In 1979 the Saturday Evening Post Company granted Rumbleseat Press, Inc. an *1193exclusive license to manufacture porcelain dolls derived from certain illustrations done by the artist Norman Rockwell and published in the Saturday Evening Post. Later the Post (as we shall call the company and the magazine interchangeably) cancelled the license; but, contrary to the terms of the license agreement, Rumbleseat continued making the dolls. This conduct led the Post (and its parent corporation, but we can ignore this detail) to sue Rumbleseat. The Post won in the district court. Rumbleseat’s appeal raises a variety of jurisdictional and procedural issues, of which the most important are whether the validity of a copyright is arbitrable and whether a provision in a copyright license that forbids the licensee to challenge the validity of the copyright is enforceable. Both are novel issues.
The original term of Rumbleseat’s license was three and a half years. Renewal for successive years was automatic, however, unless a party gave notice at least 90 days before the end of the year that it did not want to renew. The license agreement specifies that once the license is cancelled Rumbleseat has to stop making or selling the licensed goods, except that it has 275 days after cancellation to liquidate its existing inventory. The agreement also contains a warranty by the Post that it has valid copyright in the Rockwell illustrations. Actually, the Post had copyrighted each of the magazines in which the illustrations appeared but had not copyrighted the illustrations separately, and in the license negotiations Rumbleseat had questioned whether the Post had valid copyrights in the illustrations.
Rumbleseat on its part promised in the agreement to register in the Post’s name any copyrights that the Post reasonably deemed necessary to protect its own copyrights. Rumbleseat also agreed that if it acquired any copyrights in its own name it would transfer them to the Post when the license ended. Works derived from copyrighted material — “derivative works” as they are called — are copyrightable provided the derivative work has some incremental originality; the copyright in the derivative work is limited to that increment. See 17 U.S.C. §§ 101, 103(b); Gracen v. Bradford Exchange, 698 F.2d 300 (7th Cir.1983); Durham Industries, Inc. v. Tomy Corp., 630 F.2d 905, 909 (2d Cir.1980). The Rockwell illustrations were the original work; the porcelain dolls that the license authorized Rumbleseat to make and sell were derivative works.
Paragraph 9 of the license agreement provides that Rumbleseat “shall not, during the Original Term [of the agreement] or any time thereafter dispute or contest, directly or indirectly, ... the validity of any of the copyrights ... which [the Post] may have obtained.” This is the no-contest clause. The agreement also has an arbitration clause: “any controversy or claim arising out of or relating to this contract, or the breach thereof, shall be settled by arbitration in accordance with the Rules of the American Arbitration Association,” and the arbitrators’ judgment “may be entered in any court having jurisdiction thereof.”
In March 1983 the Post notified Rumble-seat that it intended to terminate the agreement at the end of the year. Termination day came and went, yet Rumbleseat continued making the dolls, and in July 1984 the Post demanded arbitration. The parties engaged in the preliminaries of arbitration for several months (e.g., picking the arbitrators and deciding where the arbitration would be held), but then Rumbleseat told the Post that it was going to try to enjoin the arbitration. That precipitated the Post’s lawsuit, filed in January 1985. The complaint charges copyright violations, and bases federal jurisdiction on 28 U.S.C. § 1338(a), which provides that “the district courts shall have original jurisdiction of any civil action arising under any Act of Congress relating to ... copyrights---Such jurisdiction shall be exclusive of the courts of the states in ... copyright cases.” The relief sought in the complaint includes a preliminary injunction against Rumbleseat’s withdrawing from the arbitration. Rumbleseat counterclaimed, charging breach of contract and copyright infringement.
*1194The district judge treated the Post’s request for a preliminary injunction against Rumbleseat’s leaving the arbitration as a request for an order to arbitrate, and issued the order, refusing Rumbleseat’s demand for a jury trial on the issue of arbitrability. The arbitration was then conducted before three lawyers in Indianapolis, who after hearing evidence handed down an award in which they enjoined Rumbleseat from making or selling procelain dolls modeled on the Rockwell illustrations covered by the license and ordered Rumbleseat to transfer to the Post all copyrights that Rumbleseat had taken out on the dolls and to pay the Post $200,000 in damages. The arbitrators did not indicate the grounds of the award. The Post moved the district court under section 9 of the federal arbitration code, 9 U.S.C., § 9, to confirm the award. The court did so, and certified its order for an immediate appeal under Rule 54(b) of the Federal Rules of Civil Procedure. This was necessary to give us jurisdiction, because Rumbleseat had impleaded other parties in the district court and its claims against them remain pending there.
The initial curiosity is how this case managed to get converted, without anybody’s noticing until we began to question counsel at oral argument, from a copyright infringement suit to a suit under the federal arbitration code. Converted it was; although an order confirming an arbitration award is an unusual terminus to a suit for copyright infringement, that is what it was here. The Post got all the relief from the arbitrators that it had sought in its suit for copyright infringement, and the award also extinguished Rumbleseat’s counterclaim. The district court and the opposing parties have all treated the judge’s order confirming the arbitration award as a final disposition of the dispute between the Post and Rumbleseat.
All this is fine provided the conversion did not deprive the federal district court of jurisdiction. A federal court cannot base its jurisdiction over a suit to enforce an arbitration clause in a copyright license on either 28 U.S.C. § 1338(a) or the general federal-question statute, 28 U.S.C. § 1331. Both statutes are limited to suits that arise under federal law, and a dispute over the terms of a copyright license is not deemed to arise under the Copyright Act (no one suggests that the dispute in this case arose under any other federal law). See T.B. Harms Co. v. Eliscu, 339 F.2d 823, 828 (2d Cir.1964) (Friendly, J.); Weinstein v. University of Illinois, 811 F.2d 1091, 1095-96 (7th Cir.1987); In re Chicago, Rock Island & Pac. R.R., 794 F.2d 1182, 1188 (7th Cir.1986) (dictum); McCall-Bey v. Franzen, 777 F.2d 1178, 1185 (7th Cir.1985) (dictum). Such a dispute is too remote from the federal grant (the copyright). See generally Gully v. First Nat’l Bank, 299 U.S. 109, 57 S.Ct. 96, 81 L.Ed. 70 (1936) (Cardozo, J.). The analogy is to a dispute that pits a tenant against a landlord who acquired the premises from the federal government. Cf. Shulthis v. McDougal, 225 U.S. 561, 569-70, 32 S.Ct. 704, 706-07, 56 L.Ed. 1205 (1912). Of course “remoteness” is just a metaphor, disguising a conclusion that state courts are competent to resolve disputes likely to be dominated by ordinary contract principles, cf. McCall-Bey v. Franzen, supra, 111 F.2d at 1186, even though the subject of the dispute is a right created by federal law and even though in particular cases issues of federal copyright law may come in by way of defense or counterclaim. But the conclusion is a reasonable one.
As noted in McCall-Bey, see 111 F.2d at 1186, and many other cases, see, e.g., Christianson v. Colt Industries Operating Corp., 798 F.2d 1051, 1059-62 (7th Cir.1986), the principle that disputes over the terms of copyright licenses do not arise under federal law has an exception for cases where establishing the plaintiff's right will require interpreting federal law. The exception does not apply to a case such as this, however, in which the question of interpretation is injected by the defendant, here by pleading the invalidity of the plaintiff’s copyrights. For a good discussion see Peay v. Morton, 571 F.Supp. 108, 115-17 (M.D.Tenn.1983). Otherwise the plaintiff — who can’t be sure what defenses the *1195defendant will plead — would not know whether to sue in state or federal court.
That federal jurisdiction depends on the complaint rather than on the answer, counterclaim, or other subsequent pleadings is an aspect of the “well-pleaded complaint” rule, on which see Franchise Tax Bd. v. Construction Laborers Vacation Trust, 463 U.S. 1, 9-12, 103 S.Ct. 2841, 2846-48, 77 L.Ed.2d 420 (1983). The only significance we attach to the exception for cases where the plaintiffs right depends on an interpretation of federal law is the following: if the complaint is actually based on federal law, the plaintiff’s effort to conceal this fact because he wants to prevent the defendant from removing the case to federal court must not be allowed to succeed. See id. at 22,103 S.Ct. at 2852-53. Making sure that plaintiffs do not get away with concealing the federal nature of their claims is particularly important in cases arising under copyright law, because federal jurisdiction over such cases is exclusive. The plaintiff cannot be allowed to defeat the congressional determination merely by omitting mention of that law from his complaint. This is just another facet of the “weZZ-pleaded complaint” rule. We question, however, the suggestion (no more than that) in Merrell Dow Pharmaceuticals Inc. v. Thompson, — U.S. -, 106 S.Ct. 3229, 3236 n. 12, 92 L.Ed.2d 650 (1986), that a case may arise under federal law if it is apparent that issues of federal law predominate, even if federal law is not the basis of the plaintiff’s claim. This seems too uncertain a test for determining whether a federal court has jurisdiction, and also one that often could not be applied at the outset of the litigation; the plaintiff might not know whether he belonged in state or federal court until he received the defendant’s answer or until pretrial discovery or a pretrial conference had clarified the issues for trial.
Still another facet of the well-pleaded complaint rule is that merely naming a federal statute in a complaint will not confer federal jurisdiction if the only relief the plaintiff actually wants is based on state rather than federal law. Cf. The Fair v. Kohler Die & Specialty Co., 228 U.S. 22, 25, 33 S.Ct. 410, 411, 57 L.Ed. 716 (1913) (Holmes, J.); Excelsior Wooden Pipe Co. v. Pacific Bridge Co., 185 U.S. 282, 287-88, 22 S.Ct. 681, 683, 46 L.Ed. 910 (1902). In such a case the invocation of federal-question jurisdiction is in bad faith and must fail. This may or may not be such a case. The Post has never pursued its claim for relief under copyright law; the entire thrust of its case has been to enforce the arbitration clause in the license agreement. In this respect the case is unlike Kamakazi Music Corp. v. Robbins Music Corp., 684 F.2d 228, 230 (2d Cir.1982), where the order to arbitrate was sought by the defendant, who had pleaded a contractual defense to the plaintiff’s suit for copyright infringement. It is possible, however, that when the complaint in this case was filed — the critical time in deciding whether there is federal jurisdiction — the Post was bending a bow with two strings: a main claim based on federal copyright law and a pendent claim under state law for breach of contract, entitling the Post to arbitration. But, given the complaint’s heavy emphasis on arbitration, it might have been advisable for the district judge to inquire at the outset whether the Post was serious about its copyright claims or was instead making a federal law tail wag a state law dog.
We need not pursue these questions. There is an alternative ground for federal jurisdiction even if the suit is conceived to be one solely to compel arbitration. The federal arbitration code, 9 U.S.C. §§ 1 et seq., authorizes federal district courts to stay proceedings before them pending arbitration, to order arbitration, and to confirm or set aside arbitration awards. Although not cited in the Post’s complaint, the code is the apparent basis for the request for an order to arbitrate. It is true that section 4 of the code allows the court to order arbitration only if the court would have jurisdiction of the dispute sought to be arbitrated. Since in this case that dispute is a contract dispute that arises under state law, the only possible basis for jurisdiction is diversity of citizenship. But by affidavits filed at our request *1196after argument, the parties established the requisite complete diversity upon the date when the complaint was filed. The parties’ sloppiness about jurisdiction is inexcusable, but not fatal.
The ultimate question on the merits is the validity of the arbitrators’ award. Against its validity Rumbleseat first makes several procedural arguments. The strongest — but it is very weak — is that the district judge should have ordered a jury trial on the issue of arbitrability, rather than ordering arbitration directly. Section 4 of the arbitration code, 9 U.S.C. § 4, provides that if a question is raised as to whether the parties had agreed to arbitrate and the party alleged to be in default demands a jury trial of the question, the district court “shall” conduct such a trial. But it also provides that “upon being satisfied that the making of the agreement for arbitration ... is not in issue, the court shall make an order directing the parties to” arbitrate. It is not true that by merely demanding a jury trial a party to an arbitration agreement can get one. He can get one only if there is a triable issue concerning the existence or scope of the agreement. If the arbitrability of the parties’ dispute involves no questions or only legal questions, a jury trial would be pointless because its outcome could not affect the judge’s decision on whether to order arbitration. See Par-Knit Mills, Inc. v. Stockbridge Fabrics Co., 636 F.2d 51, 54-55 (3d Cir.1980).
There is no question that the dispute over Rumbleseat’s allegedly violating the license agreement by continuing to make and sell dolls after the agreement was terminated (and the 275-day grace period for selling dolls made before cancellation expired) arises from the agreement. It is therefore within the scope of the arbitration clause. It is also a copyright dispute — more precisely, given the relief sought by the Post, a potential copyright dispute. The Post contends that Rumble-seat is violating the Post’s copyrights by continuing to make and sell dolls derived from the Rockwell illustrations after Rumbleseat’s license to do these things ended. But the fact that Rumbleseat may also be violating the Post’s property rights does not change the fact that the parties have a contract dispute covered by the arbitration clause. Rumbleseat’s only argument against this conclusion is that the arbitration clause expired when the agreement was terminated, but this is a ridiculous argument, and certainly does not raise a jury issue. Although the license agreement has ended, the principal clause that Rumbleseat is alleged to have violated, which requires that it stop making and selling the dolls after the license agreement ends, must, if it is to mean anything, continue in force after the agreement ends, and indeed until the requirements set forth in the clause are satisfied. This is especially clear in light of the 275-day grace period. It takes almost a year after the license ends for the cessation clause to take full effect; for though the prohibition against making takes effect immediately, the prohibition against selling takes effect only at the end of the 275 days. The Post had also accused Rumbleseat of violating the grant-back clause, and the arbitrators ordered Rumbleseat to transfer to the Post the derivative-work copyrights that it had obtained while the license was in effect. This is another contractual obligation that not only outlasts the license but cannot even come into play until the license is terminated.
The other procedural issues (including that of personal jurisdiction over Rumble-seat’s owner) merit no discussion, and so we come at last to the substantive issues. They involve the arbitrability of copyright questions and the validity of a no-contest clause in a copyright licensing agreement. Rumbleseat argued to the arbitrators that the Post, by failing to copyright the illustrations separately, never obtained valid copyrights in them and therefore breached its warranty of title and by this failure excused Rumbleseat from having to comply with the cessation clause. The Post countered with various arguments, but, curiously, omitted to mention paragraph 9 of the licensing agreement, which forbids Rumbleseat to contest the validity of the Post’s copyrights. The arbitrators must have re*1197jected Rumbleseat’s challenge to the Post’s copyrights, but we cannot tell on what ground they did so. They may have determined that the Post’s copyrights were valid; much evidence pro and con was presented. They may have decided that, even if the copyrights were invalid, this did not justify Rumbleseat’s violating the cessation clause. Or, having read the licensing agreement, and doubtless noticing paragraph 9 (the agreement is short), they may have decided they were forbidden to consider the validity of the Post’s copyrights even though the Post had not directed their attention to the paragraph.
If, as Rumbleseat argues, the Copyright Act implicitly forbids an arbitrator to determine the validity of a copyright, or if either the act or federal common law forbids an arbitrator to enforce a no-contest clause — and if either the validity of the Post’s copyrights or the validity of the no-contest clause is an essential strut of the arbitrators’ awared — then the award is illegal and the district court should not have confirmed it. Although the scope of federal judicial review of arbitration awards is excruciatingly narrow, it does cover the situation where the arbitrator has made an order that federal law forbids. See E.I. DuPont de Nemours & Co. v. Grasselli Employees Independent Ass’n of East Chicago, Inc., 790 F.2d 611, 615-17 (7th Cir.1986); id. at 617-20 (concurring opinion).
The italicized “if,” however, is a big one, as we do not know the grounds of the award. Now when a jury is given a mixture of proper and improper instructions, its verdict must be set aside even if the verdict may have been based on a theory on which the jury was properly instructed. See, e.g., Stromberg v. California, 283 U.S. 359, 368, 51, S.Ct. 532, 535, 75 L.Ed. 1117 (1931). Should the same rule be applied to the review of arbitration awards? Although we can find no decision directly on point, we point out that in Randall v. Lodge No. 1076, Int’l Association of Machinists & Aerospace Workers, 648 F.2d 462, 467-69 (7th Cir.1981), this court remanded to the arbitrator where it was unclear whether his award rested on a proper or an improper ground. Our disposition implied that the award could not be upheld merely because it may have rested on a proper ground. John T. Brady & Co. v. Form-Eze Systems, Inc., 623 F.2d 261 (2d Cir.1980), contains language suggesting that such an award would be upheld, but goes on to cite evidence that the award was not in fact based on an improper ground, for the arbitrator had issued an opinion and the opinion did not mention the improper ground. See id. at 642 (last sentence of penultimate paragraph); see also Sun Petroleum Products Co. v. Oil, Chemical & Atomic Workers Int’l Union, Local 8-901, 681 F.2d 924, 928-29 (3d Cir.1982). Western Electric Co. v. Communication Equipment Workers, Inc., 554 F.2d 135, 138 (4th Cir.1977) (per curiam), and In re Arbitration between Sobel and Hertz, Warner & Co., 469 F.2d 1211, 1215-16 (2d Cir.1972), also contain language which creates some tension with Randall.
But we sense no essential disagreement among the circuits. The fact that arbitrators are not required to give reasons for their awards, and the warm judicial regard for arbitration evident in the modern cases — a warmth that in part reflects modern caseload pressures — support a presumption that the arbitrators (who in this case, after all, were lawyers) based their award on proper grounds. Randall, which authorizes a clarifying remand only when “the reasons that are given [by the arbitrator] strongly imply that the arbitrator may have exceeded his or her authority under the submission and contract,” 648 F.2d at 468 (emphasis in original), is consistent with such a presumption.
The circumstances of the present case either rebut the presumption or make it inapplicable. To begin with, it is by no means clear that the arbitrators exceeded their authority in deciding copyright validity (we are, indeed, about to hold the contrary); and given this uncertainty it seems artificial to assume that they steered clear of the issue, erroneously believing that they had to. Second, the validity of the Post’s copyrights was a focus of the hear*1198ing before the arbitrators, and they never suggested that they thought themselves unauthorized to decide it. The Post did not even cite the no-contest clause to the arbitrators. So probably (though not certainly) the arbitrators upheld the validity of the copyrights; and had they not done so they might well have rejected the Post’s claims entirely or in part. We have no reason to suppose that any court would in such circumstances assume that the arbitrators had not decided the issue of validity, especially in a case such as this where the question is not just whether the arbitrators strayed beyond the limits of their mandate but whether in doing so they violated federal law. Finally, the probability that the arbitrators in reaching their decision relied on neither the validity of the Post’s copyrights nor on the no-contest clause must be very small, and since we do not know which they relied on, Rumbleseat is entitled to insist that we decide whether, if the arbitrators relied on either ground, they violated federal law.
We can find no cases on whether the validity of a copyright is arbitrable. Kamakazi Music Corp. v. Robbins Music Corp., supra, holds that there is no rule against submitting copyright claims in general to arbitration, but reserves the question whether the arbitrator may determine the validity of the copyright on which the claim is founded. See 684 F.2d at 231; see also 522 F.Supp. 125, 131 (S.D.N.Y.1981) (the district court’s opinion in Kamakazi). Rumbleseat's main argument against arbitrability is that Congress’s decision to give the federal courts exclusive jurisdiction of copyright actions implicitly precludes arbitration of disputes over the validity of a copyright. This is a non sequitur. The dispute that the Post wanted to arbitrate was a dispute over compliance with a copyright licensing agreement; and, as we have seen, state courts have jurisdiction over such disputes — indeed, federal courts can acquire jurisdiction over them only by virtue of diversity or pendent jurisdiction. And if in the course of a lawsuit in state court over a licensing agreement an issue of patent or copyright law arises, the state court is empowered to decide it. This conclusion is supported by the cases cited in 13B Wright, Miller & Cooper, Federal Practice and Procedure § 3582, at pp. 303-10 (2d ed. 1984); see in particular Pratt v. Paris Gaslight & Coke Co., 168 U.S. 255, 259, 18 S.Ct. 62, 64, 42 L.Ed. 458 (1897); Excelsior Wooden Pipe Co. v. Pacific Bridge Co., supra, 185 U.S. at 286-87, 22 S.Ct. at 682-83, and Peay v. Morton, supra, 571 F.Supp. at 117. None, however, involved the validity of a copyright. It is also supported by the principle that although federal courts have been held to have exclusive jurisdiction over cases arising under the federal antitrust laws, state courts are allowed to decide federal antitrust defenses to suits properly brought in those courts under state law. See, e.g., Lyons v. Westinghouse Electric Corp., 222 F.2d 184, 187 (2d Cir.1955) (L. Hand, J.). It is also supported by the practical consideration that if the state court is not allowed to decide federal defenses the litigation will be split up in an uneconomical fashion. If A, licensor, sues B, licensee, for breach of a copyright licensing agreement, and if B defends by arguing that A had no valid copyright and therefore violated the warranty clause, and if the state court may not decide the issue of validity, then the suit in state court must be stayed while the parties repair to federal court, much as in a case of primary (agency) jurisdiction.
It does not follow as the night the day that an arbitrator should also be allowed to decide the issue of validity. But it does follow that an argument against his being allowed to do so cannot be based on the exclusivity of federal copyright jurisdiction. That exclusivity just isn’t engaged by a suit over breach of a copyright license.
At argument we pressed counsel for Rumbleseat to give us a practical reason why arbitrators should not be allowed to decide issues of validity that arise in copyright license suits. He was unable to do so, and we can think of none ourselves. It is true that a copyright is a form of monopoly, so that a decision erroneously upholding the validity of a copyright might have the effect of continuing an unlawful monopoly in force. But there is no reason to think that arbitrators are more likely to err in copyright cases than state or federal judges are; the Supreme Court recently rejected such an argument in holding that the arbitration of antitrust claims arising out of international transactions is not contrary to public policy. See Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 105 S.Ct. 3346, 3357-58, 87 L.Ed.2d 444 (1985). And a copyright, even more than a patent, is a legal rather than necessarily an economic monopoly. It forbids copying the copyrighted work without the copyright holder’s permission, but *1199it does not forbid the making of close substitutes. So long as the second comer creates a work that is not substantially similar to the copyrighted features of the first work, there is no infringement. See 3 Nimmer on Copyright § 13.03[A] (1986); Nichols v. Universal Pictures Corp., 45 F.2d 119 (2d Cir.1930) (L. Hand, J.). If any monopolies harmful to the public have ever been built on invalid copyrights, we have not heard of them. Since, moreover, the opinions of commercial (as distinct from labor) arbitrators are not published, and have no precedential authority anyway— since, as this case illustrates, commercial arbitrators rarely state the grounds of their awards — erroneous decisions by arbitrators cannot establish or even assist in establishing the validity of a copyright against all potential infringers, but only against the defendant in the particular arbitration. Moreover, an award could not possibly be any sort of precedent — however attenuated — where, as in this case, it is not even certain that the arbitrators decided an issue of validity.
Rumbleseat is not the only company that can make passable porcelain copies of a Norman Rockwell illustration. If those illustrations really are in the public domain, someone else will come along and make the dolls even if this arbitration award is allowed to stand. Moreover, assuming the only question is whether the Post owns the copyrights or Rockwell’s heirs own them, the consumer could not suffer whatever the outcome of the arbitration. If these copyrighted illustrations confer lawful monopoly power on their true owner, the consumer will have to pay the monopoly price whoever that true owner is.
The danger of monopoly is more acutely posed by patents than by copyrights, yet Congress recently passed a statute expressly authorizing the arbitration of issues of patent validity. See 35 U.S.C. § 294. This is some evidence that the incremental danger of harmful monopoly from allowing arbitrators to adjudicate the validity of rights to intellectual property is not great. In the case of copyright disputes the danger seems minuscule. And against it must be weighed the damage to the utility of arbitration in disputes over copyright licenses if arbitrators may not determine issues of validity that arise in the course of the arbitration. The question is not whether the parties to a suit for copyright infringement may decide to refer the dispute to arbitration — no one doubts they may; it is whether the arbitration of a dispute arising from a copyright license must be interrupted if the licensee challenges the validity of the copyright. If so, this would toss a monkey wrench into an important means of resolving contractual disputes over intellectual property. We hold that federal law does not forbid arbitration of the validity of a copyright, at least where that validity becomes an issue in the arbitration of a contract dispute.
Last, we must decide whether a clause in a copyright licensing agreement forbidding the licensee to contest the validity of the copyright he has licensed is against public policy, as expressed in the Copyright Act or other possible sources of federal common law, and is therefore unenforceable. (Rumbleseat’s argument that it is barred by state law requires no extended discussion. The argument is based on restrictive-covenant cases; they are too remote to be illuminating.) This is another open question, see Herbert Rosenthal Jewelry Corp. v. Kalpakian, 446 F.2d 738, 739-40 (9th Cir.1971), and again the concern is with copyright monopolies. Suppose the Rockwell illustrations really were in the public domain and both the Post and Rumbleseat knew it, but, also knowing that both would be better off without competition, they agreed to give Rumbleseat an exclusive license with a no-eontest clause, hoping that no other potential competitor would discover the invalidity of the copyrights and challenge (or defy) them. Somewhat analogous practices involving patents have been alleged. See George L. Priest, Cartels and Patent Lácense Agreements, 20 J. Law & Econ. 309, 340-49 (1977). The danger of this kind of cozy deal would be less if the law forbade the Post to enforce the no-contest clause, so that Rumbleseat, if it changed its mind about the advantages of mutual forbearance, could go into competition with the Post notwithstanding the license.
We cannot call this danger nonexistent, although we suspect it is slight given our earlier remarks about the unlikelihood that a copyright (especially one that by hypothesis is invalid!) would confer an economically significant monopoly, one that would raise the price of the monopolized good well above, and depress its output well below, the competitive level. The danger is not so great, however, as to justify a rule of federal common law outlawing no-contest clauses without evidence of any monopolistic danger or effect. Such a clause serves *1200a useful purpose in most cases. Without it the licensee always has a club over the licensor's head: the threat that if there is a dispute the licensee will challenge the copyright’s validity. The threat would discourage copyright licensing and might therefore retard rather than promote the diffusion of copyrighted works. Also, a no-contest clause might actually accelerate rather than retard challenges to invalid copyrights, by making the would-be licensee think hard about validity before rather than after he signed the licensing agreement. Rumbleseat had, in fact, used its expressed doubts of the validity of the Post’s copyrights to obtain a lower royalty rate in the negotiations for the license.
What is needed is a balancing of the pros and cons of the clause in each case. That balancing is best done under antitrust law. Section 1 of the Sherman Act, 15 U.S.C. § 1, forbids contracts that restrain trade. If Rumbleseat had wanted, it could have attacked the no-contest clause under that statute. It did not do so. We decline to create a federal common law rule that would jostle uncomfortably with the Sherman Act. Noting the convergence of patent-misuse principles with antitrust principles, we said in USM Corp. v. SPS Technologies, Inc., 694 F.2d 505, 512 (7th Cir.1982): “If misuse claims are not tested by conventional antitrust principles, by what principles shall they be tested? Our law is not rich in alternative concepts of monopolistic abuse; and it is rather late in the date to try to develop one without in the process subjecting the rights of patent holders to debilitating uncertainty.” This point applies with even greater force to copyright misuse, where the danger of monopoly is less. We hold that a no-contest clause in a copyright licensing agreement is valid unless shown to violate antitrust law.
This holding is not barred by Lear v. Adkins, 395 U.S. 653, 89 S.Ct. 1902, 23 L.Ed.2d 610 (1969), which held that federal law forbids a state court to hold that a patent licensee is, by virtue of having been licensed, estopped to challenge the patent’s validity. Our case involves a negotiated clause rather than a doctrine that in effect reads a no-contest clause into every licensing agreement. The doctrine is apt to have a broader effect. We made a similar distinction in American Equipment Corp. v. Wikomi Mfg. Co., 630 F.2d 544, 549 (7th Cir.1980). Furthermore, the logic of Lear does not extend to copyright licenses. The opinion is narrowly written. It emphasizes
the important public interest in permitting full and free competition in the use of ideas which are in reality a part of the public domain. Licensees may often be the only individuals with enough economic incentive to challenge the patentability of an inventor’s discovery. If they are muzzled, the public may continually be required to pay tribute to would-be monopolists without need or justification.
Id. 395 U.S. at 670, 89 S.Ct. at 1911. A patent empowers its owner to prevent anyone else from making or using his invention; a copyright just empowers its owner to prevent others from copying the particular verbal or pictorial or aural pattern in which he chooses to express himself. The economic power conferred is much smaller. There is no need for a rule that would automatically invalidate every no-contest clause. If a particular clause is used to confer monopoly power beyond the small amount that the copyright laws authorize, the clause can be attacked under section 1 of the Sherman Act as a contract in restraint of trade. Rumbleseat does not argue that the clause here restrained trade in that sense. The fact that we can find no antitrust case — or for that matter any other reported case — that deals with a no-contest clause in a copyright license is evidence that these clauses are not such a source of significant restraints on freedom to compete as might warrant a per se rule of illegality.
A further point, already alluded to, is that the competitive interest in confining copyright and patent protection to valid copyrights and patents has to do with the originality, novelty, etc. of the work that is copyrighted or patented, rather than with the owner’s identity. Only the last question is in issue here; the validity of the Post’s copyrights, not the copyrightability of the Rockwell dolls, is in issue. Once it is decided that a work is copyrightable, the decision has been made that the additional cost per copy to the public is warranted by the encouragement that copyright protection gives to the creation of new and valuable works; whether or not the ownership of the copyright is contestable by a licensee is then a detail irrelevant to the competitive policies that underlie Lear. We have made this distinction in other contexts. See, e.g., Brunswick Corp. v. Riegel Textile Corp., 752 F.2d 261, 267 (7th Cir.1984).
*1201This overstates the case a little, because Rumbleseat argues among other things that the Post failed to register its copyrights properly. If that is right, maybe the Rockwell illustrations have fallen into the public domain. It seems more likely, though, that what happened (or may have happened, for of course the arbitrators may have been correct in upholding the validity of the Post’s copyrights — if that is what they did) is that the Post failed to perfect copyright in its derivative works— the photographs, printed in the magazine, of Rockwell’s illustrations. The Rockwell family, which owns the copyrights on the original illustrations, could still enjoin Rumbleseat’s infringement. The argument that the family’s copyrights fell into the public domain presumably relies on the old doctrine of indivisibility — a disfavored doctrine, however, and one rejected years ago for magazines. See 3 Nimmer on Copyright § 10.01[C](2) (1986).
So probably the effect of the copyrights on the price and output of porcelain dolls would be about the same whether or not Rumbleseat succeeded in knocking out the Post’s copyrights. At least this is a strong possibility. Unfortunately, because Rumbleseat failed to make the record before the arbitrators part of the record in this court, we have only the haziest idea of what was and was not in issue before the arbitrators; but doubts engendered by Rumbleseat’s failures must be resolved against Rumble-seat. At all events the basic originality and hence copyrightability of the Rockwell drawings on which the design of the infringing dolls is ultimately based seem not to be in issue.
So we have a narrow and a broad holding on no-contest clauses: they are valid in copyright licenses (broad); they are valid when no issue of copyrightability is presented (narrow). These, we emphasize, are both holdings, and therefore bind the district courts in this circuit. Of course a court of coequal or superior authority to this court might find one more persuasive than the other, and not being bound as a matter of authority to follow our decisions could decide to adopt just one.
Affirmed.