Appeal from the United States District Court for the Northern District of Georgia D.C. Docket No. 1:20-cv-00695-SDG

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

NEWSOM, Circuit Judge:

When Jay Gould was the CEO of Interface, Inc. he allegedly engaged in misconduct at an annual sales meeting. As a result, Interface’s board of directors terminated his employment for cause. Gould sued, claiming that Interface had breached his employmentPage 2 agreement by firing him. In particular, Gould contended that the board made its cause determination in bad faith, thereby overstepping the qualified discretion it enjoyed under the agreement. The district court granted summary judgment to Interface, holding that the contract gave the board absolute discretion to determine cause and that, in any event, Interface hadn’t acted in bad faith.

On appeal, Gould presents a different theory: The employment contract, he now says, gave Interface no discretion to determine the existence of cause. The question at the heart of this appeal is whether Gould’s new no-discretion theory is an “issue” of the sort that is subject to forfeiture or a subsidiary “argument” of the sort that isn't. After careful review , we hold that Gould's new theory is more the former than the latter and, accordingly , that he has forfeited it by failing to raise it below . And because Gould has presented nothing else in support of his breach-of-contract claim, we affirm the district court's judgment.

I

Jay Gould was the CEO of Interface, Inc., a carpet manufacturer. According to the company, Gould got drunk at its annual sales meeting and repeatedly called one of his employees a “fucking bitch.” Final R. & R. 12, Dkt. No. 195. And, the company says, that wasn't Gould's first breach of etiquette. Just a year earlier, Interface had sanctioned and counseled him for engaging in alcoholfueled sexual harassment. Any further infractions, the company Page 3had warned, would result in discipline “up to and including termination for Cause.' Id. at 7.

Following the latter episode, Interface hired King & Spalding LLP to conduct an investigation. The firm's analysis corroborated the allegations against Gould, and Interface's board of directors voted unanimously to fire him for cause. Under Gould's employment agreement, termination with cause entitled him to far less compensation-potentially $10 million less-than termination without cause.

Interface's authority to fire Gould is covered by Section 5(c) of his employment contract. That provision provides that, “[s]ubject to the terms of Section 5(d) below, the Company may terminate Executive's employment hereunder, in its sole discretion, whether with or without Cause, at any time upon written notice to Executive.” Am. and Restated Employment and Change in Control Agreement 10, § 5(c), Dkt. No. 4-1. Section 5(d), in turn, establishes the protocols for firing an employee without cause, but it doesn't specifically address termination with cause. Id. § 5(d). A separate provision of the agreement, Section 5(a)(i), defines the term “Cause” as follows:

(A) Executive's fraud, dishonesty , gross negligence, or willful misconduct with respect to business affairs of the Company (including its subsidiaries and affiliated companies), (B) Executive’s refusal or repeated failure to follow the established lawful policies of the Company applicable to persons occupying the same or similar positions, (C) Executive’s material breach of

Page 4this Agreement, or (D) Executive’s conviction of a felony or other crime involving moral turpitude. A termination of Executive for Cause based on clause (A), (B) or (C) of the preceding sentence shall take effect 30 days after Executive receives from the Company written notice of intent to terminate and the Company’s description of the alleged Cause, unless Executive shall, during such 30-day period, remedy the events or circumstances constituting Cause; provided, however, such termination shall take effect immediately upon the giving of written notice of termination for Cause under any of such clauses if the Company shall have determined in good faith that such events or circumstances are not remediable (which determination shall be stated in such notice).

Id. at 8, § 5(a)(i).

Section 5(a)(i)'s final sentence-and in particular the concluding proviso—makes clear that Interface possesses discretion, qualified by a duty of 'good faith,' to assess whether the events giving rise to “Cause” are “remediable.” Less clear—but central to the dispute here—is what sort of discretion Interface enjoys under the agreement to determine whether cause exists in the first place. The contract’s termination-related provisions yield three possibilities: First, the agreement might give Interface absolute discretion to determine the existence of cause, in which case, under the governing Georgia law, a reviewing court can’t evaluate the merits of the company’s determination at all. See Automatic Sprinkler Corp. of Am. v. Anderson , 243 Ga. 867, 868 (1979) (holding that, when a Page 5contract “leave[s] decisions absolutely to the uncontrolled discretion of one of the parties[,] . . . the issue of good faith is irrelevant”). Second, the agreement might give the company qualified discretion—subject to a duty of good faith—to determine the existence of cause, in which case the court must decide whether Interface’s determination was made in “bad faith.” Id. Or third, the agreement might give Interface no discretion to determine the existence of cause, in which case the court must decide whether the company's determination was 'in fact erroneous.' Id.

Gould sued Interface, alleging that it had breached his employment contract based on the “manner of its termination and [its] fail[ure] to pay him under the terms” of the agreement. Am. Compl. ¶ 81, Dkt. No. 4. Interface moved for summary judgment on the ground that it enjoyed absolute discretion to determine the existence of cause for Gould's firing. In the alternative, it argued that even if its discretion was qualified by a duty of good faith, it had discharged its duty by relying on the results of King & Spalding's investigation. In opposition to Interface's motion, Gould seemed to accept that the company had qualified discretion to determine the existence of cause, but he argued that it had failed to perform in good faith both because King & Spalding’s investigation was a 'sham' and because it didn't make 'any good faith findings of violations . . . of the 'Cause' provision.' Opp'n to Mot. for Summ. J. 37, Dkt. No. 174-1.

Page 6In his report and recommendation, the magistrate judge concluded that Interface was entitled to summary judgment for two independent reasons. First, he concluded that the company had absolute discretion to terminate Gould with cause and, therefore, that its cause determination wasn’t subject to a duty of good faith. Second, he concluded that even if Interface had only qualified discretion to determine the existence of cause, Gould had failed to create a genuine issue of material fact regarding whether the company had acted in bad faith in making that determination. Both parties filed objections to the report and recommendation, but the district court adopted it in its entirety .

Gould filed a motion for reconsideration in which, for the first time, he took a more aggressive position: Under the employment contract, he contended, Interface had no discretion whatsoever-as opposed to qualified discretion, subject to a good-faith limitation—to terminate him with cause. The district court denied relief, in part because Gould's motion relied on what the court called an “argument” that he hadn’t previously raised. Order Den. Mot. for Recons. 5-7, Dkt. No. 252. For purposes of reconsideration, the court ruled, he had “waived” that “argument” by raising it too late. Id.

Page 7Gould appealed the district court’s summary-judgment decision.¹ Before us, he reprises his contention that Interface had no discretion to determine that cause existed to fire him.

The main issue before us, as before the district court on reconsideration, is whether Gould has forfeited the theory that he now makes the centerpiece of his case—namely, that Interface had no discretion to determine the existence of cause.²

To reset the stage, Section 5(c) of Gould's employment agreement states that “[s]ubject to the terms of Section 5(d) below, the Company may terminate Executive’s employment hereunder, in its sole discretion, whether with or without Cause, at any time upon written notice to Executive.” Am. and Restated Employment and Change in Control Agreement 10, § 5(c), Dkt. No 4-1. SectionPage 8 5(d) lays out the protocols for termination without cause, but it doesn’t squarely address termination with cause. Id. § 5(d). And finally , the term 'Cause' is defined in Section 5(a)(i), as-

(A) Executive's fraud, dishonesty , gross negligence, or willful misconduct with respect to business affairs of the Company (including its subsidiaries and affiliated companies), (B) Executive’s refusal or repeated failure to follow the established lawful policies of the Company applicable to persons occupying the same or similar positions, (C) Executive’s material breach of this Agreement, or (D) Executive’s conviction of a felony or other crime involving moral turpitude.

Id. at 8, § 5(a)(i).

As already noted, we can imagine—and at various points in this litigation the parties have advanced—three interpretations of these provisions: First, the agreement gave Interface absolute discretion to determine the existence of cause; second, the agreement gave the company qualified discretion-subject to a duty of good faith—to determine cause; and third, the agreement gave it no discretion to determine cause.

In the district court-at least prior to filing his motion for reconsideration-Gould seemed to acknowledge that Interface possessed qualified discretion to determine whether cause existed to terminate him, subject to a duty of good faith. Accordingly , the entirety of his argument was that the company had acted in bad faith in finding cause. So, for instance, in his amended complaint, Gould alleged that Interface breached the employment agreement Page 9by “relying upon a sham investigation as a pretext for termination with cause.” Am. Compl. ¶ 27, Dkt. No. 4. Likewise, in his response to the company’s motion for summary judgment, Gould contended that King & Spalding’s investigation was a “sham” and that Interface didn't make 'any good faith findings of violations . . . of the ‘Cause’ provision.” Opp’n to Mot. for Summ. J. 37, Dkt. No. 174-1. And in his objections to the magistrate judge’s R&R, Gould criticized the judge (1) for interpreting the employment agreement in a way that “remove[d] any duty of [Interface] to act in good faith when making a determination of termination for ‘Cause’” and (2) for finding that 'the unduly influenced, incomplete King & Spalding investigation—upon which [Interface] contends to have relied [] in making their 'good faith' determination of termination for ‘Cause’—was in good faith.” Pl.’s Objs. to the R. & R. 32, 33, Dkt. No. 198.

As already explained, in adopting the magistrate judge's R&R, the district court held (1) that Interface had absolute discretion to terminate Gould with cause and (2) that even if the company had only qualified discretion to determine the existence of cause, Gould hadn’t created a genuine issue of material fact regarding whether it had acted in bad faith in doing so. As the magistrate judge explained—and by adopting the R&R, the district court concluded—“in terminating an employee for cause, [Interface] was expressly given sole discretion.” Final R. & R. 53, Dkt. No. 195. Accordingly , 'there is no implied duty of good faith as to the for-cause termination decision, and any claim of a breach of that implied duty must fail.' Id . 'Moreover,' and in any event, 'even ifPage 10 [Interface]'s authority to terminate [Gould] was qualified rather than absolute, . . . there is nothing to suggest that [the company]’s decision to rely on the King & Spalding investigation was arbitrary or capricious, that it was based on improper pecuniary motive, that it was predicated on dishonesty or illegality , or that there was a total absence of any factual evidence to support it.' Id. at 54.

Beginning with his motion for reconsideration, and again before us, Gould presented a new and different theory-namely, that Interface had no discretion to decide whether cause existed to terminate him. In his motion for reconsideration, Gould argued that the issue wasn’t whether Interface had found cause in good faith but, rather, “whether Interface’s determination of cause ‘was in fact erroneous.’” Pl.’s Mot. for Recons. 19, Dkt. No. 223 (quoting Automatic Sprinkler , 243 Ga. at 868). And in his brief to us, he reiterated that position: '[R]ead reasonably , nothing in Section 5(c) suggests that Interface has the discretion to decide for itselfwhether conclusively or qualified by a duty of good faith-whether cause exists.” Br. of Appellant 33. “Because Interface has no discretion under the agreement to determine the existence of cause,” Gould contended, “this case is one in which the issue . . . is whether Interface’s determination of cause ‘was in fact erroneous.’” Id. at 35-36 (quoting Automatic Sprinkler , 243 Ga. at 868).

The question is whether Gould’s new “no discretion” theory is properly before us or whether, instead, he lost the right to press that theory by failing to properly present it to the district court.

Page 11As an initial matter, we reiterate the distinction between “waiver” and “forfeiture.” Like the district court, the parties have framed the debate around whether Gould “waived” his no-discretion theory . See, e.g. , Br. of Appellee 33 ('Mr. Gould waived below his argument that Interface lacks discretion to make a good-faith determination of Cause.”); Order Den. Mot. for Recons. 5, Dkt. No. 252 (“Gould waived the argument that he now raises.”).

That's not quite right. We're not casting aspersions'waiver' and 'forfeiture' are often conflated and confused. But to be clear, as the Supreme Court has said, “[t]he terms waiver and forfeiture—though often used interchangeably by jurists and litigants-are not synonymous.' Hamer v. Neighborhood Hous. Servs. of Chi. , 583 U.S. 17, 20 n.1 (2017). Waiver is the 'intentional relinquishment or abandonment of a known right.” Id . Forfeiture, by contrast, 'is the failure to make the timely assertion of a right.' Id .; accord, e.g. , United States v . Campbell , 26 F .4th 860, 872 (11th Cir. 2022) (en banc). The distinguishing feature, then, is intentionality—waiver is purposeful; forfeiture is inadvertent.

No one has suggested that Gould meant to withhold (or jettison) his no-discretion theory in the district court. So whatever else may be true, it seems clear to us that he didn’t waive that theory. Descriptively , it seems more accurate to say that his failure to present it to the district court was inadvertent. The question, therefore, is whether he has forfeited his right to pursue his no-discretion theory on appeal.

Which leads to the second—and in many respects more vexing—question: Is Gould’s no-discretion theory even subject to forfeiture? Answering that question requires us to explore the distinction between “issues” (or “positions” or “claims”), on the one hand, and subsidiary “arguments,” on the other. That distinction is an evergreen issue in appellate law and practice, and it lurks just beneath the surface in a lot of cases.

In many respects, the Supreme Court’s decision in Yee v . City of Escondido , 503 U.S. 519 (1992), is the modern fountainhead of the issue-argument distinction. In the California state courts, the Yees, owners of a mobile-home park, had contended that a local rentcontrol ordinance effected a physical taking of their property in violation of the Fifth Amendment. They had asserted in their complaint that 'the rent control law has had the effect of depriving [them] of all use and occupancy of their real property and granting to the tenants of mobilehomes presently in [t]he [p]ark, as well as the successors in interest of such tenants, the right to physically permanently occupy and use the[ir] real property .' Yee , 503 U.S. at 525 (alteration adopted) (emphasis added). In the Supreme Court, the Yees 'attempt[ed] to challenge the ordinance on two additional grounds”: In particular, they argued that it constituted (1) “a denial of substantive due process” and (2) “a regulatory taking.” Id. at 532–33.

Page 13The Supreme Court’s treatment of the Yees’ new grounds is illuminating. The Court declined to consider their substantivedue-process claim because they hadn’t “raise[d] a substantive due process claim in the state courts, and no state court ha[d] addressed such a claim.” Id. at 533. The Yees neither “include[d] a due process claim in their complaint[,] [n]or did [they] raise a due process claim in the [California] Court of Appeal.” Id. “It was not until their petition for review in the California Supreme Court that [the Yees] finally raised a substantive due process claim.' Id.

But the Supreme Court saw the Yees’ regulatory-taking theory in a different light. The Court acknowledged that it was 'unclear whether petitioners made this argument below: Portions of their complaint and briefing can be read either to argue a regulatory taking or to support their physical taking argument.” Id. at 534. Even so, the Court explained, the Yees’ “regulatory taking argument st[ood] in a posture different from their substantive due process claim.” Id. In particular, the Court emphasized that the Yees’ physical- and regulatory-taking theories were “not separate claims ' but, 'rather, separate arguments in support of a single claim-that the ordinance effects an unconstitutional taking.' Id. at 535. And because the Yees “raised a taking claim in the state courts,” they “could have formulated any argument they liked in support of that claim.'³ Id.

Page 14Following Yee 's lead, this Court has begun to flesh out its own issue-argument jurisprudence. So, for instance, we have held that “[p]arties can most assuredly waive [or forfeit] positions and issues on appeal, but not individual arguments .” Secretary, U.S. Dep’t of Labor v . Preston , 873 F .3d 877, 883 n.5 (11th Cir. 2017) (emphasis added) (citing Yee , 503 U.S. at 534). Or, put slightly differently , we've said that '[i]f an issue is 'properly presented, a party can make any argument in support of that issue; parties are not limited to the precise arguments they made below.'' In re Home Depot Inc. , 931 F.3d 1065, 1086 (11th Cir. 2019) (alteration adopted) (emphasis added) (quoting Yee , 503 U.S. at 534).⁴

In fleshing out the general principle that '[l]itigants can waive or forfeit positions or issues through their litigation conduct Page 15in the district court but not authorities or arguments,” this Court recently observed that “a party cannot usually argue that a legal text should be read to mean something different on appeal than what it argued for below.” ECB , 113 F .4th at 1320. By contrast, a party can cite new interpretive rules and authorities to bolster its previously presented theory that a text has a particular meaning. Id. at 1321.

This makes good sense. Courts’ interpretations of legal instruments-whether public or private-depend on the premise that the words and phrases used therein have discernible meanings. See, e.g. , Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 56 (2012) ('The words of a governing text are of paramount concern, and what they convey, in their context, is what the text means.”). From that premise, it logically follows that a litigant who advances a particular reading of a word or phrase may not then put forward an entirely new interpretation on appeal. Put simply , having committed in the district court to the view that a written legal instrument means A, a party cannot thereafter urge the appellate court to conclude that the same instrument-comprising the same terms-means B.⁵

Page 16To be sure, a litigant can bring new interpretive principles to bear on appeal. A party who argues in district court, for instance, that an instrument’s “plain language” means A, can fortify that position on appeal by invoking traditional interpretive canons— e.g. , ejusdem generis, the rule against superfluity , etc. See ECB , 113 F .4th at 1321 (“[A] party on appeal can always cite a new authority—such as the canons of construction—in favor of reading a legal text to mean what the party advocated for below.”). What he may not do is repudiate his position that the instrument means A in favor of a contrary position—that, in fact, it means B.

That distinction—between allowing a litigant to strengthen an existing interpretive position and prohibiting him from advancing an altogether new one—follows not only from linguistic considerations but also practical ones. When a party sharpens his position by presenting new contentions and authorities, he improves “the quality and depth of argument . . . on appeal.” Preston , 873 F.3d at 883 n.5. By contrast, when a litigant changes his position entirely , it muddies the litigation waters. Ours is an adversarial Page 17system; we trust that “the parties know what is best for them,” and we hold them “responsible for advancing the facts and arguments entitling them to relief.” Greenlaw v. United States , 554 U.S. 237, 244 (2008) (citing Castro v. United States , 540 U.S. 375, 386 (2003) (Scalia, J., concurring in part and concurring in the judgment)). Given that foundational premise, it makes little sense to allow a party to fundamentally change his understanding of a legal instrument— i.e. , to contend in the district court that one interpretation of a text is 'best for' him, only to insist on appeal, to the contrary , that an entirely different meaning is 'best for' him. Id . So when a party attempts to do so, it sows confusion and dulls the adversarial process that we trust to produce reliable results.

To see these rules in action, we can look to our recent decision in ECB . There, the parties disputed the meaning of a key provision of an insurance policy . 113 F .4th at 1316. The insured contended that a modifier at the end of a sentence applied only to the immediately preceding term; the insurer insisted, by contrast, that it modified all the terms in the list. Id. On a motion for reconsideration, the insured cited-for the first time-two semantic canons in support of its interpretation. Id. at 1317. The district court denied the motion on the ground that the insured had forfeited its canon-based arguments by not making them before its Rule 59 motion. Id. We disagreed. Observing that the insured 'ha[d] consistently argued' for a certain meaning of the insurance policy , we concluded “[t]he canons merely provide additional authority to support that position.” Id. at 1321. So the insured hadn't forfeited itsPage 18 reliance on the canons by failing to invoke them earlier; they were simply new arguments in support of a preexisting position. Id .

In sum, Yee and this Court's precedents establish that parties can forfeit positions or issues that they fail to properly raise at the district court, but they can’t forfeit particular arguments or authorities. And when a party contends on appeal that a legal text means something different from what it posited below, it crosses over to the “issue” side of the issue-argument divide.

Applying Yee , ECB , and related decisions here is straightforward: Gould impermissibly sought to present a new position (or issue) on appeal when, having contended in the district court that his employment agreement meant one thing, he reversed course and contended on appeal that it meant something else. Let us explain.

As we’ve said, the terms of Gould’s employment contract might plausibly be read to mean any of three different things: (1) that Interface had absolute discretion to determine whether cause existed to fire him; (2) that the company had qualified discretion, subject to a duty of good faith, to make that determination; or (3) that it had no discretion at all to do so. In the district court, Gould adopted Interpretation No. 2: He acknowledged that his employment agreement—a “legal text”—“should be read to mean” that Interface had qualified discretion to determine whether cause Page 19existed for his termination, ECB , 113 F .4th at 1320, but contended that the company had (in effect) abused that discretion by making its determination in bad faith. See supra at 8-9. By contrast, in his motion for reconsideration, and before us, Gould advocates for Interpretation No. 3. Contrary to his prior view, he claims that, in fact, the agreement “should be read to mean,” ECB , 113 F .4th at 1320, that the company had no discretion to determine the existence of cause in the first instance. See supra at 10. So Gould isn't fortifying a 'consistent[]' position with new arguments, ECB , 113 F.4th at 1320; he’s insisting that his contract means something else altogether. That, he may not do.

Our conclusion is buttressed by two additional considerations: In support of his new no-discretion theory , Gould asks us to consider (1) a slew of new facts and (2) an entirely new category of law.

With respect to the former, Gould's shift-from a qualifieddiscretion-subject-to-a-good-faith-limitation theory to a no-discretion theory-alters the factual landscape entirely . In the district court, Gould's position turned centrally on whether Interface's board acted in bad faith—an inquiry rooted in the facts about the particulars of King & Spalding’s investigation and the board’s reliance on it. See Opp’n to Mot. for Summ. J. 17–21, Dkt. No. 174-1 (extensively citing the factual record to argue that King & Spalding’s investigation was motivated by animus). But once Gould shifted to a no-discretion theory , all those facts ceased to matter.

Page 20Interface’s motives, methods, and investigative process are irrelevant if the contract simply gave it no discretion to determine the existence of cause in the first place. Gould's switch jettisons the factual predicates that underlie the district-court proceedings and substitutes an altogether different inquiry in their place-and thereby suggests, we think, that we're in altogether different territory.

So too, Gould invokes before us a different category of law than he appealed to below. In the district court, Gould's qualifieddiscretion theory relied on Georgia law concerning the implied covenant of good faith and fair dealing and cases addressing whether employers had acted in “bad faith” in exercising contractual discretion. See Opp’n to Mot. for Summ. J. 34, Dkt. No. 174-1 (quoting Shelnutt v. Mayor , 776 S.E.2d 650, 657 (Ga. Ct. App. 2015), for the meaning of “bad faith” under Georgia contract law). The body of precedent relevant to that claim focuses on standards for bad faith-what counts as arbitrary , dishonest, or capricious conduct. See Shelnutt , 776 S.E.2d at 657. After Gould pivoted to his nodiscretion theory , however, that body of precedent became irrelevant. Instead, Gould’s new no-discretion theory requires only an examination of the principles of contract interpretation governing the allocation of decisionmaking authority in private agreements. See Br. of Appellant 30-33 (citing Georgia caselaw for the proposition that contracts must be given their plain and ordinary meaning and arguing that, on such a reading, the employment agreement provides Interface with no discretion to determine cause). By abandoning the bad-faith doctrinal line and substituting a plain-meaningPage 21 line in its stead, Gould crossed into a wholly different legal framework.

* * *

Regrettably , there's no clear, algorithmic formula for determining whether a newly raised theory is a forfeitable issue or a nonforfeitable argument—at least, not one that we’ve been able to divine. But we can say this much with confidence: If a party advances an altogether new interpretation of a legal text on appeal— different from the one he embraced in the district court-he impermissibly raises a new issue. And he seals his fate when, as here, he urges the appellate court to consider a universe of facts and a category of law different from those he put before the district court.

In light of these considerations, we hold that Gould's no-discretion theory is a new issue, rather than a new argument. And because Gould failed to raise that theory-qua-issue before his motion for reconsideration, we hold that he has forfeited it. In extraordinary circumstances, we have the authority to resurrect forfeited issues. See United States v. Campbell , 26 F .4th 860, 872 (11th Cir. 2022) (en banc). But, here, Gould hasn’t argued that any such circumstances exist. Nor could he: Gould had myriad opportunities to take the position that he now advances on appeal, but he failed to do so.

Because Gould has forfeited his position that Interface had no discretion to determine cause, only one claim remains that wePage 22 must address on the merits: his claim that his conduct didn’t meet the contractual definition of cause.

But as we have explained, the district court held that Interface was entitled to summary judgment for two independent reasons: (1) that Interface’s decision to terminate Gould with cause was “entirely within the discretion of the corporation” and therefore not subject to a duty of good faith, and (2) that even if the company had only qualified discretion to terminate Gould for cause, he had failed to create a material issue of fact as to whether it terminated him in bad faith. Gould's remaining claim goes to Reason No. 2 but not Reason No. 1. So, even without evaluating the merits, we can conclude that, taken alone, Gould’s remaining claim doesn't require reversal.

For the foregoing reasons, we AFFIRM the district court's judgment.