[This is the third in series of posts summarizing my new article, “The Grapes of Roth.”]
Introduction
Part I-A: Duck-Rabbits in Equity

In yesterday’s post, I summarized how infringement determinations were made in the age of equity, in the early to mid-twentieth century: judges compared the plaintiff’s and defendant’s works as an ordinary observer would, then mentally pared down the similarities by considering only protected expression. Based on that filtered comparison, the judge would arrive at a conclusion about whether the protected similarities were substantial enough to warrant an infringement determination.

That method put a lot of faith in the ability of federal judges to exercise aesthetic, legal, and policy discretion to reliably arrive at just outcomes involving infringement. But by the early 1960s, that faith was waning. Several factors were behind this shift. The first was simply generational: the active bench of the Second Circuit, the leading copyright appellate court in the mid-twentieth century, experienced a nearly complete turnover between 1951 and 1957, and by 1964, all nine active judges had been appointed within the last ten years. Learned Hand and the cadre of judges who served with him–judges who were used to equitable decision-making–were gone, replaced by judges such as Edward Lumbard, Henry Friendly, Irving Kaufman, and Thurgood Marshall. This new generation of judges had a very different conception of the judicial role, one considerably more wary of open exercises of judicial discretion.

There were several reasons for this, many of which are briefly elaborated in my article, and a full account of which could occupy its own article or even a book. I’ll just briefly list them here. Legal realism and Lochner–era judicial review had desanctified the judicial decisionmaking process in the eyes of many. The rise of the administrative state and an explosion of statutes moved the conceptual center of the law away from courts, and away from the common law, and placed it instead with legislatures and statutory text. The Erie decision reinforced that view by locating law in the command of a government, not a transcendent body of principles. Equity was fading as a familiar decision-making procedure. The federal courts and federal caseload exploded, putting pressure on lower courts to better explain their decisions for more efficient review.

The result was a revolution in legal ideology, the foundational axioms of which were encapsulated by a theory that briefly flourished at Harvard Law School in the 1950s and ’60s, and was laid out in a set of teaching materials titled “The Legal Process.” Legal Process theory had two significant components to it. First was the principle of “institutional settlement,” meaning the way in which different sorts of decisions were parceled out to different institutional players. With the rise of statutes and regulations, the content of the law was no longer mostly in the hands of judges, but rather to a large degree was the responsibility of legislatures and administrative agencies. Once it was settled who had primary authority over the content of some provision of law, that institutions decision should govern the other players–including courts. (For more on how “institutional settlement” may have influenced copyright law, see Shyam Balganesh’s article Copyright as Legal Process.)

The second component was the principle of “reasoned elaboration.” Judicial decisions are legitimate, according to this principle, only because they are explained in a series of steps that “build the bridge between the authorities they cite and the results they decree.” The 1960s witnessed a surge of multi-step tests and multi-factor balancing frameworks. Judge Hand’s statement in Sinram v. Pennsylvania R.R. that ultimately every judicial decision rests on fiat was cast aside.

One immediate impact of this ideological shift was a sharp turn towards formalized decision-making in judicial opinions. Gone was Hand’s one-sentence determination in Sheldon that “the dramatic significance of the scenes we have recited is the same.” A modern judicial opinion had to be supported with premises that led, through a syllogistic process, inexorably to the ultimate conclusion. But given the tricky and unexplainable balancing act described in yesterday’s post, where was copyright law supposed to get such a framework?

The Second and Ninth Circuit’s first attempt was to simply drop the protectability and policy questions out of the infringement inquiry, leaving only the question of how much material the defendant took. Thus, in Bradbury v. CBS (9th Cir. 1961), the Ninth Circuit reviewed a Hand-like decision from well-known district court judge Leon Yankwich. Yankwich had painstakingly gone through the two works in question — the plaintiff was Ray Bradbury, and the work he said was infringed was the precursor to “Fahrenheit 451” — and considered how much protectable material had been taken by the defendant. Hardly any, was Yankwich’s answer — the vast majority of the similarities were common themes or developments in works set in a dystopian society, going all the way back to Victor Hugo’s “Torquemada.” The remainder was not substantial enough to be infringing.

Too complicated! said the Ninth Circuit. Look at all the similarities that Bradbury read into the record — 21 of them! Such zingers as:

3. In the third scene of defendants’ teleplay, the Bookman, returning to headquarters, meets the “different” or “unorthodox” girl who provokes him into new ways of thinking about life and his work by her unorthodox approach.

In plaintiff’s work Fahrenheit 451, the Fireman meets a “different” girl, one with unorthodox ideas, outside his house on return from a burning job.

In each, the “different” girl influences and motivates the leading character and the development of the plot.

It turns out that Jill Layton from Brazil, Luna Lovegood from Harry Potter, and Wyldstyle from the Lego Movie are all infringing! Without stopping to consider whether the similarities were actually protected material, the Bradbury majority concluded that there were enough of them to make the defendant’s work “recognizable by an ordinary observer as having been taken from the copyrighted source.” The Second Circuit, in Ideal Toy Corp. v. Fab-Lu Ltd. (2d Cir. 1966), followed suit, quoting the now-familiar two-part inquiry from Arnstein v. Porter, but holding that it was satisfied if the “average lay observer would recognize the alleged copy as having been appropriated from the copyrighted work.”

Recognizability had much to recommend it for ease of application. By jettisoning the difficult legal question of protectability, and the difficult policy question of the proper threshold for infringement, all that was left was the factual question of how much was borrowed. Asking whether an ordinary observer would recognize the copied material was a relatively simple way of answering that question, one that required little knowledge of the art form or copyright law. And in the 1960s that must have seemed attractive because the sheer number of types of works subject to copyright claims was exploding.  Whereas most infringement cases prior to 1950 dealt with plays, novels, films, musical compositions, and the occasional factual compilation or instructional text — materials federal judges might have some artistic opinions about — after the Supreme Court’s decision in Mazer v. Stein (1954), courts were inundated with cases involving fabric patterns, dolls, toys, jewelry, labels, trophies, fireplace tools, equipment manuals, and greeting cards. A test that asked only about recognizability lessened judges’ burden considerably.

Roth Greeting Cards was another case in the Bradbury/Fab-Lu line, using “recognizability” as the test for infringement. The defendant, United Card Co., which had no staff artists or humorists, was churning out cards that had the same exact jokes as Roth Greeting Cards, with different (but thematically appropriate) artwork. Clearly United was “stealing” Roth’s cards. But the jokes were uncopyrightable, and the artwork was different. That didn’t stop the majority, with an opinion written by Judge Frederick Hamley, who had been on the Bradbury panel. “The remarkable similarity between the Roth and United cards in issue (with the possible exception of exhibits 5 and 6) is apparent to even a casual observer,” the Roth majority concluded. Perhaps in response to a dissenting opinion that asked, in essence, where the beef was, Judge Hamley argued it was plaintiff’s whole vibe: “It appears to us that in total concept and feel the cards of United are the same as the copyrighted cards of Roth.”

However much it satisfied the new yen for discretion-less tests, this “vaguely recognizable” standard had serious problems, as Roth Greeting Cards illustrated. Copying uncopyrightable material should not amount to infringement, even if it was paired with imitated but not copied copyrightable material. Roth was followed a few years later by a Ninth Circuit decision, written by Roth’s dissenter, holding basic luggage tags to be infringing, because they were the same shape as the plaintiff’s, shared a blank area for initials, and contained the text “if found, notify” the manufacturer’s phone number. But if recognizability was out, and equitable discretion was out, what was the alternative? Judge Hamley’s throw-away line about “total concept and feel” would provide a solution.