 Cases like cable vision in which the parties struggle over whether what the defendant has produced constitutes a copy are rare. Much more common are cases in which the parties struggle over whether the character and amount of material copied by the defendant from the plaintiff's work rise to the level necessary to constitute infringement. Once again, the burden is on the plaintiff. The plaintiff must show not only, as we've seen, that the defendant copied from the plaintiff's copyrighted work, but also that the copying, assuming it to be proved, went so far as to constitute improper appropriation. There are three more or less discreet ways in which the plaintiff can do so. First, he can show that the defendant made a verbatim copy of his entire work. This is what Melville Nimmer helpfully refers to as comprehensive copying. This happens more often than you might think. Examples include peer-to-peer sharing of sound recordings and the comprehensive replication of digital copies of films that were at issue in the cable vision case, which we just discussed. So long as the plaintiff's work qualifies for copyright protection, applying the standards we discussed in the first lecture, comprehensive copying of this sort easily gives rise to a violation of section 106.1. Things get a bit more complicated when what the defendant has copied verbatim is not the entire work, but a piece or slice of the plaintiff's work. Nimmer refers to this as, quote, fragmented literal similarity, close quote. An example can be found in this lecture. You'll recall that I illustrated the principle of subconscious copying by playing for you a segment of the chiffon's recording of Donald Mack's composition, He's So Fine. In order to enable you to hear that recording, I had to create a copy of that segment in the recording of this lecture, which you are now watching and listening to. I did not copy the entire song, so I've not engaged in comprehensive copying. Instead, I copied a piece of it, specifically a piece long enough for you to assess the plausibility of the court's conclusion that George Harrison had innocently copied Mack's composition. Thus I have engaged in, quote, fragmented literal copying, both of Mack's composition and of the chiffon's sound recording of that composition. To prevail in a fragmented literal similarity case, the plaintiff must show two things, in addition, as we've seen, to the fact of copying. First, he must show that the slice the defendant took includes some copyrightable expression, not merely ideas or facts. All the factors we discussed in lecture number one concerning the idea expression distinction become relevant here. For example, if the slice the defendant took consists of a sentence that is the only effective way of conveying a particular idea, then pursuant to the merger doctrine, the taking of that sentence does not constitute copyright infringement. Can I escape liability on this basis for copying a slice of He's So Fine? No. There's plenty of copyrightable expression in the 30-second slice that I appropriated. Second, the plaintiff must show that the slice that the defendant took is big enough. How big is that? The adjective most often used by the courts is substantial. The portion taken must be a substantial part of the plaintiff's work, measured both qualitatively and quantitatively. Relatively small slices have been deemed sufficient to pass this test, provided that they are sufficiently important to the plaintiff's creation. For example, in the 1980 Iowa State Research case, the defendant, ABC, broadcast as part of its coverage of the Olympics two brief excerpts from the plaintiff's documentary film examining the life of a wrestler. One expert was two and a half minutes long, the other 12 seconds, and the judgment of the second circuit, that was plenty. A willingness to find liability upon the taking of relatively small excerpts is not limited to the United States. For example, the courts in Australia have proven to be similarly protective of plaintiffs. In a recent high-profile case, the plaintiff owned the copyright in Crookapurl, a nursery rhyme originally written by Marion Sinclair in 1932. It was popular in Australia for decades, and indeed I can remember my mother singing it to me in the United States. The key segment of the song goes like this. I'm going to do my best here. Crookapur sits in the old gum tree, Mary, Mary, King of the bush is he. My apologies to Miss Sinclair. In the late 1970s, the Australian rock group Men at Work composed and recorded an equally famous, perhaps more famous song called Down Under. I'm going to play a short passage of that song. Listen carefully for the flute riff that comes near the end. Appears several other times in Down Under, and each time is equally short. George Ham, one of the members of Men at Work, added this riff to the song in order to inject some, quote, Australian flavor, close quote, into it. In the ensuing copyright litigation, Ham testified that, like George Harrison, he was not aware that he was lifting this segment from Crookapurl. Somewhat more specifically, here's how Judge Jacobson, who oversaw this case, summarized Ham's affidavit, quote, he said that the flute section, which he added, fitted rhythmically to Down Under, and the percussion drum section at the start of the song, which is in fact played on beer bottles with different amounts of water in them. Mr. Ham pointed out that the lyrics, vocal melody, chords, and bass lines were already established when he heard Down Under. He looked for a complementary part for his instruments, and especially one which fell into the, quote, tongue-in-cheek, quote, nature of the song. He described the flute line as an Aussie cliché melody, or what he thought was an Irish-Australian-style melody. Now, Judge Jacobson did not definitively resolve the question of whether Ham had been aware when he inserted the riff that had been taken from Crookapurl, apparently assuming that, as we saw on the George Harrison case, subconscious copying is actionable. The decisive issue in Judge Jacobson's judgment was the amount taken. Here's how the court expressed the relevant standard. Copywriting infringement arises when a defendant, quote, has copied a substantial part of the copyrighted work. The question whether he has copied a substantial part depends much more on the quality than on the quantity of what he has taken, close quote. That standard, the judge ruled, had been met in this case. The remedy? The defendants had to pay, the plaintiffs, 5% of the royalties they had earned on the song since 2002 and into the future. In 2011, the highest court in Australia refused to hear an appeal from the ruling, leaving Jacobson's decision intact. This case had a tragic coda. In April of 2012, George Ham died at age 58 of an apparent drug overdose. It seems that one factor in his decline was his dismay that his legacy had been permanently tarnished by the judgment that he had copied another musical work, suggesting that more important for him than the financial liability arising out of this case was the reputational harm it generated. So, to summarize, in a case involving fragmented literal similarity, the plaintiff must prove that the defendant took a substantial portion of the plaintiff's work, measured qualitatively and quantitatively. In practice, that requirement is forgiving, permitting plaintiffs to recover for the taking of quite small pieces. There's a limit, however. That limit is sometimes referred to as the, quote, de minimis doctrine. An excellent description and illustration of this doctrine can be found in Judge Laval's 2001 opinion in the case of Davis versus the Gap. Quote, the de minimis doctrine is rarely discussed in copyright opinions because suits are rarely broad over trivial instances of copying. Nevertheless, it's an important aspect of the law of copyright. Trivial copying is a significant part of modern life. Most honest citizens in the modern world frequently engage without hesitation in trivial copying that, but for the de minimis doctrine, would technically constitute a violation of the law. We do not hesitate to make a photocopy of a letter from a friend to show to another friend or of a favorite cartoon to post on the refrigerator. Waiters at a restaurant sing happy birthday at a patron's table. When we do such things, it's not that we are breaking the law but unlikely to be sued given the high cost of litigation. Because of the de minimis doctrine and trivial instances of copying, we are in fact not breaking the law. If a copyright owner were to sue the makers of trivial copies, judgment would be for the defendants. The case would be dismissed because trivial copying is not an infringement. Now, one warning about this otherwise illuminating passage. Judge Laval's example in the middle of this paragraph about recording television programs while we are out is not strictly speaking an example of de minimis copying. Rather, as I mentioned in connection with the cable vision case, that's an example of fair use, a doctrine that Judge Laval himself has done more to reshape and invigorate than any other judge. But if you exclude that one instance, the passage in front of you nicely illustrates the general and unsurprising principle. The law will not penalize the taking of trivial amounts. There's one important exception to the de minimis doctrine. It's illustrated by the facts of the 2005 case of Bridgeport Music. The plaintiff in that case owned the copyright and a sound recording of the song Get Off Your Ass and Jam, which was originally made by George Clinton Jr. and the Funkadelics. Clinton is pictured in the slide on your screen. The Funkadelics recording of Get Off opens with a solo guitar riff that contains three notes and lasts a total of four seconds. Here it is. Gangsta rap group NWA, depicted on this slide, subsequently recorded a song entitled A Hundred Miles and Running, which contained a sample from the Funkadelics recording. Specifically, NWA copied two seconds of the four-second riff, lowered the pitch, looped that slice, and extended it to 16 beats. The loop created by NWA lasts a total of seven seconds. That seven-second segment, derived from the original recording, appears in five places in 100 miles. I'm going to play one of those five appearances. Listen carefully, because the segment appears in the background, and it's hard to hear. A word of warning in the foreground of this excerpt are some explicit lyrics that may offend some listeners. If you don't wish to hear them, you should skip over the next 20 seconds of this lecture. Despite the brevity of the sample taken from the Funkadelics recording, the copyright owner brought an infringement suit. Really, you may think. To be sure, as the district court that heard the case observed, the segment taken by the defendants, though tiny, was arguably important to the defendant's work. In the court's words, quote, the portion of the song at issue here is an arpeggiated chord, that is, three notes that is struck together, comprise a chord, but are instead played one at a time in very quick succession, that is, repeated several times at the opening of get-off. The arpeggiated chord is played on an unaccompanied electric guitar. The rapidity of the notes and the way they are played produce a high-pitched whirling sound that captures the listener's attention and creates anticipation of what is to follow. Close quote. Well, maybe, but the fact remains that the excerpt taken is very short to repeat, two seconds long, much shorter than the excerpt in the Cucumberrow case, and much less distinctive. Indeed, if I hadn't pointed it out, almost surely you would not have noticed it. This would thus seem to be an easy case of de minimis copying, taking an amount too small to warrant the attention of the law, and the district court so ruled. Surprisingly, the Sixth Circuit Court of Appeals reversed, holding broadly, that the de minimis defense is not available to a defendant who samples sound recordings. Taking even a tiny piece of someone's sound recording will trigger a violation of Section 106.1. The rule, announced in Bridgeport Music, was widely denounced, and it may well be a bad rule, but it's somewhat less shocking than it first appears. The reason is that in U.S. law, sound recordings get narrower protection than other sorts of copyrights. Here's the relevant provision. As you can see, imitating a sound recording does not constitute infringement. With respect to all other types of copyrighted works, including musical compositions, imitation does constitute infringement, but not for sound recordings. Only capturing the actual sounds fixed in the recording gives rise to liability. Dubbing is actionable, but not imitation. This provision explains why, as we discussed in Lecture No. 3, musicians are free to make covers of commercially released songs. As we saw there, musicians have to pay the composers of those songs, or their assignees, a modest compulsory fee, but they don't have to get the composer's permission. Even more strikingly, they don't have to pay anything to the owners of the copyrights in the sound recordings and embodying the original rendition of the song, because they're not capturing its actual sounds. Now, the reason why this rule is relevant to digital sampling is that, as the Sixth Circuit observes, it creates an inexpensive alternative to sampling. As the court explained, quote, it must be remembered that if an artist wants to incorporate a riff from another work in his or her recording, he's free to duplicate the sound of that riff in the studio. The market will control the license price and keep it within bounds. The sound recording copyright owner cannot exact a license fee greater than what it would cost the person seeking the license to just duplicate the sample in the course of making the new recording. Close quote. In other words, the court argues, a blanket rule that all digital sampling, no matter how small, gives rise to liability would not quote stifle creativity, close quote, because rap artists and others can either imitate the sounds they want or can obtain a license, the fee for which will be low, because the licensor is aware that imitation is lawful. The court's last argument in support of its ruling is different. Not only would a blanket prohibition on sampling not stifle creativity, said the court, it's fair. Why? Because quote, sampling is never accidental. It's not like the case of a composer who has a melody in his head, perhaps not even realizing that the reason he hears the melody is that it's the work of another, which he had heard before. When you sample a sound recording, you know you are taking another's work product. Close quote. Notice here the echoes of the fairness theory of copyright in general. Intentionally taking someone else's quote work product, the court suggests, is wrong, morally wrong, and deserves to be punished. I leave to you the question of whether these arguments in combination are convincing. I hasten to add that the elimination of the de minimis defense for sound recordings does not necessarily mean that all digital sampling is unlawful. It's still in principle possible for samplers to avoid liability on the basis of fair use. We'll take up that option in two weeks. Incidentally, the Fair Use Doctrine, as you'll see, also provides the best line of defense for me if the funkadelics or NWA or the chiffons object to my use in this lecture of pieces of their recordings. The seemingly absolute position taken by the Sixth Circuit Court of Appeals in the Bridgeport case is not limited to the United States. In the 2008 Matal Auf Matal case, the Federal Supreme Court of Germany similarly ruled that the sampling of any part of a digital recording gives rise to infringement of the neighboring right in the recording. If you follow this link on your screen, you can read an English translation of the German decision. So the Bridgeport Music Doctrine, for better or worse, seems to be spreading. The third path to a demonstration of improper appropriation is the most interesting and complicated. It involves situations in which the defendant has not copied any portion of the plaintiff's work verbatim, but has instead created a different but similar work. Such cases of, to use numerous phrase, comprehensive non-literal similarity, are governed in the United States by the so-called, quote, substantial similarity, close quote, test. As you might expect, the defendant is liable under this approach if and only if his work is substantially similar to the plaintiff's. Before considering what that phrase might mean, we need to confront two threshold questions. First, are additions relevant? In other words, can a defendant escape liability by showing that he added sufficient things to the plaintiff's work to make the two works different? The canonical answer is no. As Judge Hand once put it, quote, no plagiarist can excuse the wrong by showing how much of his work he did not pirate, close quote. However, as we'll see, courts don't always adhere strictly to that guideline. Second, what happens when the plaintiff's work contains some material protected by copyright law and some material that's not protected, for example, that consists of facts or sends off air? Cases of this sort present courts with an analytical challenge. Similarity that arises because the defendant has mimicked unprotected material should, of course, not give rise to liability. On the other hand, if the courts or juries ignore the unprotected material altogether, they may fail to give the plaintiff appropriate protection for the original way in which he has arranged the unprotected pieces. No single way of handling this problem governs all cases. Instead, one can find in the case law a spectrum of approaches. At one extreme is what is sometimes called the totality approach. As that label suggests, it urges the trier of fact, whether a judge or a jury, to compare the two works in their totality and decide if they are viewed in this way substantially similar. At the opposite extreme is the so-called filtration approach. The essence of this strategy is that the trier of fact is instructed to remove all unprotected material from the plaintiff's work or comparing it to the defendant's. The context in which the filtration approach is most consistently applied is computer software. I'll spend a bit of time exploring how filtration works in that setting. You'll recall from lecture three that some, not all, software firms seek to control four sorts of behavior. Reproduction by consumers of the object code embodying their works. Reproduction by commercial enterprises of the object code of their works. Incorporation of parts of their source code into new programs. And finally, non-literal copying, namely the production and distribution of programs that have the same structure, sequence, or organization, even though they have none of the original source or object code. Two of these activities, I hope you now see, are forms of comprehensive literal similarity. The third is an example of fragmented literal similarity. And the fourth is an example of comprehensive non-literal similarity. It's in this fourth and last setting that the question of filtration arises. When does a program that resembles another program in structure and function but contains none of the same source code give rise to infringement? The courts that first addressed this issue analyzed it using essentially the same tools that they employed when analyzing non-literal similarity between two films or photographs. As a result, the plaintiffs with some frequency won. Then in 1992, the influential Second Circuit Court of Appeals adopted a sharply new approach. In the Altai case, the Second Circuit adopted what has come to be called the Abstraction Filtration Comparison Test for cases involving non-literal similarity of software. As its name suggests, this test has three parts. Part one, Abstraction, is adapted from the test formulated by Judge Lernan Hand long ago for determining when the plot of one literary work is too close to the plot of another. Here's how Judge Hand described that test in the 1930 Nichols case. Quote, upon any work, and especially upon a play, a great number of patterns of increasing generality will fit equally well as more and more of the incident is left out. The last may perhaps be no more than the most general statement of what the play is about and at times might consist only of its title. But there is a point in this series of abstractions where they are no longer protected since otherwise the playwright could prevent the use of his ideas to which, apart from their expression, his property is never extended. Nobody has been able to fix that boundary and nobody ever can. In some cases, the question has been treated as though a analogous to lifting a portion out of the copyrighted work. But the analogy is not a good one because, though the skeleton is a part of the body, it pervades and supports the whole. In such cases, we are rather concerned with the line between expression and what is expressed. Graphically, this approach might be represented as follows. At the lowest level of abstraction, in other words, the level with the most detail, the plot of a play might be depicted like this. Long lines represent major events, shorter lines minor events. Abstracted one level, in other words, removing some of the minor incidents, the plot would look like this. Abstracted another level, it would look like this and so forth. At the highest level of abstraction, the plot is extremely general, something like boy meets girl, they quarrel and separate, they make up and live happily ever after. Somewhere in this sequence of nested patterns lies a line hard to discern but nevertheless present. Formulations of the plot sufficiently abstract to fall above the line constitute, within the language of copyright law, mere ideas. Formulations sufficiently detailed to fall below the line consist of protected expression. Only if the defendant's plot parallels the plaintiff's at a level of detail below the line has the defendant infringed the plaintiff's work. In the Altide case, the Second Circuit adapted the so-called pattern test, originally developed to deal with plots, to software. A software program, the court held, can be arranged into levels of abstraction just like a plot. The highest level consists of a general description of the function of the program. At the next level down are major modules or subroutines. At the next level are finer divisions among parts of modules, until one reaches at the lowest level the source code. Once the various levels of the program have been separated in this fashion, the Altide test requires the trier effect to remove, in other words, to filter out features that do not enjoy copyright protection. With respect to software, the court ruled there are many reasons for disqualifying features. First, the court should identify and then filter out elements of the plaintiff's program that are necessary for the program to operate efficiently. For example, anything necessary to enable the program to sort and search efficiently should be removed. Next to go are features that are dictated by external factors. These include features necessary to meet mechanical specifications of the computer, the compatibility requirements of other programs, the design standards of the computer manufacturers, demands of the industry being served, in other words, the demands of customers, and widely accepted programming practices. Finally, and least surprisingly, it should identify and eliminate features that the original programmer had taken from the public domain. The net result is an abridged pattern of features, like a comb missing many teeth. The third and last step in the Altide test is comparison. The features of the defendants program are laid alongside the now filtered version of the plaintiff's program, and the two are compared. If and only if they appear substantially similar is the defendant liable. Since 1992, most courts confronted with cases of this general sort have employed the Altide test. It's not quite universal, and the Supreme Court has never lent Altide its imprimatur, but it is surely dominant. As you might surmise, the Altide approach favors defendants. Its adoption made it significantly more difficult for software developers to recover against competitors who deployed structurally similar programs. The Altide approach has many critics. Executives in major software firms sometimes point to this decision as one of the reasons why they have turned increasingly to patent law to protect their innovations. And some scholars, like Dennis Karjala, point out that the relationship between the idea-expression distinction that underlies the abstraction portion of the test and the filtration portion of the test is far from clear. But for better or worse, the test seems here to stay. Now, let's return to the main road. The fundamental problem I've noted is how to separate protected from unprotected material when applying the substantial similarity test. The approach most generous to copyright owners is totality analysis. The least generous approach is the abstraction filtration comparison test, now applied by most U.S. courts in the context of software. In between these two is the, quote, more discerning observer test. The way in which that approach works is best explained after we have explored the meaning of substantial similarity. So to that troublesome concept we now turn. The courts have had great difficulty defining exactly how much similarity between a defendant's work and a plaintiff's work is necessary to trigger liability. They agreed that it's not a mathematical question. The percentage of overlap is not decisive. The standard is much more impressionistic than that. But the courts have trouble capturing in words the level required. Various formulations can be found in the case law. Listed on the map, in front of you, are the most important. You should not think of these formulations as distinct tests, analogous to the avowedly distinct tests that are used to determine conceptual separability with respect to useful articles. Rather, they at least purport to be different ways of describing the same standard. Often courts say that two works are substantially similar if they have, quote, the same aesthetic appeal, close quote. For example, in the 2001 Boisant case, the Second Circuit said, quote, generally an allegedly infringing work is considered substantially similar to a copyrighted work if the ordinary observer, unless he's set out to detect the disparities, would be disposed to overlook them and regard their aesthetic appeal as the same. You may recall that in the 2005 Mannion case, the district court relied on this test to rule that a jury could find substantial similarity between this magazine photo of Kevin Garnett and this photo prepared for a billboard advertising Coors Beer. Here they are, side by side. Another slightly different formulation. In the Ideal Toy case, the same court defined the relevant test as, quote, whether an average lay observer would recognize the alleged copy as having been appropriated from the copyrighted work, close quote. Applying this language, a lower court in the 1987 Steinberg case concluded that this famous magazine cover, suggesting the myopia of residence of New York City, was infringed by this poster, publicizing the movie Moscow on the Hudson. Again, here they are, side by side. Yet another variation. Sometimes courts contend that attention should be focused on, quote, the overall concept and feel of the two works. For example, in the Croft case, the Ninth Circuit relied on that phrase in ruling that the characters' flora fauna and layout of an imaginary land depicted in an advertisement for McDonald's restaurants resembled sufficiently the corresponding features of the imaginary land depicted in the Fun Stuff Children's Show, sufficiently closely to give rise to liability. Individual features the court noted were very similar. For example, quote, both lands are governed by mayors who have disproportionately large round heads dominated by long, wide mouths, close quote. But the real basis of liability was the overall judgment that McDonald's land, quote, captured the total concept and feel of the Fun Stuff Show, close quote. In a similar vein, a district court in the Kish case concluded that a rational jury might find substantial similarity in the, quote, underlying tone or mood, close quote, of the two photos depicted on your screen. From what perspective should these judgments be made? In other words, through whose eyes should the trier effect view the two works when determining whether they are substantially similar? The dominant answer to that question is the average lay observer. In other words, not an expert in the field of art or literature at issue, but an ordinary lay person. That answer is subject to two qualifications, however. First, on occasion, courts will adopt the perspective of the members of the intended audience for the two works. For example, in the Croft case, which I just mentioned, the Ninth Circuit at one point sought to assess the degree to which the Fun Stuff Show and McDonald's land would have looked similar to the children at which both works were primarily aimed. For those of you familiar with trademark law, this maneuver should remind you of the ways in which courts sometimes assess the likelihood of consumer confusion that arises when two trademarks are similar. The second qualification is that courts will occasionally adjust their glasses when the plaintiff's work contains some unprotected material, as well as some protected material. As you recall, I described the two polar positions taken by courts when dealing with such cases, the so-called totality analysis, and the filtration approach exemplified by the Altai case. In between those two positions is the so-called more discerning observer test. The Second Circuit, in the Boisson case, explained this approach as follows. In the Folio Impressions case, one of the plaintiff's fabric was not original and therefore not protectable. We articulated the need for an ordinary observer to be, quote, more discerning in such circumstances. The ordinary observer would compare the finished product that the fabric's designs were intended to grace women's dresses and would be inclined to view the entire dress, consisting of protectable and unprotectable elements as one whole. Here, since only some of the design enjoys copyright protection, the observer's inspection must be more discerning. Shortly after Folio Impressions was decided, we reiterated that a, quote, more refined analysis, quote, is required where a plaintiff's work is not wholly original, but rather incorporates elements from the public domain. In these instances, quote, what must be shown is substantial similarity between those elements and only those elements that provide copyrightability to the allegedly infringed compilation. In contrast where the plaintiff's work contains no material imported from the public domain, the more discerning test is unnecessary. In applying this test, a court is not to dissect the works that issue into separate components and compare only the copyrightable elements. To do so would be to take the more discerning test to an extreme, which would result in almost nothing being copyrightable because original works broken down into their composite parts would usually be little more than basic unprotectable elements, like letters, colors, and symbols. This outcome, affording no copyright protection to an original compilation of unprotectable elements, would be contrary to the Supreme Court's holding in FICE publications. Okay, that's a lot of language, and after all that, it's still not entirely clear what the court has in mind. The bottom line seems to be, when comparing the two works, don't filter out the unprotected material, as the court did with software and Altai, but instead adopt a more skeptical stance than the hypothetical ordinary observer. As we've seen, the emphasis in all of these cases is on some variant of an ordinary lay observer. Does that mean that expert witnesses have no role in substantial similarity determinations? Not quite. Although the ultimate judgment must always be made from the standpoint of an ordinary observer, experts can help in framing the ultimate judgment. The Court of Appeals for the Ninth Circuit has developed the most elaborate, some scholars say convoluted, system for combining expert and lay analysis. In the Ninth Circuit, there are now two steps to a determination of substantial similarity, an extrinsic test and an intrinsic test. With respect to the first, expert opinion is relevant, but not with respect to the second. To win, the plaintiff has to satisfy both. So what are these two terms, extrinsic and intrinsic, mean? It's not entirely clear, in part because the Ninth Circuit's usage has changed over time. In the Croft case, where this approach was first deployed, the extrinsic test was defined as, quote, similarity in ideas. Intrinsic was defined as, quote, similarity in expression. This formulation, you probably noticed, is hard to reconcile with the fundamental principle that copyright law doesn't protect ideas at all. Perhaps sensing that problem, in subsequent cases, the Ninth Circuit has formulated the approach differently. For example, in the Cavalier case, extrinsic is treated as, quote, an objective comparison of specific expressive elements, close quote, while intrinsic is, quote, a subjective comparison of the total concept and feel, close quote, of the two works. In the Swarovski case, the court backed up a bit, defining the issue addressed by the extrinsic test as, quote, whether the two works share a similarity of ideas and expression as measured by external objective criteria, while the intrinsic test, again, focused on total concept and feel. Even the Ninth Circuit acknowledges the clumsiness of these maneuvers. In Swarovski, for example, the judge has admitted, quote, the extrinsic test provides an awkward framework to apply to copyrighted works like music or art objects, which lack distinct elements of idea and expression. Nevertheless, the test is our law, and we must apply it, close quote. So, too, must lid against when they appear in the courts of the Ninth Circuit. This concludes our review of the doctrine of substantial similarity. In closing, I would urge you not to put too much stock in the language the courts employ when setting forth these various formulations. At least as good a guide to the law in this area is the pattern of results reached in the cases. That's the main reason, aside from simple entertainment, that during our tour, I've been telling you the central facts of several of these cases and how they were resolved. This is not to say, of course, that the results in these cases are entirely consistent. They surely are not. But from them, you can distill an impression of the degree or kind of similarity that will likely get you in trouble and the degree or kind that will not. More guidance than that may be impossible.