The text below represents commentary, not legal advice.

Return to: Critical Copyright, and more.

Return to: Paul Edward Geller, attorney.

 

Revised version of a paper given at the interdisciplinary workshop held by the Annenberg Center for Communication, University of Southern California, on September 26-27, 1998. Copyright © Paul Edward Geller 1998. Published in longer and footnoted form in the Journal of the Copyright Society of the U.S.A., vol. 46 (1998), p. 39, and reprinted in Dear Images: Art, Copyright and Culture (Ridinghouse ICA, 2002), edited by Daniel McClean and Karsten Schubert. For comments on prior drafts, I thank Jeanne d'Andrea, Lionel Bently, Lorin Brennan, François Dessemontet, Jay Dougherty, Thomas Dreier, Bernard Edelman, Ysolde Gendreau, Paul Goldstein, Marci Hamilton, Ulrich Loewenheim, Jerome Reichman, Mark Rose, Robert Rotstein, Alain Strowel, and David Vaver.

Hiroshige vs. Van Gogh :
Resolving the Dilemma of Copyright Scope in Remedying Infringement

Paul Edward Geller

1. A Thought-Experiment

European literati and artists became captivated by Japanese art, especially prints, in the latter half of the nineteenth century. When in Paris, Vincent van Gogh, a collector of Japanese prints himself, painted the Portrait of Old Tanguy. Tanguy had an art-supply shop in which he displayed the works of avant-garde painters, among others. The background of his Portrait shows a wall covered with Japanese prints.

In the Fall of 1887, Van Gogh painted his Japonaiseries. The most notable are based on a pair of prints by Utagawa Hiroshige, the Plum Garden at Kameido and Sudden Shower over the Great Bridge near Atake, from the series One Hundred Views of Famous Places in Edo. Hiroshige, in his wood-block prints, composes in clearly delineated levels of space and subtly gradated colors; Van Gogh, in his oil paintings The Tree and The Bridge, pushes and pulls our vision with agitated brush-strokes and boldly opposed colors. (Link to these works.)

How would Hiroshige have responded had he known what this foreign enthusiast had made of his works? It might be a mistake to attribute to Hiroshige our litigious impulse to dispute Van Gogh's decision to copy his graceful prints in viscous oils. While traditional Oriental culture might have demanded that a student acknowledge a master for creating models worthy of imitation, it did not threaten the student with legal sanctions for copying such models, even badly. Like the Chinese scholar Shen Zhou before him, Hiroshige might have thought: ". . . if my poems and paintings, which are only small efforts to me, should prove to be of some aid to the forgers, what is there for me to grudge about?" (Quoted in William P. Alford, To Steal a Book is an Elegant Offense: Intellectual Property Law in Chinese Civilization, Stanford University Press, 1995, at 29.)

Suppose that one of Hiroshige's heirs, holding his rights, sues Vincent van Gogh, his occasional dealer Tanguy, or his brother Theo or sister-in-law Johanna, heirs who took charge of works after his death. This hypothetical suit would be for copying Hiroshige's prints and making Van Gogh's Japonaiseries available to the public. This case is not posited to test the state of the French law of author's rights in the late nineteenth century. Rather it shall serve as a thought-experiment to elucidate a basic dilemma common to all copyright law.

 

2. The Copyright Dilemma

The field of copyright has distinct dimensions. One dimension concerns individual creators and their works. Here we may ask: Are Hiroshige's rights in his works infringed by Van Gogh's studies? The other dimension concerns disseminating works to the public. Here the question becomes: Should the court grant Hiroshige or his heir remedies against the dissemination of Van Gogh's studies?

These dimensions, though distinct in theory, are intertwined in practice. Creators elaborate works in communicating with colleagues and critics, patrons and promoters, and larger publics. Technology has progressively globalized the communication networks into which creators have fed their works. To assess this progress, start with the fact that, two millennia ago, the Roman Empire and the Han Dynasty of China did no more than sporadically trade with each other. Only later did Greek art forms, previously brought by Alexander to the frontiers of India, filter into the Far East over the Great Silk Road. Sill later yet, Chinese inventions essential to globalization, such as the magnetic compass and paper, reached the West.

Creations are thus fed back into cultural networks at varying rates. The art worlds of modern Japan and Europe learned of each other's rapidly developing styles with significant delays through the last century. For example, the Japanese art of the "floating world" had already known such greats as Utamaro in the eighteenth century, but it most notably synthesized Chinese and European models of landscape art in the work of Hokusai and Hiroshige in the early nineteenth century. While Europe had only fragmentary contacts with the full range of artistic developments in Japan before the middle of the nineteenth century, European access and response to this art had expanded greatly by the time Van Gogh reached Paris in 1886.

This rate of feed-back is critical to the rationale of intellectual property. Farmers would be reluctant to market produce that was promptly stolen once it went on sale. But the law recognizes property in the fruits of the farmers' labor and polices the marketplace for thieves. Similarly, authors and inventors would seem to lose incentives to release their products of mind into the marketplace if, in the process, they found such products used by free-riders. Instead, armed with intellectual property, they are supposedly induced to feed more works and inventions back into communication networks, thus further stimulating culture and technology. Such feedback helps to achieve policies that have been formulated, for example, in terms of "the Progress of Science and useful Arts." (U.S. Constitution, art. I, § 8, cl. 8.)

Should the law then allow Hiroshige or his heir to stop Van Gogh from making studies of Hiroshige's works and Van Gogh's heirs from feeding the resulting new works back into world culture? If, on the one hand, we provide creators with no legal control over the fate of their creations, they might be reluctant to input such public goods into cultural networks. If, on the other, we provide them with too much control, we allow them to stop others from elaborating on these products of mind and feeding still further creations back into such networks. If the rules on point are too lax, copyright becomes ineffectual; if the rules are too stringent, such property becomes counter-productive.

Let us restate the dilemma doctrinally. Copyright arises out of creative acts: Hiroshige has copyrights in the prints he creates, and Van Gogh has them in the studies he creates. In principle, copyrights, so-called properties, entitle creators to dispose of their respective works at will; however, in our hypothetical suit, such rights could well conflict. That is, Hiroshige's right in a print could, without regard for others, lead to stopping Van Gogh or his heirs from exercising his subsequent right to dispose of his study. But the fact that Hiroshige's work comes first ought not be decisive for the simple reason that, unlike patent law, copyright law does not favor works because of priority in time. The follower, Van Gogh, may further invoke freedoms of creation and expression.

Courts, when faced with a hard case such as ours, seek equitable solutions. They are like Goldilocks who, in the home of the three bears, wanted porridge that was neither "too hot" nor "too cold." The court may well ask: Does any equitable or policy-based defense, sometimes a provision setting forth an apparent exception, excuse arguable infringement? In the United States there is the doctrine of fair use, and Germany allows transformative uses under the narrower doctrine of free utilization (freie Benutzung). Other jurisdictions permit such transformative uses as parody under specific exceptions, while still others subject such uses to general infringement analysis.

True dilemmas are not always resolved by Goldilockean compromises. Nor is it just a matter of simply reconciling particular older and younger creators' claims, but of resolving the dilemma consistently with basic copyright rationales. While such doctrines as fair use and free utilization are idiosyncratic to particular copyright laws, and specific exceptions vary considerably between all these laws, infringement analysis is common to all such laws. Accordingly, in approaching the common dilemma of copyright scope, only infringement analyses common to the major copyright systems will here be considered.

 

3. Creating Works: Infringement?

Asking when infringement takes place engages analysis along the dimension of creation. It focuses on the works of individuals: on the one hand, the work of an earlier creator, for us Hiroshige, whose heir is our plaintiff; on the other, the work of a later creator, for us Van Gogh, our defendant. It is a matter of comparing the older and younger authors' works.

 

a. Creative Options; Sliding-Scale Analysis

Courts avoid our dilemma in easy cases, where they are faced with only literal or close copies. Such copies, displaying little or nothing new above and beyond the works copied, cannot in themselves enrich the creativity that feeds culture. The dilemma becomes acute, however, when it is a matter, not of obvious copies, but rather when prior works are admitted to be the bases of later works that are therefore alleged to be infringing.

In the eighteenth century, copyright was instituted to deal only with easy cases, the pirate reprinting of books or restaging of plays. At the start of the nineteenth century, courts typically found no infringement in what leading French commentary called "[t]he transmutation of form that the translator causes the original to undergo." (Augustin-Charles Renouard, Traité des droits d'auteurs, Jules Renouard et Cie., 1838-39, vol. 2, at 37.) But as trade in books became increasingly globalized, authors and publishers started to claim rights to stop translations in foreign markets. Rights then began to expand beyond protecting any work against substitution in the markets that work targeted. The right to control deriving one work from another emerged to reach new markets in new media. That right now comes into play, for example, when literary works are staged or filmed.

In hard cases, when plaintiffs allege that defendants have derived works from their own, courts often resort to complementary doctrines. Asking what is protectible, courts may not protect "facts" or "ideas," but rather only "expression"; assessing infringement, courts ask whether plaintiff's work is copied in defendant's "substantially" similar work or whether "essential" or "characteristic traits" are taken. But such notions as "ideas" defy ready definition, and equally metaphysical notions of "substance" suggest that works of the mind are things like tables and chairs, consistently perceived by all audiences, but none of these doctrines by itself guides courts to consistent decisions. For example, in a suit for infringement of Gone with the Wind by the novel The Blue Bicycle in France, the first trial court found infringement, and the case had to be appealed and retried twice. (T.G.I. Paris, 3e ch., Dec. 6, 1989, CDA, May 1990, at 21, rev'd, CA Paris, 1re ch., Nov. 21, 1990, 147 RIDA 319 [1991], rev'd, Cass. civ. I, Feb. 4, 1992, 152 RIDA 196 [1992], on remand, CA Versailles, Dec. 15, 1993, 160 RIDA 255 [1994]) With the same works at issue in the United States, the trial court simply refused to grant a preliminary injunction without a showing of more than a taking of idea-like elements like plot themes. (Trust Company Bank v. Putnam Publishing Group Inc., 5 U.S.P.Q. 2d 1874, 1878-1879 [C.D. Ca1. 1988]).

Courts may be said to "filter" factually or idea-determined materials out of the scope of what is protectible. For example, in one French case, a researcher sued a writer for taking vocabulary from his scholarly works on Cajun culture to use in a novel set in a Cajun locale. But the trial court found no infringement in the use of such culturally typical expressions that it considered to be indispensable to giving the novel "a minimum of verisimilitude." (T.G.I. Paris, 1re ch., Jan. 16, 1991, 148 RIDA 180 (1991), aff'd, CA Paris, 1re ch., Jan. 14, 1992, 152 RIDA 198 (1992).) Moreover, to the extent that there is only one way, or very few ways, to convey facts accurately, for example, in lists, charts, etc., resulting expressions may not be protected. Obviously, there can be just one correct spelling and numbering for each listing in a telephone directory and only a few conventional signs for situating sites on maps. More generally, the scope of protection narrows to the extent creative options are either too restricted to allow for creative elaboration or indispensable for subsequent creation.

Courts often have difficulty deciding how coarsely or finely to filter out factually or idea-determined materials. It is not always easy to specify at what point a set of creative options is too restricted or indispensable to allow for protection. For example, in one case in the United States, copyright was asserted in a so-called pitching form into which such baseball statistics as game times, betting odds, pitchers' performances, etc., were compiled. But one judge on appeal effectively asked: What is the "proper level of abstraction" for formulating such data, more generally, for grasping "the ideas which underlay" the expression in the work at issue? (Kregos v. Associated Press, 937 F.2d 700, 714 [2d Cir. 1991] [Sweet, J., dissenting].) Logically, without knowing just how to focus on any idea against which creative options might be measured, it becomes impossible to say when there are too few ways to express that idea to protect corresponding expressions.

In easy cases, courts find infringement in the full reconstruction of plaintiff's work in a literal or close copy. In hard cases, the question remains open: how much of the "substance" of a work must be re-generated before there is infringement? Since creativity is protean, standards of protection, such as "originality" or "creativity," are satisfied to varying degrees from case to case. In the United States, the Supreme Court indicated that a minimally creative compilation of facts would attract "thin" protection. (Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340, 349 [1991].) By implication, a more creative work would call for "thicker" protection, not merely against literal or close copies, but against derivative works. Continental European doctrines, typically predicating degrees of creativity that vary from work to work, comparably allow for such sliding-scale analysis of infringement.

Unfortunately, judicial penchants can push and pull on such analysis. For example, in one German case, the courts, generous to a plaintiff, indulged the premise that standard industrial-drafting techniques allowed for injecting sufficient personal creativity into a technical drawing to protect it against close copying. (Explosionszeichnungen decision, BGH, Feb. 28, 1991, 1991 GRUR 529.) In another German case, judicial sympathies favored the defendant: the image of a female nude ingeniously posed by the innovative fashion photographer Helmut Newton was at issue, but a painting which creatively transformed that image, albeit recognizably, was found not to infringe Newton's copyright. (Power of Blue decision, OLG Hamburg, Oct. 12, 1995, 1996 ZUM 315.)

One commentator notes the opening for possibly inconsistent results: "The outcome in a given case might well be that, while the plaintiff's work could be protected against outright piracy by a third party, it does not receive protection as against defendant's work that incorporates parts of it less obviously or more creatively than any slavish copy." (François Dessemontet, "Switzerland" § 8[1][a], in International Copyright Law and Practice, edited by P.E. Geller, Matthew Bender & Co., 1998, vol. 2.)

 

b. A New Framework for Infringement Analysis

How to respond to these analytic vagaries and judicial ambivalences? Let us start by limiting the sense of "idea" to the statutory terms of any "procedure, process, system, [or] method" for generating works. (17 U.S.C. § 102(b) (1998).) Aristotle provides us with a classical example, explaining the idea of tragedy in just such operational terms: have a hero commit an act of hubris, and have him fall and discover his own fall, to accomplish a catharsis of pity and fear in the audience.

Next, let us situate our operational notion of "ideas" within a larger framework of analysis. Our notion encompasses a wide range of procedures and methods for generating works. At one end, distinguish as routines those information-processing procedures that leave little or no room for creative options. Moving forward from that starting point, we can distinguish the following processes along a spectrum from copying to creation. Once drawn, these distinctions will be illustrated in terms of the extent to which the processes can be satisfactorily computerized.

At the start of the spectrum, rote copying results from applying a single set of identified routines to a work. For example, to copy a text literally, with the same wording, we might change fonts, using as many routines as there are printing symbols, plus some others to reformat the text. Or, to copy an image closely, we might trace it out on transparent paper, or we might photocopy it, laying it into the photocopy machine, setting the desired number of copies and other parameters, and pressing the button. Either way, a given set of identified routines suffices to obtain a copy.

There are many different processes in the middle of the spectrum. The rubric of knowledgeable reworking seems to cover most of them. Such processes use routines that are generally known, but not necessarily all identifiable. Nor need the set of such routines used in a given case fully suffice to determine the entire work generated in that case. Consider translating a French cookbook into English: it is necessary to rely on lexical and syntactic routines known, sometimes only implicitly, by bilingual speakers. For example, while French usage regularly places adverbs between verbs and direct objects, English-speakers most often relocate the adverbs, usually to the beginning or end of clauses. In any event, a straightforward cookbook, if not almost every work-a-day text, is susceptible of only a slightly variable set of translations likely to be acceptable to bilingual speakers. As other examples of mid-range processes, consider selecting and organizing facts into a compilation, or excerpts into an anthology, or recontextualizing a work, as in appropriation art.

At the far end of the spectrum, there is innovative recasting. Here, no known set of routines suffices for moving from one work to another or even to comparable works. Consider the enterprise of translating James Joyce's Ulysses into French: at hundreds of thousands of points, the translators had to make choices that linguistic rules alone could not have dictated. The saxophonist Charlie Parker provides another example: starting from the chords of the tune Cherokee,which an accompanying guitarist had inverted just "to keep a beat going," he moved into an entirely new mode of jazz improvisation. (Quoted in Ira Gitler, Swing to Bop: An Oral History of the Transition in Jazz in the 1940s, Oxford University Press, 1985, at 69-72.)

At the start of the spectrum, rote copying can be automated. When a photocopy machine copies an image, a microprocessor in the machine executes a complex set of routines. Rote copying comes about for the simple reason that, before formalizing these procedures in any computer algorithm, the programmer must have identified them all. Imagine a more complex case: a computer, digitally recording audio input, for example, a musical work, converts all the sounds of that work into video output, which is screened as moving visual patterns. The computer program effectively reprocesses the musical work, its rhythms, melodies, chords, timbres, etc., visually by playing these materials out as lines, light and dark, colors and values, etc., all moving in space and time, apparently as an abstract cinematic work. Assume, too, that the algorithm used here is in an off-the-shelf program, which the party making the conversion does not creatively supplement, say, through interactive input. The result, despite the distinct audio and video forms, is rote copying.

Moving further along the spectrum, mid-range processes such as knowledgeable reworking can only be partially automated. For example, an expert system can be used to help with the translation of a text, but the routines of such systems succeed with variable confidence factors in translating all but the simplest texts. These factors represent the probabilities that the expert system will translate given passages to the satisfaction of bilingual speakers, and they can vary depending on the difficulty of any given passage and the languages in question. Such variability will arise even in using an expert system to translate a work-a-day text like our hypothetical cookbook from French to English, both closely related languages that share largely common roots. It will every so often prove necessary to call upon the implicit knowledge of bilingual speakers to decide between proposed translations that are arguably acceptable or equally debatable.

At the far end of the spectrum, innovative recasting largely defies automation. Suppose that a text to be translated, like our example of James Joyce's Ulysses, breaks systematically with most cultural precedents. As Umberto Eco emphasizes, such a work can challenge us "to rethink the whole language, the entire inheritance of what has been said, can be said, and could or should be said." (A Theory of Semiotics, Indiana University Press, 1976, at 274.) Innovative recasting is needed to translate Joyce's novel into French, since no bilingual consensus provides rules to guide translation at most points in this seminal English text. Of course, run-of-the-mill works can also be innovatively recast, as when popular tunes became fodder for modern jazz.

A word of caution is called for at this point. Infringement analysis need not look to the actual genesis of the works in question. It is rather a matter of ascertaining to what extent there were routines available to move from plaintiff's to defendant's work at the time of the later work. To the extent that available routines did not suffice, it may be inferred that the allegedly infringing work arose thanks to processes that ranged from knowledgeable reworking to innovative recasting. The computer metaphor, just used, provides merely one way to think about such a determination. Expert testimony, as explained below, will often be helpful in making the requisite findings.

 

c. Application: Hiroshige's and Van Gogh's Works

Our framework of analysis ostensibly applies no matter what type of work is at issue. But, once again, there is a difficulty that arises from protean creativity, which manifests itself in the ever-proliferating diversity of media and genres. Let us then initially test our framework of analysis by asking whether it can be cogently applied to diverse types of works. After that, we can apply it to the works of Hiroshige and Van Gogh in question here.

On the one hand, literary works are coded in discrete terms, such as words and phrases, arranged pursuant to grammatical rules. On the other, artistic works are embodied in continuously variable materials such as line, space, light, and color, configured in space. In computer terms, a literary work can be generated using a lexicon like the ASCII code and a sequencing or syntactic program, while an artistic work can be more or less faithfully bit-mapped. The same distinction can be reconceptualized as applying between works repeatably readable from texts and works uniquely embodied in objects.

This distinction admits of both border-zone and hybrid cases. For example, a musical work is more like a literary work when it is written down in a score, but it moves into a border zone, where it becomes more like an artistic work, when it is performed. Chinese poetry forms a hybrid case when written in calligraphic form: no translation into an alphabetic language, no matter how creative, can render the resonance of such a poetic text with its visual form. How can a court assess where works fall along the spectrum of creative processes, given that works tend to be so diversely situated relative to the basic distinction between literary and artistic works?

In a literary work, like a novel, it is a matter of analyzing, most notably, how the text is sequentially translated into another language or sequentially adapted into another medium or genre. For example, Dashiell Hammett's detective story The Maltese Falcon was made into John Huston's film by dramatizing it into dialogue, acting instructions, etc., and by visually elaborating it into sets, action, camera work, etc. Making Hammett's text into a screenplay required knowledgeable cuts and adjustments to fit the story into cinematic constraints. Huston moved up the spectrum toward innovative recasting in the ways he shot, paced, and edited the film.

By contrast, it is often a subtle matter to analyze how one artistic work is transformed into another. The materials of such a work, for example, graphics, palette, etc., can be so packed together, its visual signals so saturated, that they become inextricable to the untrained eye. Nelson Goodman, whose philosophical analysis is seminal here, gives the example of a print by Hokusai, in which "[a]ny thickening or thinning of the line, its color, its contrast with the background, its size," would change the entire impact of the work. (Languages of Art: An Approach to a Theory of Symbols, Hackettt Publ., 1976, at 229.)

Against that background, return to our initial thought-experiment: From Hiroshige's prints, Van Gogh makes his studies in oil, the Japonaiseries. Of necessity, Hiroshige's Plum Garden reappears in Van Gogh's Tree, since the prior print is the subject matter of the later study. But the bare similarity of their patterns neither proves rote copying nor disproves knowledgeable reworking or innovative recasting. Recall the test for sorting out creative processes: Can any one set of known routines suffice to move from a prior to a later work? To the extent that routines do not suffice, knowledgeable reworking or even innovative recasting might be at work. Turn, then, to the movement from Hiroshige's prints to Van Gogh's studies.

Starting from Oriental models, Hiroshige's space hints at classic Western perspective and anticipates Impressionism as well. His print The Plum Garden moves the eye back from more intensely colored foregrounds to paler backgrounds, while Van Gogh's study The Tree scrambles all such inherited visual signals, for example, drawing our eye under the red sky and bursting yellow blossoms into a constricted space punctuated by angular tree trunks. A leading expert is categorical about the revolutionary step Van Gogh is taking beyond the Impressionism which formed the background for his mature work: "The result is a kind of explosion that opens the way to Expressionism." (Klaus Berger, Japonisme in Western Painting from Whistler to Matisse, translated by D. Britt, Cambridge University Press, 1992, at 131.)

Contrast another pair of works: Hiroshige's print Sudden Shower over the Great Bridge near Atake with Van Gogh's study The Bridge. The older artist clearly blocks out a simple Spring shower which falls on passers-by who are walking across a solid bridge planted in a placid river. The younger artist brushes in driving rain, a bridge uneasily suspended on moody pilings, and a turbulent river, all with the tensions of a raging storm. Our expert wonders "what the unfortunate people in the picture have done to deserve such an elemental onslaught." (Id. at 132.)

Is Hiroshige's copyright infringed by Van Gogh? On the spectrum starting at rote copying and including knowledgeable reworking, Van Gogh ultimately reaches the stage of innovative recasting. It would, however, oversimplify matters to declare in the abstract which of these processes infringes copyright and which does not. To resolve the dilemma of copyright scope, it will prove necessary to distinguish more finely between their legal consequences.

 

4. Disseminating Works: Remedies Available?

It is time to change from the dimension of creation to that of dissemination. If the court orders Van Gogh to stop painting his studies, it nips dissemination in the bud. If it orders Van Gogh's heir not to display these studies, it stops dissemination in its tracks. The impact of awarding money to Hiroshige's heir will depend on the amount of the award.

 

a. Property versus Liability Remedies

It is helpful at this point to distinguish between property and liability regimes. When a court is ready to order you off someone else's land, whenever you set foot on it, a property remedy is available. When the court merely holds you liable to pay money after you have already trespassed, a liability remedy is imposed. The optimum mix of property and liability remedies will vary according to the branch of intellectual property and the case.

Our dilemma arises precisely because Hiroshige may claim property, not merely in the prints he created, but in the later studies Van Gogh created of those prints. Hiroshige's property claim would be most decisively implemented in an injunction to preclude Van Gogh or his heirs from exercising any further property claim to dispose of his later work freely. As already pointed out, however, not only does copyright law not contain any principle, as does patent law, favoring any initial creator because of priority in time, but it is subject to freedoms of creation and expression. In reaching an appropriate mix of remedies, it will be argued, a court may resolve our dilemma, while avoiding conflicts between prior and later creators' claims.

Typically, under copyright law, the mix of remedies has erratically varied from case to case. This variation has in part arisen because the doctrines limiting copyright scope, such as the idea-expression distinction and substantiality criteria of infringement, have been neither uniform in meaning nor coherently articulated. Furthermore, in considering remedies, courts have to take account of equitable considerations that change from case to case, for example, the force of the parties' initial showings, the eventual impact of proposed remedies on the parties' respective positions, etc. It is submitted that courts should fashion and coordinate property and liability remedies within a framework of analysis such as that proposed here. Here are some guidelines:

1.  No injunction or other coercive remedies should be issued against whoever makes a solitary copy exclusively for private enjoyment or study.

2.  Courts should (a) always be ready to enjoin and otherwise provide coercive relief against rote copying for the public, (b) exercise discretion in granting injunctions and other coercive remedies concerning works generated by mid-range processes such as knowledgeable reworking, and (c) refrain from enjoining innovative recasting, as well as the dissemination of the new works thus generated, absent strong equitable reasons for stopping them.

3.  Courts may (a) impose the full range of monetary awards, including statutory or other special damages used for punitive or deterrent purposes, in cases of rote copies, especially when these are marketed with scienter, but (b) adjust actual or statutory damages, or reasonable royalties or profit shares, to the market interests at stake in any other case.

How would these guidelines apply to our hypothetical suit? Assume that the court finds that Hiroshige's prints are innovatively recast in Van Gogh's Japonaiseries. First, no order would be issued to stop the making of these studies privately; second, no order would lead to seizing these studies or enjoin publicly displaying or marketing them. Third, a money judgment could still be obtained for such marketing, and it would be measured by considering how Hiroshige's prints contributed to the appeal that Van Gogh's studies would hold for relevant audiences.

Our dilemma would be resolved: neither creativity nor feedback into cultural networks would be blocked, while both prior and later creators' rights would be protected on the marketplace. Having just glanced at how our guidelines come together, let us now look at how they may apply more particularly in other hard cases.

 

b. Applications to Different Remedies

Works of authorship display different creative processes in endlessly variable combinations and permutations. As a result, the guidelines just proposed would still lead to different mixes of remedies from case to case, although hopefully with more rhyme and reason than before. Some comments are in order on how to reach optimum mixes in complex cases. At the level of injunctions and other coercive remedies, there is an obvious difficulty. A framework has been proposed for the sliding-scale analysis of infringement. But how does it help in making all-or-nothing decisions, notably whether to enjoin or not?

Courts and legislators have to avoid invading privacy, whether of creators or of end-users of works. Refraining from coercive relief against private copying avoids the risks both of intimidating the timid muse that inspires creation in private and of violating constitutionally protected privacy interests. For example, the law does not stop computer programmers from making private copies of programs to discover methods and codes underlying these programs. Privately made studies are also indispensable for younger artists who conduct research into artistic methods in copying older works. They need freedom to get the knack of such methods manually as well as in inspecting the works visually.

Faced with public dissemination, courts protect copyright owners' markets by seizing infringing copies and by stopping copying, showings, sales, etc. It has been proposed that discretion to issue such orders be exercised with an eye to where plaintiff's and defendant's works fall along the spectrum of creative processes. Courts are most ready to enjoin or seize literal or close copies, almost always rote copies, that threaten markets the copied work targets but that bring nothing new into the marketplace. In mid-range cases, where knowledgeable reworking generates translations, adaptations, etc., discretion has to be exercised to balance risks to protected markets against risks to the feedback of new content. For example, the Ninth Circuit thought it inappropriate to enjoin the film Rear Window when that work turned out to be subject to renewal rights in the short story on which it had been based. Such an injunction would not have advanced the right-holder's market interests, and it made little sense to deny "the public the opportunity to view a classic film for many years to come." (Abend v. MCA, Inc., 863 F.2d 1465, 1479 [9th Cir. 1988], aff'd, 495 U.S. 207, 236 [1990].)

The difficulty at the level of monetary awards is not so obvious. It would seem that such awards can be easily adjusted within our framework of sliding-scale analysis. After all, the court can award more or less money to plaintiff depending on how massive or marginal defendant's taking is found to be. It is not so simple, since the measures of monetary awards are diverse: some are granted for punitive or deterrent purposes; some, to compensate plaintiff for usurped markets or other damages; some, to recover defendant's undue gains. For punitive or deterrent purposes, many courts may assess monetary awards, especially statutory or other special damages, at levels beyond the market interests at stake in the cases before them. There is no risk that the prospect of paying such extraordinary damages might discourage the release of new creations when they are levied against rote copies, for the simple reason that such copies feed nothing creative into communication networks. But there is the risk that awards out of proportion with market stakes might burden the feedback of new content in cases where reworked or recast works are found to be infringing.

These cases call for purely compensatory awards, notably damages, royalties, or profits, or else statutory damages granted in lieu of such awards. The amount of any compensatory award has to be gauged in the light of the relative impacts of plaintiff's and defendant's works on the marketplace. It is easy enough to measure actual damages in a case where plaintiff's work finds itself forsaken because defendant's work merely serves as a substitute for it on the markets it addresses. In other cases, plaintiff might not yet have put the work at issue on the market or might not have had any market success, or the work might be transformed by defendant into a derivative work that addresses different markets, for example, as when a novel is adapted to film. The key here lies in apportioning compensatory awards to the extent to which plaintiff's work contributes to the appeal of defendant's on the marketplace.

Suppose, for example, quite hypothetically, that no license had been obtained from Dashiell Hammett to film The Maltese Falcon, as John Huston did. Not only would it not be equitable, but it would undermine incentives to make and release the film, if the writer were accorded all the film profits that would be otherwise attributable to talented stars, cinematic direction and production, promotion, etc. But the court may award Hammett reasonable royalties or shares of profits for Huston's rather thorough-going use of his story, which partially accounts for the abiding appeal of the resulting classic film noir, even while the court factors out creative filming, etc., also contributing to that appeal.

 

c. Proof: the Audience Test

In the United States, juries may find facts, while other systems largely rely on judges for that task. Of course, juries do not decide on injunctive and other equitable remedies that remain the judges' responsibilities. If called in the United States, juries find whether infringement takes place and assess monetary awards if they do so find.

Perhaps, all triers of fact are tempted to act as test audiences when hearing infringement cases. Whether applied by judges or juries, audience tests give rise to a pair of related problems: first, not all triers of fact are sure to emulate accurately the responses of the relevant audiences; second, it is not always obvious who should constitute the relevant audience. To take an example of the first problem, imagine a case in which a poem is allegedly translated: to ascertain whether there is translation, much less how the appeal of the original poem finds its way into the translation, the trier of fact must itself be bilingual or hear testimony from an expert who knows the languages of both texts. As to the second problem, it brings us back to the issue of copyright scope: the extent of the markets reserved to plaintiff, along with the markets open to defendant, will depend on just this issue. But collapsing all criteria of relevant audiences into some notion of undifferentiated audience response only obfuscates the issue.

An imaginary case, introduced above, dramatizes the problem of emulating audience response: a computer converts audio input into video output. It reprocesses plaintiff's musical work, its rhythms, melodies, etc., into what seems an abstract cinematic work, namely moving and colored patterns on screen. The defendant who, without any interactive input, mechanically uses the computer, as well as the conversion algorithm running the computer, engages in rote copying. Nonetheless, the average judge or jury might not hear, and in turn not see, much similarity in appeal between plaintiff's musical work and defendant's ostensible cinematic work. As a result, a test merely asking the trier of fact to emulate some lay listener's and ordinary observer's response might not lead to any finding of infringement at all, much less damages. Return, for a moment, to the example of the allegedly translated poem: there a bilingual expert could explain to the trier of fact how linguistic rules do, or do not, account for moving from plaintiff's text in one language to defendant's in another. Similarly, in this imaginary case, an expert could focus the trier of fact on how the conversion algorithm in the program generates, from plaintiff's musical work, the full fabric defendant's ostensible cinematic work.

Consider a real case which illustrates the problem of selecting relevant audiences. The artist Jeff Koons instructed an art studio to make a sculpture, in a number of copies, of a photograph of a couple holding a string of cute puppies. The sculpture, Koons claimed, critiqued kitsch sentimentality typified in the photograph, but no defense of fair use was allowed on that basis, and infringement was found. (Rogers v. Koons, 751 F.Supp. 474 [S.D.N.Y. 1990], amended, on reh'g, 777 F.Supp. 1 [S.D.N.Y. 1991], aff'd, 960 F. 2d 301 [2d Cir. 1992], cert. denied, 506 U.S. 934 [1992].) But from whose perspective would Koons' sculpture so obviously constitute an actionable copy of the photograph, and whose markets would it usurp, if any, for purposes of assessing damages? Plaintiff's popular audience was not relevant to the extent that it did not buy the sculpture rather than the photograph; defendant had a different, more sophisticated art market, for whom his colored, three-dimensional sculpture slyly, knowledgeably, reworked the black-and-white, two-dimensional photograph at issue. For that market, did the appeal of defendant’s sculpture in any way derive from plaintiff's photograph? An expert might well testify that any equally banal image could have been appropriated for that purpose. On that basis, the result would colorably have been no damages at all.

 

5. A Caveat and Conclusion

We started with a hypothetical suit. At issue were Hiroshige's wood-block prints. Relief was sought against Van Gogh's studies in oils of these prints. This case dramatizes the dilemma implicit in all cases where prior works form bases for later works. On the one hand, denying relief would ignore copyright rationales altogether; on the other, overly stringent remedies could betray these rationales.

We have here reframed infringement analysis with an eye to resolving this dilemma. But our framework is not definitive, but merely an attempt to pull together various strands of infringement analysis in diverse copyright systems. Furthermore, all infringement analysis has to look to current culture for its premises concerning creative processes, aesthetic perceptions, and economic realities. Indeed, as a creature of our culture, copyright law can only proceed critically from just such premises in the light of its own basic rationales. No doubt, future aesthetic and economic insights will help to reconceptualize infringement analysis in new and better forms.

With that caveat in mind, a pair of points should be stressed in concluding. On the level of theory, courts need better analytic tools than shifting sets of vague and disparate tests and criteria of protectability and infringement. The attempt to bring underlying doctrines together into some coherent framework of analysis at least helps to debug corresponding tests and criteria and to optimize infringement findings and remedies in the cases. On the level of practice, courts ought not content themselves with simply finding infringement vel non but, in hard cases, would do well to discern more finely how plaintiffs' works are reworked or recast into defendants' works. In that light, it is submitted, they could grant remedies more consistently with copyright rationales.