How the Meaning of Reproduction has Changed Over Time

Copyright is usually thought of as the right to control the reproduction of literary and creative works. Of course, today, copyright law does far more than this, but it is useful to begin with the reproduction right as the conceptual archetype of copyright law. What does it mean to “reproduce” or “copy” a work? Do partial copies count? What if a new work incorporates substantial parts of an existing work, but also extends that work in significant ways? What about temporary instantiations that exist for less than a second or two? What about versions of the work that are only intelligible by machines, or only useful as the part of a machine?

This short chapter will canvass some of the ways our concept of copying in copyright law has changed over time.

The Scope of Copyright in the Premodern Era

Even in the very early days of copyright law, the rights of the author were not strictly limited to verbatim copies of particular texts, but they were much narrower than our current conception of copyright. In the early days of copyright under the Statute of Anne, 1710, 8 Ann., c. 19, §1 (Eng.) “the author of any book or books” or his assignees, was entitled to “the sole right and liberty of printing such book and books.” Construed literally, the Act regulated only exact and entire reprinting and had nothing to say about making adaptations and derivative works based on other people’s books. Indeed, making a shortened version or abstract of a longer text called an abridgement was common in the premodern era (by which we mean from 1710 to roughly 1850). However, the lawful scope of abridgement was contested throughout this period.

Consider Gyles v. Wilcox, one of the earliest reported decisions on the doctrine of fair abridgement, a doctrine that eventually evolved into the modern law of fair use in the United States and “fair dealing” in many other jurisdictions.

Gyles v. Wilcox (1740) 26 ER 489

Lord Hardwicke (the Lord Chancellor)

… As to what has been said by Mr. Attorney General of the acts being a monopoly, and therefore ought to receive strict construction, I am quite of a different opinion, and that it ought to receive a liberal construction, for it is very far from being a monopoly, as it is intended to secure the property of books in the authors themselves, or the purchasers of the copy, as some recompence for their pains and labour in such works as may be of use to the learned world.

The question is, Whether this book of the New Crown Law, which the defendant has published, is the same with Sir Matthew Hale’s Historia Placitorum Coronae, the copy of which is now the property of the plaintiff.

Where books are colourably shortened only, they are undoubtedly within the meaning of the act of Parliament, and are a mere evasion of the statute, and cannot be called an abridgment.

But this must not be carried so far as to restrain persons from making a real and fair abridgment, for abridgments may with great propriety be called a new book, because not only the paper and print, but the invention, learning, and judgment of the author is shewn in them, and in many cases are extremely useful, though in some instances prejudicial, by mistaking and curtailing the sense of an author.

If I should extend the rule so far as to restrain all abridgments, it would be of mischievous consequence, for the books of the learned, les Journels des Scavans, and several others that might be mentioned, would be brought within the meaning of this act of parliament.

In the present case it is merely colourable, some words out of the Historia Placitorum Coronae are left out only, and translations given instead of the Latin and French quotations that are dispersed through Sir Matthew Hale’s works; yet not so flagrant as the case of Read v. Hodges, for there they left out whole pages at a time; but I shall not be able to determine this properly, unless both books were read over, and the case fairly stated between the parties.

[Rather than send it to a jury,] I think is one of those cases where it would be much better for the parties to fix upon two persons of learning and abilities in the profession of the law, who would accurately and carefully compare them, and report their opinion to the court.

Notes and questions

(1) What happened to the case? Lord Hardwicke continued the injunction, pending a master of the court’s report on the similarities between the two works. The result is not contained in the report of the case, but it is discussed in Tonson v. Walker, (1752) 36 Eng. Rep. 1017 (Ch.). Apparently, the court-assisted arbitration proceeding led to an agreement that the defendant’s work was a fair abridgment outside the Statute of Anne’s scope. Accordingly, the injunction was dissolved.

(2) Notice how Lord Hardwicke rejects the argument that the Statute of Anne should be narrowly construed as a monopoly. Lord Hardwicke goes beyond the literal words of the Statute of Anne to the purposes it entailed and rejects any “mere evasion of the statute.” So it is worth asking whether the decision in Gyles expands or contracts the scope of copyright? In The Prehistory of Fair Use, Matthew Sag argues:

Although Gyles is often cited as the origin of the fair use doctrine in England and has generally been received as a pro-abridgment decision, Lord Hardwicke’s reasoning gave as much to copyright owners as it took away. On the one hand, Gyles confirmed the legality of some abridgments (those described as fair). Yet it also entrenched a broad purposive reading of the Statute of Anne and condemned another set of abridgments (those deemed unfair) as infringing copyright.

(3) As Gyles illustrates, the premodern copyright cases tended to set the dividing line between infringing and noninfringing abridgements with a view to two factors. First, the degree of intellectual labor contributed by the defendant. This was not simply an assessment of the amount of work added by the defendant, usually the question was presented as whether the defendant’s work should fairly be seen as a new work. The second factor was whether the defendant’s work was likely to substitute for the plaintiff’s original work.

The modern doctrine of fair use has retained some of the decisional structure of the premodern English abridgment cases. An assessment of fair use is significantly influenced by the degree of newness of the defendant’s work in the sense of the question of whether the defendant’s use was transformative (this is part of the first fair use factor). Modern fair use cases are also highly sensitive to concerns of market substitution (this is part of the fourth fair use factor).

***

Over time the scope of what was considered an infringing copy expanded. One particularly important case to note in this expansion was Folsom v. Marsh, 9 F. Cas. 342, 349 (C.C.D. Mass. 1841). Folsom v. Marsh is often cited as the beginning of the modern United States doctrine of fair use, but the term does not actually appear in the decision. The earliest American report that actually uses the expression “fair use” is Lawrence v. Dana, 15 F. Cas. 26, 60 (C.C.D. Mass. 1869) (No. 8136). And as we have just seen, the fair use doctrine in the United States was in many ways a continuation of earlier English authorities on fair abridgment.

The dispute in Folsom v. Marsh was over two different books about George Washington. The plaintiffs, Folsom, Wells & Thurston, owned the rights to Jared Sparks’ extensive 12-volume collection of Washington’s writings. The defendant, Reverend Charles Upham, created a shorter book, The Life of Washington, mainly using extracts from Washington’s writings, with about a third of the content coming from Sparks’ collection. Upham’s publisher argued that their book was a fair abridgment and didn’t violate copyright because once Washington’s writings were published, anyone could use them to create a new work. However, despite similar abridgments having been allowed by numerous English authorities, the court did not accept the argument.

Justice Story framed the issue in terms of the need to find a line between thinly disguised piracy and acceptable quotation and stressed the need to consider “the comparative use made in one of the materials of the other; the nature, extent, and value of the materials thus used; the objects of each work; and the degree to which each writer may be fairly presumed to have resorted to the same common sources of information, or to have exercised the same common diligence in the selection and arrangement of the materials.”

The court noted that 353 out of 866 pages in Upham’s work were identical to passages in Sparks’ work, with 319 of those pages consisting of Washington’s letters that had not been published before Sparks’ edition. Justice Story held that the letters of Washington were proper subjects of copyright and that the rights to those letters had been acquired by the plaintiffs. Thus, the central question was whether this extensive use of Washington’s letters in Upham’s work constituted copyright infringement. Justice Story recognized that Upham’s work could not be classified as a mere abridgment or simple compilation of Sparks’ work. Nonetheless, while Upham’s narrative was either original or derived from common sources, his verbatim use of entire letters from Sparks’ collection, chosen for their significance, raised the issue of whether this constituted an infringement.

Folsom v. Marsh, 9 F. Cas. 342, 349 (C.C.D. Mass. 1841) (No. 4901)

Circuit Justice Story

… The question, then, is, whether this is a justifiable use of the original materials, such as the law recognizes as no infringement of the copyright of the plaintiffs. It is said, that the defendant has selected only such materials, as suited his own limited purpose as a biographer. That is, doubtless, true; and he has produced an exceedingly valuable book. But that is no answer to the difficulty. It is certainly not necessary, to constitute an invasion of copyright, that the whole of a work should be copied, or even a large portion of it, in form or in substance. If so much is taken, that the value of the original is sensibly diminished, or the labors of the original author are substantially to an injurious extent appropriated by another, that is sufficient, in point of law, to constitute a piracy pro tanto. The entirety of the copyright is the property of the author; and it is no defence, that another person has appropriated a part, and not the whole, of any property. Neither does it necessarily depend upon the quantity taken, whether it is an infringement of the copyright or not. It is often affected by other considerations, the value of the materials taken, and the importance of it to the sale of the original work.

In short, we must often, in deciding questions of this sort, look to the nature and objects of the selections made, the quantity and value of the materials used, and the degree in which the use may prejudice the sale, or diminish the profits, or supersede the objects, of the original work. Many mixed ingredients enter into the discussion of such questions. In some cases, a considerable portion of the materials of the original work may be fused, if I may use such an expression, into another work, so as to be undistinguishable in the mass of the latter, which has other professed and obvious objects, and cannot fairly be treated as a piracy; or they may be inserted as a sort of distinct and mosaic work, into the general texture of the second work, and constitute the peculiar excellence thereof, and then it may be a clear piracy. If a person should, under color of publishing “Elegant Extracts” of poetry, include all the best pieces at large of a favorite poet, whose volume was secured by a copyright, it would be difficult to say why it was not an invasion of that right, since it might constitute the entire value of the volume. …

In the present case, I have no doubt whatever, that there is an invasion of the plaintiffs’ copyright; I do not say designedly, or from bad intentions; on the contrary, I entertain no doubt, that it was deemed a perfectly lawful and justifiable use of the plaintiffs’ work. But if the defendants may take three hundred and nineteen letters, included in the plaintiffs’ copyright, and exclusively belonging to them, there is no reason why another bookseller may not take other five hundred letters, and a third, one thousand letters, and so on, and thereby the plaintiffs’ copyright be totally destroyed. Besides; every one must see, that the work of the defendants is mainly founded upon these letters, constituting more than one third of their work, and imparting to it its greatest, nay, its essential value. Without those letters, in its present form the work must fall to the ground. It is not a case, where abbreviated or select passages are taken from particular letters; but the entire letters are taken, and those of most interest and value to the public, as illustrating the life, the acts, and the character of Washington. It seems to me, therefore, that it is a clear invasion of the right of property of the plaintiffs, if the copying of parts of a work, not constituting a major part, can ever be a violation thereof; as upon principle and authority, I have no doubt it may be. If it had been the case of a fair and bona fide abridgment of the work of the plaintiffs, it might have admitted of a very different consideration.

I have come to this conclusion, not without some regret, that it may interfere, in some measure, with the very meritorious labors of the defendants, in their great undertaking of a series of works adapted to school libraries. But a judge is entitled in this case, as in others, only to know and to act upon his duty. I hope, however, that some means may be found, to produce an amicable settlement of this unhappy controversy. The report of the master must stand confirmed, and a perpetual injunction be awarded, restraining the defendants, their agents, servants and salesmen, from farther printing, publishing, selling, or disposing of any copy or copies of the work complained of; the “Life of Washington,” by the Rev. Charles W. Upham, containing any of the three hundred and nineteen letters of Washington, stated in the report of the master, and never before published; and that it be referred to a master, to take an account of the profits made by the defendants, in the premises; with leave for either party to apply to the court for farther directions.

Notes and questions

(1) In a 2008 article in the Yale Law Journal, The Ideology of Authorship Revisited: Authors, Markets, and Liberal Values in Early American Copyright, Oren Bracha argues that Folsom v. Marsh was a pivotal component of American copyright law’s transformation in the nineteenth century. Over the course of the nineteenth century, Bracha contends copyright changed from an exclusive right to make verbatim copies of particular texts to an abstract right of general control in which the only boundaries of a work were identified vis-à-vis its market value. Bracha argues that the concept of fair use announced in Folsom v. Marsh was a fundamental change in copyright’s baseline.

(2) In The Prehistory of Fair Use, Matthew Sag argues that Bracha’s characterization is correct in broad strokes, but possibly over-states the abruptness of the shift. He points out that fair abridgement was lawful in the United States prior to Folsom v. Marsh and abridgement did not become suddenly unlawful after Folsom v. Marsh. Nonetheless, the line between fair and unfair did seem to have shifted around this time as evidenced by the 1870 amendment to the U.S. Copyright Act that gave authors the right to control translations and dramatizations of their works.

(3) Why did the scope of fair abridgment reduce over time? In The Prehistory of Fair Use Matthew Sag argues that premodern cases illustrate a half-formed notion of the derivative work right: unauthorized derivatives could be enjoined to defend the market of the original work, but they did not constitute a separate market unto themselves. Folsom departs from the earlier English cases in that it recognizes derivatives as inherently valuable—not just something to be enjoined to defend the original work against substitution. This subtle shift is important because while the boundaries of a defensive derivative right can be ascertained vis-à-vis the defendant’s work on the plaintiff’s original market, the boundaries of an offensive derivative right can only be determined in the context of some other limiting principle. The expansion of the derivative right may reflect broader shifts in 19th century thinking and a desire to protect value as an abstract concept as opposed to merely protecting title. It might also be due to an anchoring effect: as more and more derivatives were enjoined defensively, courts and copyright owners began to see these derivatives as part of the authors’ inherent rights in their creations.

***

The scope of the reproduction right did not change overnight. Consider the famous case of Kalem Co. v. Harper Brothers, 222 U.S. 55 (1911). Harper Brothers owned the copyright in Ben-Hur: A Tale of the Christ, a very popular novel written by Lew Wallace. Kalem made a silent film based on Ben Hur without Harper Brothers’s authorization. The Supreme Court’s decision focused on whether the right to “dramatize” the work was infringed when Kalem gave copies of the film to movie theatres on the understanding that they would display the film. The Supreme Court said that it was:

In some cases where an ordinary article of commerce is sold nice questions may arise as to the point at which the seller becomes an accomplice in a subsequent illegal use by the buyer. It has been held that mere indifferent supposition or knowledge on the part of the seller that the buyer of spirituous liquor is contemplating such unlawful use is not enough to connect him with the possible unlawful consequences, but that if the sale was made with a view to the illegal resale the price could not be recovered. But no such niceties are involved here. The defendant not only expected but invoked by advertisement the use of its films for dramatic reproduction of the story. That was the most conspicuous purpose for which they could be used, and the one for which especially they were made. If the defendant did not contribute to the infringement it is impossible to do so except by taking part in the final act. It is liable on principles recognized in every part of the law.

This decision is an early example of secondary liability in copyright law, but arguably, the most interesting thing about it is what is missing. It seems remarkable now, but the court of appeals in Harper & Bros. v. Kalem Co., 169 F. 61 (2d Cir. 1909) held in essence that a film based on a book was not a copy of the book and the Supreme Court did not disagree with this conclusion. The court of appeals said:

The series of photographs taken by the defendant constitutes a single picture, capable of copyright as such; and as pictures only represent the artist’s idea of what the author has expressed in words, they do not infringe a copyrighted book or drama, and should not as a photograph be enjoined.

The notion that “pictures only represent the artist’s idea of what the author has expressed in words” and thus fall on the idea side of the idea-expression dichotomy fails to grasp that the pictures in sequence tell a story and that story was the Ben Hur story. The case illustrates how limited the scope of copyright was compared to what it is now.

The express statutory right to “prepare a derivative work based upon the copyrighted work” is a recent addition to American copyright law. Section 1 of the Copyright Act of 1790 gave domestic authors who complied with certain formalities the “sole right and liberty” to print, reprint, publish, and to vend (sell or offer for sale) qualifying maps, charts, and books, for a term of fourteen years. Section 2 added an importation right, but the Act did not expressly recognize any rights with respect to adaptations or derivative works.

In 1909, Congress provided a more expansive, but still specific, list of derivatives that authors were entitled to control. The Copyright Act of 1909, Section 1(b) conferred an exclusive right (among other things) to translate literary works, to dramatize nondramatic literary works, and to novelize a dramatic work. It is important to note the structure of the 1909 Act. The Act not only identified specific types of derivatives that authors were entitled to control—translations, dramatizations and novelizations; it also tied each type of derivative to the type of work being adapted. It was not until the Copyright Act of 1976 that United States copyright law conferred an express general right to control the preparation of derivative works based upon copyrighted works.

Must a Copy be Human-Readable?

If a tree falls in the forest, does it really make a sound? Is something a copy for the purposes of copyright law if it can’t be read or understood by human beings? You might think this question is about digital technology, but it was actually first raised over a century ago in relation to player pianos.

Emile Berliner invented the gramophone in 1887. Unlike earlier sound recording devices, such as Thomas Edison’s phonograph which used cylindrical recordings, the gramophone used flat discs to record and play sound. This eventually became the dominant format for recorded music for the analog era. Player pianos were not quite as significant as the gramophone, but they hold an important place in copyright law history. Player pianos also allowed for the mechanical reproduction of music, but in a slightly different way using perforated rolls of paper (piano rolls). For the authors of musical works, the significance of these inventions was that they could reproduce performances of compositions without the need for a live musician, and thus without the need for printed sheet music.

The question of whether mechanical encodings of music that were not human-readable should be thought of as “copies” came to the Supreme Court in White-Smith Music Publishing Co. v. Apollo Co., 209 U.S. 1 (1908).

White-Smith Music Publishing Co., a music publisher, held the copyright for various musical compositions. White-Smith argued that Apollo’s use of their compositions in the production of piano rolls was a form of unauthorized reproduction. The Copyright Act in force at the time primarily addressed written or printed materials such as books, maps, and sheet music.

The Supreme Court’s opinion explained the process of creating perforated music rolls used in mechanical piano attachments like the pianola. The rolls consisted of perforated sheets that controlled air pressure, which operated the mechanism to sound the notes as the roll moved over the tracker board. The Court outlined three main methods of producing these rolls: first, an arranger marks and cuts perforations on paper based on the musical score, creating an original stencil used to produce multiple rolls; second, a new roll can be made by reproducing an existing perforated roll; and third, an automatic recording device attached to a piano can create a perforated matrix that generates the music roll. Skilled workers use these methods to convert sheet music into perforations that, with the aid of the mechanical system, reproduce the recorded music.

White-Smith Music Publishing Co. v. Apollo Co., 209 U.S. 1 (1908)

Mr. Justice Day delivered the opinion of the court.

In the last analysis this case turns upon the construction of a statute, for it is perfectly well settled that the protection given to copyrights in this country is wholly statutory. When we turn to the consideration of the act it seems evident that Congress has dealt with the tangible thing, a copy of which is required to be filed with the Librarian of Congress, and wherever the words are used (copy or copies) they seem to refer to the term in its ordinary sense of indicating reproduction or duplication of the original. …

What is meant by a copy? We have already referred to the common understanding of it as a reproduction or duplication of a thing. A definition was given by Bailey, J., in West v. Francis. He said: “A copy is that which comes so near to the original as to give to every person seeing it the idea created by the original.”

Various definitions have been given by the experts called in the case. The one which most commends itself to our judgment is perhaps as clear as can be made, and defines a copy of a musical composition to be “a written or printed record of it in intelligible notation.” It may be true that in a broad sense a mechanical instrument which reproduces a tune copies it; but this is a strained and artificial meaning. When the combination of musical sounds is reproduced to the ear it is the original tune as conceived by the author which is heard. These musical tones are not a copy which appeals to the eye. In no sense can musical sounds which reach us through the sense of hearing be said to be copies as that term is generally understood, and as we believe it was intended to be understood in the statutes under consideration. A musical composition is an intellectual creation which first exists in the mind of the composer; he may play it for the first time upon an instrument. It is not susceptible of being copied until it has been put in a form which others can see and read. The statute has not provided for the protection of the intellectual conception apart from the thing produced, however meritorious such conception may be, but has provided for the making and filing of a tangible thing, against the publication and duplication of which it is the purpose of the statute to protect the composer.

Also it may be noted in this connection that if the broad construction of publishing and copying contended for by the appellants is to be given to this statute it would seem equally applicable to the cylinder of a music box, with its mechanical arrangement for the reproduction of melodious sounds, or the record of the graphophone, or to the pipe organ operated by devices similar to those in use in the pianola. All these instruments were well known when these various copyright acts were passed. Can it be that it was the intention of Congress to permit them to be held as infringements and suppressed by injunctions?

After all, what is the perforated roll? The fact is clearly established in the testimony in this case that even those skilled in the making of these rolls are unable to read them as musical compositions, as those in staff notation are read by the performer. It is true that there is some testimony to the effect that great skill and patience might enable the operator to read his record as he could a piece of music written in staff notation. But the weight of the testimony is emphatically the other way, and they are not intended to be read as an ordinary piece of sheet music, which to those skilled in the art conveys, by reading, in playing or singing, definite impressions of the melody.

These perforated rolls are parts of a machine which, when duly applied and properly operated in connection with the mechanism to which they are adapted, produce musical tones in harmonious combination. But we cannot think that they are copies within the meaning of the copyright act.

It may be true that the use of these perforated rolls, in the absence of statutory protection, enables the manufacturers thereof to enjoy the use of musical compositions for which they pay no value. But such considerations properly address themselves to the legislative and not to the judicial branch of the Government. As the act of Congress now stands we believe it does not include these records as copies or publications of the copyrighted music involved in these cases.

Notes and questions

(1) According to the Court, what is the difference between sheet music and a piano roll? Why does the Court take the narrow view that piano rolls were not copies?

(2) Congress responded to the White-Smith decision by amending the definition of a copy, making it clear that actual human readership or direct human readability were not required. Instructions encoded in machine-readable memory that could reproduce the sound but were unintelligible to humans qualified as “copies” in the relevant sense. The modern Copyright Act takes the same view. It gives copyright owners the exclusive right “to reproduce the copyrighted work in copies” and defines copies in Section 101 as “material objects . . . in which a work is fixed . . . and from which the work can be perceived, reproduced, or otherwise communicated.” Why do you think it is important to treat mechanical copies that humans can’t read directly as copies?

(3) The question White-Smith raised has returned in an unfamiliar shape. Is a generative AI model trained on copyrighted works itself a copy of those works? The difficulty is that we are used to thinking of a copy as something that exists in a reasonably well-defined state. A piano roll is an object; so is a hard drive sector. A trained model is a messy statistical artifact, billions of parameters that encode no particular work in any particular place, but which, given the right provocation, can sometimes reproduce one.

The question has now been litigated. In GEMA v. OpenAI, Case No. 42 O 14139/24 (Landgericht München I, 11 November 2025), a German court held that the memorization of song lyrics in a language model’s parameters is itself an act of reproduction, and that regurgitation in an output is a further reproduction; the text and data mining exception, in the court’s view, covers only the analytical phase of training. The court treated it as beside the point that the content exists “only in the language models in the form of probability values and distributed across various parameters,” asking instead whether the model could reproduce the work in recognizable form. The judgment is not final and is expected to be reviewed on appeal, and it is of course a decision under German and EU law rather than United States law.

A. Feder Cooper and James Grimmelmann argue that the mere fact that there is no identifiable copy residing in the dispersed weights that comprise a language model is not enough to deny the charge of memorization. The fact of retrieval shows that the model contains a copy, even if the existence of that copy can only be proven inferentially. On this view, even highly contingent or prompt-dependent regurgitation demonstrates that the model internally encodes expressive content that meets the statutory threshold for a “copy.” For more, see A. Feder Cooper & James Grimmelmann, The Files Are in the Computer: On Copyright, Memorization, and Generative AI, 100 Chicago-Kent Law Review 141 (2025).

Subconscious Copying

Bright Tunes Music v. Harrisongs Music, 420 F.Supp. 177 (S.D.N.Y. 1976)

In 1970 George Harrison, famed guitarist of The Beatles, released a solo album including a track entitled, “My Sweet Lord”. The copyright owner of the 1962 hit “He’s So Fine”—who had acquired the rights from the original authors—sued Harrison alleging that My Sweet Lord was copied from He’s So Fine.

The songs are mostly different, but they have some features in common. As the court described it:

The case is notable for its exploration of “subconscious copying” as copyright infringement. The judge did not contend that Harrison had set out to copy He’s So Fine, but he concluded that he had nonetheless done so subconsciously.

Notes and questions

(1) Is it fair to hold someone liable for unintentionally replicating aspects of another’s work, given that common chord progressions, rhythms, and melodies are shared across many songs?

(2) Why is the judge so sure that the author of He’s So Fine came up with that sequence of notes from his own imagination, but that George Harrison could not have done the same?

Is a Temporary Reproduction a Copy Under the Copyright Act?

If Amy sends Bob an email attaching a copyrighted book as a PDF file, it seems clear that when Bob clicks the download button, he has created a copy of the book. But did the phone company that provided Bob’s internet access also make a copy when it allowed the file to pass along its network as a series of packets of information? What if a copy of those files was stored temporarily part way along the transfer process?

If Amy sets up a buffer that takes broadcast television signals, holds them for 1.2 seconds, and then bounces them onto Bob in a new format, has she made “a copy” of the TV program? Or is that just a performance?

If Amy buys an eBook from Bob under a license that allows for her to make a single copy, does she need a new license every time her book reader calls a piece of the file up from long-term memory into short-term memory so it can be displayed?

These questions all raise the issue of how copyright law deals with temporary or transitory copies.

The RAM Copy Doctrine

The short-term memory in a computer is typically comprised of RAM (random access memory). Beginning with MAI Systems Corporation v. Peak Computer, Inc., 991 F.2d 511 (9th Cir. 1993), courts have consistently held that the creation of such “RAM copies” is an act of reproduction under the Copyright Act.

MAI sold computer hardware that required proprietary software to operate. MAI licensed the software to its customers under strict terms, allowing them to use the software only with MAI’s permission. MAI sold services relating to its hardware but faced competition from Peak Computer. Rather than competing on quality or price of service, MAI argued that Peak’s technicians were violating copyright law because in the course of repairing MAI systems, Peak technicians would load MAI’s operating software into the computer’s RAM to test and troubleshoot the system. This step was necessary to diagnose problems and fix the computers.

MAI argued that when Peak technicians loaded the MAI software into RAM, it constituted a “copy” under copyright law and the Ninth Circuit agreed.

MAI Systems Corporation v. Peak Computer, Inc., 991 F.2d 511 (9th Cir. 1993)

RAM copy doctrine revised

The RAM copy doctrine announced in MAI v. Peak had far-reaching effects, but because the copies made by Peak clearly lasted for some appreciable period of time, the decision left some ambiguity as to whether all temporary copies counted as “copies” under the Copyright Act. The Second Circuit addressed that question in Cartoon Network LP, LLLP v. CSC Holdings, Inc. (also known as the Cablevision case). Cablevision operated a remote storage digital video recording (RS-DVR) system which allowed customers to record television programs without needing a physical DVR in their homes. Instead of storing the recorded content on a local device (like a set-top DVR box or an old-fashioned VCR), the recorded shows were stored on Cablevision’s central servers. One of the plaintiff’s arguments in this case was that when Cablevision ran television programs through a temporary buffer (basically a holding pattern inside the computer) it made an unauthorized copy.

Cablevision had a statutory license that allowed it to take live broadcast signals and send them to its subscribers. To provide the additional service of recording programs on the customer’s behalf, Cablevision split the incoming signal into two streams. One stream was sent directly to subscribers’ televisions, while the other stream was sent to a buffer where it was processed (i.e., encoded and routed) for storage on the remote server if a subscriber had requested that recording. This buffering process lasted less than 1.2 seconds.

Cartoon Network v. CSC Holdings, 536 F.3d 121 (2d Cir. 2008)

Notes and questions

(1) The Second Circuit was not alone in seeking to limit the broad sweep of the RAM copy doctrine. CoStar Group Inc. v. LoopNet, Inc., 373 F.3d 544 (4th Cir. 2004) held that while temporary reproductions “may be made in this transmission process, they would appear not to be ‘fixed’ in the sense that they are ‘of more than transitory duration.’”