FOR PUBLICATION

UNITED STATES COURT OF APPEALS

FOR THE NINTH CIRCUIT

VINCENT A. AMBROSETTI, No. 24-2270 D.C. No. Plaintiff - Appellant, 3:21-cv-00211- AR v.

OREGON CATHOLIC PRESS; OPINION

BERNADETTE FARRELL,

Defendants - Appellees.

Appeal from the United States District Court for the District of Oregon Karin J. Immergut, District Judge, Presiding

Argued and Submitted June 10, 2025 San Francisco, California

Filed August 27, 2025

Before: SIDNEY R. THOMAS and MILAN D. SMITH, JR., Circuit Judges, and DOUGLAS L. RAYES, District Judge. *

Opinion by Judge Milan D. Smith, Jr.

* The Honorable Douglas L. Rayes, United States District Judge for the District of Arizona, sitting by designation. SUMMARY **

Copyright / Discovery Sanctions

The panel affirmed in part and reversed in part the district court’s judgment in favor of defendants Bernadette Farrell and her publisher Oregon Catholic Press, and remanded, in Vincent Ambrosetti’s action alleging that Farrell violated the Copyright Act by copying Ambrosetti’s liturgical work, “Emmanuel,” in her own Christian music composition, “Christ Be Our Light.” Affirming in part, the panel held that the district court did not abuse its discretion by excluding certain evidence as a sanction for Ambrosetti’s late disclosure of that evidence, as well as the theory of access premised on that evidence. The panel concluded that the discovery sanctions were not claim dispositive because Ambrosetti still could, and did, proceed with other theories of access and striking similarity on his copyright claim. Accordingly, the district court was not required to explicitly consider whether the claimed discovery misconduct involved willfulness, fault, or bad faith, nor to consider the availability of lesser sanctions. The panel further concluded that the district court did not abuse its discretion when it excluded the evidence on the grounds that Ambrosetti had not shown that his failure to produce the evidence was either substantially justified or harmless under Fed. R. Civ. P. 37(c)(1). Reversing in part, the panel held that the district court erred in granting summary judgment to defendants because,

** This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.

AMBROSETTI V. OREGON CATHOLIC PRESS 3

even without the excluded evidence, there were triable issues of fact as to (1) whether there was a reasonable possibility that Farrell had access to “Emmanuel,” under either a chain of events theory or a widespread dissemination theory, before she wrote “Christ Be Our Light” and (2) whether the works were substantially similar under the extrinsic test or were strikingly similar.

COUNSEL

Donald J. Schmid (argued), Law Offices of Donald J. Schmid LLC, South Bend, Indiana; Stephanie J. Grant, Parna A. Mehrbani, and Sasha A. Petrova, Tonkon Torp LLP, Portland, Oregon; for Plaintiff-Appellant. Bert P. Krages II (argued), Portland, Oregon; Leonard D. DuBoff, The DuBoff Law Group LLC, Portland, Oregon; for Defendants-Appellees. M. SMITH, Circuit Judge:

In this copyright infringement case, Plaintiff Vincent Ambrosetti alleges that Defendant Bernadette Farrell copied his liturgical work, “Emmanuel,” in her own Christian music composition, “Christ Be Our Light.” The district court granted summary judgment to Farrell and her publisher, Defendant Oregon Catholic Press, on the basis that Ambrosetti could show neither access nor striking similarity. On appeal, Ambrosetti challenges the district court’s exclusion of certain evidence as a sanction for late disclosure as well as its grant of summary judgment to Defendants. We conclude that the district court did not abuse its discretion by excluding Ambrosetti’s evidence and the theory of access premised on that evidence. However, we hold that the district court erred by granting summary judgment to Defendants because there are triable issues of fact as to whether there is a reasonable possibility Farrell had access to “Emmanuel” before she wrote “Christ Be Our Light” and whether the works are substantially or strikingly similar. Accordingly, we affirm in part, reverse in part, and remand.

FACTUAL BACKGROUND

Plaintiff Vincent Ambrosetti is a musician and songwriter who wrote the music and lyrics to—and owns both the rights and title to the copyright of—the 1980 song “Emmanuel.” Ambrosetti is a prolific songwriter who has composed more than 400 works of sacred music. Ambrosetti is also the president and founder of the King’s Minstrels Charitable Trust, which does business as International Liturgical Publications (ILP). ILP is a non-profit publisher of sacred music for the Catholic Church.

AMBROSETTI V. OREGON CATHOLIC PRESS 5

Ambrosetti first published “Emmanuel” in 1980 as part of a songbook called “Singing Holy.” “Singing Holy” was an anthology of 102 works—98 of which were written by Ambrosetti—published by ILP and presented at a convention of the National Association of Pastoral Music (NPM). In approximately 1985, Ambrosetti did a commercial recording of “Emmanuel” with the National Philharmonic of London as part of a collection called “I Will Sing,” which was also published by ILP. “I Will Sing” was re-released in the late 1980s by the Daughters of St. Paul. Ambrosetti also recorded and released “Emmanuel” several other times between 1980 and the early 1990s, including a commercial recording with the Baltimore Symphony Orchestra in the late 1980s and another recording for a choral collection of sacred music in 1980 and again in 1985. Ambrosetti publicly performed “Emmanuel” at hundreds of events in the United States during the 1980s through early 1990s. One of those performances was at the 1985 NPM convention in Cincinnati, Ohio, and another was at the 1988 NPM regional convention in Buffalo, New York. NPM conventions attracted around 4,000 to 5,000 attendees in the 1980s. A national convention in the 1980s would have had 25 or more showcase events during which music publishers would have one hour to perform and promote their works. These showcase events would be held concurrently in six to eight sessions, with publishers performing their works in separate rooms and often in separate buildings. At the 1985 NPM convention, Ambrosetti allegedly met Owen Alstott, then the publisher and chief executive of Oregon Catholic Press (OCP). At the meeting, Ambrosetti provided Alstott with a copy of “Emmanuel.” In 1986, Ambrosetti sent Alstott a second copy of “Emmanuel.” In 1986 and 1987, Ambrosetti received letters from Alstott, acknowledging receipt of music from Ambrosetti. Also in 1986, Alstott, who lived in the United States, traveled to London to negotiate with composers who were associated with the St. Thomas More Group. During this trip, Alstott met British musician Bernadette Farrell, a founding member of that group. Alstott and Farrell began a personal relationship in 1991 and were married in 1992. Alstott and Farrell discussed music before their marriage. In 1993, Farrell composed the song “Christ Be Our Light” (Christ) for a liturgy held in a church in East London. OCP, which has been the sole publisher of Farrell’s musical works since 1986, published “Christ.” Prior to composing “Christ,” Farrell had made a handful of trips to the United States. Her first trip was in 1987, when she conducted a workshop on pastoral liturgy at the University of Notre Dame. Her second trip was in 1988, when she attended the NPM regional convention in Buffalo and sang in the OCP Choir for the St. Thomas More Centre Showcase. Farrell states that she did not attend other showcase events at that conference. She also made a handful of trips to the United States in 1989, 1990, 1991, and 1992.

PROCEDURAL BACKGROUND

On May 8, 2020, Ambrosetti sued Farrell and OCP (Defendants) in the Northern District of Indiana, seeking damages and injunctive relief for copyright infringement. 1

1 Ambrosetti had filed a prior lawsuit against Farrell and OCP in the Northern District of Indiana on August 27, 2019, but that lawsuit was dismissed on May 7, 2020, because Ambrosetti had failed to obtain a registration for “Emmanuel” before filing the lawsuit. Ambrosetti v. Or. Cath. Press, 458 F. Supp. 3d 1013, 1020 (N.D. Ind. 2020). A plaintiff must show that he or she owns the copyright in the infringed work to prevail on a copyright claim. See Williams v. Gaye, 895 F.3d 1106, 1119

AMBROSETTI V. OREGON CATHOLIC PRESS 7

The complaint alleged that Farrell was familiar with “Emmanuel” prior to 1993 and that she copied “Emmanuel” when composing “Christ.” In February 2021, the case was transferred to the District of Oregon. The cutoff date for fact and expert discovery was December 6, 2021. The cutoff date for depositions was January 20, 2022. The depositions of Ambrosetti, Farrell, Alstott, and others were taken in January 2022. The deadline for filing dispositive motions was February 22, 2022. Five days before this deadline, during a meet-and- confer call regarding Defendants’ intended motion for summary judgment, Ambrosetti disclosed for the first time the letters he received from Alstott in 1986 and 1987. Before this point, Ambrosetti had not mentioned submitting music to OCP. Ambrosetti stated that he had discovered these letters by accident while organizing old files to prepare to sell ILP’s Tennessee offices. On February 22, 2022, Defendants filed their motion for summary judgment. On March 15, 2022, Ambrosetti filed his opposition, and on March 16, Ambrosetti filed a corrected declaration stating that he had provided a copy of music including “Emmanuel” to Alstott at the 1985 conference and sent him another copy of “Emmanuel” in 1986, and that Alstott had sent him two letters acknowledging receipt of the music. In their reply brief, Defendants requested that the district court preclude Ambrosetti from asserting a theory of access based on these

(9th Cir. 2018). Ambrosetti obtained the registration in “Emmanuel” on March 4, 2020, after the filing of his prior lawsuit but before the filing of the one underlying this appeal. Ambrosetti, 458 F. Supp. 3d. at 1020. The parties do not dispute that Ambrosetti now owns the copyright to “Emmanuel.” letters, noting that they had relied on Ambrosetti’s existing theories of access—that “Emmanuel” had been widely published and performed prior to 1993—in preparing for depositions and additional discovery. Ambrosetti did not file a sur-reply, and the issue was extensively discussed at oral argument. On January 25, 2024, the magistrate judge issued his findings and recommendations. As sanctions for the untimely disclosure, he excluded the two letters from Alstott and precluded Ambrosetti from arguing that Farrell could have accessed “Emmanuel” via the copies of “Emmanuel” that Ambrosetti had provided to Alstott. The magistrate judge found Ambrosetti failed to genuinely dispute that there was no reasonable possibility that Farell could have accessed “Emmanuel” before composing “Christ.” First, Ambrosetti could not argue the Alstott had provided Farrell with copies of “Emmanuel.” Second, his other theories of access only created a “bare” possibility that Farrell could have accessed “Emmanuel” before composing “Christ” rather than a “reasonable” possibility that she could have done so. Next, the magistrate judge found that, regardless, Ambrosetti had not sufficiently demonstrated that “Christ” and “Emmanuel” were substantially similar, and that because he had not shown substantial similarity, he could not show striking similarity either. Ambrosetti timely objected to the findings and recommendations, claiming that the magistrate judge had misapplied the test for “substantial similarity” and had “incorrectly recommended exclusion of key evidence.” On March 13, 2024, the district court issued an opinion and order adopting in part and rejecting in part the magistrate judge’s findings and recommendations. The district court agreed with the magistrate judge that Alstott’s letters and the

AMBROSETTI V. OREGON CATHOLIC PRESS 9

theory of access based on those letters should be excluded and that Ambrosetti had not raised a genuine issue of material fact as to his other theory of access. The district court disagreed with the magistrate judge’s conclusions as to substantial similarity, however, concluding that when “Christ” and “Emmanuel” were evaluated as independent combinations of elements and viewed in the light most favorable to Ambrosetti, he had demonstrated substantial similarity. Nonetheless, the district court agreed that Ambrosetti had not shown a genuine issue of material fact as to striking similarity. And without either access or striking similarity, Ambrosetti’s suit could not proceed, so the district court granted summary judgment to Defendants. Ambrosetti timely appealed.

JURISDICTION AND STANDARD OF REVIEW

The district court had jurisdiction pursuant to 28 U.S.C. § 1338(a). Transgo, Inc. v. Ajac Transmission Parts Corp., 768 F.2d 1001, 1013 (9th Cir. 1985). We have appellate jurisdiction pursuant to 28 U.S.C. § 1291. Williams v. Gaye, 895 F.3d 1106, 1116 (9th Cir. 2018). “We review the district court’s rulings concerning discovery, including the imposition of discovery sanctions, for abuse of discretion. But to the extent the imposition of sanctions turns on the resolution of an issue of law, review is de novo.” Goodman v. Staples the Off. Superstore, LLC, 644 F.3d 817, 822 (9th Cir. 2011) (internal citation omitted). “We review de novo the district court’s decision to grant summary judgment.” Ets-Hokin v. Skyy Spirits, Inc., 225 F.3d 1068, 1073 (9th Cir. 2000). Under the summary judgment standard, we “must determine whether, viewing the evidence in the light most favorable to the non-moving party, there are any genuine issues of material fact and whether the district court correctly applied the substantive law.” Leisek v. Brightwood Corp., 278 F.3d 895, 898 (9th Cir. 2002).

ANALYSIS

I. The legal framework for copyright infringement

“A copyright plaintiff must prove (1) ownership of the copyright; and (2) infringement—that the defendant copied protected elements of the plaintiff’s work.” Three Boys Music Corp. v. Bolton, 212 F.3d 477, 481 (9th Cir. 2000).2 A plaintiff may prove infringement with “circumstantial, rather than direct, evidence.” Williams, 895 F.3d at 1119. “Absent direct evidence of copying, proof of infringement involves fact-based showings that the defendant had ‘access’ to the plaintiff’s work and that the two works are ‘substantially similar.’” Three Boys Music Corp., 212 F.3d at 481 (quoting Smith v. Jackson, 84 F.3d 1213, 1218 (9th Cir. 1996)). Even if there is no evidence of access, a plaintiff can still prevail if he or she can show “striking similarity” between the works. Unicolors, Inc. v. Urb. Outfitters, Inc., 853 F.3d 980, 985 (9th Cir. 2017) (quoting Baxter v. MCA, Inc., 812 F.2d 421, 423 (9th Cir. 1987)). As such, a plaintiff can proceed on a copyright claim in one of two ways in the absence of direct evidence: by showing access and substantial similarity, or by showing striking similarity alone.

2 We overturned part of Three Boys and many of the other copyright cases cited here in Skidmore v. Led Zeppelin, 952 F.3d 1051, 1066 (9th Cir. 2020) (en banc) (holding that a high degree of access to a copyrighted work does not justify a lower standard of proof for substantial similarity). For simplicity, our opinion omits this subsequent history.

AMBROSETTI V. OREGON CATHOLIC PRESS 11

II. The district court did not abuse its discretion by

imposing discovery sanctions Ambrosetti first argues that the district court abused its discretion by failing to apply the proper legal standard when imposing its “claim dispositive” discovery sanctions. Under our precedent, when a sanction results in the dismissal of a claim, a district court must explicitly consider “whether the claimed noncompliance involved willfulness, fault, or bad faith, and also to consider the availability of lesser sanctions” before imposing the sanction. R & R Sails, Inc. v. Ins. Co. of Pa., 673 F.3d 1240, 1247 (9th Cir. 2012) (internal citation omitted). Ambrosetti argues that because the exclusion of Alstott’s letters and the related theory of access resulted in the dismissal of his otherwise legally sufficient claim for copyright infringement, the district court abused its discretion when it failed to engage in the R & R Sails analysis. Ambrosetti’s argument fails for a simple reason: the sanctions were not claim dispositive. As our precedent makes clear, a discovery sanction is not claim dispositive just because the sanction made it “much more difficult, perhaps almost impossible” for a party to proceed with their claim or argument. Yeti by Molly, Ltd. v. Deckers Outdoor Corp., 259 F.3d 1101, 1106 (9th Cir. 2001). Rather, a sanction is claim dispositive only where it is “fatal” to the claim. R & R Sails, 673 F.3d at 1247. For example, in R & R Sails, the district court’s preclusion of the invoices at issue amounted “in practical terms” to a dismissal of R & R’s claim because without those invoices, it had no evidence with which to support its Brandt fees claim or its request for punitive damages. Id. Therefore, the sanction was claim dispositive, and the extra analysis as to bad faith and lesser sanctions was required. Id. Meanwhile, in Yeti, although the district court excluded the testimony of the defendant’s only damages expert, that sanction was not claim dispositive because “although onerous, [it] was less than a dismissal.” 259 F.3d at 1106. Here, while the exclusion of Alstott’s letters certainly makes Ambrosetti’s path forward more difficult, it did not “in practical terms” amount to the dismissal of Ambrosetti’s claim because Ambrosetti could—and did—proceed with other theories of access. Indeed, both before and after Ambrosetti discovered the letters, he argued that Farrell had accessed “Emmanuel” via her and Alstott’s attendance at the NPM conferences or via the song’s widespread dissemination. The exclusion of the letters was not fatal to Ambrosetti’s copyright infringement claim; it only limited his arguments about access to those he had timely developed before the discovery deadlines. And even if Ambrosetti could not proceed on any theory of access, he could still argue that “Christ” and “Emmanuel” were strikingly similar because a copyright infringement claim based on striking similarity does not require a separate showing of access. Unicolors, Inc., 853 F.3d at 985. In short, nothing about the sanctions here was a “fatal blow” to Ambrosetti’s copyright claim. R & R Sails, 673 F.3d at 1247. Moreover, although Ambrosetti has abandoned any challenge to the exclusion of these letters outside of the R & R Sails argument articulated above, we conclude that the district court did not abuse its discretion when it decided to exclude the letters on the grounds that Ambrosetti had not shown that his failure to produce the letters was either “substantially justified” or “harmless.” Fed. R. Civ. P. 37(c)(1). As the magistrate judge noted, Ambrosetti and OCP had been “ensnarled in litigation” for nearly eight years by that point, so there was good cause to be skeptical that

AMBROSETTI V. OREGON CATHOLIC PRESS 13

critical documents could still be unearthed. Additionally, Ambrosetti had not shown that introducing a new theory after the close of discovery had not harmed Defendants. See Torres v. City of Los Angeles, 548 F.3d 1197, 1213 (9th Cir. 2008) (burden is on the party facing the sanction to demonstrate the failure to comply with Rule 26(a) is substantially justified or harmless); Ollier v. Sweetwater Union High Sch. Dist., 768 F.3d 843, 862 (9th Cir. 2014) (explaining that the goal of Rule 26 “is to encourage parties to try cases on the merits, not by surprise, and not by ambush”). Accordingly, even if Ambrosetti had not abandoned any challenge to the district court’s determinations as to harm and justification, we would conclude that the district court did not abuse its discretion by excluding the letters and theory of access on those grounds.

III. The district court erred by granting summary

judgment to Defendants Nonetheless, we conclude that the district court erred by granting summary judgment to Defendants. Even without Alstott’s letters, genuine issues of material fact remain as to whether Farrell had access to “Emmanuel” before composing “Christ.” Moreover, although we agree with the district court that there is a genuine issue of material fact as to whether the works are substantially similar, we disagree with the district court’s conclusions as to striking similarity and conclude that there is a genuine dispute of material fact as to striking similarity as well. A. There is a genuine issue of material fact as to whether Farrell had access to “Emmanuel” before composing “Christ” Even excluding the two letters from Alstott, Ambrosetti has demonstrated a genuine issue of material fact as to whether Farrell otherwise could have accessed “Emmanuel” before she composed “Christ.” “To show circumstantial evidence of ‘access,’ the plaintiff may generally either provide (a) ‘evidence of a “chain of events . . . between the plaintiff’s work and defendants’ access to that work” or’ (b) ‘evidence that “the plaintiff’s work has been widely disseminated.”’” Woodland v. Hill, 136 F.4th 1199, 1207 (9th Cir. 2025) (alteration in original) (quoting Unicolors, 853 F.3d at 985). “To prove access, a plaintiff must show a reasonable possibility, not merely a bare possibility, that an alleged infringer had the chance to view the protected work.” Art Attacks Ink, LLC v. MGA Ent. Inc., 581 F.3d 1138, 1143 (9th Cir. 2009). Although Ambrosetti is not terribly clear as to this point, he appears to raise both a “chain of events” theory and a “widespread dissemination” theory. See id. at 1143–44 (determining that the plaintiff implicitly raised a chain of events argument in addition to its widespread dissemination argument). Under a “chain of events” theory, circumstantial evidence can be used to prove access by “establishing a chain of events linking the plaintiff’s work and the defendant’s access.” Id. at 1143. Meanwhile, under a “widespread dissemination” theory, a plaintiff may show access by demonstrating that the work has been broadly distributed; evidence of such widespread dissemination often “centers on the degree of a work’s commercial success and on its distribution through radio, television, and other relevant mediums.” Loomis v. Cornish, 836 F.3d 991, 997 (9th Cir. 2016). Where the “relevant market” is smaller, the issue is whether the allegedly copied work “saturat[ed]” that market. Id.

AMBROSETTI V. OREGON CATHOLIC PRESS 15

Under his “chain of events” theory, Ambrosetti argues that Farrell could have accessed “Emmanuel” either through her attendance at the 1988 NPM convention or via Alstott’s attendance at the 1985 NPM conference. Ambrosetti argues that their attendance at these events demonstrated a genuine issue of material fact as to whether he had raised a reasonable possibility of access and that the district court erred by failing to view the evidence in the light most favorable to him. We agree. In this case, there is no dispute that Alstott attended the 1985 NPM convention and that Farrell attended the 1988 NPM convention. 3 At both conventions, Ambrosetti performed “Emmanuel.” These were not large events: rather, a regional NPM convention at the time would have had about 1,000 attendees, while a main convention would have had about 4,000 to 5,000 attendees. Taken together, those facts are sufficient, viewing the evidence in the light most favorable to Ambrosetti, to demonstrate a “reasonable” possibility that Farrell could have accessed “Emmanuel” before composing “Christ” rather than a “bare” possibility that she could have done so. See 4 Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 13D.05[B][2] (2025) (a reasonable possibility of access arises if the evidence “ma[kes] plausible a realistic scenario whereby defendant heard the protected work prior to composing its own song,” such as evidence that the plaintiff and defendant both performed at the same music venue at the same time).

3 Although Ambrosetti does not mention this fact in his appellate briefing, an OCP publisher, John Limb, also attended the 1985 NPM convention in Cincinnati. Like Alstott, Limb claims he did not attend ILP’s showcase. Indeed, Farrell’s and Alstott’s admitted attendance at the NPM conventions distinguishes this case from those in which we have determined that the plaintiffs had shown only a bare possibility of access. In Art Attacks, there was no evidence that the parties ever attended the Los Angeles County Fair at the same time, which destroyed the plaintiffs’ chain of events theory. 581 F.3d at 1144. Moreover, in Briggs v. Blomkamp, a district court case upon which the magistrate judge also relied, there was no evidence that the defendant had accessed the website where the plaintiff had posted his screenplay; the plaintiff merely speculated that this website was the kind of website that the defendant would likely have perused. 70 F. Supp. 3d 1155, 1166 (N.D. Cal. 2014), aff’d sub nom. Briggs v. Sony Pictures Ent., Inc., 714 F. App’x 712 (9th Cir. 2018). In both cases, the plaintiffs could only theorize that the defendants might have been in the right place (or on the right website) at the right time. Here, the parties actually attended, and performed at, the same small events. Although Farrell and Alstott have denied that they attended Ambrosetti’s showcases at those conventions, it is up to the trier of fact to determine whether to believe their stories. See Three Boys Music Corp., 212 F.3d at 484 (a party’s credibility “should be left to the jury” (quoting Arnstein v. Porter, 154 F.2d 464, 469 (2d Cir. 1946), abrogated on other grounds as recognized in Heyman v. Com. & Indus. Ins. Co., 524 F.2d 1317, 1319 (2d Cir. 1975)). It is not for us or the district court to determine the merits of their denials. “In other words, [a] defendant cannot obtain summary judgment based on lack of access through his own unilateral denial of that access, in the face of plaintiff’s evidence for its existence.” Nimmer on Copyright § 13D.05[C][2] n.77.

AMBROSETTI V. OREGON CATHOLIC PRESS 17

Meanwhile, as to Ambrosetti’s “widespread dissemination” theory, he argues that the worldwide distribution and publication of “Emmanuel” prior to 1993 is sufficient to demonstrate a reasonable possibility that Farrell could have heard “Emmanuel” prior to composing “Christ.” The magistrate judge rejected this theory primarily because Ambrosetti had failed to establish the “amount of sales” that he had generated and so failed to show that “Emmanuel” had enjoyed sufficiently substantial commercial success before 1993. Although some of our past decisions have distinguished “bare possibility” and “reasonable possibility” in part based on numerical analysis of sale units, there is ultimately no “bright-line rule” where some number of units distributed always equals—or does not equal—access. See Nimmer on Copyright § 13D.05[C][1] (first citing Rice v. Fox Broad. Co., 330 F.3d 1170, 1178 (9th Cir. 2003) (17,000 copies sold was insufficient to show widespread dissemination); then citing Art Attacks, 581 F.3d at 1144 (2,000 T-shirts sold per year was insufficient); and then citing L.A. Printex Indus., Inc. v. Aeropostale, Inc., 676 F.3d 841, 848 (9th Cir. 2012) (local sale of 50,000 yards of fabric over four-year period created a reasonable inference that defendants had opportunity to view and copy the design)). Rather, each case must be evaluated based on its “actual facts” and whether those facts show a “reasonable opportunity” for a defendant to access a plaintiff’s work. See id. Critically, we have held that saturation of a relevant market can create a reasonable possibility of access. See Loomis, 836 F.3d at 997. Here, Ambrosetti, Alstott, and Farrell were all members of the same part of the same liturgical music community during the relevant period, as demonstrated by their attendance at the same niche conventions. Cf. id. at 998 (no evidence that the alleged infringers were part of the local Santa Barbara music scene during a brief ten-day stay in the city). Ambrosetti does not need to show widespread dissemination among musicians generally, but only among members of this small scene. 4 To that end, he has highlighted that “Emmanuel” has been in the music library of the Archbishop of Canterbury (London) since before Farrell composed “Christ”; that the Daughters of Saint Paul, which distributed “I Will Sing,” has the broadest Catholic distribution of any publishing house in the world; that he did several commercial recordings of “Emmanuel” with various orchestras; and that he performed “Emmanuel” at hundreds of events at parishes, dioceses, and other organizations, among other assertions about radio play and the importance of “Emmanuel” within ILP’s catalogue. Ambrosetti’s declaration contains several more indicia of widespread dissemination as well. Accordingly, although “distinguishing ‘reasonable’ from ‘bare’ poses a complicated line-drawing exercise,” Nimmer on Copyright § 13D.05[B][2] (footnote omitted), we must remember that summary judgment asks only “if the evidence is such that a reasonable jury could return a verdict for the nonmoving party.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). Considering the record holistically, given the limited size of the relevant liturgical music scene

4 Additionally, although Farrell was in the United Kingdon for much of the relevant period, she did make several trips to the United States during this time, often related to her work as a composer of liturgical music. Moreover, Alstott seemingly lived and worked in the United States for most of this period, as did Limb. Therefore, to the extent that some of Ambrosetti’s evidence of widespread distribution is local to the United States, that does not snuff out his dissemination theory of access in this particular case.

AMBROSETTI V. OREGON CATHOLIC PRESS 19

during the applicable time period and the variety of ways in which Ambrosetti has shown that Farrell could have come across “Emmanuel,” even without actual sales numbers for ILP’s catalogues, we conclude that Ambrosetti has adduced sufficient facts to create a triable issue as to his dissemination theory as well. See Three Boys Music Corp., 212 F.3d at 483–84 (explaining several attenuated but plausible bases for the plaintiffs’ widespread dissemination theory). Of course, Ambrosetti need only rely on one theory of access: “showing both simply cements access.” Nimmer on Copyright § 13D.05[C]. Still, a plaintiff must also show sufficient similarity to demonstrate copying. Id. For the reasons explained below, we conclude Ambrosetti has met his burden at this stage on the issues of substantial and striking similarity as well. B. There is a genuine issue of material fact as to whether “Emmanuel” and “Christ” are substantially similar Although we disagree with the district court’s conclusions as to access, we agree with the district court that, when viewed in the light most favorable to Ambrosetti, “Christ” and “Emmanuel” are substantially similar under the extrinsic test, so Ambrosetti survives summary judgment. We apply a two-part test for substantial similarity. 5 Williams, 895 F.3d at 1119. Under the extrinsic test, we

5 “We do not . . . have a ‘well-defined standard for assessing when similarity in selection and arrangement becomes “substantial,’” largely because it is a fact-driven and context-dependent inquiry.” Hanagami v. Epic Games, Inc., 85 F.4th 931, 943 (9th Cir. 2023) (quoting Rentmeester v. Nike, Inc., 883 F.3d 1111, 1121 (9th Cir. 2018). “We have suggested generally that the ‘selection and arrangement of elements “consider[] whether two works share a similarity of ideas and expression as measured by external, objective criteria.” Swirsky v. Carey, 376 F.3d 841, 845 (9th Cir. 2004). This test “often requires analytical dissection of a work and expert testimony.” Three Boys Music Corp., 212 F.3d at 485. Such “‘[a]nalytical dissection’ requires breaking the works ‘down into their constituent elements, and comparing those elements for proof of copying as measured by “substantial similarity.”’” Swirsky, 376 F.3d at 845 (quoting Rice v. Fox Broad Co., 148 F. Supp. 2d 1029, 1051 (C.D. Cal. 2001), rev’d in part on other grounds, 330 F.3d 1170 (9th Cir. 2003)). Meanwhile, the intrinsic test “test[s] for similarity of expression from the standpoint of the ordinary reasonable observer, with no expert assistance.” Jada Toys, Inc. v. Mattel, Inc., 518 F.3d 628, 637 (9th Cir. 2008) (quoting Apple Comput., Inc. v. Microsoft Corp., 35 F.3d 1435, 1442 (9th Cir. 1994)). Courts apply only the extrinsic test at summary judgment. See Funky Films, Inc. v. Time Warner Ent. Co., 462 F.3d 1072, 1077 (9th Cir. 2006). “Because the requirement is one of substantial similarity to protected elements of the copyrighted work, it is essential to distinguish between the protected and unprotected material in a plaintiff’s work.” Swirsky, 376 F.3d at 845 (emphasis in original). “Still, ‘substantial similarity can be found in a combination of elements, even if those elements are individually unprotected.’” Williams, 895 F.3d at 1119– 20 (quoting Swirsky, 376 F.3d at 848). “Each note in a scale, for example, is not protectable, but a pattern of notes in a tune may earn copyright protection.” Metcalf v. Bochco, 294

must be similar enough that “the ordinary observer, unless he set out to detect the disparities, would be disposed to overlook them.”’” Id. (quoting Rentmeester, 883 F.3d at 1121).

AMBROSETTI V. OREGON CATHOLIC PRESS 21

F.3d 1069, 1074 (9th Cir. 2002). “We have extended copyright protection to ‘a combination of unprotectable elements . . . only if those elements are numerous enough and their selection and arrangement original enough that their combination constitutes an original work of authorship.’” Skidmore v. Led Zeppelin, 952 F.3d 1051, 1074 (9th Cir. 2020) (en banc) (quoting Satava v. Lowry, 323 F.3d 805, 811 (9th Cir. 2003)). Ambrosetti presented expert analysis by Dr. Lawrence Ferrara, who opined that “Emmanuel” and “Christ” have “significant similarities.” Dr. Ferrara ultimately compiled a list of twenty-three aggregate similarities between the works. Defendants did not include the reports of their own expert musicologist in their motion papers, purportedly because they could present the facts and argument more precisely in their own briefing and because they believed that Dr. Ferrara’s report on its own did not create triable issues of fact. To that end, Defendants argue that Dr. Ferrara’s extensive reports and testimony “do not alter the fact that” “Emmanuel” and “Christ” have markedly different pitch sequences, 6 rhythms, and metrical placement and that many of the other similarities between the works are simply commonplace building blocks within liturgical music. They also argue that the district court erred in determining that Dr.

6 “Pitch” (or scale degree) refers to how high or low a note is. Gray v. Hudson, 28 F.4th 87, 92 (9th Cir. 2022). While notes in a scale can be referred to with alphabetic names (A, B, C, etc.), as is common in music copyright cases, see, id. at 93, the parties here have opted here to refer to notes by numerical degrees. In the key of E minor, the pitch “E” is scale degree 1, the pitch F# is scale degree 2, the pitch “G” is scale degree 3, the pitch “A” is scale degree 4, the pitch “B” is scale degree 5, the pitch “C” is scale degree 6, and the pitch “D” is scale degree 7. Ferrara’s list of similarities was sufficient to meet our criteria for a selection and arrangement theory. We agree with the district court that there are genuine issues of material fact as to whether the two works are substantially similar. As we emphasized in Williams, “at summary judgment, so long as the [plaintiffs] ‘presented “indicia of a sufficient disagreement concerning the substantial similarity of [the] two works,” then the case must be submitted to a trier of fact.’” 895 F.3d at 1137 (second alteration in original) (emphasis omitted) (quoting Swirsky, 376 F.3d at 844). Defendants and Dr. Ferrara have presented principled disagreements as to whether differences in, for example, rhythm and metrical placement, make “Christ” and “Emmanuel” sufficiently similar in the context of the overall compositions. And there was “nothing inherently unsound” about Dr. Ferrara’s methodology; he clearly explained, for example, why (in his view) the differences in pitch and rhythm could not be separated from the similar melodic development of the two works. Swirsky, 376 F.3d at 846– 47. While defendants would have us break down the works into pitch sequences, rhythms, and metrical placements and analyze each element alone, doing so would be improper. See Swirsky, 376 F.3d at 847–48 (“Objective analysis of music under the extrinsic test cannot mean that a court may simply compare the numerical representations of pitch sequences and the visual representations of notes to determine that two choruses are not substantially similar, without regard to other elements of the compositions. Under that approach, expert testimony would not be required at all[.]”). While each item in Dr. Ferrara’s list of twenty-three similarities might not be protectable, together, their unique combination is. Skidmore, 952 F.3d at 1074. We cannot

AMBROSETTI V. OREGON CATHOLIC PRESS 23

separate out each piece of glass within the kaleidoscope of an overall work: rather, we must view the work as the sum of all the relevant elements together. Relatedly, critiquing the district court for relying on Dr. Ferrara’s list of aggregate similarities overlooks the substantial explanation Dr. Ferrara provided as to how those individually unprotectable elements—such as pitch sequence, rhythmic durations, and scale degrees—combine to form significantly similar works. Assuredly, under Skidmore, a plaintiff cannot simply toss together a list of common building blocks that two works purportedly share and call it a day. Id. at 1075 (“Presenting a ‘combination of unprotectable elements’ without explaining how these elements are particularly selected and arranged amounts to nothing more than trying to copyright commonplace elements.” (quoting Satava, 323 F.3d at 811–12)). But Dr. Ferrara did explain how these elements were selected and arranged in unique form, explaining the relationship among the pitch sequences, rhythmic choices, melodic development, and so forth. His concluding list of twenty- three aggregate similarities is merely one part of his overall analysis. Moreover, while the parties agree that certain elements in “Christ” and “Emmanuel”—like the key of E Minor, 3/4 meter, and the pitch sequence 5-1-2-3—are commonplace building blocks within liturgical music, those similarities make up only a minor part of the overlap between the works. And Defendants have not identified any prior works that share the same combination of pitches, rhythm, meter, and other choices that Dr. Ferrara explains makes the works very similar. See Granite Music Corp. v. United Artists Corp., 532 F.2d 718, 720 (9th Cir. 1976) (“Evidence of similar musical phrases appearing in prior works . . . demonstrates that the . . . probability of independent, coincidental production was great.”). Rather, Defendants identify several works with pitch sequences similar to “Emmanuel” and “Christ” and argue that the differences in rhythm and metrical placement make those works sound different from both “Christ” and “Emmanuel.” 7 This argument, however, merely supports Ambrosetti, given that Dr. Ferrara generally agrees but concludes that “Christ” and “Emmanuel” are similar to each other, just not those works. Defendants’ inability to point to prior art that shares all or most of the relevant similarities outside of pitch sequence further undercuts their argument. Ultimately, as we have repeatedly observed, “[s]ummary judgment is ‘not highly favored’ on questions of substantial similarity in copyright cases.” L.A. Printex, 676 F.3d at 848 (quoting Shaw v. Lindheim, 919 F.2d 1353, 1355 (9th Cir. 1990)); see also Twentieth Century-Fox Film Corp. v. MCA, Inc., 715 F.2d 1327, 1330 n.6 (9th Cir. 1983) (“Since substantial similarity is usually an extremely close question of fact, summary judgment has traditionally been disfavored in copyright litigation.”). While “no one magical combination” of factors like melody, rhythm, and pitch “will automatically substantiate a musical infringement suit,”

7 Defendants also assert—insufficiently and only in passing—that Ambrosetti has at best “thin” copyright protection. But we have emphasized that “[a] selection and arrangement copyright is not always thin.” Skidmore, 952 F.3d at 1076 n.13. That Defendants themselves identify a variety of other songs that share similar pitch sequences but that sound dissimilar because of different rhythms, metrical placements, meters, and so forth illustrates the “wide range of possible expression and broad creative choices” involved in creating the works at issue. Hanagami, 85 F.4th at 947. Accordingly, “Emmanuel” merits broad copyright protection. Id. at 947–48; see also Williams, 895 F.3d at 1120.

AMBROSETTI V. OREGON CATHOLIC PRESS 25

Ambrosetti has sufficiently shown via Dr. Ferrara’s testimony that the similarity here is “substantial.” Swirsky, 376 F.3d at 849. Accordingly, like the district court, we conclude that Ambrosetti has demonstrated “‘indicia of a sufficient disagreement concerning the substantial similarity of [the] two works’ so that the issue . . . should [be] presented to a jury.” Id. at 853 (quoting Brown Bag Software v. Symantec Corp., 960 F.2d 1465, 1472 (9th Cir. 1992)). The extrinsic test requires nothing more. See id. C. There is a genuine issue of material fact as to whether “Emmanuel” and “Christ” are strikingly similar Moreover, we conclude that the district court erred by granting summary judgment to Defendants because there is a genuine issue of material fact as to whether “Christ” and “Emmanuel” are strikingly similar. The district court had concluded that, although Ambrosetti had shown triable issues of fact as to whether the works were substantially similar, he had not shown triable issues of fact as to striking similarity. We disagree. “Two works are strikingly similar when the similarities between them are so great that they are ‘highly unlikely to have been the product of independent creation.’” Malibu Textiles, Inc. v. Label Lane Int’l, Inc., 922 F.3d 946, 953 (9th Cir. 2019) (quoting Rentmeester v. Nike, Inc., 883 F.3d 1111, 1124 (9th Cir. 2018)). Where a plaintiff can show striking similarity, he or she does not have to show access, because “access may be inferred.” Id. at 952. “‘Striking’ similarities from which such an inference may be drawn are similarities of the kind that cannot be explained, in the normal course of human events, by the possibilities of independent creation, coincidence or prior common source.” 5 Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 19D.07[C][1][d] (2025). For largely the same reasons articulated above, Ambrosetti has demonstrated genuine issues of material fact as to whether the works are strikingly similar. Dr. Ferrara’s expert reports and testimony thoroughly outline numerous similarities between “Christ” and “Emmanuel” that continue not just for a short sequence but for several musical phrases in a row. Moreover, in his testimony, Dr. Ferrara commented that, in his extensive experience, it “is rare that you had a melody of this length and this substantiality and similarity between two works.” This assertion is buoyed by the hundreds of pages of reasoned analysis that Dr. Ferrara provided. While he admitted that certain aspects of the two works were commonplace—such as the 5-1-2-3 pitch sequence that both works employ—he repeatedly asserted that the overall similarities between the two works were “not even remotely commonplace” and that he believed there was “very strong musicological evidence of copying.” While Dr. Ferrara could, of course, not opine as to any legal conclusions, see Nationwide Transp. Fin. v. Cass Info. Sys., Inc., 523 F.3d 1051, 1058 (9th Cir. 2008), 8 his overall analysis demonstrates genuine issues of material fact as to whether the two works are so similar as to preclude the possibility of independent creation. Moreover, while there are also some noted differences between the works, such as certain differences in pitch

8 Similarly, although Ambrosetti also asserts that he has sufficiently shown that the works are strikingly similar because he averred so in a declaration, we “can disregard a self-serving declaration that states only conclusions and not facts that would be admissible evidence.” Nigro v. Sears, Roebuck & Co., 784 F.3d 495, 497 (9th Cir. 2015).

AMBROSETTI V. OREGON CATHOLIC PRESS 27

sequence and rhythm, those do not necessarily preclude a finding of striking similarity. As we have previously noted, “a jury could find these to be knowing modifications, which could be evidence of willful copying.” Malibu Textiles, 922 F.3d at 953–54 (“While the trial court placed great emphasis on the minor differences between the two patterns, we feel that the very nature of these differences only tends to emphasize the extent to which the defendant has deliberately copied from the plaintiff.” (quoting Concord Fabrics, Inc. v. Marcus Bros. Textile Corp., 409 F.2d 1315, 1316 (2d Cir. 1969) (per curiam))). Here, a reasonable juror could conclude that the differences between “Christ” and “Emmanuel” mean that each composer created their work independently; he or she could also conclude that Farrell subtly but intentionally altered “Emmanuel.” That determination is for the trier of fact to make. Accordingly, while “[w]e do not have a well-defined standard for assessing when similarity in selection and arrangement becomes . . . striking,” id. at 953 (internal quotation marks omitted) (quoting Rentmeester, 883 F.3d at 1121), we conclude that the facts of this case demonstrate a genuine issue of material fact as to whether it was “highly unlikely” that Farrell composed “Christ” without having heard “Emmanuel,” Rentmeester, 883 F.3d at 1124. As one of our sister circuits stated in a similar case reversing summary judgment (in which Dr. Ferrara served as an expert for the defendants), “[i]t was not for the district court to make this factual finding where such strong competing evidence was before it.” Repp v. Webber, 132 F.3d 882, 891 (2d Cir. 1997). Instead, “[t]he issue of ‘striking similarity,’ by virtue of the supported opinions of the experts, including that of Professor Ferrara for the defendants, was shown to be a genuine issue of material fact.” Id. So too here. In conclusion, the district court did not abuse its discretion by excluding Alstott’s letters and the theory of access premised on those letters. However, even without those letters, Ambrosetti has demonstrated genuine issues of material fact as to whether the Farrell had a reasonable possibility of accessing “Christ” before composing “Emmanuel”; whether the works are substantially similar; and whether the works are strikingly similar. We therefore affirm the district court’s exclusion of the letters and the related theory of access but reverse its grant of summary judgment to Defendants. We remand for proceedings consistent with this opinion.

AFFIRMED IN PART, REVERSED IN PART, AND

REMANDED.

Each side shall bear its own costs on appeal.