by Ryan Law­son 1

There is cur­rent­ly a cir­cuit split on the issue of whether merg­er and scènes-à-faire should be eval­u­at­ed as bars to copy­righta­bil­i­ty or as defens­es to infringe­ment. The dis­tinc­tion deter­mines which par­ty bears the bur­den of proof dur­ing lit­i­ga­tion and which par­ty that bur­den favors. Because these defens­es are impor­tant in many copy­right suits and result in con­flict­ing results among the fed­er­al courts, it is impor­tant that the cir­cuit split be resolved.

In this arti­cle, I pro­pose that gene­r­i­cide, a trade­mark doc­trine, can be used to guide the analy­sis of when to apply merg­er and scènes-à-faire. By clar­i­fy­ing how a par­ty can demon­strate that a copy­right holder’s work is a prod­uct of merg­er or includes scènes-à-faire, par­ties will be bet­ter posi­tioned to use these doc­trines dur­ing lit­i­ga­tion, and bet­ter able to adjust their behav­ior in accor­dance with expect­ed outcomes.


The Copy­right Act pro­tects expres­sion and not ideas, a prin­ci­ple known as the idea/expression dichoto­my. 2 Merg­er is a com­mon law doc­trine in copy­right that “func­tions as an excep­tion to the idea/expression dichoto­my.” 3 Merg­er doc­trine rec­og­nizes that there are cer­tain sce­nar­ios in which there are not enough ways of express­ing an idea. The Sec­ond Cir­cuit has explained that “expres­sion is not pro­tect­ed in those instances where there is only one or so few ways of express­ing an idea that pro­tec­tion of the expres­sion would effec­tive­ly accord pro­tec­tion to the idea itself.” 4

Although every Court of Appeals rec­og­nizes merg­er as a com­mon-law copy­right doc­trine, they are split over whether it is a defense to infringe­ment or a bar to copy­righta­bil­i­ty. The Sec­ond and Fed­er­al Cir­cuits apply merg­er as a defense to copy­right infringe­ment. 5 By con­trast, the Fifth and Sixth Cir­cuits have deter­mined merg­er stands as a bar to copy­right pro­tec­tion in the first instance. 6 The dis­agree­ment between the Courts of Appeals can be large­ly sum­ma­rized as a dif­fer­ence between which par­ty should bear the bur­den of proving/disproving merg­er and at what stage of the lig­a­tion it should be applied. 7

Scènes-à-faire are tropes and oth­er expres­sive devices that are per­va­sive or essen­tial to cer­tain cat­e­gories of expres­sion. 8 For exam­ple, “the war between rival clans, devel­op­ing love between chil­dren of the clans, [and] death and rec­on­cil­i­a­tion after the rejec­tion of revenge,” from Shakespeare’s Romeo and Juli­et are con­sid­ered scènes-à-faire, because the play has served as the inspi­ra­tion for many oth­er works and these ele­ments have been used in oth­er works count­less times. Because these ele­ments have become ubiq­ui­tous in lit­er­a­ture, they are in the pub­lic domain and can­not be pro­tect­ed by copy­right; they are scènes-à-faire. 9

Although every Court of Appeals rec­og­nizes scènes-à-faire, they are—as in the merg­er debate—split over whether it is a defense to infringe­ment or a bar to copy­righta­bil­i­ty. The Sev­enth Cir­cuit recent­ly held that “[t]he dif­fer­ence between the doc­trine and copy­right inva­lid­i­ty is vital to main­tain.” 10 By con­trast, the Sixth and Tenth Cir­cuits view scènes-à-faire as a bar to copy­right pro­tec­tion. 11

The pro­posed meth­ods of deal­ing with merg­er and scènes-à-faire dif­fer as to when courts should con­sid­er the doc­trines. If they are bars to copy­right pro­tec­tion, then the plaintiff’s choic­es at the time of orig­i­nal cre­ation should gov­ern whether merg­er or scènes-à-faire apply. For exam­ple, in Ora­cle v. Google, 12 merg­er as a bar to copy­right pro­tec­tion would depend on Oracle’s mind­set regard­ing the inclu­sion of the Java pro­gram­ming mod­ules. A court would need to deter­mine what moti­vat­ed the cre­ation of the mod­ules. 13 If, on the oth­er hand, merg­er and scènes-à-faire are affir­ma­tive defens­es to infringe­ment, then it is the defendant’s choic­es at the time of their sub­se­quent cre­ation that gov­ern whether merg­er or scènes-à-faire should apply. Return­ing to Ora­cle, it would depend on why Google chose to use the pro­gram­ming mod­ules in the Android oper­at­ing sys­tem. The options open to Google, such as how stan­dard the pro­gram­ming mod­ules had become, would be con­trol­ling. Expert tes­ti­mo­ny would play an impor­tant role in deter­min­ing the out­come of this analysis.

I pro­pose that copy­right doc­trine should look to trade­mark law’s gene­r­i­cide doc­trine for the answer on how to apply the merg­er and scènes-à-faire. One goal of trade­mark law is to sup­port a sys­tem of easy iden­ti­fi­ca­tion of prod­ucts so that con­sumers can find their pre­ferred prod­uct with rel­a­tive ease. 14 How­ev­er, one of the major prob­lems with using trade­marks as sources of infor­ma­tion cap­i­tal 15 is that courts must pro­tect against cer­tain trade­marks that would cre­ate unfair com­pe­ti­tion. One way courts do this is by pro­hibit­ing the reg­is­tra­tion of gener­ic marks. An exam­ple of a gener­ic mark is some­thing like the phrase “apple” which “is a gener­ic name for the edi­ble fruit of the apple tree.” 16 Accord­ing­ly, no enti­ty could trade­mark the phrase “apple” with respect to the fruit. 17 One sim­ple test that courts have devised for deter­min­ing if a mark is gener­ic is the “Who are you—What are you?” test. 18 Gen­er­al­ly, if a mark answers the ques­tion “What are you?” then it is like­ly gener­ic, because it is describ­ing the prod­uct itself. 19

A relat­ed prob­lem is “gene­r­i­cide,” which occurs when a valid trade­mark becomes so ubiq­ui­tous in col­lo­qui­al usage and its own­er fails to police the mark ade­quate­ly, with the result that the mark itself achieves gener­ic sta­tus. 20 Such a sce­nario is prob­lem­at­ic for the hold­er of the gener­ic trade­mark because they may be unable to ensure that con­sumers are using their mark as a source indi­ca­tor rather than in its gener­ic sense. 21 In the case of “ther­mos,” the trade­mark hold­er, Ther­mos LLC, lost their trade­mark because con­sumers had begun to equate the word “ther­mos” with any vac­u­um-sealed liq­uid con­tain­er. 22 The the­o­ret­i­cal under­pin­ning of the pro­hi­bi­tion of gener­ic trade­marks is fair com­pe­ti­tion: once gene­r­i­cide occurs, a trade­mark own­er is pre­clud­ed from con­tin­u­ing to uti­lize their mark, incen­tiviz­ing them to enforce their trade­mark rights pri­or to the occur­rence of gene­r­i­cide. 23

Copy­right law should adopt a fair com­pe­ti­tion analy­sis and apply a test sim­i­lar to trademark’s gene­r­i­cide. 24 Courts should begin by exam­in­ing whether ideas and expres­sions had merged or scènes-à-faire exist­ed in the rel­e­vant indus­try at the time of the sec­ondary copy­right owner’s cre­ation. Instead of con­sid­er­ing the choic­es that Google made, the Ora­cle court should con­sid­er mar­ket con­di­tions at the time of Google’s incor­po­ra­tion of the mod­ules into its own work. If at the time of Google’s cre­ation of the Android oper­at­ing sys­tem the Java pro­gram­ming mod­ules had become an indus­try stan­dard, then Google would not be liable for incor­po­rat­ing them. If, on the oth­er hand, the mod­ules were not an accept­ed indus­try stan­dard, then Google’s use of them would be infring­ing. By adopt­ing this approach, copy­right law would give bet­ter def­er­ence to the way that indus­tries have evolved to com­pete. 25

A sim­i­lar approach has already been applied in a sem­i­nal copy­right case: Lotus v. Bor­land. 26 In Lotus, the ques­tion was whether the menu com­mand hier­ar­chies in Lotus’ spread­sheet pro­gram were copy­rightable. In reach­ing its deci­sion that the menu com­mand hier­ar­chies were not copy­rightable, the First Cir­cuit rea­soned: “Lotus wrote its menu com­mand hier­ar­chy so that peo­ple could learn it and use it. Accord­ing­ly, it falls square­ly with­in the pro­hi­bi­tion on copy­right pro­tec­tion estab­lished in Bak­er v. Selden and cod­i­fied by Con­gress in § 102(b).” 27

Sim­i­lar­ly, Sun—later pur­chased by Oracle—created Java pro­gram­ming mod­ules that could be used by pro­gram­mers eas­i­ly and adopt­ed wide­ly. Although an ana­logue to trademark’s gene­r­i­cide in copy­right would con­sid­er mar­ket con­di­tions at the time of the sec­ondary artist or defendant’s creation—and not the ini­tial author’s intent—the Lotus case still focus­es on a sim­i­lar inquiry: whether the dis­put­ed mech­a­nism has become so ubiq­ui­tous in the indus­try as to become a stan­dard unwor­thy of copy­right pro­tec­tion. Specif­i­cal­ly in Lotus, the court dis­cerned that the pri­ma­ry author’s intent was for the mar­ket to devel­op around its orig­i­nal cre­ation and to incor­po­rate it as a com­mon ele­ment, and that the pri­ma­ry author was suc­cess­ful to that end.

Rely­ing on author’s intent is some­what con­tro­ver­sial in many legal fields, 28 but there are rea­sons for focus­ing on stan­dard­iza­tion with­in markets—to which an author’s intent or fail­ure to take pro­tec­tive mea­sures may point. Many tech­ni­cal mar­kets rely on stan­dard ele­ments, and com­put­er pro­grams, which can be pro­tect­ed by copy­right pro­tec­tion, are no excep­tion. Google’s Android oper­at­ing sys­tem incor­po­rat­ed a stan­dard tech­no­log­i­cal ele­ment by using the Java sys­tem, which “rapid­ly grew in pop­u­lar­i­ty and now comprise[s] a large share of the Unit­ed States mar­ket.” 29 The rea­son that Oracle’s pro­gram­ming mod­ules should lose copy­right pro­tec­tion is because they had become wide­ly used and their ini­tial cre­ators had not tak­en steps to pro­tect them from infringe­ment by oth­ers. 30


There are sev­er­al prob­lems with apply­ing the gene­r­i­cide doc­trine in the copy­right con­text. The first is a mechan­i­cal issue: the doc­trine does not map onto copy­right direct­ly. In the trade­mark con­text, it is con­sumers who make a trade­mark gener­ic by using it in a giv­en mar­ket as a gener­ic mark. One famous exam­ple is “ther­mos,” which became a gener­ic mark for any vac­u­um insu­lat­ed liq­uid con­tain­er. 31 In copy­right, it is not as clear how expres­sive con­tent would trans­form into scènes-à-faire or become the object of merg­er. Because authors con­trol the con­tent in copy­right­ed mate­r­i­al, it is ulti­mate­ly sec­ondary artists who are respon­si­ble for turn­ing some­thing into scènes-à-faire or the object of merg­er. Like gene­r­i­cide, iden­ti­fy­ing the exact point when this occurs can be chal­leng­ing. The pro­gram­ming mod­ules in Ora­cle may be described as merged ele­ments in the indus­try, but scènes-à-faire may present a trick­i­er analy­sis. If sec­ondary authors are respon­si­ble for deter­min­ing when ele­ments have become scènes-à-faire, many of these authors will be incen­tivized to pro­mote works that arguably infringe on pri­or works but do not oth­er­wise result in law­suits, in order to ensure that sub­se­quent authors have access to copy­ing cer­tain expres­sive ele­ments through the scènes-à-faire doc­trine. On the oth­er hand, if mar­ket con­di­tions con­trol the analy­sis, then it seems unlike­ly that any­thing would ever become scènes-à-faire, because law­suits would arise to pre­vent that from happening.

Notably, the orig­i­nal copy­right own­er does not lose all pro­tec­tion once a court has found that merg­er or scènes-à-faire applies. Instead, it is only scènes-à-faire or por­tions sub­ject to merg­er that would be unprotected—for exam­ple, the Java pro­gram­ming mod­ules in Ora­cle. 32 Although this would rep­re­sent a reduc­tion in the val­ue of some copy­rights, a ben­e­fit might be that it would help elim­i­nate com­plaints about the term of copy­rights, which many view as being too long. 33 Addi­tion­al­ly, if a copy­right­ed work has become the object of merg­er or scènes-à-faire, this may indi­cate that the copy­right has enjoyed sig­nif­i­cant pop­u­lar­i­ty and the author has oth­er­wise recov­ered fair com­pen­sa­tion for their copy­right monop­oly in spite of the sub­se­quent reduc­tion in value.

To some extent, the copy­right doc­trines of merg­er and scènes-à-faire already incor­po­rate ideas under­pin­ning gene­r­i­cide, because they do not con­tain cer­tain groups of expres­sions that are fixed in time, but instead evolve as some forms of expres­sion are repeat­ed and move into our com­mon lex­i­con. How­ev­er, the Courts of Appeals have diverged in their treat­ment of scènes-à-faire and merg­er, and it is like­ly that at some point the Supreme Court will resolve the issue. One way to ensure that these doc­trines remain flex­i­ble but pow­er­ful would be to incor­po­rate a form of gene­r­i­cide analy­sis with­in them. As a result, courts would place less empha­sis on whether copy­right pro­tec­tion pre­ced­ed a sub­se­quent work, and more empha­sis on the nature of the rel­e­vant mar­ket at the time of the sec­ondary artist’s cre­ation. This may serve as an impor­tant lim­i­ta­tion on the scope of copy­right pro­tec­tions, giv­en that the Supreme Court seems hes­i­tant to reign in expan­sions in copy­right terms. 34



  1. This Con­tri­bu­tion is based on my expe­ri­ence writ­ing a prob­lem for the 40th Edi­tion of NYU Moot Court Board’s Case­book, pub­lished in 2016. The Con­tri­bu­tion is a com­men­tary on the copy­right doc­trines of merg­er and scènes-à-faire. I pro­pose to sim­pli­fy both doc­trines through the impor­ta­tion of gene­r­i­cide doc­trine from trade­mark law.
  2. 17 U.S.C. § 102.
  3. Ora­cle Am., Inc. v. Google Inc., 750 F.3d 1339, 1359 (Fed. Cir. 2014).
  4. Kre­gos v. Asso­ci­at­ed Press, 937 F.2d 700, 705 (2d Cir. 1991).
  5. See Ora­cle, 750 F.3d at 1358 (“[C]oncepts of merg­er and [scènes-à-faire] are affir­ma­tive defens­es to claims of infringe­ment.”); Kre­gos, 937 F.2d at 705 (“[T]his so-called ‘merg­er’ doc­trine [is applied] in deter­min­ing whether action­able infringe­ment has occurred, rather than whether a copy­right is valid.”).
  6. See Lex­mark Int’l, Inc. v. Sta­t­ic Con­trol Com­po­nents, Inc., 387 F.3d 522, 537–39 (6th Cir. 2004); Mason v. Mont­gomery Data, Inc., 967 F.2d 135, 138–39 (5th Cir. 1992).
  7. See gen­er­al­ly Andrew B. Hebl, A Heavy Bur­den: Prop­er Appli­ca­tion of Copyright’s Merg­er and Scenes A Faire Doc­trines, 8 Wake For­est Intell. Prop. L.J. 128, 148–50 (2008) (explain­ing that the bat­tle between the Courts of Appeals is over whether the plain­tiff or the defen­dant should bear the bur­den of apply­ing the merg­er doc­trine). See also San­dro Oca­sio, Prun­ing Para­copy­right Pro­tec­tions: Why Courts Should Apply the Merg­er and Scènes à Faire Doc­trines at the Copy­righta­bil­i­ty Stage of the Copy­right Infringe­ment Analy­sis, 3 Seton Hall Cir­cuit Rev. 303, 331–37 (2006).
  8. See 2 Patry on Copy­right § 4:24 (quot­ing Judge Yankwich from the Cen­tral Dis­trict of Cal­i­for­nia: “The French use a very expres­sive phrase in dra­mat­ic lit­er­a­ture: “scenes a faire” that is, scenes which ‘must’ be done.”).
  9. Reed-Union Corp. v. Tur­tle Wax, Inc., 77 F.3d 909, 914 (7th Cir. 1996).
  10. Id.
  11. See Lex­mark Int’l, Inc. v. Sta­t­ic Con­trol Com­po­nents, Inc., 387 F.3d 522, 535 (6th Cir. 2004) (explain­ing scènes-à-faire does not play any role in the infringe­ment analy­sis but is instead a bar to copy­right pro­tec­tion); Gates Rub­ber Co. v. Ban­do Chem. Indus., Ltd., 9 F.3d 823, 838 (10th Cir. 1993).
  12. See Ora­cle, 750 F.3d 1339 (Google want­ed to incor­po­rate the well-known and fre­quent­ly used Java pro­gram­ming modules—which Ora­cle owned—into its Android oper­at­ing sys­tem).
  13. See Ora­cle, 750 F.3d at 1361 (“The focus is, there­fore, on the options that were avail­able to Sun/Oracle at the time it cre­at­ed the API pack­ages.”)
  14. William M. Lan­des & Richard A. Pos­ner, Trade­mark Law: An Eco­nom­ic Per­spec­tive, 30 J.L. & Econ. 265, 268–69 (1987).
  15. Id.
  16. 2 McCarthy on Trade­marks and Unfair Com­pe­ti­tion § 12:1 (4th ed.).
  17. Of course, the phrase “Apple” can serve as a trade­mark in oth­er com­mer­cial realms, as it does in the per­son­al com­put­ing mar­ket. 2 McCarthy on Trade­marks and Unfair Com­pe­ti­tion § 12:1 (4th ed.).
  18. Id.
  19. See U. S. Jaycees v. San Fran­cis­co Jr. Cham­ber of Com­merce, 513 F.2d 1226, 1228–29 (9th Cir. 1975) (explain­ing the appli­ca­tion of the “Who are you—What are you?” test with regards to the Junior Cham­ber of Com­merce, which was found to be too gener­ic to reg­is­ter as a trade­mark).
  20. See 2 McCarthy on Trade­marks and Unfair Com­pe­ti­tion § 12:1 (4th ed.) (explain­ing the process of gene­r­i­cide using exam­ples such as ther­mos and aspirin).
  21. Id.
  22. Id.
  23. See Wayne F. Oso­ba, The Leg­isla­tive Response to Anti-Monop­oly: A Missed Oppor­tu­ni­ty to Clar­i­fy the Gener­ic­ness Doc­trine, 1985 U. Ill. L. Rev. 197, 198 (1985) (explain­ing that the pro­hi­bi­tion against gener­ic trade­marks rep­re­sents one bound­ary beyond which trade­marks would become too anti-com­pet­i­tive to be grant­ed); Rebec­ca Tush­net, Why the Cus­tomer Isn’t Always Right: Pro­duc­er-Based Lim­its on Rights Accre­tion in Trade­mark, 116 Yale L.J. Pock­et Part 352, 353 (2007) (explain­ing that the goal of pro­hibit­ing gener­ic marks is to improve com­pe­ti­tion). See gen­er­al­ly Aber­crom­bie & Fitch Co. v. Hunt­ing World, Inc., 537 F.2d 4, 9 (2d Cir. 1976) (find­ing that depriv­ing cer­tain com­peti­tors of gener­ic names would give too much com­pet­i­tive advan­tage to who­ev­er held the gener­ic mark).
  24. See gen­er­al­ly Jamie Lund, Copy­right Gene­r­i­cide, 42 Creighton L. Rev. 131 (2009) (explain­ing how to apply gene­r­i­cide to copyright’s scènes-à-faire and merg­er doc­trines).
  25. This is slight­ly dif­fer­ent than accus­ing an author of sleep­ing on his or her rights. An author whose copy­right becomes the stan­dard is not being pun­ished because they have slept on their rights, because they have no right to assert against their copy­right becom­ing scènes-à-faire in what­ev­er area their copy­right exists in. Con­tra Lund, supra note 24, at 136 (argu­ing that a copy­right owner’s pro­tect­ed expres­sion becom­ing scènes-à-faire is the result of sit­ting on their copy­right and not assert­ing it).
  26. 49 F.3d 807, 817–18 (1st Cir. 1995), aff’d, 516 U.S. 233 (1996) (hold­ing that the copy­right at issue was not allowed as a “method of oper­a­tion”).
  27. Id. at 817.
  28. See gen­er­al­ly Amy M. Adler, Against Moral Rights, 97 Cal. L. Rev. 263, 275 (2009) (explain­ing the rela­tion­ship between artists’ intent and moral rights, and why dis­re­gard­ing the for­mer strength­ens argu­ments against the lat­ter).
  29. 750 F.3d at 1351.
  30. Lund, supra note 24, at 147–49 (sug­gest­ing sev­er­al steps Lotus could have tak­en to pro­tect its pro­gram, includ­ing using patents to pro­tect their soft­ware).
  31. Am. Ther­mos Prod. Co. v. Aladdin Indus., Inc., 207 F. Supp. 9, 14 (D. Conn. 1962), aff’d sub nom. King-See­ley Ther­mos Co. v. Aladdin Indus., Inc., 321 F.2d 577 (2d Cir. 1963), and decree mod­i­fied sub nom. King-See­ley Ther­mos Co. v. Aladdin Indus., Inc., 320 F. Supp. 1156 (D. Conn. 1970).
  32. See Ora­cle, 750 F.3d at 1349.
  33. 17 U.S.C. §§ 302–305.
  34. See Eldred v. Ashcroft, 537 U.S. 186, 208–10 (2003) (explain­ing how Con­gress’ exten­sions of the copy­right term do not con­sti­tute per­pet­u­al copy­right).