that state law governs privilege claims in defamation suits brought in diversity: “Although [New York Times ] placed constitutional limits on state libel claims, it did not itself create a federal cause of action for libel. The ‘rule of decision’ in this case, therefore, is defined by state law.”).

But in this instance we need not disturb the parties’ choice. When the parties agree on the substantive law that should govern, “we may hold the parties to their plausible choice of law, whether or not that choice is correct.” Perry v. Blum, 629 F.3d 1, 8 (1st Cir.2010); see Moores v. Greenberg, 834 F.2d 1105, 1107 n. 2 (1st Cir.1987) (noting that when the parties agree on what substantive law controls, a federal court “ordinarily should” honor the agreement). Here, all parties indicate, at least implicitly, that federal law controls, and they have briefed and argued the privilege question on that basis. In all events, the parties have not identified any material difference between federal and state (presumably Massachusetts) law on this point, nor has our research found any. We think it practical under these circumstances to abide by the parties’ choice and decide the question under federal law.

We begin with the basics. By safeguarding communications between attorney and client, the privilege encourages disclosures that facilitate the client’s compliance with law and better enable him to present legitimate arguments when litigation arises. United States v. Mass. Inst. of Tech., 129 F.3d 681, 684 (1st Cir.1997); see Upjohn v. United States, 449 U.S. 383, 389, 101 S.Ct. 677, 66 L.Ed.2d 584 (1981). The privilege is not limitless, however, and “courts must take care to apply it only to the extent necessary to achieve its under *24 lying goals.” In re Keeper of Records (Grand Jury Subpoena Addressed to XYZ Corp.), 348 F.3d 16, 22 (1st Cir.2003). In other words, “the attorney-client privilege must be narrowly construed because it comes with substantial costs and stands as an obstacle of sorts to the search for truth.” Id. (citing United States v. Nixon, 418 U.S. 683, 709-10, 94 S.Ct. 3090, 41 L.Ed.2d 1039 (1974)).

The contours of the privilege are reasonably well honed. It protects “only those communications that are confidential and are made for the purpose of seeking or receiving legal advice.” Id.; see also In re Grand Jury Subpoena (Mr. S.), 662 F.3d 65, 70 (1st Cir.2011) (outlining necessary prerequisites to the privilege); 8 John H. Wigmore, Evidence § 2292, at 554 (John T. McNaughton rev. 1961). That protection ceases, or is often said to be “waived,” when otherwise privileged communications are disclosed to a third party. Mass. Inst. of Tech., 129 F.3d at 684. The rationale is that such disclosure “destroys the confidentiality upon which the privilege is premised.” In re Keeper of Records, 348 F.3d at 22; see generally 2 Paul R. Rice, Attorney-Client Privilege in the U.S. § 9:79, at 357 (2d ed. 1999). The party invoking the privilege must show both that it applies and that it has not been waived. In re Keeper of Records, 348 F.3d at 22.

There is a possible extension of the privilege when a third party helps the lawyer give legal advice. See United States v. Kovel, 296 F.2d 918 (2d Cir.1961) (Friendly, J.). In Kovel, the Second Circuit held that third-party specialists, such as accountants, hired to assist lawyers in complicated matters, cannot be compelled to testify about client confidences. Id. at 919. The court reasoned that “the presence of an accountant, whether hired by the lawyer or by the client, while the client is relating a complicated tax story to the lawyer, ought not destroy the privilege” because “the accountant is necessary, or at least highly useful, for the effective consultation between the client and the lawyer which the privilege is designed to permit.” Id. at 922. 20

The filmmakers try to fit the Bardsley documents into the Kovel mold. They argue that Bardsley was “necessary, or at least highly useful” because Leopold required her report in order to analyze and assess the legal risks associated with the film. That may be true (though a point we do not resolve here), but it is not sufficient. The key, it seems to us, involves considering the source and nature of the information contained in the documents. If the communication contains only client confidences made in pursuit of legal advice — or legal advice based on such client confidences — that communication, if intended to remain confidential, should be covered by the privilege, regardless of whether it came from the client, his attorney, or an agent of either one. 21 If, however, the transmitted information consists *25 largely of facts acquired from non-client sources, those facts are not privileged. 22 See generally In re Grand Jury Subpoena (Mr. S.), 662 F.3d at 71 (remarking that real estate transaction documents generated by attorney are ordinarily non-confidential by nature and generally do not involve the giving of legal advice).

On the record before us, at least for the most part, we seem not to be dealing with client confidences. According to the filmmakers, Bardsley functioned as a type of fact-checker of the assertions in the film. The very nature of that role would require her to verify those assertions with outside sources, and the filmmakers do not claim otherwise. Insofar as the filmmakers’ script itself is concerned, it is not confidential but is intended for public consumption. And there is evidence that Bardsley’s report — -an annotated version of the script following her fact-check — was designed to be disclosed to third parties (prospective insurers), even if it ultimately was not. Indeed, a contemporaneous document produced by the filmmakers indicates that they hired Bardsley because “Fred [Leopold] needs a 3rd party report to put in front of insurance co. [sic].”

We acknowledge that the face of Bardsley’s report contains declarative markings of confidentiality, and evidence indicates that the electronic version was password protected. Yet deposition testimony fairly displays the filmmakers’ tacit acknowledgment that they had no expectation that Leopold would keep Bardsley’s annotated script confidential, even if his actual disclosure of it was never brought to bear by prospective insurers. In short, the filmmakers have not met their burden of showing that all of the materials they sought to protect come within the Kovel rubric.

Turning briefly to the email communications between the filmmakers and Bardsley, a less clear answer emerges. Simply put, the record does not provide much help on the nature of these communications. It may be that Bardsley in large part was obtaining the filmmakers’ sources in order to conduct her independent evaluation, or perhaps she was providing the filmmakers updates on her progress. While it is questionable whether such information would fall within the privilege, it is at least conceivable that some of the email communications may involve client confidences to some degree.

The Yicinis rest heavily on the apparent business purpose of the filmmakers’ relationship with Leopold, and Bardsley’s involvement to further that relationship, to argue that no privilege attaches to any of the sought after documents. Yet it is not entirely clear from this record that such purpose did not incorporate a meaningful attorney-client privilege component. See generally In re Grand Jury Subpoena (Mr. S.), 662 F.3d at 72 n. 3 (noting that *26 while documents generated by attorney facilitating a real estate transaction may not typically fall within the privilege, “[i]t takes little imagination to- conceive instances in which a particular communication regarding a real estate closing may satisfy all of the requirements of the attorney-client privilege”). Notably, early on in the litigation the district court displayed an initial instinct that live testimony may be required for proper court resolution of the privilege question in this case, and possible in camera review was preliminarily discussed. While we understand that the filmmakers bore the burden of providing sufficient evidence to establish privilege, our careful review of the procedural history of this case — a tale we need not detail here^ — counsels against rendering a decision based on this legal stance. See id. (“a blanket assertion of privilege is generally insufficient”).

One final word on the substance. The doctrine construing the attorney-client privilege narrowly seems to favor production in this instance. That doctrine strikes us as particularly applicable in defamation cases, such as this one, involving public figures. Cf. Lando, 441 U.S. at 175-76, 99 S.Ct. 1635; Bruno & Stillman, 633 F.2d at 594-97. Actual malice must be proven with “convincing clarity,” N.Y. Times, 376 U.S. at 285-86, 84 S.Ct. 710, and this same standard applies whether the matter is resolved on summary judgment or at trial, Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 244, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). Mindful of this hefty burden, upholding the district court’s decision on this record to withhold the sought documents which seemingly bear directly on state of mind would be incompatible with the “search for truth.” Nixon, 418 U.S. at 710, 94 S.Ct. 3090.

This conclusion prevents us from reaching the issue of actual malice. Whatever documents must be produced, the trial judge did not examine them when he granted summary judgment in favor of the filmmakers on that issue. And because the documents have not been submitted to us, we cannot determine whether summary judgment was -warranted despite them. We are cognizant, nevertheless, that even should some documents or portions thereof be disclosed in the end, some of the district court’s rulings on the seven putative defamatory statements may still stand. And yet we deem it is unwise to embark on a piecemeal approach to these statements at this juncture. On remand, the actual malice issue will have to be readdressed should any documents be disclosed.

We recognize the possibility that some documents, or portions of some documents, may contain information that is privileged under the framework set forth above. Rather than risk disclosure of such information, the district court has the option within its discretion on remand to review these documents in camera, allow the filmmakers to withhold any documents covered by privilege, and redact prior to production any portions of admissible documents that it finds are privileged. See In re Grand Jury Subpoena (Mr. S.), 662 F.3d at 70 (emphasizing “prudential purpose” of in camera review to resolve legitimate privilege disputes, and noting that district court may be “well advised” to conduct review even absent an explicit request). The court also may need to consider on remand the Vicinis’ waiver argument pertaining to the sufficiency of the privilege log, and choose to entertain other legitimate arguments relative to disclosure, waiver and privilege which the parties seek to raise.

*27 We underscore that we are not intending to set any precedent for requiring district courts to engage in a particular procedure for discerning the validity of a privilege claim, either in the defamation context or in another type of litigation. Rather, the particular procedural history before the district court and other unique circumstances in this case call us to the present course. See In re Grand Jury Subpoenas, 123 F.3d 695, 699-700 (1st Cir.1997) (remanding for in camera review where record lacked sufficient information for appellate court to form a reasoned judgment whether requested records were privileged).

III. CONCLUSION

For these reasons, we affirm the limited purpose public figure status determination but otherwise vacate the dispositive judgment, vacate the denial of the motion to compel insofar as the Bardsley documents are concerned, and remand for further proceedings consistent with this opinion. We take no position on the actual-malice issue. The parties shall bear their own costs.

So ordered.