EASTERBROOK, Circuit Judge.

John Dillinger, Public Enemy No. 1, died on July 22, 1934, at the Biograph Theater in Chicago. He emerged from the air conditioned movie palace into as sweltering evening accompanied by two women, one wearing a bright red dress. The "lady in red," Anna Sage, had agreed to betray his presence for $10,000. Agents of the FBI were waiting. Alerted by Polly Hamilton, the other woman, Dillinger wheeled to fire, but it was too late. A hail of bullets cut him down, his .45 automatic unused. William C. Sullivan, The Bureau 30-33 (1979). Now a national historic site, the Biograph bears a plaque commemorating the event. It still shows, movies, and the air conditioning is not better now than in 1934.

Jay Robert Nash believes that Dillinger did not die at the Biograph. In Dillinger: Dead or Alive? (1970), and The Dillinger Dossier (1983), Nash maintains that Dillinger learned about the trap and dispatched Jimmy Lawrence, a small-time hoodlum who looked like him, in his stead. The FBI, mortified that its set-up had no sting, kept the switch quiet. Nash points to discrepancies between Dillinger's physical characteristics and those of the corpse: Dillinger had a scar on his upper lip and the corpse did not; Dillinger lacked a tooth that the corpse possessed; Dillinger had blue eyes, the corpse brown eyes; Dillinger's eyebrows were thicker than those of the corpse. Although Dillinger's sister identified the dead man, Nash finds the circumstances suspicious, and he is struck by the decision of Dillinger's father to encase the corpse in concrete before burial. As part of the cover-up, according to Nash, the FBI planted Dillinger's fingerprints in the morgue. After interviewing many persons connected with Dillinger's gang and the FBI's pursuit of it, Nash tracked Dillinger to the west coast, where Dillinger married and lay low. Nash believes that he survived at least until 1979. The Dillinger Dossier contains pictures of a middle-aged coupled and then an elderly man who, Nash believes, is Dillinger in dotage. Nash provides capsule versions of his conclusions in his Bloodletters and Badmen: A Narrative Encyclopedia of American Criminals from the Pilgrims to the Present (1973), and his expose Citizen Hoover (1972).

Nash's reconstruction of the Dillinger story has not won adherents among historians — or the FBI. Someone in Hollywood must have read The Dillinger Dossier, however, because in 1984 CBS broadcast an episode of its Simon and Simon series entitled The Dillinger Print, Simon and Simon featured brothers Rick and A.J. Simon, private detectives in San Diego. The district court summarized the episode, 704 F. Supp. 823, 828-29 (N.D.Ill. 1989):

The opening scene of The [Dillinger] Print shows Ty Becker, a retired FBI agent, telling his grandchildren about Dillinger's life and the shooting outside the Biograph Theater on July 22, 1934. Ty also mentions that he doubts Dillinger was the man who was shot that night and vows to track him down some day. After the grandchildren leave with his daughter Addie, an intruder breaks into Ty's home, steals an old gun which once belonged to Dillinger, and kills Ty with the gun.

Concerned that the police would regard Ty's death as a typical murder incident to burglary rather than related to moon-lighting she suspected he was performing for the FBI, Addie hires the Simons. Next, Addie goes to the bank to remove her father's safety deposit box. While she is in the vault, a masked man, wearing 1930s style spectator shoes, shoots tear gas into the bank and steals Ty's safety deposit box. The Simons arrive at the bank soon thereafter and pick up the thief's gun, which turns out to be the Dillinger gun Stolen from Ty's house. Police investigation later reveals that the gun bears the fresh fingersprint of John Dillinger.

The Simons are next seen purchasing newspapers which are carrying the story that Dillinger may be alive. As the Simons discuss the case, A.J. reads from a book entitled "Twentieth Century Desperadoes." He implies that some evidence supports the idea Dillinger is alive and relates to Rick several physical discrepancies between Dillinger and the Corpse described in the 1934 autopsy. Nash cites the same discrepancies in his books.

Numerous Dillinger impostors soon come forward, and the FBI enters the case. However, FBI Agents Kinneman, who is a friend of A.J.'s, informs the Simons that Ty Becker was not working for the FBI at the time of his death.

The scene then switches to a health club. As A.J. and Kinneman are playing racquetball, a man wearing a trench coat and spectator shoes enters the club and, from the viewing area, sprays a salvo of bullets over A.J. and Kinneman. Afterwards, A.J. Solemnly swears revenge against Dillinger of whoever tried to shoot him.

The next day, the Simons visit the police station and discover that "leads" regarding Dillinger's whereabouts have poured in from all over the world. They agree to check out a few leads, including one that takes them to a dentist in San Diego. The dentist rejects the Simons' suggestion that Dillinger lives in his house and attributes the police tip to a crazy woman who lives across the street.

The Simons and Addie then pay a visit to Ty Becker's old secretary, who informs them that Ty was working on an internal FBI investigation at the time he was killed. Immediately thereafter, the Simons and Addie receive a call from Kinneman, who tells them that "there is more truth to this" Dillinger affair than anyone had imagined and that the Simons should meet him at a closed-down theater.

When the Simons and Addie enter the theater, a gangster documentary is playing. Kinneman and a man wearings spectator shoes shoot at the trio. The Simons eventually kill the man in spectator shoes. They then subdue Kinneman, who admits to killing Ty Becker in order to stop the internal investigation which, it turns out, was directed at Kinneman, and to leaving the fake Dillinger fingerprint on the gun at the bank.

In the penultimate scene, the FBI thanks the Simons for their help in arresting Kinneman and solving the Dillinger mystery. Rick nevertheless insists Dillinger may be alive and perhaps living in Oregon. . . . The episode closes with a teaser: the dentist, one of the leads whom the Simons had interviewed earlier in the program, is seen pushing his elderly father in a wheelchair and admonishing him to refrain from discussing the "old days in Chicago" anymore.

Nash filed this suit seeking damages on the theory that The Dillinger Print violates his copyrights in the four books setting out his version of Dillinger's escape from death and new life on the west coast. The district court determined that the books' copyrighted material consists in Nash's presentation and exposition, not in any of the historical events. 691 F. Supp. 140 (N.D.Ill. 1988). CBS then moved for summary judgment, conceding for this purposes both accesses to Nash's books and copying of the books' factual material. The court granted this motion, 704 F. Supp. 823, holding that The Dillinger Print did not appropriate any of the material protected by Nash's copyrights.

CBS's concession removes from this case two questions that bedevil copyright litigation. See Selle v. Gibb, 741 F.2d 896, 901-02 (7th Cir. 1984). It leaves the questions whether the copier used matter that the copyright law protects and, if so, whether it took "too much" (that is, more than allowed by the "fair use" doctrine codified in 17 U.S.C. § 107). See Atari, Inc. v. North American Philips Consumer Electronics Corp., 672 F.2d 607, 614-17 (7th Cir. 1982), Toksvig v. Bruce Pub. Co., 181 F.2d 664 (7th Cir. 1950). These latter questions have proven especially difficult when, as in this case, the copier works in a medium different from the original.

Learned Hand, whose opinions still dominate this corner of the law, observed in Nichols v. Universal Pictures Corp., 45 F.2d 119, 121 (2d Cir. 1930), that all depends on the level of abstraction at which the court conceives the interest protected by the copyright. If the court chooses a law level (say, only the words the first author employed), then a copier may take the plot, exposition, and all other original material, even though these may be the most important ingredients of the first author's contribution. As a practical matter this would mean that anyone could produce the work in a new medium without compensating the original author, despite the statute's grant to the author of the privilege to make "derivative works." If on the other hand the court should select a high level of abstraction, the first author may claim protection for whole genres of work ("the romantic novel" or, more modestly, any story involving doomed young lovers from warning clans, so that a copyright on Romeo and Juliet would cover West Side Story too). Even a less sweeping degree of abstraction creates a risk of giving copyright protection to "the idea" although the statute protects only "expression". 17 U.S.C. § 102(b); Holmes v. Hurst, 174 U.S. 82, 86, 19 S.Ct. 606, 607, 43 L.Ed. 904 (1989); Mazer v. Stein, 347 U.S. 201, 217-18, 74 S.Ct. 460, 470, 98 L.Ed. 630 (1954); Rockford Map Publishers, Inc. v. Directory Service Co., 768 F.2d 145 (7th Cir. 1985); Miller v. Universal City Studios, Inc., 650 F.2d 1365 (5th Cir. 1981); United Telephone Co. v. Johnson Publishing Co., 855 F.2d 604, 608-09 (8th Cir. 1988).

Sometimes called the "abstractions test," Hand's insight is not a "test" at all. It is a clever way to pose the difficulties that require courts to avoid either extreme of the continuum of generality. It does little to help resolve a given case, even when melded with Hand's further observation, in cases such as Peter Pan Fabrics, Inc., v. Martin Weiner Corp., 274 F.2d 487, 489 (2d Cir. 1960), that one may ask what the "ordinary observer" would think are the essential parts of the two works' "aesthetic appeals." See also Seller, 741 F.2d at 903-05. Who is the "ordinary" observer, and how does this person choose the level of generality? Ordinary observers, like reasonable men in torts, are fictitious characters of the law, reminders that judges must apply objectives tests rather than examine their own perceptions. They do not answer the essential question: at what level of generality? After 200 years of wrestling with copyright questions, it is unlikely that courts will come up with the answer any time soon, if the indeed there is "an" answer, which we doubt.

Hand returned again and again to the opposing forces that make the formulation of a single approach so difficult. Intellectual (and artistic) progress is possible only if each author builds on the work of others. No one invents even a tiny fraction of the ideas that make up our cultural heritage. Once a work has been written and published, any rule requiring people to compensate the author slows progress in literature and art, making useful expressions "too expensive," forcing authors to re-invent the wheel, and so on. Every work uses scraps of thought from thousands of predecessors, far too many to compensate even if the legal system were frictionless, which it isn't. Because any new work depends on others even if unconsciously, broad protection of intellectual property also creates a distinct possibility that the costs of litigation — old authors trying to get a "piece of the action" from current successes — will prevent or penalize the production of new works, even though the claims be rebuffed. Authors as a group therefore might, prefer limited protection for their writings — they gain in the ability to use others' works more than they lose in potential royalties. See William M. Landes Richard A. Posner, An Economic Analysis of Copyright Law, 18 J. Legal Studies 325, 332-33, 349-59 (1989).

Yet to deny authors all reward for the value their labors contribute to the works of others also will lead to inefficiently little writing, just as surely as excessively broad rights will do. The prospect of reward is an important stimulus for thinking and writing, especially for persons such as Nash who are full-time authors. Before the first work is published, broad protection of intellectual property seems best; after it is published, narrow protection seems best. At each instant some new works are in progress, and every author is simultaneously a creator in part and a borrower in part. In these roles, the same person has different objectives. Yet only one rule can be in force. This single rule must achieve as much as possible of these inconsistent demands. Neither Congress nor the courts has the information that would allow it to determine which is best. Both institutions must muddle through, using not a fixed rule but a sense of the consequences of moving dramatically in either direction.

If Nash had written a novel that another had translated into a screenplay, this would be a difficult case. Although The Dillinger Print is substantially original, it does not matter that almost all of the second author's expression is new. "[N]o plagiarist can excuse the wrong by showing how much of his work he did not pirate." Sheldon v. Metro-Goldwyn Pictures Corp., 81 F.2d 49, 56 (2d Cir. 1936) (L. Hand, J.). The TV drama took from Nash's works the idea that Dillinger survived and retired to the west coast, and employed many of the ingredients that Nash used to demonstrate that the man in the Cook County morgue was not Dillinger. CBS even used one of Nash's books as a prop: "Twentieth Century Desperadoes" in The Dillinger Print is a ringer for Nash's Bloodletters and Badmen. To see that The Dillinger Print is in a sense a "derivative work", we need only imagine how we would react if Nash had written a short story based on the premise that Dillinger was really a woman masquerading as a man, and CBS has used a switch in sex roles as the centerpiece of a drama. In such an event we would need to decide, as Hand did in Sheldon, whether the portions CBS took over were qualitatively so important that the original author's market would be diminished excessively by a rule allowing similar appropriations in the regular course.

Nash does not portray The Dillinger Dossier and its companion works as fiction, however, which makes all the difference. The inventor of Sherlock Holmes controls that Character's fate while the copyright lasts; the first person to conclude that Dillinger survived does not get dibs on history. If Dillinger survived, that fact is available to all. Nash's rights lie in his expression: in his words, in his arrangement of facts (his deployment of narration interspersed with interviews, for example), but not in the naked "truth". The Dillinger Print does not use any words from The Dillinger Dossier or Nash's other books; it does not take over any of Nash's presentation but instead employs a setting of its own invention with new exposition and development. Physical differences between Dillinger and the corpse, planted fingerprints, photographs of Dillinger and other gangsters in the 1930s, these and all the rest are facts as Nash depicts them. (Nash did not take the photographs and has no rights in them; The Dillinger Print used the photos but not Nash's arrangement of them.)

The cases closest to ours are not plays translated to the movie screen (as in Sheldon) but movies made from speculative works representing themselves as fact. For example, Universal made a motion picture based on the premise that an idealistic crewman planted a bomb that destroyed the dirigible Hindenburg on May 6, 1937. The theory came straight from A.A. Hoehling's Who Destroyed the Hindenburg? (1962), a monograph based on exhaustive research. The motion picture added subplots and development, but the thesis and the evidence adduced in support of it could be traced to Hoehling. Nonetheless, the Second Circuit concluded that this did not infringe Hoehling's rights, because the book placed the facts (as opposed to Hoehling's exposition) in the public domain. Hoehling v. Universal City Studios, Inc., 613 F.2d 972 (1980). See also Miller v. Universal City Studios (facts about a notorious kidnapping are not protected by copyright). Cf. Musto v. Meyer, 434 F. Supp. 32 (S.D.N.Y. 1977) (idea for The Seven Per Cent Solution derived from article in medical journal).

Hoehling suggested that "[t]o avoid a chilling effect on authors who contemplate tackling an historical issue or event, broad latitude must be granted to subsequent authors who make use of historical subject matter, including theories or plots". 618 F.2d at 978. As our opinion in Toksvig shows, we are not willing to say that "anything goes" as along as the first work is about history. Toksvig held that the author of a biography of Hans Christian Andersen infringed the copyright of the author of an earlier biography by using portions of Andersen's letters as well as some of the themes and structure. Hoehling rejected Toksvig, see 618 F.2d at 979, concluding that "[k]nowledge is expanded . . . by granting new authors of historical works a relatively free hand to build upon the work of their predecessors." Id. at 980 (footnote omitted). With respect for our colleagues of the east, we think this goes to the extreme of looking at incentives only ex post. The author is in Hoehling and Toksvig spent years tracking down leads. If all of their work, right down to their words, may be used without compensation, there will be too few original investigations, and facts will not be available on which to build.

In Toksvig the first author, who knew Danish, spent three years learning about Andersen's life; the second author, who knew no Danish, wrote her biography in less than a year by copying out of the first book scenes and letters that the original author discovered or translated. Reducing the return on such effort, by allowing unhindered use, would make the initial legwork less attractive and so less frequent. Copyright law does not protect hard work (divorced from expression), and hard work is not an essential ingredient of copyrightable expression (see Rockford Map), to the extent Toksvig confuses work or ideas with expression, it has been justly criticized. Rosemont Enterprises, Inc., v. Random House, Inc., 366 F.2d 303, 310 (2d Cir. 1966); Miller v. Universal City Studios, 650 F.2d at 1371; William F. Party, The Fair Use Privilege in Copyright Law 65-70 (1985). See also Eisenschiml v. Fawcett Publications, Inc., 246 F.2d 598, 604 (7th Cir. 1957) (recognizing both that any two treatments of the same historical subject will be similar because the facts limit the author's freedom, and that a scholar is entitled to use a predecessor's work). We need not revisit Toksvig on its own facts to know that it is a mistake to hitch up at either pole of the continuum between granting the first author a right to forbid all similar treatments of history and granting the second author a right to use anything he please of the first's work. Cf. New Era Publications v. Henry Holt Co., 873 F.2d 576 (2d Cir.), rehearing en banc denied, 884 F.2d 659 (1989).

Authors of fiction do not (necessarily) need greater incentives than authors of non-fiction. Users of and elaborators on works of non-fiction are not (necessarily) more easily dissuaded than are those who use or elaborate on works of fiction. Decisions such as Hoehling do not come straight from first principles. They depend, rather, on the language of what is now 17 U.S.C. § 102(b): "In no case does copyright protection for an original work . . . extend to any idea, . . . or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work." Long before the 1976 revision of the statute, courts had decided that historical facts are among the "ideas" and "discoveries" that the statute does not cover. International News Service v. Associated Press, 248 U.S. 215, 234, 39 S.Ct. 68, 71, 63 L.Ed. 211 (1918). This is not a natural law; Congress could have made copyright broader (as patent law is). But it is law, which will come as no surprise to Nash. His own books are largely fresh expositions of facts looked up in other people's books. Consider the introduction to the bibliography in Murder, America: Homicide in the United States from the Revolution to the Present 447 (1980):

The research for this book was done in libraries and achieves throughout the United States, in addition to interviews and lengthy correspondence. The author's own files, exceeding more than a quarter of a million separate entries and a personal crime library of more than 25,000 volumes, were heavily employed.

The producers of Simon and Simon used Nash's work as Nash has used others': as a source of facts and ideas, to which they added their distinctive overlay. As the district court found, CBS did no more than § 102(b) permits. Because The Dillinger Print uses Nash's analysis of history but none of his expression, the judgment is

AFFIRMED.