Philippe Calderon v. Sixt Rent a Car, LLC

5 F.4th 1204

Opinion

[PUBLISH]

IN THE UNITED STATES COURT OF APPEALS

FOR THE ELEVENTH CIRCUIT

________________________

No. 20-10989

________________________

D.C. Docket No. 0:19-cv-62408-AHS

PHILIPPE CALDERON, on behalf of themselves and all others similarly situated, ANCIZAR MARIN, on behalf of themselves and all others similarly situated,

Plaintiffs - Appellees,

versus

SIXT RENT A CAR, LLC, Defendant - Appellant,

SIXT FRANCHISE USA, LLC, Defendant.

________________________

Appeal from the United States District Court for the Southern District of Florida

________________________

(July 14, 2021)

Before JILL PRYOR, NEWSOM, and MARCUS, Circuit Judges.

NEWSOM, Circuit Judge, delivered the opinion of the Court, in which MARCUS, Circuit Judge, joined, and JILL PRYOR, Circuit Judge, joined in Parts I and II.

JILL PRYOR, Circuit Judge, filed a concurring opinion. NEWSOM, Circuit Judge, filed a concurring opinion. NEWSOM, Circuit Judge:

A customer making an airline, hotel, or car-rental reservation on Orbitz.com agrees to a contract that includes an arbitration provision. That provision requires the customer to arbitrate disputes related to, among other things, “any services or products provided.” In this case, we must decide whether that phrase refers to services and products provided (1) by Orbitz or (2) by anyone. Reading the “any services or products provided” clause in the light of neighboring provisions and the larger contractual context—and applying a dose of common sense—we conclude that it refers only to services and products provided by Orbitz. Because the underlying dispute in our case doesn’t relate to services or products provided by Orbitz, but only to those provided by Sixt Rent A Car, a company that does business through Orbitz, we will affirm the district court’s denial of Sixt’s motion to compel arbitration.

3

I

Ancizar Marin used Orbitz.com to book a rental car from Sixt. Toward the end of his reservation process with Orbitz, Marin clicked on a big “Reserve Now” button immediately below a statement that said, “By selecting to complete this booking I acknowledge that I have read and accept the . . . Terms of Use.” The words “Terms of Use” were accompanied by a hyperlink prompting Marin to read and accept them. Marin clicked “Reserve Now,” indicating that he agreed to Orbitz’s Terms of Use.

Orbitz’s Terms of Use, which describe themselves as “constitut[ing] the entire agreement between [the customer] and Orbitz,” contain a provision that mandates arbitration of certain disputes. This case turns on the meaning of that arbitration provision—and in particular its use of the word “Claims.” The arbitration provision says that—

Any and all Claims will be resolved by binding arbitration, rather than in court . . . . This includes any Claims you assert against us, our subsidiaries, travel suppliers or any companies offering products or services through us, including Suppliers, (which are the beneficiaries of this arbitration agreement).

The arbitration provision thus applies only to capital-C “Claims,” which, importantly, the Terms of Use define as follows:

[A]ny disputes or claims relating in any way to [1] the Services, [2]

any dealings with our customer service agents, [3] any services or products provided, [4] any representations made by us, or [5] our Privacy Policy.

4

Among the five categories of activities listed in the definition, this case centers on the third—“any services or products provided.”

The Terms of Use also include several other provisions relevant to this case.

First, the Terms of Use define capital-S “Services”—the first of the five “Claim[]” categories—to mean “the Web sites, mobile applications, call center agents, and other products and services provided by Orbitz, including any Content,” and not “products or services that are provided by third parties.” Second, the Terms of Use provide that, whenever a customer asserts a “Claim[],” he must “give [Orbitz] an opportunity to resolve” it “by contacting ‘Orbitz Legal: Arbitration Claim Manager’” and then waiting 60 days before proceeding. And third, the Terms of Use explain that the “use” of “products or services that are provided by third parties, and that are available through a link from the Services . . . is subject to the terms set forth by their respective owners or operations.”

Marin had no complaints about any of his interactions with Orbitz, which by all accounts went smoothly. A few weeks after securing his reservation through Orbitz, Marin picked up his car from Sixt. When he did so, Marin signed an entirely separate agreement with Sixt—which, notably, didn’t contain an arbitration provision. Marin drove the rental car and returned it (he says) damage- free. Later, though, Sixt sent him an email alleging that he had damaged the car, followed by a collection letter seeking more than $700.

5

Marin sued Sixt in federal court on behalf of a putative class of Sixt customers. Marin alleged that Sixt breached its own contract with him and violated two state consumer-protection statutes. He didn’t sue Orbitz, nor did he complain of any wrongdoing by Orbitz or allege any violation of Orbitz’s Terms of Use by anyone. His complaint mentioned Orbitz just once, in passing.

Sixt moved to compel arbitration of Marin’s lawsuit. It didn’t invoke its own contract—there being no arbitration provision in its contract to invoke—but rather Orbitz’s Terms of Use. Sixt argued that when Marin accepted Orbitz’s Terms of Use at the reservation stage, he agreed to arbitrate actions against Sixt concerning damage fees that it had imposed.

The district court denied Sixt’s motion. The court held that Marin’s lawsuit fell outside of the scope of the arbitration provision because the suit concerned Sixt’s practices, not Orbitz’s. Alternatively, it determined that Sixt had no authority to enforce the arbitration provision because it wasn’t a third-party beneficiary of Orbitz’s Terms of Use and didn’t meet the conditions for equitable estoppel.

Sixt appealed the denial of its motion to compel arbitration. 1

1 We review questions of law in the denial of a motion to compel arbitration de novo. Lawson v. Life of the South Ins. Co., 648 F.3d 1166, 1170 (11th Cir. 2011).

6

II

A

The parties agree that Florida law governs our interpretation of Orbitz’s contract. Under Florida law, the meaning of an arbitration provision is a “matter of contractual interpretation” and thus turns on the “intent of the parties to [the] contract, as manifested in the plain language of the arbitration provision and contract itself.” Jackson v. Shakespeare Found., Inc., 108 So. 3d 587, 593 (Fla. 2013).

Accordingly, we turn to the text of the arbitration provision in Orbitz’s Terms of Use. Again, that provision says that—

Any and all Claims will be resolved by binding arbitration, rather than in court . . . . This includes any Claims you assert against us, our subsidiaries, travel suppliers or any companies offering products or services through us, including Suppliers, (which are the beneficiaries of this arbitration agreement).

For our purposes, the key term in the arbitration provision is “Claims.” The

provision requires arbitration of “[a]ny and all Claims.” So, on the one hand, if Marin’s suit doesn’t constitute a “Claim[],” then the provision doesn’t compel arbitration here. On the other hand, if Marin’s suit constitutes a “Claim[],” the provision does require arbitration here because the other preconditions are clearly satisfied—Sixt, as a company providing rental-car services through Orbitz, is

7

either (or both) a “travel supplier[]” or a “compan[y] offering products or services through” Orbitz.

Because arbitrability turns entirely on whether Marin’s suit constitutes a “Claim[],” we train our attention to the meaning of that word. The Terms of Use define “Claims” as follows:

[A]ny disputes or claims relating in any way to [1] the Services,

[2] any dealings with our customer service agents, [3] any services or products provided, [4] any representations made by us, or [5] our Privacy Policy.

Our analysis will focus on the definition’s third clause—“any services or products provided.” Sixt’s rental-car service, which gave rise to this dispute, doesn’t fall within any of the other four “Claim[]” categories: Category (1), capital-S “Services,” which (as already explained) are defined to mean Orbitz’s own services; Category (2), “dealings with our [i.e., Orbitz’s own] customer service agents”; Category (4), “representations made by us [i.e., Orbitz]”; or Category (5), “our [i.e., Orbitz’s own] Privacy Policy.” Sixt contends, though, that its rental-car service fits squarely within Category (3)—“any services or products provided.”

8

B

We disagree. Although its meaning is not perfectly pellucid or free from all doubt, we conclude that the phrase “any services or products provided” refers to services or products provided by Orbitz, not services or products by anyone. Accordingly, Sixt’s rental-car service doesn’t fall within the category of arbitrable “Claims.”

We come to that conclusion for three principal reasons.

1

First, the other items in the series that includes the phrase “any services or products provided” indicate that it refers to services or products provided by Orbitz. Florida courts recognize the familiar rule of contract interpretation that “the meaning of particular terms may be ascertained by reference to other closely associated words.” City of Homestead v. Johnson, 760 So. 2d 80, 84 (Fla. 2000). Cf. Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 195 (2012) (“When several nouns or verbs or adjectives or adverbs—any words—are associated in a context suggesting that the words have something in common, they should be assigned a permissible meaning that makes them similar.”). As just explained, the Terms of Use define “Claims” as disputes relating to five categories of activity. The other four categories all indisputably describe spheres of Orbitz’s own activities: (1) “Services,” contractually defined as

9

a range of things done “by Orbitz” and not “by third parties”; (2) “any dealings with our customer service agents,” which means Orbitz’s own agents; (4) “any representations made by us,” which means made by Orbitz; and (5) “our Privacy Policy,” which means Orbitz’s privacy policy.

Given the close spatial and conceptual relationship among Categories (1), (2), (4), and (5)—as describing spheres of Orbitz’s own activities—it would be odd if Category (3) referred to a massive and seemingly boundless set of activity—Sixt calls it “unlimited”—encompassing services or products provided by anyone. That category—especially sandwiched as it is between two Orbitz-related items on one side and two Orbitz-related items on the other—reads more naturally as referring to another sphere of Orbitz’s own activities, namely, its own “services or products.”

Second, two other provisions in Orbitz’s Terms of Use strongly indicate that the term “Claims” doesn’t refer to “services or products provided” by anyone, but rather to those provided by Orbitz. The first requires that any customer asserting a “Claim[]” must “give [Orbitz] an opportunity to resolve” it “by contracting ‘Orbitz Legal: Arbitration Claim Manager’” and then waiting 60 days before instituting further proceedings. Were we to accept Sixt’s definition of “Claims” as including disputes relating to “services or products provided” by anyone, it would mean that customers bringing lawsuits against third-parties for third-party misconduct wholly

10

unrelated to Orbitz had to route their litigation through Orbitz’s legal department. It would mean that Marin himself was obligated to route this lawsuit through Orbitz’s legal department, even though Orbitz has no discernible interest in it and no authority to resolve it. (What would Orbitz’s “Arbitration Claim Manager” do with 60 days to contemplate a claim about another company’s damage fees?) Indeed, it would presumably mean that we would have to dismiss Marin’s suit for his failure to comply with this pre-suit contractual obligation. Tellingly, Sixt hasn’t asked us to do that—likely because it seems so bizarre. But we think that would be the unavoidable consequence of Sixt’s boundless definition of the term “Claims.”

The other relevant provision states that the “use” of “products or services that are provided by third parties, and that are available through a link from the Services . . . is subject to the terms set forth by their respective owners or operations.” That provision indicates that, at least at some level, the Terms of Use conceive of the customer’s relationship with third-party services—like Sixt’s rental-car service—as being governed by the separate contracts that those third parties make, not by Orbitz’s Terms of Use.

Finally, there is common sense. Recall that Orbitz’s Terms of Use mandate arbitration of “Claims” against, among other entities, “travel suppliers” and “any companies offering products or services through” Orbitz. Were we to accept Sixt’s

11

definition of “Claims” as including disputes relating to “services or products provided” by anyone, then a staggering range of lawsuits that don’t concern Orbitz would be subject to Orbitz’s mandatory claim-processing and arbitration rules. On Sixt’s reading, every customer who signed Orbitz’s Terms of Use would have agreed to arbitrate every dispute it ever had with any “travel supplier[]” or any “compan[y] offering products or services through” Orbitz, regardless of whether the customer booked his reservations with that entity through Orbitz.

Imagine, for instance, that after signing Orbitz’s Terms of Use and renting his car from Sixt, Marin drove to some faraway city, physically walked into a hotel, and asked the front-desk clerk for a room. Upon arrival, he signed the hotel’s one-page contract, which contained no arbitration provision. Days later, the hotel evicted him because of his race, so he sued the hotel in federal court under Title II of the Civil Rights Act of 1964. Under Sixt’s interpretation, Marin’s lawsuit against the hotel would be subject to mandatory arbitration. The hotel is undoubtedly a “travel supplier” because hotel rooms are travel-related amenities. And assuming that the hotel—like most others—lists rooms on Orbitz.com, then it would also be a “compan[y] offering products or services through” Orbitz. And according to Sixt, Marin’s lawsuit would qualify as a “Claim[]” for the simple reason that it related to someone’s (i.e., the hotel’s) provision of services or products—even though Marin hadn’t used Orbitz to book the room. By

12

interpreting the phrase “services or products provided” to refer only to those provided by Orbitz, our interpretation avoids that extreme and unlikely implication. Cf. James v. Gulf Life Ins. Co., 66 So. 2d 62, 63–64 (Fla. 1953) (“The inconvenience, hardship, or absurdity of one interpretation of a contract . . . is weighty evidence that such meaning was not intended when the language is open to an interpretation which is neither absurd nor frivolous.”).2

2

In response, Sixt asserts that our interpretation renders parts of the arbitration provision meaningless. As just noted, the provision mandates arbitration of “Claims” brought not only against Orbitz, but also against third parties such as “travel suppliers” and “companies offering products or services through” Orbitz. Sixt insists that if the phrase “services or products provided” is read to refer only to Orbitz’s own services and products, then the arbitration provision’s references to third parties have no effect—because, Sixt says, there’s no such thing as a suit against a third party concerning Orbitz’s own activities.

But it’s not an empty set. We can envision circumstances in which such suits could arise, and we can understand why Orbitz might want to require arbitration of them. For instance, if a customer booked an international flight on

2 Note, as well, that at least in the case of a hotel that had listed rooms on Orbitz.com, it could, under Sixt’s theory, enforce the arbitration clause as a third-party beneficiary of Orbitz’s Terms of Use.

13

Orbitz.com and then showed up at the airport only to be told that the airline had no record of his reservation, he might well sue the airline. But his suit could relate, as well, to Orbitz’s own activities that facilitated the airline reservation—if, say, he sought redress for something that went wrong in Orbitz’s communications with the carrier. Because such a suit, although brought against a third party, would relate to Orbitz’s own activities, it would be an arbitrable “Claim[]” under our interpretation. And Orbitz would have good reason to want to maintain some degree of involvement in such a suit because it could implicate its practices, subject it to discovery requests and subpoenas, or lead to its own downstream liability. So it’s not true that the agreement’s references to third-party defendants makes no sense and has no independent bite under our interpretation.

Sixt also contests our interpretation on the ground that it creates redundancy.

We acknowledge that if, as we hold, the phrase “services or products provided” refers only to those provided by Orbitz, it basically duplicates the term “Services,” which, as already explained, is contractually defined to refer to a variety of “products and services provided by Orbitz.” But we also note that some redundancy is inevitable. Were we to accept Sixt’s definition of the phrase “services or products provided” as referring to those provided by anyone, then it would swallow and subsume the category of “Services.” After all, every “product[] and service[] provided by Orbitz” is also a “product[ or] and service[]”

14

provided by someone. The arguments from redundancy, therefore, are at worst a wash and, in any event, are hardly dispositive.3 We therefore conclude that the phase “any services or products provided” is most naturally read to refer to services and products provided by Orbitz rather than those provided by anyone. 4

3 Indeed, if one squints at the Terms of Use hard enough, other redundancies emerge. For instance, the Terms of Use define “we”—and thus presumably “us”—to include Orbitz and “its subsidiaries,” but then go on to require arbitration of disputes involving either “us” or “our subsidiaries.” They define “Content” to include “text, documents, information, data, articles, images, photographs, graphics, software, applications, video recordings, audio recordings, sounds, designs, features, and other materials,” even though several of those categories seem to completely subsume others. And they even define “Claims” in such a way that Categories (2), (4), and (5) might well duplicate Category (1), capital-S “Services.” After all, what dispute relating to “dealings with [Orbitz’s] customer service agents,” Orbitz’s “Privacy Policy,” or “representations made by [Orbitz]” wouldn’t also relate to Orbitz’s “Services,” which (once again) are defined to include all of Orbitz’s products and services? The fact is that “[s]ometimes drafters do repeat themselves and do include words that add nothing of substance, either out of a flawed sense of style or to engage in the ill-conceived but lamentably common belt-andsuspenders approach.” Scalia & Garner, supra, at 176–77. 4 We also reject Sixt’s alternative argument that, even if the phrase “services or products provided” refers to those provided by Orbitz, this lawsuit is a “Claim[]” because “Claims” need only “relat[e] in any way to” one of the five covered categories of activity. This lawsuit, Sixt reasons, “relat[es] in any way to” Orbitz’s own services or products because Marin reserved the rental car that led to the dispute through Orbitz.com. The problem is that Florida courts categorize relational language in arbitration provisions into two basic groups: “(1) provisions with language and application narrow in scope, and (2) provisions with language and application broad in scope.” Jackson, 108 So. 3d at 593; see also Seifert v. U.S. Home Corp., 750 So. 2d 633, 637–38 (Fla. 1999). Even for broad provisions, they require that the lawsuit have a “significant relationship” to the object of the preposition—which here means to Orbitz’s own services or products. Jackson, 108 So. 3d at 593. A “significant relationship” requires that the success or failure of the lawsuit depend on facts concerning the object to which the suit must relate. Id. Here, Marin’s claims against Sixt don’t have a significant relationship to Orbitz’s own services or products because Sixt’s liability for breach of contract and improper consumer practices depends in no way on any facts about Orbitz’s own services or products—only Sixt’s.

15

III

A

In an attempt to overcome that most natural reading, Sixt invokes the canon of construction that “any doubts concerning the scope of arbitrable issues should be resolved in favor of arbitration.” Moses H. Cone Mem’l Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 24–25 (1983). If the Moses H. Cone canon applied, it would no doubt bolster Sixt’s position. But we don’t think it applies.

Moses H. Cone’s strong pro-arbitration canon emanates from the Federal Arbitration Act and thus applies only to “arbitration agreement[s] within the coverage of the Act.” Id. at 24. In order to invoke the canon, then, a party must show that the FAA governs the arbitration agreement at issue. The Act governs an arbitration agreement only to the extent that it compels arbitration of “controvers[ies]” that “aris[e] out of” the “contract” containing the arbitration agreement or the “transaction” evidenced thereby. 9 U.S.C. § 2. Accordingly, the Act’s—or more accurately, Moses H. Cone’s—strong pro-arbitration canon of construction applies here only to the extent that Marin’s lawsuit against Sixt “aris[es] out of” his contract with Orbitz. 5

5 We note that Orbitz’s Terms of Use themselves state that they are “governed by the Federal Arbitration Act.” Sixt hasn’t argued that this statement affects the applicability of the Moses H. Cone canon, and we don’t think it does. We don’t disagree that the Terms of Use are governed by the FAA. But the FAA, in turn, applies only to written arbitration agreements concerning disputes that “aris[e] out of” the underlying contract—which, for reasons we will explain, this one doesn’t.

16

As we have explained in the course of interpreting similar language in an arbitration provision itself—rather than, as here, a statute governing a certain subset of arbitration provisions—the term “arising out of” is “broad, but it is not all encompassing.” Doe v. Princess Cruise Lines, Ltd., 657 F.3d 1204, 1218 (11th Cir. 2011). A dispute “does not arise out of or in connection with a contract” for the purposes of arbitration “just because the dispute would not have arisen if the contract had never existed.” Int’l Underwriters v. Triple I: Int’l Inv. Inc., 533 F.3d 1342, 1347 (11th Cir. 2008) (cleaned up). Rather, when determining whether a “dispute ‘arises out of’ . . . an underlying contract” for arbitration purposes, “we generally consider whether the dispute in question was an immediate, foreseeable result of the performance of contractual duties.” Hearn v. Comcast Cable Comm’ns, LLC, 992 F.3d 1209, 1213 (11th Cir. 2021) (cleaned up). Thus, for instance, we have held that a dispute related to the leasing of railroad cars between two companies didn’t “aris[e]” out of a contract governing the ongoing, separate supply of railroad cars between the same two companies—and therefore wasn’t arbitrable. Seaboard Coast Line R.R. Co. v. Trailer Train, 690 F.2d 1343 (11th Cir. 1982). Although the cases we draw on here interpreted variations on the phrase “controvers[ies] . . . arising out of [the] contract” as they appeared in contracts themselves, we can think of no reason why the meaning of the phrase would be broader in the provision of the FAA that determines whether the statute

17

governs such contracts. Accordingly, Moses H. Cone’s pro-arbitration canon of construction applies here only if Marin’s lawsuit against Sixt was an immediate, foreseeable result of the performance of his accepting Orbitz’s Terms of Use.

We conclude that it wasn’t. Marin’s lawsuit alleges that Sixt breached its own contract and violated two state consumer-protection statutes. The complaint doesn’t name Orbitz as a defendant, identify any wrongdoing by Orbitz, or allege any violation of Orbitz’s Terms of Use by anyone. In fact, it mentions Orbitz only once, and there only in passing. The relationship between Marin’s lawsuit against Sixt and Orbitz’s Terms of Use is even more attenuated than the relationship between the lawsuit and the related contract in Seaboard, in which we held that the lawsuit did not arise out of that contract. See 690 F.2d 1343.

Sixt has provided no authority for the proposition that Moses H. Cone’s pro-

arbitration canon applies to claims so tangentially related to the underlying contract. In fairness, the question has rarely come up. But we are persuaded by Judge O’Scannlain’s analysis of the issue in his concurring opinion in Revitch v. DIRECTV, LLC, 977 F.3d 713 (9th Cir. 2020). Although “the ‘arising out of’ language in § 2 has generated little judicial attention,” he explained there, it confines the FAA’s application to the arbitration of controversies with a sufficiently close “relationship” to the underlying contract. See id. at 722 (O’Scannlain, J., concurring). Accordingly, he reasoned that the Act didn’t apply

18

to the dispute before him, which pitted a customer against a satellite-television company that was a distant affiliate of the wireless service company with whom the customer agreed to arbitrate disputes. See id. at 722–24; see also David Horton, Infinite Arbitration Clauses, 168 U. Pa. L. Rev. 633, 643, 678–80 (2020) (explaining that the FAA’s “arising out of” language means that it applies only when the dispute has a sufficient “nexus” to the underlying contract and that the drafters of the Act consciously adopted this narrowing language); Recent Case, Revitch v. DIRECTV, LLC, 977 F.3d 713 (9th Cir. 2020), 134 Harv. L. Rev. 2871, 2878 (2021) (observing that “Judge O’Scannlain’s revival of the FAA’s ‘arising out of’ limitation may prove invaluable” because, among other reasons, it is “ground[ed] in statutory language” and “can coexist with the Court’s zealously enforced command to resolve scope ambiguities in favor of arbitration, as it merely recognizes an outer limit to that scope”). For similar reasons, we hold that Marin’s suit against Sixt does not “aris[e] out of” his contract with Orbitz within the meaning of § 2—and, accordingly, that Moses H. Cone’s pro-arbitration canon of construction does not apply.

B

Sixt suggests as a backup that even if Moses H. Cone doesn’t apply, Florida law embodies the same pro-arbitration interpretive rule. But while Florida law

19

recognizes that arbitration agreements are valid, enforceable, and irrevocable, Florida’s courts haven’t gone full-on Moses H. Cone.

To be sure, Florida courts “try to resolve . . . ambiguit[ies] in . . . arbitration provision[s] in favor of arbitration.” Jackson, 108 So. 3d at 593. But they typically both derive this rule from federal law and state it in weaker terms than the Supreme Court expressed it in Moses H. Cone. See, e.g., Seifert, 750 So. 2d at 641 (construing “ambiguity” in an arbitration provision against arbitration); Citigroup, Inc. v. Amodio, 894 So. 2d 296, 298 (Fla. Dist. Ct. App. 2005) (acknowledging the “federal policy” favoring arbitration but explaining that a court must “mak[e] its initial construction of an agreement to arbitrate” under “general state contract law”). 6 Florida courts also emphasize that whether a dispute is subject to arbitration depends primarily on “the plain language of the arbitration provision and contract itself.” Jackson, 108 So. 3d at 593. And in similar contexts, they have expressed reluctance to employ substantive canons of construction in the face of mere doubts over meaning. To the contrary, Florida courts have said that they will first “resort to the ordinary rules of construction,” and only after exhausting those determine whether “genuine inconsistency, uncertainty, or ambiguity in meaning remains.” Excelsior Ins. Co. v. Pomona Park Bar & Package Store, 369

6 We have found one Florida court that derived the strong pro-arbitration rule of construction from Florida law, see Ronbeck Const. Co. v. Savanna Club Corp., 592 So. 2d 344, 346 (Fla. Dist. Ct. App. 1992), but its approach doesn’t appear to reflect the general state of the law.

20

So. 2d 938, 942 (Fla. 1979). We hold, therefore, that Sixt hasn’t demonstrated that Florida law independently requires us to resolve “any doubts” about the scope of an arbitration provision in favor of arbitrability.

Accordingly, we adhere to what we believe to be the best, most ordinary, most sensible interpretation of the contract, under which the arbitration provision does not encompass Marin’s lawsuit against Sixt.7

IV

To summarize: The phrase “any services or products provided” in Orbitz’s Terms of Use doesn’t refer to services or products provided by anyone, but rather only to those provided by Orbitz itself. It follows that this lawsuit doesn’t qualify as a “Claim[]” under the Terms of Use and therefore isn’t subject to the provision’s arbitration mandate. The Moses H. Cone canon of construction doesn’t change the outcome because Marin’s dispute with Sixt didn’t “aris[e] out of” Orbitz’s Terms of Use. Accordingly, we AFFIRM the district court’s denial of Sixt’s motion to compel arbitration.

7 Because we hold that Marin’s arbitration agreement with Orbitz didn’t compel him to arbitrate this lawsuit against Sixt, we needn’t reach the alternative argument that Marin’s later contract with Sixt extinguished any previous agreements that he had made with Sixt and therefore invalidated the arbitration agreement as it applied to their relationship. We also have no occasion to address the district court’s alternative holding that Sixt had no authority to enforce the arbitration provision because it wasn’t a third-party beneficiary and didn’t meet the conditions for equitable estoppel.

21

JILL PRYOR, Circuit Judge, concurring:

I concur in Parts I and II of the majority opinion. I do not join Part III because I find it unnecessary to decide whether the Moses H. Cone canon applies in this case. For the reasons explained by the majority opinion, I have no doubt that Marin’s claims against Sixt do not fall within the arbitration provision of the Orbitz agreement.

22

NEWSOM, Circuit Judge, concurring:

I suppose it goes without saying that I agree with the Court’s conclusions (1)

that Orbitz’s Terms of Use didn’t require arbitration of Ancizar Marin’s lawsuit against Sixt Rent A Car and (2) that Moses H. Cone’s pro-arbitration canon of construction has no application here because Marin’s suit didn’t “aris[e] out of” the underlying contract within the meaning of the Federal Arbitration Act. I write separately to voice my skepticism of the Moses H. Cone rule, which strikes me as among the most dubious of the so-called “substantive” interpretive canons.

In Moses H. Cone Memorial Hospital v. Mercury Construction Corp., the Supreme Court held that contractual arbitration provisions should be broadly construed—in particular, that “any doubts concerning the scope of arbitrable issues should be resolved in favor of arbitration.” 460 U.S. 1, 24–25 (1983) (emphasis added). The Court purported to derive this strong presumption from the Federal Arbitration Act. But the Act’s pertinent text suggests nothing of the sort—it says only that certain written arbitration agreements “shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” 9 U.S.C. § 2. So far as I can tell, the Moses H. Cone canon is just made up. We should rethink it.

23

I

First, some history. Before Congress enacted the Federal Arbitration Act in 1925, American law made arbitration contracts functionally unenforceable. Courts permitted contracting parties to freely revoke their agreements to arbitrate, and they refused to stay lawsuits concerning issues that the parties had expressly agreed to resolve through arbitration. See Tobey v. Cnty. of Bristol, 23 F. Cas. 1313, 1321 (C.C.D. Mass. 1845) (No. 14,065) (Story, J.); Hiro N. Aragaki, Arbitration’s Suspect Status, 159 U. Pa. L. Rev. 1233, 1250–53 (2011). On one account, this early judicial hostility to arbitration stemmed from the English practice of paying judges based on the number of cases that they adjudicated. See Aragaki, supra, at 1252. A more charitable account attributed it to defects in the old arbitration system—among them that arbitrators “possess[ed] no authority whatsoever, even to administer an oath, or to compel the attendance of witnesses,” and were “not ordinarily well enough acquainted with the principles of law or equity, to administer either effectually.” Tobey, 23 F. Cas. at 1321.

Over time, many came to see the judiciary’s antipathy for arbitration as “anomalous and unjust.” Berkovitz v. Arbib & Houlberg, 130 N.E. 288, 292 (N.Y. 1921) (Cardozo, J.). Congress enacted the Federal Arbitration Act to overcome “the judiciary’s long-standing refusal” to enforce arbitration agreements and, in particular, to place such agreements “upon the same footing as other contracts.”

24

Volt Info. Scis., Inc. v. Bd. of Trustees of Leland Stanford Junior Univ., 489 U.S. 468, 478 (1989) (quotation marks and citations omitted). The Act thus aimed to “make arbitration agreements as enforceable as other contracts, but not more so.” Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U.S. 395, 404 n.12 (1967).

Section 2 of the FAA operationalized this new, equal-footing approach to interpreting and enforcing arbitration agreements. It states, in full, that—

A written provision in any maritime transaction or a contract evidencing a transaction involving commerce to settle by arbitration a controversy thereafter arising out of such contract or transaction, or the refusal to perform the whole or any part thereof, or an agreement in writing to submit to arbitration an existing controversy arising out of such a contract, transaction, or refusal, shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.

9 U.S.C. § 2, Pub. L. 68–401, § 2, 43 Stat. 883, 883 (1925). Stripping away the scope-defining language, then, the core of § 2 guarantees that written arbitration agreements shall be “valid, irrevocable, and enforceable.” Id. Simple enough, right?

B

Wrong. Decades after Congress enacted the FAA, lower federal courts started reading it to do much more than just place arbitration agreements on the “same footing” as other contracts. Instead, they interpreted it to beget a potent “canon of construction” requiring that “every doubt . . . be resolved in favor of arbitration.” Hanes Corp. v. Millard, 531 F.2d 585, 598 (D.C. Cir. 1976); see also

25

Metro Indus. Painting Corp. v. Terminal Constr. Co., 287 F.2d 382, 385 (2d Cir. 1961). In devising this pro-arbitration canon, the courts proceeded along two essentially parallel tracks, which I’ll explore in turn.

1

First, a surprising number of courts pretty nakedly transplanted the canon from an altogether different statute. Section 203(d) of the Taft–Hartley Act of 1947 provides—explicitly—that when a labor dispute arises out of a collective- bargaining agreement, the “desirable method for settlement” is “[f]inal adjustment by a method agreed upon by the parties.” 29 U.S.C. § 173(d), Pub. L. 80–101, § 203(d), 61 Stat. 136, 154 (1947). The Supreme Court thus sensibly said that when faced with a case arising out of a collective-bargaining agreement, it would favor the parties’ agreed-upon method of dispute resolution, including arbitration. See United Steelworkers of Am. v. Am. Mfg. Co., 363 U.S. 564, 566 (1960). More particularly—and more forcefully—the Court announced that qualifying labor disputes should be arbitrated in accordance with the parties’ contract “unless it may be said with positive assurance that the arbitration clause is not susceptible of an interpretation that covers the asserted dispute.” United Steelworkers of Am. v. Warrior & Gulf Nav. Co., 363 U.S. 574, 582–83 (1960). “Doubts,” the Court said, “should be resolved in favor of coverage”—and therefore in favor of arbitration. Id. at 583. Section 203(d) thus established a bona fide substantive preference for

26

arbitration—but only with respect to labor disputes arising out of collective- bargaining agreements.

At least initially, § 203(d)’s narrow scope was well understood and faithfully observed. The Supreme Court made clear that the provision’s preference for the parties’ own agreed-upon dispute-resolution method—and thus, where they chose it, for arbitration—derived from labor law and accordingly governed only certain labor disputes arising out of collective-bargaining agreements. The Court explained that, as a common-sense matter, “arbitration of labor disputes has quite different functions from arbitration under an ordinary commercial agreement” because “arbitration of labor disputes under collective bargaining agreements is part and parcel of the collective bargaining process itself” and advances all parties’ “common goal of uninterrupted production under the agreement.” Id. at 578, 582. And—more to the point—the Court emphasized § 203(d)’s language that (within its particular ambit) expressed a clear preference for agreed-upon arbitration. Am. Mfg. Co., 363 U.S. at 566. 1 Underscoring the provision’s narrow scope, the

1 The Supreme Court has continued to recognize that, in labor cases, the pro-arbitration rule of interpretation derives from § 203(d). See Wright v. Universal Mar. Serv. Corp., 525 U.S. 70, 78 (1998) (“That presumption [of arbitrability], however, does not extend beyond the reach of the principal rationale that justifies it, which is that arbitrators are in a better position than courts to interpret the terms of a CBA. This rationale finds support in the very text of the [Taft–Hartley Act], which announces that ‘[f]inal adjustment by a method agreed upon by the parties is declared to be the desirable method for settlement of grievance disputes arising over the application or interpretation of an existing collective-bargaining agreement.’” (emphasis omitted)).

27

Supreme Court emphasized that the approach “evinced by courts toward arbitration of commercial agreements has no place here.” Warrior & Gulf Nav. Co., 363 U.S. at 578. So, too, it seemed to follow, the approach applicable “here”—i.e., in labor- law cases—had no place in cases concerning the arbitration of non-labor-related commercial disputes.

Somewhere along the way, though, lower federal courts decided that § 203(d)’s pro-arbitration rule should govern the interpretation of ordinary, non- labor-related contracts. In run-of-the-mill commercial disputes, they began citing labor cases for the proposition that a reviewing court should construe “every doubt” about the scope of an arbitration clause in favor of arbitration. Dickinson v. Heinold Sec., Inc., 661 F.2d 638, 643 (7th Cir. 1981); see also, e.g., Becker Autoradio U.S.A., Inc. v. Becker Autoradiowerk GmbH, 585 F.2d 39, 44 (3d Cir. 1978); Wick v. Atl. Marine, Inc., 605 F.2d 166, 168 (5th Cir. 1979). The courts didn’t even acknowledge that they were changing lanes, let alone justify their basis for doing so. They thereby broadened the scope of the § 203(d)-based pro- arbitration canon and unmoored it from the text of the Taft-Hartley Act whence it had originated. Soon after, they grafted the canon onto the FAA, completing the

28

transplant. See, e.g., McAllister Bros. v. A & S Transp. Co., 621 F.2d 519, 522 (2d Cir. 1980).2

2

Second, and separately, many courts almost simultaneously developed the same pro-arbitration canon through an aggressively purposivist reading of the FAA. They reasoned that whatever its text says, “the policy of the Federal Arbitration Act is to encourage arbitration and to relieve congestion in the courts.” Seaboard Coast Line R. Co. v. Nat’l Rail Passenger Corp., 554 F.2d 657, 660 (5th Cir. 1977) (emphasis added). They accordingly imposed an interpretive presumption in favor of arbitration in order to “implement[]” what they saw as “important policies,” like alleviating “court congestion” and avoiding “delay[s] in dispute resolution.” Hanes Corp. v. Millard, 531 F.2d 585, 598 (D.C. Cir. 1976); see also Gen. Guar. Ins. Co. v. New Orleans Gen. Agency, Inc., 427 F.2d 924, 928 (5th Cir. 1970) (“Any doubts as to the construction of the Act ought to be resolved in line with its liberal policy of promoting arbitration . . . to help ease the current congestion of court calendars.”). Based on an unvarnished policy preference for arbitration over litigation, these courts all announced some variation of the canon

2 I’m not the first to note this strange development. See Margaret L. Moses, Statutory Misconstruction: How the Supreme Court Created a Federal Arbitration Law Never Enacted by Congress, 34 Fla. St. U. L. Rev. 99, 123 (2006) (noting that the federal judiciary “may have indiscriminately superimposed on the FAA the national labor policy favoring collective bargaining agreements”).

29

that all “doubts” about the meaning of contractual arbitration provisions “are to be resolved in favor of arbitration.” Galt v. Libbey-Owens-Ford Glass Co., 376 F.2d 711, 714 (7th Cir. 1967).

B

Enter Moses H. Cone, in which the Supreme Court endorsed the pro-

arbitration canon in the strongest possible terms. “Section 2 [of the Federal Arbitration Act],” it explained, “is a congressional declaration of a liberal federal policy favoring arbitration agreements.” Moses H. Cone Mem’l Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 24 (1983). Thus, the Court held, the FAA “establishes that, as a matter of federal law, any doubts concerning the scope of arbitrable issues should be resolved in favor of arbitration, whether the problem at hand is the construction of the contract language itself or . . . [a] defense to arbitrability.” Id. at 24–25 (emphasis added); see also id. at 25 n.31 (relying on lower-court FAA decisions that derived the pro-arbitration canon from labor law).

The Court has since doubled (and tripled, and quadrupled . . .) down on the Moses H. Cone canon. The FAA, it has said, “reflects an emphatic federal policy in favor of arbitral dispute resolution.” KPMG LLP v. Cocchi, 565 U.S. 18, 21 (2011) (per curiam) (quotation marks omitted). Accordingly, the Court has repeatedly incanted, “any doubts” concerning the scope of contractual arbitration clauses “should be resolved in favor of arbitration.” Mitsubishi Motors Corp. v.

30

Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 626 (1985) (quoting Moses H. Cone, 460 U.S. at 24–25); see also, e.g., Volt Info. Scis., Inc., 489 U.S. at 475–76.

Lower federal courts and state supreme courts have since run with Moses H.

Cone, understanding it to “require[ them] to construe arbitration clauses as broadly as possible,” Collins & Aikman Prod. Co. v. Bldg. Sys., Inc., 58 F.3d 16, 19 (2d Cir. 1995) (quotation marks omitted), and to “resolve any . . . uncertainties in favor of arbitration,” Old Republic Ins. Co. v. Lanier, 644 So. 2d 1258, 1260 (Ala. 1994). Some have gone so far as to hold that Moses H. Cone’s pro-arbitration canon “trumps” other interpretive rules—like, for instance, the old contra proferentem principle that ambiguities in a contract should be construed against the drafter. Kristian v. Comcast Corp., 446 F.3d 25, 62 (1st Cir. 2006); see also Hudson v. Conagra Poultry Co., 484 F.3d 496, 503 (8th Cir. 2007) (“[T]his common-law rule of construction would favor the [plaintiffs’] interpretation, but it cannot overcome the statutory rule of construction favoring arbitration as embodied by the Federal Arbitration Act.”).3 Thanks to the Moses H. Cone canon, then, the resolution of ambiguity in a contract can hinge solely on whether the disputed provision concerns arbitration. Imagine that an employer drafts a contract with two

3 See generally Richard Frankel, The Arbitration Clause As Super Contract, 91 Wash. U.L. Rev. 531, 556 (2014) (stating that most courts have “read Moses H. Cone’s policy favoring arbitration to trump the doctrine of contra proferentem and to require ambiguities to be interpreted in favor of arbitration, even if it is the drafting party that seeks to enforce the arbitration clause”). But see Paul Revere Variable Annuity Ins. Co. v. Kirschhofer, 226 F.3d 15, 25 (1st Cir. 2000) (applying the common-law rule anyway).

31

provisions, Provision X about benefits and Provision Y about arbitration. If X gets litigated, the reviewing court would construe any murky language against the employer. But if the employer seeks arbitration, then the court would construe any doubt about Y in his favor. Weird—Congress put arbitration agreements on equal footing with other contracts, only to have the courts put them on a pedestal.

C

Three characteristics of the Moses H. Cone canon—which, as I said, has now thoroughly permeated American law and practice—warrant particular attention.

1

First, and perhaps most obviously, the Moses H. Cone rule is a “substantive”

interpretive canon, in that it directs courts to depart from a contract’s most natural interpretation in favor of—and to further—a policy preference for arbitration. For the uninitiated, canons of interpretation are conventionally divided between the “semantic” and the “substantive”—or some variation on that dichotomy. See, e.g., Brett M. Kavanaugh, Fixing Statutory Interpretation, 129 Harv. L. Rev. 2118, 2121 (2016) (book review); Caleb Nelson, What Is Textualism?, 91 Va. L. Rev. 347, 394 n.140 (2005) (“descriptive” vs. “normative”); William Baude & Stephen E. Sachs, The Law of Interpretation, 130 Harv. L. Rev. 1079, 1123 (2017) (“linguistic” vs. “legal”). Semantic canons do exactly what their name implies—

32

they provide “the general rules by which we understand the English language.” Kavanaugh, supra, at 2145. They help courts ascertain the ordinary meaning of a legal text—such as by reminding us that that “[t]he expression of one thing implies the exclusion of others,” that “and combines items while or creates alternatives,” and that when words “are associated in a context suggesting that [they] have something in common, they should be assigned a permissible meaning that makes them similar.” Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 107, 116, 195 (2012).

Substantive canons are an altogether different kettle of fish. They have little (if anything) to do with a text’s ordinary meaning, but rather instruct courts to favor certain substantive policies in interpreting that text. See John F. Manning, Textualism and the Equity of the Statute, 101 Colum. L. Rev. 1, 96 (2001); Nelson, supra, at 394. They express the law’s supposed preferences when certain close interpretive calls arise. Thus, the contra proferentem canon expresses a preference that an ambiguity in a contract provision be interpreted against its drafter. See Lamps Plus, Inc. v. Varela, 139 S. Ct. 1407, 1417 (2019). Likewise, the rule of lenity expresses a preference that an ambiguity in a criminal statute be interpreted in the defendant’s favor. See Yates v. United States, 574 U.S. 528, 547–48 (2015).

Substantive canons have been the subject of debate among textualists.

Some, including then-Professor Barrett, have written that “[s]ubstantive canons”—

33

at least those that operate as more than mere tiebreakers—“are in significant tension with textualism . . . insofar as their application can require a judge to adopt something other than the most textually plausible meaning of a statute.” Amy Coney Barrett, Substantive Canons and Faithful Agency, 90 B.U. L. Rev. 109, 123–24 (2010). While “[t]extualism, in its purest form, begins and ends with what the text says and fairly implies,” Scalia & Garner, supra, at 16, substantive canons “often require judges to depart from a [text’s] most natural interpretation,” Barrett, supra, at 121.

Others have suggested that a substantive canon’s validity depends not so much on whether it diverts courts from the most textually plausible reading, but rather on its legal pedigree. See Baude & Sachs, supra, at 1122–24. On this account, substantive canons may require courts to depart from the most natural interpretation of a legal text, but only when the common law, a statute, or a constitution commands that departure. So, for instance, the common-law rule of will construction “mak[ing] it difficult to disinherit one’s children,” although it “do[es]n’t necessarily track actual linguistic usage,” remains “binding on the parties simply because [it is] the law.” See id. at 1094–95; see also, e.g., Restatement (Second) of Contracts § 201(2) (1981) (providing default rules when parties attach different meanings to a contract term); 1 U.S.C. § 1 (“In determining the meaning of any Act of Congress, unless the context indicates otherwise . . . the

34

words ‘person’ and ‘whoever’ include corporations, companies, associations, firms, partnerships, societies, and joint stock companies[.]”) ); U.S. Const., amend. XI (“The Judicial power of the United States shall not be construed to extend to [certain suits against States].”). Importantly, though, even the defenders of substantive canons reject them to the extent that judges just make them up. See Baude & Sachs, supra, at 1138–39.

Accordingly, whichever of these two camps has it right—or however much daylight really exists between them—substantive canons not firmly grounded in the written or common law are, in my view, on extremely thin ice.

2

Which brings me to the second and third—and to me more problematic—

characteristics of Moses H. Cone’s pro-arbitration canon: It is both especially potent and especially made up. As for the former, the Moses H. Cone rule is, among substantive canons, about as strong as they come. Often, a substantive canon kicks in only within a relatively narrow range of uncertainty. Contra proferentem, for instance, “applies ‘only as a last resort’ when the meaning of a provision remains ambiguous after exhausting the ordinary methods of interpretation.” Varela, 139 S. Ct. at 1417 (quoting 3 Arthur Corbin, Contracts § 559 (1960)). So too, “[t]he rule of lenity applies only if, after seizing everything from which aid can be derived, . . . we can make no more than a guess as to what

35

Congress intended.” Reno v. Koray, 515 U.S. 50, 65 (1995) (quotation marks and citations omitted). Other substantive canons operate more broadly, by applying, for instance, whenever a substantively favored interpretation is “plausible.” See Barrett, supra, at 117–18 (listing as examples the rules that statutes should be interpreted to avoid constitutional doubt and to comport with international law). But the Moses H. Cone canon operates even more broadly than that, applying whenever “any doubts” exist about the arbitration provision’s scope. 460 U.S. at 24 (emphasis added). It thus diverts courts from the best reading of the text at the first hint of uncertainty, and thereby works a massive alteration of written contracts in America.

Moses H. Cone’s pro-arbitration canon is also, so far as I can tell, a judicial invention. Unlike, say, the rule that “if Congress intends to alter the ‘usual constitutional balance between the States and the Federal Government,’ it must make its intention to do so ‘unmistakably clear in the language of the statute,’” Will v. Michigan Dep’t of State Police, 491 U.S. 58, 65 (1989) (quoting Atascadero State Hosp. v. Scanlon, 473 U.S. 234, 242 (1985)), which has at least a plausible foothold in the Tenth and Eleventh Amendments, the Moses H. Cone canon has no basis in the positive law. In pertinent part, the FAA mandates only that certain written arbitration agreements “shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation

36

of any contract.” It says nothing about giving arbitration agreements favored status or treatment. Indeed, I’m not aware of a single straight-faced attempt to defend Moses H. Cone as an interpretation of the FAA’s text.

Nor, so far as I’m aware, does the canon have any basis in any history that might inform the FAA’s meaning. It emerged decades after the Act became law and (as already explained) as a result of the overextension of an unrelated labor statute and a bald appeal to policy considerations. Early courts applying the FAA didn’t invoke a pro-arbitration presumption. Instead, they interpreted the Act to mean pretty much exactly what it says—that arbitration clauses are on the same footing as other contractual provisions. See, e.g., The Anaconda v. Am. Sugar Refin. Co., 322 U.S. 42, 44 (1944) (“Within the spheres of its operation . . . the Arbitration Act rendered a written [arbitration agreement] specifically enforceable. Thereby Congress overturned the existing rule that performance of such agreements could not be compelled by resort to courts of equity or admiralty.”).

III

I tend to think we might be better off without the Moses H. Cone canon, which can put a court in the untenable position of having to ignore the best evidence of a contractual provision’s meaning. Even when an arbitration agreement is most properly read not to require arbitration of a particular dispute, a court must compel arbitration if it has “any doubts” about the agreement’s scope.

37

Rather than employing the traditional tools of textual interpretation, courts are made to forgo meaningful interpretation in the name of, among other things, reducing court congestion.

I agree with Justice Kagan that, to one extent or another, “we’re all textualists now.”4 Because that’s so, we shouldn’t be atextually interpreting a statute in a manner that, in turn, requires us to atextually interpret contractual provisions. Rather, we should interpret the FAA to do exactly what it says, which is to make arbitration agreements valid and enforceable on the same terms as other contracts—and therefore to require us to interpret arbitration agreements, first and foremost, according to their plain text.

4 Justice Elena Kagan, The Scalia Lecture: A Dialogue with Justice Kagan on the Reading of Statutes at 8:28 (Nov. 17, 2015), http://today.law.harvard.edu/in-scalia-lecture-kagan-discussesstatutory -interpretation [http://perma.cc/3BCF-FEFR].