POSNER, Chief Judge.
This appeal from the grant of a preliminary injunction in a $100 million dollar suit over insurance coverage raises difficult and important questions concerning the power of a federal district judge to enjoin a party before it from litigating a suit in a foreign country. There is also an esoteric jurisdictional question.
Groupe Bull is one of Europe’s largest manufacturers of computers and computer software. The parent company in the Groupe, Compagnie des Machines Bull (CMB), is a French corporation 90 percent of whose stock is owned by the French state. In 1989, CMB bought Zenith Data Systems (ZDS), the former microcomputer business of Zenith Electronics Corporation. In anticipation of this acquisition, CMB’s U.S. subsidiary, Bull Data Systems, Inc. (BDS), obtained from Allendale Mutual Insurance Company worldwide property insurance coverage for Groupe Bull. Allendale is a U.S. company and the insurance contract was negotiated and signed in the United States, having been obtained for BDS by an American broker, Alexander & Alexander.
After acquiring Zenith Data Systems, Groupe Bull decided to consolidate its European inventory of microcomputers in a leased warehouse in Seclin, France. BDS added the Seclin warehouse to the Allendale insurance policy as a specifically scheduled location. Later either it or Allendale decided it would be good to have an insurance policy governed by the French insurance code for the contents of the Seclin warehouse and of any other French locations at which Groupe Bull’s property was stored. So Allendale had Factory Mutual International (FMI), a British subsidiary of Allendale authorized to write French insurance policies, write a [427]*427French policy for the French locations; but it appears that all the negotiations for this policy were conducted between BDS and Al-lendale in the United States. The Allendale and FMI policies overlap, as the former provides worldwide coverage and was never rewritten to exclude French sites.
On June 15, 1991, a fire at the warehouse destroyed the huge inventory of microcomputers, valued by BDS (as we shall refer to the defendants collectively) at some $100 million. In August, BDS filed its claim of loss, under both policies, with Allendale and FMI. The next month, Allendale and FMI filed this lawsuit in the federal district court in Chicago. The suit asked for a declaration that the cause of the fire was arson committed by the insured, which was financially troubled, and hence that the damage from the fire was excluded from coverage; but that if this was wrong and the fire damage was covered, the insurer’s’ liability was limited to $48 million, the limit in FMI’s policy. BDS responded by filing its own suit in the district court, against Allendale and Alexander & Alexander (the broker) but not against FMI, and a suit in the Commercial Court of Lille, France, against FMI alone; under French law, that court may have exclusive jurisdiction over suits to enforce insurance policies governed by the French insurance code, although this is not entirely clear. Judge Ma-rovich consolidated the two U.S. suits (Allen-dale-FMI’s and BDS’s) and BDS filed counterclaims against Allendale, FMI, and Alexander & Alexander in the consolidated suit.
Meanwhile, in France, Allendale (and FMI — but we shall refer to the pair as “Al-lendale” except where their separate identities are relevant) had pressed for a criminal investigation of the fire. The matter had been assigned to an examining magistrate {juge d’instruction). Allendale asked the Commercial Court to stay its proceeding pending the completion of the criminal investigation, and that court agreed to do so, over BDS’s objection.
Discovery now began in the consolidated suit and proceeded on an expedited basis, generating hundreds of depositions and hundreds of thousands of documents. Most of the discovery requests were by Allendale and many of them were aimed at obtaining evidence of arson. In February 1993, with the district court suit moving rapidly toward trial, BDS unexpectedly filed a motion in the Commercial Court of Lille to lift the stay and proceed to trial in that court, even though the examining magistrate had not yet concluded her investigation. BDS argued that the investigation was on the verge of completion and that the examining magistrate would conclude that there had been no arson. The timing of the motion is still peculiar — if the end was so imminent, why not wait for it rather than speculate about it?' — and eight months later the investigation is still not over. Allendale asked Judge Marovich to issue a preliminary injunction against BDS’s litigating its case in the Commercial Court, the injunction to remain in effect until the consolidated suit in the district court eventuates in a final judgment. At that point,-Allendale intends to ask the judge, if it is a judgment favorable to Allendale, to make the injunction permanent. Apparently the motion to lift the stay remains pending before the Commercial Court of Lille and has not been acted upon.
Judge Marovich issued the preliminary injunction, precipitating this appeal. Before we consider its merits, we must satisfy ourselves that the district court had jurisdiction of the case. There is no problem with regard to the suit against the two Zenith entities. The Judicial Code, in 28 U.S.C. § 1330(a), confers federal jurisdiction over civil suits against foreign states as defined by the Foreign Sovereign Immunities Act, and the Act’s definition includes companies a majority of whose shares are owned by a foreign state. 28 U.S.C. § 1603(b)(2). But there is an exception for companies that are also citizens of a state of the United States, § 1603(b)(3), and BDS is a corporate citizen of Delaware and Illinois — so what is the basis for jurisdiction over it? The district court said diversity of citizenship, but this runs into the problem that one of the plaintiffs, FMI, is a foreign company, so that foreigners are on both sides of the litigation. We must consider whether this destroys the complete diversity of citizenship that is a prerequisite to maintaining a suit under the diversity [428]*428jurisdiction. Strawbridge v. Curtiss, 7 U.S. (3 Cranch) 267, 2 L.Ed. 435 (1806); Fidelity & Deposit Co. v. City of Sheboygan Falls, 713 F.2d 1261, 1264-68 (7th Cir.1983). Many cases, such as Spearing v. National Iron Co., 770 F.2d 87, 90 (7th Cir.1985); Corporacion Venezolana de Fomento v. Vintero Sales Corp., 629 F.2d 786, 790 (2d Cir.1980), and IIT v. Vencap, Ltd., 519 F.2d 1001, 1015 (2d Cir.1975), state that the presence of foreign parties on both sides of a litigation indeed destroys complete diversity. Yet the statement is puzzling. The presence of citizens of different states on both sides of a lawsuit obviously does not destroy diversity; it is the precondition of diversity. So why should the presence of citizens of foreign states destroy diversity unless (as in Spearing) they are citizens of the same foreign state? The answer is found in the details of the statutory framework. The cases in which the statement appears are ones in which one side of the litigation had only foreign parties and the other had a mixture of foreign and domestic parties, so that the case did not fit any of the possibly applicable jurisdictional pigeonholes: 28 U.S.C. § 1332(a)(2) (suits between “citizens of a State and citizens or subjects of a foreign state”), § 1332(a)(3) (suits between “citizens of different States and in which citizens or subjects of a foreign state are additional parties”), or § 1332(a)(4) (suits between foreign states and citizen defendants). The point was not so much that there were foreigners on both sides — for this is permitted by (a)(3) — as that there was no citizen on one side, which took it out of (a)(3); and (a)(2), when read in light of (a)(3), does not permit a suit between foreigners and a mixture of citizens and foreigners. Eze v. Yellow Cab Co., 782 F.2d 1064 (D.C.Cir.1986) (per curiam). Exactly what sense all this makes rather eludes us. But we need not worry about the matter here. A case such as this, in which citizens of states are on both sides of the litigation (Allendale and Alexander & Alexander on one side and BDS on the other), and are completely diverse, fits section 1332(a)(3) to a t. Transure, Inc. v. Marsh & McLennan, Inc., 766 F.2d 1297, 1298-99 (9th Cir.1985). This is plain, but worth stating because it is sometimes overlooked. See Field v. Volkswagenwerk AG, 626 F.2d 293, 295-96 (3d Cir.1980), and cases cited in Gary B. Born & David Westin, International Civil Litigation in United States Courts: Commentary and Materials 411-12 (1991).
We come to the merits, where at first glance the action of an American judge in enjoining what is practically an arm of the French state — what in the case of the Zenith entities American law says is an arm of the French state (while allowing the arm to be sued in the courts of the United States because engaged in a commercial activity, 28 U.S.C. §§ 1605, 1607; Rush-Presbyterian-St. Luke's Medical Center v. Hellenic Republic, 877 F.2d 574, 578, 580 (7th Cir.1989))— from litigating a suit on a French insurance policy in a French court may seem an extraordinary breach of international comity. But the matter is more complex. To begin with, although the French government is the majority stockholder in the parent company of Groupe Biill, even BDS does not argue that foreign commercial enterprises deserve more solicitude from American judges when they are public rather than private. There is no suggestion that French officials are mixed up in the alleged arson. Although it is common knowledge that the French government has tried to foster a French computer industry, BDS makes no claim that the government’s role in Groupe Bull is more than that of a passive investor. So we can lay the ownership structure of the defendants on one side and decide the case as if BDS were an entirely private French company (as it may soon be, in light of the French government’s privatization program).
Until February 1993, BDS seemed content if not necessarily ecstatic to litigate its entire dispute with Allendale and FMI in the Northern District of Illinois. All the parties were before the court and an enormous amount of discovery had been conducted, while the parallel suit in France — from which an important party, Allendale, was missing because the French suit was only against FMI — had since its inception been on hold because of the criminal investigation, which was proceeding at a stately pace and is now a year and a half old. Although the filing of Allendale’s and FMI’s suit would have forced [429]*429BDS to file its claims against the insurers as counterclaims, lest failure to do so bar it under the compulsory-counterclaim rule from pursuing the claims in a separate suit, it did more: it filed a separate suit in the Northern District of Illinois against Allendale (but not FMI) and joined the broker — which had not joined in Allendale’s suit and, more to the point, which BDS had not attempted to join as a defendant before the Commercial Court of Lille — -as an additional defendant. The decision to file a U.S. suit against Allendale and the broker but not FMI, in tandem with a French suit against FMI alone, indicates concern on BDS’s part that it might not get complete relief in the French suit, even though that suit would be tried on home turf, because of the $48 million limit in FMI’s policy. Once FMI sued it in the United States, however, BDS could hope for complete relief there by adding a counterclaim against FMI to its claim against Allendale.
Why in these circumstances did BDS decide to try to reactivate the French suit? The district judge believed, not without basis, that the reason was the evidence of arson which pretrial discovery in the consolidated suit had uncovered. The evidence was far from conclusive. But there was enough of it to give Allendale a fighting chance of persuading a finder of fact by the required preponderance of the evidence that, astounding as it might seem, BDS had torched an immense inventory of microcomputers because the enterprise was in financial straits, the inventory was obsolete (maybe that was why it was so large — it was unsalable), and its destruction would enable BDS to impress its creditors by writing up the value of the inventory pending receipt of the insurance proceeds (based on that inflated value) from Allendale.
BDS vigorously denies both that it committed arson and that it had any motive to do so. More to the point — for those are issues to be sorted out by a trier of fact — it argues that if Allendale can prove arson to an American jury it can prove it to the Commercial Court of Lille just as well, and, after all, the fire occurred in France and most of the evidence concerning it is in French, so why shouldn’t the case be tried there?
To answer this question we must consider the institutional characteristics of the Commercial Court of Lille, Although called a “court,” it is actually a panel of arbitrators, composed of businessmen who devote part time to arbitrating. According to an affidavit of a French legal expert that the district judge was entitled to and did credit, the Commercial Court of Lille rarely if ever hears live witnesses. No problem, argues BDS; just ship the court the hundreds of depositions and hundreds of thousands of documents obtained in pretrial discovery in the district court. But it is difficult enough for courts staffed with legal professionals to cope with massive documentary records; it borders on the inconceivable that businessmen serving as part-time arbitrators could do so. As far as we are aware, the members of the Commercial Court do not have masters, magistrates, law clerks, externs, or other staff that might enable them to assimilate the voluminous materials that have been collected in the district court. Much of this stuff could no doubt be dispensed with in a trial with witnesses, but if there are to be no witnesses and the arbitrators are to be remitted to rummaging through deposition transcripts and other massive documentation, we do not see how Allendale could get them to consider its defense of arson.
It is true that juries have difficulty assimilating massive documentation, too, but in a jury trial the documents normally play a secondary role to live testimony. It is also true that parties to commercial agreements often agree that disputes arising under such agreements will be arbitrated, and such arbi-trations can involve complex issues, for example of antitrust law, as in Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 105 S.Ct. 3346, 87 L.Ed.2d 444 (1985). See also Shearson/American Express, Inc. v. McMahon, 482 U.S. 220, 107 S.Ct. 2332, 96 L.Ed.2d 185 (1987). So nothing in the character of arbitration excludes competence to handle big-record cases. But the parties to this case did not agree to the arbitration of their disputes, whether by the Commercial Court of Lille or anyone else. (The FMI insurance policy does specify that it shall be governed by the French insurance [430]*430code, but there is no designation of the forum to resolve disputes between the parties arising under the policy.) Nor, whatever may be true of some arbitral tribunals, does it appear that the Commercial Court of Lille is equipped to resolve a massive document case. An obvious and perhaps critical difference between it and other arbitral bodies is that most arbitrators hear live testimony; the Commercial Court apparently does not.
Allendale argues plausibly that the only thing that would persuade the Commercial Court to consider the defense of arson (unless there is a confession) is if the French authorities prosecute and convict BDS or its agents for arson. If BDS succeeds in getting the Commercial Court to adjudicate its claim against FMI before any such conviction — and none is imminent (for even if the examining magistrate in France finds probable cause to believe that BDS committed arson, this would be the equivalent only of an indictment, and BDS would be entitled to a trial)— it will prevail on the vital issue of arson by undeserved default. After the oral argument in this court, BDS’s lawyers submitted a French judicial opinion, rendered in 1987, which authorizes a civil court to make a finding of arson even if the examining magistrate declines to recommend criminal prosecution. The fire was in a discotheque, not a warehouse containing a $100 million inventory, and the lower court is not identified, so we don’t know whether it was the Commercial Court of Lille or some other arbitral body rather than a court staffed by professional judges.
In contrast to the Commercial Court of Lille, there is no doubt that the Northern District of Illinois is a suitable forum for the trial of this dispute. The principal insurance policy was written in the United States following negotiations among a U.S. insurer, a U.S. broker, and a U.S. insured (BDS). Chicago is BDS’s headquarters, and Allendale’s side of the negotiations was handled in Allen-dale’s Chicago office. It is not an ideal forum, because the principal issue is the cause of a fire that occurred in France. But precisely because of the importance that the arson issue has assumed in the parties’ dispute, and, critically, the institutional differences between a federal district court and the Commercial Court of Lille, the district court provides a more appropriate forum for the resolution of the parties’ dispute. There will be no prejudice to BDS from being forced to go to trial in the district court, but there would be clear prejudice to Allendale from being forced to go to trial in the Commercial Court, simply because of the way in which the Commercial Court is constituted in relation to the shape that the litigation has taken. This conclusion has nothing to do with the relative merits of the French and American procedural systems, an issue on which it would be impertinent for us to express a view. We have arbitral bodies, some with exclusive jurisdiction, such as the National Railroad Adjustment Board ordained by the Railway Labor Act, 45 U.S.C. § 153, and the French have courts staffed by professional judges. We can imagine a mirror-image case in which a French court was asked to enjoin an American firm from proceeding in the National Railroad Adjustment Board because that board was not equipped to do justice between the parties in the particular circumstances of their dispute.
When a federal court is asked to abstain in favor of a parallel litigation pending in another court, the presumption is against abstention, Colorado River Water Conservation District v. United States, 424 U.S. 800, 813, 96 S.Ct. 1236, 1244, 47 L.Ed.2d 483 (1976); New Orleans Public Service, Inc. v. Council of City of New Orleans, 491 U.S. 350, 359, 109 S.Ct. 2506, 2513, 105 L.Ed.2d 298 (1989); cf. Lynk v. LaPorte Superior Court, 789 F.2d 554, 559 (7th Cir.1986) — the general rule being that a federal court has a duty to exercise the jurisdiction that Congress has given it. In a case in which the presumption is reinforced rather than rebutted by practical considerations bearing on the choice of which court to proceed in — how far advanced the litigation is in one court and which court can by virtue of its jurisdiction or structure render the more complete justice — abstention would be out of the question. But that means that if Judge Marovieh had refused to grant the preliminary injunction requested by Allendale, and the Commercial Court had lifted its stay, these parallel lawsuits would be proceeding full tilt in [431]*431two tribunals 4,000 miles apart at the same time, entailing an absurd duplication of effort. This is an argument in favor of the issuance of the injunction, although an alternative would be to hope that the other tribunal would stay the case before it (or in this case refuse to lift a stay already granted).
There is nothing to the counterargument that Judge Marovich should have dismissed Allendale’s suit altogether because the suit sought a declaratory judgment, characterized by BDS as a “preemptive strike.” What is true is that because the issuance of a declaratory judgment is discretionary, Tempco Electric Heater Corp. v. Omega Engineering, Inc., 819 F.2d 746, 747 (7th Cir.1987), a suit for declaratory judgment aimed solely at wresting the choice of forum from the “natural” plaintiff will normally be dismissed and the case allowed to proceed in the usual way. Id. at 749-50. But the forum in question, the Northern District of Illinois, was chosen by the insured as well as by the insurer.
All the considerations bearing on the grant of preliminary relief support Judge Maro-vich’s order — with the possible exception of international comity. The weight of this factor has now to be assessed. A recent opinion of this court canvasses the precedents, Philips Medical Systems International B.V. v. Bruetman, 8 F.3d 600, 604-05 (7th Cir.1993), expressing an inclination toward the “laxer” standard, which allows an injunction against litigating in a foreign forum upon a finding that letting the two suits proceed would be gratuitously duplicative, or as the cases sometimes say “vexatious and oppressive.” E.g., Seattle Totems Hockey Club, Inc. v. National Hockey League, 652 F.2d 852, 855-56 (9th Cir.1981). Courts of equity have long issued injunctions against the use of litigation, including litigation in foreign courts, not to obtain a decision on the merits but to harass a party, 2 Joseph Story, Commentaries on Equity Jurisprudence §§ 1224-27 at pp. 578-83 (14th ed. 1918); Born & Westin, supra, at 242; Asset Allocation & Management Co. v. Western Employers Ins. Co., 892 F.2d 566, 571 (7th Cir.1989); see Cole v. Cunningham, 133 U.S. 107, 116-121, 10 S.Ct. 269, 272-74, 33 L.Ed. 538 (1890), at least when the party enjoined is a resident of the enjoining jurisdiction, id. at 119, 10 S.Ct. at 273, which is approximately true here given BDS’s role in the procurement of the insurance and in the litigation. But the courts that follow the stricter standard believe that considerations of international comity require that the domestic and foreign cases be distinguished and that a greater showing of need be required for an injunction in the latter type of case. E.g., Laker Airways Ltd. v. Sabena, 731 F.2d 909 (D.C.Cir.1984).
When we say we lean toward the laxer standard we do not mean that international comity should have no weight in the balance; we do not interpret the “lax” cases as assigning it no weight. The difference between the two lines of case has to do with the inferences to be drawn in the absence of information. The strict cases presume a threat to international comity whenever an injunction is sought against litigating in a foreign court. The lax cases want to see some empirical flesh on the theoretical skeleton. They do not deny that comity could be impaired by such an injunction but they demand evidence (in a loose sense — it needn’t be evidence admissible under the Federal Rules of Evidence) that comity is likely to be impaired in this case. When every practical consideration supports the injunction, it is reasonable to ask the opponent for some indication that the issuance of an injunction really would throw a monkey wrench, however small, into the foreign relations of the United States. A representation by the State Department would be one method of conveying such information. A representation by the foreign office of France, or by the French agency that administers the French government’s investment in BDS, would be another. No doubt there are others, and the Commission de Controle des Assurances, the French agency that regulates the insurance business in that country, has submitted a brief amicus curiae urging reversal. It urges the interests of French insureds in being able to litigate their claims before a French tribunal, and it defends the competence of the Commercial Court of Lille against what it describes as the “insulting” assessment of that competence by the district judge. We do not question the competence of that court, only [432]*432its capacity relative to a U.S. district court to resolve this particular dispute given the unusual turn that the litigation has taken, an issue that the amicus curiae brief does not discuss. We are given no indication, moreover, that the French commission is authorized to speak for the French state.
Just as we don’t think the “lax” cases would refuse to consider tangible evidence of a threat to comity, so we don’t think the “strict” cases would refuse to weigh against such a threat substantial U.S. interests. Groupe Bull is French, but Allendale is American, and the United States has an interest in protecting its citizens, including its corporate citizens, from trumped-up multimillion dollar claims. In particular it has an interest, well recognized in American insurance law (including the insurance law of Illinois, where the policy was negotiated), in providing a forum that will enable an insurer to establish by a preponderance of the evidence (unless, as in Gulliver’s East, Inc. v. California Union Ins. Co., 118 Ill.App.3d 589, 74 Ill.Dec. 234, 455 N.E.2d 264 (1983), it agrees to bear a heavier burden, and that is not argued here) that it is being victimized by a fraudulent claim, 18 Couch on Insurance 2d §§ 74:663, 74:667, at pp. 976, 982-83 (2d ed. 1983); Schultz v. Republic Ins. Co., 124 Ill.App.3d 342, 79 Ill.Dec. 863, 866, 464 N.E.2d 767, 770 (1984); 720 ILCS 5/20-1(b), in a situation in which a foreign forum for the presentation of such evidence is, as a practical matter, unavailable. So even if BDS had brought a “note from home” (the relevant home being the Quai d’Orsay), and brought it to a court that takes the strict view, it is far from cleat' that it would prevail. Cf. Laker Airways Ltd. v. Sabena, supra, 731 F.2d at 932.
It cannot help — -it can only hurt— BDS that French law may (we have not been able to discover whether it does) make the jurisdiction of the Commercial Court of Lille to enforce insurance policies governed by the French insurance code exclusive; and not only because there are two insurance policies, only one of which (the FMI policy) is governed by that code. Within a single jurisdiction, the specification of a particular court as having exclusive jurisdiction over some class of disputes is conclusive. But a state or nation cannot, by designating its own courts as the exclusive fora for the resolution of the class, prevent another state or nation from allowing its own courts to resolve these disputes if the other state or nation has an interest in them, as the United States does here by virtue of the citizenship of Allendale. Anyway BDS waived its objection to the jurisdiction of the Northern District of Illinois when it filed its counterclaim against FMI in that court without asserting that the court lacked jurisdiction. Restatement (Third) of the Foreign Relations Law of the United States § 421(3) (1987). Indeed, in its brief in this court BDS argues that the Commercial Court of Lille does not have exclusive jurisdiction, because they are worried that if it did this would strengthen the case for the preliminary injunction. For it might then be argued that the injunction was necessary to protect the federal district court’s jurisdiction, a situation in which Laker itself suggests that comity must yield. Laker Airways Ltd. v. Sabena, 731 F.2d at 927.
It is true that if the FMI policy were construed as expressing FMI’s consent to resolve any dispute arising under the policy in the Commercial Court of Lille, an American court would, under the Supreme Court’s foreign-arbitration cases cited earlier, be obliged to relinquish jurisdiction in favor of that court. But BDS does not take this position. Indeed, as we have been at pains to stress, until discovery turned up evidence of arson, BDS was content to litigate its dispute with FMI as well as Allendale in the Northern District of Illinois.
If Allendale were seeking an order that would run against a foreign official or agency, cf. 28 U.S.C. § 2283, or that would impede a foreign criminal prosecution, cf. Younger v. Harris, 401 U.S. 37, 91 S.Ct. 746, 27 L.Ed.2d 669 (1971) — but it was Allendale that activated the foreign criminal proceedings — there would be no need for evidence that the antisuit injunction would ruffle the smooth surface of our relations with France. There is nothing of that sort. The injunction merely prevents a French company from seeking to revive a dormant proceeding before an arbitral tribunal in France. The only [433]*433concern with international comity is a purely theoretical one that ought not trump a concrete and persuasive demonstration of harm to the applicant for the injunction, if it is denied, not offset by any harm to the opponent if it is granted.
We have so far taken for granted that whatever complications are injected by the international character of this litigation, the relief that Allendale is seeking is within the power of a court of equity to grant. We think it is, although the question is not free from doubt. Allendale is seeking a preliminary injunction, and it is preliminary, we said, to a permanent injunction, should Allen-dale prevail on the merits of its dispute over insurance coverage, against BDS’s prosecuting its French suit against FMI. So we must consider whether such an injunction would be within the power of the district court to issue. Ordinarily a plaintiff who obtains a final judgment in a mirror-image suit uses the judgment as the basis for a plea of res judicata in the parallel proceeding. If Allendale obtained a judgment that its and FMI’s policies do not cover the fire loss at Seclin, it would interpose the judgment in BDS’s suit against FMI in the Commercial Court of Lille. But there is no assurance that that court — -that any French tribunal— would accord res judicata effect to a foreign judgment against a foreign national (BDS’s French codefendants — the Zenith entities that are members of Groupe Bull). See Restatement, supra, at 591-92, 602; Thomas E. Carbonneau, “The French Exequatur Proceeding: The Exorbitant Jurisdictional Rules of Articles 14 and 15 (Code Civil) as Obstacles to the Enforcement of Foreign Judgments in France,” 2 Hastings Int’l & Comp. L.Rev. 307, 316-29 (1979). In these circumstances, where inability to plead res judicata might deprive Allendale of the benefit of its judgment, considerations of comity would not prevent a federal court, even under the strict cases, from enjoining the foreign defendants (and even more clearly a nominally American one, BDS) from proceeding, in defiance of the judgment, in a foreign court. Laker Airways v. Sabena, supra, 731 F.2d at 926-27; Gau Shan Co. v. Bankers Trust Co., 956 F.2d 1349, 1352 (6th Cir.1992); 2 Story, supra, § 1227 at pp. 582-83; cf. Ingersoll Milling Machine Co. v. Granger, 833 F.2d 680 (7th Cir.1987).
So Allendale, if it prevails on the merits of its suit, will have a plausible case for a permanent injunction. And while its chances of prevailing are — such is the factual uncertainty of the arson question — at present unknowable, this is immaterial to the question of a preliminary injunction in a case when denial of the preliminary injunction would cause irreparable harm to the movant by making Allendale proceed in two courts at once and the grant of it would cause no irreparable harm to the nonmovant. BDS, by moving to lift the stay in France but not to obtain a stay in Chicago, has signified its preference that the case proceed on these parallel tracks. Indeed this has been its strategy from the start, for we recall that it wanted to litigate against FMI in France and against Allendale in Chicago.
With so one-sided a balance of irreparable harm, even a slight showing of possible merit is enough to justify preliminary relief. Roland Machinery Co. v. Dresser Industries, Inc., 749 F.2d 380, 386-88 (7th Cir.1984); American Hospital Supply Corp. v. Hospital Products Ltd., 780 F.2d 589, 593 (7th Cir.1986); Green River Bottling Co. v. Green River Corp., 997 F.2d 359, 361 (7th Cir.1993). If material circumstances, concerning for example the French criminal proceeding, change, BDS can of course seek the modification or dissolution of the preliminary injunction.
AFFIRMED.